article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
in recent years the influence of special and general relativity on quantum information processing has begun to be investigated @xcite .
staying just within the realm of special relativity , one of the natural questions is how lorentz transformations affect the properties of both massive and massless particle states .
the basic approach is through wigner s little group machinery @xcite as the poincar group is non - compact and thus without any finite - dimensional unitary representations @xcite . in this context
, the entanglement properties of bipartite states @xcite composed of massive as well as massless particles have been extensively analyzed . the fundamental impact on entanglement
when the owners of both entangled subsystems are lorentz transformed @xcite has been recognized since , for physically plausible states , entanglement depends on the properties of the frame where it is measured @xcite . in this paper
, we go back to single - qubit transformations and discuss quantum cloning from the relativistic point of view .
we study how the requirement of lorentz covariance affects the optimality of a cloning protocol .
quantum cloning has come a long way since the discovery of the no - cloning theorem @xcite and one can find an extensive variety of cloners in two recent review articles @xcite .
lorentz covariance means that the particular cloning map must be equally effective irrespective of how any input state is rotated or boosted .
more precisely , choosing the fidelity between an input and output state as a figure of merit to measure the quality of the clones , we demand that its value be maximal and independent on the input qubit .
the additional requirement of maximality provides an optimal cloner . as a striking example of
how the relativistic covariance constraint modifies the optimality of the fidelity results , we investigate the state dependent @xmath0 cloner of , generally non - orthogonal , qubits presented in ref .
@xcite . in the relativistic domain
it is necessary to distinguish between particular particle states for which the effect of the little group generally differ .
photons polarization defines the logical basis of qubits and it is a natural choice due to their use in quantum communication protocols such as bb84 .
there are relatively few previous studies devoted to relativistic effects in either classical or quantum channels . for the classical ones we highlight ref .
@xcite where the channel capacity between two moving observers is studied . for quantum channels , there has been a recent growing interest on quantum information processing in black holes @xcite as well as on how the unruh effect assists quantum state encryption @xcite .
as usual , a standard momentum light - like 4-vector @xmath2 with @xmath3 and @xmath4 is chosen .
we can transform this vector into an arbitrary light - like 4-vector @xmath5 by a standard lorentz transformation .
the most general little group element ( stabilizer subgroup which leaves @xmath6 invariant ) is @xmath7 @xcite and consists of rotations and/or translations in a plane , a group which is isomorphic to the euclidean group ( @xmath8 ) .
the corresponding hilbert space is spanned by vectors with two indices since , together with the angular momentum @xmath9 , the translation operator of the poincar group @xmath10 yields the complete set of commuting operators . in
any given reference frame , a rotation around the direction defined by the standard vector induces a phase on the corresponding state in hilbert space : @xmath11 where for simplicity we took @xmath12 and @xmath13 is the component of the angular momentum in the direction of @xmath14 ( helicity ) . as is well known @xcite
, massless particles have only integer or semi - integer @xmath13 values and for photons @xmath15 holds . any other state
is obtained from the reference state @xmath16 by applying a standard lorentz transformation .
thus , if @xmath17 is obtained from @xmath14 by a rotation with longitudinal angle @xmath18 and azimuthal angle @xmath19 , then @xmath20 transforming this state vector by an arbitrary lorentz transformation @xmath21 we arrive at the unitary representation , which turns out to be one - dimensional @xmath22 . here
@xmath23 is an angle of the rotation dependent on the lorentz transformation @xmath21 and the initial 4-vector @xmath24 ( the explicit form of @xmath25 can be found in @xcite ) .
then , we have @xmath26 we suppose that a wave packet is prepared in a state @xmath27 where @xmath28 is a lorentz - invariant measure and @xmath29 is a normalized weight function @xmath30 that describes the superposition of modes with different frequencies but a common direction of propagation @xmath24 .
this selection is made to avoid problems coming from diffraction effects occurring for a general wave packet so that one can not simply define a polarization matrix , then to keep the validity of the sharp inequality we have to limit the distribution of lorentz boosts .
this restriction does not affect the phase distribution discussed here . ] .
let us examine the action of an arbitrary lorentz transformation on @xmath31 .
we know @xcite that the phase angle does not depend on the magnitude of @xmath24 but just on its direction .
so making the transformation @xmath32 the phase @xmath33 is common for the whole wave packet . considering the choice of our wave packet and also the discussion in @xcite , after the lorentz transformation and tracing over the momenta degree of freedom ,
we get @xmath34 where @xmath35 , @xmath36 and @xmath37 .
the helicity basis is the logical basis @xmath38 for our qubits ( we thus do not use the lorentz invariant logical basis composed of two physical photons proposed in @xcite - the task is to clone an unknown single - photon state ) .
for the rest of the article , we assume the following spacetime arrangement . in her reference frame
, alice prepares a state which travels in the @xmath24-direction .
although this direction is well - defined by the outgoing state , for a subject in another inertial reference frame who receives the state ( bob ) it is not sufficient information .
the reason is that there is the whole group of transformations ( rotations around the @xmath24-direction ) which leaves the given light - like vector intact .
this is exactly the rotational part of wigner s little group responsible for inducing the wigner phase @xmath25 in ( [ unitary_repre ] ) and both angles ( rotation and wigner phase ) coincide @xcite .
we consider bob s rotation to be completely unknown and uniformly distributed .
let us proceed to analyze how the state - dependent cloning setup investigated by bru _ et al . _
@xcite is affected if relativistic covariance is incorporated .
first , let us remember the original problem and later we formulate how relativistic covariance enters the game
. from now on @xmath39 and @xmath40 denote the pauli @xmath41 and @xmath42 matrices .
the original problem solved in @xcite is , in some sense , an opposite extreme compared to the universal cloner @xcite , where all possible pure states are distributed according to the haar measure . here , for a fixed @xmath43 , one of just two real states @xmath44
is prepared at random .
the authors solved the problem assuming several reasonable invariance constraints .
the output states were symmetric with respect to the bit flip @xmath45 and were also permutationally invariant . for later comparison with our results , the fidelity function obtained in ref .
@xcite is shown in fig .
[ fig_clone_pc ] .
the angles of the qubit parametrization are re - scaled to conform the parametrization used here .
now we evaluate ( and later generalize ) the same setup when the relativistic covariance condition is imposed .
( [ wavepacket_transf ] ) describes the effect of lorentz transformation on the photonic states that are considered here . on the bloch sphere
, this transformation becomes @xmath46 .
since our requirement is that the actual angle of the rotation is unknown and uniformly distributed so are the states on the bloch sphere .
hence , in addition to the symmetries described in the previous paragraph , we require the invariance of the output with respect to the operator @xmath47 .
the invariance reflects the ignorance of the rotation angle that induces the phase angle @xmath25 in eq .
( [ wavepacket_transf ] ) .
let us pause here and describe the physical situation .
we suppose that alice s covariant operation is @xmath45 ( plus some additional operations which we wo nt mention again ) which is combined with another covariant operation @xmath47 induced by the wigner phase @xmath48 ( that is , alice sends one of two possible states which could be transformed by bob s rotation ) so we need to compare the action of @xmath47 and @xmath49 .
the order is important because the operators do not commute .
it can be easily shown that for single qubits , @xmath50 what will prove to be very useful for later calculations .
if we want to go beyond the setup studied in @xcite and suppose that alice may prepare a general pure qubit in the form @xmath51 we find that @xmath47 and @xmath49 ( our covariant operations ) have a curious behavior since when they are applied to @xmath52 these transformations appear in general as two asymmetric oriented arcs on opposite hemispheres ( parametrized by @xmath25 ) . to get a symmetric relativistic transformation we have to assume a different covariant operation , namely @xmath53 where @xmath54=u\varrho u^{-1}$ ] is the conjugation operation so we are in an adjoint representation of a group whose members are @xmath55 @xcite .
@xmath56 is the transposition of the density matrix in the standard ( logical ) basis @xmath57 ( because of this transformation we traveled into the adjoint representation ) .
the reason for incorporating @xmath56 becomes evident when we compare the action of @xmath58 and @xmath53 ( our new covariant couple ) on @xmath52 . in this case
, we will make use of the following identity ( see proof in appendix ) @xmath59 note that @xmath60=0 $ ] .
the motivation for introducing the identity is purely computational ( just as for commutator ( [ covn1 ] ) ) but the physical interpretation is interesting . since @xmath61=0 $ ] holds the order of the covariance operations does not matter .
thus , in the next we will investigate both relativistic covariance effects , i.e. when covariance is required with respect to @xmath62 for real states @xmath63 , and @xmath53 for general states @xmath64 . except where really necessary , we will omit the symbol @xmath65 for the conjugation operation to avoid the excessive notation but we have to remember that we keep working in the adjoint representation .
let us rephrase the invariance requirements from the previous paragraph in an appropriate formalism .
the jamiokowski isomorphism @xcite between positive operators and cp maps @xcite is a traditional tool for the calculation of optimal and group covariant completely positive ( cp ) maps .
one appreciates the representation even more by realizing that an implementation of the mentioned transposition operation is particularly easy .
let @xmath66 be a cp map , then the corresponding positive operator @xmath67 is related by @xmath68\right)r^{(n)}_\mathcal{m}\right],\ ] ] with @xmath69 denoting the above discussed alternatives ( without and with the transposition , respectively ) and @xmath70 .
the expression @xmath71\equiv\varrho^t_{in}$ ] stands for the transposition of the density matrix @xmath72 .
it is important to stress that the case @xmath73 must not be in a contradiction with the definition of the isomorphism ( @xmath74 ) .
consequently , the net effect is that we require @xmath75 to be invariant with respect to the transposition of @xmath72 . if @xmath66 is a cloning cp map then using eq .
( [ covn1 ] ) for @xmath74 and eq .
( [ covn2 ] ) for @xmath73 we can first start by requiring covariance with respect to @xmath76 . then the covariance conditions in both representations ( standard and jamiokowski , respectively ) read [ covcond ] @xmath77=0\label{covcondn1}\\ \mathcal{m}^{(2)}(\varrho)=(p_{\vartheta_w}\otimes p_{\vartheta_w})^\dagger\mathcal{m}\left(p^*_{\vartheta_w}\varrho^t p_{\vartheta_w}^t\right)(p_{\vartheta_w}\otimes p_{\vartheta_w } ) \rightleftharpoons [ r^{(2)}_\mathcal{m},p_{\vartheta_w}\otimes p_{\vartheta_w}\otimes p_{\vartheta_w}]=0.\label{covcondn2}\end{gathered}\ ] ] note that @xmath78 .
let us explain the use of the jamiokowski isomorphism .
for @xmath74 , utilizing eq .
( [ covn1 ] ) we apply the covariance condition coming from the structure of the phase operator @xmath79 .
we get the basic structure of the jamiokowski operator @xmath80 and we apply the bit - flip and the output state symmetry covariance conditions .
similarly for @xmath73 , the strategy is to summon the rhs of eq .
( [ covn2 ] ) to find how the covariance condition coming from @xmath47 defines the basic structure of @xmath75 . then , in addition to the previously mentioned covariant conditions , we require the covariance regarding the transposition of an input state .
this is the reason why on the lhs of eq .
( [ covcondn2 ] ) there is @xmath81=p^*_{\vartheta_w}\varrho^t p_{\vartheta_w}^t$ ] . finally , we calculate single - copy fidelities of the cloned state for both cases .
the operator @xmath80 is just a unitary modification of @xmath75 as seen from eqs .
( [ covcond ] ) so it is sufficient to analyze the structure of the case @xmath73 and for @xmath74 to subsequently modify the operator by @xmath82
it is a simple permutation of basis states .
one of the schur lemmas gives us the structure of the positive operator @xmath75 .
it is a sum of the isomorphisms between all equivalent irreducible representations , which in the case of @xmath83 are all one - dimensional and are distinguished by the character values @xmath84 with @xmath85 .
more specifically , @xmath86 is composed of four irreducible representations .
two of them are one dimensional ( spanned by @xmath87 ) and two are three dimensional @xmath88 where @xmath89 is a decimal record of the 3-qubit basis .
taking into account the above discussed additional symmetries of @xmath90 the number of independent parameters gets limited and we arrive to the following form of @xmath75 @xmath91,\end{aligned}\ ] ] where @xmath92 ( for @xmath93 ) are coefficients of the isomorphisms @xmath94 and @xmath95,c_{12b}=\im[c_{12}]$ ] .
two additional conditions come from the trace - preserving constraint @xmath96=\openone\rightarrow c_{00}+c_{11}+c_{22}+c_{33}=1 $ ] ( common for both @xmath97 ) and , of course , from the positivity condition @xmath98 .
since we are cloning a pure qubit , our figure of merit to be maximized is the single copy fidelity between the input states @xmath99 ( for @xmath74 ) and @xmath100 ( for @xmath73 ) and the target states of the same form [ fidelity ] @xmath101\right ) r^{(1)}_\mathcal{m}\right]\label{fidelityn1}\\ f^{(2)}={\mathop{{\mathrm{tr } } _ { } } } \left[\left({| \psi\rangle\!\langle \psi |}_{gen}\otimes\openone\otimes { | \psi\rangle\!\langle \psi |}_{gen}\right ) r^{(2)}_\mathcal{m}\right]\label{fidelityn2}.\end{gathered}\ ] ] observe that in eq .
( [ fidelityn2 ] ) the transposition operator was additionally applied .
[ [ case - n1-covariance - w.r.t .- p_vartheta_w - and - p_vartheta_wsigma_x ] ] * case @xmath74 * ( covariance w.r.t . @xmath47 and @xmath49 ) + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + if we apply the covariant operations on an arbitrary real @xmath102 then for different values of the wigner phase @xmath48 we generate two symmetric trajectories on the opposite hemispheres of the bloch sphere .
reformulating the search for the fidelity as a semidefinite program using the sedumi solver @xcite in the yalmip environment @xcite the number of parameters is reduced and @xmath80 can be diagonalized .
this leads to the full analytical derivation of the fidelity function ( [ fidelityn1 ] ) as a function of the input state @xmath102 @xmath103.\ ] ] the function is depicted in fig .
[ fig_clone_pc ] and we notice several interesting things . obviously , the fidelity is lower than the original state dependent fidelity .
we observe that the minimum moved from @xmath104 @xcite to the angle @xmath105 but more interesting point is that the fidelity attains @xmath106 .
this value is reserved for the @xmath0 universal symmetric cloner @xcite , i.e. the cloning map covariant with respect to the action of @xmath107 ( or , equivalently , to the cloning of all mutually unbiased states of the bloch sphere @xcite ) .
such a low value for a kind of phase - covariant cloner we are investigating may be surprising . for @xmath108
we recover the result from @xcite where @xmath109 .
this is expected because the bit flip ( one of our additional conditions ) is unnecessary on the equator ( due to the presence of @xmath47 ) .
[ [ case - n2-covariance - w.r.t .- p_vartheta_w - and - p_vartheta_wvargammacircsigma_x ] ] * case @xmath73 * ( covariance w.r.t . @xmath47 and @xmath110 ) + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + using methods similar to those in the previous paragraph we arrive with the help of eq .
( [ fidelityn2 ] ) and @xmath52 at the following form of the fidelity function @xmath111\biggr\}},\ ] ] which is independent on the input state phase @xmath112 .
this result is no less interesting and the function is again depicted in fig .
[ fig_clone_pc ] .
the minimum angle is common with the previous case but the corresponding fidelity drops to @xmath113 .
this low value can be justified if we realize what kind of operation corresponds to @xmath73 .
we combine two impossible operations , quantum cloning and finding the universal - not operation , into what is together known as the anti - cloning operation @xcite .
this combined requirement is apparently stronger than the universal ( i.e. @xmath107 ) covariance and the reason for the low fidelity values is that the map @xmath75 must be of the same form for both @xmath114 and @xmath115 as a result of eq . ( [ jamiolk ] ) .
could this result tell us something about , for instance , the security of quantum key distribution ( qkd ) ? looking at the most studied protocol bb84
@xcite ( of course , implemented by the polarization encoding which is preferred for a free - space communication for which the relativistic effects may be very relevant ) we see that four qubits equidistantly distributed on the meridian are used . for @xmath74 , they form the @xmath116plane of the bloch sphere ( ` real meridian ' ) and for @xmath73 it is an arbitrary grand circle intersecting the north and south pole ( ` complex meridian ' ) . from the viewpoint of an eavesdropper without the knowledge of the wigner phase @xmath25 and decided to clone the quantum states to get some information , we can now demonstrate that not all quadruples are equally good . if the states @xmath117 are used for the qkd purposes then by inserting @xmath118 into eqs .
( [ fidelanaln1 ] ) and ( [ fidelanaln2 ] ) , we get @xmath119 and @xmath120 . on the other hand , using the quadruple @xmath121 we get the fidelity @xmath122 because by phase - rotating the quadruple states ( that is , applying @xmath47 ) we pass the mutually unbiased states of the bloch sphere .
we see that @xmath123 corresponds to the fact that for @xmath73 the eavesdropper has less information about the input state .
another interesting question is how the relativistic covariance affects the optimality of the universal cloner .
here the situation is different . in the analysis above we combined two covariant operations ( @xmath124 and @xmath47 ) which are not generally subsets of each other . on the other hand , as we saw , every wigner rotation is a @xmath125 covariant rotation and since @xmath126 we may conclude that the optimality of the universal cloner will remain unchanged .
pictorially , it corresponds to the situation where alice sends a completely unknown photon ( @xmath107 covariance ) to bob who , in addition , does not know how the whole bloch sphere rigidly rotates ( his rotation with respect to alice ) .
however , this is again a kind of @xmath107 rotation .
in conclusion , we investigated the role played by the requirement of relativistic covariance in the problem of the optimality of one of the most prominent forbidden quantum - mechanical process as quantum cloning . observing that the effect of wigner s little group can be translated into the language of so - called phase - covariant processes we studied how the effectiveness of the cloning process becomes modified .
particularly , we considered an observer in a different reference frame with no knowledge of the parameters of the reference frame where the state designated for cloning was produced .
here we focused on the class of state - dependent cloners where the effect is especially appreciable .
first , as a direct application of the relativistic considerations on the cloning setup studied in @xcite where one of two real states is prepared in one inertial frame and cloned in another inertial frame whose transformation properties regarding the first one are completely unknown .
second , we went beyond this setup and supposed that in the first frame two general pure qubits related by a common action of the pauli @xmath127 matrix and the density matrix transposition operator might be prepared .
again , we wanted such a state to be cloned in another inertial frame without knowledge of which state was actually sent and how it was relativistically transformed .
in both cases , we brought analytical expressions for local fidelities of the output states asking the fidelity to be maximal and optimal .
one of the intriguing results is that in the second case the fidelity drops even below the universal cloner limit .
the reason is that we combined the mentioned cloning invariance conditions with another forbidden process - finding the orthogonal complement of an unknown state .
note that even without the relativistic context we generalized the previous research on the phase - covariant cloning maps and at the same time we studied optimal covariant processes considering covariance operations which do not commute .
as an example of the consequences for communication security issues we have shown that for an eavesdropper determined to get some information by cloning a bb84 quadruple of states , not all possibilities are equally good and some provide him with more information .
let @xmath128 be invertible linear transformations .
we define the conjugation operation @xmath129\equiv{\mathop{{\mathrm{ad}}(b)}}[\varrho]=b\varrho b^{-1}$ ] ( and similarly @xmath130 ) satisfying @xmath131=0,[a , c(c)]\not=0,[c(b),c(c)]\not=0 $ ] . then if @xmath132 we have @xmath133 .
_ noting that @xmath134 similarly , we get @xmath135 . equalling these two expressions we immediately see that @xmath136 holding if @xmath132.@xmath137 + now we identify @xmath138 and @xmath139 so we have @xmath140=\sigma_xp_{\vartheta_w}\vargamma[\varrho]$ ] and to get eq .
( [ covn2 ] ) we apply @xmath56 on the equation from the left and right using the fact that @xmath141=\openone$ ] .
yalmip : a toolbox for modeling and optimization in matlab . j. lfberg .
proceedings of the cacsd conference , taipei , taiwan , 2004 . | the influence of the relativistic covariance requirement on the optimality of the symmetric state - dependent @xmath0 cloning machine is studied .
namely , given a photonic qubit whose basis is formed from the momentum - helicity eigenstates , the change to the optimal cloning fidelity is calculated taking into account the lorentz covariance unitarily represented by wigner s little group .
to pinpoint some of the interesting results , we found states for which the optimal fidelity of the cloning process drops to @xmath1 which corresponds to the fidelity of the optimal classical cloner . also , an implication for the security of the bb84 protocol is analyzed . |
the deviation of the light rays in the gravitational fields is referred to gravitational lensing .
the gravitational lensing ( gl ) in the weak limit has been used to test the general relativity since its beginning @xcite .
but , this theory in the weak limit was not able to describe the high bending and looping of the light rays . hence , scientist community stated this phenomenon in the strong filed regime . in the strong field limit ,
the light rays pass very close to black hole and one set of infinitive relativistic `` ghost '' images would be produce on each side of black hole .
these images are produced due to the light rays wind one or several times around the black hole before reaching to observer . at first , this phenomenon was proposed by darwin @xcite .
several studies of null geodesics in the strong gravitational field have been done in the past years @xcite-@xcite . in 2000 ,
virbhadra and ellis showed that a source of light behind a schwarzschild black hole would product an infinitive series of images on each side of the massive object @xcite .
theses relativistic images are formed when the light rays travel very close to the black hole horizon , wind several times around the black hole before appearing at observer . by an alternative method , frittelli et al . obtained an exact lens equation , integral expression for deflection and compared their results with virbhadra et al @xcite .
a new technic was proposed by bozza et al . to find the position of the relativistic images and their magnification @xcite .
they used the first two terms of approximation to study schwarzschild black hole lensing .
this method was applied to other works such as eiroa , romero and torres studied a reissner - nordstrom black hole lensing@xcite ; petters calculated relativistic effects on microlensing events @xcite .
afterward , the generalization of bozza s method for spherically symetric metric was developed in @xcite .
bozza compared the image patterns for several interesting backgrounds and showed that by the separation of the first two relativistic images , we can distinguish two different collapsed objects .
further studies were developed for other black holes and metrics @xcite-@xcite .
+ the gravitational lenses are important tools for probing the universe . in refs .
@xcite narasimha and chitre predicted that the gravitational lensing of dark matter can give the useful data about the position of the dark matter in the universe .
also , the gravitational lens are used to detect the exotic objects in the universe , such as cosmic strings @xcite-@xcite .
+ on the other hand , the gravitational theories in higher dimensions have attracted considerable attention .
one of these higher dimension gravities is the supersymmetric string theory .
einstein - gauss - bonnet ( @xmath6 ) theory , which emerges as the low - energy limit of this theory , can be considered as an effective model of gravity in higher dimensions .
this theory yields a correction to einstein - hilbert action .
the gauss - bonnet term involves up to second- order derivatives of the metric with the same degrees of freedom as the einstein theory @xcite .
the variation of egb action has different solutions and the spherically symmetric solution in the presence of gauss - bonnet gravity was obtained by boulwar and deser @xcite and charged black hole one is found by wiltshire @xcite .
the properties of the gauss - bonnet black holes have been studied in refs . @xcite-@xcite .
+ in this paper we study the strong gravitational lensing and obtain the logarithmic deflection angle and corresponding coefficients . in the final we investigated some properties of relativistic images .
+ the paper is organized as follows : in section 2 , we briefly present the einstein - gauss - bonnet gravity .
section 3 is devoted to investigate the strong gravitational lensing in the presence of gauss - bonnet term .
we consider the logarithmic term which was proposed by bozza , and obtain its parameters @xmath1 and @xmath2 . in section 4
, some properties of relativistic images will be studied . finally , in the last section we present summery .
the action of einstein - gauss - bonnet gravity in five dimensional is given by @xcite , @xmath7 where @xmath8 and @xmath9 are ricci scalar and gauss - bonnet constant respectively .
@xmath10 is five - dimensional newton s constant and @xmath11 is the gauss - bonnet term as follows , @xmath12 here @xmath13 and @xmath14 are ricci tensor and riemann tensor respectively .
note that the indexes run over the components of five dimensional space .
the exact and spherically metric solution of the above action have been founded by boulware and deser @xcite , @xmath15 where @xmath16 here @xmath17 is related to @xmath18 mass and note that we set @xmath19 . for simplicity
, we introduce the dimensionless quantities as @xmath20 and @xmath21 .
so , we have , @xmath22 when @xmath23 tends to zero the warp factor of the myers - perry metric is obtained @xcite .
the solution of @xmath24 , @xmath25 is the horizon radius of the black hole .
the variation of the horizon is plotted with respect to @xmath26 in figure 1 .
the lens equation for a source of light and an observer situated at large distances from a lens(deflector ) is given by @xcite , @xmath27 where , @xmath28 and @xmath29 stand for the lens - source and observer - source diameter distance , respectively .
the angular positions of source and images with respect to the optical axis ( the line joining the observer and center of the lens ) are represented by @xmath30 and @xmath31 .
the deflection of the light rays denotes by @xmath32 which can be positive , @xmath33 ( bending toward the lens ) or be negative , @xmath34 ( bending away from the lens ) . in the next section
, we will obtain the deflection angle .
the particular distance from the center of the lens to the null geodesic at the source position is called impact parameter , which is given by following expression , @xmath35 we can find the angular positions of images by the intersection of two functions @xmath36 and @xmath37 vs @xmath31 for the same side and vs @xmath38 for opposite side .
in addition to the primary and secondary image positions ( due to the weak limit ) , there is a sequence of intersections that show the angular positions of the relativistic images .
these points are very close to each other , so they are not distinguishable .
for this reason , we call them relativistic images .
these images are due to the bending of light rays more than @xmath39 .
+ now , we are going to investigate the deflection angle in the presence of gauss - bonnet correction gravity . by using the null geodesic equation for the following standard background metric ,
@xmath40 one can find the following equations , @xmath41 @xmath42.\ ] ] where @xmath43 is the energy of photon and @xmath44 and @xmath45 are angular momentums in @xmath46 and @xmath47 directions . here
a dot denotes derivation with respect to affine parameter . if we consider the @xmath31 component of geodesic equations in the equatorial plane @xmath48 ) , we have @xmath49= \dot{\phi}l_\psi=0.\end{aligned}\ ] ] here , if we consider @xmath50 , the deflection angle of light ray becomes zero and this is illegal , therefor we set @xmath51 . for a light ray coming from infinity the deflection angle in the directions @xmath46
is given by @xcite , @xmath52 where @xmath53^ { -\frac{1}{2}}\frac{dx}{x},\end{aligned}\ ] ] where @xmath54 is the closet approach distance for the light ray when it passes near the lens .
the impact parameter for the closet approach is expressed by , @xmath55 the above relation is obtained from the null geodesic equation ( [ rgeo ] ) with setting @xmath56 by using ( [ j ] ) and ( [ impact ] ) one can relate the image position to the closet approach and this relation allows us to write the deflection angle as a function of image position .
the image position is plotted as a function of the closest approach in figure 2 .
we see that the image positions and distances between relativistic images reduce by increasing the gauss - bonnet parameter . for the large values of @xmath54
the curves coincide for any value of gauss - bonnet parameter and this means that primary and secondary image position remain unchanged .
there is a minimum value for the closest approach that is called the photon sphere radius and is a @xmath57const null geodesic .
the photon sphere is the root of derivative of the impact parameter with respect to @xmath54 which is given by , @xmath58 the dashed curve in the figure 1 shows the variation of photon sphere radius .
it decreases with increasing the gauss - bonnet parameter and tends to zero at @xmath59 .
when @xmath54 asymptotically approaches the photon sphere radius , the photon reveals around the lens more times and the deflection angle diverges as @xmath54 tends to photon sphere .
we can rewrite the equation ( [ alpha ] ) as , @xmath60 and @xmath61 where @xmath62 .
the function @xmath63 diverges as @xmath64 approaches to zero .
therefore , we can split the integral ( [ iz ] ) in two parts , the divergent part @xmath65 and the regular one @xmath66 , as follows @xcite @xmath67 @xmath68\,dz.\ ] ] here we expand the argument of the square root in @xmath63 up to the second order in @xmath64 @xmath69 where @xmath70\nonumber\\ & = -\frac{2\left(-x_0 ^ 4 - 2a+2x_0 ^ 2\sqrt{\frac{x_0 ^ 4 + 2a}{x_0 ^ 4}}\right)}{x_0 ^ 4 + 2a}\end{aligned}\ ] ] @xmath71 \nonumber\\ & = \frac{-x_0 ^ 8 - 4x_0 ^ 4a-4a^2 + 6x_0 ^ 6\sqrt{\frac{x_0 ^ 4 + 2a}{x_0 ^ 4}}+4x0 ^ 2\sqrt{\frac{x_0 ^ 4 + 2a}{x_0 ^ 4}}a}{(x_0 ^ 4 + 2a)^2}.\end{aligned}\ ] ] as @xmath23 goes to zero , @xmath72 and @xmath73 tend to five dimensional schwarzschild ones , @xmath74 and @xmath75 . for @xmath76 , @xmath77 is nonzero and the leading order of the divergence in @xmath78 is @xmath79 , which have a finite result . as @xmath80 , @xmath77 approaches zero and
the divergence is of order @xmath81 , that makes the integral divergent logarithmically .
therefor , the deflection angle can be approximated in the following form @xcite @xmath82 where @xmath83}{u_{ps}\sqrt{\mathcal{f}^3(x_{ps})\mathcal{c}(x_{ps } ) } } \approx0.6902 + 0.154\,a+0.373\,a^2\,,\nonumber\\ & b_r = i_r(x_{ps}),\,\,\,\,\,\,u_{ps}=\sqrt{\frac{\mathcal{c}(x_{ps})}{\mathcal{f}(x_{ps})}}\,.\end{aligned}\ ] ] when @xmath23 tends to zero , we have @xmath84 and @xmath85 , that these values belong to myers - perry metric @xcite . using ( [ deflection ] ) and ( [ a ] ) , we can investigate the properties of strong gravitational lensing in the presence of gauss- bonnet correction .
the variations of the @xmath86 is shown in figure 1 .
also , coefficients @xmath1 , @xmath2 , and the deflection angle @xmath32 have been plotted with respect to the gauss- bonnet correction in figures 3 - 4 .
we see that by increasing @xmath9 , the deflection angle @xmath32 and @xmath1 increase and @xmath2 decreases .
the deflection angle becomes diverge as @xmath87 .
in the previous section , we investigated the strong gravitational lensing by using a simple and reliable logarithmic formula for deflection angle that was obtained by bozza et al . and obtained corresponding parameters @xmath1 and @xmath2 .
now we study some properties of relativistic images in the presence of gauss - bonnet gravity .
when a source , lens , and observer are highly aligned , we can write the lens equation in strong gravitational lensing , as following @xcite @xmath88 where @xmath89 is the offset of deflection angle in which all the loops are subtracted , and the integer @xmath90 indicates the @xmath90-th image .
the image position @xmath91 and the image magnification @xmath92 can be approximated as obtained in ref @xcite , @xmath93 @xmath94 where @xmath95 @xmath96 is the angular position of @xmath97 .
they separate the outer most image @xmath98 from the others images which are packed together at @xmath3 .
therefore , the separation between @xmath98 and @xmath3 and ratio of their magnification can be considered by , @xmath99 the asymptotic position of the set of images @xmath3 can be obtained from the minimum of the impact parameter as , @xmath100 by considering equation ( [ thetainfty ] ) , we can approximate equations ( [ sr ] ) as , @xmath101 another property that can be defined for relativistic images is the relative magnification of the outermost relativistic image with the other ones .
this is shown by @xmath5 which is related to @xmath102 as , @xmath103 if we suppose a five dimensional black hole with mass @xmath104 ( galaxy center mass ) and the distance between the observer and black hole is @xmath105 ( the distance between the sun and galaxy center ) @xcite , we can study the effect of the gauss - bonnet parameter on these quantities .
our results are presented in figure 4 - 6 and table 1 .
+ [ s ] [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ]
the light rays can be deviated from a straight way in the gravitational field as predicted by general relativity in which this deflection of light rays is known as gravitational lensing . in the strong field limit
, the deflection angle of the light rays passing very close to the black hole , becomes so large that , the light rays wind several times around the black hole before appearing at the observer .
therefore the observer would detect two infinite set of faint relativistic images produced on each side of the black hole . on the other hand ,
the gravitational theories in higher dimensions have been attracting considerable attention in recent decades .
einstein - gauss - bonnet theory that emerges as the low - energy limit of supersymmetric string theory , is one of the candidates for higher dimension theory .
we considered five dimensional metric with gauss - bonnet correction and studied the strong gravitational lensing and obtained the deflection angle and corresponding parameters @xmath1 and @xmath2 .
we saw that by increasing @xmath9 , the deflection angle @xmath32 and @xmath1 increase and @xmath2 decreases .
the deflection angle became diverge as @xmath87 .
finally , we estimated some properties of relativistic images which can be detected by astronomical instruments .
our results have been presented in figures 5 - 7 . in figures 5 and 7 ,
the variations of @xmath3 and @xmath5 , shown that the position of compacted images and relative magnification reduce with increasing @xmath9 .
also the angular separation is an increasing function and diverges , as @xmath9 tends to two ( figure 6 ) .
furthermore , we saw that the position of images reduces with @xmath9 ( see figure 2 ) .
but you should note the decreasing rate of @xmath3 is more than @xmath98 , therefore @xmath4 will be increasing function .
11 a. einstein , _ science _ ,
* 84 * , 506 ( 1936 ) .
p. schneider , j. ehlers , and e. e. falco `` gravitational lenses '' _ springer - verlag , berlin _ , ( 1992 ) . c. darwin , _ proc . of the royal soc . of london
_ * 249 * , 180 ( 1959 ) .
j. m. bardeen , `` black holes '' , ed .
c. dewitt and b. s. dewitt , _ gordon and breach _ , * 215 * ( 1973 ) .
viergutz , _ a and a _ * 272 * , 355 ( 1993 ) .
r. j. nemiroff , _ amer .
_ 61,619 ( 1993 ) .
h. falcke , f. melia and e. agol , _ apj letters _ * 528 * l13 ( 1999 ) .
k. s. virbhadra and g. f. r. ellis , _ phys .
d _ * 62 * , 084003 ( 2000 ) .
s. frittelli , t. p. kling and e. t. newman , _ phys .
d _ * 61 * , 064021 ( 2000 ) .
v. bozza , s. capozziello , g. lovane and g. scarpetta , _ gen .
grav_. * 33 * , 1535 ( 2001 ) .
e. f. eiroa , g. e. romero and d. f. torres , _ phys .
d _ * 66 * , 024010 ( 2002 ) ; e. f. eiroa , _ phys .
d _ * 71 * , 083010 ( 2005 ) ; e. f. eiroa , _ phys .
d _ * 73 * , 043002 ( 2006 ) .
e. f. eiroa , _ phys.rev .
d _ * 71 * , 083010 ( 2005 ) .
a. o. petters , _ mnras _ * 338 * , 457 ( 2003 )
. v. bozza , _ phys .
d _ * 66 * , 103001 ( 2002 ) . v. bozza , f. de luca , g. scarpetta and m. sereno , _ phys .
d _ * 72 * , 08300 ( 2005 ) ; v. bozza , _ phys .
d _ * 67 * , 103006 ( 2003 ) .
v. bozza , _ phys .
* 78 * , 103005 ( 2008 ) .
a. bhadra , _ phys .
d _ * 67 * , 103009 ( 2003 ) .
t. ghosh and s. sengupta , _ phys .
d _ * 81 * , 044013 ( 2010 ) .
a. n. aliev and p. talazan , _ phys .
d _ * 80 * , 044023 ( 2009 ) .
galin n. gyulchev and ivan zh .
stefanov , _ phys .
k. s. virbhadra and g. f. r. ellis , _ phys .
d _ * 65 * , 103004 ( 2002 ) .
k. s. virbhadra , _ phys .
d _ * 79 * , 083004 ( 2009 ) .
k. s. virbhadra and c. r. keeton , _ phys .
d _ * 77 * , 124014 ( 2008 ) .
k. s. virbhadra , d. narasimha and s. m. chitre , _ astron . astrophys_.
* 337 * , 1 - 8 ( 1998 ) .
87 , 063005 ( 2013 ) .
y. liu , s. chen and j. jing , _ phys .
d _ * 81 * , 124017 ( 2010 ) .
s. chen , y. liu and j. jing , _ phys .
d _ * 83 * , 124019 ( 2011 ) .
j. sadeghi , a. banijamali and h. vaez , _ astrophys . space .
sci _ * 343 * , 559 ( 2013 ) .
j. sadeghi , and h. vaez , arxiv:1310.4486 [ gr - qc ] .
d. narasimha and s. m. chitre , _ astro .
j_. * 97 * 327 ( 1989 ) .
d. narasimha , _ icnapp _ , ed .
r. cowsik , * 251 * ( 1994 ) . c. j. hogan , r. narayan , _ mnras _ * 2111 * , 575 ( 1984 ) .
a. vilenkin , _ apjl _ * 51 * , 282 ( 1984 ) .
a. vilenkin , _ nat _ * 322 * , 613 ( 1986 ) .
b. zwiebach , _ phys .
b _ 156 , 315 ( 1985 ) .
d. g. boulware and s. deser , _ phys .
rev . lett_. * 55 * , 2656 ( 1985 ) .
d. wiltshire , _ phys .
d _ * 38 * , 2445 ( 1988 ) .
rong - gen cai ei , qi guo , _ phys .
d _ * 69 * , 104025 ( 2004 ) .
rong - gen cai , _ phys .
* 65 * , 084014 ( 2002 ) .
j. crisostomo , r. troncoso , j. zanelli , _ phys .
d _ * 62 * , 084013 ( 2000 ) .
sachiko ogushi , misao sasaki , _ prog .
phys_. * 113 * 979 ( 2005 ) .
tim clunan , simon f. ross , douglas j. smith , _ class .
quant . grav_.
* 21 * , 3447 ( 2004 ) .
a. barrau , j. grain , s.o .
alexeyev , _ phys.lett.b _ * 584 * , 114 ( 2004 ) .
maximo banados , _ phys .
b _ * 579 * , 13 ( 2004 ) .
m. cvetic , s. nojiri , s. d. odintsov , _ nucl .
b _ * 628 * , 295 ( 2002 ) .
s. nojiri , s. d. odintsov , _ phys .
b _ * 523 * , 165 ( 2001 ) .
s. nojiri , s. d. odintsov , s. ogushi , _ phys .
d _ * 65 * , 023521 ( 2002 ) .
gustavo dotti , reinaldo j. gleiser , _ class .
* 22 * , l1 ( 2005 ) .
ishwaree p. neupane , _ phys . lett .
d _ * 69 * , 084011 ( 2004 ) .
ishwaree p. neupane , _ phys .
d _ * 67 * , 061501 ( 2003 ) .
s. gillessen , f. eisenhauer , s. trippe , t. alexander , r. genzel , f. martins , and t. ott , _ apj _ * 692 * , 1075 ( 2009 ) . | in this paper we investigate the strong gravitational lensing in a five dimensional background with gauss - bonnet gravity , so that in 4-dimensions the gauss - bonnet correction disappears . by considering the logarithmic term for deflection angle
, we obtain the deflection angle @xmath0 and corresponding parameters @xmath1 and @xmath2 .
finally , we estimate some properties of relativistic images such as @xmath3 , @xmath4 and @xmath5 . + + * keywords : * gravitational lensing ; gauss - bonnet correction + |
since the foundation of quantum mechanics , a huge amount of data on energy levels , linewidths and other properties of atoms have been recorded . very often , such compilations are still waiting for a qualitative analysis , based on simple models . in recent articles
@xcite , on the basis of the scaling suggested by thomas - fermi theory @xcite , we have demonstrated universality in the ionization potentials and the correlation energies of atomic ions . in the present paper , we construct an analytical expression for the ionization energy of atomic ions , which is to be used in order to detect problematic values @xcite in the numbers provided by the nist database.@xcite our expression is a regularized perturbative series ( rps ) , previously employed in other contexts @xcite .
we use perturbation theory in @xmath8 @xcite in order to compute the first two coefficients of the energy series in the large-@xmath1 region .
additionaly , we require our rps to reproduce the value of the ionization potential at @xmath6 ( i.e. the electron affinity ) and the slope of the curve at this point . the latter is computed from a kind of zero - range forces theory that uses as input the electron affinity and the covalent radius of the neutral atom with @xmath7 electrons .
@xcite the rps continuosly interpolates between the @xmath9 and large-@xmath1 limits for a given isoelectronic sequence .
hereunder , we provide the main formulae entering the rps expression .
a few isoelectronic sequences are studied in quality of examples . a systematic analysis of the nist database is contained in refs . .
it is well known that , for large @xmath1 , the attraction of the electrons by the nucleus is stronger than electron - electron repulsion . on the contrary , for the neutral atom
both contributions are more or less balanced , and in the anionic domain this balance may even be broken at a given @xmath4 , where the atom spontaneously autoionizes . first - principle calculations @xcite and some extrapolations @xcite indicate that @xmath5 is indeed very close to @xmath7 , excluding the possibility of doubly charged negative ions .
a recent result by gridnev,@xcite on the other hand , rigorously states that the wave function is normalized at threshold .
if we combine this result with perturbation theory , we get that the binding energy exhibits a linear dependence on @xmath1 near @xmath5 . in a previous paper,@xcite
we compute the slope of the curve not at @xmath5 , but at @xmath6 . at this value of @xmath1
the outermost electron weakly interacts with the neutral core and the interaction is short - ranged .
it can be shown that conditions are fulfilled for the application of zero - range forces theory .
@xcite the slope of the curve may be computed from:@xcite @xmath10 where @xmath11 and @xmath12 is the electron affinity of the neutral system with @xmath7 electrons .
atomic units are to be used everywhere in the paper .
@xmath13 is related to the size of the core , containing nuclear charge @xmath1 and @xmath7 electrons . for computational purposes
, we use the covalent radius of the @xmath7 electron atom as an estimation of @xmath13 .
( [ eq1 ] ) will be used in sect .
[ regular ] , where we construct a rps for the ionization potential .
in the following , we shall construct the large-@xmath1 series for the atomic energy .
this is , in fact , a formal limit . in nature ,
atomic ions become unstable for large @xmath1 , showing a threshold for electron - positron pair production at @xmath14.@xcite performing the scaling @xmath15 in the non - relativistic hamiltonian , we get : @xmath16 notice that the expression inside brackets has a one - particle contribution ( kinetic energy plus nuclear attraction ) and the two - particle repulsion between electrons .
the latter is of order @xmath17 . at large values of @xmath1
, the atom can be described as a system of non - interacting electrons in the central coulomb field of the nucleus .
the energy in this leading approximation is : @xmath18 next , we shall include electron repulsions in first order perturbation theory .
the energy is written as : @xmath19 where @xmath20 is the slater determinant made of hydrogenic functions .
corrections are explicitly given by : @xmath21 note that the sums runs over the occupied orbitals @xmath22 and @xmath23 in the slater determinant , and that @xmath24 and @xmath25 denote , respectively , direct and exchange two - electron coulomb integrals involving orbitals @xmath26 and @xmath27 .
their explicit expression can be found in ref . [ ] .
once we constructed a series for the total energy : @xmath28 one can find also a similar expression for the ionization potential , defined as @xmath29 .
we get : @xmath30 where , @xmath31 in these equations , @xmath32 is the principal quantum number of the last electronic shell , and @xmath33 the last occupied orbital . to end up this section ,
we shall stress that , in the large-@xmath0 limit : @xmath34 eq .
( [ largenb2 ] ) comes from analytical estimations , whereas eq .
( [ largenb1 ] ) comes from a fit to the numerical results .
these functional forms are consistent with the dependence @xmath35 , suggested by thomas - fermi theory.@xcite at large @xmath1 , a relativistic approach is required . in the leading approximation
, one should solve the dirac equation for an electron in a central coulomb field .
we choose a simpler approach in which relativistic corrections are computed in first order degenerate perturbation theory , @xmath36 both @xmath37 and @xmath38 are eigenstates of the non - relativistic one - electron hamiltonian .
greek indices label states for which the total angular momentum ( orbital plus spin ) is a good quantum number .
the relativistic perturbation , @xmath39 , includes the kinetic ( @xmath40 ) , spin - orbit ( @xmath41 ) , and the darwin term ( @xmath42 ) terms.@xcite the sum in eq .
( [ wf_rel ] ) runs over the space orthogonal to @xmath43 . the first order correction to @xmath44
can be obtained by replacing the non - relativistic states @xmath23 and @xmath33 in eq .
( [ b1 ] ) by the expression ( [ wf_rel ] ) , yielding : @xmath45 where @xmath46 is the last occupied state .
the @xmath47 coefficient must be changed also in accordance with ( [ e_rel ] ) .
the final expression for @xmath47 reads : @xmath48 , \label{b2_rel}\ ] ] where @xmath49 is the total angular momentum quantum number of the last occupied state and @xmath50 is the fine structure constant .
a summary of matrix elements @xmath51 is given in the appendix .
details on the derivation of @xmath52 can be found in ref .
once the region near the anionic threshold and the large-@xmath1 limit are described , one may try to find an interpolation between them . to this end
, we use a regularization of the perturbative series , eq .
( [ z_series]).@xcite the next two formal terms of the series @xmath53 coming , in principle , from higher order perturbative corrections , are instead used to force that , at @xmath6 , @xmath54 and @xmath55 .
that is : @xmath56 we get a linear system of two equations and two variables ( @xmath57 and @xmath58 ) , yielding : @xmath59
we would like to show how eq .
( [ reg_series ] ) , with the coefficients @xmath44 and @xmath47 given in ( [ b1_rel ] ) and ( [ b2_rel ] ) , respectively , and @xmath57 and @xmath44 coming from ( [ sol_b ] ) , can be used to detect inconsistencies in the nist data for the ionization potential of atomic ions .
we study four isoelectronic systems ( fig .
[ fig1]-[fig3 ] ) in quality of examples .
an exhaustive revision will be published elsewhere.@xcite ( color online ) the case of ne - like ions ( @xmath60 ) .
upper panel : ionization potential taken from the nist compilation along with our nonrelativistic ( discontinuous , blue ) and relativistically corrected ( continuous , red ) rps predictions versus atomic charge .
curves seem smooth at any scale .
lower panel : the difference between the nist reported values and the rps relativistic curve versus atomic charge . a 5-points running average curve ( continuous , blue ) is used as a reference for measuring deviations ( see the inset ) .
inconsistencies are detected at @xmath61 , 22 , 27 , 74 and @xmath62 .
we stress also that there is a great dispersion of the data for @xmath63 , which is however within the hughe error bars reported for these points . ] in figs .
[ fig1]-[fig3 ] , we plot the non - relativistic and the relativistic rps , along with the nist data , for these sequences .
the lower panels show the difference nist - rps(relativ ) .
the maximum relative errors are near 1% for @xmath60 , 2% for @xmath64 , and rises to around 10% for @xmath65 . in the first case , @xmath60 , our approximate treatment of relativity does not reproduce the correct asymptotics at large @xmath1 . in the rest of the systems , both asymptotics ( @xmath6 and large @xmath1 ) are correct , and the maximum errors
are reached at intermediate @xmath1 , as it is common with interpolants @xcite .
( color online ) the same as fig .
[ fig1 ] for ni- ( @xmath66 , left panels ) and cu - like ( @xmath67 , right panels ) ions . in the @xmath68 system , we find inconsistencies at @xmath69 , 74 , and 79 , whereas for @xmath70 small deviations at @xmath71 , and 79 are noticed .
relativistic effects are significant for cu - like ions at large @xmath1 values , in which the last electron occupies a 4s orbital . ]
we claim that , in spite of the fact that our relativistic rps does not have spectroscopic precision , abrupt changes in the difference nist - rps may be a sign of inconsistency .
indeed , abrupt changes in @xmath72 are related to rearrangements of the electronic spectrum . in the interval between rearrangements or for large enough @xmath1 ,
the occupancy of orbitals is fixed , and @xmath72 should be smooth .
the difference with our smooth rps interpolant should also be a smooth function of @xmath1 .
ne - like systems are closed shell , and do not exhibit rearrangements at any @xmath1 . in order to make evident inconsistent points in the nist data , we construct an average nist - rps curve by means of a 5-points running average . in fig .
[ fig1 ] , the @xmath73 point is so far from the average curve , for example , that it should be corrected .
we can even give an estimate of the needed correction by measuring the distance to the average curve , which in this case is -0.034 a.u .
in ni - like ions , the 3d@xmath74 electronic configuration is reached already for @xmath75 .
thus , we expect a smooth dependence from this point on . the @xmath76 point , for example ,
is deviated from the average curve in 0.141 a.u . , and its error bar is only 0.115 a.u .
cu - like ions , on the other hand , show a 3d@xmath744s configuration at any @xmath1 .
nd - like ions experience rearrangements at various @xmath1 values , but in the neighbourhood of @xmath76 , a problematic point , the difference should be smooth .
thus , we can undoubtedly distinguish this point .
( color online ) the same as fig .
[ fig1 ] but for nd - like ions ( @xmath65 ) .
only the @xmath76 point is clearly inconsistent . ]
the main result of the present paper is an analytical expression for the ionization energies of @xmath0-electron ions ( @xmath77 ) based on first - principles .
this formula is exact in the @xmath6 and large @xmath1 regions . in the transition region ,
the error is only a few percents of the total ionization potential .
we show that our rps expression may help identifying problems in a large database , such as the nist compilation .
a detailed analysis of the data is presented in refs and .
the authors are grateful to the caribbean network for quantum mechanics , particles and fields ( ictp ) for support .
. also acknowledge financial support from the european community s fp7 through the marie curie itn - index .
we can write the relativistically corrected hamiltonian as : we consider first - order perturbative corrections due to @xmath39 .
the relevant matrix element is @xmath83 , where @xmath84 is an eigenstate of @xmath85 , @xmath86 , @xmath87 , @xmath88 , @xmath89 , and @xmath90 an eigenstate of @xmath85 , @xmath86 , @xmath88 , @xmath91 , @xmath92 .
only states such that @xmath93 enter eq .
( [ wf_rel ] ) , thus we restrict ourselves to this case .
we have : notice that @xmath95 states can be labelled by @xmath99 ( principal quantum number ) , @xmath82 , @xmath27 and @xmath100 ( total angular momentum projection on @xmath101 ) , whereas for @xmath102 we need @xmath99 , @xmath82 , @xmath103 ( orbital angular momentum projection on @xmath101 ) , and @xmath104 ( spin angular momentum projection on @xmath101 ) .
we can expand @xmath105 in terms of @xmath106 by means of the clebsch - gordan coefficients : 99 r. carcasses and a. gonzalez , phys .
a 80 , 024502 ( 2009 ) .
a. odriazola , a. gonzalez and e. rasanen , phys .
a 90 , 052510 ( 2014 ) .
kirzhnits , yu.e .
lozovik , and g.v .
shpatakovskaya , sov .
18 , 649 ( 1975 ) ; e.h . lieb , rev .
53 , 603 ( 1981 ) ; l. spruch , rev . mod .
63 , 1512 ( 1991 ) .
g. gil and a. gonzalez , the nist compilation of ionization potentials revisited ( i ) : from he - like to xe - like ions , arxiv:1405.6831 [ physics.atom-ph ] .
g. gil and a. gonzalez , the nist compilation of ionization potentials revisited ( ii ) : from la - like to ac - like ions , in preparation .
a. kramida , yu .
ralchenko , j. reader , and nist asd team ( 2012 ) .
nist atomic spectra database ( ver .
5.0 ) , [ online ] . available :
http://physics.nist.gov/asd [ 2013 , march 4 ] .
national institute of standards and technology , gaithersburg , md .
a. matulis and f.m .
peeters , j. phys . : cond .
matter 6 , 7751 ( 1994 ) ; a. gonzalez and i. mikhailov , int . j. mod
b 11 , 3469 ( 1997 ) .
g. senatore and n. h. march , j. chem .
83 , 1232 ( 1985 ) .
g. gil and a. gonzalez , mod .
b 27 , 1350178 ( 2013 ) .
h. hogreve , j. phys .
b 31 , l439 ( 1998 ) .
a.v . sergeev and s. kais , int .
75 , 533 ( 1999 ) ; s. kais and p. serra , int .
19 , 97 ( 2000 ) .
gridnev , j. math .
53 , 102108 ( 2012 ) .
bethe , phys .
76 , 38 ( 1949 ) ; l.d .
landau and ya.a .
smorodinsky , _ lectures on nuclear theory _ ( plenum press , new york , 1959 ) .
popov , sov .
jetp 32 , 3 ( 1971 ) ; w. greiner , _ relativistic quantum mechanics _ ( springer - verlag , berlin , 2000 ) .
g. gil , stability of atoms in the anionic domain ( @xmath3 ) , diploma thesis ( faculty of physics , university of havana , 2012 ) .
also in g. gil , a. gonzalez , arxiv:1309.3468 [ physics.atom-ph ] .
bransden and c.j .
joachain , _ physics of atoms and molecules _ ( longman , essex , 1983 ) . | we study @xmath0-electron atoms with nuclear charge @xmath1 . it is well known that , in the cationic ( @xmath2 ) high-@xmath1 region
, the atom behaves as a weakly interacting system .
the anionic ( @xmath3 ) regime , on the other hand , is characterized by an instability threshold at @xmath4 , below which the atom spontaneously emits an electron .
we construct a regularized perturbative series ( rps ) for the ionization potential of ions in an isoelectronic sequence that exactly reproduces both , the large @xmath1 and the @xmath1 near @xmath5 limits .
the large-@xmath1 expansion coefficients are analytically computed from perturbation theory , whereas the slope of the energy curve at @xmath6 is computed from a kind of zero - range forces theory that uses as input the electron affinity and the covalent radius of the neutral atom with @xmath7 electrons .
relativistic effects , at the level of first - order perturbation theory , are considered .
our rps formula is to be used in order to check the consistency of the ionization potential values for atomic ions contained in the nist database . |
we will treat a class of models of ` dense ' random graphs , that is , simple graphs on @xmath0 vertices in which the average number of edges is of order @xmath1 .
more specifically we will consider exponential random graphs in which dependence between the random edges is defined through some finite graph , in imitation of the use of potential energy to provide dependence between particle states in a grand canonical ensemble of statistical physics .
exponential random graphs have been widely studied ( see @xcite for a range of recent work ) since the pioneering work on the independent case by erds and rnyi @xcite .
we will concentrate on the phenomenon of phase transitions which can emerge for dependent variables .
following analyses using mean - field and other uncontrolled approximations ( see @xcite ) there has recently been important progress by chatterjee and diaconis @xcite , including the first rigorous proof of singular dependence on parameters .
we will extend their result both in the class of models and parameter values under control and provide an appropriate formalism of phase structure for such models .
we consider the class of models in which the probability of the simple graph @xmath2 on @xmath0 vertices is given by : @xmath3},\ ] ] where : @xmath4 is an edge , @xmath5 is any finite simple graph with @xmath6 edges , @xmath7 is the normalization constant , @xmath8 is the density of graph homomorphisms @xmath9 : @xmath10 and @xmath11 denotes the vertex set .
expectation of a real function of a random graph is denoted @xmath12 .
our main results are the following . .1truein
[ one ] restrict to @xmath13 . then the pointwise limit @xmath14 exists and is analytic in @xmath15 and @xmath16 off a certain curve @xmath17 which includes the end point @xmath18 the derivatives @xmath19 and @xmath20 are both discontinuous across the curve , except at @xmath21 where , however , all the second derivatives @xmath22 , @xmath23 , and @xmath24 diverge . .1truein
[ two ] if in theorem [ one ] the graph @xmath5 is a @xmath25-star , @xmath26 , the analogous result also holds for @xmath27 .
a @xmath25-star has @xmath25 edges meeting at a vertex . .1truein
[ corr ] the parameter space @xmath28 of each of the models consists of a single phase with a first order phase transition across the indicated curve and a second order phase transition at the critical point @xmath21 . to explain the language of phase transitions in corollary [ corr ] we first give a superficial introduction to the formalism of classical statistical mechanics within @xmath29-dimensional lattice gas models ; for more details see for instance @xcite .
assume each point in a @xmath29-dimensional cube @xmath30 is randomly occupied ( by one particle ) or not occupied , and assume there is a ( many - body ) potential energy of value @xmath31 associated with every occupied subset of @xmath32 congruent to a certain @xmath33 .
the interaction is attractive if @xmath34 and repulsive if @xmath35 .
we define the probability that the occupied sites in @xmath32 are precisely @xmath36 by @xmath37 } } { { \mathbb z}_n},\ ] ] where the parameter @xmath38 is called the inverse temperature , the parameter @xmath39 is called the chemical potential , the normalization constant @xmath40 is called the partition function , @xmath4 and @xmath5 are subsets of @xmath32 with @xmath4 a singleton and the cardinality @xmath41 , and @xmath42 is the number of copies of @xmath43 in @xmath36 .
( we are using ` free ' boundary conditions . )
one of the basic features of the formalism is that the free energy density , @xmath44/n^d$ ] , contains all ways to interact with or influence the system , so that ` all ' physically significant quantities can be obtained by differentiating it with respect to @xmath45 and @xmath46 .
for instance @xmath47 the ( average ) particle density . to model materials in thermal equilibrium , calculations in this formalism normally require that the system size be sufficiently large , and in practice one often resorts to using @xmath48 . with this as motivation we tentatively define a ` phase ' as a set of states ( i.e. probability distributions ) corresponding to a connected region of the @xmath49 parameter space , which is maximal for the condition that @xmath50 are analytic in @xmath45 and @xmath46 for all @xmath51 .
intuitively one associates a ` phase transition ' with singularities which develop in some of these quantities as the system size diverges .
an important simplification was proven by yang and lee @xcite who showed that the limiting free energy density @xmath52 always exists and that certain limits commute : @xmath53 this implies that phases and phase transitions can be determined from the limiting free energy density , and so a phase is commonly defined ( see for instance @xcite ) as a connected region of the @xmath54 parameter space maximal for the condition that @xmath55 is analytic . using the obvious analogues for random graphs , with @xmath15 playing the role of @xmath56 and @xmath16 the role of @xmath57 ( and therefore positive if and only if the model is ` attractive ' ) , @xmath7 plays the key role of free energy density .
we will show in theorems [ a ] and [ b ] below that the limiting free energy density @xmath58 exists , and the proof of theorem 2 by yang and lee @xcite , on the commutation of limits , then goes through without any difficulty in this setting , so we can again define phases and phase transitions through the limiting free energy density , as follows .
a phase is a connected region , of the parameter space @xmath28 , maximal for the condition that the limiting free energy density , @xmath58 , is analytic .
there is a @xmath59-order transition at a boundary point of a phase if at least one @xmath59-order partial derivative of @xmath60 is discontinuous there , while all lower order derivatives are continuous .
theorems [ one ] and [ two ] thereby justify our interpretation in corollary [ corr ] that each of our models consists of a single phase with a first order phase transition across the indicated curve except at the end ( or ` critical ' ) point @xmath61 , where the transition is second order , superficially similar to the transition between liquid and gas in equilibrium materials .
chatterjee and diaconis have proven that the main object of interest , @xmath62 , exists for all @xmath15 and nonnegative @xmath16 and is the solution of a certain optimization problem : [ cd ] fix one of our models and assume @xmath5 has @xmath26 edges .
then for any @xmath63 and @xmath64 , the pointwise limit @xmath65 exists : @xmath66 } \big(\beta_1u+\beta_2u^p-\frac{1}{2}u\log u-\frac{1}{2}(1-u)\log(1-u)\big).\ ] ] the following detailed analysis of this maximization problem is therefore fundamental to understanding the phase structure of our models , and we now address it .
[ max ] fix an integer @xmath67 .
consider the maximization problem for @xmath68 on the interval @xmath69 $ ] , where @xmath70 and @xmath71 are parameters .
then there is a v - shaped region in the @xmath72 plane with corner point @xmath21 such that outside this region , @xmath73 has a unique local maximizer ( hence global maximizer ) @xmath74 ; whereas inside this region , @xmath73 always has exactly two local maximizers @xmath75 and @xmath76 . moreover , for every @xmath15 inside this v - shaped region ( @xmath77 ) , there is a unique @xmath17 such that the two local maximizers of @xmath78 are both global maximizers .
furthermore @xmath79 is a continuous and decreasing function of @xmath15
. plane . graph drawn for @xmath80.,width=384 ] by the lebesgue differentiation theorem , @xmath79 being monotone guarantees that it is differentiable almost everywhere . the location of maximizers of @xmath73 on the interval @xmath69 $ ] are closely related to properties of its derivatives @xmath81 and @xmath82
: @xmath83 @xmath84 we first analyze properties of @xmath82 on the interval @xmath69 $ ] . consider instead the function @xmath85 simple optimization shows @xmath86 and the equality holds if and only if @xmath87 .
thus @xmath88 the graph of @xmath89 is concave up with two ends both growing unbounded , and the global minimum is achieved at @xmath87 .
this implies that for @xmath90 , @xmath91 on the whole interval @xmath69 $ ] ; whereas for @xmath92 , @xmath82 will take on both positive and negative values , and we denote the transition points by @xmath93 and @xmath94 ( @xmath95 ) .
based on properties of @xmath82 , we next analyze properties of @xmath81 on the interval @xmath69 $ ] . for @xmath90
, @xmath81 is monotonically decreasing . for @xmath92
, @xmath81 is decreasing from @xmath96 to @xmath93 , increasing from @xmath93 to @xmath94 , then decreasing again from @xmath94 to @xmath97 .
based on properties of @xmath81 and @xmath82 , we analyze properties of @xmath73 on the interval @xmath69 $ ] .
independent of the choice of parameters @xmath15 and @xmath16 , @xmath73 is a bounded continuous function , @xmath98 , and @xmath99 , so @xmath73 can not be maximized at @xmath96 or @xmath97 . for @xmath90 ,
@xmath81 crosses the @xmath100-axis only once , going from positive to negative .
thus @xmath73 has a unique local maximizer ( ( hence global maximizer ) @xmath74 .
for @xmath101 , the situation is more complicated and deserves a careful analysis .
if @xmath102 ( resp .
@xmath103 ) , @xmath73 has a unique local maximizer ( hence global maximizer ) at a point @xmath104 ( resp .
@xmath105 ) . if @xmath106 , then @xmath73 has two local maximizers @xmath75 and @xmath76 , with @xmath107 . notice that @xmath93 and @xmath94 are solely determined by the choice of parameter @xmath92 , and vice versa . by ( [ beta2 ] ) ,
@xmath108 @xmath109 consider the function @xmath110 it is not hard to see that @xmath111 , @xmath112 , @xmath113 is decreasing from @xmath96 to @xmath114 , then increasing from @xmath114 to @xmath97 , and the global minimum value is @xmath115 this implies in particular that @xmath102 for @xmath116 .
the only possible region in the @xmath72 plane where @xmath106 is thus bounded by @xmath117 and @xmath92 .
we now analyze the behavior of @xmath118 and @xmath119 more closely when @xmath15 and @xmath16 are chosen from this region .
recall that by construction , @xmath95 . by monotonicity of @xmath113 on the intervals @xmath120 and @xmath121 , there exist continuous functions @xmath122 and @xmath123 of @xmath15 , such that @xmath124 for @xmath125 and @xmath126 for @xmath127 .
@xmath122 is an increasing function of @xmath15 , whereas @xmath123 is a decreasing function , and they satisfy @xmath128 also , as @xmath129 , @xmath130 and @xmath131 . by ( [ beta2 ] ) , the restrictions on @xmath93 and @xmath94 yield restrictions on @xmath16 .
we have @xmath124 for @xmath132 and @xmath126 for @xmath133 . notice that @xmath134 and @xmath135 are both decreasing functions of @xmath15 , and as @xmath129 , they both grow unbounded . by construction , for every parameter value @xmath72 , @xmath136 .
also , for fixed @xmath15 , @xmath134 is the value of @xmath16 for which @xmath137 , and @xmath135 is the value for which @xmath138 .
thus the curve @xmath135 must lie below the curve @xmath134 . and
together they generate the bounding curves of the v - shaped region in the @xmath72 plane where two local maximizers exist for @xmath73 .
it is not hard to see that the corner point is given by @xmath139 .
( see figures 16 . ) fixing an arbitrary @xmath77 , we examine the effect of varying @xmath16 on the graph of @xmath81 .
it is clear from ( [ l ] ) that @xmath81 shifts upward as @xmath16 increases . as a result , as @xmath16 gets large , the positive area bounded by the curve @xmath81 increases , whereas the negative area decreases . by the fundamental theorem of calculus ,
the difference between the positive and negative areas is the difference between @xmath140 and @xmath141 , which goes from negative ( @xmath138 , @xmath75 is the global maximizer ) to positive ( @xmath137 , @xmath76 is the global maximizer ) as @xmath16 goes from @xmath135 to @xmath134 .
thus there must be a unique @xmath142 such that @xmath75 and @xmath76 are both global maximizers .
we denote this @xmath16 by @xmath143 ; see figures [ v - shape ] and [ qcurve ] . by analyzing the graph of @xmath81
, we see that the parameter values of @xmath144 are exactly the ones for which positive and negative areas bounded by @xmath81 equal each other .
an increase in @xmath15 will induce an upward shift of @xmath81 , which must be balanced by a decrease in @xmath17 .
similarly , a decrease in @xmath15 will induce a downward shift of @xmath81 , which must be balanced by an increase in @xmath17 .
this justifies that @xmath79 is monotonically decreasing in @xmath15 .
furthermore , the continuity of @xmath81 as a function of @xmath15 and @xmath16 implies the continuity of @xmath79 as a function of @xmath15 .
the transition curve @xmath17 displays a universal asymptotic behavior as @xmath145 : @xmath146 by proposition [ max ] , it suffices to show that as @xmath147 , @xmath148 has two global maximizers @xmath75 and @xmath76 .
this is easy when we realize that as @xmath149 , @xmath150 for every @xmath100 in @xmath151 .
the limiting maximizers on @xmath152 $ ] are thus @xmath153 and @xmath154 , with @xmath155 .
so far we have used results from @xcite but have avoided specific reference to the framework of graph limits , developed by lovsz et al , which was used to prove those results .
we now need to refer directly to graph limits ; for the notation and an introduction to this material see for instance @xcite or @xcite .
[ gen ] let @xmath2 be a random graph on @xmath0 vertices in one of our models . for parameter values of @xmath72 in the upper half - plane @xmath156 ,
the behavior of @xmath2 in the large @xmath0 limit is as follows : @xmath157 where @xmath158 is the set of maximizers of ( [ l ] ) . the assumptions of theorems 4.2 and 6.1 in @xcite are satisfied for parameter values @xmath156 . by proposition [ max ] , along the curve @xmath144 , the maximization problem ( [ l ] ) is solved at two values @xmath75 and @xmath76 ; whereas off this curve , it is solved at a unique value @xmath74 .
thus in the large @xmath0 limit , along the curve @xmath144 , @xmath2 behaves like an erds - rnyi graph @xmath159 ( @xmath100 picked randomly from @xmath75 and @xmath76 ) ; whereas off this curve , @xmath2 is indistinguishable from the erds - rnyi graph @xmath160 .
[ q ] fix any @xmath161 .
let @xmath162 be an edge , so @xmath8 is the edge density of @xmath2 .
then there exists a continuous and decreasing function @xmath163 such that @xmath164 and @xmath165 here @xmath93 and @xmath94 are defined as in the proof of proposition [ max ] : @xmath166 . as @xmath167 , @xmath168 and @xmath169 and
the jump is noticeable even for relatively small values of @xmath16 . as @xmath143 is a continuous and decreasing function of @xmath15 , the inverse function @xmath163 exists and is also continuous and decreasing .
we examine the effect of varying @xmath15 on the graph of @xmath81 ( and hence on the global maximizers of @xmath73 ) .
first note that varying @xmath15 does not change the shape of @xmath81 .
inside the v - shaped region , there are three cases . recall that @xmath107 . for @xmath170 ,
positive and negative areas bounded by @xmath81 equal each other , thus @xmath75 and @xmath76 are both global maximizers . for @xmath171 , the graph of @xmath81 shifts downward
, negative area exceeds positive area , thus @xmath75 is the global maximizer . for @xmath172 , the graph of @xmath81 shifts upward , positive area exceeds negative area ,
thus @xmath76 is the global maximizer . outside the v - shaped region ,
there are two cases . below the lower bounding curve
, @xmath81 has a unique local maximizer @xmath105 . above the upper bounding curve
, @xmath81 has a unique local maximizer @xmath104 .
our conclusion then follows from theorem [ gen ] .
[ p ] assume that in one of our models @xmath5 is a @xmath25-star ( @xmath67 ) .
for all parameter values @xmath72 , the behavior of @xmath2 in the large @xmath0 limit is as follows : @xmath157 where @xmath158 is the set of maximizers of ( [ l ] ) .
this follows from related results in @xcite .
we separate the parameter plane @xmath173 into upper and lower half - planes .
the upper half - plane ( @xmath174 ) satisfies the assumptions of theorem 4.2 , and the lower half - plane ( @xmath175 ) satisfies the assumptions of theorem 6.4 . by similar reasoning as in theorem [ gen ] , the rest of the proof follows .
.1truein real and complex analyticity are both defined in terms of convergent power series .
any problem in the real analytic category may be complexified and thereby turned into a complex analytic one , and any complex analytic situation with real coefficients is obviously real analytic and can thus be treated with real analytic techniques .
the following analytic implicit function theorem may be interpreted in either the real or complex setting .
[ kp ] suppose that the power series @xmath176 is absolutely convergent for @xmath177 , @xmath178 .
if @xmath179 and @xmath180 , then there exist @xmath181 and a power series @xmath182 such that ( [ f ] ) is absolutely convergent for @xmath183 and @xmath184 .
[ ana1 ] off the end point @xmath21 , the local maximizer @xmath74 for @xmath185 ( @xmath75 and @xmath76 if inside the v - shaped region ) is an analytic function of the parameters @xmath15 and @xmath16 .
a local maximizer @xmath74 for @xmath73 is a zero for @xmath81 with the additional property that @xmath81 would change sign from positive to negative across @xmath186 .
fix a choice of parameters @xmath187 .
set @xmath188 and @xmath189 .
the function @xmath190 is thus transformed into a function @xmath191 .
it is clear that @xmath190 is analytic for @xmath192 , @xmath193 , and @xmath194 .
recall that @xmath73 bounded continuous , @xmath98 , and @xmath99 implies that @xmath74 can not be @xmath96 or @xmath97 .
it follows that the transformed function @xmath191 has the desired domain of analyticity , and is locally absolutely convergent . as for the coefficients , @xmath195 by construction , and @xmath196 as can be seen from the proof of proposition [ max ] . in more detail , for @xmath197
, @xmath82 is positive for all @xmath100 in @xmath198 . and for @xmath199 , @xmath200 only for @xmath87 , which coincides with @xmath74 only for @xmath201 .
lastly , for @xmath202 , @xmath200 only for @xmath203 and @xmath204 .
three possible situations might occur .
outside the v - shaped region , @xmath73 has a unique local maximizer @xmath74 , with @xmath104 if above or along the upper bounding curve , and @xmath105 if below or along the lower bounding curve . inside the v - shaped region , @xmath73 has two local maximizers @xmath75 and @xmath76 , with @xmath107 .
all the conditions of theorem [ kp ] are satisfied , thus @xmath205 converges for @xmath72 close to @xmath206 .
[ ana2 ] off the phase transition curve , @xmath207 ( @xmath141 or @xmath140 if inside the v - shaped region ) is an analytic function of the parameters @xmath15 and @xmath16 .
it is clear that @xmath185 is analytic for @xmath208 , @xmath193 , and @xmath209 . outside the v - shaped region
, @xmath73 has a unique local maximizer @xmath74 in @xmath198 , which is analytic in @xmath15 and @xmath16 by proposition [ ana1 ] . inside the v - shaped region , @xmath73 has two local maximizers @xmath75 and
@xmath76 , both have values in @xmath198 and are analytic in @xmath15 and @xmath16 by proposition [ ana1 ] . below the phase transition curve , @xmath210 is given by @xmath141 , which coincides with @xmath211 along the lower bounding curve .
above the phase transition curve , @xmath210 is given by @xmath140 , which coincides with @xmath211 along the upper bounding curve .
our claim follows by realizing that compositions of analytic functions are analytic as long as the domains and ranges match up .
[ a ] let @xmath2 be a random graph on @xmath0 vertices in one of our models .
the limiting free energy density @xmath212 is an analytic function of the parameters @xmath15 and @xmath16 off the phase transition curve in the upper half - plane @xmath156 .
the assumptions of theorems 4.2 and 6.1 in @xcite are satisfied for parameter values @xmath156 .
our claim then follows from proposition [ ana2 ] .
[ b ] assume that in one of our models @xmath5 is a @xmath25-star ( @xmath67 ) .
the limiting free energy density @xmath212 is an analytic function of the parameters @xmath15 and @xmath16 off the phase transition curve .
the assumptions of theorems 4.2 and 6.4 in @xcite are satisfied .
our claim again follows from proposition [ ana2 ] .
[ ls ] let @xmath213 ^ 2\rightarrow [ 0,1]$ ] be two symmetric integrable functions . then for every finite simple graph @xmath214 , @xmath215 .1truein _ proof of theorems [ one ] and [ two ]
_ the stated analyticity is proven in theorems [ a ] and [ b ] , so we only need to examine the situation along the phase transition curve .
we know from theorems [ gen ] and [ p ] that @xmath216 converges in probability to @xmath74 , off the curve . by lemma [ ls ] , @xmath217 then converges in probability to @xmath218 .
as @xmath217 is uniformly bounded in @xmath0 , this implies that @xmath219 therefore @xmath220 similarly , @xmath221 by corollary [ q ] , these two first derivatives @xmath222 and @xmath223 are discontinuous across the curve ( except at the end point ) .
let us now take a closer look at the behavior of @xmath224 at the critical point .
recall that @xmath225 is monotonically decreasing on @xmath69 $ ] , and the unique zero is achieved at @xmath114 .
take any @xmath226 .
set @xmath227 .
consider @xmath72 so close to @xmath21 such that @xmath228 .
for every @xmath100 in @xmath69 $ ] , we then have @xmath229 .
it follows that the zeros @xmath230 ( @xmath75 and @xmath76 if inside the v - shaped region ) must satisfy @xmath231 , which easily implies the continuity of @xmath232 and @xmath233 at @xmath21 . to see that the transition at the critical point is second - order , we check the second derivatives of @xmath234 in its neighborhood . off the phase transition curve , @xmath235 @xmath236 @xmath237 but as was explained in proposition [ max ] , @xmath238 converges to zero as @xmath72 approaches @xmath239 ; the desired singularity is thus justified .
much of the literature on phase transitions in exponential random graph models uses techniques , such as mean - field approximations , which are mathematically uncontrolled . as such they have been useful in discovering interesting behavior but they can be misleading in detail .
for instance , although phase transitions have been discovered in this way for the two - star ( @xmath5 a two - star ) @xcite and edge - triangle ( @xmath5 a triangle ) models @xcite , the approximation leads to an error in the qualitative nature of the transition , attributing phase coexistence to the full v - shaped region of figure [ v - shape ] rather than just the curve @xmath240 ; in other words it does not distinguish the local maxima in the region from the global maxima .
chatterjee and diaconis @xcite gave the first rigorous proof of singular behavior in an exponential random graph model , the edge - triangle model .
our paper is an extension of this important first step ; besides extending the models and parameters under control we have provided a mathematical framework of ` phases ' which we hope will be useful in motivating future mathematical work in this subject .
our results show that all models with ` attraction ' ( @xmath241 ) , and also @xmath25-star models with ` repulsion ' ( @xmath242 ) exhibit a transition qualitatively like the gas / liquid transition : a first order transition corresponding to a discontinuity in density , with a second order criticial point . in theorem 7.1 of @xcite chatterjee and
diaconis suggest that , quite generally , models with repulsion exhibit a transition qualitatively like the solid / fluid transition , in which one phase has nontrivial structure , as distinguished from the ` disordered ' erds - rnyi graphs , which have independent edges .
we have not yet been able to extend our results to this regime , which is an important open problem .
it is a pleasure to acknowledge useful discussions with persi diaconis and sourav chatterjee introducing us to the subject of random graphs and giving us early access to , and explanations of , their preprint @xcite , as well as suggestions for improving early drafts of this paper .
cr also acknowledges support under nsf grant dms-0700120 . | we derive the full phase diagram for a large family of exponential random graph models , each containing a first order transition curve ending in a critical point . .4truein |
understanding how the solar corona is heated to high temperatures is one of the most important problems in solar physics . in principle
, observations of emission from the solar corona should reveal important characteristics of the coronal heating mechanism , such as the time scale and location of the energy release . in practice , however , relating solar observations to physical processes in the corona has proved to be very difficult .
one of the more significant obstacles is the complexity of the solar atmosphere , which makes it difficult to isolate and study individual loops , particularly in the cores of solar active regions .
one circumstance in which the line of sight confusion can be reduced is observations of coronal `` moss . ''
the moss is the bright , reticulated pattern observed in many euv images of solar active regions .
these regions are the footpoints of high temperature active region loops , and are a potentially rich source of information on the conditions in high temperature coronal loops ( see @xcite ) .
a particularly useful property of the moss is that its intensity is proportional to the total pressure in the coronal loop @xcite , which is an important constraint for coronal loop modeling . for steady heating models
the pressure and the loop length completely determine the solution up to a filling factor ( see , @xcite ) .
recently @xcite used this property of the moss in conjunction with a magnetic field extrapolation to infer the volumetric heating rate for each field line in the core of an active region .
this allowed them to simulate both the soft x - ray and euv emission independent of any assumptions about the relationship between the volumetric heating rate and the magnetic field .
one limitation of this approach is that the filling factor is left as a free parameter and is adjusted so that the soft x - ray and euv emission best match the observations .
observations of both the intensity and the electron density in the moss would determine the filling factor and remove this degree of freedom .
a more fundemental question is how relevant steady heating models are to coronal heating .
recent work has shown that high temperature coronal emission ( @xmath0mk ) can be modeled successfully with steady heating ( e.g. , @xcite ) . however , pervasive active region doppler shifts ( e.g. , @xcite ) and the importance of non - equilibrium effects in lower temperature loops ( e.g. , @xcite ) , suggest an important role for dynamical heating in the solar corona . the purpose of this paper is to address these two issues with new observations from the extreme ultraviolet imaging spectrometer ( eis ) on _ hinode_. eis has an unprecedented combination of high spatial , spectral , and temporal resolution and provides a unique view of the solar corona . here
we use eis density sensitive line ratios to infer the density and pressure of the moss and determine the filling factor of high temperature coronal loops .
we also use eis measurements of several emission lines formed near 1mk to test the consistency of steady heating models with moss observations .
in this section we discuss the application of steady heating models to observations of the moss with eis . to begin we calculate the plasma emissivites as a function of both temperature and density for ions relevant to eis using the chianti atomic physics database ( e.g. , * ? ? ?
we use the ionization fractions of @xcite and the coronal abundances of @xcite .
the plasma emissivities are related to the observed intensity by the usual expression @xmath1 where @xmath2 and @xmath3 are the electron density and temperature and @xmath4 is a coordinate along the field line . note that what we refer to as the emissivity ( @xmath5 ) is actually the emissivity divided by @xmath6 . for many strong emission lines
this quantity is largely independent of the density . in this context , however , including the density dependence explicitly is important since many of the eis emission lines have emissivities that are sensitive to the assumed density .
our calculations cover the range @xmath713 and @xmath88 . to explore the variation of the observed emission with base pressure we have calculated a family of solutions to the hydrodynamic loop equations using a numerical code written by aad van ballegooijen ( e.g. , @xcite ) .
we consider total loop lengths in the range @xmath9100 mm and maximum temperatures in the range @xmath107.5mk , which are typical of active region cores .
the loops are assumed to be oriented perpendicular to the solar surface and to have constant cross sections . for consistency with our emissivity calculations we use the more recent radiative loss calculations given in @xcite .
each solution to the loop equations gives the density and temperature as a function of position along the loop ( @xmath11 ) .
we interpolate to find the emissivity from the calculated density and temperature in each computational cell .
we then integrate the emissivity at heights below 5 mm to determine the total footpoint intensity using equation [ eq : ints ] .
the resulting intensities are displayed in figure [ fig : calc ] and are generally consistent with earlier work which showed that the moss intensities are proportional to the pressure and independent of the loop length ( e.g. , * ? ? ?
* ; * ? ? ?
the linear relationship between the intensity and the pressure breaks down for some of the hotter lines for the lowest pressures we ve considered . for these solutions the peak temperature of formation for the line is closer to the apex temperature and there is some emission from along the loop leg . for each line we perform a fit of the form @xmath12 to the calculated intensities .
only pressures above @xmath13@xmath14 k are considered in the fits .
the exponent , which is also shown in figure [ fig : calc ] , is generally close to 1 .
one approximation that has been made in earlier theoretical work that is not strictly valid is that the emissivity is independent of the density .
our calculations account for this and are more accurate .
these model calculations suggest the following recipe for deriving the loop pressure and filling factor from the observations . the 186.88/195.12 ratio can be used to infer the pressure .
the pressure then can be used to determine the expected intensities in each of the lines .
the filling factor is derived from the ratio of the simulated to observed intensity .
to test these model calculations we use observations from eis on _ hinode _ , which was launched 23 september 2006 .
the eis instrument produces high resolution solar spectra in the wavelength ranges of 170210 and 250290 .
the instrument has 1 spatial pixels and 0.0223 spectral pixels .
further details are given in @xcite and @xcite . despite the fact that _ hinode _
was launched close to the minimum in the solar activity cycle there have been several active regions available for observation . for this work we analyzed _ hinode _ observations of noaa active region 10940 from 2007 february 2 at about 10:42 ut .
these observations show areas of moss that are relatively free from contamination from other emission in the active region . the observing sequence for this period consisted of stepping the 1 slit over a @xmath15 region taking 15s exposures at each position .
these data were processed to remove the contribution of ccd background ( pedestal and dark current ) , electron spikes , and hot pixels , and converted to physical units using standard software .
we then fit the calibrated data assuming gaussian profiles and a constant background .
the resulting rasters for some of the strongest lines are shown in figure [ fig : rast ] . to identify the moss in this region
we adopt the very simple strategy of looking for the brightest 186.88 emission .
this line is very sensitive to density .
the line intensity relative to 195.12 rises by about a factor of 6 as the densities rises from @xmath16 , which is typical of active region loops at this temperature , to @xmath17@xmath14 , which is at the high end of what we expect to observe in the moss .
the result of using a simple intensity threshold is shown in figure [ fig : rast ] .
for each of the 1416 spatial pixels identified as moss we use the 186.88/195.12 ratio to determine the base pressure in the loop from the calculation shown in figure [ fig : calc ] .
the base pressure is then used to infer the intensities in all of the emission lines , again using the calculations of this type also shown in figure [ fig : calc ] . as has been found in many previous studies ,
the simulated intensities are much higher than what is observed . to reconcile the observed and simulated intensities we introduce a filling factor for each point derived from the 195.12 intensities . as shown in figure [ fig : filling ] , the distribution of filling factors is approximately gaussian with a peak at about 16% and standard deviation of about 4% .
the mean filling factor we derive here is generally consistent with previous results that have been about 10% ( e.g. , * ? ? ?
* ; * ? ? ?
we also find that the derived filling factor is nonuniform and inversely proportional to the base pressure . to compare the model with the observation we have made scatter plots of observed intensity versus simulated intensity ( including the filling factor ) for a number of emission lines .
these plots are shown in figure [ fig : compare ] and indicate a reasonable agreement between the steady heating model and the observations .
the agreement is particularly good for 188.23 and 203.82 .
the simulated 184.54 intensities are about 20% too high and the simulated 202.04 are about 10% too low .
these discrepancies are not alarming considering the uncertainties in the atomic data .
there are , however , rather signficant discrepancies in the 275.35 intensities .
the simulated intensities are generally about a factor of 4 larger than what is predicted from the steady heating model .
the correlation between the simulated and observed intensities is also poor .
it is possible that considering loop expansion or nonuniform heating may provide a better match at these temperatures , but we have not yet performed these simulations .
it is also possible that there are errors in the atomic data for this line or that the instrumental calibration is not correct at this wavelength .
more extensive analysis , including the observation of other lines formed at similar temperatures , will be required to resolve this issue .
we have presented an initial analysis of active region moss observed with the eis instrument on _ hinode_. we find that the intensities predicted by steady , uniformly heated loop models are too high and a filling factor is required to bring the simulated intensities into agreement with the observations .
the mean filling factor we derive here ( @xmath18% ) is similar to that determined from earlier work with steady heating models .
furthermore , we also find that the filling factor in the moss must be nonuniform and varies inversely with the loop pressure .
the next step is to use the methodology outlined here to simulate all of the active region emission ( including the high temperature loops ) and compare with eis observations .
finally , we emphasize that while we find reasonable agreement between the moss intensities and steady heating , we can not rule out dynamical heating processes , such as nano - flares ( e.g. , * ? ? ?
* ; * ? ? ?
the x - ray emission observed in this region clearly has a dynamic component @xcite and it would be surprising if non - equilibrium effects did not play some role in the heating of this active region .
hinode is a japanese mission developed and launched by isas/ jaxa , with naoj as domestic partner and nasa and stfc ( uk ) as international partners .
it is operated by these agencies in co - operation with esa and nsc ( norway ) . | observations of transition region emission in solar active regions represent a powerful tool for determining the properties of hot coronal loops . in this letter we present the analysis of new observations of active region moss taken with the extreme ultraviolet imaging spectrometer ( eis ) on the _ hinode _ mission .
we find that the intensities predicted by steady , uniformly heated loop models are too intense relative to the observations , consistent with previous work . to bring the model into agreement with the observations a filling factor of about 16%
is required .
furthermore , our analysis indicates that the filling factor in the moss is nonuniform and varies inversely with the loop pressure . |
the aim of this work is to study the dynamics of the gravitational instability , _ i.e. _ , the evolution of scalar density perturbations , in presence of viscous effects .
viscosity can be divided in two different kinds : _ bulk viscosity _ ( bv ) @xmath0 and _ shear viscosity _ @xmath4 . in this paper , we will concentrate on the analysis of isotropic ( or almost quasi - isotropic ) cosmological models . in this respect , we can safely neglect shear viscosity since there is no displacement of matter layers wrt each other ( in the zeroth - order motion ) and this kind of viscosity represents the energy dissipation due to this feature . indeed , in presence of small inhomogeneities , such effects should be taken into account but we focus on the study of scalar density perturbations and volume changes of a given mass scale are essentially involved .
we expect that the non - equilibrium dynamics of matter compression and rarefaction is more relevant than frictions , thus , we concentrate on bv only since it outcomes from the difficulty for a thermodynamical system to follow the equilibrium configuration . in this paper ,
we assume bv as function of the state parameters of the fluid following the line of the fundamental analysis due to the _ landau school _ @xcite .
in particular , we implement the so - called _ hydrodynamical description _ of the fluid , _ i.e. _ , an arbitrary state is consistently characterized by the particle - flow vector and the energy momentum tensor alone and viscosity is fixed by the macroscopic fluid parameters . in the homogeneous models
@xmath0 depends only on time and we assume it as a power - law of the density of the fluid @xmath1 , _ i.e. _ , @xmath5 with @xmath3 , @xmath6
the study of the density - perturbation dynamics during the matter - dominated era can be consistently described using the newtonian picture , as soon as sub - horizon - sized scales are treated .
the fundamental result is the _ jeans mass _ @xcite ( @xmath7 ) , which is the threshold value for the fluctuation masses to condense , generating a real structure .
if masses greater than @xmath7 are addressed , the density contrast ( @xmath8 ) diverges ( in time ) giving rise to the gravitational collapse . in the following
, we brief discuss how bv affects such a scheme .
the starting point is the eulerian set of motion equations on which a perturbative theory is developed ( the background is assumed static and uniform ) .
if we introduce bv effects in the first - order analysis , the density - contrast growth results to be dumped by viscosity , suppressing the structure formation _ without changing the threshold value of @xmath7_. in particular , the grater @xmath3 is , the slower @xmath8 diverges in time and , furthermore , viscosity generates a decreasing exponential regime and a damped oscillatory one in place of the standard pure oscillatory behavior @xcite . in this respect
, we now analyze the _ top - down mechanism _ of sub - structure formation , _
i.e. _ , the comparison of one collapsing agglomerate with @xmath9 and an internal non - collapsing sub - structure with @xmath10 . the sub - structure mass must be compared with a decreasing @xmath7 since the latter is inversely proportional to the collapsing agglomerate background mass .
as soon as such value reaches the sub - structure mass , this begins to condense . in the standard jeans model , such mechanism
is always allowed since the amplitude for perturbations characterized by @xmath10 remains constant in time but , the presence of decreasing @xmath8 in the viscous model , requires a discussion on the effective damping and on the efficacy of the top - down scheme . as shown in figure 1 , in correspondence of a very small viscosity coefficient ( case a ) , we can show how the sub - structure survives in the oscillatory regime during the background collapse and the fragmentation occurs . on the other hand , if bv is strong enough ( case b ) , the damping becomes very strong and the sub - structure vanishes during the agglomerate evolution .
thus , the top - down mechanism results to be deeply unfavored . a b moreover ,
if the static and uniform background solution is corrected for the expansion of the universe @xcite , a jeans - like relation and a considerable damping of the density contrast growth can be found .
in 1963 @xcite , e.m . lifshitz and i.m . khalatnikov firstly proposed the quasi - isotropic solution ( qis ) based on the idea that , as a function of time , the _ 3_-metric @xmath11 ( where @xmath12 , @xmath13=1 , 2 , 3 ) is expandable in powers of @xmath14 , _
i.e. _ , a taylor expansion is addressed : @xmath15 where @xmath16 ( space contraction ) and @xmath17 ( consistence of the perturbation scheme ) .
we now focus on the relevance of dealing with bv properties of the cosmological fluid approaching the big - bang singularity @xcite .
as far as we characterize the bv coefficient like @xmath18 , it is easy to realize that the choice @xmath19 prevents dominating viscous effects . on the other hand ,
simple considerations indicate that the case @xmath20 leads to negligible contributions of bv in the asymptotic regime towards the big - bang . as a consequence , in studying the singularity physics ,
the most appropriate form of the power - law is @xmath21 and our aim is to determine the conditions on the parameter @xmath3 ( _ i.e. _ , on the viscosity intensity ) which allows for the existence of a qis for the radiation dominated universe . to this purpose , we separate zeroth- and first - order terms into the 3-metric tensor and the whole analysis of einstein equations follows this scheme of approximation . in the search for a self - consistent solution , we make use of the hydrodynamical equations in view of fixing the form of the energy density . of course , in our solution the power - law for the leading _
3_-metric term is sensitive to @xmath3 . as a result , by guaranteeing the consistence of the model , we find that the qis exists if and only if @xmath3 remains smaller than a certain critical value , _
i.e. _ , @xmath22 ( here @xmath23 ) .
in fact , for values of the viscous parameter greater than @xmath3 , the perturbative expansion towards the singularity can not be addressed , since fluctuations would grow more rapidly than zeroth - order terms .
it is worth noting that the study of the perturbation dynamics in a pure isotropic picture yields a very similar asymptotic behavior when viscous effects are taken into account @xcite .
the friedmann scheme is preserved only if we deal with limited values of the viscosity parameter , obtaining the condition @xmath24 .
this smaller constraint is physically motivated if we consider , _ as it is _ , the friedmann model a particular case of the qis .
finally , analyzing @xmath8 , we confirm and generalize the result obtained in @xcite about the damping of density perturbations by the viscous correction . | * abstract : * this paper focuses on the analysis of the gravitational instability in presence of bulk viscosity both in newtonian regime and in the fully - relativistic approach . the standard jeans mechanism and the quasi - isotropic solution are treated expressing the bulk - viscosity coefficient @xmath0 as a power - law of the fluid energy density @xmath1 , _
i.e. _ , @xmath2 . in the newtonian regime ,
the perturbation evolution is founded to be damped by viscosity and the top - down mechanism of structure fragmentation is suppressed .
the value of the jeans mass remains unchanged also in presence of viscosity . in the relativistic approach
, we get a power - law solution existing only in correspondence to a restricted domain of @xmath3 . ' '' '' ' '' '' |
following the seminal work of markowitz @xcite , the problem of optimization of an investor s portfolio based on different criteria and market assumptions are being studied by several authors . in the mean - variance optimization approach ,
as done by markowitz , either the expected value of the portfolio wealth is optimized by keeping the variance fixed or the variance is minimized by keeping the expectation fixed .
though markowitz s mean - variance approach to portfolio is immensely useful in practice , its scope is limited by the fact that only gaussian distributions are completely determined by their first two moments . in a pioneering work merton @xcite , @xcite has introduced the utility maximization to the optimal portfolio selection .
merton s approach is based on applying the method of stochastic optimal control via an appropriate hamilton - jacobi- bellman ( hjb ) equation .
the corresponding optimal dynamic portfolio allocation can also be obtained from the same equation .
although this approach has greater mathematical tractability but does not capture the tradeoff between maximizing expectation and minimizing variance of portfolio value . there is another approach , namely risk sensitive optimization , where a tradeoff between the long run expected growth rate and the asymptotic variance is captured in an implicit way .
the aforesaid utility maximization method can be employed to study the risk - sensitive optimization by choosing a parametric family of exponential utility functions . in such optimization ,
an appropriate value of the parameter is to be chosen by the investor depending on the investors degree of risk tolerance .
we refer @xcite , @xcite , @xcite , @xcite for this criterion under the geometric brownian motion ( gbm ) market model .
risk sensitive optimization of portfolio value in a more general type of market is also studied by various authors .
jump diffusion model is one such generalization , which captures the discontinuity of asset dynamics .
empirical results support such models @xcite .
terminal utility optimization problem under such a model assumption is studied in @xcite . in all these references
, it is assumed that the market parameters , i.e. , the coefficients in the asset price dynamics , are either constant or deterministic functions of time .
we study a class of models where these parameters are allowed to be finite state pure jump processes .
we call each state of the coefficients as a regime and the dynamics as a regime switching model
. the regime switching can be of various types .
it is known that for a markov switching model , the sojourn or holding times in each state are distributed as exponential random variables , whereas the holding time can be any positive random variable for the semi - markov case .
thus the class of semi - markov processes subsumes the class of markov chains .
there are some statistical results in the literature ( see @xcite and the references therein ) , which emphasize the advantage of the applicability of semi - markov switching models over simple homogeneous markov switching models .
it is mainly useful to deal with the impact of a changing environment , which exhibits duration dependence . to understand this , consider a market situation where ,
if the volatility of certain stock price remains low for longer than certain duration , then that observation discourages increasingly more traders to trade on that depending on the length of the duration . in that case
, this type of duration dependence mass - trading behaviour might cause further low volume trading resulting in lack of volatility boost . in this type of market behaviour
, the density function of holding time of low volatility regime should exhibit heavier tail than exponential .
it is important to note that , a markov chain either time homogeneous or in - homogeneous does not exhibit such age dependent transition , whereas a generic semi - markov process may exhibit this phenomenon .
this motivate us to consider age - dependent transition of the regimes .
risk sensitive portfolio optimization in a gbm model with markov regimes is studied in @xcite whereas @xcite studies the same problem in a semi - markov modulated gbm model . in @xcite the market parameters ,
@xmath0 , @xmath1 and @xmath2 are driven by a finite - state semi - markov process @xmath3 , where @xmath1 and @xmath2 denote the drift and volatility parameters of @xmath4-th asset in the portfolio . strictly speaking ,
the assumption that all the parameters from different assets are governed by a single semi - markov process is rather restrictive .
ideally those could be driven by independent or correlated processes in practice .
although two independent markov processes jointly becomes a markov process , the same phenomena is not true for semi - markov processes .
for this reason the case of independent regimes are important where regimes are not markov .
in general a pure jump process need not be a semi - markov process .
in particular the class of age - dependent processes ( as in @xcite ) is much wider than the type of age independent semi - markov processes studied in @xcite . in a recent paper @xcite
, option pricing is studied in a switching market where the regimes are assumed to be an age - dependent process .
an age - dependent process @xmath5 on a finite state space @xmath6 is specified by its instantaneous transition rate @xmath7 , which is a collection of measurable functions @xmath8 where @xmath9 and @xmath10 .
indeed , embedding @xmath11 in @xmath12 , an age dependent process , @xmath13 on @xmath11 is defined as the strong solution to the following system of stochastic integral equations @xmath14}{\displaystyle\int\limits}_{\mathbb{r } } h_{\lambda}(x_{u- } , y_{u-},z)\wp(du , dz),\\ y_t & = y_0 + t- { \displaystyle\int\limits}_{(0,t ] } { \displaystyle\int\limits}_{\mathbb{r } } g_{\lambda}(x_{u- } , y_{u-},z ) \wp(du , dz ) , \end{array}\right\}\ ] ] where @xmath15 is the poisson random measure with intensity @xmath16 , independent of @xmath17 , and @xmath18 and for every @xmath19 , @xmath20 , using a strict total order @xmath21 on @xmath22 .
we refer to @xcite for a proof that @xmath7 indeed represents instantaneous transition rate of @xmath13 . in this paper
we consider a regime switching jump diffusion model of a financial market , where an observed euclidean space valued pure jump process drives the regimes of every asset .
further , we assume that every component of that pure jump process is an age dependent semi - markov process and the components are independent .
we study the finite horizon portfolio optimization via the risk sensitive criterion under the above market assumption .
the optimization problem is solved by studying the corresponding hjb equation where we employ the technique of separation of variables to reduce the hjb equation to a system of linear first order pdes containing some non - local terms . in the reduced equation ,
the nature of non - locality is such that the standard theory of integro - pde is not applicable to establish the existence and uniqueness of the solution . in this paper , to show well - posedness of this pde , a volterra integral equation(ie ) of the second kind is obtained and then the existence of a unique @xmath23 solution is shown . then it is proved that the solution to the ie is a classical solution to the pde under study .
the uniqueness of the pde is proved by showing that any classical solution also solves the ie . in the uniqueness part , we use conditioning with respect to the transition times of the underlying process .
besides , we also obtain the optimal portfolio selection as a continuous function of time and underlying switching process .
the expression of this function does not involve the functional parameter @xmath7 .
thus the optimal selection is robust .
our approach of solving the pde also enables us to develop a robust numerical procedure to compute the optimal portfolio wealth using a quadrature method .
the rest of the paper is organized as follows . in the next section
we give a rigorous description of the model of a financial market dynamics and then derive the wealth process of an investor s portfolio .
the problem of optimizing the portfolio wealth under the risk sensitive criterion on the finite time horizon is also stated in section 2 . in section 3
we have established a characterization of the optimal wealth using the corresponding hamilton - jacobi - bellman equation .
an optimal portfolio strategy is also shown to exist in the class of markov feedback control . furthermore , an optimal feedback control is produced as a minimizer of a certain functional associated with the hjb equation .
we illustrate the theoretical results by performing numerical experiments with an example and obtain some relevant results in section 4 .
section 5 contains some concluding remarks .
the proofs of certain important lemmata are given in an appendix .
let @xmath11 denote a finite subset @xmath24 .
without loss of generality , we choose @xmath25 .
consider for each @xmath26 , @xmath27 a continuously differentiable function with @xmath28 and @xmath29 assume that for each @xmath30 , @xmath31 denotes a finite borel measure on @xmath24 .
let @xmath32\times \mathcal{x}^{n+1 } \rightarrow [ 0,\infty)$ ] , @xmath33\times \mathcal{x}^{n+1 } \rightarrow \mathbb{r}$ ] , and @xmath34\times \mathcal{x}^{n+1 } \rightarrow ( 0,\infty)^{1\times m_1}$ ] be continuous functions of the time variable for each @xmath35 , where @xmath36 and @xmath37 are positive integers .
we also consider a collection of measurable functions @xmath38 for each @xmath39 , @xmath40 .
we further introduce some more notations .
fix @xmath41 and @xmath42 $ ] and we denote @xmath43_{1 \times n}$ ] , and @xmath44_{n \times m_1}$ ] , where @xmath45 is the @xmath46-th component of @xmath2 function . for each @xmath47 , we denote @xmath48_{n \times m_2}$ ] .
we use @xmath49^*$ ] to denote transpose of a vector .
let @xmath50 be a complete probability space on which @xmath51 @xmath11 valued random variables @xmath52 , @xmath51 non negative random variables @xmath53 , a standard @xmath36-dimensional brownian motion @xmath54^{*}\}_{t \ge 0}$ ] , @xmath37 poisson random measures @xmath55 on @xmath56 with intensity @xmath57 , for @xmath58 , and another set of poisson random measures @xmath59 on @xmath56 with intensities @xmath60 defined such that all the random variables , processes and measures are independent .
we denote the compensated measures by @xmath61 for @xmath58 and @xmath62 for @xmath63 . for each @xmath64 , let @xmath65 be the solution to with @xmath66 replaced by @xmath67 , @xmath68 by @xmath69 , @xmath70 by @xmath71 and @xmath72 by @xmath73 . in other words
@xmath74}{\displaystyle\int\limits}_{\mathbb{r } } h^l(x^l_{u- } , y^l_{u-},z_0)\wp^l(du , dz_0)\text{\label{1}}\\ y_t^l & = & y_0^l+t- { \displaystyle\int\limits}_{(0,t ] } { \displaystyle\int\limits}_{\mathbb{r } } g^l(x^l_{u- } , y^l_{u-},z_0 ) \wp^l(du , dz_0)\text{\label{2}},\end{aligned}\ ] ] where @xmath75 and @xmath76 .
we denote the tuple @xmath77 by @xmath78 and @xmath79 by @xmath80 .
hence , @xmath81 , @xmath82 and @xmath13 are independent .
the process @xmath83 is a time homogeneous strong markov process .
let the filtration @xmath84 be the right continuous augmentation of the filtration generated by @xmath85 such that @xmath86 contains all the @xmath87-null sets .
we consider a frictionless market consisting of @xmath51 assets whose prices are denoted by @xmath88 and @xmath89 and are traded continuously .
we model the hypothetical state of the assets at time @xmath90 by the pure jump process @xmath5 .
the state of the asset indicates its mean growth rate and volatility .
we assume @xmath91 thus the corresponding asset is ( locally ) risk free , which refers to the money market account with the floating interest rate @xmath92 at time @xmath90 corresponding to regime @xmath93 .
the other @xmath94 asset prices are assumed to be given by the following stochastic differential equation @xmath95 , \\ s_0^l & = s_l , \quad s_l \geq 0 , ~ l=1,2,\ldots , n. \nonumber\end{aligned}\ ] ] these prices correspond to @xmath94 different risky assets . therefore , @xmath1 represents the growth rate of the @xmath4-th asset and @xmath96 the volatility matrix of the market . here
we further assume the following . * assumptions * : * for each @xmath39 and @xmath40 , we assume @xmath97 . * for each @xmath39 and @xmath40 , we further assume @xmath98 . *
let @xmath99 denote the diffusion matrix .
assume that there exist a @xmath100 such that for each @xmath90 and @xmath93 , @xmath101 , where @xmath102 denotes the euclidean norm .
the next lemma asserts the existence and uniqueness of the solution to the sde ( [ eq : sde ] ) . the proof is deferred to the appendix . [ sdesol ] under the assumption _ * ( a2 ) * _ the equation ( [ eq : sde ] ) has a strong solution , which is adapted , a.s .
unique and an rcll process .
we note that _ * ( a1 ) * _ and _ * ( a2 ) * _ follow for the special case where @xmath103 by _ * ( a3 ) * _ the diffusion matrix @xmath104 is uniformly positive definite , which ensures that @xmath104 is invertible .
we will use this condition in section 3 .
this condition also implies that @xmath105 .
consider an investor who is employing a self - financing portfolio of the above @xmath51 assets starting with a positive wealth .
if the portfolio at time @xmath90 comprises of @xmath106 number of units of @xmath107 asset for every @xmath108 , then for each @xmath109 the value of the portfolio at time @xmath90 is given by @xmath110 we allow @xmath106 be real valued , i.e. , borrowing from the money market and short selling of assets are allowed .
we further assume that @xmath111 is a @xmath112 adapted , rcll process for each @xmath4 .
then the self - financing condition implies that @xmath113 if @xmath106 are such that @xmath114 remains positive , we can set @xmath115 , the fraction of investment in the @xmath4-th asset .
then we have @xmath116 and hence @xmath117 .
we call @xmath118^{*}$ ] as the portfolio strategy of risky assets at time @xmath90
. then the wealth process , @xmath119 , now onward denoted by @xmath120 , takes the form @xmath121 thus we would consider the following sde for the value process , @xmath122 u_t^l\right ) dt \\ { \nonumber } & \quad + v_t^{u } \sum\nolimits_{l=1}^{n } \sum\nolimits_{j=1}^{m_1 } \sigma_{lj}(t , x_t ) ~u_t^l dw_t^j \\ { \nonumber } & \quad + v_{t-}^{u } \sum\nolimits_{l=1}^{n } \sum\nolimits_{j=1}^{m_2}u_{t-}^l { \displaystyle\int\limits}_{\mathbb{r } } \eta_{lj}(z_j ) n_j(dt , dz_j ) \\ { \nonumber}&= v_t^{u } ( ~r(t , x_t ) + b(t , x_t)u_t ) dt + v_t^{u } u_t^{*}\sigma(t , x_t)dw_t \nonumber \\ & \quad + v_{t-}^{u } \sum\nolimits_{j=1}^{m_2 } \int_{\mathbb{r } } \left[u_{t-}^ { * } \eta(z)\right]_j n_j(dt , dz_j),\end{aligned}\ ] ] where @xmath123_{1 \times m_2}$ ] .
note that , some additional assumptions on @xmath124 are needed for ensuring a strong solution of .
[ rem1 ] as before , we need to assume that @xmath125 is such that for each @xmath58 , and @xmath126 , @xmath127_j > -1 $ ] to ensure a positive solution to . for some technical reasons we require a stronger condition on @xmath125 .
we would require that the the portfolio should be chosen from @xmath128_j \geq -1+\delta , \forall j , z\ } ~\quad \text{for some}~ 0<\delta\leq 1.\end{aligned}\ ] ] it is clear from the definition and above derivation that @xmath129 , the portfolio wealth process , is a controlled process .
let @xmath130 be a non empty convex set , denoting the range of portfolio .
the range is determined based on investment restrictions .
for example , @xmath131 in the case of unrestricted short selling . the restrictions on short selling makes @xmath132 , where @xmath133 for @xmath134 .
clearly , @xmath135 for @xmath134 , correspond to no short selling .
[ defi1 ] an rcll and adapted process @xmath136}$ ] is said to be _
admissible _ portfolio strategy if : * the process @xmath124 takes values from the convex set @xmath137 , where @xmath138 is as in , * has an almost sure unique strong solution , * @xmath139}\|u_t(\omega)\|<\infty$ ] . [ solstate ] under * ( a1 ) * and with admissible control @xmath124 , ( i ) the sde has an almost sure unique positive strong solution , ( ii ) the solution has finite moments of all positive and negative orders , which are also bounded on @xmath140 $ ] uniformly in @xmath124 .
\(i ) we first note that , since @xmath141 and satisfies definition [ defi1](iii ) , @xmath142_j)|<\max\left(|\ln \delta| , c\|\eta_{\cdot j}(z_j)\|\right),\ ] ] where @xmath143}\|u_t(\omega)\|$ ] and @xmath144 is the @xmath46-th column of the matrix @xmath145 . again using * ( a1 ) * and the finiteness of measure @xmath31 , the integration of rhs with respect to @xmath146 has finite expectation .
this implies that @xmath147_j)\,n_j(ds , dz_j)<\infty$ ] .
therefore in the similar line of proof of lemma [ sdesol ] , we can show under the assumption * ( a1 ) * and admissibility of @xmath124 , ( [ eq : pov ] ) has an a.s .
unique positive rcll solution , which is an adapted process , and the solution is given by @xmath148_j)\,n_j(ds , dz_j ) \bigg ] .
\end{aligned}\ ] ] ( ii ) we first consider the first order moment . to prove for each @xmath90 , @xmath149 has a bounded expectation , we first note that the rhs can be written as a product of a log - normal random variable and @xmath150_j)\,n_j(ds , dz_j)\right)$ ] , where both are conditionally independent , given the process @xmath124 .
we further note that the log - normal random variable has bounded parameters on @xmath140 $ ] uniformly in @xmath124 .
therefore it is sufficient to check if @xmath151 , \end{aligned}\ ] ] is bounded on @xmath140 $ ] , for all @xmath152 . using finiteness of @xmath31 and * ( a1 )
* , one can show that the above expectation is bounded .
thus @xmath153 has bounded expectation on @xmath140 $ ] , uniformly in @xmath124 .
now for moments of general order , we note that for any @xmath154 can also be written in a similar form of where each of the integrals inside the exponential would be multiplied by the constant @xmath155 .
thus the rest of the proof follows in a similar line of that of first order case , given above .
our goal is to study risk sensitive optimal control problem on the above wealth process .
we would see in the next section that , in order to obtain a classical solution to the corresponding hjb equation to be defined shortly , certain regularity of the conditional c.d.f of holding time of @xmath78 is needed .
we devote the next subsection to establishing some smoothness of relevant density functions .
we define the function @xmath156 $ ] as @xmath157 and let @xmath158 and for each @xmath159 with @xmath160 for all @xmath161 and @xmath162 .
set @xmath163 we assume further conditions on the transition rate so that the unconditional transition probability matrix is irreducible . +
* assumption : * * ( a4 ) * the matrix @xmath164 is irreducible . from the definition of @xmath165 and the assumptions on @xmath7 , we observe @xmath166 , for all @xmath167 . we also note that @xmath168 hold for all @xmath169 .
now assume that @xmath170 denotes the time of @xmath94-th transition of the @xmath4-th component of @xmath78 , whereas @xmath171 and @xmath172 . for a fixed @xmath90 ,
let @xmath173 .
hence @xmath174 and @xmath175 .
it is shown in @xcite that @xmath176 is the conditional c.d.f of the holding time of @xmath177 and @xmath178 is the conditional probability that @xmath177 transits to @xmath46 given the fact that it is at @xmath161 for a duration of @xmath162 .
let @xmath179 the remaining life of @xmath4-th component i.e. , the time period from time @xmath90 after which the @xmath4-th component of @xmath78 would have the first transition .
note that @xmath180 is independent of every component of @xmath13 other than @xmath4-th one .
we denote the conditional c.d.f of @xmath180 given @xmath181 and @xmath182 as @xmath183 .
it is important to note that this c.d.f does not depend on @xmath90 mainly because @xmath184 is time - homogeneous .
we also notice that @xmath185 is the duration of stagnancy of @xmath186 at present state before it moves to another .
from now we denote @xmath187 by @xmath188 and the corresponding conditional expectation as @xmath189 .
let @xmath190 be the component of @xmath78 , where the subsequent transition happens .
therefore , @xmath191 represents the conditional probability of observing next transition to occur at the @xmath4-th component given that @xmath192 and @xmath193 .
we find the expressions of the c.d.f and the probability defined above and obtain some properties in the following lemma .
the proof is deferred to the appendix . in order to state the lemma , we introduce some notations .
we define an open set @xmath194 and a linear operator @xmath195 where dom(@xmath196 ) , the domain of @xmath196 is the subspace of @xmath197 such that for each @xmath198 dom(@xmath196 ) above limit exists for every @xmath199 and @xmath200 .
[ theo1 ] consider @xmath201 as given above .
+ ( i ) for each @xmath4 , @xmath202 ( ii ) let @xmath203 be the conditional c.d.f of @xmath180 given @xmath204 and @xmath205 .
then @xmath206 and is @xmath207 in @xmath0 variable .
\(iii ) @xmath208 is differentiable with respect to @xmath0 .
\(iv ) @xmath209 and @xmath191 are in _
dom_(@xmath196 ) .
@xmath210 \(v ) @xmath211 .
in this paper we consider a risk sensitive optimization criterion of terminal portfolio wealth corresponding to a portfolio @xmath124 , that is given by @xmath212 \nonumber \\ \nonumber & = -\left(\frac{2}{\theta}\right ) ~\ln\mathbb{e}\left [ ( v_t^u)^{-\frac{\theta}{2 } } ~\biggr|~ x_0=x , y_0=y , v^u_0=v\right],\end{aligned}\ ] ] which is to be maximized over all admissible portfolio strategies with constant risk aversion parameter @xmath213 .
since logarithm is increasing , it suffices to consider the following cost function @xmath214,\ ] ] which is to be minimized . for
all @xmath215 , let @xmath216,\\ { \varphi}_\theta(t , x , y , v ) : = \inf_{u}\tilde{j}_\theta^{u , t}(t , x , y , v ) , \end{array}\right\}\ ] ] where the infimum is taken over all admissible strategies as in definition [ defi1 ] .
hence , @xmath217 represents the optimal cost .
we also define the following class of functions @xmath218 [ defig ] let @xmath219 be such that for every @xmath220 the following hold : * @xmath221 is twice continuously differentiable with respect to @xmath222 for all @xmath223 and @xmath224 is in dom(@xmath196 ) for each @xmath225,@xmath93 , * for fixed @xmath226 , @xmath227 , * for each @xmath228 , @xmath229 is in @xmath230 .
let @xmath136}$ ] be an admissible strategy such that it has the following form @xmath231 for some measurable @xmath232 .
we call such controls as markov feedback control
. then the augmented process @xmath233}$ ] is markov where , @xmath234 are as in , , ( [ eq : pov ] ) .
we note that for any measurable @xmath232 , the equation may not have a strong solution .
however , we will show the existence of a markov feedback control which is optimal and under which has an a.s .
unique strong solution .
let @xmath235 be the infinitesimal generator of @xmath236}$ ] , and @xmath224 be a @xmath237 function with compact support , then we have @xmath238 \frac{\partial } { \partial v } { \varphi}(t , x , y , v ) \nonumber\\ & \quad + \frac{1}{2 } v^2 \left[\tilde u^{*}(t , x , y , v)a(t , x)\tilde u(t , x , y , v ) \right]\frac{\partial^2 } { \partial v^2}{\varphi}(t , x , y , v)\nonumber \\ & \quad + \sum\nolimits_{j=1}^{m_2 } { \displaystyle\int\limits}_{\mathbb{r } } \!\left[{\varphi}\left(t , x , y , v\left(1 + [ \tilde u^{*}(t , x , y , v ) \eta(z)]_j\right)\right ) - { \varphi}(t , x , y , v ) \right]\ , \nu_j(dz_j ) \nonumber \\ & \quad + \sum_{l=0}^{n}\sum_{j\neq x^l}\lambda^l_{x^lj}(y^l)\left[{\varphi}(t , r^l_jx , r^l_0y , v ) - { \varphi}(t , x , y , v)\right],\end{aligned}\ ] ] where the linear operator @xmath239 is given by @xmath240 , @xmath63 , @xmath241 and @xmath242 is the standard basis of @xmath243 . for a given @xmath244 , by abuse of notation , we write @xmath245 , when @xmath246 for all @xmath247 .
we consider the following hjb equation @xmath248 with the terminal condition @xmath249^{n+1 } , \quad v>0.\end{aligned}\ ] ] we now define a classical solution to the problem - .
[ defcl ] we say @xmath250 is a classical solution to - if @xmath220 and for all @xmath251 , @xmath252 satisfies - .
we look for a solution to - of the form @xmath253 where @xmath254 .
clearly the left hand side of is in class @xmath255
. we will establish the following result in first two subsections .
[ theo2 ] the cauchy problem - has a unique classical solution , @xmath256 , of the form .
substitution of ( [ eq : rstransf ] ) into , yields @xmath257 + h_\theta(t , x)\psi(t , x , y)=0,\end{aligned}\ ] ] for each @xmath199 with the condition @xmath258 where the map @xmath259\times \mathcal{x}^{n+1}\rightarrow \mathbb{r}$ ] is given by @xmath260,\ ] ] the infimum of a family of continuous functions @xmath261
+ \frac{1}{2 } \left(-\frac{\theta}{2}\right)\left(-\frac{\theta}{2}-1\right)\left[u^{*}~a(t , x)~u \right ] \nonumber \\ & \quad + \sum\nolimits_{j=1}^{m_2 } { \displaystyle\int\limits}_{\mathbb{r } } \!\left(\left(1 + [ u^ { * } \eta(z)]_j\right)^{\left(-\frac{\theta}{2}\right)}- 1 \right)\ , \nu_j(dz_j).\nonumber\end{aligned}\ ] ] it is important to note that the linear first order equation is nonlocal due to the presence of the term @xmath262 in the equation .
it implies that @xmath263 depends on the value of @xmath264 at the point @xmath265 , which does not lie in the neighbourhood of @xmath266 .
we now define a classical solution to - below .
[ deffcl ] we say @xmath267 is a classical solution to - if @xmath268 dom@xmath269 and for all @xmath199 , @xmath252 satisfies - .
it is interesting to note that other than the terminal condition , no additional boundary conditions are imposed .
the remaining parts of the boundary is @xmath270\}$ ] .
we note from that , @xmath271 , for all @xmath42 $ ] .
hence @xmath272 does not cross the boundary .
thus the pde would have no solution for any boundary condition which is not obtained from the terminal condition .
[ theo3 ] the cauchy problem - has a unique classical solution .
note that theorem [ theo2 ] may be treated as a corollary of theorem [ theo3 ] in view of the substitution ( [ eq : rstransf ] ) and subsequent analysis .
thus it suffices to establish theorem [ theo3 ] .
we establish theorem [ theo3 ] in the subsection [ lfoe ] via a study of an integral equation which is presented in subsection [ tvie ]
. the following result would be useful to establish well - posedness of - .
[ hcontinuityrs ] consider the map @xmath259\times \mathcal{x}^{n+1}\rightarrow \mathbb{r}$ ] , given by , .
then under _ * ( a3 ) * _ , we have * @xmath273 is continuous , negative valued and bounded below ; * @xmath274 is @xmath23 in both @xmath275 and @xmath276 for each @xmath93 ; * for every @xmath277 , there exists a unique @xmath278 such that @xmath279 . and
@xmath280\times \mathcal{x}^{n+1}\to \mathbb{a}_1 $ ] is continuous in @xmath90 ; * @xmath281 is admissible .
\(i ) we recall that , @xmath282 , the range of portfolio includes the origin .
therefore @xmath283 thus @xmath273 is negative valued . by the continuity assumptions on @xmath284 and @xmath285 , for fixed @xmath124 and each @xmath286 , @xmath92 , @xmath287 and @xmath104 are bounded on @xmath140 $ ] .
let @xmath288 be such that @xmath289}\{|r(t , x)| , ||b(t , x)|| , ||a(t , x)||\ } \leq m.\ ] ] we also observe that for each @xmath290 , @xmath291_j\right)^{-\frac{\theta}{2}}-1)\ , \nu_j(dz_j ) & \geq & -\sum_j { \displaystyle\int\limits}_{\mathbb{r } } \nu_j(dz_j ) \\ & = & -\sum_j\nu_j(\mathbb{r})>-\infty , \end{aligned}\ ] ] using the finiteness of the measure @xmath31 .
also , * ( a3 ) * gives @xmath292 hence by using the above mentioned bounds , we can write , @xmath293 , where @xmath294 since @xmath295 is independent of @xmath90 and @xmath296 as @xmath297 , @xmath298 is bounded below . now we will show that for fixed @xmath90 and @xmath93 , @xmath299 is a strictly convex function of variable @xmath244 .
for fixed @xmath90 and @xmath93 , let @xmath300 denote the hessian matrix for @xmath301 .
then @xmath302-th element of @xmath300 , @xmath303_j\right)^{-\frac{\theta}{2}-2}\ , \nu_j(dz_j).\end{aligned}\ ] ] since @xmath124 is in @xmath304 , @xmath305_j$ ] is bounded below by a positive @xmath306 .
hence , in addition to that using * ( a3 ) * , there exists @xmath307 such that @xmath308 is a positive definite matrix and this proves the strict convexity of @xmath299 on variable @xmath124 .
therefore @xmath309\right)$ ] is a non - empty convex compact set .
hence , @xmath310 is a compact - valued correspondence . since @xmath273 is negative , from ,
we can write @xmath311 we also note that @xmath312 is jointly continuous .
since @xmath310 is continuous , then it follows from the maximum theorem ( @xcite , th .
@xmath313 ) that @xmath314 is continuous with respect to @xmath277 .
hence ( i ) is proved .
\(ii ) follows from the continuity of @xmath314 .
\(iii ) the set of minimizers is defined by @xmath315 again by using ( @xcite , th .
@xmath313 ) , @xmath316 is upper semi - continuous .
since @xmath299 is strictly convex in @xmath124 , for each @xmath317 $ ] and @xmath286 there exist only one element in @xmath318 . by abuse of notation
, we denote that element by @xmath318 itself .
since a single - valued upper semi - continuous correspondence is continuous , @xmath318 is a continuous function .
\(iv ) since @xmath319 is continuous in @xmath90 , there exists a positive constant @xmath320 such that @xmath321 for all @xmath42 $ ] , @xmath322 .
thus @xmath323 is bounded .
since @xmath323 does not depend on @xmath225 , the lipschitz conditions of theorem 1.19 of @xcite are satisfied .
again since @xmath323 is bounded , all growth conditions are also satisfied
. therefore definition [ defi1](ii ) is satisfied and this completes the proof . in order to study - we consider the following integral equation with the previous notations @xmath324dr , \end{aligned}\ ] ] for all @xmath325 , where @xmath326 with each component being @xmath327 .
.2 in equation is a volterra integral equation of second kind .
we note that the boundary of @xmath328 has many facets . for @xmath329 , we directly obtain from , @xmath330 . hence no additional terminal conditions are required .
although the values of @xmath264 in facets @xmath270\}$ ] are not directly followed but can be obtained by solving the integral equation on the facets .
[ iesolnrs ] ( i ) the integral equation has a unique solution in @xmath331 , and ( ii ) the solution is in the _ dom_(@xmath196 ) .
\(i ) we first observe that the solution to the integral equation is a fixed point of the operator @xmath332 , where @xmath333dr.\end{aligned}\ ] ] it is easy to check that for each @xmath334 is continuous . now since @xmath335 by proposition [ hcontinuityrs](i ) , @xmath336f_{\tau^l|l}(r|x , y)dr]|\\ & \leq & \sum_{i=1}^{n}p_{t , x , y}(\ell(t)=l)\int_{0}^{t - t}e^{h_\theta(t , t+r , x)}\sum_{j\neq x^l}p^l_{x^lj}(y^l+r)f_{\tau^l|l}(r|x , y)dr\|\psi-\tilde{\psi}\|\\ & < & k_1\|\psi-\tilde{\psi}\|,\end{aligned}\ ] ] where @xmath337 .
since @xmath338 is strictly less than @xmath327 , implies that @xmath339,for all @xmath340 . hence @xmath341 .
therefore , @xmath332 is a contraction .
thus a direct application of banach fixed point theorem ensures the existence and uniqueness of the solution to .
\(ii ) we denote the unique solution by @xmath264 .
next we show that @xmath342 dom@xmath269 . to this end , it is sufficient to show that @xmath343 .
the first term of @xmath344 is in dom@xmath269 , which follows from lemma [ theo1 ] ( iv ) and proposition [ hcontinuityrs ] ( ii ) . now to show that the remaining term @xmath345dr,\end{aligned}\ ] ] is also in the dom(@xmath196 ) for any @xmath346 , we need to check if the following limit @xmath347,\end{aligned}\ ] ] exists and , the limit is continuous in @xmath348 .
if the limit exists , the limiting value is clearly @xmath349 . by a suitable substitution of variables in the integral ,
the expression in the above limit can be rewritten , using , as @xmath350 by lemma [ theo1 ] ( iv ) , @xmath351 is in dom@xmath269 .
thus @xmath352 is bounded on @xmath353 $ ] by a positive constant @xmath354 .
hence by the mean value theorem on @xmath351 , the integrand of the first integral of is uniformly bounded . therefore , using the bounded convergence theorem , the integral converges as @xmath355 .
the second integral of converges as the integrand is continuous at @xmath356 .
now we compute @xmath357\big)dr\\ & & -\sum_{j\neq x^l}p^l_{x^lj}(y^l)\psi(t , r^l_jx , r^l_0y)f_{\tau^l|l}(0|x , y),\end{aligned}\ ] ] using lemma [ theo1 ] ( iii ) . from we know @xmath358 , therefore @xmath349 can be rewritten using as @xmath359\beta_l(t , x , y)-\sum_{j\neq x^l}p^l_{x^lj}(y^l)\psi(t , r^l_jx , r^l_0y)f_{\tau^l|l}(0|x , y).\end{aligned}\ ] ] clearly is in @xmath197 .
hence @xmath360 is in the dom(@xmath196 ) .
hence the right hand side of is in the dom(@xmath196 ) for any @xmath361 .
thus ( ii ) holds .
[ ivpuniqrs ] the unique solution to also solves the initial value problem - .
let @xmath264 be the solutions of the integral equation .
then by substituting @xmath329 in , follows .
using the results from the proof of lemma [ theo1 ] , proposition [ iesolnrs ] , lemma [ theo1](iv ) and , we have @xmath362[1-f_{\tau^l|l}(t - t|x , y)]e^{h_\theta(t , t , x)}\\ & -\sum_{l=0}^{n}p_{t , x , y}(\ell(t)=l)\big[f_{\tau^l|l}(0|x , y)(f_{\tau^l|l}(v|x , y)-1)\big]\\ & \times e^{h_\theta(t , t , x)}-h_\theta(t , x)\sum_{l=0}^{n}p_{t , x , y}(\ell(t)=l)[1-f_{\tau^l|l}(t - t|x , y)\\ & \times e^{h_\theta(t , t , x)}+\sum_{l=0}^{n}\big[\sum_r f_{\tau^r}(0|x^r , y^r)p_{t , x , y}(\ell(t)=l)-f_{\tau^l}(0|x^l , y^l)\big]\beta_l(t , x , y)\\ & + \sum_{l=0}^{n}p_{t , x , y}(\ell(t)=l)\big(-h_\theta(t , x)+f_{\tau^l|l}(0|x , y)\big)\beta_l(t , x , y)\\ & -\sum_{j\neq x^l}p^l_{x^lj}(y^l)\psi(t , r^l_jx , r^l_0y)f_{\tau^l|l}(0|x , y)\big).\end{aligned}\ ] ] using the equality in lemma [ theo1 ] ( v ) , the right hand side of above equation can be rewritten as @xmath363-h_\theta(t , x)\psi(t , x , y)\\ & = -\sum_l\sum_{j\neq x^l}{\lambda}^l_{x^lj}(y^l)\big[\psi(t , r^l_jx , r^l_0y)-\psi(t , x , y)\big]-h_\theta(t , x)\psi(t , x , y).\end{aligned}\ ] ] hence @xmath264 satisfies .
[ classicalrs ] let @xmath264 be a classical solution to - .
then @xmath264 solves the integral equation .
if the pde has a classical solution @xmath264 , then @xmath264 is also in the domain of @xmath364 , where @xmath364 is the infinitesimal generator of @xmath365 .
then we have from @xmath366 consider @xmath367 then by it s formula , @xmath368 where @xmath369 is a local martingale with respect to @xmath370 , the usual filtration generated by @xmath371 .
thus from @xmath372 is a local martingale . from definition of @xmath373 ,
@xmath374}n_t<\|\psi\|e^{\|h_\theta\|t}$ ] a.s .
thus @xmath372 is a martingale .
therefore by using , we obtain @xmath375.\end{aligned}\ ] ] hence using the markov property of @xmath184 and * ( a4 ) * , @xmath376,~\forall ( t , x , y)\in\bar{\mathscr{d}}.\ ] ] by conditioning on the component of @xmath78 where the transition happens , @xmath377}|\ell(t)]]\\ & = & \sum_{l=0}^{n}p_{t , x , y}(\ell(t)=l)\mathbb{e}_{t , x , y}[e^{[\int_{t}^{t}h_\theta(s , x_s)ds]}|\ell(t)=l ] \end{aligned}\ ] ] where @xmath190 is described in subsection 2.4 after * ( a4)*. next by conditioning on @xmath180 we rewrite @xmath378\\ { \nonumber}&=&\mathbb{e}_{t , x , y}[\mathbb{e}_{t , x , y}[e^{\int_{t}^{t}h_\theta(s , x_s)ds}|\ell(t)=l,\tau^l(t)]|\ell(t)=l]\\ { \nonumber}&=&p_{t , x , y}(\tau^l(t)>t - t|\ell(t)=l)e^{\int_{t}^{t } h_\theta(s , x)ds}\\ & & + \int_{0}^{t - t}\mathbb{e}_{t , x , y}[e^{\int_{t}^{t}h_\theta(s , x_s)ds}|\ell(t)=l,\tau^l(t)=r ] f_{\tau^l|l}(r|x , y ) dr . \end{aligned}\ ] ] since @xmath379 is constant on @xmath380 provided @xmath381 , the above expression is equal to @xmath382e^{h_\theta(t , t , x)}+\int_{0}^{t - t}e^{h_\theta(t , t+r , x)}\\ & & \times \mathbb{e}_{t , x , y}[\mathbb{e}_{t , x , y}[e^{\int_{t+r}^{t}h_\theta(s , x_s)ds}|x^l_{t+r},\ell(t)=l,\tau^l = r]|\ell(t)=l,\tau^l = r]f_{\tau^l|l}(r|x , y)dr\\ & = & [ 1-f_{\tau^l|l}(t - t|x , y)]e^{h_\theta(t , t , x)}\\ & & + \int_{0}^{t - t}e^{h_\theta(t , t+r , x)}\times \sum_{j\neq x^l}p^l_{x^lj}(y^l+r)\psi(t+r , r^l_jx , r^l_0(y+r\mathbf{1}))f_{\tau^l|l}(r|x , y)dr .
\end{aligned}\ ] ] from and the above expression , the desired result follows .
the result follows from proposition [ iesolnrs ] , proposition [ ivpuniqrs ] , and proposition [ classicalrs ] .
now we are in a position to derive the expression of optimal portfolio value under risk sensitive criterion .
the optimal value is given by @xmath383 where the existence and uniqueness of the classical solution to - follows from theorem [ theo3 ] .
we note that the study of - becomes much simpler if the coefficients @xmath384 are independent of time @xmath90 . for time
homogeneous case , proposition _ [ hcontinuityrs ] _ is immediate .
furthermore , the proof of theorem _ [ theo3 ] _ does not need the results given in proposition _ [ iesolnrs ] _ , proposition _ [ ivpuniqrs ] _ , and proposition _ [ classicalrs]_. indeed theorem _ [ theo3 ] _ can directly be proved by noting the smoothness of terminal condition .
we conclude this section with a proof of the verification theorem for optimal control problem . the main result is given in theorem [ markov&admissible ] .
[ verificationth ]
let @xmath256 be as in theorem _ [ theo2 ] _
, then * @xmath385 for every admissible markov feedback control @xmath386 . *
let @xmath323 be as in proposition _ [ hcontinuityrs]_(iv ) , then @xmath387 \(i ) consider an admissible markov feedback control @xmath388 , where @xmath389 and @xmath256 , the classical solution to - as in . now by it s formula @xmath390 dr \nonumber\\ & = \displaystyle\sum_{j=1}^{m_1}\int_t^s \frac{\partial}{\partial v}{\varphi}_m(r , x_r , y_r , v^{\bar{u}}_r ) v^{\bar{u}}_r[\tilde{u}(r , x_r , y_r , v_r)^*{\sigma}(r , x_r)]_j dw^j_r\nonumber \\ & + \displaystyle\sum_{j=1}^{m_2}\int_t^s \int_\mathbb{r } \bigg[{\varphi}_m(r , x_r , y_r , v^{\bar{u}}_{r-}(1+[\tilde{u}(r , x_{r-},y_{r-},v_{r-})^*\eta(z)]_j))- { \varphi}_m(r , x_r , y_r , v^{\bar{u}}_{r-})\bigg ] \tilde{n}_j(dr , dz_j)\nonumber\\ & + \sum_{l=0}^{n}\int_t^s \int_\mathbb{r } \bigg[{\varphi}_m(r , r^l_{x^l_{r- } + h^l(x^l_{r-},y^l_{r-},z_0)}(x_{r-}),r^l_{y^l_{r-}-g^l(x^l_{r-},y^l_{r-},z_0)}(y_{r-}),v^{\bar{u}}_{r- } ) \nonumber\\ & -{\varphi}_m(r , x_{r-},y_{r-},v^{\bar{u}}_{r-})\bigg ] \tilde{\wp}^l(dr , dz_0).\end{aligned}\ ] ] we would first show that the right hand side is an @xmath391 martingale .
since @xmath386 is admissible , using definition [ defi1](iii ) , it is sufficient to show , the following square integrability condition @xmath392 ^ 2dr < \infty , \end{aligned}\ ] ] to prove that the first term is a martingale . again since @xmath393 , @xmath394 . thus using the boundedness of @xmath264
the above would follow if @xmath395^{-\theta}dr<\infty\end{aligned}\ ] ] holds .
now we consider the second integral .
rewriting that term , we obtain @xmath396_j)^{-\frac{\theta}{2}}-1\right]\tilde{n}_j(dr , dz_j).\end{aligned}\ ] ] we first observe that @xmath397_j)>\delta$ ] , and this implies @xmath398_j)^{-\frac{\theta}{2}}<\delta^{-\frac{\theta}{2}}.\ ] ] thus the integrand of is a product of a bounded function and @xmath399 .
since @xmath31 , the lvy measure of @xmath400 is a finite measure for each @xmath46 , to show is an @xmath391 martingale , it is enough to verify .
similarly the third integral can be rewritten as @xmath401\tilde{\wp}^l(dr , dz_0).\end{aligned}\ ] ] in the integrand is a product of a bounded function with compact support and @xmath399 . since , the compensator of @xmath402 is @xmath403 , is also an @xmath391 martingale if holds .
thus is the sufficient condition for the right side of to be a martingale .
however readily follows from the lemma [ solstate](ii ) and an application of tonelli s theorem .
taking conditional expectation on both sides of given @xmath404 and letting @xmath405 , we obtain @xmath406-{\varphi}_m(t , x , y , v)\nonumber\\ & = \mathbb{e}\int_t^t\bigg[\mathscr{a}^{\tilde{u}}{\varphi}_m(r , x_r , y_r , v_r^{\bar{u}})\bigg|x_t = x , y_t = y , v^{\bar{u}}_t = v\bigg]dr\geq 0.\end{aligned}\ ] ] the above non - negativity follows , since @xmath256 is the classical solution to - and @xmath407 for all @xmath0 . and implies result ( i ) .
\(ii ) the right hand side of becomes zero by considering @xmath408 and this completes the proof of ( ii ) .
finally we show in the following theorem that @xmath256 as in theorem [ theo2 ] indeed gives the optimal performance under all admissible controls .
[ markov&admissible ] let @xmath256 be as in theorem _ [ theo2 ] _ and @xmath409
. then @xmath410 .
we first note that in the proof of proposition [ verificationth](i ) , we have only used the properties ( ii ) and ( iii ) of definition [ defi1 ] of the markov control .
since these two properties are true for a generic admissible control @xmath124 , we can get as in proposition [ verificationth](i ) .
@xmath411 for every admissible control @xmath124 . by taking infimum
, we get @xmath412 .
the other side of inequality is rather straight forward . using proposition [ verificationth](ii ) and theorem [ hcontinuityrs](iv ) ,
@xmath323 is admissible , and @xmath413 .
thus @xmath414 .
hence the result is proved .
now we establish a characterisation of @xmath256 using the hjb equation in the following proposition .
[ comparison ] let @xmath224 be any classical solutions to - .
let @xmath256 be as in theorem _
[ theo2]_. then @xmath415 , for all @xmath247 .
thus the unique solution @xmath256 obtained in theorem [ theo2 ] is maximal among all classical solution to - .
note that in the proof of proposition [ verificationth](i ) , to show that the right hand side of is a martingale , we have only effectively used the fact that @xmath256 satisfies conditions ( i),(ii ) and ( iii ) of definition [ defig ] .
hence for any @xmath220 and @xmath323 as in proposition [ hcontinuityrs](iv ) , @xmath416 dr,\end{aligned}\ ] ] is an @xmath417 martingale .
taking conditional expectation in , given @xmath418 and letting @xmath405 , we have @xmath419-{\varphi}(t , x , y , v)\nonumber\\ & = \mathbb{e}\int_t^t\bigg[\mathscr{a}^{u^\star}{\varphi}(r , x_r , y_r , v_r^u)\bigg|x_t = x , y_t = y , v^{\bar{u}^\star}_t = v\bigg]dr , \end{aligned}\ ] ] using @xmath420 . now using rhs is nonnegative and theorem [ verificationth](ii ) , we obtain @xmath421 .
we have seen that the optimal portfolio value with risk sensitive criterion is given by ( [ iefinalrs ] ) and - . for illustration purpose , we are considering a simple model in which all the parameters for all assets are governed by a single semi - markov process .
then @xmath422 if @xmath423 , and we denote that value as @xmath424 where @xmath425 and @xmath426 are the first components of @xmath93 , and @xmath427 respectively . hence implies @xmath428 provided @xmath429 and @xmath430 . in other words
@xmath431 depends only on @xmath432 . in view of this
, we may introduce a new function @xmath433 to denote @xmath434 .
therefore gets reduced to @xmath435+\bar{h}_\theta(t , x)\psi(t , x , y)=0 , \end{aligned}\ ] ] for every @xmath436 , @xmath437 , @xmath438 .
we further assume that @xmath439 , i.e. , the portfolio includes a single stock and a money market instrument .
we also specify the state space @xmath440 , i.e. , the semi - markov process has three regimes . the drift coefficient
, volatility and instantaneous interest rate at each regime are chosen as follows : @xmath441 the transition rates for @xmath169 are assumed to be given by @xmath442 where @xmath443 hence the holding time of the first component in each regime has the conditional probability density function @xmath444 and the conditional c.d.f @xmath445 .
we also assumed @xmath446 and @xmath447}(z)}{b - a } dz$ ] .
it is shown separately in @xcite that the classical solution to with @xmath448 , satisfies the following integral equation @xmath449 + { \displaystyle\int\limits}_{0}^{t - t}\exp\left[{\displaystyle\int\limits}_{t}^{t+r}\bar h_\theta(s , x)\,ds\right]\times \nonumber \\ & \quad\sum\nolimits_{j \neq x } p_{xj}(y+r)\bar{\psi}(t+r , j,0)\frac{f(y+r \mid x)}{1-f(y \mid x)}dr,\end{aligned}\ ] ] which also follows from . here we compute @xmath450 by discretization of above integral equation using an implicit step - by - step quadrature method as developed in @xcite .
we take @xmath451 , @xmath452 so @xmath453 .
the discretization is given by @xmath454 therefore from ( [ eq : iers ] ) we get @xmath455 + \delta t \sum_{l=0}^{m}w_m(l)\nonumber \\ & \frac{f(y+l\delta t \mid i)}{1-f(y \mid i ) } \left(\exp\left[h_\theta^{m - l}(i ) - h_\theta^m(i)\right ] \sum_{j \in\mathcal{x } , j \ne i}p_{ij}\psi^{m - l}(j,0)\right),\end{aligned}\ ] ] where @xmath456 are weights , chosen as below @xmath457 and @xmath458 + \frac{1}{2 } \left(-\frac{\theta}{2}\right ) \left(-\frac{\theta}{2}-1\right)u^2\sigma^2(t , i ) \right.\nonumber \\ & \quad \quad \quad \left . -1 + \frac{(1+bu)^{1-\frac{\theta}{2 } } - ( 1+au)^{1-\frac{\theta}{2}}}{u(1-\frac{\theta}{2})(b - a ) } \right].\end{aligned}\ ] ] for a given initial portfolio value @xmath225 , from and we get @xmath459 thus the numerical approximation of risk sensitive optimal wealth is given by - . in proposition [ hcontinuityrs ] we have seen that there exists a unique @xmath244 which gives @xmath460 and that we can find by using any convex optimization technique .
here we have used the interior - point method to find the optimal @xmath124 .
.49 .49 we use above mentioned numerical scheme to compute the risk sensitive optimal wealth function given in ( [ eq : phitildeiter ] ) .
figure [ fig1 ] describes the behaviour of risk sensitive optimal wealth for different values of initial portfolio wealth and maturity . the left side plot in figure [ fig1 ]
show that the optimal wealth is monotonically increasing with the value of initial investment . on the other hand
the right side plot shows monotonically increasing behaviour of the optimal wealth with respect to the the maturity of investment .
figure [ fig : fig2rs ] shows the movement in risk sensitive optimal wealth for different values of risk aversion parameter .
the plot shows the strict diminishing behaviour of risk sensitive optimal wealth for increasing risk aversion parameter value .
in this paper a portfolio optimization problem , without any consumption and transaction cost , where stock prices are modelled by multi dimensional geometric jump diffusion market model with semi - markov modulated coefficients is studied .
we find the expression of optimal wealth for expected terminal utility method with risk sensitive criterion on finite time horizon .
we have studied the existence of classical solution of hjb equation using a probabilistic approach .
we have obtained the implicit expression of optimal portfolio .
it is important to note that , the control is robust in the sense that the optimal control does not depend on the transition function of the regime . we have also implemented a numerical scheme to see the behaviour of solutions with respect to initial portfolio value , maturity and risk of aversion parameter .
the results of the numerical scheme are in agreement with the theory of financial market .
the corresponding problem in infinite horizon is needs further investigation .
this would require appropriate results on large deviation principle for semi - markov processes which need to be carried out .
first we show the uniqueness by assuming that the sde ( [ eq : sde ] ) admits a solution , @xmath461 , say , the stopping time @xmath462 . using it^ o lemma ( theorem 1.16 of @xcite ) for @xmath463
we get , @xmath464 -\frac{1}{2}(s_{s-}^l)^{-2}(s_{s-}^l)^2 a_{ll}(s , x_{s-})ds \\ & \quad + \sum\nolimits_{j=1}^{m_2}{\displaystyle\int\limits}_{\mathbb{r}}\ ! \left[\ln(s_{s-}^l + \eta_{lj}(z_{j})s_{s-}^l ) -\ln(s_{s-}^l)\ , \right]\ , n_j(ds , dz_j ) .
\end{aligned}\ ] ] integrating both sides from 0 to @xmath465 yields , @xmath466 where all the integrals have finite expectations almost surely by using * ( a2)*. @xmath467 \end{aligned}\ ] ] thus any solution to ( [ eq : sde ] ) has the above expression . under * ( a2 ) * , @xmath468 has finite expectation for any finite stopping time @xmath469 .
let @xmath470 .
now if possible , assume @xmath471 . by letting @xmath472 in the above expression
, we obtain that @xmath473 is exponential of a random variable which is finite for almost every @xmath474 . thus @xmath475 .
but for almost every @xmath476 @xmath477 .
hence non - positivity occurred only by jump .
in other words @xmath478 for some @xmath479 .
but that is contrary to the assumption on @xmath145 .
hence @xmath480 .
therefore , @xmath481 a.s . for all @xmath482 and is given by @xmath483 .
\end{aligned}\ ] ] thus by equation ( [ eq : sdesol ] ) , @xmath484 is an adapted and rcll process and is uniquely determined with the initial condition @xmath485 .
hence the solution is unique .
\(i ) one can compute the conditional c.d.f @xmath183 in the following way @xmath487 we also denote the derivative of @xmath488 by @xmath489 , given by @xmath490 from the definition of @xmath491 we have , @xmath492 we also introduce a new variable @xmath493 .
we denote the conditional c.d.f of @xmath494 given @xmath192 and @xmath193 as @xmath495 which is equal to @xmath496 .
it is easy to see that @xmath497 . to compute this probability we use a conditioning on @xmath180 . thus @xmath498\\ { \nonumber } & = & { \displaystyle\int\limits}_0^v p_{t , x , y}(\tau^{-l}(t)>\tau^l(t)|\tau^l(t)=s ) f_{\tau^l}(s|x^l , y^l)ds\\ { \nonumber } & = & { \displaystyle\int\limits}_0^v(1-p_{t , x , y}(\tau^{-l}(t)\leq s))f_{\tau^l}(s|x^l , y^l)ds\\ & = & { \displaystyle\int\limits}_0^v\prod_{m\neq l}(1-f_{\tau^m}(s|x^m , y^m))f_{\tau^l}(s|x^l , y^l)ds.\end{aligned}\ ] ] again , @xmath499 and from , we have ( i ) .
\(iv ) in order to show that @xmath191 and @xmath209 belong to @xmath196 we introduce a new function @xmath504 and @xmath505 . consider another function @xmath506 we note that @xmath507 is the derivative of @xmath508 with respect to @xmath225 and it is continuous .
now we show that @xmath508 is @xmath23 in @xmath427 . to this end
we first show the existence of the following limit @xmath509.\end{aligned}\ ] ] by a suitable substitution of variable , the expression in the above limit is @xmath510.\end{aligned}\ ] ] using the above expression converges to @xmath511 as @xmath355 and the limit is continuous in @xmath427 .
thus @xmath512 if @xmath225 is a differentiable function of @xmath90 , then @xmath513 hence @xmath514 since @xmath515 it follows from lemma [ theo1 ] ( i ) , ( ii ) and the above notations @xmath516 and @xmath517 .
hence @xmath191 and @xmath209 are in the dom(@xmath196 ) . now operating @xmath196 on @xmath191 and using , we have @xmath518 operating @xmath196 on @xmath209 @xmath519 this completes the proof of ( iv ) .
* acknowledgement : * the authors are grateful to mrinal k. ghosh and anup biswas for very useful discussions .
the authors also do acknowledge the referee and the associate editor for their insightful suggestions which have helped to improve the quality of this work .
we are also thankful to the editor for encouraging us to revise the initial manuscript .
ghosh m. k. , goswami a. , and kumar , s. , _ portfolio optimization in a semi - markov modulated market _
, appl math optim , 60(2009 ) , 275 - 296 .
ghosh m. k. , goswami , a. , and kumar , s. , _ portfolio optimization in a markov modulated market _ ,
modern trends in controlled stochastic processes , 181 - 195 , 2010 .
ghosh m. k. and saha , s. , stochastic processes with age - dependent transition rates , stoch .
29(2011 ) , 511 - 522 .
goswami , a. , patel , j. and shevgaonkar , p. , a system of non - parabolic pde and application to option pricing , stoch .
34(2016 ) , 893 - 905 .
hunt j. and devolder p. , semi - markov regime switching interest rate models and minimal entropy measure , physica a : statistical mechanics and its applications 390 , 15(2011 ) , 3767 - 3781 .
kallsen jan , optimal portfolios for exponential lvy processes . math .
methods oper .
51 ( 2000 ) , 357 - 374 . | this article studies a portfolio optimization problem , where the market consisting of several stocks is modeled by a multi - dimensional jump diffusion process with age - dependent semi - markov modulated coefficients .
we study risk sensitive portfolio optimization on the finite time horizon .
we study the problem by using a probabilistic approach to establish the existence and uniqueness of the classical solution to the corresponding hamilton - jacobi - bellman ( hjb ) equation .
we also implement a numerical scheme to investigate the behavior of solutions for different values of the initial portfolio wealth , the maturity and the risk of aversion parameter . 1 true cm * key words * portfolio optimization , jump diffusion market model , semi - markov switching , risk sensitive criterion . |
dirichlet branes , discovered by dai , leigh and polchinski in 1989 , have been a central part of the dramatic progress in superstring theory of recent years .
they are the simplest solitons in string theory ; the basic definition of dirichlet brane ( d - brane ) is that it is an allowed boundary condition for a string .
@xcite although they play many physical roles , from a mathematical point of view their most salient feature is that they provide the natural context in which yang - mills theory is embedded in string theory .
this has allowed physicists to rederive and better understand some of the most beautiful mathematical constructions which have been discovered in this area , such as the adhm construction of instantons and nahm s construction of monopoles .
d - branes on calabi - yau ( cy ) manifolds have been the focus of a number of recent works ( see @xcite for additional references ) .
early on it was realized that this theory had a close connection with mirror symmetry and especially kontsevich s homological mirror symmetry proposal @xcite , which interprets mirror symmetry as an equivalence between the derived category of coherent sheaves ( naturally associated to yang - mills on khler manifolds ) and a derived category proposed by fukaya , associated to isotopy classes of lagrangian submanifolds .
these two classes of objects ( usually called `` b '' and `` a''-type branes respectively ) correspond to the two classes of supersymmetry preserving ( or bps ) branes on a cy , and thus the physical understanding of mirror symmetry indeed should imply such an equivalence . in physical terms
, kontsevich s proposal addresses only the structure of topologically twisted open string theory ( we explain this below ) , while one expects more structure to appear in discussing the full open string theory which governs the physics of d - branes .
one approach to this extra structure is to study the additional conditions on the d - brane world - volumes which follow from space - time supersymmetry . in the large volume limit , these two conditions are the hermitian yang - mills equations ( for b branes ) and the special lagrangian condition ( for a branes ) . in this context ,
strominger , yau and zaslow formulated mirror symmetry as the statement that the mirror @xmath0 to a cy @xmath1 could be obtained as the moduli space of a particular a brane mirror to the b brane which is the point on @xmath1 .
@xcite this brane is expected to come in families which give a @xmath2 fibration structure to @xmath1 , a claim which has been proven in many cases by gross and by ruan .
@xcite although this approach clearly captures some essential part of the truth , it should be realized that the specific additional conditions which were used , the hermitian yang - mills and special lagrangian conditions , are not believed to be the correct physical conditions except in the large volume limit .
this is related to the point that in string theory , cy manifolds are not really ricci - flat manifolds except in the large volume limit ; the einstein equations are known to be corrected . indeed ,
the stringy version of cy khler moduli space , which is best thought of as the complex structure moduli space of the mirror , looks quite different from the conventional geometric version .
the correct replacements for either ricci - flatness or the supersymmetry conditions on d - branes are not known at this writing , but some things are known , which we summarize here .
in particular , one has strong reasons to believe that the metric defined by the syz construction ( which is the `` d@xmath3-metric '' of @xcite on the mirror manifold ) is _ not _ ricci - flat .
having cast doubt on the usual mathematical starting point for the discussion of d - branes , we are obliged to say something about what replaces it .
now the fundamental physical definition is quite clear a d - brane is a boundary condition in the world - sheet conformal field theory ( cft ) ; the a and b branes each correspond to boundary conditions preserving a particular half of world - sheet supersymmetry .
this definition has some conceptual advantages ; notably , as in the discussion of greene and plesser @xcite , mirror symmetry is manifest .
the problem with it is that it is rather hard to make precise contact with geometry except in the large volume limit .
however , a number of general results have been obtained , which we outline . in keeping with the cft framework
, we freely use mirror symmetry in our considerations , while regarding the precise connection with geometry as following from subsequent considerations which are not yet completely understood .
an example of a geometrical statement which is on a sound footing in string theory is the statement that the holomorphic properties of b branes are independent of khler moduli , and thus are `` the same '' as in the large volume limit @xcite .
the precise meaning of `` same '' turns out to be subtle however and this is the point where the derived category must enter the discussion .
a recent development is a precise contact with the graded categories which appear in kontsevich s proposal , which shows that ( if we consider b branes ) , as the khler moduli are varied , the gradings undergo a `` flow '' determined by the variation of certain phases associated with the branes ( in a brane language , the phase appearing in the special lagrangian condition ) .
@xcite this puts us in a position to explain perhaps the simplest physical structure carried by d - branes which is not already implicit in @xcite .
this is the concept of `` marginal stability '' which governs the possible variation of the spectrum of bps branes .
this concept reduces to variation of @xmath4-stability for b branes in the large volume limit and based on this analogy , the work of joyce on such transitions for a branes @xcite , and many other considerations , a deformation of stability called @xmath5-stability was proposed in @xcite which governs the spectrum of bps branes on arbitrary cy s in string theory .
this proposal has undergone a number of tests ; we also briefly describe work in progress which attempts to complete the proposal by associating natural @xmath6-structures to regions in khler moduli space .
another front on which physical considerations have made progress is in the explicit construction of these bps branes and their moduli spaces .
a particularly simple class of constructions are the orbifold constructions .
these can be exactly solved for orbifolds of flat space , and provide a physical context which contains the structures found both in orbifold resolution by quiver varieties @xcite and in the study of the generalized mckay correspondence @xcite , including generalizations of beilinson s construction of holomorphic bundles on @xmath7 .
@xcite more recently , it has been found that results from the study of explicit gepner model boundary states @xcite can be rederived starting with a `` landau - ginzburg orbifold theory . ''
@xcite geometrically , this uses an embedding of the cy in a weighted projective space , and describes bundles on cy by using the generalized mckay approach to describe bundles on the ambient space , and then restricting .
one new point which emerges is that the physical construction naturally contains moduli which appear only on restriction .
another striking result from this approach is an explicit prediction for the connection formula which relates the natural bases of periods at the large volume and gepner points .
the most general definition of a dirichlet brane in weakly coupled type string theory is that it is a conformal boundary condition in the world - sheet cft .
this definition is discussed at this level of generality in @xcite .
it is quite analogous to the discussion of branes in a geometrical framework as calibrated submanifolds carrying appropriate vector bundles in the sense that they are auxiliary objects ; one requires some description of the ambient space ( the cy for us ) to get started on their definition . in the most general case , where we take as the `` ambient space '' an abstract cft defined in terms of a hilbert space , hamiltonian and operator product coefficients ,
our starting point is a priori non - geometric : it does not come with any local coordinates or other conventional geometric data . on general grounds , the cft s which are easy to describe in this abstract way are `` highly stringy ''
the conventional equations of physics such as the einstein and yang - mills equations will not be valid ( except in special cases with high symmetry ) but will be drastically modified in string theory ( we will say more about this below ) . in some cases , such as the gepner models , we can argue that they are still connected to solutions of einstein s equations by varying parameters ( this is the `` large volume limit '' ) , and the interesting question is then what aspects of this conventional geometrical interpretation survive the stringy modifications and how do other aspects change .
however , there are other cases such as asymmetric orbifolds @xcite where we have no such large volume limit or other geometrical interpretation at present . in these cases
, the abstract definition of d - brane makes sense , but the identification of d - branes with cycles carrying bundles on some space can not even get started .
although general , these remarks are intended to make the point that many of the statements about branes which are taken as definition or otherwise manifest in the mathematical treatments , actually require proof or at least justification in the physical discussion starting from string theory .
while we do nt want to belabor this point , it should be kept in mind .
we now specialize to the particular case of string compactification on a calabi - yau threefold , for which the cft is a @xmath8 superconformal field theory with @xmath9 .
these have been discussed in @xcite and elsewhere .
the notation @xmath8 means there are two independent @xmath10 superconformal algebras describing left and right - moving excitations on the string world - sheet .
each @xmath10 algebra has two supercharges @xmath11 and @xmath12 which anticommute to produce the hamiltonian and a @xmath13 charge @xmath14 , which provides a grading on the hilbert space .
a distinguishing feature of the @xmath15 algebra is that it admits a topological twisting , a redefinition of the generators after which one of the supercharges can be used as a `` brst charge , '' a differential @xmath16 which squares to zero .
the cohomology of the sum @xmath17 of these operators in the two @xmath10 algebras is finite dimensional and is the hilbert space of the topologically twisted theory .
this hilbert space also parameterizes a space of `` operators '' in the theory which correspond to linearized deformations of the theory ; these always turn out to integrable in the @xmath8 case so these theories come in families with locally smooth moduli spaces .
physicists are particularly interested in these models not because they admit topological twisting but because they lead to string theories with space - time supersymmetry .
seeing this requires discussing a bit more structure , namely the `` bosonized @xmath13 . ''
the main point here is that there are natural `` spectral flow '' operators denoted @xmath18 ( on left and right ) which provide automorphisms between sectors of different @xmath13 charge @xmath19 and @xmath20 .
sectors with integral ( resp .
half - integral ) @xmath13 charge correspond to space - time bosons ( fermions ) and this one - to - one relation guarantees space - time supersymmetry .
the topologically twisted world - sheet theories are important in this context mainly because their observables correspond to certain distinguished amplitudes in this space - time supersymmetric theory .
the case of threefolds with @xmath9 is further distinguished in that the correlator of three operators ( the `` yukawa coupling '' ) is always well - defined ; one can further show that such correlators are the third derivatives of a function on moduli space @xmath21 known in the physics literature as the `` @xmath15 prepotential '' and in the math literature as the `` gromov - witten potential . ''
we will discuss the physical interpretation of this potential below .
one made a choice in this construction of whether to identify @xmath11 or @xmath12 as the brst operator @xmath16 ; it turns out that only the relative choice between the two @xmath10 algebras matters in the final result .
this choice distinguishes `` a '' and `` b '' topologically twisted theories , which although they arise from the same cft , need have little else in common . on the level of abstract cft
, one can also exchange @xmath11 and @xmath12 in one of the @xmath15 algebras to define a `` new '' cft , all of whose physical predictions would be the same , but in which the a and b twisted theories would be exchanged . thus there is no a priori difference between a and b , and this is the sense in which mirror symmetry is manifest in cft .
this formalism arises from several more concrete constructions .
contact with geometry and the large volume limit is most direct in what is called the `` nonlinear sigma model '' definition .
this starts from a calabi - yau threefold with ricci - flat metric @xmath22 , and maps of the string world - sheet into this `` target space . ''
the ricci - flat metric is determined by a choice of complex structure ( let the moduli space of these be @xmath23 ) and complexified khler class ( moduli space @xmath24 ) ; physical arguments strongly suggest that this produces a @xmath8 cft determined by these moduli .
the two gradings provided by left and right @xmath13 charge essentially correspond to the bigrading of dolbeault cohomology , and the b and a twisted theories as above correspond respectively to a theory describing variation of hodge structure ( the yukawa couplings are the second derivatives of the periods of the holomorphic three - form ) and to the `` quantum cohomology theory '' whose prepotential generates the gromov - witten invariants .
mirror symmetry equates the a model on @xmath1 , which sees only @xmath25 , with the b model on its mirror @xmath0 , which only sees @xmath26 .
in particular , the complexified khler cone @xmath24 is only a large volume approximation to the `` true '' stringy khler moduli space @xmath25 , which is best defined simply as @xmath26 .
much work has been done on the physical interpretation of this point .
since there is such a close connection between these topologically twisted theories and the complex geometry , the claim that these originate from physical cft s for which mirror symmetry is manifest places this formulation of mirror symmetry essentially beyond doubt ; various explicit arguments have confirmed this .
@xcite beyond these results which can be convincingly justified in the topologically twisted string theory , what one actually has in this definition is a way to compute general observables as a series expansion in powers of the curvature of the metric @xmath22 multiplied by a distinguished scale of length @xmath27 , the `` string length . ''
one also has prescriptions for adding `` instanton '' corrections associated to non - trivial holomorphic maps from world - sheets of various genus into the cy .
it should be realized that this expansion by itself does not provide a very convincing argument for existence of these models on finite size cy , because such series expansions often do not converge .
however , given evidence from other directions that models do indeed exist , we can interpret them as valid descriptions at least of the asymptotics of these observables for large cy s .
( whenever we talk about `` large '' cy s in the following , we simply mean the case in which the leading terms are good approximations . )
perhaps the most basic observation one can make from these results is that a cy is actually not ricci flat in string theory @xcite .
one traditionally derives ricci - flatness as the condition for `` zero beta function '' required for conformal invariance ; however this is just the leading term and one sees the next correction at order @xmath28 ; the true condition takes the form @xmath29 with the term @xmath30 given in @xcite , and a higher order term shown to be non - zero in @xcite .
it has also been shown that the corrections are such that a solution can be found at each order in the series @xcite . a conceptual argument that ricci - flatness is not to be expected is simply that the structure of the khler moduli space predicted by mirror symmetry is quite different from the geometric khler cone or a naive complexification of this , and we would expect the `` true '' stringy metric to depend naturally on this true moduli space ; the space of ricci - flat metrics does not .
however , it is difficult to go further with this definition , simply because neither the equation ( [ beta - function ] ) nor the metric @xmath31 it describes are canonically defined ; the theory of renormalization which led to it does not distinguish in any convincing way between @xmath31 and any other metric @xmath32 which would be produced by an arbitrary local functional redefinition @xmath33 .
the study of d - branes will provide a better definition of the metric , which we discuss in the next section .
let us turn now to alternatives to the non - linear sigma model .
the most famous of these is the gepner model , also described in the references above .
this starts off from completely non - geometric data , but is believed to be equivalent to the `` analytic continuation '' of the cy non - linear sigma model to a specific highly stringy point in the true khler moduli space .
the fact that such models do exist and are described by the @xmath15 formalism is our best reason to believe that the series expansions described earlier can indeed be summed in some way to define the models under discussion .
the best understanding we have at present of how the gepner model is connected to the geometric picture comes through witten s `` linear sigma model '' definition ( @xcite ; see @xcite for work on d - branes in this framework ) .
one can think of this as defining the cy through a specific embedding in @xmath34 .
this embedding is described in two steps .
first , one constructs a weighted projective space as a symplectic quotient of @xmath34 by a @xmath13 action , in coordinates @xmath35 for chosen weights @xmath36 .
one then defines the cy as a complex hypersurface in this ambient space .
( one can similarly get more general toric varieties , etc . )
the symplectic quotient is the ordinary quotient by @xmath13 of the stable points in the preimage @xmath37 of a moment map @xmath38 defined by @xmath39 to get weighted projective space with weights @xmath36 , one takes @xmath40 for @xmath41 , @xmath42 such that @xmath43 , and @xmath44 .
actually , the quotient is a line bundle @xmath45 over the weighted projective space ; one then imposes the complex equations @xmath46 with @xmath47 , which reduce to @xmath48 and @xmath49 within this zero section .
the important point is now that one can argue that varying the true khler parameter to the gepner point corresponds to an analytic continuation to the `` landau - ginzburg phase , '' the limit @xmath50 . here
the symplectic quotient is rather different : it is the orbifold @xmath51 , and the equations @xmath52 now constrain the coordinates to @xmath53 for @xmath41 .
although on the surface this looks quite different from the original nonlinear sigma model , since the orbifold is birational to the original line bundle , from the point of view of complex geometry the two phases are not so different , and further physical considerations bear this out , leading to a moduli space @xmath54 covering both phases with only isolated singularities , and no immediately evident distinction between the phases .
this discussion motivates the idea that a good starting point for defining cy is orbifold resolution , and indeed an even simpler class of cy s can be defined this way , as the resolution of orbifolds @xmath55 with @xmath56 a discrete subgroup of @xmath57 .
this produces non - compact cy s , which have also been much studied as local models of singularities in cy s , with the general conclusion being that mirror symmetry and almost all of the qualitative features of cy physics are visible in these simpler examples .
the simplest case is @xmath58 , studied in @xcite .
the coarsest classification of boundary conditions in cft is according to the portion of the superconformal algebra they preserve . one way to specify
a boundary condition is in terms of a linear transformation implementing the `` reflection '' from left - moving to right - moving excitations ; this transformation must contain an automorphism of the entire symmetry algebra and thus we need to know these . in the @xmath8 context , there are two basic automorphisms one can use to relate left and right moving algebras .
@xcite the trivial choice @xmath59 and @xmath60 is referred to as a `` b - type '' boundary condition as it is compatible with the b topological twisting .
such boundary conditions can be obtained in the non - linear sigma model by fixing the boundary to sit on a holomorphic submanifold of the cy carrying a holomorphic bundle .
the other basic choice is @xmath61 and @xmath62 or `` a - type , '' which is obtained by fixing the boundary to sit on a lagrangian submanifold , carrying a flat bundle .
this is all the data which is required to specify a boundary condition in the topologically twisted open string theory @xcite , and in this context mirror symmetry can be interpreted as a one - to - one equivalence between the two classes of boundary conditions and all topological correlators .
this is the physical background behind kontsevich s proposal ; however there is much more to say about its physical underpinnings and meaning which has only been addressed recently . from a physical point of view , the most basic question is what physical observables could one compute if one knew something about the derived category of coherent sheaves or its mirror .
this question is not entirely trivial as the categorical framework is quite different from the usual physics language , which assumes that the objects involved come in moduli spaces with explicit coordinatizations , etc .
there are various physical realizations of the bps branes we are talking about , depending on what string theory we use and how many flat dimensions the brane extends in .
all are essentially identical in terms of the underlying string theory ; the primary distinction is between those which can be treated purely classically ( so , objects are solutions of equations of motion derived from some action ) , and those in which quantum effects are taken into account ( so objects are zero energy wave functions or typically harmonic forms on the classical moduli space ) . in any case
, they will have half of the supersymmetry available in the original type theory on cy , because in that theory left and right - movers led to independent supersymmetries , which are now related by the boundary conditions .
two realizations seem to be particularly useful to keep in mind . in the first ,
the physical picture is of open strings free to move in the @xmath63 dimensions not taken up by the calabi - yau ; this leads to a world - volume theory with @xmath64 supersymmetry in @xmath65 dimensions , which we treat classically .
these theories are gauge theories and the basic specification of such a theory is a choice of gauge group ( a semi - simple lie group ) @xmath66 , a khler manifold @xmath67 of `` chiral matter '' configurations admitting a @xmath66-action by isometries , a @xmath66-invariant holomorphic function @xmath68 called `` superpotential , '' and finally moment maps for the @xmath66 action . although all of this data is physical , for the primary question of finding the bps branes and their moduli space ( as a complex variety ) , the data only gets used in the following way : a configuration of the brane system is a solution of @xmath52 in the symplectic quotient @xmath69 .
we refer to the set of these as the moduli space of brane configurations .
since quotient commutes with restriction , one can also think of this as the symplectic @xmath66-quotient of the variety @xmath52 .
the second realization takes the branes to be particles in the extra @xmath65 dimensions .
this description is related to the first as follows : given a moduli space of simple brane configurations ( i.e. with endomorphism group @xmath13 ) ; one identifies particles with zero energy states in a supersymmetric quantum mechanics on this moduli space . by standard arguments
, these will be cohomology classes .
( if the moduli space is singular , it turns out that the quotient / restriction definition gives it a natural embedding in a non - singular space , which can be used to make the definition . )
these are the `` bps branes '' of @xcite etc .
from now on we will consider only the first realization of brane , but use the second to define what we will call `` bps central charge '' and explain the concept of marginal stability .
it can be shown @xcite that in an @xmath15 supersymmetric theory ( such as type on cy ) , the mass of a particle ( in the usual space - time sense ) satisfies a bound @xmath70 with equality attained only for bps particles .
the quantity @xmath71 is the `` bps central charge '' and depends only on the `` charge '' or topological type of the brane and on the cy moduli .
it does not depend on the particular point or cohomology class in the moduli space of brane configurations . for a branes
, @xmath71 is the integral of the holomorphic three - form over the the lagrangian submanifold , and thus depends only on the homology class of the submanifold and the moduli @xmath72 .
the bps bound is the same as that coming from the calibrated geometry of special lagrangian submanifolds .
@xcite for b branes , @xmath71 is defined in terms of the prepotential @xmath21 , a function on @xmath25 , usually computed by invoking mirror symmetry , and depends on the k - theory class of the brane and the moduli @xmath54 . from now on ,
when we speak of `` cy moduli space '' we mean the part ( @xmath72 or @xmath54 ) which appears in these considerations , and `` brane charge '' means the homology or k - theory class .
the most basic physical role of this prepotential is its appearance in ( [ bps - bound ] ) . implicit in this setup is a flat ( gauss - manin ) connection on the cy moduli space , related to the choice of a fixed basis on homology .
this allows transporting a bps brane between points in moduli space and following the variation of its central charge .
naively , this transport allows us to identify the complete spectrum ( list ) of bps branes at each point in moduli space .
however , this is not true , because of the existence of `` lines of marginal stability '' on which the spectrum changes .
there is a strong constraint on these changes : a brane with charge @xmath16 can undergo a decay ( and thus disappear from the spectrum ) into constituents with charges @xmath73 ( satisfying @xmath74 ) only if the phases @xmath75 are equal .
( the normalization factor @xmath76 will be explained later . )
this follows from energy conservation , which implies that @xmath77 , and ( [ bps - bound ] ) , and no other assumptions .
we will return to this point ; for now we stress that the spectrum of bps branes does change with moduli , in both realizations . in the first realization
, it will turn out that this is the direct generalization of `` wall crossing '' phenomena related to the behavior of @xmath4-stability under variations of khler class .
once we know that moduli spaces of brane configurations exist , of course it would be quite interesting to know more about them and in particular to find their khler metrics . in particular , the problem of how to give a canonical definition of the metric of a cy in string theory has a natural answer in this context : it is just the moduli space of the `` d0-brane , '' the b brane which embeds in a point .
this presupposes that such a brane exists and has a moduli space which is the cy , which although manifest at large volume can not be taken for granted on a string - scale cy .
furthermore , there might not be a unique candidate ; one can choose one by using the flat connection of khler moduli space to carry the large volume d0-brane along some path to the point of interest , a prescription which in general is path dependent .
anyways , assuming that such a d0-brane exists , we stress that for each point in the true khler moduli space ( the complex structure moduli space of the mirror ) , there exists a canonical khler metric on the cy , which away from the large volume limit is not expected to be ricci flat ( we will mention some explicit results bearing this out below @xcite ) .
this metric can also be thought of as the metric on the moduli space of the d3-brane wrapped on the @xmath2 fibration of syz mirror symmetry ; again from this point of view , except in special limits , it has no reason to be ricci flat .
( in the explicit computations of @xcite , this would come about after instanton corrections . )
it is not even clear that it has the same khler class as the original metric ; indeed once one goes from @xmath78 to @xmath79 one loses any clear sense of what this would mean .
this metric or even the question of what equations determine it seems rather inaccessible at present and we return to the question of what kontsevich s ideas should help physicists to compute . in some sense , the answer is the space @xmath67 , group @xmath66 and superpotential @xmath80 ; however it has taken some time for physicists to recognize any of this data in kontsevich and fukaya s rather abstract derived categories , and indeed this relation has not been completely spelled out in the literature .
the most basic point is simply that brane configurations can naturally be thought of as objects in a category whose morphisms are the linearized variations of the gauge theory data described above ( tangent vectors to a point in @xmath67 modulo gauge directions ) associated to the system which is the direct sum of two brane subsystems .
the superpotential enters because such variations typically correspond to obstructed deformations ; the non - trivial point seems to be that in cases associated to calabi - yau threefolds this obstruction theory can always be summarized in terms of gradients of a potential .
this point has been recognized in numerous special cases for example , the functional @xmath80 for holomorphic bundles is the holomorphic chern - simons functional , and for holomorphic curves is the functional proposed in @xcite but a mathematical argument as simple and general as the physical argument based on @xmath64 supersymmetry does not seem to have appeared in the literature .
having seen what kontsevich s framework can describe , we also can see in this language what it does not describe : namely , the details of the symplectic quotient which are determined by the specific moment maps . in particular , not all holomorphic objects ( solutions of @xmath52 ) correspond to points in this quotient ; only stable objects do ( in the git sense , which depends on the specific moment maps ) .
thus , the question of characterizing stable bps branes ( in the physical sense ) could be answered by proposing the mathematical stability condition which they satisfy .
one can show by physical arguments that just as the derived category of coherent sheaves does not depend on the khler structure of the cy , the moment maps should only depend on the khler structure and not on the complex structure .
@xcite on large cy s , branes correspond to bundles satisfying the hermitian yang - mills equations , so the stability condition is just the one given by the donaldson and uhlenbeck - yau theorems , which is slope or @xmath4-stability : a bundle @xmath81 ( or coherent sheaf ) is @xmath4-stable if , for all subsheaves @xmath82 , one has @xmath83 with the slope @xmath84 .
this condition depends on both khler and complex structure but in a clearly separated way : @xmath4 depends on khler structure , and the subobject relation on complex structure .
this will be the prototype for our stringy stability condition .
a first step towards such a condition is to consider a deformed hermitian yang - mills equation derived from string theory by marino et .
@xcite as the condition for unbroken supersymmetry from the `` born - infeld action . ''
their equation is a deformation of the yang - mills action which involves higher powers of the yang - mills curvature @xmath21 , and takes the form @xmath85 on a @xmath86-fold with khler form @xmath87 .
considerations of geometric invariant theory in @xcite show that such an equation will have a solution for `` @xmath88-stable '' bundles ; i.e. a stability condition with @xmath89 playing the role of the slope .
although very explicit , this equation only takes into account power - like corrections in @xmath27 , while it is known that further instanton corrections are present .
related to this , it is not clear which khler form one should take for @xmath87 ( the ricci - flat or some stringy version . )
this point could also be addressed by d - brane considerations , but it is not clear that this is the best approach to take as one would prefer to have a condition which depends not on @xmath87 but on the point in @xmath79 , i.e. on the complex structure of the mirror . one can identify the precise quantity which generalizes slope by appealing to mirror symmetry to relate this question to work of joyce on invariants counting special lagrangian manifolds .
@xcite among the many results of this work is an analysis of the local stability of a special lagrangian manifold under variations of the complex structure ; this is determined by an inequality involving ( in physics language ) the phase of the bps central charge ( [ phase - def ] ) .
this fits very well with the known physical considerations on marginal stability , and the sign of the inequality is new information from this point of view .
this brings us to the proposal of @xcite , that the bps branes at a specific point in ( true ) khler moduli space are the `` @xmath5-stable '' branes , i.e. each brane based on a holomorphic object @xmath81 such that for all subobjects @xmath82 one has @xmath90 .
compared to joyce s criterion , the main difference is to use morphisms and subobjects in the definition .
there is also a conceptual difference : instead of working with special lagrangians , we take the stability condition back to the b picture and apply it there .
this has the great practical advantage that the definition of subobject is fairly clear in the b picture ( though we will have to say more about this below ) while computing the morphisms in the a picture is a harder and still not fully understood problem . besides the arguments we cited , in @xcite the proposal
is compared with more explicit results in the large volume and in orbifold limits ( more below ) and finds agreement .
it was studied in the example of the @xmath58 orbifold and its resolution in @xcite , about which more below , and seems to produce sensible results there .
as we mentioned , on physical grounds it is impossible for the spectrum to change except on lines of marginal stability as defined earlier , and it is hard to come up with any competing proposal which satisfies this constraint .
having said this , the proposal has two ill - defined points .
first , one must lift the phases from the interval @xmath91 ( which is what ( [ phase - def ] ) gives us ) to @xmath92 .
the geometry underlying this ( in the a picture ) is explained in @xcite ; in our work we have only postulated this lifting ad hoc .
second , it requires objects to live in an abelian category ( as do all git notions of stability ) , but the only universal category associated to a calabi - yau seems to be the derived category used by kontsevich , which is not abelian .
the two points are connected and we will come back to them below . for completeness , we should say that ( in our opinion ) the assumption that a branes are in fact special lagrangian submanifolds is on less firm a footing than most of the other elements of the picture .
the basic problem is the one that we mentioned , that the appropriate definition of the metric and khler form of the cy in string theory has not in fact been settled , but is almost certainly not the ricci flat metric .
one possibility is that one still has the special lagrangian condition but with respect to a preferred non ricci flat khler form , possibly the d0-metric described above .
some evidence for this idea can be found in the work of leung , yau and zaslow @xcite which shows that the special lagrangian condition is related to the mmms equation ( [ mmms - equation ] ) , in which the metric could be derived from d - brane considerations , by a fourier - mukai transform of the type which should describe t duality in string theory . at present
however this is just a guess , and more physical work is needed to understand this point ; for example it is clear that with some work a series expansion analogous to ( [ beta - function ] ) could be found for the corrections to the special lagrangian condition .
we now turn to a point which emerges quite clearly from physics and cft but does not seem to have appeared in the mathematical discussion , namely a `` flow of gradings '' which is induced under variation of khler moduli ( in the b picture ) or complex structure ( in the a picture ) . consider the b branes for definiteness ; in the large volume limit each brane corresponds to some stable coherent sheaf and thus for each pair of branes @xmath81 and @xmath21 we have graded morphisms @xmath93 .
one often denotes this space as @xmath94)$ ] , a notation which finds its justification in homological algebra .
suppose we now follow a general path in khler moduli space between two points @xmath95 and @xmath96 .
the new claim is that a morphism of degree @xmath97 at the starting point @xmath95 will undergo `` flow of grading '' determined by the phases ( [ phase - def ] ) : its degree at the point @xmath96 will be @xmath98 ( this is the reason for the @xmath76 in ( [ phase - def ] ) , which first appeared in @xcite . ) in particular , the gradings are @xmath92-valued . this rule can be expressed more simply by saying that the grading of objects varies with @xmath99 , so that the starting @xmath100,f[n+\varphi_k(f)])$ ] turns into @xmath101,f[n+\varphi_l(f)])$ ] . in itself this is just a definition but it can be checked physically if we have more than one definition of the underlying cft . as we discuss in the next section , this check has been made between large volume and orbifold points in various models .
the phases are known from mirror symmetry results and flow by integers between these points , and the gradings of morphisms agree with these predictions .
the argument for the flow from cft is quite simple . @xcite
it relies on the identification of a morphism with a `` winding string in the bosonized @xmath13 , '' and has the following intuitive picture .
one can think of each brane as having location @xmath99 on a circle of circumference @xmath102 ( the bosonized @xmath13 ) and a morphism as an open string stretched between the pair of branes ; grading corresponds to length , so the flow is simply induced by motion of the branes on the circle .
once we grant this point , we quickly see that no single abelian category ( such as the category of coherent sheaves ) could possibly describe the branes throughout khler moduli space .
this is because a general path will cause gradings to flow below zero , and convert objects from even to odd grading , both of which lead to violations of the axioms of an abelian category .
however , the derived category and its distinguished triangles still make sense with these flows .
this is perhaps the fundamental reason why one is forced to the derived category in these considerations .
recent work of seidel and thomas and of horja @xcite provides even more concrete motivation for this point ; they show that the monodromies associated to general paths in true khler moduli space act naturally on the derived category , not the category of coherent sheaves . one does not have
a notion of subobject in the derived category and thus no git definition of stability can apply in this context .
the most direct way out of this problem is to propose that each point in khler moduli space comes with a preferred abelian category .
this could be defined by a choice of @xmath6-structure @xcite , which might be determined by the phases as well : one basically wants to keep all objects whose phases lie in the interval @xmath103 $ ] and which are not involved in morphisms of negative degree . this point is presently under investigation
the most accessible physical definitions of calabi - yau manifolds are the orbifolds @xmath55 and the landau - ginzburg orbifolds described in the introduction . as it turns out , the basic constructions involved in defining d - branes in these theories are already fairly standard in mathematics and thus we can be relatively brief . for b branes ,
one clearly wants to take some sort of @xmath56-equivariant bundles on @xmath104 , and this can be implemented in physical terms by starting with the gauge theory lagrangian for branes in @xmath104 and applying a quotient which acts simultaneously on @xmath104 and in the gauge group .
@xcite this leads directly to the quiver theories which appear in the constructions of kronheimer and subsequent work .
@xcite in particular , b branes are objects in a `` mckay '' quiver category , whose primitive objects ( nodes ) correspond to irreps of @xmath56 , whose links correspond to terms in the tensor product with the representation defining the @xmath56 action on @xmath104 , and with specified quadratic relations . the physical application requires @xmath105 and such orbifold singularities can often be resolved ( always for @xmath106 ) to smooth manifolds .
this resolution is visible as the moduli space of a d0-brane and one can even compute the metric ( actually , a controlled approximation in the small blow - up limit ) ; for @xmath55 one obtains explicit non - ricci - flat metrics of the type discussed earlier .
@xcite more generally , the question arises of how b branes as quiver objects or irreps of @xmath56 are related to the large volume definition of b branes as coherent sheaves . in recent mathematical work
@xcite , a generalization of the mckay correspondence has been developed , which in particular provides a natural basis for k theory with compact support ( thus supported on the exceptional divisor ) labelled by irreps of @xmath56 .
the natural conjecture is that this is the physical correspondence we are looking for .
@xcite this conjecture can be tested against results from mirror symmetry which determine the gauss - manin connection on khler moduli space and lead to a connection formula which relates the orbifold `` charge '' basis to the large volume basis .
this gives an explicit expression for the chern character of the bundle corresponding to each irrep , and this has been checked to agree with the conjecture for @xmath58 .
@xcite a somewhat more intricate version of this applies to the landau - ginzburg orbifolds ; one uses the mckay correspondence to get the natural basis for the k theory of the weighted projective space , and then restricts this to the calabi - yau . again
one can compare with results from mirror symmetry and find agreement .
@xcite these constructions are intimately related to and generalize beilinson s construction of quiver categories from sheaves on @xmath7 @xcite , and it is these categories of quiver representations which are compared to the category of coherent sheaves in the test of `` flow of gradings '' mentioned above .
more generally , all this provides a fairly concrete relation between the problems of classifying coherent sheaves and of solving certain @xmath64 supersymmetric gauge theories .
although neither problem is easy in general , the second problem is not only more familiar to physicists but is a more appropriate language for the subsequent considerations than an explicit classification of sheaves would have been .
it would be even more valuable if such constructions could be found for all sheaves , not just those which arise on restriction from the ambient space .
the first step in this generalization already appears in the physical theories
it turns out that these include additional moduli which can be shown to correspond to linearized deformations of bundles which appear after restriction to the cy .
@xcite presumably one could go beyond this linearized level by computing exact superpotentials ; this may be possible using physical methods .
it would be quite interesting to know more about the variation of the spectrum of bps branes with khler moduli .
some simple results of this type were found for @xmath58 in @xcite ; for example it was shown that there are lines of marginal stability arbitrarily close to the large volume limit , by constructing bundles which were @xmath107 away from violating slope stability .
our general conclusion has to be that kontsevich s proposal , coming as it did before the study of d - branes , has turned out to be remarkably prescient .
although it has taken physicists some time to catch up , we are finally extending this picture to provide both more concrete pictures of branes on calabi - yau manifolds and new concepts which should be of interest to mathematicians .
we would like to thank d .-
diaconescu , b. fiol and c. rmelsberger for enjoyable collaborations , and f. bogomolov , m. kontsevich , m. marino , d. morrison , a. polishchuk , p. seidel , and r. p. thomas for valuable discussions .
we thank r. p. thomas for comments on the manuscript . | we give an overview of recent work on dirichlet branes on calabi - yau threefolds which makes contact with kontsevich s homological mirror symmetry proposal , proposes a new definition of stability which is appropriate in string theory , and provides concrete quiver categories equivalent to certain categories of branes on cy . |
in recent years there has been an increasing interest in phase transitions between fluid phases of different electrolyte concentrations in ionic solutions .
two different regimes have been identified experimentally @xcite .
the `` solvophobic '' regime occurs for large solvent dielectric constants that effectively turn off coulombic interactions .
consequently , solvophobic phase transitions are primarily driven by unfavourable interactions between solute and solvent .
this behavior is well described by the usual theory of nonelectrolyte solutions with short range interactions , and clearly leads to ising - like critical behavior . by contrast , in the `` coulombic '' regime the solvent has a low dielectric constant and electrostatic interactions between the solute ions drive the phase separation . in this case , the phase diagrams are quite asymmetric , and apparently mean - field critical behavior has been claimed @xcite , although ising - mean field crossover has also been seen within a narrower range of temperatures around the critical than in non - ionic fluids @xcite .
some of these experimental studies suggest that there exists a new characteristic length in these systems that competes with the correlation length for density fluctuations @xcite .
electrolyte systems in the coulomb regime are often modelled as charged hard spheres embedded in a uniform dielectric continuum ( primitive models ) .
most studied is the `` restricted primitive model '' ( rpm ) , in which the ions are of equal size .
a vapor - liquid phase transition at very low temperatures and densities was predicted theoretically 25 years ago @xcite and by early computer simulation studies @xcite .
improvement of computer simulation techniques has allowed an increasingly precise determination of the coexistence parameters @xcite .
there have also been recent studies of primitive models with asymmetry in size @xcite and charge @xcite .
very recent results suggest that the critical behavior of the rpm belongs to the ising universality class @xcite . from a theoretical point of view
, different approaches have been used in order to explain the vapor - liquid transition of the primitive models .
integral equations such as the mean spherical approximation ( msa ) @xcite , as well as the debye - hckel theory @xcite and poisson - boltzmann approaches @xcite have succesfully been applied to it . for the rpm ,
the most succesful theories are the pairing theories , that consider the ionic fluid as a mixture of bound pairs and free ions in chemical equilibrium @xcite , in the spirit of the bjerrum s ideas @xcite . however , in all these theories the transition is driven by the free ions , even when in same cases , as in the debye - hckel - bjerrum approach , the associated pairs are the dominant species .
analytical @xcite and computer simulation @xcite results , on the other hand , show that the structure of the vapor phase is dominated by neutral clusters , mostly dimers and tetramers .
computer simulations have also demonstrated that the phase envelope of the rpm resembles that of the charged hard dumbbell model @xcite .
this result has been confirmed by theoretical studies @xcite .
the question that the present work examines in detail is the influence of ionic association on the vapor - liquid transition of primitive model electrolytes .
in contrast to an earlier study @xcite , which only considered tightly bound dimers , here we examine a range of models with varying values of @xmath0 , the maximum separation between ions in a dimer .
@xmath2 corresponds to the stillinger - lovett description @xcite of the free ion system .
the structure of this paper is as follows . in section [ sec2 ]
, we examine the microscopic structure of the coexisting phases of the ionic fluid and compare the correlation functions to those of a tightly tethered dimer model . in section [ sec3 ] , an exact chemical representation of the ionic fluid as a mixture of associated pairs and free ions is introduced . the role of the associated pairs on the phase coexistence
is studied in detail in section [ sec4 ] and the paper closes with discussion and conclusions .
in this section we analyse the role of pairing on gas - liquid coexistence of primitive model ionic fluids .
we have studied by computer simulation an 1:1 size - asymmetric primitive model , in which the ions are modelled as hard spheres of diameters @xmath3 and @xmath4 , and carrying charges @xmath5 and @xmath6 , respectively , embedded on a dielectric continuum of dielectric constant @xmath7 ( @xmath8 for the vacuum ) .
the interaction potential between two ions separated by a distance @xmath9 is given by : @xmath10 where @xmath11 is the hard - core potential that takes the value @xmath12 if @xmath13 and 0 otherwise .
the size asymmetry is characterized by the parameter @xmath14 , defined as : @xmath15 monte carlo simulations in the neutral grand - canonical ensemble were performed , characterized by a temperature @xmath16 and the _ configurational _ chemical potential for a pair of unlike ions @xmath17 , with @xmath18 and @xmath19 are the thermal de broglie wavelengths of the ionic species : @xmath20 as usual , cubic boxes of length @xmath21 under periodic boundary conditions are used .
the long - ranged character of the coulombic interactions is handled by the use the ewald summation tecnique with conducting boundary conditions , with 518 fourier - space wavevectors and real - space damping parameter @xmath22 .
the relative error due to the infinite sums truncation in the electrostatic energy is less than @xmath23 for random configurations in small systems @xcite , and this choice has been also validated by direct simulations of the rpm @xcite . in order to speed up the simulations , the basic steps of the simulations ( insertion and deletions of pairs of unlike ions ) are biased following @xcite .
moreover , a fine - discretization approximation is used : the positions available to each ion are the sites of a simple cubic lattice of spacing @xmath24 .
this methodology has succesfully been applied to the rpm @xcite , to 1:1 size - asymetric primitive models @xcite and to @xmath25:1 size - asymetric primitive models @xcite , and allows a speedup relative to the continuum calculations of a factor of 100 for small systems .
the results are almost indistinguishable from the continuum ones for a discretization parameter @xmath26 @xcite , where @xmath27 is the unlike - ion collision diameter .
this value of the discretization parameter ( @xmath28 ) was used in the present study .
histogram reweighting techniques @xcite and mixed - field finite size scaling methods @xcite were used to obtain the vapor - liquid envelopes and the effective critical points , respectively . for the sake of comparison ,
we have also studied tightly tethered dipolar dimer systems @xcite , consisting of @xmath29 pairs of a positive and negative ion restricted to remain at separations @xmath30 satisfying @xmath31 . simulation details and some preliminary results for a range of values of @xmath14 are presented in ref .
@xcite .
the unlike - ion collision diameter @xmath1 provides the basic length scale appropriate for defining both the reduced temperature and reduced density via @xmath32 where @xmath33 is the particle number of each ionic species @xcite .
the reduced simulation box length is defined similarly via @xmath34 , and the reduced energies and chemical potential as @xmath35 and @xmath36 .
the value of the asymmetry parameter considered in this paper is @xmath37 . for that case ,
the gas - liquid coexistence of the ionic fluid shows a shift in both temperature and density respect to the tethered dimer fluid ( see fig .
[ fig1 ] ) , which is a general feature when comparing ionic and tethered dimer systems @xcite . on the other hand
the asymmetry is not high so as to favor large chain - like neutral clusters , as occurs for bigger values of @xmath14 @xcite .
these features qualify this case to be a typical example for moderately asymmetric 1:1 electrolytes , including the rpm . fig .
[ fig1 ] makes the similarity between the phase diagram of the ionic and the tethered dimer fluids clear .
it also suggests in an indirect way that the pairing plays a decisive role on the gas - liquid coexistence in the ionic fluid . in order to clarify such a role ,
we have studied the ion - ion radial distribution functions @xmath38 corresponding to the ionic systems , and the corresponding ones for the tethered dimer systems . for this purpose , we have considered the gas and liquid states at coexistence for a temperature @xmath39 , with @xmath40 the corresponding critical temperature .
the ion - ion radial distribution functions of the ionic systems , for @xmath41 and the coexisting densities @xmath42 and @xmath43 are plotted in figs .
[ fig2 ] and [ fig3 ] , respectively . in both cases , the unlike - ion radial distribution function becomes very large close to contact , indicating the association in bound pairs of unlike ions .
moreover , the like - ion radial distribution functions show maxima around @xmath44 ( in the case of @xmath45 , this peak coincides with the contact value ) , and there is a secondary maximum in @xmath46 for @xmath47 .
these observations allow us to conclude that there are high correlation between pairs of associated unlike ions .
we recall that the range of densities in which the gas - liquid coexistence occurs prevents packing effects , so the structure is completely given by the coulombic interactions .
the ion - ion radial distribution functions for the tethered dimer fluid are plotted on the fig .
[ fig4 ] for the gas branch ( @xmath48 , @xmath49 ) and fig .
[ fig5 ] for the liquid branch ( @xmath48 , @xmath50 ) .
the comparison between the ionic fluid and tethered dimer fluid microscopic structures confirms the qualitative similarity between both systems .
a further test of this similarity is found in the comparison of the neutral cluster populations in the gas branch .
we use gillan s definition of a cluster @xcite . two ions @xmath51 and @xmath52
are directly bound when the distance between them is less than @xmath53 .
this condition defines mathematically an equivalence relationship , and the equivalences classes in which the ions group are the clusters .
as the interaction between like ions is repulsive , the dependence of the cluster definition on @xmath54 or @xmath55 should be very weak ( if they are taken smaller than the mean distance between two like ions ) .
on the other hand , the cluster definition is not very sensitive to the value of @xmath56 if that value lies between the first minimum of @xmath46 and the mean distance between aggregates . in this work
we have used @xmath57 . as expected ,
the microscopic structure of the gas phase is dominated by the @xmath29-ion neutral clusters @xcite .
their fractions @xmath58 for the ionic and tethered ion systems at the gas phase in coexistence for @xmath59 are quite similar ( see fig . [ fig6 ] ) , although the tethered dimer systems have slightly higher fractions of neutral clusters than do ionic systems , specially for large @xmath29 , in agreement with the results previously reported @xcite . despite the similarities found between the ionic fluid and the tethered dimer fluid , there are some differences that can rationalize the quantitative differences between their phase diagrams .
the unlike - ion radial distribution functions of the tethered dimer fluid differ qualitatively from the ionic counterparts close to the contact value , since in the former there is a jump at the maximum allowed separation of two ions of a dimer , while the ionic @xmath46 takes smoothly higher values on that range of values of @xmath60 .
consequently , the condition on the confinement of the ions that compose a tethered dimer must be relaxed in order to refine the pairing concept in the ionic fluids . in the next section
we will introduce an appropriate framework for such a goal .
we consider an ionic fluid ( @xmath61 ) be contained in a volume @xmath62 and in equilibrium with a reservoir at a temperature @xmath16 and chemical potentials @xmath63 and @xmath64 . for simplicity , the neutral grand - canonical ensemble , in which only neutral configurations are allowed , will be considered .
this ensemble has been shown to be equivalent to the usual grand - canonical ensemble in the thermodynamic limit @xcite .
the neutral grand - canonical partition function can be written as : @xmath65 where @xmath66 is the number of ions of each species , @xmath67 , @xmath68 , @xmath69 and @xmath70 are the thermal de broglie wavelengths corresponding to each species , and @xmath71 is the canonical configurational partition function corresponding to a fixed number of ions at the temperature @xmath16 and enclosed in the volumen @xmath62 : @xmath72 with @xmath73 the total potential energy and @xmath74 the hard - core contribution , equal to @xmath12 if two particles overlap , and @xmath75 otherwise ( other tempered potentials can be used , but the main results of this section will remain unchanged ) . if the fluid particles are strongly associated into bound @xmath76 pairs , as occurs in the low - temperature and low - density region in which the vapor - liquid transition occurs for the ionic fluids , the `` physical '' representation described above will be inadequate , and it can replaced by a `` chemical '' picture , in which the fluid is composed by associated pairs and free ions .
first , a rule that unequivocally identifies bound pairs for each configuration of the ionic fluid ( up to a set of null measure in the configurational space ) is needed .
once such a rule is defined , we can write the canonical configurational partition function as : @xmath77 where @xmath78 is the number of bound pairs , @xmath79 is the number of free ions of each species , and @xmath80 is the canonical configurational partition function corresponding to a system of @xmath78 associated pairs and @xmath81 free ions of each species : @xmath82 where @xmath83 are the positions of the ions that compose the associated pair @xmath51 , and @xmath84 corresponds to the coordinates of a @xmath85 free ion @xmath51 , respectively .
the potential energy @xmath86 _ does not _ coincide , in general , with the physical potential energy @xmath87 , since different configurations of the `` chemical '' mixture can be compatible with a given ionic configuration . substituting equation ( [ decompose ] ) into the equation ( [ partition ] ) and rearranging the resulting expression , the neutral grand - canonical partition function can be written as : @xmath88 with @xmath89 and @xmath90 .
these last definitions correspond to the chemical equilibrium conditions between the free ions and the associated pairs @xcite .
we remark that this derivation is _ independent _ of the criterion chosen to define the pairs
. however , the choice must be such that matches the microscopic structure of the fluid .
as we have seen in the previous section , the vapor structure is dominated by neutral aggregates , mostly dimers and tetramers .
this fact is consistent with the bjerrum ideas of pairing @xcite .
consequently , a distance - based criterion seems to be the most appropriate : two unlike ions that are closer than @xmath0 ( @xmath0 being a suitable cutoff distance ) are considered as an associated pair .
it is easy to see that such a criterion _ does not _ define uniquely the associated pairs when the population of tetramers and higher order neutral clusters is not negligible , as it can be seen from fig .
[ fig7 ] , since there are different arrangements of associated pairs that are compatible with the same ionic configuration .
hence , a more systematic criterion is needed , in order to be able to define associated pairs for _ almost every _ ionic configuration , while keeping the intuitive definition of an associated pair as encompassing unlike ions that are at the closest distances .
we use a suitable modification of the stillinger - lovett pair definition on a given ionic configuration @xcite . in this prescription , all the distances between two unlike ions are computed , and then the first pair is _ defined _ as the two unlike ions at closest distance .
this step is repeated , taking into account only ions that remain unpaired from previous steps for the evaluation of @xmath76 distances .
we stop when the minimum distance between two unlike ions is greater than @xmath0 , considering the remaining ions as free ions ( no free ions were considered in @xcite , and consequently @xmath91 in that case ) .
this protocol provides an unique configuration of associated pairs and free ions for almost each ionic configuration , since the method is ambiguous only in a subset of the ionic configurations in which at least two @xmath76 distances are equal , and the measure of such a set is null in the configurational space . obviously , different pairing prescriptions can be used .
however , this protocol provides an explicit expression for the `` chemical '' potential energy @xmath92 : @xmath93 where @xmath87 is the physical potential energy due to the hard - core and eletrostatic interactions , and the other terms correspond to effective interactions of entropic origin needed to reduce the configurational space of the `` chemical '' system .
the term @xmath94 depends only on the positions of the ions that compose an associated pair and confines them to be at a distance shorter than @xmath0 : @xmath95 the pairwise potential energy @xmath96 corresponds to an steric hindrance condition between two associated pairs @xcite , since the distances between two unlike ions corresponding to different pairs can not be shorter than the minimum distance between the ions that compose each pair : @xmath97 with @xmath98 .
the interaction between the associated pairs and the free ions is modified by the term @xmath99 , that prevents the free ion to be closer to the unlike ion of the associated pair than its partner : @xmath100 finally , two unlike free ions can not be at a shorter distance than @xmath0 due to the term @xmath101 : @xmath102 it is not hard to see that a mixture of associated pairs and free ions with a potential energy given by the eq .
( [ hameffec ] ) is completely equivalent to the original ionic system .
it is interesting to note that , except the term on @xmath103 that ties the ions that compose an associated pair , the other terms are pairwise , short - ranged modifications of the hard - core conditions , so the potential energy of an allowed mixture configuration is the same as in the corresponding ionic configuration .
moreover , every configuration of associated pairs and free ions obtained from an ionic configuration by the pairing procedure described above fulfil the constrains induced by the added potentials .
conversely , the mixture configuration is the same as the one obtained from the corresponding ionic configuration by the stillinger - lovett protocol .
this analysis provides an exact chemical representation of the ionic system as a mixture of bound pairs and free ions in chemical equilibrium .
the equivalence between representations of the system is not only at the level of thermodynamic properties , but also in the microscopic structure , unlike previous studies based on the matching of the `` chemical '' and the `` physical '' free energies , e.g. via the virial coefficients @xcite . in those studies the effective potentials given by eqs .
( [ vp ] ) and ( [ v+- ] ) could be guessed , but the other terms ( that arise from matching third and higher order virial coefficients ) were not clearly identified .
furthermore , they have not being used at all in pairing theories of electrolytes .
we must stress the importance of the effective new term given by eq .
( [ vpp ] ) in the potential energy , that reduces the configuration space available to the associated pairs .
this effect is specially important for high values of @xmath0 .
an extreme case case that illustrates the effect of missing the steric hindrance term is the loosely tethered dimer fluid , in which the ions that compose the pair can be at any relative distance greater than @xmath1 . in this case , the canonical partition function of the dimer fluid is @xmath104 , where @xmath78 is the number of dimers , @xmath105 is given by eq .
( [ partition2 ] ) and the @xmath106 factor is the number of different ways of pairing the unlike ions in each ionic configuration and lead to different dimer configurations . as a consequence , in the thermodynamic limit the helmholtz free energy per particle @xmath107 of the loosely tethered dimer
fluid diverges as : @xmath108 where @xmath109 is the ionic helmholtz free energy per ion , that it is well - behaved in the case of neutral systems @xcite .
the association degree of the ionic fluid in the coexistence region has been studied by computer simulation in the framework of the previous analysis . in the usual grand - canonical simulations , we have identified the associated pairs by the stillinger - lovett rule with @xmath110 , i.e. all the ions are associated , and the minimum image convention is used in order to calculate the distances between unlike ions .
the probability distributions @xmath111 of having a pair an internal separation distance @xmath112 between ions is plotted in the fig .
[ fig8 ] for the gas and liquid phases at @xmath113
. surprisingly , both distributions are practically identical , despite the fact that the corresponding densities are quite different .
the distributions show a very pronounced peak for @xmath114 , and a local maximum around @xmath115 , approximately where @xmath46 presents the second local maximum . for large values of @xmath112 , the distributions @xmath111 decay to zero .
these results confirm the strong association in the ionic fluid at low temperatures , and that the structure is weakly affected by variations in density , at least in the range in which the gas - liquid coexistence occurs . for values of @xmath112 close to the contact value
, the distribution is well described by the non - interacting pair fluid probability distribution , given by the following expression : @xmath116\right)\end{aligned}\ ] ] valid for @xmath117 and an internal cutoff @xmath118 . however , for @xmath119 , @xmath111 deviates from the ideal expression as a consequence of the interaction between associated pairs .
the local maximum showed by the distribution function indicates that the most bound pairs are solvated by less bound pairs , to form stable neutral tetramers ( and higher order clusters , but their inclusion does not seem to affect the conclusions of this section ) .
this fact is in agreement with the features observed in the pair correlation functions in the previous section .
we must stress that the structure that @xmath111 presents is only due to the coulombic interactions .
it is instructive to compare our results in the coexistence region to the higher temperature ones .
we have computed @xmath111 at @xmath120 and the same range of densities ( see fig .
[ fig9 ] ) .
the qualitative behavior of @xmath111 at high temperatures is similar to that predicted by stillinger and lovett @xcite .
first , it shows a maximum localized in @xmath121 . on the other hand , @xmath111 takes significant values in a wider range than the corresponding functions at lower temperatures , decaying for large values of @xmath112 as @xmath122 , with @xmath123 .
the latter prediction differs from the value @xmath124 given in ref .
@xcite and requires further study to elucidate it . the probability distribution @xmath125 of having a pair an internal separation distance between ions when the cutoff distance to define a pair is equal to @xmath0 can be obtained from @xmath111 as : @xmath126 where @xmath127 is the cumulant distribution corresponding to @xmath111 .
the density of associated pairs is equal to @xmath128 , and the total density of free ions is @xmath129 $ ] , where @xmath130 is the total density of ions in the physical picture .
it can be seen from fig .
[ fig8 ] that for @xmath131 , more than 95% of ions are associated into pairs in both vapor and liquid phases .
so , if the associated pair fluid phase separate in the same @xmath132 region as the ionic fluid , it is expected than the free ions do play a mere perturbative role in the phase coexistence .
actually , the free ionic subsystem is a non - additive binary charged hard - sphere mixture , where @xmath0 plays the role of the unlike - ion collision diameter .
however , the interactions between like ions are purely repulsive , and it is not expected differences between the behavior of this mixture and the rpm as soon as @xmath3 and @xmath4 are less than @xmath0 . furthermore , this system is in a polar environment given by the associated pair subsystem , and consequently the effective dielectric constant @xmath133 of the background will be increased .
then the free ion subsystem is expected to have the same behavior as the rpm at the effective reduced temperature @xmath134 : @xmath135 where @xmath136 is the reduced temperature . for @xmath137 ,
the critical temperature reduces at least to one third of the rpm reduced critical temperature .
consequently , any possible free ion - driven vapor - liquid transition should happen at much lower temperatures , and thus the free ion subsystem should play no role in the vapor - liquid transition . in order to check this hypothesis , the phase diagram of the associated pair fluid has to be obtained .
this issue will be addressed in the next section .
the results obtained in the previous section suggest that the `` chemical '' picture introduced above is a very convenient description of the ionic fluid structure . furthermore , we can analyse the role played by the associated pairs in the phase equilibrium by eliminating the free ions .
we have performed grand - canonical monte carlo simulations of the associated dimer system for @xmath37 and different values of @xmath0 .
no free ions are allowed ( @xmath138 ) and the potential energy of one configuration is given by the eq .
( [ hameffec ] ) .
the grand - canonical free energies corresponding to these systems will consequently provide an upper bound to the ionic fluid grand - canonical free energy at the same temperature and pair chemical potential .
as for the ionic system , a fine - discretization approach with a refinement parameter @xmath28 , and ewald summation technique with conducting boundary conditions is used to take into account the long range character of the coulombic interations .
the basic steps are either insertion or deletion of associated pairs ( chosen randomly with the same probability ) , biased with a boltzmann distribution that depends on the separation between the ions that compose the dimer .
an associated pair is inserted in the following way : the negative ion is placed at a random place in the box , and its counterion is placed at a relative position @xmath139 ( @xmath140 ) following a probability distribution proportional to @xmath141 , where @xmath142 is the ewald potential energy between two unlike ions of unit charge . on the other hand ,
a pair is deleted with a probability proportional to @xmath143 , being @xmath139 the relative position of the positive ion of the pair with respect to the negative one .
the value of @xmath144 has been adjusted to improve the sampling during the simulation . in this work
we have set @xmath145 .
the acceptance probabilities of a pair insertion @xmath146 and a pair deletion , @xmath147 are the following : @xmath148 \left[\sum_{k=1}^{n_p + 1 } \exp(\beta_0 q^2\phi_0(\mathbf{r}_\pm^k))\right]^{-1 } \right)\\ w_{ij}^{d}&=&\textrm{min}\left(1 , \frac{\exp(-\beta \mu - \beta \delta u)}{(l^*)^3 } \left[\sum_{\sigma_\pm\le |\mathbf{r}_\pm| < r_c}\frac{\exp(\beta_0 q^2\phi_0 ( \mathbf{r}_\pm))}{\zeta^3 } \right]^{-1 } \left[\sum_{k=1}^{n_p } \exp(\beta_0 q^2\phi_0(\mathbf{r}_\pm^k))\right]\right)\end{aligned}\ ] ] where @xmath51 and @xmath52 are the initial and final configurations , respectively , @xmath78 is the number of pairs at the configuration @xmath51 , @xmath67 , @xmath149 is the configurational chemical potential , and @xmath150 is the `` chemical '' potential energy variation in the movement , including the steric hindrance terms given by eq .
( [ vpp ] ) .
as it can be seen , the acceptance probabilities reduce to the usual ones as @xmath151 .
effective critical points for different values of @xmath152 were estimated by using mixed - field finite - size scaling methods @xcite and assuming ising - like criticality .
although recent results @xcite indicate that the pressure should also enter the field mixing , our approach should be satisfactory to discern the dependence of the critical parameters on @xmath0 .
moreover , a systematic study for the rpm case has shown that the assumed universality class is very likely to be the right one @xcite .
we use histogram reweighting techniques @xcite to combine the histograms from different runs ( typically three ) and estimate the critical parameters and their standard errors . very long runs
are needed in order to overcome the critical slowing down and the low acceptance ratios due to the low temperatures involved . defining a step as a try of a pair insertion / deletion
, we have performed simulations of about @xmath153 equilibration steps and @xmath154 sampling steps for @xmath155 ; @xmath156 equilibration steps ans @xmath157 sampling steps for @xmath158 ; and @xmath159 equilibration steps and @xmath160 sampling steps for @xmath161 , in order to get good statistics for the histograms .
the ( effective ) critical parameters are obtained by minimizing the deviation of the appropriately scaled mixed - field @xmath162 marginal probability distribution ( @xmath163 is the potential energy density and @xmath164 is the mixing parameter ) with respect to the corresponding critical 3-dimensional ising universal function @xcite . as also occurs in the ionic and tethered dimer systems , the matching improves as @xmath152 increases ( see fig .
[ fig10 ] ) .
our estimations of the critical parameters for different values of @xmath152 and @xmath165 are listed in table [ tablecritpar ] and plotted in figs .
[ fig11 ] and [ fig12 ] .
the smallest value of @xmath166 is the same as the maximum allowed separation in the tethered dimer system , and as expected the results obtained for the associated pairs fluid are practically indistinguishable from the tethered dimer system . as @xmath166 increases , it is observed a sharp increase in the critical temperature but the critical density remains almost unaffected .
however , for @xmath167 the critical temperature reaches a maximum and then decreases , converging towards the free ion critical temperature . on the other hand
, the critical density also decreases towards the free ion critical density , although the statistical uncertainties are bigger for this parameter and the exact path of convergence is less defined .
we must also point out that the statistical uncertainties are big enough not to make possible an estimation of correction - to - scaling effects .
for an heuristic explanation for such a behavior , we have to consider the energetically favoured configurations .
it is important to consider the interaction between two different associated pairs , that has its minimum energy configuration in a square ring conformation for low values of @xmath14 , with a close - energy secondary minimum conformation for the linear @xmath76 chain @xcite .
if @xmath0 is too small , by deforming the square it is possible to find energetically relevant configurations ( i.e. the potential energy difference per ion of such configurations with respect to the minimum one is of the same order as the thermal energy ) in which one of the associated pairs has an internal ionic separation less than @xmath0 , but the other one does not fulfil such a condition . consequently , a greater value of @xmath0 also increases the number of energetically relevant configurations between two associated pairs .
then the effective interaction between associated pairs is enhanced , and thus an increase of the critical temperature should be expected until all the relevant configurations are allowed ( that occurs for @xmath168 ) . on the other hand ,
the average size of the associated pairs and their aggregates is increased , leading to a decrease on the critical density ( in order to keep the reduced density in terms of the effective particle size constant ) .
these arguments are not longer valid for @xmath169 .
our results show the inclusion of associated pairs with larger internal ionic separations are not crucial to the gas - liquid phase transition but also that they bar the phase transition , as it can inferred from the critical parameters decrease . as for the ionic and tethered dimer models , histogram reweighting techniques @xcite allow us to obtain the coexistence curve up to @xmath170 .
as the finite - size effects are not important far from the critical point , we have considered @xmath158 .
taking advantage of the wide range of densities covered by near - critical histograms , we combine them with liquid subcritical state histograms in order to extend the density range .
the extra simulations involve shorter runs ( typically @xmath159 steps after equilibration ) .
the gas - liquid coexistence curves for different values of @xmath166 are plotted in fig .
[ fig13 ] , showing the same tendency as the observed one in the critical parameters .
it is interesting to note that the vapor branches coincide ( except close to the critical point ) for @xmath171 .
we conclude from this observation that is in the liquid branch where the ionic fluid differs mostly from the associated pair system .
in summary , an exact `` chemical '' representation of the ionic system as a mixture of @xmath76 associated ion pairs and free ions has been introduced .
this representation , closely related to the stillinger - lovett pairing procedure for the rpm , has the advantage that not only provides an exact matching of between the physical and chemical representation thermodynamics , but also from a microscopic point of view .
it also avoids an entropy catastrophe that occurs for the tethered dimer model studied at large values of the tether length .
the addition to the physical hamiltonian of some new pairwise hard - core interactions between the `` chemical '' components provides the suitable hamiltonian for the `` chemical '' representation . in the low temperature and low density regime
in which the ionic vapor - liquid transition occurs , such a representation provides a faithful characterization of the microscopic structure of the ionic fluid , and it can be the basis of new theoretical approaches .
the analysis of the phase behavior of the system only composed by associated pairs indicates strongly that the ionic fluid vapor - liquid transition is driven by them . for @xmath172 , the critical temperature increases with @xmath0 , as more energetically favoured configurations are allowed . for @xmath131 ,
the critical parameters converge smoothly _ from above _ towards the ionic fluid critical parameters as @xmath0 increases .
this fact indicates not only that the free ions do not drive the phase transition , but also have the opposite effect in the transition .
the value of @xmath173 corresponding to the maximum on the critical temperature , can be regarded as the optimal size of the associated pair .
some remarks on the limitations of the present work are appropriate at this point .
we have focussed only on the effect of the associated pairs in the thermodynamical properties .
however , the conducting character of the ionic fluid is completely driven by the free ions . on the other hand , the remarkable success of pairing theories ( even when the `` chemical '' representation they implicitly use is not completely correct ) remains unexplained .
it is possible that the associated pair solvation , in addition to the ion - ion correlations , can mimic the pair - pair interactions .
further studies are needed to completely solve the origin of the ionic fluid vapor - liquid phase transition .
j.m.r .- e . and l.f.r .
gratefully acknowledge financial support for this research by grant no .
pb97 - 0712 from dgicyt ( spain ) and no .
fqm-205 from pai ( junta de andaluca ) .
e . also wishes to thank an fpi scholarship from ministerio de educacin y cultura ( spain ) .
azp acknowledges financial support from the department of energy ( grant de - fg02 - 01er1512 ) and acs - prf ( grant 38165-ac9 ) .
199 m. e. fisher , j. stat .
phys . * 75 * , 1 ( 1994 ) ; j. phys . :
matter * 8 * , 9103 ( 1996 ) .
r. r. singh and k. s. pitzer , j. chem .
phys . * 92 * , 6775 ( 1990 ) .
h. weingrtner , s. wiegand and w. schrer , j. chem . phys .
* 96 * , 848 ( 1992 ) .
k. c. zhang , m. e. briggs , r. w. gammon and j. m. h. levelt sengers , j. chem . phys . * 97 * , 8692 ( 1992 ) . t. narayanan and k. s. pitzer , phys
lett . * 73 * , 3002 ( 1994 ) ; j. phys . chem . * 98 * , 9170 ( 1994 ) .
j. jacob , a. kumar , m. a. anisimov , a. a. povodyrev and j. v. sengers , phys .
e * 58 * , 2188 ( 1998 ) .
m. a. anisimov , j. jacob , a. kumar , v. a. agayan and j. v. sengers , phys .
lett . * 85 * , 2336 ( 2000 ) .
k. gutkowski , m. a. anisimov and j. v. sengers , j. chem . phys .
* 114 * , 3133 ( 2001 ) .
g. stell , k. c. wu and b. larsen , phys .
lett . * 37 * , 1369 ( 1976 ) .
p. n. vorontsov - velyaminov and v. p. chasovskikh , high temp .
( ussr ) * 13 * , 1071 ( 1975 ) .
g. orkoulas and a. z. panagiotopoulos , j. chem . phys .
* 101 * , 1452 ( 1994 ) .
j. m. caillol , d. levesque and j. j. weis , j. chem .
* 107 * , 1565 ( 1997 ) .
q. yan and j. j. de pablo , j. chem .
111 * , 9509 ( 1999 ) .
a. z. panagiotopoulos and s. k. kumar , phys .
lett . * 83 * , 2981 ( 1999 ) .
a. z. panagiotopoulos , j. chem .
phys . * 116 * , 3007 ( 2002 ) .
e. luijten , m. e. fisher and a. z. panagiotopoulos , phys .
lett . * 88 * , 185701 ( 2002 ) .
j. m. romero - enrique , g. orkoulas , a. z. panagiotopoulos and m. e. fisher , phys .
lett . * 85 * , 4558 ( 2000 ) .
q. yan and j. j. de pablo , j. chem .
phys . * 114 * , 1727 ( 2001 ) ; phys .
lett . * 86 * , 2054 ( 2001 ) .
a. z. panagiotopoulos and m. e. fisher , phys .
* 88 * , 045701 ( 2002 ) .
q. yan and j. j. de pablo , phys .
lett . * 88 * , 095504 , ( 2002 ) ; j. chem . phys . * 116 * , 2967 ( 2002 ) . y. levin and m. e. fisher , physica a * 225 * , 164 ( 1996 ) .
e. gonzalez - tovar , mol .
phys . * 97 * , 1203 ( 1999 ) .
e. gonzalez - tovar and c. w. outhwaite , mol
. phys . * 83 * , 1273 ( 1994 ) ; a. k. sabir , l. b. bhuiyan and c. w. outhwaite , molec . phys . * 93 * , 405 ( 1998 ) . f. o. raineri , j. p. routh and g. stell , j. phys .
iv ( france ) * 10 * , 99 ( 2000 ) .
n. bjerrum , kgl .
selsk . mat .-
* 7 * , 1 ( 1926 ) .
m. j. gillan , mol .
49 * , 421 ( 1983 ) .
f. bresme , e. lomba , j. j. weis and j. l. f. abascal , phys .
e * 51 * , 289 ( 1995 ) .
j. c. shelley and g. n. patey , j. chem .
phys . * 103 * , 8299 ( 1995 ) .
y. v. kalyuzhni , mol . phys . * 94 * , 735 ( 1998 )
. j. jiang , l. blum , o. bernard , mol .
phys . * 99 * , 1765 ( 2001 ) ; j. jiang , l. blum , o. bernard , j. m. prausnitz and s. i. sandler , j. chem . phys . * 116 * , 7977 ( 2002 ) . f. h. stillinger and r. lovett , j. chem . phys . *
48 * , 3858 ( 1968 ) .
g. hummer , chem .
. lett . * 235 * , 297 ( 1995 ) .
a. m. ferrenberg and r. h. swendsen , phys .
61 * , 2635 ( 1988 ) ; _ _ i__bid . * 63 * , 1195 ( 1989 ) . n. b. wilding and a. d. bruce , j. phys . : condens .
matter * 4 * , 3087 ( 1992 ) .
e. h. lieb and j. l. lebowitz , adv . math .
* 9 * , 316 ( 1972 ) .
d. chandler and l. r. pratt , j. chem .
* 65 * , 2925 ( 1976 )
. t. l. hill , _
_ s__tatistical mechanics ( mcgraw - hill , new york , 1956 ) , sec .
27 m. e. fisher and d. m. zuckerman , chem .
* 293 * , 461 ( 1998 ) ; m. e. fisher and d. m. zuckerman , j. chem
. phys . * 109 * , 7961 ( 1998 ) . f. h. stillinger and r. lovett , j. chem .
phys . * 48 * , 3869 ( 1968 ) .
m. e. fisher and g. orkoulas , phys .
lett . * 85 * , 696 ( 2000 ) ; g. orkoulas , m. e. fisher and c. stn , j. chem . phys . * 113 * , 7530 ( 2000 ) .
a. z. panagiotopoulos , j. phys .
: condens .
matter * 12 * , r25 ( 2000 ) . .
the size asymmetry parameter is set @xmath37 .
the critical parameters have been obtained from the extrapolation to @xmath175 of the results for @xmath176 and @xmath177 while the subcritical coexistence curves were obtained using @xmath158.[fig1],width=321 ] .
the cation 4 can be paired with either the anion 1 or the anion 2 . in the first case
the ions 2 and 3 remain free , and in the latter case the ions 1 and 3 constitute another associated pair.[fig7],width=321 ] of having a pair an internal separation distance @xmath60 between ions at @xmath113 and reduced densities @xmath179 ( solid line ) , corresponding to the gas phase , and @xmath180 ( dashed line ) , corresponding to the liquid phase .
the dot - dashed line corresponds to the ideal dimer fluid approach ( see text ) . in the inset , the cumulant distribution @xmath181 .
the meaning of the symbols is the same as before .
[ fig8],width=321 ] for the associated pairs fluid characterized by @xmath37 and @xmath185 .
the solid line is the universal function corresponding to the 3-dimensional ising universality class , and the symbols correspond to the best - matching simulation results : the squares correspond to @xmath155 , the diamonds to @xmath158 and the triangles to @xmath161 .
[ fig10],width=321 ] | we present a systematic study of the effect of the ion pairing on the gas - liquid phase transition of hard - core 1:1 electrolyte models . we study a class of dipolar dimer models that depend on a parameter @xmath0 , the maximum separation between the ions that compose the dimer .
this parameter can vary from @xmath1 that corresponds to the tightly tethered dipolar dimer model , to @xmath2 , that corresponds to the stillinger - lovett description of the free ion system .
the coexistence curve and critical point parameters are obtained as a function of @xmath0 by grand canonical monte carlo techniques .
our results show that this dependence is smooth but non - monotonic and converges asymptotically towards the free ion case for relatively small values of @xmath0 .
this fact allows us to describe the gas - liquid transition in the free ion model as a transition between two dimerized fluid phases .
the role of the unpaired ions can be considered as a perturbation of this picture . |
phase locking phenomena are found in a wide variety of nonlinear driven systems in condensed matter physics.@xcite it takes place when an internal frequency of the system locks to a rational multiple of the frequency of an external ac - drive .
a simple example of this is the case of an overdamped particle moving in a tilted washboard potential , where the frequency of motion of the particle over the periodic potential can be locked to multiples of the frequency of a superimposed ac force for a finite range of the dc force ( tilt of the washboard ) . since the internal ( washboard ) frequency is proportional to the mean velocity of the particle , phase locking results in a constant mean velocity for a certain range of dc - force curve when the interference condition is satisfied .
a particularly well known realization of this effect is shapiro steps@xcite in the dc current - voltage ( @xmath0 ) characteristics of a single small area josephson junction driven by a time periodic current . within the washboard analogy outlined above
, a simple analysis provides expressions for the appearance of shapiro steps at specific voltages @xcite corresponding to integer multiples of the driving frequency
. driven systems with many degrees of freedom can also exhibit phase locking .
this has attracted broad scientific and technological interest since phase - locking in complex systems can either be induced by collective effects , providing for a low dimensional interpretation of the phenomenon , or itself induce collective ( low dimensional ) behavior in the complex system .
phase - locking experiments have provided information about such dynamical response of non - equilibrium collective states , where dimensionality , thermal fluctuations , quenched disorder , and the magnitude of external fields can be very relevant .
a particular well known example is the large josephson junction array ( jja ) , with @xmath1 junctions , driven by an external current @xmath2 with frequency @xmath3 and with an applied magnetic field density @xmath4 , where @xmath5 is the magnetic field , @xmath6 the lattice period of the josephson array and @xmath7 the quantum of flux .
_ giant _ shapiro steps at voltages @xmath8 , @xmath9 being an integer , have been observed experimentally in zero magnetic field ( @xmath10 ) .
fractional giant _
shapiro steps at voltages @xmath11 , were observed experimentally @xcite for strongly commensurate magnetic fields , @xmath12 , with @xmath13 integers , and extensively investigated in numerical simulations.@xcite also , _ subharmonic _
giant shapiro steps at voltages @xmath14 were observed experimentally @xcite for zero magnetic field ( @xmath10 ) , and attributed to the nucleation of complex collective dynamical states @xcite induced by disorder or inductance effects .
shapiro - like phase - locking is also observed in the case of driven vortex lattices in bulk superconductors with two - dimensional periodic pinning arrays , as recently reported both experimentally @xcite and theoretically.@xcite also superconductors , where vortices are driven over a one - dimensional potential generated by thickness modulations,@xcite or are confined to move through mesoscopic channels , @xcite show shapiro - like phase - locking .
moreover , systems with many degrees of freedom in the presence of quenched disorder can also exhibit phase - locking when there is a dynamically induced periodicity , like charge density waves @xcite and vortex lattices in superconductors with random pinning.@xcite in the phase - locking examples mentioned above , the ac - drive is applied parallel to the dc - drive .
however , it was recently shown that a different type of phase - locking , distinct from shapiro phase - locking , is possible in vortex lattices if the ac - force is applied _ perpendicular _ to the dc - force.@xcite in this case the interference effect is due to an effective _ parametric _ ac drive in the longitudinal direction , which is induced by the transverse ac drive . in several systems , like charge density waves or single degree of freedom systems ( _ e.g. _ , the single josephson junction ) , the dynamical variables are such that perturbations or displacements can be induced in only one direction ( _ i.e. _ , the displacement field is a scalar ) .
an important characteristic of vortex lattices in superconductors is that the displacement field is two dimensional .
in particular , the behavior of displacements in the direction perpendicular to the driving force shows phenomena like a transverse critical current @xcite and a transverse freezing transition @xcite at high velocities . phase locking in ac - driven vortex lattices , where the ac and dc forces are perpendicular ,
arises as a direct consequence of the nonlinear coupling between the two directions of motions .
transverse phase - locking has been reported for vortex lattices moving in rectangular or triangular pinning arrays @xcite as well as in arrays of randomly distributed pinning centers.@xcite in this paper we investigate the possibility of transverse phase locking in a two dimensional ( 2d ) fully frustrated jja , where the average of the external magnetic field corresponds to one half flux quantum per plaquette , @xmath15 .
this system has several attractive properties .
the presence of a magnetic field ( @xmath16 ) breaks the axial symmetry in the direction of the bias current , and 2d - cooperative behavior may come into play .
this leads to the well - known fractional giant shapiro steps @xcite induced by a longitudinal ac current .
it also , as we will demonstrate in this paper , allows for _ transverse _ phase locking when the ac current is perpendicular to the dc current , since the two directions of motion become coupled .
non - equilibrium dynamical phases for fully frustrated jjas driven by a dc current have previously been studied.@xcite phase locking can be used to characterize temporal order in the different dynamical phases of the jja at high velocities by their ac - response , as was done in ref . for bulk superconductors . here , we report transverse phase locking steps in the @xmath0 characteristics , similar to the longitudinal giant shapiro steps , but with very different characteristic dependencies on external ac - drive amplitude @xmath17 and frequency @xmath3 .
the critical depinning current in the system with transverse ac drive is _ larger _ than the critical current of the dc driven system . for @xmath18 the depinning critical current @xmath19 and the phase - locked step width @xmath20 for @xmath21 increase quadratically with @xmath17 .
for @xmath22 we find windows of @xmath23 where depinning is hysteretic and the periodic phase - locked dynamical states become unstable .
we characterize the dynamical states around the phase - locking interference condition in the @xmath0 curve with the voltage noise , lyapunov exponents and poincar sections .
we find that zero temperature phase - locking behavior in large fully frustrated arrays is well described by an effective four plaquette model .
the remainder of this paper is outlined as follows . in section
ii we present the model used for simulating the dynamics of the fully frustrated jja . in section iii
we develop an analytical framework for predicting critical current and phase - locking properties of the fully frustrated jja .
section iv presents simulated transverse phase - locking steps in typical @xmath0 curves obtained from an effective four plaquette model for the jja .
we calculate numerically the dependence of the critical current and the magnitude of phase - locking steps with the amplitude @xmath17 and frequency @xmath3 of the external ac - drive .
the results obtained are analyzed in more detail by studying voltage noise , lyapunov exponents and poincar sections for the dynamical states around the phase - locking interference condition .
we also compare numerical simulations results for large arrays with those obtained using the effective four plaquette model .
the discussions and conclusions of the investigation are presented in sections iv and v.
we study a current driven jja with an ac current perpendicular to the dc current , as shown in figure [ fig : fig1]a . a magnetic field , @xmath5 ,
is applied such that half a flux quantum , @xmath24 with @xmath25 being the area of a plaquette and @xmath26 being the flux quantum , penetrates each plaquette ; corresponding to the fully frustrated xy model,@xcite where the ground state is a `` checkerboard '' vortex lattice , in which a vortex ( flux quantum ) penetrates every other square grid ( see fig . [
fig : fig1]b ) . in such ground state ,
current and phase differences in the junctions are described by a repeated two - junction by two - junction ( @xmath27 plaquette ) superlattice unit cell .
numerical simulations @xcite of large driven arrays suggest that this spatial periodicity is preserved when the dynamics is phase - locked to an external ac - perturbation applied in parallel to the dc force .
we will show later , in section iv , that this is also a good approximation when the ac current is _ perpendicular _ to the dc current .
we will therefore consider the simple system of a @xmath27 superlattice unit cell of the array and the associated gauge - invariant phase differences in a field of @xmath15 , with the dc current ( per plaquette ) @xmath28 perpendicular to an ac current ( per plaquette ) @xmath29 , as shown in fig .
[ fig : fig2 ] .
flux quantization , total current conservation at the central node , and the applied total currents in the two directions give the following governing equations , @xmath30 where @xmath9 is an integer , @xmath31 is the normalized time in units of @xmath32 , @xmath33 being the normal state single junction resistance , @xmath3 is the normalized frequency in units of @xmath34 , @xmath17 and @xmath28 are the normalized external currents in units of the single junction critical current @xmath35 .
this model was introduced by benz _
@xcite ( for @xmath36 ) to study the dc current - voltage curve of the @xmath15 array .
they obtained analytically that the critical current per junction of this model is @xmath37 .
@xcite the same model was later used in ref . to study the ( longitudinal ) shapiro steps . since the analysis done in the work of refs .
did not include the transverse ac current with accompanying transverse voltage drop , an additional constraint of @xmath38 was implied , reducing the model system to two dynamical degrees of freedom .
in contrast , our model system of a four plaquette unit cell consists of three effective dynamical variables .
we calculate the instantaneous longitudinal @xmath39 and transverse @xmath40 ( normalized ) voltages per junction as , @xmath41 and the @xmath0 characteristics , @xmath42 as a function of @xmath28 , where @xmath43 is a time average .
the total longitudinal voltage @xmath44 for an @xmath1 array , built with this @xmath45 superlattice unit cell , is @xmath46 .
when vortices move with a mean velocity @xmath47 in such @xmath48 superlattice structure , we can obtain the normalized voltage @xmath49 using the relation @xmath50 , where @xmath6 is the array periodicity ( see fig . [
fig : fig1]a ) .
the intrinsic washboard frequency for vortices moving with velocity @xmath47 in the periodic potential of the jja is @xmath51 .
phase - locking in the longitudinal direction is obtained when the frequency @xmath3 of the ac drive locks to a rational multiple of the intrinsic washboard frequency . for the @xmath9-th
harmonic this corresponds to @xmath52 , _
i.e. _ @xmath53 . this leads to phase - locking at voltages @xmath54 for fully frustrated jja . in general , for @xmath12 , the ground state has @xmath55 superlattice structure , therefore the voltage for vortices moving with velocity @xmath47 is @xmath56 , and phase - locking for the @xmath9-th harmonic is obtained at voltages @xmath57 .
this is the condition for the so - called `` fractional giant shapiro steps '' @xcite .
we will in this section assume that the dynamics of the system is represented by the simple two - plaquette degrees of freedom as shown in figure [ fig : fig2 ] .
we will apply the following linear transformation of the phase - variables of eqs .
( [ eq : eq_1])-([eq : eq_4 ] ) : @xmath58 with these variables we can represent the constraint ( @xmath59 ) of eq .
( [ eq : eq_1 ] ) as @xmath60 , and thereby write the relevant three degrees of freedom in either of the two following forms , eliminating @xmath61 : @xmath62 or eliminating @xmath63 : @xmath64 the normalized voltages are given by , @xmath65 and @xmath66 .
we will in the spirit of the usual shapiro analysis assume that @xmath67 which is the solution to eqs .
( [ eq : eq_12 ] ) and ( [ eq : eq_15 ] ) for large @xmath17 and @xmath3
. we will here look at eqs .
( [ eq : eq_14 ] ) and ( [ eq : eq_16 ] ) . assuming @xmath68 and @xmath69 , where @xmath70 and @xmath71 are constants and @xmath72 ,
the static contribution to equation ( [ eq : eq_16 ] ) is : @xmath73 where @xmath74 is the @xmath9th order bessel function of the first kind . inserting this into the static part of equation ( [ eq : eq_14 ] ) yields @xmath75\right\}\sin{\phi_x^{(0 ) } } \ ; . \label{eq : eq_20}\end{aligned}\ ] ] this expression provides a unique relationship between the constant phase , @xmath70 , and the dc current , @xmath28 .
however , for increasing @xmath28 , there exists a critical value , @xmath76 , for which no real @xmath70 can satisfy eq .
( [ eq : eq_20 ] ) .
this value is given by @xmath77\right\}\sin{\phi_x^{(0)}}\right ] \ ; , \label{eq : eq_21}\end{aligned}\ ] ] which is the predicted critical dc current for static states ( @xmath78 ) .
we notice that the identical expression for the critical current , eq .
( [ eq : eq_21 ] ) , can be obtained from the _ anisotropic _ dc driven system , @xmath79 where the anisotropy , @xmath80 ( suppression of transverse critical current ) , is given by the standard shapiro @xcite critical current , @xmath81 .
we will here use equations ( [ eq : eq_11])-([eq : eq_13 ] ) , since @xmath82 for @xmath83 provides for a simple description of the dynamics .
we will assume the ansatz , @xmath84 and @xmath85 .
the equation for @xmath63 , ( [ eq : eq_13 ] ) , now reads : @xmath86 where we will use the approximations : @xmath87 and @xmath88 . with the ansatz for @xmath63 above
, this equation has no static component .
the time varying component , at frequency @xmath3 , yields the coefficients @xmath89 and @xmath90 @xmath91\sin{\phi_x^{(0)}}}{4\omega^2+j_0 ^ 2\left(\frac{i_{ac}}{\omega}\right)-j_2 ^ 2\left(\frac{i_{ac}}{\omega}\right ) } j_0\left(\sqrt{a_0 ^ 2+b_0 ^ 2}\right ) \label{eq : eq_27 } \\
b_1 & = & b_0j_0\left(\sqrt{a_0 ^ 2+b_0 ^ 2}\right ) \ ; = \ ; \frac{2\omega\sin{\phi_x^{(0)}}+\left[j_0\left(\frac{i_{ac}}{\omega}\right)-j_2\left(\frac{i_{ac}}{\omega}\right)\right]\cos{\phi_x^{(0)}}}{4\omega^2+j_0 ^ 2\left(\frac{i_{ac}}{\omega}\right)-j_2 ^ 2\left(\frac{i_{ac}}{\omega}\right ) } j_0\left(\sqrt{a_0 ^ 2+b_0 ^ 2}\right ) \ ; , \label{eq : eq_28}\end{aligned}\ ] ] where @xmath92 and @xmath93 are the solutions for @xmath94 .
thus , the solution @xmath95 is correct up to order @xmath96 and @xmath97 is correct up to @xmath98 .
inserting this solution @xmath99 for @xmath63 into the @xmath63-linearized equation ( [ eq : eq_11 ] ) gives the static properties : @xmath100 thus , the locking range for the this step can be found to second order in @xmath101 .
the dominant part of this expression for the range in phase - locking yields : @xmath102 the expression displays quadratic growth of the phase - locked step size for small @xmath17 .
this is consistent with the particle ( pancake ) model results @xcite for vortices in rectangular pinning arrays , and it is different from the known longitudinal ( shapiro ) phase - locking of jjas.@xcite in addition to the range of phase - locking , @xmath20 , equation ( [ eq : eq_30 ] ) provides information about the offset , @xmath103 , of the phase - locked step relative to the ohmic ( linear ) curve ( see inset in figure [ fig : fig3 ] ) .
the offset is given by the part of the equation that does _ not _ depend on @xmath70 . from eq .
( [ eq : eq_30 ] ) we have : @xmath104 expressions ( 31 ) and ( 32 ) provide a second order ( in @xmath96 ) description of the phase - locking step magnitude and location as a function of the system parameters , @xmath3 and @xmath17 , for large @xmath3 .
we will here show the results of numerical simulations of the system analyzed in the previous section .
the simulations are conducted with numerical parameters corresponding to the model parameters , using a fourth order runge - kutta method such that the normalized time step typically is no larger than 1% of the period of the driving frequency , and often smaller .
since we are mostly concerned with dc @xmath0 characteristics , we choose to acquire data for averaging over many ac - periods of motion ( typically 10@xmath105 - 10@xmath106 ) after a sufficient initial time of interval allowed for transient behavior .
simulated @xmath0 characteristics are obtained by performing the necessary averages as described , and then changing the dc current , @xmath28 , slightly to acquire the next point on the @xmath0 curve .
all simulations are conducted for the fully frustrated case of @xmath59 .
figure [ fig : fig3 ] shows a simulated @xmath0 characteristic , @xmath49 as a function of @xmath28 , simulated for @xmath107 and @xmath108 , for the simple four plaquette model showed in figure [ fig : fig2 ] , described by the equations ( [ eq : eq_1])-([eq : eq_4 ] ) . as is obvious from the figure , we obtain clear signatures of critical current(s ) and phase - locked steps .
specifically , we observe the @xmath20 step at @xmath109 , and steps @xmath110 and @xmath111 at @xmath112 and @xmath113 , respectively .
we observe a critical current larger than the previously predicted @xcite value of @xmath114 for a fully frustrated dc - driven system .
the characteristics of this plot are very similar to the behavior observed in jja with parallel ac @xmath115 dc drives , obtained both by simulations @xcite and experiments,@xcite as well as analytically for the four plaquette model.@xcite comparing simulations of the full - rsj dynamical equations @xcite for different large arrays ( @xmath1 junctions ) we will later ( below ) demonstrate that the simple model of a @xmath27 array gives very good description of the jja dynamics .
however , we will first compare the predictions of the analytical treatment of the previous section to numerical results . in order to verify the simple theory for critical current and phase - locking behavior developed above , we have conducted numerical simulations of the four plaquette system described by eqs .
( [ eq : eq_1])-([eq : eq_4 ] ) for @xmath117 .
the first set of simulations are conducted to investigate the critical current , @xmath19 , as it is described by eq .
( [ eq : eq_21 ] ) .
this expression provides an estimate of the critical current , @xmath118 , for which the jja switches from a zero - voltage state ( @xmath119 and @xmath120 ) to a non - zero voltage state .
the simulations are conducted accordingly , starting the system at rest for small @xmath28 and slowly increasing the dc - bias until non - zero average voltage is detected .
the results are shown in figure [ fig : fig4]a , where the solid curve represents the expression , eq .
( [ eq : eq_21 ] ) , and the markers represent the simulation results for several frequencies @xmath121 as a function of the characteristic ratio , @xmath122 . the size of the markers are larger than the error on the estimated critical current .
it is obvious that the agreement is very good for all simulated data sets , and we conclude that the critical current , as given by eq .
( [ eq : eq_21 ] ) , is a relevant estimate for @xmath3 not smaller than 1 .
figure [ fig : fig4]b shows the critical current , @xmath123 , evaluated from numerical simulations when the dynamical system switches from the non - zero voltage state to the zero - voltage state ( see inset in figure [ fig : fig3 ] ) .
we have here shown the results of numerical simulations with markers as for figure [ fig : fig4]a , together with the solid curve of eq .
( [ eq : eq_21 ] ) .
however , it is clear from the figure that the critical current , @xmath123 , for decreasing dc - bias may be smaller than for increasing dc - bias ( @xmath124 ) .
since this is a multi - dimensional system , the critical current may be hysteretic , such that decreasing the dc - current , @xmath28 , for non - zero voltage states ( @xmath125 ) is subject to different critical characteristics .
one simple way of investigating this is to assume a non - phase - locked state of voltage @xmath126 , such that @xmath127 .
a primitive analysis can provide a hint to this hysteresis .
the critical current analysis of the previous section , is obviously a critical current for a system operated at the @xmath128 branch of the @xmath0-curve .
we may instead analyze what may happen for a _ non - phase - locked _ @xmath129 state .
we will still assume ( 17 ) to be an appropriate description of the transverse current .
however , equation ( [ eq : eq_13 ] ) becomes ( for small @xmath101 and with no resonance to @xmath3 ) : @xmath130 inserting the @xmath63 solution into equation ( [ eq : eq_11 ] ) yields the static component : @xmath131 where the second term on the left hand side is the result of the resonant mixing between the propagation , @xmath132 and the transverse oscillation , @xmath63 .
however , the overdamped dc driven pendulum equation is also subject to the following simple relationship , @xcite @xmath133 combining the two expressions provides the relationship @xmath134
^ 2-v_x^2 } \ ; \ ; \le \ ; \ ; 0.826591 \ ; . \label{eq : eq_37}\end{aligned}\ ] ] as we have indicated , the critical current , @xmath123 has a maximum value at around 0.82 ( for @xmath135 and @xmath136 ) .
thus , we can argue that propagating ( @xmath137 ) solutions may exist for @xmath138 , which provides for a hysteretic @xmath0 characteristic switching between zero and non - zero voltages in the range @xmath139 , when @xmath140 .
notice that when @xmath141 , the relevant critical current for both zero and non - zero voltage states must be @xmath118 , since no static states exist for @xmath142 . however , when @xmath143 , the actual critical current for switching into a zero - voltage state may be anywhere in the interval @xmath144 $ ] .
figure [ fig : fig4 ] clearly indicates the hysteretic switching in the @xmath0 characteristics when @xmath139 , which is the case for small @xmath145 .
we note that the above rather primitive analysis of the hysteresis provides a fairly good agreement with the results of numerical simulations .
the results of the analysis are not completely consistent with its assumptions in that the resulting amplitude of @xmath63 for the optimized @xmath136 is about 1.5 , which is not a small number . however , a more detailed ( nonlinear ) analysis of a single frequency representation of @xmath63 yields quantitatively similar and qualitatively identical results ( @xmath146 ) as the above , and we therefore conclude that the simple explanation for hysteresis presented here is relevant . for @xmath135 we can provide an explicit approximate expression for @xmath147 by assuming the critical current is given by the value of @xmath49 which optimizes
this leads to @xmath149 , which inserted into the above equation yields the optimized coordinates : @xmath150 .
it is noteworthy that we observe , as predicted by the expression for @xmath118 , @xmath151 to be _ larger _ than the dc - driven system , @xmath152 .
hence , a transverse ac driving leads to an _ enhancement _ of the critical current .
this is contrary to the case with the ac - current parallel to the dc - current , where the critical current is reduced , _
i.e. _ @xmath153.@xcite the predicted range , @xmath20 , in @xmath28 of phase - locking , as given by eq .
( [ eq : eq_31 ] ) , is investigated through simulations similar to the above study of the critical current .
comparisons between the predicted expression and results of numerical simulations are shown in figure [ fig : fig5]a , which displays the largest magnitude of the range in dc current for which phase - locking is observed as a function of the characteristic ratio , @xmath122 , for different values of @xmath3 in the interval @xmath121 .
markers represent results of numerical simulations and solid curves represent the corresponding predicted results of eq .
( [ eq : eq_31 ] ) .
it is obvious that the simulated parameter sets provide very good overall validation of the perturbation analysis , with the larger of the simulated frequencies providing better agreement than the smaller , as expected .
however , an observation common to all simulated frequencies is that large deviations from the expected behavior are found for parameter values ( @xmath3 and @xmath17 ) leading to @xmath154 .
the reason for this discrepancy is likely due to a dynamical instability , which can be explained by the perturbation analysis above .
the average equilibrium position , @xmath155 , of the variable @xmath63 can be observed from eq .
( [ eq : eq_26 ] ) if we write @xmath156 , where @xmath71 is varying slowly in time ( much slower than @xmath3 ) , and @xmath157 represents all high frequency ( including @xmath3 ) contributions ( @xmath158 ) . the slow evolution of equation ( [ eq : eq_26 ] ) can then be written , @xmath159 thus , we find that the stable position of @xmath63 is : @xmath160 the consequence of this abrupt transition in @xmath63 is that the locking phase , @xmath70 , must experience a similar abrupt transition of @xmath161 , as can be seen from equation ( [ eq : eq_11 ] ) .
we therefore claim that the apparent discrepancy observed between the numerical simulations and the perturbation theory near the roots of @xmath162 , is a result of dynamical instabilities arising from switching the average phase , @xmath155 between @xmath163 and @xmath161 .
we finally show the comparisons of the center of the phase - locked step as a function of the characteristic ratio , @xmath122 , for different values of @xmath3 in the interval @xmath121 .
the predicted behavior , eq .
( [ eq : eq_32 ] ) , is subject to the same issues as the predicted range of phase - locking since the origin of both expressions is eq .
( [ eq : eq_30 ] ) .
figure [ fig : fig5]b shows the offset , @xmath164 , between the center of the step , @xmath165 , and the ohmic curve . as is the case for the phase - locking range shown in figure [ fig : fig5]a , the comparison between numerical simulations ( markers ) and the corresponding predicted offsets ( solid curves ) is very good , except for the instabilities near the roots of @xmath162 . notice that the phase - locking analysis leading to the predictions eqs .
( [ eq : eq_31 ] ) and ( [ eq : eq_32 ] ) does not depend on the sharp @xmath63 transition between @xmath163 and @xmath161 .
the reason is that this transition provides only a sign change in the effective equations of phase - locking , and the magnitudes of locking range and offset are therefore unaffected as long as @xmath166 .
based on the above presented comparisons between the numerical simulations of critical currents , range of phase - locking and position of the phase - locked step in the @xmath0 characteristics of a transversely ac - driven jja and the corresponding results from simple perturbation analysis , we conclude that the high frequency behavior is well described by the presented analytical treatment . in fig .
[ fig : fig6 ] we show the critical current behavior for intermediate and low frequencies . for intermediate frequencies ( figure [ fig : fig6]a )
we observe how the critical current , @xmath118 , increasingly deviates from the high frequency behavior outlined above .
even so , we notice that the overall behavior of the critical current is qualitatively well described by the analysis leading to equation ( [ eq : eq_21 ] ) for @xmath167 .
we have , for comparison , included an example of the critical current for the longitudinally ac - driven jja as an inset .
not surprisingly , decreasing the frequency further ( see figure [ fig : fig6]b ) results in rather large discrepancy between the high frequency analysis of section iiia and the numerical simulations , and no universal behavior of the critical current as a function of the characteristic ratio , @xmath122 , can be found .
however , we do observe that the critical current does seem to increase quadratically for small @xmath17 . in fig .
[ fig : fig7 ] we show the phase - locking range , @xmath20 , at @xmath109 as a function of @xmath23 for intermediate ( @xmath168 ) and low frequencies ( @xmath169 ) . again , as for the critical current we observe that the intermediate frequency range provides for reasonably good qualitative comparisons between numerical simulations and the high frequency analysis of section iiia .
we have , for comparison , included an example of the comparable range of phase - locking for the longitudinally ac - driven jja as an inset . a noticeable feature of figure [ fig : fig7]a is that the dynamical instability discussed above around @xmath135 seems to widen as the frequency is lowered . figure [ fig :
fig7]b shows how this instability provides for increasing discrepancy between high frequency analysis and numerical simulations .
however , we notice that even the very low frequencies retain the basic feature of quadratic growth of the phase - locking range as a function of @xmath17 for small @xmath17 .
let us now analyze in detail the dynamics of the voltage responses ( increasing and decreasing dc - current ) to elaborate on our previous results : critical current hysteresis and windows without transverse phase locking .
we calculate @xmath0 curves and lyapunov exponents as a function of @xmath17 and @xmath3 . in order to distinguish between periodic or quasi - periodic dynamics and chaotic dynamics
we calculate the maximum lyapunov exponent , @xmath170 , following the standard methods of nonlinear dynamics.@xcite this means that a small perturbation , @xmath171 , to the initial condition will displace the new trajectory by an amount @xmath172 .
the lyapunov exponent is then defined as @xmath173 to recognize a chaotic trajectory we evaluate the maximum lyapunov exponent .
if @xmath174 the trajectory is locally unstable ; i.e. , initial points that are arbitrarily close to each other are macroscopically separated by the flow after a sufficiently long time and the attractor is chaotic .
negative lyapunov exponents are obtained when trajectories that start sufficiently close to a subset are attracted to it . here
we will show two particular cases : @xmath175 , corresponding to a set of parameters where no hysteresis is obtained , and @xmath176 , corresponding to the hysteretic regime . in fig .
[ fig : fig9 ] we plot the @xmath0 curves and maximum lyapunov exponents for @xmath177 and @xmath178 ( a - b ) and @xmath179 ( c - d ) .
the exponents are estimated from @xmath180 after a finite time @xmath181 , with @xmath182 .
for @xmath178 we show a range in @xmath28 where a wide transverse phase locking step exists ( fig .
[ fig : fig9]a ) , and the corresponding maximum lyapunov exponent , @xmath170 , is shown in fig . [
fig : fig9]b .
we see that within the step we have @xmath183 , with the most negative value of @xmath170 at the center of the step . outside the steps ,
the lyapunov exponent is nearly equal to zero , @xmath184 , corresponding to quasiperiodic behavior .
a different behavior is obtained for @xmath179 , shown in fig .
[ fig : fig9]c , where we see that the step disappears and thus there is no transverse phase locking in the same @xmath28 range where we find a step in fig .
[ fig : fig9]a . the maximum lyapunov exponent , plotted in fig . [ fig : fig9]d , is small but positive for the @xmath28 range around the expected location of the step
, the smallness of @xmath170 implies that the dynamics can be either chaotic ( @xmath185 ) or quasiperiodic ( @xmath186 ) in this case .
the @xmath0 curves near the corresponding critical currents are shown as insets .
we see that the absence of hysteresis in critical current is associated with the occurrence of transverse phase locking .
inversely , hysteresis in critical current is obtained for approximately the same parameters for which transverse phase locking is absent .
this is in agreement with the above analysis that indicates the critical current hysteresis is present in the vicinity of @xmath135 , which is also the location of the dynamical instabilities of the locked phase of the @xmath20 step . in summary , around the transverse phase
locking step region we can distinguish three different voltage responses : a , b and c , which are indicated in fig .
[ fig : fig9 ] .
we now calculate the voltage power spectrum and poincar sections to distinguish these three types of dynamical behavior .
this way to characterize dynamical behaviors was used before in capacitive rf - biased jja .
@xcite we analyze both transverse and longitudinal voltage power spectra . from the instantaneous transverse voltage
we obtain the transverse voltage power spectrum : @xmath187 where @xmath188 . from the instantaneous longitudinal voltage , @xmath39
, we obtain the longitudinal voltage power spectrum : @xmath189 for studying the nature of the attractor in the different regimes it is useful to consider a poincar section of the phase - space trajectories .
@xcite we consider the stroboscopic poincar section of the trajectories in the @xmath190 vs. @xmath191 plane , recording the values taken by these variables each period of the ac drive . in fig .
[ fig : fig10 ] we show the power spectra and poincare sections for @xmath177 and for the @xmath17 and @xmath28 values corresponding to the a , b and c regimes . for each case
we show the longitudinal , @xmath192 , and transverse , @xmath193 , voltage power spectra as a function of @xmath194 and their corresponding poincar sections .
let us first discuss the case corresponding to the regime b , in which there is transverse phase - locking .
this is shown in fig .
[ fig : fig10]b for @xmath195 and @xmath178 , which corresponds to the step with mean voltage @xmath196 ( see fig .
[ fig : fig9]a , regime b ) .
we see that the longitudinal power spectrum , @xmath192 , presents a delta - like peak for @xmath197 .
thereby , the first harmonic of longitudinal voltage fluctuations is locked to @xmath198 , as expected for this step , since it corresponds to @xmath199 in the phase - locking condition @xmath52 .
the phase - locking with a double frequency corresponds to the case when the vortex lattice oscillates in full synchrony with the transverse ac - current , and the ground state repeats itself after one period of the ac drive . in the transverse voltage power spectrum , @xmath200
, there is a sharp peak at @xmath3 .
this is characteristic of transverse phase - locking : the dynamics in the transverse direction locks at half the frequency than the dynamics of the longitudinal direction .
this is so because in a single period of the ac - drive , @xmath182 , the longitudinal component moves forward @xmath9 steps in the lattice period @xmath6 , while the transverse component completes only the first half of its oscillation . in fig .
[ fig : fig10]e , we show the poincar section corresponding to this case in the regime b. the figure shows a very localized poincar section since the trajectory always comes back approximately to the same location in phase space in each ac cycle , since the trajectory is periodic ( closed orbit ) .
now we analyze the case corresponding to the a regime , which is for a current outside the step , @xmath201 , see fig .
[ fig : fig9]a . in this case
we see again in @xmath192 a peak at @xmath198 and in @xmath193 a peak at @xmath3 .
however , the peaks now have a small broadening , and small amplitude satellite peaks have appeared at neighboring frequencies .
this is evidence of another kind of long - term behavior , namely quasiperiodic dynamics .
we can corroborate this with the corresponding poincar section shown in fig .
[ fig : fig10]d .
it consists now on a closed one dimensional curve , which means that trajectories wind around on a torus , never intersecting itself and yet never quite closing , typical of a quasiperiodic orbit .
we have also looked at the time dependent estimates of the lyapunov exponent , @xmath202 .
we find that @xmath203 for finite @xmath31 , but its absolute value tends to zero for long times as @xmath204 , consistent with quasiperiodic behavior .
let us now study the last case , corresponding to regime c. this is done for @xmath195 and @xmath179 in fig . [
fig : fig10]c .
we see that there are broad peaks in the spectrum in both directions , @xmath192 and @xmath193 , and that there is a marked increase in the power spectra for low frequencies . moreover , in fig .
[ fig : fig10]f we show the corresponding poincar section which consists on successive points jumping from one region of phase space to another and forming a complex curve , which does not seem to close on itself .
it is rather difficult to decide from this plot if it corresponds to a quasiperiodic orbit or to a low dimensional attractor of a weakly chaotic orbit .
we have obtained also the time dependent estimate of the lyapunov exponent also for this case .
we find that @xmath205 for all @xmath31 , but its magnitude is decreasing with time as @xmath204 as far as we have been able to observe . the fact that @xmath202 is possitive for finite @xmath31 means that there is a dynamical instability that causes a seemingly chaotic behavior at intermediate times and a large noise as seen in the low frequency power spectrum .
however , for long times it is very likely that the system will settle in a quasiperiodic dynamics with @xmath206 . in any case ,
the regime c is very different from the regime a , as can be seen by comparing the power spectra of fig .
[ fig : fig10]a and fig .
[ fig : fig10]c . another view of the dynamics can be obtained by looking at the behavior of the lyapunov exponent and the noise in the region , where a step is expected , as a function of @xmath23 .
we proceed as follows : for a given @xmath207 , we compute the set of values of lyapunov exponents @xmath170 and low frequency longitudinal noise @xmath208 that correspond to currents @xmath28 in the region of voltage where a step is expected .
( we look at @xmath28 values for which @xmath209 , we consider @xmath210 ) .
we plot the resulting set of values of @xmath170 as a function of @xmath23 in fig .
10a and the values of @xmath211 as a functions of @xmath23 in fig .
the vertical lines in the plot correspond to the zeroes of @xmath212 .
we see clearly that near these values there are windows of dynamical instability where @xmath213 and where the noise @xmath211 is large . in the regions of phase - locking
we find a couple of interesting results that are worth mentioning .
( i ) the most negative value of the lyapunov exponent occurs in the middle of the phase - locked step and its magnitude is proportional to the step width @xmath20 , as given by eq .
( [ eq : eq_31 ] ) .
( ii ) the largest value of the noise @xmath211 occurs at the edge of the phase - locked step ; its magnitude is also proportional to the step width @xmath20 , as given by eq .
( [ eq : eq_31 ] ) .
we will now consider the quality of the simple @xmath27 model as representing the dynamics of large @xmath214 jjas .
it is known that collective effects at high currents may come into play . at high currents , the @xmath215 symmetry of the ground state
can be broken because a driving current can induce domain walls .
@xcite simulations of @xmath0 curves with the rsj model and free boundary conditions , for @xmath15 and @xmath216 have reported a chaotic regime at @xmath217 related to the motion of domain walls .
@xcite it has been shown that open boundary conditions nucleate domain walls leading to a critical current lower than de analytic value @xmath218 at @xmath216.@xcite moreover , ciria and giovanella @xcite have shown microscopically that different dynamical states are possible for the longitudinal shapiro steps . besides the checkerboard ground state configuration , other stable solutions with domain - walls are possible . then , depending on dc current value and history , domain walls can appear , which are not permitted in the four plaquette model .
therefore , in order to evaluate to what extent the four plaquette model is valid in the transverse ac driven case , we have calculated numerically @xmath0 curves for @xmath214 arrays , for @xmath219 , with the full rsj model used before in ref . .
we use periodic boundary conditions in both directions in the presence of an external dc current , @xmath28 , plus a perpendicular ac current , @xmath220 .
we solve the dynamical equations with time step @xmath221 ( @xmath222 ) and total integration time @xmath223 after a transient @xmath224 .
we calculate @xmath0 curves as a function of @xmath17 and @xmath3 , increasing dc current , @xmath225 , from checkerboard ground state at @xmath226 and then decreasing dc current , @xmath227 , from the phase configuration obtained at high current .
we use a dc current step @xmath228 to obtain @xmath19 and @xmath229 to calculate the step width .
one of the relevant results with the four plaquette model is the dependence of the critical current with @xmath23 for high frequencies , as shown in fig .
[ fig : fig4]a and fig .
[ fig : fig6]a .
we have also calculated @xmath19 as a function of @xmath23 for high @xmath3 in large jja arrays . in fig .
[ fig : fig11]a we show the case for a particular high frequency value in a @xmath230 array .
we see that it has the same behavior as observed in the four plaquette model : @xmath231 , ranges of @xmath23 around the maxima of @xmath19 where there is hysteresis , and a quadratic increase with @xmath17 for @xmath18 .
we compare with the analytical results expected for @xmath118 , eq .
( [ eq : eq_21 ] ) and @xmath123 , eq .
( [ eq : eq_37 ] ) , which are represented by dot - dashed lines .
we see that @xmath118 obtanained numerically for a large array is in excellent agreement with the analytical result for the @xmath27 system .
this is quite reasonable , since @xmath118 corresponds to the limit of stability of the checkerboard ground state , which is well represented by the @xmath27 model . on the other hand
, the @xmath123 shows some small deviation from the @xmath27 result , @xmath232 . also the range in @xmath23 where there is hysteresis is bigger in a large system .
the current @xmath123 corresponds to the low current limit of stability of the moving ( non - zero voltage ) state . in large systems
, the moving state can have domain walls , as was found in ref . , and the presence of domain walls can lead to a lower @xmath123 . in order to analyze
more quantitatively in which @xmath23 ranges the collective effects could be more relevant , we focus on two cases : case @xmath6 corresponding to values that do not show hysteresis in the critical current in a small system , but are close to the edge of the @xmath23 range of hysteresis , and case @xmath233 corresponding to values that show hysteresis in the @xmath27 system . for each case
we calculate the critical current both increasing and decreasing the dc drive , and therefore they correspond to @xmath234 , @xmath235 , @xmath236 and @xmath237 in fig .
[ fig : fig11]a .
we show in the inset of fig .
[ fig : fig11]a the critical currents obtained for all these cases as a function of system size , @xmath238 . in case @xmath6 , corresponding to the non - hysteretic region , we see that there is no size effect in @xmath239 up to @xmath240 .
also we see that @xmath241 for @xmath242 , while for @xmath240 we find that hysteresis has appeared and @xmath243 . in the hysteretic region ,
case @xmath233 , size dependent critical currents are obtained for @xmath244 , while @xmath245 is size independent . moreover ,
the amplitude of the hysteresis , @xmath246 weakly increases with system size . in fig .
[ fig : fig11]b we show the range of phase locking @xmath20 as a function of @xmath23 for the @xmath247 array .
we find that , when there is phase - locking , the numerically obtained @xmath20 is very accurately described by the analytical result of eq .
( [ eq : eq_31 ] ) for the @xmath27 model . in the inset of fig .
[ fig : fig11]b we show the size dependence of @xmath20 for the case marked as @xmath248 in the plot ( it corresponds to the same @xmath23 of case @xmath6 of fig .
[ fig : fig11]a ) .
there is no appreciable size dependence .
as observed in the simulations of the @xmath27 system , we also find here that the phase - locking is lost near the zeros of @xmath212 due to the presence of dynamical instabilities .
also we observe that the presence of hysteresis in the critical current is nearly coincident with the absence of phase locking .
we find that with increasing system size these regimes of dynamical instability are amplified in their extension both in their @xmath28 dependence and in their range of @xmath23 around the zeros of @xmath212 .
this means that the dynamical instabilities detected in the four plaquette system can lead to an increased spatiotemporal chaos in larger systems where collective effects are important .
it is important to point out that there are no trivial connections between vortex dynamics in the fully frustrated jja and that of a commensurate vortex lattice moving in a rectangular pinning potential in a bulk superconductor .
@xcite this is so because the london model with vortices interacting through pair potentials apply to jjas only in the limit of very low vortex density , such that @xmath249
. @xcite the fully frustrated case represents , in this last respect , an interesting limit for studying , where the complete phase field , rather than just the positions of vortices , should be taken into account to describe the dynamics .
we have found transverse phase locking steps in fully frustrated jja .
this new type of ( fractional ) giant phase - locking steps presents marked differences with the well known longitudinal fractional giant shapiro steps .
particularly , the presence of the transverse ac force increases the critical depinning current with respect to the case without ac drive ( or with a longitudinal ac drive ) .
we have analyzed both analytically and numerically the behavior of the steps as a function of ac amplitude @xmath17 and frequency @xmath3 . for @xmath18 , the depinning critical current and the phase - locked step width @xmath20 for @xmath21 ,
increase quadratically with @xmath17 .
for @xmath22 we have found windows of @xmath23 where depinning is hysteretic and phase locking is destroyed due to dynamical instabilities .
the emergence of a weakly chaotic behavior at zero temperature , in a system with non capacitive junctions , is another particular characteristic of transverse phase locking which is absent in longitudinal phase locking in overdamped jja . comparing with the behavior of large fully frustrated arrays
we have found that transverse phase locking can be well described by an effective four plaquette model , and that collective effects become more important close to the regions of dynamical instability of the four - plaquette model .
our results could be observed experimentally in jja . in particular
, the enhancement of the critical depinning current with a transverse ac drive could be an interesting experimental consequence of the phenomena reported here .
we acknowledge discussions with g. carneiro , h. pastoriza , c. reichhardt and d. shalom .
parts of this work were supported by the director , office of advanced scientific computing research , division of mathematical , information and computational sciences of the u.s .
department of energy under contract number de - ac03 - 76sf00098 .
the work in argentina was supported by cnea , conicet and anpcyt ( pict99 - 03 - 06343 ) . | we study , analytically and numerically , phase locking of driven vortex lattices in fully - frustrated josephson junction arrays at zero temperature .
we consider the case when an ac current is applied _ perpendicular _ to a dc current .
we observe phase locking , steps in the current - voltage characteristics , with a dependence on external ac - drive amplitude and frequency qualitatively different from the shapiro steps , observed when the ac and dc currents are applied in parallel .
further , the critical current increases with increasing transverse ac - drive amplitude , while it decreases for longitudinal ac - drive . the critical current and the phase - locked current step width , increase quadratically with ( small ) amplitudes of the ac - drive . for larger amplitudes of the transverse ac - signal , we find windows where the critical current is hysteretic , and windows where phase locking is suppressed due to dynamical instabilities .
we characterize the dynamical states around the phase - locking interference condition in the @xmath0 curve with voltage noise , lyapunov exponents and poincar sections .
we find that zero temperature phase - locking behavior in large fully frustrated arrays is well described by an effective four plaquette model . |
three long - baseline laser interferometer gw detectors have been , or are nearly , constructed .
the us ligo ( laser interferometer gravitational - wave observatory ) has started observation with two 4-km arm detectors situated at hanford , washington , and livingston , louisiana ; the hanford detector also contains a 2-km interferometer .
the italian / french virgo project is commissioning a 3-km baseline instrument at cascina , near pisa .
there are detectors being developed at hannover ( the german / british geo project with a 600-m baseline , which had its first test runs in 2002 ) and near perth ( the australian international gravitational observatory , aigo , initially with an 80-m baseline ) .
a detector at tokyo ( tama , 300-m baseline ) has been in operation since 2001 .
the astrophysical detection rates are expected to be low for the current interferometers , such as ` initial ligo ' , but second - generation observatories with high optical power are in the early stages of development ; these ` advanced ' interferometers have target sensitivities that are predicted to provide a practical detection rate .
double neutron star ( dns ) binary mergers are potentially among the strongest gw sources and are a key search target for interferometric observatories @xcite .
eight close dns binary systems are known to exist in the galaxy as a subset of the observed radio binary pulsar population @xcite .
energy loss from gw emission @xcite causes an orbital in - spiral until the binary system coalesces , resulting in a burst of gws usually described as a ` chirp ' signal .
the discovery of the dns binary systems containing psrs j07373039 @xcite and j1756@xmath12251 @xcite using the parkes radio telescope brings the number of known dns systems in the galactic disk to merge within the hubble time to four ; this excludes the psr b2127 + 11c system as it is in a globular cluster and probably not formed by binary evolution ( see table 1 ) . with an orbital period of only 2.45 hr , the double pulsar system j07373039a
, b will coalesce in only 87 myr , a factor of 3.5 shorter than the coalescence time of the hulse - taylor psr b1913 + 16 system , which yielded the first evidence for gravitational radiation . predicted merger rates are dominated by the j07373039 system , because of its proximity , short coalescence time and the difficulty of its discovery @xcite .
in contrast , psr j17562251 s parameters are similar to those of previously known pulsars ; thus , its very recent addition is nt expected to significantly alter the predicted merger rate . & @xmath2 & @xmath3 & ref .
+ & period & period & & mass & & & + & ( ms ) & ( hr ) & & ( m@xmath4&(myr ) & ( myr ) & + + * j0737@xmath13039a , b * & 23 , 2773 & 2.45 & 0.088 & 2.58 & 210,50 & 87 & ( 1,2 ) + b2127 + 11c & 31 & 8.04 & 0.681 & 2.71 & 969 & 220 & ( 3 ) + * b1913 + 16 * & 59 & 7.75 & 0.617 & 2.83 & 108 & 310 & ( 4 ) + * j1756@xmath12251 * & 28 & 7.67 & 0.181 & 2.57 & 444 & 1690 & ( 5 ) + * b1534 + 12 * & 38 & 10.10 & 0.274 & 2.75 & 248 & 2690 & ( 6 ) + j1811@xmath11736 & 104 & 449 & 0.828 & @xmath5 & 900 & @xmath6 & ( 7 ) + j1518 + 4904 & 41 & 207 & 0.249 & 2.61 & 16,000 & @xmath7 & ( 8) + j1829@xmath82456 & 41 & 28.2 & 0.139 & 2.5 & & @xmath9 & ( 9 ) + in order to estimate the dns coalescence rate , kalogera et al . ( 2001 ) used a semi - empirical approach , based on the physical properties of known dns systems and pulsar survey selection effects , to obtain scale factors that correct for the unobserved fraction of existing systems . in an alternative approach , kim et al . ( 2004 ) generated the probability distribution of merger rates using monte carlo simulations in which they populate a model galaxy with a specified number of ns binaries whose properties are those of known systems ; they also model survey selection effects to deduce detection rates and their statistical significance . with this model and
more up - to - date pulsar survey results that include j07373039 , @xcite presented updated bounds for the coalescence rate for galactic disk dns systems , yielding @xmath10 myr@xmath11 , and derived detection rates for initial and advanced ligo of @xmath12 yr@xmath11 and @xmath13 yr@xmath11 respectively , factors 57 higher than previous estimates . others @xcite assume a simple parametrization of merger rates with merger probability @xmath14 , where @xmath15 is the time from formation of the binary and @xmath16 is a constant ; the distribution of mergers as a function of time is simply given by the convolution of the star formation rate with the distribution @xmath17 .
@xcite , using numerical simulations , derive a probability distribution for dns merger time , @xmath18 , of the form @xmath19 , and find that mergers are possible between @xmath20 yr and the age of the universe . using both evolutionary and statistical models , @xcite find a galactic merger rate of @xmath21 yr@xmath11 , similar to the lower bound calculated by @xcite .
a purely theoretical approach to estimate coalescence rates can be based on the predictions of population synthesis models that attempt to simulate the entire pulsar population @xcite .
however , the size of the parameter space and associated uncertainties involved seriously limit this way of inferring merger rates .
for instance , new theoretical models of merger formation can have considerable impact : recent simulations by chaurasia and bailes ( 2005 ) show that significant kick velocities imparted to neutron stars at birth give rise to highly eccentric orbits resulting in accelerated orbital decay and greatly reduced lifetimes .
their work indicates that a non - negligible fraction of dns systems could merge within times as short as 10 yrs .
this implies that there could be a selection effect favouring the observation of those binary pulsar systems that coalesce over relatively long times . using the lower galactic rate of 17 myr@xmath11 from kalogera et al .
( 2004 ) and four merger times , @xmath22 ( 0,1,2,5 ) gyr , assumed the same for all systems .
the curves are calculated using the sfr evolution factor obtained from the parametrized model labelled sf2 in porciani & madau ( 2001 ) and a flat-@xmath23 cosmology with @xmath24 and @xmath25 , assuming a sfr cut - off at @xmath26 ; the cut - off projects forward in time to the end points of the time - delayed curves .
a time delay of 15 gyr has minimal effect on @xmath27 at small @xmath28 compared to the differential rate locked to the evolving sfr . ]
our interest is in the detection of a gw background generated by dns merger events throughout the cosmos . as a starting point , we adopt the galactic coalescence rates of @xcite as a basis for estimating the rate density in our local intergalactic neighbourhood . to project this throughout the universe , we assume that the dns system formation rate follows the evolving star formation rate ( sfr ) , for dns systems with relatively short merger times of tens to hundreds of millions of years . in this way , we employ the updated galactic dns merger rate estimates as input to a model that describes the observer s evolving record of dns mergers in terms of a ` probability event horizon ' ( peh ) .
we describe how the detectability of dns mergers evolves with observation time .
we use the peh concept @xcite to describe the cumulative gw signal from cosmic dns mergers and show that the signal is dominated by the spatial and temporal distribution of events .
instead of describing such a signal as ` stochastic ' , we argue that the signal , although temporally random , has a unique signature .
we use the term ` geometrical noise ' to distinguish it from the stationary noise usually associated with a primordial gw stochastic background .
in the standard friedmann cosmologies , one can express the differential dns merger rate as the merger rate in the redshift shell @xmath28 to @xmath29 : @xmath30 where @xmath31 is the cosmology - dependent co - moving volume element and @xmath32 is the all - sky ( @xmath33 solid angle ) dns merger rate , as observed in our local frame , for sources out to redshift @xmath28 .
source rate density evolution is accounted for by the dimensionless evolution factor @xmath34 ; this is normalized to unity in the present - epoch universe @xmath35 so @xmath36 is the @xmath37 rate density .
the galactic dns merger rate is converted to a rate per unit volume , @xmath36 , using the conversion factor @xmath38 from @xcite for the number density of galaxies in units of mpc@xmath39 .
the @xmath40 factor accounts for the time dilation of the observed rate by cosmic expansion , converting a source - count equation to an event rate equation .
we assume a ` flat-@xmath23 ' cosmology , in which the density parameters of matter and vacuum energy sum to unity ; we use @xmath24 and @xmath25 for their present - epoch values and take @xmath41 km s@xmath11 mpc@xmath11 for the hubble parameter at @xmath37 . with the assumption that the dns system formation rate follows the evolving star formation rate ,
@xmath34 is obtained by normalizing a sfr evolution model to its @xmath37 value .
presently there is no consensus among astronomers on a model of star formation history at high @xmath28 , so we shall use the parametrized model labelled sf2 of @xcite scaled to the flat-@xmath23 cosmology ; see @xcite for details . the cumulative dns merger rate @xmath32 is calculated by integrating the differential rate from the present epoch to redshift @xmath28 : @xmath42 whereas core - collapse supernova events ( because of their short - lived progenitors
) closely follow the evolving star formation rate , dns merger events are generally significantly delayed with respect to star formation . in order to test the sensitivity of @xmath27 to this delay , we have calculated it under the simplified assumption that all such events are delayed by the same merger time @xmath18 .
figure [ frates ] plots @xmath27 for four choices of @xmath18 , namely 0 , 1 , 2 and 5 gyr .
the three delayed curves were obtained by converting @xmath34 to a function of cosmic time ( or , equivalently , look - back time ) , shifting it by @xmath18 and then converting back to a function of @xmath28 . cosmic time and @xmath28 are related by the evolving hubble parameter @xmath43 , which can be expressed in terms of the contributions of matter and vacuum energy @xcite : @xmath44^{1/2}\ ] ] for a flat-@xmath23 cosmology , and cosmic time is expressed by the formula : @xmath45^{-1}\mathrm{d}z\,.\ ] ] figure 1 shows that a 12 gyr merger time does not significantly alter @xmath27 , especially for small @xmath28 .
increasing @xmath18 has the effect of reducing the available volume for observing dns mergers and , for @xmath46 gyr , coalescences are restricted to the volume bounded by @xmath47 .
so , for systems such as those used by @xcite in their rate estimates , we can safely assume that they quite closely follow the evolving star formation rate .
for definiteness , in the following calculations , we assume that the dns merger rate approximately follows the sfr evolution factor @xmath34 . as the dns mergers throughout the universe are independent of each other , their distribution is a poisson process in time : the probability @xmath48 for at least one event to occur in a volume out to redshift @xmath28 during observation time @xmath49 at a mean rate @xmath32 is given by an exponential distribution : @xmath50 with mean number of events @xmath51 .
the peh is defined by the minimum distance , or redshift @xmath52 , for at least one event to occur over some observation time @xmath49 , with probability above some selected threshold @xmath48 .
we find @xmath53 by fixing @xmath48 and solving the above condition numerically , thus defining the peh . in practice
, we shall take @xmath54 , corresponding to a 95% probability of observing at least one event within @xmath28 , so @xmath55 is the mean number of events . converting @xmath53 to luminosity distance @xmath56 , and differentiating with respect to @xmath49 then yields @xmath57 : the peh velocity , which describes the rate at which observations penetrate into the low - probability ` tail ' of the event distribution .
[ tbl-2 ] .the probability for at least one dns merger to occur within the volume bounded by the luminosity distance @xmath58 mpc for one year of observation ( initial ligo ) and by @xmath59 mpc for one day ( advanced ligo ) assuming the galactic dns merger rate limits , @xmath60 , from kalogera et al .
( 2004 ) . [ cols= " < , <
, < , < " , ] table [ tbl-2 ] uses equation ( [ prob2 ] ) to calculate the probability for at least one dns merger to occur within the volume bounded by @xmath58 mpc for one year of observation ( initial ligo ) and within @xmath59 mpc for one day ( advanced ligo ) using the maximum and minimum values of the galactic dns merger rate @xmath60 as estimated by @xcite , namely 292 and 17 myr@xmath11 .
the two distances , 20 and 200 mpc , correspond approximately to the detection thresholds for dns merger detection for initial and advanced ligo respectively .
figure [ peh1 ] plots @xmath56 , using the upper and lower galactic merger rate limits from kalogera et al .
for a dns merger at distance @xmath61 from the detector , the optimal gw signal - to - noise ratio , @xmath62 , assuming matched filtering is @xcite : @xmath63 where @xmath64 and 200 mpc for initial and advanced ligo respectively and @xmath65 , the angular sensitivity function of the detectors , averages to 2.56 over all sky and source positions . by substituting @xmath56 for @xmath61 in equation ( [ snr1 ] )
, we obtain the dns merger probability event horizon of the optimal signal - to - noise ratio : @xmath66 ( dashed curve ) and 17 myr@xmath11 ( solid line ) , and merger rate evolution following the instantaneous sfr .
the horizontal lines show the 200 and 20 mpc horizons , indicating the sensitivity limits for dns coalescence detection assuming optimal filtering for advanced and initial ligo . ] for advanced and initial ligo using the local galactic rate limits for dns mergers from kalogera et al .
( 2004 ) : 292 myr@xmath11 ( dashed curve ) and 17 myr@xmath11 ( solid line ) . for advanced ligo with detection threshold @xmath67 ( assuming optimal filtering )
, at least one dns merger will be detectable during 12211 days of observation . for initial ligo sensitivities
, dns mergers will be detectable in 22380 yrs of observation and , for a @xmath68 , in 468 days . in the region below @xmath69 ,
individual mergers are unlikely to be detected , but the cumulative signal will manifest as ` geometrical noise ' in interferometric data . ] assuming @xmath70 , corresponding to the detection threshold for a dns merger in advanced ligo for an ` average ' source at 200 mpc @xcite , one can model the evolution of @xmath62 as a function of duty cycle or observation time .
figure [ fig3 ] plots @xmath71 , showing the peh for at least one dns merger to occur during observation time @xmath49 , with @xmath72 , for the upper and lower uncertainty limits of the local dns rate densities @xcite . for advanced ligo ,
at least one dns merger with a @xmath73 11.7 will be detectable using a matched filter every 12211 days with a 95% probability . for initial ligo sensitivities , events with the same @xmath62
will be detectable at a rate of one per 22380 yr and for @xmath68 , one every 468 days .
although dns mergers with @xmath74 will be present in the data , they are unlikely to be temporally resolved . here
, we consider a signal comprised of an ensemble of chirp signals that result from the final mergers of cosmological dnss . even though the individual signals will not overlap , the cumulative background is stochastic because one can not predict with certainty the exact temporal evolution of the signal .
but , the gw amplitude distribution will be very different from a continuous astrophysical gw background .
for a signal that consists of a sum of many continuously emitting , randomly distributed sources , such as the gw signal from the galactic population of white dwarf - white dwarf or white dwarf - ns binaries , the central limit theorem predicts that the gw amplitude distribution will be gaussian .
the term ` confusion noise ' is often used to describe the signal from the galactic population of white dwarf - white dwarf binaries .
in contrast , the gw background from dns mergers has a non - gaussian signature that is characterized by the spatial distribution of the sources , so we use the term geometrical noise to describe it .
@xcite developed a procedure to simulate time series of the gw signals from supernovae occurring throughout the universe .
the method has provided a tool to probe the statistical signature of cosmological ` standard candle ' gw sources that follow the evolving star formation rate . in its simplest form , the simulation outputs a distribution of events in redshift as a function of observation time . here
, we use this procedure to extract the peh signature from a distribution of observer - source distances and observation times .
the peh technique is a procedure that records the time and distance of successively closer events , ( @xmath75 , @xmath76 ) , that satisfy the condition @xmath77 : each event in the peh time series occurs at a shorter distance than the preceding one .
a key feature of the procedure is that it probes the rate of an observer s penetration into the small probability ` tail ' of an event distribution .
@xcite applied this technique to the observed gamma - ray burst ( grb ) redshift and observation - time distribution and found that it can be used to set limits on the local rate density of grbs . because it is sensitive to the distribution of events in the small probability tail , in this case corresponding to smaller volumes , the procedure ` tracks ' the temporal evolution of close events as detected by an observer .
if a time series is comprised of more than one class of event , then the peh technique can identify these distributions based on the distribution of peh data .
this was shown by @xcite , where the peh model fitted to the data indicated that the anomalously close grb 980245 probably does not belong to the ` classical ' grb population .
as a further proof of concept , we apply the peh technique to a distribution of gw amplitudes from a simulated distribution of dns mergers throughout the universe .
we use a dns dimensionless characteristic gw amplitude @xmath78 at 20 mpc @xcite and a non - redshifted characteristic frequency @xmath79 hz . the results are plotted in figure [ algor ] .
the 5% and 95% peh curves are shown along with the output from the peh simulation shown as data points .
the 5% peh curve is interpreted as the ` null peh ' , as there is a 95% probability that no events will be observed at an amplitude above this threshold . in order to develop the peh concept into a technique for gw data analysis
, we need to incorporate a method for treating detector noise . for ` standard candle ' sources ,
this can be done by working in terms of progressively increasing amplitudes rather than progressively decreasing redshifts , as shown in figure 5 .
noise transients of varying amplitudes mimic sources at varying redshifts , enabling the introduction of a ` noise peh ' . here
we use a ` gaussian - noise peh ' ; this is based on a standard zero - mean gaussian distribution with standard deviation @xmath80 , modelled on idealized interferometric detector noise with @xmath81 denoting the root - mean - square amplitude of the noise and @xmath82 the sampling frequency see @xcite .
figure 5 shows the 5% and 95% gaussian - noise pehs . from kalogera et al .
we use a dns characteristic gw amplitude @xmath78 at 20 mpc ( schutz 1991 ) and a non - redshifted characteristic frequency @xmath79 hz .
the 5% ( or 95% probability of no events occurring ) and 95% peh curves for dns mergers are shown , along with the output of the peh simulation applied to a distribution of characteristic gw amplitudes modelled on dns mergers , shown as filled circles .
also shown are the 5% and 95% ` gaussian - noise pehs ' , modelled on idealized interferometric detector noise
see text for details .
the figure shows that the dns merger pehs evolve much faster than the gaussian - noise pehs . ]
[ algor ] it is clear from the small gradient of the gaussian - noise peh curves that the temporal evolution of transients , originating from the tail of the distribution is slow .
the striking feature of this figure is the gradient of the peh curves for the source amplitude distribution : it is much bigger than for the gaussian - noise case , because of the distance - amplitude relation .
we define this type of peh data as ` geometrical ' , because it is linked to the spatial distribution of sources .
we note that although we have incorporated sfr evolution and a cosmological model to calculate the dns merger rate at high @xmath28 , the peh will be most sensitive to events that are comparable to or greater than the detection threshold of the detector . for advanced ligo , this will be of order several hundred mpc ( @xmath83 ) , implying that the peh will be dominated by sources in the local universe where cosmological effects are minor .
given that a geometrical signature is present in a time series , one can fit a peh model to the data , with the local rate density as a free parameter . in the presence of noise ,
this signature will be modified , but if the detector noise is well characterized , a fitting function constructed from the sum of both noise and geometrical distributions can be used . for small signal - noise ratios , the noise will dominate the peh data for short observation times .
but as observation time increases , the geometrical signature will grow at a faster rate than the noise .
it is possible that the gw signal from dns mergers can be identified before that of a single merger , occurring above the detector threshold , as shown in figure [ fig3 ] .
we propose that there exists an intermediate detection regime , with pre - filtered signal - noise - ratio of less than unity , where the dns merger rate is high enough that the geometrical signature could be identified in interferometer data .
this type of search provides an opportunity to set upper limits on the dns merger rate .
the cumulative signal from transient gw sources at cosmological distances is commonly described as a stochastic background because of the temporal randomness of the individual events .
but we have shown here that , for a cosmic population of dns mergers , the gw cumulative signal will be dominated by the spatial distribution and temporal evolution of the sources .
we use the term ` geometrical ' noise , because the gw amplitude distribution composed of dns mergers is dominated by the spatial distribution of the sources .
interestingly , geometrical noise from cosmic dns mergers ( and similar transient gw sources ) is linked to the sensitivity of the detector and the cumulative event rate .
if a gw detector could probe dns mergers to distances corresponding to the time when they first occurred in the universe , the detector could potentially resolve all events and there would be no geometrical noise from unresolved signals .
however , if the cumulative rate measured in our frame were to be so high that the individual chirps would overlap , corresponding to a signal of duty cycle greater than unity , then the resulting signals would manifest as geometrical noise .
this is unlikely for dns mergers : the individual events are not expected to overlap but to be nonetheless not individually detectable because of the small signal - to - noise ratio .
we highlight the concept that the detectability of a gw stochastic background from dns mergers throughout the universe and of individual dns mergers is linked to the observer via the instrument duty cycle , the cumulative event rate and the sensitivity of the detector .
the peh model is applied above to the detectability of dns mergers with advanced and initial ligo ; it shows the temporal evolution of the gw signal from many faint unresolved events converging to detectable single events of high signal - to - noise ratio .
detection strategies based on the peh model could utilize the geometrical signature of the temporal evolution of the chirp gw amplitude distribution .
we plan to investigate the peh technique further using simulated gw background and non - gaussian noise data .
the peh method could also be used to set constraints on the black hole - black hole ( bh - bh ) merger rate , which is presently highly uncertain .
although occurring at a lower intrinsic rate than dns mergers , the gw emission from bh - bh mergers is expected to be significantly higher , implying an enhanced detection rate because of the larger detection volume .
estimates range from 304000 yr@xmath11 assuming a bh mass of 10 m@xmath84 and a detection horizon of @xmath85 @xcite for advanced ligo sensitivity .
we plan to use the bh - bh mass and merger rate distributions to determine if a peh model could be developed and applied to interferometer data to set limits on the rate and mass distributions .
although we have shown how the cumulative gw signal from dns mergers should manifest in a time series , we have not yet tested how robust the above procedure is to the non - stationarity of the data .
for instance , if a time series is plagued with non - gaussian transients , then the effectiveness of the fitting procedure will be reduced , unless a good model for the transients is available .
we assume optimal ( matched ) filtering can be applied to the candidate transients , but the filters that will be most effective will not be exactly matched to individual events .
the ` norm ' and ` mean ' filters , although not optimal , should perform at a significant fraction of a matched filter for transients where the signal is not completely characterized @xcite .
we emphasize that using the peh in a signal processing context is based on fitting to outliers using a model distribution and not ` detecting ' individual events .
nonetheless , it is important to filter the data to select as many transients that are candidates for dns mergers as possible . as a test of the robustness of the procedure using different filters , we plan to inject simulated cosmologically distributed gw transients into non - gaussian noise .
we note that the simulated gw amplitude distribution will be modified by the detector antenna pattern , which we have not included in our simulations , where we used the mean value . in applying a peh model to real data
, account must be taken of the modulation and distortion of the signal by detector characteristics and by both detector and environmental noise .
we plan to extend the current simulations by incorporating such effects as part of a program to test the sensitivity of the peh fitting procedure .
d. m. coward is supported by an australian research council fellowship and m. lilley is supported from the australian research council grant dp0346344 .
this research is part of the research program of the australian consortium for interferometric gravitational astronomy .
we thank dr paul saulson ( ligo ) and the referee for several helpful comments and suggestions .
99 ando s. 2004 , journal of cosmology & astroparticle physics , 0406 , 7 arnaud n. , et al .
2003 , phys .
d , 67 , 062004 belczynski k. , kalogera v. , bulik t. 2002 , apj , 572 , 407 burgay m. , et al . , 2003 , nature , 426 , 531 bromm j.s .
, loeb a. 2002 , apj , 575 , 111 chaurasia h.k . , bailes m. 2005 astro - ph/0504021 champion d.j .
, lorimer d.r .
, mclaughlin m. a. , cordes j.m . , arzoumanian z. , weisberg j.m . , taylor j.h .
2004 , mnras , 350 , l61 coward d.m .
, burman r.r .
, blair d. g. 2002a , mnras , 329 , 411 coward d. , burman r. , blair d. 2002b , class
. quantum grav .
19 , 1303 coward d.m .
, burman r.r .
mnras , 2005 , 361 , 362 faulkner a.j . , et al .
2005 , apj , 618 , l119 hughes s.a .
2002 , phys .
d , 66 , 102001 hulse r.a .
, taylor j.h .
1975 , apj , 195 , l51 kalogera v. , narayan r. , spergel d.n .
, taylor j.h .
2001 , apj , 556 , 340 kalogera v. , et al .
2004 , 614 , 137 kim c. , et al .
2004 astro - ph/0405564 lipunov v.m . ,
postnov k.a . , prokhorov m. e. 1997 , mnras , 288 , 245 nice d.j . ,
sayer r.w .
, taylor j.h .
1996 , apj , 466 , l87 lyne a.g . ,
mckenna j. 1989 , nat , 340 , 367 lyne a.g . , et al .
2000 , mnras 312 , 698 lyne a.g . ,
2004 , science , 303 , 1153 peebles p.j.e .
1993 , principles of physical cosmology , ( princeton , new jersey : princeton univ . press ) , p. 312
piran t. 1992 , apj , 389 , l45 porciani c. , madau p. 2001 , apj .
, 548 , 522 schutz b.f .
1991 , the detection of gravitational waves , ed .
blair , d. g. ( cambridge : cambridge university press ) , p. 411
taylor j.h . ,
weisberg , j.m .
1989 , apj , 345 , 434 thorne k.s . ,
cutler c. 2002 , general relativity and gravitation , ed . n. bishop & s. d. maharaj ( singapore : world scientific ) , p. 72
wolszczan a. 1991 , nature , 350 , 688 | gravitational waves generated by the final merger of double neutron star ( dns ) binary systems are a key target for the gravitational wave ( gw ) interferometric detectors , such as ligo , and the next generation detectors , advanced ligo .
the cumulative gw signal from dns mergers in interferometric data will manifest as ` geometrical noise ' : a non - continuous stochastic background with a unique statistical signature dominated by the spatial and temporal distribution of the sources . because geometrical noise is highly non - gaussian ,
it could potentially be used to identify the presence of a stochastic gw background from dns mergers .
we demonstrate this by fitting to a simulated distribution of transients using a model for the dns merger rate and idealized gaussian detector noise . using the cosmological ` probability event horizon ' concept and recent bounds for the galactic dns merger rate
, we calculate the evolution of the detectability of dns mergers with observation time . for advanced ligo sensitivities and a detection threshold assuming optimal filtering ,
there is a 95% probability that a minimum of one dns merger signal will be detectable from the ensemble of events comprising the stochastic background during 12211 days of observation . for initial ligo sensitivities ,
we identify an interesting regime where there is a 95% probability that at least one dns merger with signal - to - noise ratio @xmath0 unity will occur during 468 days of observation .
we propose that there exists an intermediate detection regime with pre - filtered signal - noise - ratio less than unity , where the dns merger rate is high enough that the geometrical signature could be identified in interferometer data .
[ firstpage ] gravitational waves binaries : neutron stars pulsars : general supernovae . |
belle collaboration observed a new resonance @xmath1 in the invariant mass spectrum of @xmath4 of @xmath5 @xcite .
its mass and width are @xmath6 mev and @xmath7 mev respectively . this charged state stimulated extensive studies of its nature @xcite .
theoretical explanations of @xmath1 include the s - wave threshold effect @xcite , the @xmath8 molecular state @xcite , the tetraquark state @xcite , the cusp effect @xcite and the @xmath9 bound state @xcite .
recently we studied whether @xmath1 could be an s - wave @xmath10 ( @xmath11 ) molecular state considering both the pion and sigma meson exchange potentials @xcite .
our numerical results show that there may exist an s - wave @xmath12 molecular state with @xmath13 .
if one ignores the width of @xmath14 , there may also exist an s - wave @xmath15 molecular state with @xmath16 .
however , @xmath14 will rapidly decay into @xmath17 before the formation of the @xmath11 bound state because of its large width around 384 mev .
in this work we assume @xmath1 is a @xmath12 molecular state with @xmath13 and study its two - body open charm decay modes , which may be used to distinguish the tetraquark and molecule picture .
this work is organized as follows .
after introduction , we discuss the hidden _ vs _ open charm decays of @xmath1 . in section [ amp ] , the decay amplitude of the two - body open charm decay of @xmath1 is given .
the last section is the numerical results and discussion .
assuming @xmath1 is an s - wave @xmath18 molecular state , we illustrate its strong decays allowed by okubo - zweig - iizuka ( ozi ) rule in fig .
[ haha ] . here
indices 1 , 2 , 3 , and 4 are @xmath19 , @xmath20 , @xmath21 and @xmath22 quarks respectively .
the dashed - line box denotes the s - wave @xmath23 molecular state .
[ haha ] ( a ) describes the hidden charm decay process @xmath24 by exchanging one charmed meson between @xmath25 and @xmath26 .
@xmath1 was observed in the @xmath4 channel .
if it s the s - wave @xmath18 molecular state , its quantum number is @xmath27 @xcite .
kinematically allowed hidden charm decays can be classified into ( 1 ) p - wave modes : @xmath28 , @xmath29 , @xmath30 , @xmath31 , @xmath32 , @xmath33 , @xmath34 , @xmath35 ; ( 2 ) s - wave modes : @xmath3 .
the neutral partner of @xmath1 may also decay into @xmath36 via p - wave and @xmath37 via s - wave .
@xmath1 was observed only in the @xmath4 channel up to now , which is very puzzling .
one of the possible reasons is the mismatch of the q - values of the initial and final states @xcite .
if so , one should also expect a signal in the @xmath38 channel since there is nearly no mismatch of the q - value .
another potential reason is the specific nodal structure in the wave functions of the final states , which was discussed by bugg recently @xcite .
meng and chao adopted the re - scattering mechanism to explain the suppression of the @xmath39 decay mode @xcite .
c + + ( a ) + + + + ( b ) besides the discovery mode @xmath4 , we want to emphasize that the s - wave @xmath40 mode is very important since there is no additional suppression factor @xmath41 where @xmath42 is the decay momentum and @xmath43 is the mass of @xmath1 .
naively one may expect that the orbital excitation in the @xmath44 meson easily transfers to @xmath45 .
moreover , the @xmath46 mode may help distinguish the @xmath47 of @xmath1 and test the @xmath18 molecular state assignment .
if the @xmath46 mode is observed , one can exclude the @xmath48 possibility for the @xmath18 system .
we strongly urge experimental colleagues to search @xmath1 in the @xmath49 channel .
both @xmath1 and @xmath44 mesons have finite widths .
@xmath1 can decay through the upper tail of its mass distribution into @xmath50 , especially the lower part of the @xmath44 mass distribution . as pointed out in ref .
@xcite , the dominant decay mode of @xmath1 is @xmath51 .
besides the above dominant mode , there exists another class of open charm decay mode through the re - scattering mechanism @xcite , which is shown in fig .
[ haha ] ( b ) .
one pion or @xmath52 ( @xmath53 ) meson exchange occurs between the light quarks .
these two - body decay modes include @xmath54 and @xmath55 .
the @xmath56 decay mode is forbidden for a pseudoscalar meson .
this kind of open charm decay is strongly suppressed if @xmath57 is a molecular state , as will be shown in the following sections .
however , @xmath57 will fall apart into these two - body final states very easily with large phase space if @xmath57 is a tetraquark state as proposed in refs .
@xcite . therefore
, these two - body decay modes can be used to distinguish the molecular and tetraquark pictures .
this is the key motivation of the present work .
assuming @xmath1 is an s - wave @xmath23 molecular state with @xmath48 @xcite , we derive the two - body open charm decay amplitude using the optical theorem . from the unitarity of the s - matrix @xmath58 , one obtains @xmath59 where @xmath60 .
we sandwich the above equation between the one - particle state @xmath61 and two - particle state @xmath62 . by inserting a complete set of intermediate states to the right - hand side
, one gets @xmath63 now the absorptive part reads @xmath64\mathcal{m}^*[p\to \{p_{i}\}].\label{abs}\end{aligned}\ ] ] the optical theorem relates the absorptive part of the @xmath65 decay amplitude to the sum of all possible @xmath66 decays and the re - scattering process @xmath67 . thus , for the two - body charm decay @xmath68 , the absorptive part of the decay amplitude can be written as @xmath69 \nonumber\\&=&\frac{|\mathbf{p}|}{32\pi^2 m_z}\int d\omega \mathcal{m}^*[z^+\to d^*\bar{d}_1]\nonumber\\&&\times\mathcal{m } [ d^*\bar{d}_1\to d^{(*)}\bar d^{(*)}]\end{aligned}\ ] ] where @xmath70 denotes the three momentum of intermediate state in the center of mass frame of @xmath1 . because the mass of @xmath1 is close to the sum of the masses of @xmath44 and @xmath71 ,
the dispersive part of @xmath72 amplitude is very important as shown in the case of x(3872 ) by meng and chao @xcite .
the dispersive part of the amplitude is related to the absorptive part through the dispersion relation , @xmath73 with @xmath74\nonumber\\&&\times\exp(-\eta @xmath75^{1/2}/2m_z$ ] is the three momentum of the intermediate state in the rest frame of @xmath1 .
@xmath76 is kllen function .
one chooses @xmath77 .
the exponential not only describes the dependence of the interaction between @xmath1 and @xmath12 on @xmath78 , but also plays the role of the cutoff .
the factor @xmath79 is related to the interaction radius @xmath80 by @xmath81 @xcite . in order to get @xmath82 $ ] and @xmath83 $ ]
, we use the following lagrangians @xcite @xmath84 @xmath85 @xmath86+h.c.,\end{aligned}\ ] ] where @xmath87 and @xmath88 are pseudoscalar and vector heavy mesons respectively , i.e. @xmath89=((@xmath90 , @xmath91 , @xmath92 ) . @xmath93 and @xmath94 denote the octet pseudoscalar and the nonet vector meson matrices .
the values of the coupling constants will be given in section [ numerical ] .
the lagrangian relevant to the interaction of @xmath44 with @xmath95 and @xmath96 mesons is given in ref .
@xcite .
the flavor wave function of @xmath1 is @xcite @xmath97 thus @xmath1 with @xmath48 can decay into @xmath98 , @xmath99 and @xmath100 via the intermediate states @xmath101 and @xmath102
. if one ignores the mass difference between neutral and charged charmed mesons , the contribution from the intermediate state @xmath101 is same as that from @xmath102 .
cc & + + ( a)&(b ) + + & + + ( c ) & fig . [ haha1 ] ( b ) describes the open charm decay @xmath104 .
the absorptive part of the decay amplitude of @xmath105 via the @xmath106 meson exchange is @xmath107\nonumber\\&=&\frac{|\mathbf{p}|}{32\pi^2 m_z}\int d\omega \;i [ i g_{z^+d_1 d^*}%\underline{\epsilon_1\cdot % \epsilon_2 } ] \nonumber\\&&\times[i(-i)\frac{g_{d^*dp}}{\sqrt { 2}}(iq^{\mu})%\underline{\epsilon_{2\mu } } ] ig_{d_1d^*\pi}[-3q^{\beta}q^{\nu}\nonumber\\&&+g^{\beta\nu}q^2-\frac{g^{\nu\beta}}{m_{d^*}m_{d_1}}(p_3\cdot q)(p_1\cdot q)]\epsilon_{3\beta}%\underline{\epsilon_{1\nu } } \nonumber\\&&\times \big[-g_{\nu\rho}+\frac{p_{1\nu}p_{1\rho}}{m_{d_1}^2}\big]\big[-g_{\mu}^{\rho}+\frac{p_{2}^{\rho}p_{2\mu}}{m_{d^*}^2}\big ] \nonumber\\&&\times\frac{\mathcal{f}^2(q^2,m_i^2)}{q^2-m_\pi^2}.\end{aligned}\ ] ] the amplitude from the @xmath95 exchange is @xmath108\nonumber\\&=&\frac{|\mathbf{p}|}{32\pi^2 m_z}\int d\omega \ ; i [ i g_{z^+d_1 d^*}%\underline{\epsilon_1\cdot % \epsilon_2 } ] \nonumber\\&&\times[-2i_1^{(i)}i f_{d^*dv}\varepsilon_{\mu\nu\alpha\beta}(iq^{\mu})%\underline{\epsilon_{\rho}^{\nu } } ( -ip_4^{\alpha}-ip_2^{\alpha } ) % \underline{\epsilon_2^{\beta } } ] \nonumber\\&&\times i_2^{(i)}%\underline{\epsilon_{\rho}^{\tau}}%\underline{\epsilon_{1}^{\eta } } [ i\varepsilon_{\tau\sigma\xi\kappa } \frac{p_1^{\kappa}}{m_{d_1}}\epsilon_{3}^{\xi}\theta^{\sigma}_{\eta}(g_{2,2}'+g_{1,2 } ' ) \nonumber\\&&+i\varepsilon_{\tau\sigma\eta\kappa } \frac{p_1^{\kappa}}{m_{d_1}}\epsilon_{3\gamma}\theta^{\sigma\gamma}(g_{2,2}'-g_{1,2 } ' ) + i\varepsilon_{\tau\eta\xi\kappa}\epsilon_{3}^{\xi}\frac{p_1^{\kappa}}{m_{d_1}}g_{1,0}']\nonumber\\ & & \times\big[-g^{\tau\nu}+\frac{q^{\tau}q^{\nu}}{m_{i}^2}\big]\big[-g^{\eta}_{\lambda}+\frac{p_1^{\eta}p_{1\lambda}}{m_{d_1}^2}\big ] \nonumber\\&&\times\big[-g^{\beta\lambda}+\frac{p_{2}^{\beta}p_{2}^{\lambda}}{m_{d^*}^2}\big ] \frac{\mathcal{f}^2(q^2,m_i^2)}{q^2-m_i^2}\end{aligned}\ ] ] with @xmath109 where the index @xmath110 in @xmath111 and @xmath112 denotes the @xmath95 meson . @xmath113 and @xmath114 .
the same notations are also used in the following subsections .
the open charm decay @xmath116 is depicted in fig .
[ haha1 ] ( c ) .
the absorptive part of the decay amplitude of @xmath117 from the @xmath95 meson exchange is @xmath118\nonumber\\&=&\frac{|\mathbf{p}|}{32\pi^2 m_z}\int d\omega \;i [ i g_{z^+d_1 d^*}%\underline{\epsilon_1\cdot % \epsilon_2 } ] \nonumber\\&&\times i_{1}^{(i)}[-2ig_{d^*d^*v}p_{2\nu}\epsilon_{4}^{\mu}-4if_{d^*d^*v}(q^{\mu}\epsilon_{4\nu}-g_{\nu}^{\mu } q\cdot\epsilon_4)]%\underline{\epsilon_{\rho}^{\nu}\epsilon_{2\mu } } ] \nonumber\\&&\times i_{2}^{(i)}%\underline{\epsilon_{1\kappa}}\underline{\epsilon_{\rho\xi } } \big[ig_{2,2 } \theta^{\kappa\xi}+ig_{1,0}(g^{\kappa\xi}-\frac{p_1^{\kappa}p_1^{\xi}}{m_{d_1}^2})\big]\big[-g^{\nu}_{\xi}+\frac{q_{\xi}q^{\nu}}{m_{i}^2}\big]\nonumber\\ & & \times\big[-g^{\sigma}_{\kappa}+\frac{p_1^{\sigma}p_{1\kappa}}{m_{d_1}^2}\big ] \big[-g_{\sigma\mu}+\frac{p_{2\sigma}p_{2\mu}}{m_{d^*}^2}\big]\frac{\mathcal{f}^2(q^2,m_i^2)}{q^2-m_i^2}.\nonumber\\\end{aligned}\ ] ] fig .
[ haha1 ] ( d ) corresponds to the process @xmath120 .
the absorptive part of the amplitude from the @xmath106 exchange is @xmath121\nonumber\\&=&\frac{|\mathbf{p}|}{32\pi^2 m_z}\int d\omega \;i [ i g_{z^+d_1 d^*}%\underline{\epsilon_1\cdot % \epsilon_2 } ] \nonumber\\&&\times[-\frac{ig_{d^*d^*p}}{2\sqrt { 2}}\varepsilon_{\mu\tau\alpha\xi}(iq^{\tau})%\underline{\epsilon_{2}^{\mu } } ( ip_2^{\alpha}+ip_4^{\alpha})\epsilon_4^{\xi } ] ig_{d_1d^*\pi}\nonumber\\&&\times[-3q^{\beta}q^{\nu}+g^{\beta\nu}q^2-\frac{g^{\nu\beta}}{m_{d^*}m_{d_1}}(p_3\cdot q)(p_1\cdot q)]\epsilon_{3\beta}%\underline{\epsilon_{1\nu } } \nonumber\\&&\times \big[-g_{\nu\rho}+\frac{p_{1\nu}p_{1\rho}}{m_{d_1}^2}\big]\big[-g^{\mu\rho}+\frac{p_{2}^{\rho}p_{2}^{\mu}}{m_{d^*}^2}\big ] \frac{\mathcal{f}^2(q^2,m_i^2)}{q^2-m_\pi^2}.\nonumber\\\end{aligned}\ ] ] for the @xmath52 meson exchange , the amplitude is @xmath122\nonumber\\&=&\frac{|\mathbf{p}|}{32\pi^2 m_z}\int d\omega \;i [ i g_{z^+d_1 d^*}%\underline{\epsilon_1\cdot % \epsilon_2 } ] \nonumber\\&&\times i_{1}^{(i)}[-2ig_{d^*d^*v}p_{2\nu}\epsilon_{4}^{\mu}-4if_{d^*d^*v}(q^{\mu}\epsilon_{4\nu}-g_{\nu}^{\mu } q\cdot\epsilon_4)]%\underline{\epsilon_{\rho}^{\nu}\epsilon_{2\mu } } % \underline{\epsilon_2^{\beta } } \nonumber\\&&\times i_2^{(i)}%\underline{\epsilon_{\rho}^{\tau}}\underline{\epsilon_{1}^{\eta } } [ i\varepsilon_{\tau\sigma\xi\kappa } \frac{p_1^{\kappa}}{m_{d_1}}\epsilon_{3}^{\xi}\theta^{\sigma}_{\eta}(g_{2,2}'+g_{1,2 } ' ) \nonumber\\&&+i\varepsilon_{\tau\sigma\eta\kappa } \frac{p_1^{\kappa}}{m_{d_1}}\epsilon_{3\gamma}\theta^{\sigma\gamma}(g_{2,2}'-g_{1,2 } ' ) + i\varepsilon_{\tau\eta\xi\kappa}\epsilon_{3}^{\xi}\frac{p_1^{\kappa}}{m_{d_1}}g_{1,0}']\nonumber\\ & & \times\big[-g^{\nu\tau}+\frac{q^{\tau}q^{\nu}}{m_{i}^2}\big]\big[-g^{\sigma\eta}+\frac{p_1^{\sigma}p_{1}^{\eta}}{m_{d_1}^2}\big ] \big[-g_{\sigma\mu}+\frac{p_{2\sigma}p_{2\mu}}{m_{d^*}^2}\big]\nonumber\\&&\times\frac{\mathcal{f}^2(q^2,m_i^2)}{q^2-m_i^2}.\end{aligned}\ ] ] throughout this section , @xmath123 is the form factor to describe the structure effect in the vertex of the re - scattering process @xmath124 .
we use the monopole form @xcite @xmath125 where @xmath126 with @xmath127 mev and @xmath128 .
we collect the values of these coupling constants below @xmath129 where @xmath130 , @xmath131 gev@xmath132 , @xmath133 are the parameters in the effective lagrangian @xcite .
casalbuoni and his collaborators extracted @xmath134 gev@xmath132 with the available experimental data @xcite . in ref .
@xcite , zhu and dai gave @xmath135 , @xmath136 , @xmath137 , @xmath138 , @xmath139 , @xmath140 and @xmath141 gev@xmath142 in the framework of qcd sum rule .
for the parameter @xmath79 , we use @xmath143 gev@xmath142 @xcite .
the meson masses are from pdg : @xmath144 mev , @xmath145 mev .
@xmath146 mev , @xmath147 mev , @xmath148 mev , @xmath149 mev , @xmath150 mev , @xmath151 mev @xcite . with the above parameters
, we plot the dependence of the branching ratio of the two - body open charm decays @xmath152 on @xmath153 in fig .
[ decay ] . in table
[ table ] , we also list show the branching ratios of @xmath154 , @xmath155 , @xmath156 modes with typical values of @xmath153 .
c in short summary , we have calculated the two - body open charm decay widths assuming @xmath1 is a pseudoscalar molecular state .
our numerical results show that the branching ratios of the two - body decays @xmath157 are around @xmath158 , which is strongly suppressed compared to the main decay modes @xmath159 if it s a molecular state with @xmath48 .
however , they are only moderately suppressed compared with the discovery mode @xmath4 . as an excellent candidate of the exotic states ,
the observation of @xmath1 has stimulated several speculations about its structure .
the tetraquark scheme seems an interesting choice . however , there is no reliable dynamics forbidding the quarks and anti - quarks to regroup and fall apart . with plentiful phase space @xmath1
will easily fall apart into final states such @xmath28 , @xmath29 , @xmath32 , @xmath34 , @xmath3 , @xmath160 etc if @xmath1 is a tetraquark .
especially the two - body open charm modes @xmath161 etc will become one of the main decay modes , in sharp contrast with the case of @xmath1 being a molecular state .
thus these modes provide a smoking gun to distinguish the molecular and tetraquark schemes .
we strongly urge experimentalists to look for these two - body open charm decay modes together with the hidden charm mode @xmath40 .
.the branching ratios with typical values of @xmath153.[table ] [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ]
this project was supported by the national natural science foundation of china under grants 10625521 , 10721063 , 10705001 and the china postdoctoral science foundation ( 20060400376 ) .
was also supported by the _ fundao para a cincia e a tecnologia of the ministrio da cincia , tecnologia e ensino superior _ of portugal ( sfrh / bpd/34819/2007 ) .
rosner , phys .
d 76 * , 114002 ( 2007 ) . c. meng and k.t .
chao , arxiv:0708.4222 [ hep - ph ] .
x. liu , y.r .
liu , w.z .
deng and s.l .
zhu , phys . rev .
* d 77 * , 034003 ( 2008 ) , arxiv:0711.0494 .
x. liu , y.r .
liu , w.z .
deng , s.l .
zhu , arxiv:0803.1295 [ hep - ph ] .
ding , arxiv:0711.1485 [ hep - ph ] .
lee , a. mihara , f.s .
navarra and m. nielsen , arxiv:0710.1029 [ hep - ph ] .
l. maiani , a.d .
polosa and v. riquer , arxiv:0708.3997 [ hep - ph ] .
k. cheunga , w.y .
keung and t.c .
yuan , phys . rev .
* d 76 * , 117501 ( 2007 ) .
gershtein , a.k . likhoded and g.p .
pronko , arxiv:0709.2058 [ hep - ph ] .
bugg , arxiv:0709.1254 [ hep - ph ] .
bugg , arxiv:0802.0934 [ hep - ph ] .
qiao , arxiv:0709.4066 [ hep - ph ] .
e. braaten and m. lu , arxiv:0712.3885 [ hep - ph ] .
y. li , c.d .
l and w. wang , phys . rev . *
d 77 * , 054001 ( 2008 ) .
liu , q. zhao and f.e .
close , arxiv:0802.2648 [ hep - ph ] .
b. zhang , x. liu and s.l .
zhu , chin .
* 24 * , 2537 - 2539 ( 2007 ) , arxiv:0705.3082 [ hep - ph ] ; x. liu , b. zhang , l.l .
shen and s.l .
zhu , phys . rev . * d 75 * , 074017 ( 2007 ) , arxiv : hep - ph/0701022 [ hep - ph ] . | the two - body open charm decays @xmath0 occur through the re - scattering mechanism and their branching ratios are strongly suppressed if @xmath1 is a @xmath2 molecular state .
in contrast , @xmath1 falls apart into these modes easily with large phase space and they become the main decay modes if @xmath1 is a tetraquark state .
experimental search of these two - body open charm modes and the hidden charm mode @xmath3 will help distinguish different theoretical schemes . |
experimental studies of quantum information processing ( qip ) has progressed much in recent years . in studies of qip using trapped ions , realization of small scale computation and entanglement generation with relatively high fidelity has been reported @xcite , which reveals that there is no fundamental obstacle to scaling a large number of ions .
the upcoming challenges are large - scale operations and high - fidelity gate operations toward fault - tolerant quantum computation . as a way to realize high - fidelity gate operations , quantum gates and quantum computation using geometric phase factors have recently been studied .
this originates from holonomic quantum computation ( hqc ) proposed by zanardi _
et al._@xcite in hqc , degenerate multiple quantum states are utilized and unitary operations are performed by varying the system hamiltonian along a closed path in the parameter space .
the final state in the hqc is dependent only on the global property of the closed path ; therefore , hqc is considered to be robust against certain types of errors . even when diabatic evolutions of the system or nondegenerate quantum states are used , similar advantages can be expected as long as the unitary operations performed are determined by geometric phase factors [ geometric quantum computation ( gqc)@xcite ] .
there have been a number of experiments in different systems related to gqc @xcite .
it is known that universal quantum computation can be realized with a combination of single - qubit and two - qubit operations @xcite . as demonstrations of two qubit operations in trapped - ion systems , `` geometric - phase gate '' by leibried _
et al._@xcite and the mlmer - srensen gate @xcite have been realized .
a gate fidelity of 99.3% has been realized using such a scheme @xcite .
single qubit operations in trapped - ion systems have been performed to date using a dynamical method with variable pulse lengths and phases . by replacing such a method with those that use geometric phases , gate operations that are robust against variation of parameters such as pulse intensity and lengths can be expected . in this work
, we report the realization of purely geometric single qubit operations using a four - level system in a single @xmath0ca@xmath2 ion .
three transitions in the four - level system are excited either by three optical fields or by two optical fields and one rf magnetic field .
rotation operations by arbitrary angles along two different axes , @xmath3 and @xmath4 in the bloch sphere for the qubit , are demonstrated by utilizing stimulated raman adiabatic passage ( stirap ) in this four - level system .
robustness against parameter variations is also demonstrated .
this work is based on proposals using a _ tripod _ system comprising of one upper state and three lower states connected by three oscillating fields @xcite .
there are two dark states in the system , and those dark states are adiabatically manipulated with the intensities and phases of the oscillating fields to perform single - qubit operations .
there was also a proposal and demonstrations of single - qubit operations using geometric phase factors using a two - level system and square pulses @xcite .
however , to the best of our knowledge no adiabatic ( or holonomic ) demonstration of single - qubit operations using dark states has been reported so far .
here the formalization of the scheme is summarized based on kis and renzoni @xcite for later reference .
a four - level system comprising of @xmath5 is considered [ see fig . [
fig : fig1](a ) ] .
@xmath6 and @xmath7 span a qubit manifold , while @xmath8 represents an upper state and @xmath9 represents an auxiliary state used in stirap .
three near resonant oscillating fields are applied to this system between @xmath8 and @xmath10 , and the hamiltonian in the interaction picture with the rotating wave approximation is given as @xmath11 , % \]\ ] ] where @xmath12 is the detuning of the lasers , which is assumed to be common to all the three transitions .
the rabi frequencies for the transitions between @xmath8 and the qubit states are @xmath13 ( assumed to be real ) and @xmath14 , where a real quantity @xmath15 is the common envelope function . @xmath16 and @xmath17 are assumed to be independent of time . the rabi frequency between @xmath8 and @xmath9 is described as @xmath18 with @xmath19
. the following states in the qubit manifold are useful for describing the hamiltonian and explaining the gate procedure : @xmath20 and @xmath21 , and similarly , the following states in the total lower - state manifold spanned by @xmath22 : @xmath23 and @xmath24 , where @xmath25 . based on these
, the hamiltonian can be simplified as follows : @xmath26 where @xmath27^{1/2}$ ] . when @xmath28 , this hamiltonian has four eigenvectors @xmath29 with eigenvalues of @xmath30 , respectively , where @xmath31 .
we proof the ability to perform arbitrary single - qubit operations by first starting from an arbitrary initial state in the qubit manifold , @xmath32 , which can be rewritten using @xmath33 and @xmath34 as @xmath35
. this initial state will be transferred in the first stirap sequence that makes use of @xmath36 [ the left half in fig.[fig :
fig1](b ) , with @xmath37 and @xmath38 to the intermediate state @xmath39 .
this intermediate state will then be transferred using the second stirap sequence [ the right half in fig .
[ fig : fig1](b ) , with @xmath40 and @xmath41 to the final state @xmath42 , which can be rewritten using the identity operator @xmath43 and the pauli operators @xmath44 on the computational manifold @xmath45 as @xmath46\left|\psi_0\right\rangle$ ] with @xmath47 .
this operation corresponds to a rotation by angle @xmath48 around @xmath49 , which can be taken arbitrarily by selecting the values for @xmath50 and @xmath51 , hence arbitrary single - qubit operations can be performed .
this scheme is based on adiabatic population transfer using @xmath36 , and diabatic transitions from this state to @xmath52 are among the possible causes of infidelity in gate operations .
the probability for such diabatic transitions is calculated to be in the order of @xmath53$]@xcite , which should be set to be much smaller than 1 to maintain high fidelity .
the experimental setup has been previously described @xcite and only a brief description is given here .
a single @xmath0ca@xmath1 is trapped in vacuum ( @xmath54 pa ) using a linear paul trap .
the trap used here is a conventional linear trap with an operating frequency of 23 mhz and secular frequencies of @xmath55 mhz . a bias magnetic field of @xmath56 t
is applied to define a quantization axis , which results in a zeeman splitting of @xmath574.9 mhz between @xmath58 sublevels . in experiments that employ three optical fields ,
a titanium sapphire laser at 729 nm stabilized to a high - finesse low - thermal - expansion cavity is used for the excitation of ions between @xmath59 and @xmath58 .
the amplitudes and frequencies of the three optical fields at 729 nm are changed by varying the three rf fields that are combined and fed to an acousto - optic modulator .
the rf fields are generated by three direct - digital synthesis ( dds ) boards that are controlled by a field - programmable gate array .
polarization of the optical fields is adjusted so that there are polarizations both parallel and perpendicular to the bias magnetic field ; therefore , the transition to be excited is selected by changing the frequencies of the dds boards . in experiments that employ two optical fields and one rf magnetic field , an rf coil in the vicinity of the trap is used to generate the rf field .
the details of the rf excitation procedure are similar to that described in ref .
@xcite .
we have chosen the @xmath59@xmath58 electric - quadrupole transition of @xmath0ca@xmath1 for realizing the geometric phase gate .
the encoding of the tripod system to the sublevels in @xmath59 and @xmath58 is as follows : @xmath60 as the `` upper '' state @xmath8 and three zeeman sublevels in @xmath58 ( @xmath61 ) as the lower states @xmath6 , @xmath7 and @xmath9 , respectively ( see fig . [
fig : fig1](c ) ) .
the reason that a level in @xmath59 is adopted instead of one in @xmath62 as @xmath8 is to avoid the effects of spontaneous emissions when all the fields are resonant .
however , unwanted couplings between @xmath63 and @xmath6 , @xmath7 and @xmath9 must still be considered .
since these couplings are off resonance , varying @xmath64 leads to time - dependent ac - stark shifts that disturb the null eigenenergies of the qubit states .
this effect was numerically evaluated with realistic parameters and the phase accumulated during the gate operation was confirmed to be smaller than 0.5 rad under typical experimental conditions in the present work .
we essentially do not compensate such ac - stark shifts for the present work . in the experiments in [ sec : xzrotations ] and of rotations by angle @xmath65 in [ sec : fidelity ] , we adjust one of the detunings ( one for @xmath8@xmath6 ) so that fringe shift resulting from ac - stark shifts are apparently canceled .
this cancelation depends on the details of the stirap pulses , including such conditions as peak rabi frequencies , pulse shapes or total time ( it may not depend on such conditions as input states or rotation angles since ac stark shifts do not depend on these ) .
we also evaluate gate performances without cancellation of this type by examining rotations by angle @xmath66 with relatively low rabi frequencies in [ sec : fidelity ] and [ sec : discussions ] .
the procedure for gate operations is as follows .
firstly , the ion is cooled to near the motional ground state with doppler cooling by 397 and 866 nm and with sideband cooling of the axial motion by lasers at 729 and 854 nm .
the average motional quantum number along axial direction after sideband cooling is @xmath67 .
optical pumping using a 397 nm @xmath68 transition is then performed to initialize the ion in @xmath60 ( @xmath8 ) .
the ion is transferred from @xmath8 to the computational manifold spanned by @xmath45 using square @xmath65 pulses , and then a gate stirap pulse sequence is applied between @xmath6 and @xmath7 .
the pulse shape of the gate consists of partially overlapping sinusoidal curves and constant values , as in fig .
[ fig : fig1](d ) . a step variation of @xmath69 by @xmath70 , which brings a geometric phase factor , is given at the middle of the gate sequence where @xmath71 .
after the gate sequence , a square @xmath65 pulse is applied to map the state of @xmath6 to @xmath8 .
an additional pulse may also be applied , depending on the element of the density matrix to be observed .
state discrimination is performed by detecting fluorescence from the ion with a photomultiplier tube during a period of 7 ms when 397 and 866 nm fields are applied@xcite .
the time dependence of the optical pulses is chosen as follows ( @xmath72 , @xmath73 , @xmath74 and @xmath75 where @xmath76 ) : @xmath77 @xmath78 where @xmath79 and @xmath80 are the maximum values for @xmath81 and @xmath82 , respectively . in the experimental results given in this article , the values of @xmath83
were empirically chosen to in between 0.55 and 0.85 so that effective population transfer was attained in experiments and/or numerical simulations .
it is noted that the values of @xmath83 used in this article are not thoroughly optimized ones but those that give relatively good performances among tested .
we show that arbitrary one - qubit operations can be realized with the geometric method introduced here .
this can be accomplished by showing @xmath3- and @xmath4-rotations independently , because arbitrary operations can be decomposed into sequences of such rotations .
the procedure for @xmath3-rotations is as follows .
the ion is first prepared in @xmath8 by sideband cooling and optical pumping , then initialized to @xmath6 , and a stirap sequence is applied to perform a geometric gate .
after this , the populations of the final states are analyzed by mapping the qubit states to the optical transition using a rectangular @xmath65 pulse on @xmath84 ( @xmath85 ) .
the populations in @xmath9 and @xmath8 are also measured to verify that the gate operation is closed to the manifold spanned by the qubit states @xmath45 .
@xmath86 and @xmath87 are adjusted to be equivalent , and the peak values for @xmath88 are @xmath89 khz .
the total time for the stirap sequence is set to be @xmath57120 @xmath90s by taking appropriate account of the adiabaticity .
@xmath91 is used for the results given in this section .
figure [ fig : xzrotations](a ) shows the results for @xmath3-rotations .
a sinusoidal oscillation of population between @xmath6 and @xmath7 is observed .
the sum of populations in the other states @xmath9 and @xmath8 is below 0.05 over the entire region , with which we confirm that the rotation operations are almost limited to the computational manifold spanned by @xmath45 .
the contrasts of the populations of @xmath6 and @xmath7 are obtained by fitting to @xmath92 and @xmath93 , respectively .
the procedure for observation of @xmath4-rotations is similar to ramsey interferometry .
a superposition of the qubit states @xmath6 and @xmath7 is first prepared by application of a @xmath66 pulse on @xmath8@xmath7 and a @xmath65 pulse on @xmath8@xmath6 .
a @xmath4-rotation is then performed with stirap , which is followed by a @xmath65 pulse on @xmath8@xmath6 and a @xmath66 pulse on @xmath8@xmath7 .
the last pulse causes an interference between @xmath8 and @xmath7 , which is detected by a projection measurement using fluorescence . here
the peak values for @xmath88 are set to be @xmath57@xmath94 khz and the total time for the stirap sequence is @xmath57120 @xmath90s .
figure [ fig : xzrotations](b ) represents the results for @xmath4-rotations , which shows a population oscillation against @xmath70 with an almost unit contrast . from a least - squares fit
, the contrast is obtained to be @xmath95 .
in order to investigate the action of gate operations more quantitatively , we also performed estimation of gate fidelities by using quantum state tomography .
we performed fidelity estimation in the following three cases : ( a ) x - rotation by angle @xmath65 , ( b ) z - rotation by angle @xmath65 , and ( c ) hadamard gate [ rotation around @xmath96 in the bloch sphere by angle @xmath65 ] .
as explained before , general operations of the stirap gate can be described as the following unitary operator : @xmath98\end{aligned}\ ] ] where @xmath99@xmath100@xmath101@xmath102@xmath103 . using this general expression ,
the unitary operators for the three cases given above are respectively written as follows : @xmath104 the initial states in the three cases are respectively prepared as follows : ( a)@xmath105 , ( b)@xmath106 , and ( c)@xmath107 . in order to take into account imperfect preparation of these initial states
, we performed measurement of these states using the technique of quantum state tomography . in the encoding adopted here , both the two states @xmath6 and @xmath7 in the computational subspace are in the same electronic state @xmath58 . in order to discriminate these two states , we applied a mapping pulse between @xmath6 and @xmath8 before performing fluorescence detection .
the population out of the computational manifold ( population in @xmath7 and @xmath8 ) before application of the mapping pulse , which is typically below 0.05 as described before , is ignored here for simplicity .
the mapping pulse itself produces a geometric phase relative to the other states ( @xmath7 and @xmath9 ) , which should be taken into account properly in the fidelity analysis .
the mapping operation can be described as follows : @xmath108 this corresponds to a rotation by angle @xmath65 around the @xmath109 axis in the bloch sphere of the two - level system @xmath110 .
the density matrix after application of the mapping pulse were reconstructed ( using linear reconstruction ) by fluorescence detection and an optional @xmath66 pulse prior to that .
the populations along three orthogonal axes in the bloch sphere , @xmath111,@xmath112 and @xmath113 , are measured , where @xmath111 ( @xmath112 ) is the population in @xmath58 after the mapping pulse and a @xmath66 pulse with the phase @xmath114 ( @xmath115 ) on @xmath8@xmath7 , and @xmath113 is the population in @xmath58 state immediately after the mapping pulse .
using the bloch vector @xmath116@xmath117@xmath118 the density operator after the mapping pulse is expressed as @xmath119 and by using this the initial density operator is described as follows : @xmath120 in a similar way , the final density operator just after stirap gate operations , @xmath121 , is obtained in terms of populations @xmath122,@xmath123 and @xmath124 ( defined in the same way as above ) after a mapping pulse that follows the gate operation .
the gate fidelities in the three cases can be described using the density operators and the unitary operators , as follows : @xmath125 the fidelities are explicitly written in terms of the initial and final populations , as follows : @xmath126 in order to determine the confidence intervals of the fidelities , we obtained variances of the fidelities considering propagation of uncertainty based on the above expression , as follows : @xmath127 where @xmath128 represents the variance of each quantity .
the variances of the populations [ such as @xmath129 are simply determined here as the variances in binomial distributions ; _ e.g. _
@xmath130 , where @xmath131 is the number of experiments per one measurement condition . the confidence intervals ( 68% ) for the fidelities
are determined as the square roots of the variances of the fidelities .
table [ tab1 ] shows the results for the initial and final population measurements .
the measurements were performed in essentially the same conditions as used in the previous section . using the measured populations ,
the fidelities are estimated and shown in table [ tab2 ] .
the second column in table [ tab2 ] shows the fidelities reflecting both initial and final populations .
we should note that these values are largely affected by imperfect initialization and analysis and therefore should be rather taken as the lower limits for the fidelities . for the purpose of reference
, we also estimated the fidelities when ideal initial or final states are assumed ( the third and fourth columns in table [ tab2 ] , respectively ) .
the fidelities for ideal initial states ( the third column ) represent the goodness of the generated states without considering initialization errors .
these values that are better than those in the second column can be considered as the upper limits for the fidelities ( when all operations other than the gates are assumed to be perfect ) .
the fidelities for ideal final states ( the fourth column ) represent the goodness of the initialization process .
these values support the observation that the values in the second column is largely affected by imperfect initialization ( and possibly imperfect analysis ) . lllllll gate type&@xmath111&@xmath112&@xmath113&@xmath122&@xmath123&@xmath124 + x-@xmath65 & 0.510&0.549&0.026&0.463&0.555&0.985 + z-@xmath65 & 0.509&0.045&0.502&0.472&0.973&0.499 + hadamard&0.497&0.515&0.025&0.485&0.989&0.488 + llll gate type&fidelity&fidelity & fidelity + & & ( ideal initial&(ideal final + & & states)&states ) + x-@xmath65 & [email protected]&[email protected]&[email protected] + z-@xmath65 & [email protected]&[email protected]&[email protected] + hadamard&[email protected]&[email protected]&[email protected] + we also performed fidelity analysis for stirap gates with rotation angle of @xmath66 , which are more sensitive to coherence between stirap and square - pulse operations and hence to ac stark shifts .
table [ tab3 ] shows the results for the initial and final population measurements for x and z gates .
either @xmath133 or @xmath134 as specified are used as the initial state .
it should be noted that this time ac stark shifts are not compensated at all , and relatively small values of the rabi frequencies are used to suppress their effect .
the peak values for @xmath88 are set to be @xmath57@xmath135 khz for x gates and @xmath57@xmath136 khz for z gates .
the total time for the stirap sequence is @xmath57290 @xmath90s and @xmath137 is used .
the results of population measurements are shown in table [ tab3 ] . using these populations , the fidelities and their confidence intervals
are estimated in the same way as above and shown in table [ tab4 ] .
it should be noted that the fidelities in the case of ideal final states in table [ tab4 ] are higher than what are given in table [ tab2 ] .
when taking the results in table [ tab3 ] , we used a noise eater with a sample and hold capability to reduce fluctuations in the amplitudes of the square pulses .
we speculate that this helped improving the operational fidelity of the initialization .
lllllll gate type&@xmath111&@xmath112&@xmath113&@xmath122&@xmath123&@xmath124 + x-@xmath66(@xmath138 prep . ) & 0.009 & 0.505 & 0.515 & 0.029 & 0.611 & 0.622 + x-@xmath66(@xmath139 prep . ) & 0.526 & 0.540 & 0.006 & 0.595 & 0.916 & 0.363 + z-@xmath66(@xmath138 prep . ) & 0.009 & 0.505 & 0.515 & 0.592 & 0.055 & 0.515 + z-@xmath66(@xmath139 prep . ) & 0.526 & 0.540 & 0.006 & 0.546 & 0.534 &
0.041 + llll gate type&fidelity&fidelity & fidelity + & & ( ideal initial&(ideal final + & & states)&states ) + x-@xmath66(@xmath138 prep . ) & [email protected] & [email protected] & [email protected] + x-@xmath66(@xmath139 prep . ) & [email protected] & [email protected] & [email protected] + z-@xmath66(@xmath138 prep . ) & [email protected] & [email protected] & [email protected] + z-@xmath66(@xmath139 prep . ) & [email protected] & [email protected] & [email protected] +
we have also attempted encoding of the qubit into optically separated levels .
use of optically separated levels may be advantageous when considering a combination with optical two - qubit gate schemes @xcite and use of less magnetic - field - sensitive transitions , such as @xmath140@xmath141 .
this encoding uses an rf transition between the ground zeeman sublevels @xcite .
figure [ fig : fig1](e ) shows the levels used for this encoding .
@xmath5 are encoded into @xmath140 , @xmath142 , @xmath143 and @xmath144 , respectively .
[ fig : xzrotations](c ) shows the results for @xmath3-rotations with qubit encoding to optically separated levels .
blue hollow circles ( red filled circles ) represent the populations in @xmath6 when @xmath6 ( @xmath7 ) is initially prepared .
the contrasts are @xmath145 and @xmath146 for the preparation of @xmath6 and @xmath7 , respectively .
[ fig : xzrotations](d ) shows the result for @xmath4-rotations , with a contrast of @xmath147 . in these results ,
the peak values of @xmath88 are set to be @xmath148 khz and @xmath149 khz for @xmath3- and @xmath4-rotations , respectively .
the total time for the stirap sequence is approximately 128 @xmath90s . in ideal cases of @xmath3-rotations ,
the peak values of @xmath86 and @xmath87 should be equal , which is not the case in the experiment described above .
for a technical reason concerning the difference in the amplitude modulation of rf and optical fields , the temporal shapes of @xmath86 and @xmath87 are not proportional to each other in the present setup .
the peak values given above are determined empirically so that they give correct @xmath3-rotations .
this imperfection may be avoided by simply calibrating the amplitude modulation process for either the rf or optical field so that @xmath86 and @xmath87 give exactly the same time dependency .
the feature of the geometric phase gate is that it is expected to be robust against variations in the pulse area , namely the rabi frequency and the pulse length . to study this feature ,
two experiments were conducted .
one is the measurement of @xmath3-rotations with variation of the peak rabi frequency and the illumination period .
the other is the measurement of @xmath4-rotations with variation of the peak rabi frequency , while the illumination period is held constant .
the ratios of the peak rabi frequencies are held constant in each case .
both experiments are performed by measuring population oscillations as the rotation angle @xmath150 is varied .
the results for the former experiment is shown in fig.[fig : robustness ] ( a ) and ( b ) for three values of the illumination period , with ( a ) fringe contrasts and ( b ) shifts .
the ratios of the rabi frequencies are held constant , and the peak values of @xmath151 are shown in the horizontal axes as representatives of the rabi frequencies .
@xmath152 was used for the results given in fig .
[ fig : robustness ] , which gave relatively high contrasts for the case of @xmath153 khz .
there is a region ( above 10@xmath5730 khz ) over which the absolute values of the fringe contrasts are almost independent of the rabi frequency , which demonstrates the robustness .
the fringe shifts are expected to start from zero and increase quadratically as the rabi frequency is increased due to the increase of ac stark shifts .
the observed shifts shown in fig . [ fig : robustness ] ( b ) basically follow this expectation .
the results for the latter experiment are shown in fig.[fig : robustness](c ) and ( d ) with ( c ) fringe contrasts and ( d ) shifts .
the fringe contrasts do not change appreciably as the peak rabi frequency is varied in the region above 20 khz .
we have performed numerical simulations trying to explain the loss of fringe contrasts and the shifts .
simulations are performed using a liouville equation with decay terms representing laser - frequency and magnetic - field flucutuations .
ac stark shifts due to different zeeman components and electric - dipole - allowed transitions ( @xmath59@xmath154 , @xmath59@xmath62 and @xmath58@xmath62 ) are fully taken into account as time - dependent variation of the detunings .
the results are shown as curves in fig .
[ fig : robustness ] , where overall qualitative and partial quantitative agreements between experimental and numerical results are obtained .
the possible factors for loss of fidelity in the previously described results includes laser frequency and magnetic field fluctuations .
the laser linewidth in the current setup is @xmath155 hz and the magnetic field fluctuations corresponds to fluctuations of resonance frequencies of up to @xmath156 hz .
the effects of thermal distribution of motional quantum numbers and intensity fluctuations are negligible , because the scheme used here is robust against variations in rabi frequencies .
we performed numerical simulations ( similar to what is described in the previous section ) to evaluate the fidelity of some of the gates given in in [ sec : fidelity ] . for x - rotation by angle @xmath65 and z - rotation by angle @xmath65 ,
the dominant infidelity factor was the effect of magnetic field fluctuations ( @xmath571 % ) .
infidelity due to diabaticity was estimated to be @xmath570.2% .
the contribution from laser - frequency fluctuations in this case was below 0.1 % .
this can be explained by the fact that the scheme is not sensitive to one - photon detunings but only to two - photon detunings ( as will be described later in this section ) .
the contribution from laser - frequency fluctuations may be larger when the encoding as in [ sec : rf ] is used .
ac stark shifts were compensated in this case by detuning one of the beams as described before .
the possible fidelity loss from not performing compensation would amount to 2030 % in our simulation .
for the results of x- and z - rotations by angle @xmath66 , the largest infidelity factor was diabaticity ( 25 % ) , and the effect of magnetic field fluctuations was 13 % .
although no ac - stark - shifts compensation was performed in this case , infidelity due to ac stark shifts was as small as 01 % , and the effect of laser - frequency fluctuations was 0.20.3 % .
the scheme demonstrated here is useful not only for atomic systems ( neutral or ionic ) , but also for other systems including solid - state systems .
the scheme is effective for those systems where the coherence time is relatively long , but inhomogeneities of the excitation field intensities can not be avoided , or for those where slow intensity fluctuations of excitation fields occur .
it can also be used effectively in assuring equal operations in a large ensemble of particles .
the requirements for the scheme are that a sufficient number of levels are available for implementation , and that energy shifts due to excitation fields ( such as ac stark shifts ) can be avoided .
elimination of ac stark shifts is not a straightforward task for the scheme presented here since a number of fields with time - dependent amplitudes are used .
it can be confirmed from numerical simulations that the operations of the stirap gates are not sensitive to one - photon detunings ( i.e. detunings for @xmath157@xmath8 where @xmath158 and @xmath159 ) but to every two - photon detuning , and hence all relative shifts between two of @xmath6 , @xmath7 and @xmath9 should be minimized .
this might be relatively complicated for _ e.g. _ @xmath0ca@xmath1 since , in addition to adjacent zeeman components in @xmath59@xmath58 , dipole - allowed transitions such as @xmath59@xmath154 , @xmath59@xmath62 and @xmath58@xmath62 also give rise to appreciable ac stark shifts@xcite .
the scheme demonstrated here realizes single - qubit operations that are in general noncommutable to each other by using adiabatic manipulation of dark states .
these amount to implementations of noncommutable ( non - abelian ) holonomies , which can be a building block of hqc by zanardi _
et al._@xcite and leads to application in other fields including neutral - atom systems @xcite .
this work was supported by the kakenhi `` quantum cybernetics '' project of the ministry of education , culture , sports , science and technology ( mext ) in japan and the japan society for the promotion of science ( jsps ) through its funding program for world - leading innovative r&d on science and technology ( first program ) .
j. p. home _
et al . _ ,
science * 325 * , 1227 ( 2009 ) ; d. hanneke _ et al .
_ , nature phys .
* 6 * , 13 ( 2010 ) ; b. p. lanyon _
et al . _ ,
science * 333 * , 57 ( 2011 ) ; t. monz _
et al . _ ,
106 * , 130506 ( 2011 ) .
k. osterloh , m. baig , l. santos , p. zoller , and m. lewenstein , phys .
lett . * 95 * , 010403 ( 2005 ) ; j. ruseckas , g. juzeliunas , p. ohberg , and m. fleischhauer , phys .
95 * , 010404 ( 2005 ) ; j. dalibard , f. gerbier , g. juzeliunas , and p. ohberg , rev .
phys . * 83 * , 1523 ( 2011 ) .
ca@xmath1 in the experiment .
( d ) experimentally used pulse sequence .
( e ) level scheme that implements another four - level system containing one rf and two optical transitions , with qubit levels @xmath160 separated by an optical frequency .
, width=302 ] - and @xmath4-rotations ) .
variations of population are plotted against the phase shift in the middle of the gate pulse .
the number of experiments per data points is 200 for experiments using three optical transitions [ ( a ) and ( b ) ] and 100 for those using one rf and two optical transitions [ ( c ) and ( d ) ] .
the error bars represent errors in projection measurements , which are derived as standard deviations in binomial distributions with numbers of samples as just given .
( a ) results for @xmath3-rotations using three optical transitions . blue hollow circles , red filled circles , magenta crosses and black asterisks represent the populations in @xmath6 , @xmath7 , @xmath9 and @xmath8 , respectively . ( b ) results for @xmath4-rotations using three optical transitions
. the population in @xmath6 is plotted .
( c ) results for @xmath3-rotations using one rf and two optical transitions .
red filled circles ( blue hollow circles ) represent the populations in @xmath7 and @xmath9 when @xmath6 ( @xmath7 ) is initially prepared .
( d ) results for @xmath4-rotations using one rf and two optical transitions .
the population in @xmath7 and @xmath9 is plotted.,width=302 ] in 17 steps , and the number of experiments per angle step is 50 .
the errors in the population measurements are estimated in the same way as in fig .
[ fig : xzrotations ] , and the fringe contrasts and shifts are obtained through weighted fits considering those errors .
the error bars in the figure represents errors in the fitted parameters .
@xmath152 is used here .
( a)fringe contrasts and ( b ) shifts for @xmath3-rotations as the peak rabi frequencies are varied . here
the ratios of the peak values are fixed .
the horizontal axis represents the peak value of @xmath161 , and the blue circles , red triangles and green squares represent the results for pulse durations of 145 , 290 and 435 @xmath90s , respectively . numerically calculated results for pulse durations of 145 , 290 and 435 @xmath90s are also plotted as blue solid curves , red dashed curves and green dotted curves , respectively .
( c ) fringe contrasts and ( d ) shifts for @xmath4-rotations as the peak rabi frequencies are varied
. the ratios of the peak values are fixed .
numerically calculated results are also plotted as a blue solid curve .
, width=302 ] | universal single - qubit operations based on purely geometric phase factors in adiabatic processes are demonstrated by utilizing a four - level system in a trapped single @xmath0ca@xmath1 ion connected by three oscillating fields .
robustness against parameter variations is studied .
the scheme demonstrated here can be employed as a building block for large - scale holonomic quantum computations , which may be useful for large qubit systems with statistical variations in system parameters . |
the structure of the pulsar magnetosphere has been an active area of research for more that 30 years , starting with the pioneering work by goldreich & julian ( 1969 ) . despite the fact that real astrophysical radio - pulsars are believed to be oblique rotating magnetic dipoles ,
much of the theoretical effort has been devoted to a significantly simpler case of an aligned rotating magnetic dipole ( i.e. , a `` non - pulsing pulsar '' ) , in which case the problem becomes stationary and axisymmetric .
the situation becomes even more tractable if one also assumes the space around the pulsar to be filled with a plasma that , on one hand , is dense enough to shorten out the longitudinal electric fields , thus providing the basis for the ideal - magnetohydrodynamic approximation , and , on the other hand , is at the same time tenuous enough for all the non - electromagnetic forces to be negligible , thus enabling one to regard the plasma as force - free .
all these simplifications have lead early on to the derivation , simultaneously by several researchers , of the main equation governing the structure of the magnetosphere , the so - called pulsar equation ( scharlemann & wagoner 1973 ; michel 1973b ; okamoto 1974 ) , which is essentially a special - relativistic generalization of the well - known force - free grad
shafranov equation .
this equation is a quasi - linear elliptic second - order partial differential equation with a regular singular surface , the so - called light cylinder . despite the fact that this , in general non - linear , equation was first derived thirty years ago ,
most of the attempts to solve it , both analytical and numerical , have been limited , until very recently , to its linear special cases and to the region inside the light cylinder ( e.g. , scharlemann & wagoner 1973 ; michel 1973a ; beskin et al . 1983 ; beskin & malyshkin 1998 ) .
these self - imposed restrictions can of course be attributed to the considerations of mathematical convenience and tractability ; however , they are not very well physically motivated and thus may be too restrictive to be relevant to real astrophysical systems . at the same time , by the late 1990s , the available computing resources and numerical techniques have become sufficiently powerful for the most general nonlinear problem , involving the regions on both sides of the light cylinder , to become tractable . in their pioneering work , contopoulos et al .
( 1999 , hereafter ckf ) have obtained the first ( and apparently unique ) numerical solution of the general problem , albeit with a rather poor spatial resolution .
their approach has later been used by ogura & kojima ( 2003 ) who have obtained essentially the same results , but with a higher numerical resolution .
both of these groups have focussed on the global structure of the solution ; consequently , they have not paid enough attention , in our opinion , to some key but subtle issues regarding the separatrix between the open and closed field - line regions and , especially , the point of intersection of this separatrix with the equatorial plane . as a result ,
the separatrix has not , we believe , been treated correctly , and , in particular , the separatrix equilibrium condition ( okamoto 1974 ; lyubarskii 1990 ) has not been satisfied close to the intersection point .
we thus suspect that their solution is probably not quite correct . in this paper
we try to clarify how the magnetic field near the separatrix and especially around the separatrix equator intersection point should be treated , as this question appears to be extremely important for setting up the _ correct _ boundary conditions for the global problem . in
[ sec - basic ] we review the basic equations describing the ideal - mhd force - free magnetosphere of an aligned rotating dipole ; we pay special attention to the so - called light cylinder regularity condition and discuss how this condition could be used to find the correct form of the poloidal - current function and , simultaneously , to obtain the solution that passes smoothly across the light cylinder .
then , in [ sec - y - point ] , we investigate some general properties of the magnetic field near the putative separatrix between the open and closed field regions ; we use the separatrix equilibrium condition to show that , if the separatrix intersects the equator at the light cylinder ( as has been assumed in the numerical simulations by ckf and by ogura & kojima 2003 ) , then the poloidal current and hence the toroidal magnetic field have to vanish on the last open field line above the separatrix .
we then use these findings in [ sec - analysis ] to construct a unique self - similar asymptotic solution of the pulsar equation in the vicinity of such an intersection point ; in particular , we find all the power - law exponents describing the field near this point and also the angle between the separatrix and the equator . at the end of this section
we show that a light surface ( a surface where the electric field becomes equal to the magnetic field and therefore where the particle drift velocity reaches the speed of light ) has to appear right outside the light cylinder ( it starts at the intersection point and extends outward at a finite angle with respect to the equator ) . we thus conclude that the only way to get a force - free solution that would be valid at least some finite distance beyond the light cylinder , is to consider the case of a separatrix equator intersection point lying some finite distance inside the light cylinder .
we investigate the magnetic field structure around such a point in [ sec - t - point ] .
we find that in this case the poloidal - current function may stay finite , but its derivative with respect to the poloidal flux has to go to zero on the last open field line ; we then find the asymptotic behavior of the magnetic field around the intersection point corresponding to this case .
finally , we present our conclusions in [ sec - conclusions ] , where we also discuss the implications of our results for the past and future numerical studies of the axisymmetric pulsar magnetospheres .
we consider the magnetosphere of an aligned rotating magnetic dipole under the assumptions of stationarity , axial symmetry with respect to the rotation axis @xmath0 , and reflection symmetry with respect to the equator @xmath1 .
we shall include the effects of special relativity but will ignore general - relativistic effects ( i.e. , work in euclidean space ) .
we shall assume that the dipole is surrounded by a very tenuous , but highly conducting plasma .
the first of these assumptions enables us to neglect all non - electromagnetic forces ( i.e. , gravity , inertial , and pressure forces ) , and thus to conclude that the structure of the magnetosphere is governed by the relativistic force - free equation : @xmath2 here , the electric charge density @xmath3 and the electric current density @xmath4 are related to the electric and magnetic field through the steady - state maxwell equations : @xmath5 the second of our assumptions concerning the magnetospheric plasma , i.e. , the assumption of infinite conductivity , enables us to use ideal magnetohydrodynamics ( mhd ) : @xmath6 these equations ( with the appropriate boundary conditions ) should be sufficient for determining the structure of the magnetosphere .
as is well known , an axisymmetric magnetic field @xmath7 can be described in terms of two functions , the poloidal magnetic flux ( per one radian in the azimuthal direction ) @xmath8 and the poloidal current @xmath9 , as @xmath10 where @xmath11 is the azimuthal ( or toroidal ) angle . in cylindrical coordinates
the magnetic field components then are : @xmath13 next , by applying faraday s law @xmath14 in a steady state , we see that @xmath15 , and hence @xmath16 ; together with the assumption of axial symmetry , this gives @xmath17 that is , the electric field is purely poloidal .
since , as it follows from equation ( [ eq - force - free ] ) or ( [ eq - ideal - mhd ] ) , the electric field must be perpendicular to the magnetic field , we can write it as @xmath18 where @xmath19 is the poloidal magnetic field , and @xmath20 is the unit vector in the @xmath11-direction . by substituting this relationship into @xmath15 , we find that @xmath21 , which has the meaning of the angular velocity of magnetic field lines , is constant along the field lines , @xmath22 .
this is the well - known ferraro isorotation law . in the problem
we are interested in , all the field lines are tied , at least at one end , to the surface of the pulsar , which is assumed to be rotating as a solid body with some uniform angular velocity @xmath23 .
we also assume that our ideal - mhd assumption is valid all the way up to the pulsar surface , i.e. , that there is no appreciable gap between this surface and the ideal - mhd region .
therefore , from now on , we shall assume that all the field lines rotate with the same angular velocity was considered , e.g. , by beskin et al .
( 1983 ) and by beskin & malyshkin ( 1998 ) . ]
@xmath24 from equation ( [ eq - e ] ) we find @xmath25 where @xmath26 is the cylindrical radius normalized to the radius of the light cylinder ( lc ) @xmath27 .
thus , we see that @xmath28 inside the lc and @xmath29 outside the lc
. now let us turn to the force - free equation ( [ eq - force - free ] ) .
the toroidal component of this equation , together with equation ( [ eq - e_phi=0 ] ) , gives : @xmath30 \times [ \nabla\psi\times\nabla\phi ] = 0 \quad \rightarrow \quad \tilde{i}=\tilde{i}(\psi ) \
, , \label{eq - i = i(psi)}\ ] ] and the poloidal component ( perpendicular to @xmath31 ) of equation ( [ eq - force - free ] ) can be written as @xmath32 - { { 1+x^2}\over x}\ , { \partial\psi\over{\partial x } } = - i i'(\psi ) \ , , \label{eq - pulsar}\ ] ] where we have introduced @xmath33 and where we also incorporated @xmath34 into @xmath35 on the right - hand side ( rhs ) by defining @xmath36 then the magnetic field components can be rewritten in terms of @xmath37 as @xmath38 equation ( [ eq - pulsar ] ) is the famous _ pulsar equation _ ( scharlemann & wagoner 1973 ; michel 1973b ; okamoto 1974 ) , also known in literature as the relativistic force - free grad shafranov equation ( e.g. , beskin 1997 ) .
it is an elliptic second - order partial differential equation ( pde ) for the flux function @xmath37 ; the left - hand side ( lhs ) of this equation is linear , whereas the rhs is , in general , nonlinear .
one very important feature of equation ( [ eq - pulsar ] ) is that it has a regular singular surface at the light cylinder @xmath39 .
both indicial ( or characteristic ) exponents for this equation are equal to zero and so a general solution of equation ( [ eq - pulsar ] ) diverges logarithmically at this surface ( e.g. , bender & orszag 1978 ) . on physical grounds ,
one imposes an additional condition that both the function @xmath37 and its 1st and 2nd derivatives remain finite near the lc . for any given function @xmath35
, it is indeed possible to find such non - divergent solutions separately on each side of the lc .
then , as can be seen from equation ( [ eq - pulsar ] ) itself , the contribution of the first term on the lhs [ the term proportional to @xmath40 vanishes as @xmath41 , and thus such non - divergent solutions on both sides of the lc have to satisfy the well - known lc regularity condition ( e.g. , scharlemann & wagoner 1973 ; okamoto 1974 ; beskin 1997 ) : @xmath42 this condition is very important and merits a few extra words of discussion . our force - free equation with a constant field - line angular velocity @xmath21 is a 2nd - order equation with one , a priori unknown , integral of motion , @xmath35 , and one singular surface , the lc . according to general theory ( beskin 1997 ) , one then needs two boundary conditions to set up the problem properly .
we believe that the correct approach here is to set the boundary conditions for the function @xmath8 at both the inner boundary ( inside the lc : on the surface of the star , the symmetry axis , the separatrix between the open and closed field - line regions , and maybe the equator ) and at whatever outer boundary ( outside the lc : the equator and infinity or the light surface ) . at the same time the poloidal current function @xmath35 should not be prescribed explicitly at any boundary of the domain ; instead , it is to be determined from the matching condition at the singular surface , i.e. , the lc ( see below ) .
thus , our position differs from that of beskin ( see beskin et al . 1983 , and beskin 1997 ) who proposed prescribing @xmath35 explicitly at the surface of the pulsar . indeed , if one considers the region inside the lc only , then , upon prescribing both @xmath8 and @xmath35 on the inner boundary ( the pulsar surface , etc . ) , one should be able to obtain a solution that is regular at the lc by using the regularity condition ( [ eq - regularity - lc ] ) at the other boundary of this inner region , i.e. , at the lc .
thus , the solution in this inner region is then completely determined ; as such , it is totally independent of what happens outside the lc .
this does not seem physical ; indeed , one should then be able to take this solution and continue it smoothly across the lc , thus prescribing both @xmath43 and @xmath44 at the outer side of the lc .
but then the problem of finding a solution in the region outside the lc becomes over - determined , as we now have two conditions at @xmath39 , in addition to any conditions at the outer boundary .
for example , in the very important particular case of the domain under consideration extending all the way to radial infinity , one can not actually prescribe any specific boundary conditions at infinity because equation ( [ eq - pulsar ] ) has a regular singularity there .
instead , however , one imposes a regularity condition at infinity . in spherical polar coordinates ) .
we hope that this does not cause any confusion . ] ( @xmath45 ) this regularity condition can be written as @xmath46 this regularity condition has a very simple physical meaning : it is the condition of the force balance in the @xmath47-direction between the toroidal magnetic field @xmath48 and the poloidal electric field @xmath49 ( which both become much larger than the poloidal magnetic field @xmath50 as one approaches infinity ) .
thus , in this case the region outside the lc does not have any boundary conditions at all ( apart from the condition @xmath51 at the equator ) , but has two regularity conditions , one at the lc , and the other at infinity .
these two regularity conditions are already sufficient to uniquely determine the solution @xmath52 for a given @xmath35 .
therefore , if , in addition to these two regularity conditions , one also tries to impose the function @xmath43 along the lc , the system becomes over - determined .
thus we come to the conclusion that prescribing @xmath35 at the inner boundary is not appropriate .
one could , however , try to use the outer boundary condition ( or the regularity condition at infinity ) as the condition that fixes @xmath35 . to do this
, one first sets the inner boundary conditions , picks an initial guess for @xmath35 , and uses ( [ eq - regularity - lc ] ) to obtain a regular solution inside the lc ; one then continues this solution smoothly across the lc and solves equation ( [ eq - pulsar ] ) in the outer region as an initial - value problem ( with both @xmath8 and @xmath53 specified at @xmath39 ) ; as a result , one gets a mismatch at the outer boundary between the obtained solution and the desired outer boundary condition [ or , in the case of an infinite domain , one would presumably fail to get a convergent solution satisfying the regularity condition ( [ eq - regularity - infinity ] ) ] .
then one iterates with respect to @xmath35 until this outer - boundary mismatch is zero ( or until a solution regular at infinity is achieved ) . in reality , however , such an approach may not be practical as it involves solving an initial - value problem for an elliptic equation , which is not a well - posed problem . instead , we advocate the approach adopted by ckf . in this approach
one considers the regions inside the lc ( @xmath54 ) and outside the lc ( @xmath55 ) separately .
first , one makes an initial guess for @xmath35 and prescribes the corresponding boundary conditions for @xmath8 at the inner boundary of the the region inside the lc and at the outer boundary of the the region outside the lc [ or , if this outer boundary is at the radial infinity , imposes the regularity condition ( [ eq - regularity - infinity ] ) there ] .
then , one uses , again separately in each region , the regularity condition ( [ eq - regularity - lc ] ) in lieu of a boundary condition at the lc [ in practice , one can think of ( [ eq - regularity - lc ] ) as of a mixed - type dirichlet
von neumann boundary condition relating the value of @xmath8 at the surface @xmath39 to the value of its derivative normal to this surface ] .
thus one obtains two solutions , one inside and the other outside the lc , that correspond to the same function @xmath35 and are both regular at @xmath39 .
however , these solutions also depend on their respective boundary or regularity conditions set at the inner / outer boundaries of the domain .
hence , in general , although they are both regular at the lc , these solutions are not going to coincide at @xmath39 .
the mismatch @xmath56 depends on both the chosen function @xmath35 and on the inner and outer boundary conditions .
one then iterates with respect to @xmath35 in order to minimize this mismatch .
we believe that for a given set of boundary conditions there should be only one choice of @xmath35 for which the mismatch @xmath57 vanishes and hence the function @xmath8 becomes continuous along the entire lc , i.e. , @xmath58 .
once this special function @xmath35 is found , it then follows from ( [ eq - regularity - lc ] ) , which is satisfied separately on both sides of the lc , that @xmath53 is also continuous , and so the entire solution passes smoothly across the lc .
thus , one can say that the function @xmath35 is determined by the regularity condition ( [ eq - regularity - lc ] ) applied separately on both sides of the lc , plus the condition of matching of the two solutions .
another , equivalent way to put it , is to say that @xmath35 is determined by the ( non - trivial ! ) requirement that the derivative @xmath59 actually _ exists _ at the lc [ whereby @xmath35 is expressed in terms of this derivative via equation ( [ eq - regularity - lc ] ) ] .
this approach has been implemented successfully in the pioneering numerical work by ckf and then subsequently by ogura & kojima ( 2003 ) .
both these groups have used a relaxation procedure to arrive at @xmath35 that corresponded to a unique solution that was both continuous and smooth at the lc .
the main focus of this paper is the behavior of the magnetic field in the vicinity of the point of intersection of the separatrix @xmath60 between the region of closed field lines ( region i ) and the region of open field lines ( region ii ) with the equator @xmath61 ( see fig . [ fig - geometry ] ) .
we shall generally call this point the _ separatrix intersection point_. our interest in this non - trivial problem is fueled by the belief that understanding the key features of this behavior is absolutely crucial to devising the proper boundary conditions and providing verification benchmarks for any future global numerical investigations of a force - free pulsar magnetosphere . at the same time , the treatment of the magnetic field around this very special point is expected to require a certain degree of subtlety and delicacy , as noted , for example , by beskin et al .
( 1983 ) and by ogura & kojima ( 2003 ) .
among the questions that need to be answered are : what is the radial position @xmath62 of the intersection point ( namely , whether this point lies at the lc , @xmath63 , of inside the lc , @xmath64 ) ? and what is the angle @xmath65 at which the separatrix approaches the equator at this point , the three a priori possibilities being @xmath66 ( in which case we ll call this point the cusp - point ) , @xmath67 ( the y - point ) , and @xmath68 ( the t - point ) ? in our analysis , we shall assume that equations ( [ eq - force - free ] ) and ( [ eq - ideal - mhd ] ) that describe our system are valid almost everywhere in the vicinity of the separatrix intersection point , i.e. , everywhere with perhaps the exception of measure - zero regions . thus , we shall allow for the presence in our system of current sheets of infinitesimal thickness , across which the magnetic field can experience a finite jump .
we limit our consideration to the situation where such current sheets can be present only along the separatrix between regions i and ii for @xmath69 and along the equatorial ( @xmath61 ) separatrix between the upper and lower open - field regions for @xmath70 .
thus , we have one current sheet that lies on the equator at @xmath70 and at @xmath71 splits into two symmetrical current sheets ( one in each hemisphere ) lying along the separatrix @xmath72 ( see , e.g. , okamoto 1974 ) .
at the same time we assume that our equations apply everywhere else at least in some region around the intersection point , including the portion of this region that lies outside the lc ( in the case @xmath63 ) .
let us first assume that the separatrix between closed and open field - line regions reaches the lc ( i.e. , @xmath63 ) , and let us consider the separatrix equilibrium condition ( e.g. , okamoto 1974 ; lyubarskii 1990 ) .
indeed , whether or not the separatrix contains a current sheet ( which we assume to be infinitesimally thin ) , it must satisfy the condition of force balance , which is obtained by integrating equation ( [ eq - force - free ] ) across the separatrix : ) may not be valid inside a current sheet as other forces may also be important in the pressure balance there . however , gravity and the plasma inertia are small because the separatrix current layer is assumed to be very thin ; and the plasma pressure gradient , while not necessarily small inside the current layer , gives , when integrated across the layer , just the difference between the values of the pressure on both sides outside the separatrix , both of which are small by assumption . ] @xmath73 by using equation ( [ eq-|e| ] ) and also the fact that @xmath74 in region i , we can rewrite this condition as @xmath75\ , ( 1-x^2 ) = ( b_\phi^{ii})^2 = { 1\over{r_{lc}^4}}\ , { i^2(\psi_s)\over{x^2 } } \ , , \label{eq - sepx - equil-2}\ ] ] where @xmath76 marks the distance from the intersection point ( @xmath61 ) along the separatrix . now ,
as the separatrix approaches the lc ( @xmath41 , @xmath77 ) , we get @xmath78 = { { i_s^2}\over{r_{lc}^4 } } \ , , \label{eq - sepx - equil - endpt}\ ] ] where we have defined @xmath79 as the value of the poloidal current of the last open field line @xmath80 ( above the separatrix current sheet ) .
since @xmath81 is constant along this last open field line , and hence is independent of @xmath76 , we immediately see that in order to satisfy equation ( [ eq - sepx - equil - endpt ] ) in the limit @xmath41 , we must have either : + _ ( i ) _
@xmath82 , i.e. , all poloidal current that flows out of the pulsar must return back to the pulsar along open magnetic field lines , with no finite poloidal current flowing along the separatrix ; of the compatibility condition by beskin et al .
( 1983 ) and beskin & malyshkin ( 1998 ) . ] simultaneously , from equation ( [ eq - sepx - equil - endpt ] ) it then follows that @xmath83 , and hence no finite toroidal current flows along the separatrix either ( we make a very natural assumption that @xmath84 is in the same direction as @xmath85 ) ; thus the separatrix in this case is actually not a current sheet ; + or , + _ ( ii ) _ a finite @xmath86 but a divergent poloidal magnetic field in the closed - field region , @xmath87 , as @xmath77 , @xmath88 . note that this situation can not be dismissed automatically ; indeed , close to the lc the magnitude of the electric field becomes very close to that of the poloidal magnetic field , with the infinitely strong electric forces balancing out the infinitely strong magnetic ones to a high degree .
then , the separatrix current sheet would have to carry back to the star a finite poloidal current @xmath81 ( thus closing the poloidal current circuit ) , and also a toroidal corotation current with the surface current density @xmath89 divergent near the lc : @xmath90 ; there would also have to be a divergent surface charge density on the separatrix : @xmath91 .
note that these conclusions , derived from the separatrix equilibrium condition near the lc , are valid regardless of the angle @xmath65 between the separatrix and the equator at the intersection point .
in our analysis in the next section we shall assume that there are no infinitely strong fields anywhere in the system , and thus shall dismiss situation _
( ii ) _ ( @xmath63 , @xmath92 , finite @xmath86 ) as unphysical . we shall thus concentrate our attention on situation _ ( i ) _ ( @xmath82 ) .
we shall try to analyze the structure of the pulsar equation ( [ eq - pulsar ] ) and obtain the asymptotic solution of this equation in the vicinity of the intersection point ( @xmath63 , @xmath93 ) .
in this section we perform an asymptotic analysis of the pulsar equation ( [ eq - pulsar ] ) in the vicinity of the separatrix intersection point under the assumption that it is located at the lc , i.e. , @xmath63 , @xmath93 .
our approach is actually very similar to the analysis of the magnetic field near the endpoint of a non - relativistic reconnecting current layer , performed previously by uzdensky & kulsrud ( 1997 ) .
in addition to assuming @xmath63 , we shall , in this section , make use of the following two assumptions : + 1 ) the intersection point is a finite - angle y - point , i.e. , the separatrix approaches the equator at an angle @xmath65 , @xmath67 , where we measure angles from the radial vector lying on the equator and directed _ toward _ the star , as shown in figure [ fig - geometry ] ; + 2 ) the poloidal current is fully closed in the open - field region , i.e. , @xmath94 ; this assumption is motivated by the arguments presented in the previous section .
furthermore , we also assume symmetry with respect to the equator and therefore consider only the upper half - space , @xmath95 . when considering the @xmath96-dependent coefficients in equation ( [ eq - pulsar ] ) in the vicinity of the y - point , we can , to lowest order in @xmath97 , replace @xmath96 by 1 everywhere except where it appears in a combination like @xmath98 ; thus we can rewrite this equation as @xmath99 + 2 \ , { { \partial\psi}\over{\partial\xi } } = - ii'(\psi ) \
, , \label{eq - pulsar - endpt}\ ] ] where we introduced a new coordinate , @xmath100 . instead of using coordinates @xmath101 , it is actually more convenient to use polar coordinates @xmath102 defined by @xmath103 when making the transition to these coordinates , we shall make use of the following expressions : @xmath104 as well as the usual expression for two - dimensional ( 2-d ) laplacian in polar coordinates : @xmath105 in these coordinates the poloidal magnetic field components are written as follows ( we do not need these expressions now but will need them later ) : @xmath106 finally , by using expressions ( [ eq - coord - derivatives - xi])([eq-2d - laplacian ] ) , we can rewrite the pulsar equation ( [ eq - pulsar - endpt ] ) in our polar coordinates ( [ eq - polar_coords-1])([eq - polar_coords-2 ] ) as @xmath107 now we need to solve this equation separately in region i ( region of closed field lines , @xmath108 ) and in region ii ( region of open field lines , @xmath109 ) and match the two solutions together at the separatrix @xmath110 . since in this paper we are dealing only with the local structure of the magnetic field near the y - point , we can , without any loss of generality , adopt the convention of counting the poloidal magnetic flux @xmath8 from the separatrix @xmath110 ( instead of the usual convention of counting @xmath8 from the rotation axis ) .
thus , we shall set @xmath111 and , correspondingly , @xmath112 in region i and @xmath113 in region ii . very close to the y - point , i.e. at distances much smaller than the light cylinder radius ( @xmath114 )
, the system lacks any natural length scale .
hence , we can expect the radial dependence of @xmath8 to be a power law , which enables us to make the following _ self - similar ansatz _ : @xmath115 [ we put a minus sign in equation ( [ eq - psi_ii ] ) in order to have @xmath116 . ]
then , the magnetic field components are given by : + in region i : @xmath117 and in region ii : @xmath118 notice that the condition that magnetic field does not diverge near @xmath119 imposes the restriction @xmath120 whereas for the case where the poloidal magnetic flux function in region i diverges as @xmath121 [ i.e. , case _
( ii ) _ in [ sec - y - point ] ] , we would have @xmath122 . as for the poloidal current function @xmath35 that appears on the rhs of equation ( [ eq - pulsar - polar ] )
, it takes very different forms in regions i and ii . in region
i there should be no toroidal field , so @xmath123 in the open - field region ii , @xmath35 is not zero but does approach zero in the limit @xmath124 .
since there is no natural a priori magnetic field scale in the vicinity of the y - point , we shall again employ a self - similar ansatz , i.e. , assume that @xmath125 scales as a power of @xmath126 near @xmath127 : @xmath128 so that the right - hand side of equation ( [ eq - pulsar - polar ] ) becomes @xmath129 where @xmath130 upon substituting relationships ( [ eq - psi_i ] ) and ( [ eq - i - regioni ] ) ( for the closed - field region ) and ( [ eq - psi_ii ] ) and ( [ eq - rhs - regionii ] ) ( for the open - field region ) into the main equation ( [ eq - pulsar - polar ] ) , we obtain two ordinary differential equations ( odes ) for the functions @xmath131 and @xmath132 , along with a relationship between the power - law indices @xmath133 and @xmath134 . in particular , in region i we get a homogeneous linear 2nd - order ode for @xmath131 : @xmath135 which is to be supplemented by two boundary conditions : @xmath136 similarly , in region ii we get a non - linear 2nd - order ode for @xmath132 : @xmath137 together with the relationship @xmath138 that follows from the requirement that the lhs of equation ( [ eq - pulsar - polar ] ) scale as the same power of @xmath139 as the rhs , i.e. , @xmath140 .
[ at this point , however , we have to remark that one can not a priori exclude the possibility @xmath141 ; in such a case the poloidal current would go to zero near the separatrix so rapidly that its contribution to the pulsar equation would become completely negligible .
however , as we shall discuss later in this section , it is possible to show that no continuous solutions exist in this case . ]
the boundary conditions for equation ( [ eq - pulsar - ii ] ) are @xmath142 notice that the constant @xmath143 that appears on the rhs of equation ( [ eq - pulsar - ii ] ) is in fact unimportant ; it just sets the scale of variation of the function @xmath132 and , hence , the overall scale of the magnetic field strength .
thus , we can rescale @xmath143 away by incorporating it into the solution ; we do this by defining a new variable @xmath144 then , equation ( [ eq - pulsar - ii ] ) becomes @xmath145 with the same boundary conditions @xmath146 thus , we see that in our problem the magnetic field structure near the y - point is completely characterized by four finite dimensionless parameters : the separatrix angle @xmath65 and the three power - law indices @xmath147 , @xmath133 , and @xmath134 .
our goal is to obtain a unique solution of our problem , that is to determine the values of these parameters and , simultaneously , determine the functions @xmath131 and @xmath132 .
in fact , equation ( [ eq - beta - alpha2 ] ) already gives us one relationship between the power exponents , but we still need three more relationships to fix all four parameters . therefore , we shall now discuss the conditions that will help us obtain these three additional relationships .
the second condition [ the first being equation ( [ eq - beta - alpha2 ] ) ] that links our dimensionless parameters is rather obvious ; it is the condition of force - balance across the separatrix @xmath110 and it will give us a relationship between @xmath147 and @xmath133 .
we have already discussed this condition in [ sec - y - point ] [ see eqs .
( [ eq - sepx - equil-1])([eq - sepx - equil - endpt ] ) ] ; with @xmath148 , @xmath149 , and @xmath82 , this condition can be written simply as @xmath150^i = [ b_{\rm pol}^2 ( r,\theta_0)]^{ii } \ , .
\label{eq - sepx - equil-3}\ ] ] since in our @xmath102-coordinates the separatrix is a radial line @xmath110 , and since we expect the poloidal magnetic field on the two sides of the separatrix to be in the same direction , condition ( [ eq - sepx - equil-3 ] ) simply means that @xmath151 .
then , using equations ( [ eq - br - i])([eq - btheta - ii ] ) , we immediately obtain a second relationship between the power exponents : @xmath152 and also a relationship between the normalizations of the functions @xmath131 and @xmath132 , cast in terms of their derivatives at @xmath110 : @xmath153 the remaining two relationships between the dimensionless parameters come from the properties of equations ( [ eq - pulsar - i ] ) and ( [ eq - pulsar - ii - g ] ) themselves and can not , unfortunately , be written out as explicit algebraic equations containing these parameters .
this fact , however , can prevent us neither from describing and discussing the physical and mathematical conditions on which these two additional relationships are based , nor from using these conditions to obtain the actual unique values of the parameters .
first , we shall show that there is a unique one - to - one relationship between the separatrix angle @xmath65 and the power exponent @xmath154 .
this relationship comes from analyzing equation ( [ eq - pulsar - i ] ) for the closed - field region i. for any given @xmath65 this homogeneous linear equation with boundary conditions ( [ eq - bc - i ] ) is an eigen - value problem for the coefficient @xmath155 .
it is in fact very easy to see why this must be the case .
indeed , the boundary conditions ( [ eq - bc - i ] ) do not give us any scale for @xmath156 ; if some function @xmath131 is a solution of the problem ( [ eq - pulsar - i])([eq - bc - i ] ) , then @xmath157 with an arbitrary multiplier @xmath158 will also be a valid solution .
thus , the normalization of @xmath131 is arbitrary : if , for given @xmath65 and @xmath154 , a non - trivial solution of the problem ( [ eq - pulsar - i ] ) ([eq - bc - i ] ) exists , then there will be infinitely many such solutions ; in particular , there will be a solution with @xmath159 .
thus , we can impose an additional boundary condition @xmath159 .
then , however , we have three boundary conditions for a 2nd - order differential equation , which means that for arbitrary @xmath65 and @xmath154 the system is over - determined .
hence , such a solution will exist not for all values of @xmath65 and @xmath154 ; the condition that it actually does exist gives us a certain relationship between @xmath65 and @xmath154 : for a given @xmath65 we would get an infinite discrete spectrum of the allowed values of @xmath154 . of these , however , we are interested only in the lowest one because we want a solution without direction reversals of the magnetic field , i.e. , with @xmath160 everywhere . from the practical point of view , the easiest way to determine
the relationship between @xmath154 and @xmath65 is the following .
we scan over the values of @xmath154 ; for each given @xmath154 we use our freedom of normalization of @xmath131 to set @xmath159 .
thus we now have an initial - value problem with two conditions at @xmath161 : @xmath159 and @xmath162 . with these two conditions and with the value of @xmath163 given
, we can integrate equation ( [ eq - pulsar - i ] ) forward in @xmath47 ( we do it numerically using a 2nd - order runge kutta scheme ) until the point where @xmath131 becomes zero .
this point is then declared the correct value of the separatrix angle @xmath65 for the given @xmath154 .
the function @xmath164 we have obtained as a result of this procedure is shown in figure [ fig - theta0 ] , whereas figure [ fig - f ] shows the behavior of the function @xmath131 for several selected values of @xmath154 .
we see that @xmath164 is a monotonically decreasing function with @xmath165 rad [ instead of @xmath166 that one would get if the y - point were at a finite distance inside the lc ! ] , and @xmath167 .
, corresponding to the case of zero poloidal current ( see also beskin et al . 1983 ) ; this is because in his analysis he had in fact assumed that @xmath168 and then derived the corresponding value of @xmath65 . ]
the asymptotic behavior of @xmath164 in the limit @xmath169 is @xmath170 , which is in fact very easy to obtain analytically . , we expect @xmath171 , and then we can neglect the @xmath172 term in equation ( [ eq - pulsar - i ] ) . as a result , we get a simple harmonic equation @xmath173 , with @xmath162 , @xmath174 ; the solution of this equation , positive everywhere in the domain @xmath175 , is @xmath176 ; it corresponds to @xmath177 .
] actually , the exact analytical fit for our numerical curve @xmath164 appears to be @xmath178 with @xmath179 . with the dependence @xmath164 thus determined , and with @xmath134 and @xmath133 related to @xmath154 via equations ( [ eq - beta - alpha2 ] ) and ( [ eq - alpha1=alpha2 ] ) , respectively
, we now have everything expressed in terms of @xmath154 .
the fourth ( and final ! ) condition that will help us determine @xmath154 and hence @xmath65 and the rest of the power exponents is the light cylinder regularity condition .
this condition states that at the lc @xmath180 [ which is a regular singular point for equation ( [ eq - pulsar - ii - g ] ) ] the function @xmath181 should be regular , namely , it should be a continuously differentiable function , with finite 1st and 2nd derivatives .
this condition can be written as @xmath182 and it is just a particular manifestation of the general lc regularity condition ( [ eq - regularity - lc ] ) .
we use this condition to fix the unique value of @xmath154 ( see the general discussion of this issue at the end of [ sec - basic ] ) . here
is how we do it in practice .
first , we divide up region ii into two sub - regions : region ii@xmath183 ( inside the lc : @xmath184 ) and region ii@xmath185 ( outside the lc : @xmath186 ) .
then , we scan over @xmath154 ; for each value of @xmath154 , we first determine the corresponding value of @xmath65 using the procedure outlined above .
once @xmath65 for a given @xmath154 is found , we solve equation ( [ eq - pulsar - ii - g ] ) separately in regions ii@xmath183 and ii@xmath185 ( again , using a numerical shooting method in conjunction with the 2nd - order runge - kutta integration scheme ) . in each of these regions we use regularity condition ( [ eq - lc - regularity - g ] ) at @xmath180 and one of the boundary conditions ( [ eq - bc - ii - g ] ) at the other end of the region , i.e. , @xmath187 in region ii@xmath183 and @xmath188 in region ii@xmath185 .
note that the use of the regularity condition ( [ eq - lc - regularity - g ] ) guarantees that the solutions in _ each _ of the two regions are regular .
however , in general ( i.e. , for an arbitrarily chosen @xmath154 ) , the two solutions obtained in this manner do not match each other at the lc , the mismatch @xmath189 [ and hence the mismatch in @xmath190 related to @xmath191 via eq .
( [ eq - lc - regularity - g ] ) ] being @xmath154-dependent .
we find that there is only one , special value of @xmath154 for which @xmath191 vanishes and the solution continues smoothly across the lc as one passes from region ii@xmath183 into region ii@xmath185 .
this special value , which we call @xmath192 , is declared the correct value of @xmath154 .
numerically , we find @xmath193 and , correspondingly , @xmath194 this is the way in which the unique correct solution of the problem is obtained .
the functions @xmath131 and @xmath195 corresponding to this value of @xmath154 are plotted in figures [ fig - f_0 ] and [ fig - g_0 ] , respectively .
now let us make a little digression and step back to discuss the possibility of finding a solution in the case @xmath196 . recall that in this case the dominant balance of the lowest - order ( in @xmath139 ) terms in the pulsar equation does not include the contribution @xmath197 due to the poloidal current .
at the same time , note that relationships ( [ eq - alpha1=alpha2 ] ) and ( [ eq - sepx - f=-g ] ) , derived from the separatrix force - balance condition , still hold .
hence , the solution in region ii would have to satisfy a linear equation that is identical to equation ( [ eq - pulsar - i ] ) for @xmath131 on the other side of the separatrix .
this fact , combined with the condition ( [ eq - sepx - f=-g ] ) , implies that the solution just has to pass smoothly across @xmath110 , and thus this point [ which is just an ordinary point for equation ( [ eq - pulsar - i ] ) ] does not have any special significance ; it is just the point where the solution changes sign .
thus , the general solution in this case takes the form @xmath198 , where @xmath131 satisfies the homogeneous linear equation ( [ eq - pulsar - i ] ) in the entire domain @xmath199 $ ] with the homogeneous boundary conditions @xmath162 and @xmath200 .
in addition , the solution has to pass smoothly across the regular singular point @xmath180 ( the light cylinder ) ; the regularity condition at this point is simply @xmath201 .
such a solution would start positive with a zero derivative at @xmath161 , then decrease and change sign at some angle @xmath202 , reach a minimum at the lc [ since @xmath201 ] , and finally increase and go back to zero at the equator @xmath204 .
however , the extra regularity condition at @xmath180 makes the system over - determined . indeed ,
counting two boundary conditions at @xmath205 , an arbitrary normalization of the solution [ e.g. , @xmath159 ] , and the lc regularity condition , we now have four conditions for our second - order equation with one free parameter @xmath154 .
thus , unless we are somehow extremely lucky , the solution of this problem does not exist .
one can show that this is indeed the case ; it turns out that a solution of equation ( [ eq - pulsar - i ] ) in the region inside the lc ( @xmath207 $ ] ) satisfying both the boundary condition @xmath162 and the lc regularity condition @xmath201 exists only for even values of @xmath208 , whereas a solution in region ( @xmath209 $ ] ) satisfying both the boundary condition @xmath200 and the lc regularity condition exists only for odd values of @xmath210 .
this demonstrates that , in the case @xmath196 , it is impossible to obtain a solution that is continuous at the lc .
this leaves us with the case @xmath211 ( for which we have just found a unique suitable solution ) as the only possibility .
thus we have managed to obtain a _ unique _ asymptotic solution of the ideal - mhd , force - free system ( [ eq - force - free])([eq - ideal - mhd ] ) in the vicinity of the separatrix intersection point ( @xmath63 , @xmath93 ) .
this solution , characterized by dimensionless parameters ( [ eq - alpha_0 ] ) and ( [ eq - beta&theta_0 ] ) , satisfies the separatrix equilibrium condition and is continuous and smooth at the lc .
let us now discuss the physical relevance of our solution to the pulsar problem .
note that the obtained unique solution of our problem has a finite value of the derivative @xmath212 at the equator @xmath213 , namely , @xmath214 , and thus has a non - zero radial magnetic field just above ( and just below ) the equatorial current sheet .
this actually spells bad news for the applicability of our solution .
indeed , one more condition that needs to be satisfied for the solution to be physically relevant and that we have not discussed so far is the condition @xmath215 with @xmath216 related to @xmath217 via equation ( [ eq-|e| ] ) , this condition can be expressed as @xmath218 we see that this condition is satisfied trivially everywhere inside the lc , but is not automatically satisfied outside the lc . in particular , in our solution , and in fact in any solution characterized by @xmath82 , the toroidal field on the last open field line @xmath72 is zero , whereas the poloidal magnetic ( and hence electric ) field is not zero , as our explicit solution shows .
this means that the inequality ( [ eq - b > e-2 ] ) is violated immediately outside the lc as one moves along the equator .
more generally , this inequality is violated beyond the so - called light surface ( defined as the surface where @xmath219 ) that emanates from the y - point .
it is in fact not difficult to find the location of this light surface in our solution . indeed , according to equations ( [ eq - br - ii])([eq - btheta - ii ] ) , we have @xmath220\ , ,
\label{eq - bpol^2}\ ] ] and , according to equations ( [ eq - b_phi-2 ] ) , ( [ eq - i - regionii ] ) , and ( [ eq - beta - alpha2 ] ) , @xmath221 using the result ( [ eq - bpol^2 ] ) , we can rewrite this as @xmath222 upon comparing this result with @xmath223 we conclude that the light surface @xmath224 emanates from the y - point at a finite angle @xmath225 which is determined from the algebraic equation [ provided that the solution @xmath181 is known ] : @xmath226\ , \cos\theta_{ls } \ , .
\label{eq - theta_ls-1}\ ] ] for our solution described by the parameters ( [ eq - alpha_0])([eq - beta&theta_0 ] ) , we find @xmath227 which corresponds to the angle @xmath228
. beyond the ls ( @xmath229 ) the force - free equation ( [ eq - force - free ] ) is not applicable .
the solution we have obtained in this section is valid in the region @xmath230 , but , because the process of obtaining this solution also involved the region @xmath229 [ e.g. , via the boundary condition @xmath231 , our solution may not be the correct solution of the overall problem .
in fact , we can make an even more general statement : any relevant to our problem finite - field magnetic configuration with the separatrix intersecting the equator at the lc can not be almost everywhere ( that is , again , everywhere except at a finite number of infinitesimally thin current sheets ) ideal and force - free beyond the lc !
indeed , as we saw previously , if one insists on having the separatrix intersection point at the lc , then one has to contend either with having @xmath232 divergent near @xmath39 or with having @xmath233 .
the latter case , however , is characterized by zero toroidal magnetic field everywhere along the last open field line @xmath72 .
that the poloidal ( radial ) magnetic field does not vanish identically near the equatorial current sheet , it then follows that @xmath234 , and so the force - free equation ( [ eq - force - free ] ) becomes inapplicable . and
, as our particular solution shows , the condition @xmath235 is violated not only on some singular surfaces of measure zero , such as the equatorial current sheet ( that would be acceptable to us at this stage , since we do expect our simple force - free , ideal - mhd assumptions to break down there anyway ) , but in a volume of non - zero extent in both poloidal directions , e.g. , in the region ( @xmath236,@xmath237 ) .
thus we have to conclude that if one is to keep the hope of finding a force - free solution that would be valid all the way up to infinity ( in the limit of vanishing plasma density ) , or at least to some finite distance beyond the lc , then one has to make concessions regarding the anticipated location of the separatrix / equator intersection point .
in other words , the requirement that @xmath238 outside the lc means that @xmath81 must be non - zero and this , in turn , forces one to conclude that the intersection point must be located at some finite distance inside the lc , i.e. , @xmath64 .
this scenario has actually been the subject of several studies in the past 20 years ( e.g. , beskin et al .
1983 ; lyubarskii 1990 ; beskin & malyshkin 1998 ) . according to equation ( [ eq - sepx - equil-2 ] ) , the corresponding finite-@xmath81 configuration has to be characterized by @xmath239 staying finite in the limit @xmath240 , and , therefore , the separatrix should approach the equator at a right angle , @xmath68 , corresponding to a t - point ( see lyubarskii 1990 ) .
let us now ask what other characteristic features should such a solution possess ? here we present a few simple facts that can be gleaned immediately by inspecting the force - free equation ( [ eq - pulsar ] ) and its regularity condition ( [ eq - regularity - lc ] ) .
first , since beyond @xmath71 the last open field line , @xmath72 , is assumed to lie along the equator @xmath61 , the derivative @xmath241 should be zero everywhere along this line ( for @xmath70 ) .
in particular , we should have @xmath242 , but then it follows from equation ( [ eq - regularity - lc ] ) that either @xmath243 or @xmath244 has to be zero . since here we are considering the case @xmath86
, we conclude that @xmath35 has to approach @xmath80 with a zero slope , @xmath245 .
thus , we see that @xmath246 can not be a simple linear function , unless it is constant everywhere .
next , by substituting the result @xmath247 , together with @xmath248 , into equation ( [ eq - pulsar ] ) , we find that @xmath249 for all @xmath70 .
this means that the radial magnetic field also approaches the equator @xmath61 with a zero slope , @xmath250 .
now let us see what one can deduce by investigating the regularity condition ( [ eq - regularity - lc ] ) just above the equatorial current sheet .
according to what we have just established , above the equator the function @xmath35 can be expanded as @xmath251 where @xmath252 in order to satisfy the condition @xmath253 , as @xmath124 ( here we again set @xmath111 to simplify notation )
. then we find , to lowest order in @xmath126 , @xmath254 now , the radial magnetic field just above the equator is , in general , not zero , so we can expand @xmath255 to lowest order in @xmath256 as @xmath257 where @xmath258 is in fact just the rescaled electric field above the equator , as can be seen from equations ( [ eq - e ] ) and ( [ eq - b_r-2 ] ) : @xmath259 upon substituting equations ( [ eq - tpoint - rhs ] ) and ( [ eq - tpoint - psi - z>0 ] ) into the regularity condition ( [ eq - regularity - lc ] ) , we get @xmath260^{\beta-1 } \ , , \label{eq - tpoint - regularity - lc - z>0}\ ] ] and , therefore , @xmath261 unless @xmath262 [ and @xmath263 if @xmath264 . we can now use this result to analyze the magnetic field in the vicinity of the point @xmath265 . equation ( [ eq - pulsar ] ) in the vicinity of this point can be written as @xmath266 + { { 1+x_0 ^ 2}\over
x_0}\ , { \partial\psi\over{\partial\zeta } } = -\ , ii'(\psi ) \ , , \label{eq - tpoint - pulsar}\ ] ] where @xmath267 .
now , just as we did in [ sec - analysis ] , we can make a transition to the polar coordinates : @xmath268 , @xmath269 . in these polar coordinates equation ( [ eq - tpoint - pulsar ] ) becomes @xmath270 + { { 1+x_0 ^ 2}\over x_0}\ , \biggl[{\partial\psi\over{\partial r}}\cos\theta- { \partial\psi\over{r\partial\theta}}\sin\theta \biggr ] = -\ , ii'(\psi ) \ , , \label{eq - tpoint - pulsar - polar}\ ] ] again , similarly to what we did in
[ sec - analysis ] [ see eq .
( [ eq - psi_ii ] ) ] , let us assume that the magnetic flux function in region ii ( the region of open field lines , @xmath271 ) is a power law of @xmath139 near the point ( @xmath272 ) : @xmath273 then , using equations ( [ eq - tpoint - rhs ] ) and ( [ eq - tpoint - beta=2 ] ) , we can express the rhs of equation ( [ eq - tpoint - pulsar - polar ] ) as @xmath274 whereas the lhs of this equation can be expressed as @xmath275 \nonumber \\ & - & \ , { { 1+x_0 ^ 2}\over x_0}\ , r^{\alpha_2 - 1}\ , \biggl [ \alpha_2 g(\theta)\cos\theta - g'(\theta ) \sin\theta \biggr ] \ , .
\label{eq - tpoint - lhs - polar}\end{aligned}\ ] ] in this expression , terms of the lowest order in @xmath139 are those proportional to @xmath276 ; at the same time we see from equation ( [ eq - tpoint - rhs - polar ] ) that the rhs of equation ( [ eq - tpoint - pulsar - polar ] ) scales with @xmath139 as @xmath277 .
therefore , we see that in this case of the separatrix intersecting the equator inside the lc , @xmath64 , the contribution of the rhs is completely negligible ( as @xmath278 ) .
thus , the dominant balance in equation ( [ eq - tpoint - pulsar - polar ] ) dictates that the terms of order @xmath276 should balance each other and hence we get a very simple homogeneous linear ode for the function @xmath132 : @xmath279 the boundary conditions for this equation are @xmath280 and so the solution with the lowest @xmath133 ( corresponding to the simplest magnetic field topology in this region ) obviously is @xmath281 and , correspondingly , @xmath282 thus , @xmath283 . must be equal to zero according to the lc regularity condition , whereas lyubarskii assumed that @xmath284 and hence had a finite , non - vanishing @xmath285 . ] as for region i ( the region of closed field lines ) , the magnetic field there is , to lowest order in @xmath139 , purely vertical and finite at @xmath240 , so we can write @xmath286 where @xmath287 is related to @xmath81 via the separatrix equilibrium condition ( [ eq - sepx - equil-2 ] ) : @xmath288 we express hope that these simple findings will be helpful in setting up or checking the correctness of future numerical attempts to solve this nontrivial problem .
in this paper we have considered the axisymmetric force - free magnetosphere of an aligned rotating magnetic dipole , under the additional assumptions of ideal mhd and of the uniformity of the field - line angular velocity [ @xmath289 .
this fundamental model problem is of great importance to any attempts to understand the workings of radio - pulsars .
more specifically , we have focussed most of our attention on the structure of the magnetic field in the vicinity of the point of intersection of the separatrix ( between the closed- and open - field regions ) and the equator .
we call this point the separatrix intersection point
. the unique singular nature of this point makes it play an extremely important role in the overall global problem ; in particular , without a thorough understanding of the subtleties of the magnetic field behavior near this point , it is impossible to prescribe the correct global boundary conditions in any sensible way .
we start , however , by discussing ( in [ sec - basic ] ) the basic equations governing the global force - free pulsar magnetosphere .
we give special attention to the role played by the light cylinder regularity condition in determining the poloidal electric current .
after this general discussion , the rest of the paper is devoted entirely to the analysis of the separatrix intersection point .
we first consider the separatrix equilibrium condition in the vicinity of this point ( see [ sec - y - point ] ) and find that if it lies at the light cylinder , then all the poloidal current @xmath35 has to return back to the pulsar in the open - field region above the equator , i.e. , there should be no singular current running in a current sheet along the equator and the separatrix .
we then perform ( in [ sec - analysis ] ) an asymptotic analysis of the relativistic grad shafranov , or pulsar , equation in the vicinity of such an intersection point located at the light cylinder .
we find a unique self - similar solution that can be described by the power law @xmath290 , where @xmath139 is the distance form the intersection point and @xmath291 , and by the equator
separatrix angle @xmath292 rad .
however , a further analysis of this solution in the region outside the light cylinder shows that a light surface ( a surface where @xmath293 ) appears just outside the light cylinder ; in particular , we find that this light surface originates right at the intersection point and extends outward at a finite angle with respect to the equator
. the appearance of a light surface in this case is , of course , not surprising , taking into account the fact that ( in this case of the separatrix intersection point lying at the light cylinder ) the poloidal electric current @xmath125 and hence the toroidal magnetic field have to become zero on the last open field line .
we therefore conclude that the only possibility for an ideal - mhd force - free magnetosphere above the putative equatorial current sheet to extend at least some finite distance beyond the light cylinder , is for the separatrix intersection point to be located inside , as supposed to right at , the light cylinder .
( of course we understand that the exact location of this point can only be found as a part of a global solution of the pulsar equation and that it is impossible to determine it from our local analysis ) .
these findings give us the motivation to consider the case when the intersection point lies at some finite distance inside the light cylinder . in
[ sec - t - point ] we examine the behavior of the function @xmath35 near the last open field line @xmath72 ( above the equatorial current sheet ) for this case ; we find that the derivative @xmath294 has to go to zero as @xmath295 , whereas the current itself can approach a finite value @xmath86 .
we then perform an asymptotic analysis of the magnetic field around the intersection point ; it is essentially a somewhat simplified and trivialized analogue of our analysis in [ sec - analysis ] .
we find that the separatrix approaches the equator at a right angle , @xmath68 , and that the field in the region of open field lines behaves simply as @xmath296 , while in the closed - field region the magnetic field is simply vertical and finite near the intersection point ( we call this configuration the t - point ) .
finally , we would like to make several remarks regarding the numerical simulations by ckf and by ogura & kojima ( 2003 ) : \1 ) the midplane boundary conditions ( @xmath297 , @xmath298 ; @xmath51 , @xmath299 ) adopted by both groups have automatically assumed that the separatrix intersection point lies at the lc , and thus have precluded them from even considering the possibility @xmath64 .
it is interesting to note , however , that from the magnetic contour plots presented by ogura & kojima ( 2003 ) it does seem that the actual intersection point lies a little bit inside the lc .
\2 ) neither ckf , nor ogura & kojima ( 2003 ) have discussed or even mentioned the separatrix equilibrium condition ; thus it is not clear whether this condition has been satisfied in their simulations . in the light of our present findings and of the fact that both of these groups have found @xmath86
, we suspect that the equilibrium condition has not , in fact , been satisfied in their studies , at least close to the separatrix intersection point .
this suspicion is strengthened by the fact that both groups have reported having experienced some difficulties near the separatrix .
\3 ) ogura & kojima ( 2003 ) have reported that they had found @xmath300 at some finite distance outside the lc , whereas ckf have claimed that in their solution @xmath301 everywhere . the origin of this discrepancy
is not clear .
it may be attributed to the difference in numerical resolution , although neither of the two groups have conducted very extensive convergence studies .
i am grateful to leonid malyshkin and anatoly spitkovsky for interesting and insightful discussions .
this research was supported by the national science foundation under grant no .
phy99 - 07949 . | we investigate the axisymmetric magnetosphere of an aligned rotating magnetic dipole surrounded by an ideal force - free plasma .
we concentrate on the magnetic field structure around the point of intersection of the separatrix between the open and closed field - line regions and the equatorial plane .
we first study the case where this intersection point is located at the light cylinder .
we find that in this case the separatrix equilibrium condition implies that all the poloidal current must return to the pulsar in the open - field region , i.e. , that there should be no finite current carried by the separatrix / equator current sheet .
we then perform an asymptotic analysis of the pulsar equation near the intersection point and find a unique self - similar solution ; however , a light surface inevitably emerges right outside the light cylinder .
we then perform a similar analysis for the situation where the intersection point lies somewhere inside the light cylinder , in which case a finite current flowing along the separatrix and the equator is allowed .
we find a very simple behavior in this case , characterized by a 90-degree angle between the separatrix and the equator and by finite vertical field in the closed - field region .
finally , we discuss the implications of our results for global numerical studies of pulsar magnetospheres . |
a _ spherical curve _ , or a _ knot projection _ , is a closed curve on the 3-sphere , where each crossing is a double point and crosses transversely . in this paper
we consider spherical curves with at least one crossing .
we call each part of the sphere bounded by a spherical curve a _ region_. in particular , we call a region which has @xmath0 edges on the boundary an _ @xmath0-gon_. a crossing of a spherical curve is _ reducible _ if there are just three regions around the crossing
. a spherical curve @xmath1 is reducible if @xmath1 has a reducible crossing .
@xmath1 is _ reduced _ otherwise .
the reductivity @xmath2 , defined by the second author in @xcite , of a spherical curve @xmath1 represents how far @xmath1 is from a reducible spherical curve .
the precise definition of reductivity will be given in section 2 in this paper .
it is shown in @xcite that the reductivity is four or less for every spherical curve although we have not found a spherical curve with reductivity four .
our next step is to prove that the reductivity is three or less for every spherical curve , or to find a spherical curve whose reductivity is four . + it is shown in @xcite that every reduced spherical curve has a bigon ( 2-gon ) or trigon ( 3-gon ) .
this is derived from the following formula , which came from the euler characteristic , @xmath3 which holds for every reduced spherical curve , where @xmath4 is the number of the @xmath0-gons .
is an unavoidable set for reduced spherical curves .
the set @xmath5 is an unavoidable set for spherical curves with reductivity three or four .
, width=340 ] in other words , the set @xmath6 in fig .
[ st ] is an _
unavoidable set _ for reduced spherical curves . as mentioned in @xcite ,
the reductivity of a spherical curve @xmath1 is two or less if @xmath1 has a bigon .
therefore we can say that the set @xmath5 in fig .
[ st ] consisting of a trigon is an unavoidable set for spherical curves with reductivity three or four .
moreover , we can say that every spherical curve with reductivity three or four has at least eight trigons because of the formula ( [ 238formula ] ) .
+ using the `` discharging method '' with the formula ( [ 238formula ] ) ( inspired by the story of the four color theorem ) , it is shown in @xcite that the set @xmath7 shown in fig .
[ s1 ] is also an unavoidable set for reduced spherical curves , that is , every reduced spherical curve has at least one part in @xmath7 .
is an unavoidable set for reduced spherical curves .
, width=377 ] in this paper we show the following : if there exists a spherical curve @xmath1 with reductivity four , @xmath1 has at least one of the 21 parts depicted in fig .
[ unavoidable - r4 ] , and does not have any other type of the parts in @xmath7 in fig . [ s1 ] .
[ thm - r4 ] note that dotted arcs in the figures represent the outer connection in this paper .
the theorem above implies that the set consisting of the 21 parts in fig .
[ unavoidable - r4 ] is an unavoidable set for spherical curves with reductivity four .
the rest of this paper is organized as follows : in section 2 , we review the reductivities . in section 3 , we classify 4-gons and prove theorem [ thm - r4 ] .
in this section we review the reductivity .
a _ half - twisted splice _ is a local transformation on a spherical curve illustrated in fig .
[ halftwisted ] ( @xcite ) .
we call @xmath8 the inverse of a half - twisted splice
. .,width=302 ] the _ reductivity _ , denoted by @xmath2 , of a spherical curve @xmath1 is the minimal number of @xmath8 required to obtain a reducible spherical curve from @xmath1 . considering the outer connections , bigons and trigons
are divided into the two and four types as illustrated in fig . [ 2and3-di ] .
[ 2and3-co ] for chord diagrams , where a chord diagram of a spherical curve is the preimage of the spherical curve with two points corresponding to the same crossing connected with a segment when we consider the spherical curve as an immersion of a circle to the sphere .
the following propositions are shown in @xcite : let @xmath1 be a spherical curve . * if @xmath1 has an incoherent bigon , @xmath9 . * if @xmath1 has a coherent bigon , @xmath10 . *
if @xmath1 has a trigon of type @xmath11 , @xmath10 . * if @xmath1 has a trigon of type @xmath12 , @xmath13 .
* if @xmath1 has a trigon of type @xmath14 , @xmath13 .
[ bitri ] remark that the converse of proposition [ bitri ] does not hold .
for example , the spherical curve in fig . [ no - bi ] has no bigons , but the reductivity is one . in @xcite , the definitions of spherical curves with reductivity one and two are rephrased diagramatically ( see fact 1 and theorem 1 in @xcite ) and that is useful to determine reductivity .
for example , we can say that the spherical curve @xmath1 depicted in fig .
[ r3ex ] has the reductivity three because @xmath2 is not two or less , and @xmath13 because @xmath1 has trigons of type b and c. is three.,width=188 ] we also have the following proposition : the standard projection of a @xmath15-torus knot has reductivity three ( @xmath16 ) . [ 3n+1 ] note that all the trigons of the standard projection of a @xmath15-torus knot are type b. from the proposition above , we can say that there are infinitely many spherical curves with reductivity three .
similarly , we also have the following .
the standard projection of a @xmath17 and @xmath18-torus knot has the reductivity two .
the standard projection of a @xmath19-torus knot has the reductivity three ( @xmath16 ) .
[ 3n-1 ] note that the standard projection of a @xmath17 and @xmath18-torus knot has a trigon of type a. for spherical curves @xmath1 and @xmath20 , we call the spherical curve obtained from @xmath1 and @xmath20 by connecting their arcs as depicted in fig . [ conn ] a _ connected sum _ of @xmath1 and @xmath20 , and denote it by @xmath21 .
we say a spherical curve is _ prime _ if it is not any connected sum of two ( nontrivial ) spherical curves .
since one part can be regarded as a small 1-tangle for the other one , we have the following inequality : @xmath22 in @xcite , the chord diagrams for all the prime spherical curves with up to ten crossings are listed .
we have the following : all the spherical curves with 10 or less crossings except the standard projection of a @xmath23-torus knot have the reductivity two or less .
all the prime spherical curves with ten or less crossings except the standard projections of the @xmath24 and @xmath25-torus knot ( see fig .
[ 34-torus ] and [ 43 - 35-torus ] ) have a bigon . by propositions
[ bitri ] , [ 3n+1 ] and [ 3n-1 ] , the claim holds for prime spherical curves . for non - prime spherical curves , since they are connected sums of spherical curves up to nine crossings , the claim holds .
note that the reductivity of a connected sum of the standard projection of the @xmath23-torus knot and the spherical curve with one or two crossings is zero .
-torus knot.,width=151 ] and @xmath25-torus knot.,width=302 ]
in this section we classify 4-gons and prove theorem [ thm - r4 ] .
we first classify 4-gons into type 1 to 4 with respect to the relative orientations of the boundary as illustrated in fig .
[ 4-ori ] . then we classify them with respect to the outside connection and name them @xmath26 and @xmath27 as shown in fig .
[ 4-co ] for chord diagrams .
we have the following : if a spherical curve @xmath1 has a 4-gon of type @xmath28 or @xmath29 , then @xmath13 .
[ prop3 - 4gon ] the 4-gons of type @xmath28 and @xmath29 are illustrated in fig .
[ 3d-4-di ] .
.,width=340 ] by applying @xmath8 at the marked crossing , we obtain a trigon of type a from each 4-gons .
now we prove theorem [ thm - r4 ] .
_ proof of theorem [ thm - r4 ] _
if there exists a spherical curve @xmath1 with @xmath30 , @xmath1 has at least one parts of @xmath7 in fig .
[ s1 ] because @xmath1 is a reduced spherical curve . here
, @xmath1 can not have any bigon , trigon of type a , b or c , or 4-gon of type @xmath28 or @xmath29 by propositions [ bitri ] and [ prop3 - 4gon ] .
then @xmath1 can not have the first part of @xmath7 because it is a bigon .
and @xmath1 can not have the second part of @xmath7 because even if one of the two trigons is type d , then another one should be a trigon of type a or b because the boundary is incoherent .
hence it is sufficient to consider the rest three parts of @xmath7 .
if @xmath1 has the third part of @xmath7 , the both two trigons should be type d , and we have just one suitable outer connection illustrated in fig . [ unavoidable - r4 ] .
if @xmath1 has the fourth part of @xmath7 , the trigon should be type d , and then the orientation of the 4-gon should be type 2 or 4 ( see fig .
[ 4ofu ] ) .
more precisely , the 4-gon should be type @xmath31 , @xmath32 or @xmath27 by proposition [ prop3 - 4gon ] .
.,width=113 ] , @xmath32 and @xmath27.,width=264 ] for the case of @xmath31 , the thick edge of @xmath31 in fig .
[ 44edge ] is to be shared with the trigon because of the orientation .
now we consider the outer connection .
[ 4ofu ] . considering the outer connection of d and @xmath31 , @xmath33 and @xmath34
should be connected outside . then @xmath35 and @xmath36 , @xmath37 and @xmath38 should be connected outside .
thus we have just one suitable outer connection of d and @xmath31 .
next we consider the case of @xmath32 . the thick edge of @xmath32 in fig .
[ 44edge ] is to be shared with the trigon .
considering the outer connection of @xmath39 , @xmath40 should be connected outside with @xmath41 or @xmath34 . on the other hand ,
considering the outer connection of @xmath32 , @xmath40 should be connected outside with @xmath36 .
hence there are no suitable connection of d and @xmath32 .
next we consider the case of @xmath27 . from a viewpoint of the orientation ,
each edge of @xmath27 may be shared with the trigon , and it is sufficient to consider the two thick edges in fig .
[ 44edge ] by symmetry .
first , we consider the case the left - side edge of @xmath27 in fig .
[ 44edge ] is shared with the trigon .
considering the outer connection of d and @xmath27 , @xmath40 and @xmath35 should be connected outside . then @xmath34 and @xmath41 , @xmath33 and @xmath36 , and @xmath37 and @xmath38 should be connected outside .
thus we obtain just one suitable connection in this case .
next we consider the case the upper - side edge of @xmath27 in fig .
[ 44edge ] is shared with the trigon .
considering the outside connection of d , @xmath40 should be connected outside with @xmath41 or @xmath35 . on the other hand ,
considering the outer connection of @xmath27 , @xmath40 should be connected outside with @xmath36 .
hence there are no suitable connection in this case . the parts on the fifth part of @xmath7 are similarly obtained . @xmath42
the second author was partially supported by the sumitomo foundation . 9 c. c. adams , r. shinjo , k. tanaka , complementary regions of knot and link diagrams , _ ann .
comb . _ * 15 * ( 2011 ) , 549563 .
j. a. calvo , knot enumeration through flypes and twisted splices , j. knot theory ramifications * 6 * ( 1997 ) , 785798 . m. chmutov , t. hulse , a. lum , p. rowell , plane and spherical curves : an investigation of their invariants , proceedings of the research experiences for undergraduates program in mathematics , oregon state university , 2006 , 192 . n. ito and a. shimizu , the half - twisted splice operation on reduced knot projections , j. knot theory ramifications * 21 * , 1250112 ( 2012 ) [ 10 pages ] . n. ito and y. takimura ,
knot projections with reductivity two , topology appl .
* 193 * ( 2015 ) , 290301 .
a. shimizu , the reductivity of spherical curves , topology appl .
* 196 * ( 2015 ) , 860867 . | the reductivity of a spherical curve represents how reduced the spherical curve is .
it is unknown if there exists a spherical curve whose reductivity is four . in this paper
we give an unavoidable set for spherical curves with reductivity four by considering 4-gons . |
although the standard model ( sm ) @xcite is the simplest model that successfully explains most of the phenomena and experimental observations in particle physics , there are still some unexplained observations and theoretical issues that leaves unanswered .
in particular , the astrophysical evidences of electrically neutral , non - baryonic and weakly interacting matter , i.e. dark matter ( dm ) can be naturally understood if the sm is extended .
since the evidences of dm are based only on its gravitational effects , its nature and microscopic properties remains unknown .
thus , in particle physics many extensions of the sm can be proposed with new particles that naturally fits the current astrophysical data and experimental constraints in direct dm detection . among the different candidates for dm , extensions with stable and non - relativistic ( cold ) particles with masses between a few gev and 1 tev , and interacting through the weak nuclear force ( wimps ) are currently the best motivated models which provides a natural scenery for the cold dark matter ( cdm ) .
for example , supersymmetry provides a wimp candidate through the lightest neutralino @xcite . in universal extra dimension models ,
the lightest kaluza - klein partner is stable and a possible dm candidate @xcite .
other models with wimp candidates have been recently proposed , as for example branons in large extra dimensions @xcite , t - odd particles in little higgs models @xcite , and excited states in warped extra dimensions @xcite . also , scalar candidates in scalar extensions of the sm through singlets and doublets have been widely considered @xcite in the literature , and more recently in 3 - 3 - 1 models @xcite .
in particular , after the observation of the higgs boson at the lhc @xcite , the existence of scalar elementary particles in the universe have been experimentally established .
thus , extensions of the sm with additional scalar sectors are interesting alternatives that may provide natural scenarios for cdm that match the experimental data . on the other hand , models with extra @xmath0 symmetry are one of the most studied extensions of the sm , which implies many phenomenological and theoretical advantages including flavor physics @xcite , neutrino physics @xcite , dark matter @xcite , among other effects . a complete review of the above possibilities can be found in reference @xcite .
in particular , family nonuniversal @xmath0 symmetry models have many well - established motivations .
for example , they provide hints for solving the sm flavor puzzle , where even though all the fermions acquire masses at the same scale , @xmath8 gev , experimentally they exhibit very different mass values . these models also imply a new @xmath1 neutral boson , which contains a large number of phenomenological consequences at low and high energies @xcite .
in addition to the new neutral gauge boson @xmath1 , an extended fermion spectrum is necessary in order to obtain an anomaly - free theory .
also , the new symmetry requires an extended scalar sector in order to ( i ) generate the breaking of the new abelian symmetry and ( ii ) obtain heavy masses for the new @xmath1 gauge boson and the extra fermion content .
in this paper we consider an nonuniversal @xmath0 extension of the type from reference @xcite , which exhibits an extended scalar sector with two scalar doublets and two singlets with nontrivial @xmath0 charges ( labeled for this particular model as @xmath9 ) , where the lightest scalar singlet ( denoted as @xmath10 ) is taken as candidate for cdm .
some phenomenological consequences of this model have been studied in the above reference , with special emphasis in the neutral gauge and yukawa sectors .
motivated by the problem of dm in astrophysics , further analyses of this model is continued . among the most interesting features that the model exhibits , we mention the following related to dm : 1 .
the additional @xmath1 neutral boson produces low energy deviations through a @xmath11-@xmath1 mixing angle .
restrictions to these deviations from the cern - lep and the slac collider impose restrictions on the mixing angle and the @xmath9 coupling , as we will show . at the same time , since @xmath10 scatter off via the @xmath9 interaction , these parameters have effects on the dm relic density through its annihilation to neutral gauge boson @xmath11 .
thus , the model can impose constraints on astrophysical observables from low energy particle physics .
although data from atomic parity violation additionaly imposes lower constraints on the mass of @xmath1 around @xmath12 tev , the current limit on extra neutral gauge boson on direct research at lhc imposes a stronger limit above @xmath13 tev , which have effects in the couplings of dm with the gauge boson @xmath11 .
3 . the nonuniversal feature of the quark sector with different @xmath9 charges , could provide new signatures to search for dm presence in hadron colliders .
4 . by implementing
appropriate conserved discrete or continuos symmetries , we can both remove large flavor changing neutral current couplings and odd power couplings which ensure the stability of the dm candidate . at the same time , these symmetries preserve the couplings that allow the dm annihilation to sm particles required to fit the relic density observed in the universe .
taking into account the most recent calculations of the z pole observables in the sm , the new data from atomic parity violation , and the current limit on the extra neutral gauge boson research at lhc , we find constraints on some parameters associated to the @xmath11-@xmath1 mixing angle and the @xmath9 gauge coupling .
taking into account the above constraints , we obtain allowed regions for the dm relic density according to the last results from planck experiment @xcite .
this paper is organized as follows .
is devoted to summarizing the spectrum of the model , the interactions with the light scalar singlet , and the basic condition for this scalar to be a cdm candidate .
we also show the weak couplings , taking into account the small @xmath11-@xmath1 mixing angle . in sec . 3 , we perform a global @xmath14 fit at 95% c.l to obtain constraints from the z pole parameters .
in sec . 4 , after summarizing the basic equations for the thermal relic density , we identify the fundamental interactions that determine the thermal average dm annihilation cross section . using the constraints obtained previously from the z pole analysis and using the software micromegas @xcite , we determine the relic density predicted by the model , and constraints on the dm mass for different ranges of the space of parameters according to the observed relic density by assuming that the scalar singlet provides the total composition of dm in the universe .
the particle content of the model @xcite is composed of ordinary sm particles and new exotic non - sm particles , as shown in tables [ tab : sm - espectro ] and [ tab : exotic - espectro ] , respectively , where column @xmath15 indicates the transformation rules under the sm gauge group @xmath16 , column @xmath9 contains the values of the new quantum number @xmath17 , and in the column labeled _ feature _ , we describe the type of field .
some properties of this spectrum are : 1 .
the @xmath9 symmetry is nonuniversal only in the left - handed sm quark sector : the quark family @xmath18 has @xmath19 while families @xmath20 have @xmath21 . to distinguish the top and bottom quarks from the other flavors , we use the following assignation for the phenomenological families : + @xmath22 2 .
the sm leptons and scalar particles obtain nontrivial values for the new charge @xmath17 .
3 . in order to ensure cancellation of the gauge chiral anomalies ,
the model includes in the quark sector three extra singlets @xmath23 and @xmath24 , where @xmath25 .
they are quasichiral , i.e. chiral under @xmath9 and vector - like under @xmath15 .
4 . in addition , to obtain a realistic model compatible with oscillation data , we include new neutrinos @xmath26 and @xmath27 which may generate seesaw neutrino masses . 5 .
an additional scalar doublet @xmath28 identical to @xmath29 under @xmath15 but with different @xmath9 charges is included in order to avoid massless charged fermions , and where the individual vacuum expectation values ( vevs ) are related to the electroweak vev through the relation @xmath30 .
an extra scalar singlet @xmath31 with vev @xmath32 is required to produce the symmetry breaking of the @xmath9 symmetry .
we assume that it happens at a large scale @xmath33 .
another scalar singlet @xmath10 is introduced .
since it is not essential for the symmetry breaking mechanisms , we may choose any value for its vev .
however , it is necessary that @xmath34 to ensure its stability .
finally , an extra neutral gauge boson @xmath35 is required to obtain a local @xmath9 symmetry . with the above spectrum
, we will determine all the interactions with the scalar particles of the model .
the most general , renormalizable and @xmath36 invariant potential is @xmath37 \nonumber \\ & + & \left| \phi _ 2 \right|^2\left [ \lambda _ 7 \left| \chi _ 0\right| ^2 + \lambda ' _ 7 \left| \sigma _ 0\right|^2 + \lambda ' ' _ 7 \left ( \chi _ 0^*\sigma _ 0+h.c .
\right ) \right ] \nonumber \\ & + & \lambda _ 5 \left| \phi _ 1 \right|^2\left| \phi _ 2 \right|^2+\lambda ' _ 5 \left| \phi _ 1^{\dagger } \phi _ 2 \right|^2 + \lambda _ 8 \left| \chi _ 0\right|^2\left| \sigma _ 0\right|^2+\lambda ' _ 8 \left[\left(\chi _ 0^ * \sigma _ 0\right)^2+h.c.\right ] .
\label{higgs - pot-1}\end{aligned}\ ] ] however , this potential is not suitable for scalar dm candidates . if we assume that the singlet @xmath10 corresponds to the dm portion of the model , we must impose the following properties : 1 . since @xmath10 acquires nontrivial charge @xmath9 , it must be complex in order to obtain massive particles necessary for cdm .
terms involving odd powers of @xmath10 ( i.e. , the @xmath38 , @xmath39 and @xmath40 terms ) induce decay of the dm , which spoils the prediction of the model for the dm relic density .
thus , we must impose either a global discrete or continuos symmetry .
in particular , to reduce additional free parameters , we can eliminate the @xmath41 term by demanding that the potential respect the global symmetry + @xmath42 3 . in order to avoid the above symmetry to break spontaneously or new sources of decay
, @xmath10 must not generate vev during the evolution of the universe .
thus , we demand @xmath43 . with the above conditions , the potential is the same as in ( [ higgs - pot-1 ] ) but with @xmath44 .
when we apply the minimum conditions @xmath45 for each scalar vev @xmath46 , the following relations are obtained : @xmath47 with the above relations replaced in the potential , we can obtain the squared mass matrices . since the singlet @xmath10 decouples from the other sectors at the quadratic order , we do not take into account this particle for the mass mixing .
first , for the charged sector in the basis @xmath48 , we find @xmath49 where : @xmath50 in the above expressions , we defined the following parameters : @xmath51 for the neutral real sector in the basis @xmath52 , written as @xmath53 we find : @xmath54 where the indices @xmath55 label the scalar indices @xmath56 and @xmath57 . the symbol @xmath58 is the kronecker s delta . also , we use for the couplings the following definitions : @xmath59 finally , for the neutral imaginary sector , in the basis @xmath60 and the form @xmath61 we find : @xmath62 it is interesting to observe that by comparing eq . (
[ real - constant ] ) and ( [ imag - constant ] ) , the real and imaginary components relate to each other according to : @xmath63 in particular , for @xmath64 and @xmath65 , the difference between both neutral sectors appears only at the electroweak order .
thus , some particles of the neutral heavy sector become degenerated if we take only the dominant term @xmath66 , which is also the scale of the charged sector .
after diagonalization , we obtain the following mass eigenvectors : @xmath67 where @xmath68 is the identity , and the rotation matrices are defined according to @xmath69 the rotation angles are the angle @xmath70 defined in ( [ parameters ] ) and @xmath71 obtained from the elements of the real matrix ( [ real - scalar - mass ] ) as @xmath72^{-1 } \label{alpha - angle}\end{aligned}\ ] ] where we have taken the dominant contribution assuming that @xmath73 .
we emphasis that although the second term into the inverse of ( [ alpha - angle ] ) is proportional to the small ratio @xmath74 , it is not negligibly small if @xmath75 and @xmath76 . as for the eigenvalues , we obtain for the physical particles the following squared masses at dominant order : @xmath77 from the above results , it is interesting to observe the following properties : 1 . in eq .
( [ alpha - angle ] ) , we can use the approximation @xmath78 as dominant contribution only if the scalar doublets @xmath79 exhibit either * the same self - interacting couplings , i.e. @xmath80 , or * different vevs , i.e. @xmath81 .
2 . as we predicted below eq .
( [ real - imag ] ) , the fields with masses given by the first line of ( [ scalar - mass ] ) are degenerated at dominant order .
3 . in order to obtain masses with real values for the above degenerated particles , we require that either @xmath82 or @xmath83 . on the other hand , we obtain all the couplings of the dm scalar @xmath10 with the above mass eigenstates .
the sector of the potential associated to @xmath10 is : @xmath84 after rotation to mass eigenvectors according to ( [ scalar - eigenvectors ] ) , we obtain all the interactions of @xmath10 with the scalar matter :
@xmath85 where the mass of @xmath10 is : @xmath86 the kinetic terms of the higgs lagrangian reads : @xmath87 the covariant derivative is defined as @xmath88 where @xmath89 corresponds to the pauli matrices for the doublets @xmath90 and @xmath91 for the singlets @xmath92 , while @xmath93 and @xmath94 corresponds to the hypercharge and @xmath9 charge according to values of tabs .
[ tab : sm - espectro ] and [ tab : exotic - espectro ] .
after the symmetry breaking , we obtain the charged eigenstates @xmath95 with mass @xmath96 , while for the neutral sector , we obtain the following squared mass matrix in the neutral gauge basis @xmath97 : @xmath98
\\ \end{pmatrix},\end{aligned}\ ] ] where @xmath99 . taking into account that @xmath100
, we can diagonalize the above matrix with only two rotation angles , obtaining the following mass eigenstates : @xmath101 with : @xmath102 where @xmath103 defines the weinberg angle , and @xmath104 is a small @xmath11-@xmath1 mixing angle between the sm neutral gauge boson @xmath11 and the @xmath9 gauge boson @xmath1 such that in the limit @xmath105 , we obtain @xmath106 and @xmath107 .
we find for the mixing angle that @xmath108 where the neutral masses are : @xmath109 on the other hand , the lagrangian in ( [ higgs - kinetic ] ) also contains the interactions among scalar and vector bosons . in particular , for the dm candidate @xmath10 and considering the mixing @xmath11-@xmath1 angle
, we find the following couplings : @xmath110 where we approximate @xmath111 . in the trilineal terms , we introduce the vectors @xmath112 ( @xmath113 ) corresponding to the @xmath10 ( @xmath114 ) momentum .
the couplings between fermions and gauge bosons , are described by the dirac lagrangian , which reads : @xmath115 where @xmath116 contains sm and non - sm fermions . in particular , for the neutral weak sector , and taking into account the mass states in ( [ gauge - eigenvec ] ) , we obtain @xcite @xmath117f^i \nonumber \\ & -&z_{2\mu } \sum _ { f , i}\overline{f^i}\gamma ^{\mu}\left[\frac{g}{2c_w}(v_i^{sm}-\gamma _ 5a_i^{sm})s_{\theta}+\frac{g_x}{2}(v_i^{nsm}-\gamma _ 5a_i^{nsm})c_{\theta}\right]f^i , \label{weak current}\end{aligned}\ ] ] where the vector and axial couplings for the weak neutral currents are defined according to tab .
[ tab : vector - axial - couplings ] .
we observe that a small non - sm coupling ( @xmath118 ) with the lighter weak boson @xmath119 arises through the mixing angle @xmath120 .
the above lagrangian can be written in a shorter form by defining rotations to modified vector and axial couplings as @xmath121 where : @xmath122 obtaining @xmath123 \label{weak - neutral - lag}\end{aligned}\ ] ]
we obtain the yukawa lagrangian compatible with the @xmath124 symmetry . for the quark sector
we find : @xmath125 where @xmath126 are conjugate scalar doublets , and @xmath127 . for the leptonic sector
we obtain : @xmath128 in particular , we can see in the quark lagrangian from eq .
( [ quark - yukawa-1 ] ) that due to the non - universality of the @xmath9 symmetry , not all couplings between quarks and scalars are allowed by the gauge symmetry , which lead us to specific zero - texture yukawa matrices as studied in ref . @xcite . for the purpose of this paper
, we can assume diagonal yukawa matrices .
in addition , if the global symmetry in ( [ global - symm ] ) is required only for @xmath10 , its couplings with fermions are forbidden , thus we assume that @xmath129 in ( [ quark - yukawa-1 ] ) and ( [ yukawa - leptons-1 ] ) . however , the @xmath10 particle can still scatter off to fermions indirectly through processes of the form @xmath130 , where the decays of the scalar particles into ordinary matter are described by the couplings in the first line of ( [ quark - yukawa-1 ] ) and the last of ( [ yukawa - leptons-1 ] ) . by considering the dominant contributions
, we can decouple the heavy and the light sectors of the fermions to obtain the mass eigenvalues after the symmetry breaking .
in particular , with the above assumptions and under the assignation from ( [ quark - assignation ] ) , the yukawa couplings of scalars to ordinary matter is the same as in a two higgs doublet model type iii .
since these models contains contributions to flavor changing neutral currents , which are severely suppresed by experimental data at electroweak scales , we will assume that the yukawa couplings relate to the mass of quarks in the same way as in a type ii model , avoiding the introduction of unnecesary additional free parameters .
as a result of the mixing in the couplings with the weak boson @xmath119 in ( [ weak - neutral - lag ] ) , small deviations of the sm observable arise .
in particular , we can use the z pole observables shown in tab . [
tab : observables ] with their experimental values from the cern collider ( lep ) , the slac liner collider ( slc ) and data from atomic parity violation @xcite , and the theoretical predictions from sm and the deviations predicted by the @xmath9 model , as shown in the appendix [ app : z - pole ] .
we use @xmath131 gev and @xmath132 .
all the technical details including the electroweak corrections considered are summarized in the appendix .
we perform a @xmath14 fit at @xmath133 c.l . , where the free quantities @xmath134 and the ratio @xmath135 can be constrained at the z peak .
we assume a covariance matrix with elements @xmath136 among the z pole observables , where @xmath137 is the correlation matrix and @xmath138 the quadratic root between the experimental and sm errors .
the @xmath139 statistic with three degrees of freedom ( d.o.f ) is defined as @xcite @xmath140 ^{t}v^{-1}\left [ \mathbf{y}-\mathbf{f}(t_{\beta } , m_{z'},r_g ) \right ] , \label{chi}\ ] ] where @xmath141 represents the 22 experimental observables from table [ tab : observables ] , and @xmath142 the corresponding prediction from the @xmath9 model .
table [ tab : correlation ] displays the symmetrical correlation matrices taken from ref .
the function in ( [ chi ] ) imposes constraints to the free variables ( @xmath143 ) by requiring that @xmath144 . for three d.o.f the @xmath133 c.l .
corresponds to @xmath145 . with the family assignation in ( [ quark - assignation ] ) and the z pole parameters , we find for the minimum that @xmath146 .
we obtain two - dimensional projections of the @xmath14 function . fig .
[ fig1 ] shows the limits for the @xmath1 mass in the ( @xmath147 ) plane .
the shaded areas correspond to allow regions for different values of @xmath148 , where we identify bounds for @xmath149 and the ratio @xmath150 .
thus , for example , around the experimental lowest limit of @xmath151 tev , we see that the model is allowed if @xmath152 and @xmath153 when @xmath154 and @xmath155 , respectively .
these limits on the coupling constant ratio increase for larger @xmath1 masses . in particular , for a model with @xmath156
, we find that @xmath149 must be above @xmath157 and @xmath158 tev if @xmath159 and @xmath160 , respectively . on the other hand , fig .
[ fig2 ] shows the constraints on @xmath150 as function of @xmath148 .
we observe that for @xmath2 tev , the model is completely excluded for @xmath161 .
we also see that for a @xmath13 tev @xmath1 boson , values near @xmath162 are not allowed in the range @xmath163 . in the scenario with @xmath156 , the range @xmath164 is excluded for a @xmath165 tev @xmath1 boson , while for a @xmath166 tev gauge boson , the model exhibits allowed points only at large @xmath148 values .
we found that the above constraints do not change significantly for values above @xmath167 .
the boltzmann equation for the ratio of the number density for dm particles is @xcite : @xmath168 where three sources for variation of the density are identified : first , a negative contribution due to the expansion of the universe described by the hubble constant ( @xmath169 ) .
second , there is another negative contribution coming from the annihilation of dm into ordinary matter .
this contribution is described by a thermally - averaged pair - annihilation cross sections times the relative velocity between the dm particles @xmath170 , and the instantaneous squared density @xmath171 of dm ( the higher the density , the greater the annihilation probability ) .
finally , there is a positive contribution which describes dm creation through matter collisions , which exhibits the same thermal cross section .
this contribution depends on the number of particles in thermal equilibrium ( @xmath172 ) .
the boltzmann equation is conveniently written in terms of the abundance @xmath173 and the temperature @xmath23 , with @xmath174 the entropy of the universe , obtaining : @xmath175 where @xmath176 is the planck mass and @xmath177 the number of degrees of freedom in thermal equilibrium with the photons and with masses smaller than @xmath23 . by solving the above equation , it is possible to calculate the relic density , which is defined as : @xmath178 where @xmath179 is the critical density and @xmath180 the mass of the dm candidate , while @xmath181 and @xmath182 are the entropy and abundance evaluated at the today temperature @xmath183 k measured from the microwave background . in our case , we will consider that the complex @xmath10 particle is the single dm component of the universe , thus @xmath184 and the thermal cross section is obtained from all the dispersions of @xmath10 according to the couplings obtained in section 2 . to calculate the freeze - out temperature , the thermal cross section and the relic density
, we use the numerical code micromegas @xcite , where we assume that the abundance accomplish the condition @xmath185 for high temperatures until freeze - out , and @xmath186 for temperatures below the freeze - out @xcite .
for the thermal cross section , we identify the following parameters : 1 . for direct dispersions into neutral gauge bosons , @xmath187 described by the couplings in eq .
( [ gauge - interaction ] ) , we include the restrictions derived from the z pole constraints for the parameters @xmath148 , @xmath118 and @xmath149 . in particular , according to fig .
[ fig1 ] , the larger value of the gauge coupling ratio is around @xmath188 for a @xmath13 tev extra neutral gauge boson .
thus , we reduce the number of parameters by choosing the above values , which are compatible with the z pole constraints . according to eq .
( [ mixing - angle ] ) , these values lead us to a @xmath11-@xmath1 mixing angle as large as @xmath189 .
2 . for indirect dispersions through the scalar particles @xmath190 , @xmath191 described by ( [ sigma - couplings ] ) , we can reduce the number of parameters by considering a model with two identically self - interacting scalar doublets , where @xmath192 .
this assumption lead us that the angle @xmath71 in eq .
( [ alpha - angle ] ) become + @xmath193 + thus , if we assume that @xmath194 is the experimentally observed boson at lhc ( where @xmath195 gev ) , both angles are related by only one free parameter , i.e the mass of the neutral higgs boson @xmath196 . since no positive signal in the search for extra neutral higgs bosons at lhc has been reported , we choose a large value @xmath197 gev .
the other free parameters in ( [ sigma - couplings ] ) are the couplings @xmath198 and @xmath199 .
we recall that the pseudo scalar @xmath200 and the charged scalar @xmath201 have masses at the same order as @xmath196 at dominant order , according to eq .
( [ scalar - mass ] ) .
3 . using a two higgs doublet model type ii for the yukawa couplings ,
the decay of the scalar matter into ordinary fermions @xmath202 involve the fermion masses , the electroweak vev and the angle @xmath203 .
thus , these couplings do not introduce additional free parameters .
the decay of scalars into gauge bosons @xmath204 only involve the known @xmath3 coupling constant @xmath205 .
thus , these interactions do not introduce more free parameters neither .
thus , our space of parameters is composed by the four variables @xmath206 .
taking into account that the current limit for dm relic density is @xmath207 from the planck experiment @xcite , we search for limits on the mass of the scalar dm candidate @xmath208 for different combinations of the parameters .
we obtain the following results . 1 . in order to explore the contribution of the gauge @xmath119 and @xmath209 bosons , we first consider the _ higgsphobic _ case , by fixing @xmath210 .
[ fig3ab]-(a ) shows the relic density as function of @xmath211 without @xmath11-@xmath1 mixing . according to eq .
( [ gauge - interaction ] ) , the light @xmath106 gauge boson decouple from the dm when @xmath105 , thus the dominant channel for dm annihilation is through the @xmath107 gauge boson .
we see that below @xmath212 gev , very large dm relic density is obtained which means that there are few dm annihilation through the freeze - out due to the small coupling to ordinary matter
. however , when the mass of the @xmath10 particle increases until the electroweak threshold at @xmath213 gev , labeled in figure ( a ) as * a * , the probability of annihilation into ordinary matter grows .
we see that at @xmath214 gev ( labeled as * b * ) , there arises a resonance due to the pole @xmath215 , which produces an extra excess of annihilation through the process @xmath216 .
the horizontal line is the experimental limit for the relic density .
thus , for this case , the appropriate mass is @xmath212 gev to fit the data . in fig .
[ fig3ab]-(b ) we take values for the mixing angle . in particular , we choose random values for the parameter @xmath148 , which according to the z pole limits gives values as large as @xmath217 for @xmath2 tev .
the plot is the same as the curve in ( a ) , but with a new small resonance at @xmath218 gev ( label * c * ) corresponding to the pole @xmath219 , which is consequence of the coupling of @xmath11 bosons to dm .
although the relic density decreases at this resonance , the annihilation ratio is not large enough to reach the observable density .
thus , the z pole data produce strong constraints to fit the relic density for higgsphobic models . 2 .
let us take @xmath220 , which open the scalar channels .
we consider small values for @xmath148 .
in particular , we choose random values between @xmath160 and @xmath221 , obtaining the plot from fig .
first , we see for light dm that the relic density decreases in relation to the above case .
thus , the coupling with the scalar matter produces larger annihilation rates than the coupling with gauge bosons .
in particular , for @xmath222 , the scalar channels produce lower relic densities that the resonance of the @xmath119 boson , thus this resonance does not appear in the plot . in more detail , we see that there are multiple kinematical thresholds through the curve with observable drops of the density .
the first threshold happens at @xmath223 gev ( * d * ) , due to dispersions of @xmath138 into charm quarks .
later , there is another drop at @xmath224 gev ( * e * ) from production of bottom quarks .
finally , there is a third large drop due to dispersions into @xmath225 and @xmath11 gauge bosons at @xmath226 gev ( * a * ) . on the other hand ,
we also identify sharp resonances associated to production of intermediate scalar particles .
the first at @xmath227 gev ( * f * ) corresponding to the process @xmath228 through the sm - like higgs boson with mass of @xmath229 gev .
we see that below the electroweak threshold , the density is above the experimental value , except for the resonance , that exhibits allowed points around @xmath230 gev .
the other allow point is at @xmath5 gev which is the same as the one obtained in figure [ fig3ab ] . above the electroweak threshold
, the relic density drops below the observable value .
in particular , we see that there are two others resonances at @xmath231 gev ( * g * ) and @xmath232 gev ( * b * ) .
3 . in fig .
[ fig5 ] , we show the relic density for random values in the ranges @xmath233 and @xmath234 .
we see that the general form of the plot from fig .
[ fig4 ] is preserved .
however , solutions associated to large values of @xmath148 and @xmath235 appears through the horizontal line , as shown .
the allow values for @xmath180 is an important contraint for dm research in direct detection and production in collider experiments . on the other hand ,
, we see that the resonance due to sm - like higgs boson production at * f * gives a relic density below the observable value , producing a deep loss of dm in the window @xmath236 gev .
we also see that there are solutions thorugh the observable relic density in the range @xmath237 gev 4 .
finally , we open our space of parameters by extending the @xmath238 between @xmath160 to @xmath155 , and the @xmath235 between @xmath160 to the limit for unitarity of @xmath239 . fig .
[ fig6 ] shows the results for random values .
we identify again the three old kinematical threshold @xmath240 , @xmath241 and @xmath242 and the same resonances @xmath243 , @xmath244 and @xmath245 .
in addition , another peak at @xmath246 gev appears ( * h * ) , which is hardly visible in fig . [ fig5 ] .
this peak is not due to a new resonance but a fourth kinematical threshold from dispersions of @xmath10 into higgs bosons @xmath194 due to large coupling values @xmath235 that increases the dispersion rates into the higgs .
roughly , we see that there exist values of the space of parameters that reproduce the experimental relic density for the range @xmath247 gev and @xmath248 gev . by comparing figs .
[ fig5 ] and [ fig6 ] , we see that the larger the scalar couplings , the larger the allow ranges for the dm mass .
the puzzle of the dm in the universe can be related with particles that exhibit additional non - gravitational interactions in the framework of particle physics theories . in particular , extensions with an extra non - universal abelian @xmath0 symmetry
are very well - motivated models that can introduce different dm candidates with new weak interactions through the addition of a neutral gauge boson @xmath1 . in this work , we studied some consequences of this neutral current in an specific @xmath0 model with two scalar doublets and two scalar singlets that exhibit nontrivial @xmath0 charges , where one of the singlet is taken as a dm candidate .
we first explore some consequences of the @xmath11-@xmath1 mixing using data from lep , slac and atomic parity violation .
we found for the lower experimental limit @xmath2 tev , that the @xmath0 coupling constant is constrained to values as large as @xmath249 for @xmath250 and at @xmath133 c.l .
this limit increases for larger @xmath149 values .
second , the @xmath11-@xmath1 mixing induces dispersions of the dm candidate @xmath10 through @xmath11 bosons .
however , due to the lep constraints , this interaction is largely suppressed by small mixing angles ( of the order of @xmath251 ) .
although in the resonance @xmath252 the annihilation of dm enhance , it is not enough to reduce the relic density to the observable values . however , there arises a threshold where dm can annihilate to ordinary matter through interactions with the heavy gauge boson @xmath1 , obtaining realistic relic density at @xmath5 gev . on the other hand , by considering dispersions through scalar matter , we found smaller relic densities than through gauge interactions . for small higgs couplings , we still obtain large relic densities for light dm .
however , near the higgs boson resonance @xmath227 gev , the annihilation rate sharply enhance to experimental limits on the relic density .
if we extend the values of the parameters to wider ranges , we obtain larger allowed regions , with masses as low as @xmath253 gev , and as large as @xmath254 gev . in particular
, we obtain that the larger the scalar coupling constants , the larger the range for @xmath211 . the constraints on @xmath211 and the thresholds exhibited by the relic denstiy may have observable consequences for dm research in experiments from direct detection and production in colliders , providing possible signatures in final states distributions that allow identify the presence of dm particles .
the z pole parameters with their experimental values from cern collider ( lep ) , slac liner collider ( slc ) and data from atomic parity violation taken from ref .
@xcite , are shown in table tab : observables , with the sm predictions and the expressions predicted by the extra @xmath9 model . the corresponding correlation matrix from ref .
@xcite is given in table [ tab : correlation ] .
we will use for the mass the approximation @xmath255 . in the sm
, the partial decay widths of @xmath256 into fermion species @xmath257 is described by @xcite : @xmath258 r_{qed}r_{qcd } , \label{partial - decay}\ ] ] where @xmath259 , 3 for leptons and quarks , respectively . @xmath260 and @xmath261 are qed and qcd corrections , and @xmath262 considers kinematical corrections only important for the @xmath263-quark .
universal electroweak corrections sensitive to the top quark mass are taken into account in @xmath264 and in @xmath265 which is written in terms of an effective weinberg angle data @xmath266 with @xmath267 .
nonuniversal vertex corrections are also taken into account in the @xmath268 vertex with additional one - loop leading terms which leads to @xmath269 and @xmath270 @xcite .
for the top and bottom quark masses , we use the following values calculated at the z pole scale @xcite : @xmath271 for the partial sm partial decay given by eq .
( [ partial - decay ] ) , we use the following values taken from ref .
@xcite @xmath272 for the @xmath9 model in the fourth column of tab .
[ tab : observables ] , we define from ( [ modified - coup ] ) the coupling deviations @xmath273 and we approximate @xmath274 for the mixing angle .
thus , the analytical expressions for the deviations of the z pole observables are : @xmath275 where for the light fermions @xmath276 while for the @xmath263-quark @xmath277 where @xmath278 . the above expressions are evaluated in terms of the effective weinberg angle from eq .
( [ effective - angle ] ) .
the weak charge is written as @xmath279 where @xmath280 .
the deviation @xmath281 is @xcite @xmath282 \delta \rho _ { m}+\delta q_{w}^{\prime } , \label{dev}\]]and @xmath283 which contains new physics gives @xmath284 s_{\theta } \notag \\ & & -16\left [ \left ( 2z+n\right ) { a}_{e}^{nsm}{v}_{u}^{nsm}+\left ( z+2n\right ) { a}_{e}^{nsm}{v}_{d}^{nsm}\right ] \frac{m_{z}^{2}}{m_{z'}^{2}}. \label{new}\end{aligned}\ ] ] for cesium , and for the first term in ( [ dev ] ) we take the value @xmath285 \delta \rho _ { m}\simeq -0.01 $ ] @xcite .
this work was supported by the departamento administrativo de ciencia , tecnologa e innovacin ( colciencias ) in colombia .
glashow , nucl .
22 , 579 ( 1961 ) ; s. weinberg , phys .
19 , 1264 ( 1967 ) ; a. salam , in _ elementary particle theory : relativistic groups and analyticity ( nobel symposium no .
8) _ , edited by n.svartholm ( almqvist and wiksell , stockholm , 1968 ) , p. 367 . g. servant and t. m. p. tait , nucl .
b650 , 391 ( 2003 ) ; h .- c .
cheng , j. l. feng , and k. t. matchev , phys .
. 89 211301 ( 2002 ) ; f. burnell and g. d. kribs , phys .
d73 , 015001 ( 2006 ) ; k. kong and k. t. matchev , jhep 01 , 038 ( 2006 ) ; m. kakizaki , s. matsumoto , and m. senami , phys . rev .
d74 , 023504 ( 2006 ) .
john mcdonald , phys . rev .
d 50 , 3637 ( 1994 ) ; c.p .
burgess , maxim pospelov , and tonnis ter veldhuis , nucl .
b619 , 709 ( 2001 ) ; riccardo barbieri , lawrence j. hall , and vyacheslav s. rychkov , phys . rev .
d 74 , 015007 ( 2006 ) ; laura lopez honorez , emmanuel nezri , josep f. oliver , and michel h. g. tytgat , jcap 0702 , 028 ( 2007 ) ; laura lopez honorez and carlos e. yaguna , jhep 09 , 046 ( 2010 ) .
stefano profumo , farinaldo s. queiroz , eur .
j. 74 , 2960 ( 2014 ) ; chris kelso , c. a. de s. pires , stefano profumo , farinaldo s. queiroz , p. s. rodrigues da silva , eur . phys .
j. c74 , 2797 ( 2014 ) ; d. cogollo , alma x. gonzalez - morales , farinaldo s. queiroz , patricia r. teles , arxiv : hep - ph/1402.3271 . the aleph collaboration , the delphi collaboration , the l3 collaboration , the opal collaboration , the sld collaboration , the lep electroweak working group , the sld electroweak and heavy flavour groups , phys.rept .
427 ( 2006 ) 257 - 454 [ arxiv : 0509008 [ hep - ex ] ] . @xmath289u(1)_x@xmath290_z@xmath291@xmath292@xmath293@xmath294_z^sm ( 1+_z ) @xmath295_had@xmath291@xmath296@xmath297@xmath294_had^sm ( 1+_had ) @xmath295 _ ( ^+^- ) @xmath291mev@xmath298@xmath299@xmath300 _ ( ^+^- ) ^sm ( 1 + _ ) @xmath295_had@xmath291@xmath301@xmath302@xmath303_had^sm ( 1 + _ ) @xmath295r_e@xmath304@xmath305@xmath306r_e^sm ( 1+_had+_e ) @xmath295r_@xmath307@xmath308@xmath306r_^sm ( 1+_had+ _ ) @xmath295r_@xmath309@xmath310@xmath306r_^sm ( 1+_had+ _ ) @xmath295r_b@xmath311@xmath312@xmath313r_b^sm ( 1+_b-_had ) @xmath295r_c@xmath314@xmath315@xmath313r_c^sm ( 1+_c-_had ) @xmath295a_e@xmath316@xmath317@xmath318a_e^sm ( 1+a_e ) @xmath295a_@xmath319@xmath320@xmath318a_^sm ( 1+a _ ) @xmath295a_@xmath321@xmath320@xmath318a_^sm ( 1+a _ ) @xmath295a_b@xmath322@xmath323@xmath324a_b^sm ( 1+a_b ) @xmath295a_c@xmath325@xmath326@xmath327a_c^sm ( 1+a_c ) @xmath295a_s@xmath328@xmath329@xmath324a_s^sm ( 1+a_s ) @xmath295a_fb^ ( 0,e ) @xmath330@xmath331@xmath332a_fb^(0,e)sm ( 1 + 2a_e ) @xmath295a_fb^ ( 0 , ) @xmath333@xmath334@xmath335a_fb^(0,)sm ( 1+a_e+a _ ) @xmath336a_fb^ ( 0 , ) @xmath337@xmath338@xmath335a_fb^(0,)sm ( 1+a_e+a _ ) @xmath295a_fb^ ( 0,b ) @xmath339@xmath340@xmath341a_fb^(0,b)sm ( 1+a_e+a_b ) @xmath295a_fb^ ( 0,c ) @xmath342@xmath343@xmath344a_fb^(0,c)sm ( 1+a_e+a_c ) @xmath295a_fb^ ( 0,s ) @xmath345@xmath346@xmath341a_fb^(0,s)sm ( 1+a_e+a_s ) @xmath295q_w(cs)@xmath347-@xmath348@xmath349-@xmath350@xmath351q_w^sm ( 1+q_w ) @xmath352 in the @xmath353 plane for different values of @xmath238 , with @xmath354 the ratio between the @xmath9 and @xmath3 gauge couplings .
the shaded areas show the allowed points from z pole constraints.,width=415,height=302 ] in the @xmath355 plane for different values of @xmath356 , with @xmath354 the ratio between the @xmath9 and @xmath3 gauge couplings .
the shaded areas show the allowed points from z pole constraints , width=415,height=302 ] | in the framework of an nonuniversal @xmath0 extension of the standard model , we propose an scalar candidate for cold dark matter which exhibits interactions with ordinary matter through higgs and gauge bosons .
using limits from low energy observables , we find constraints on the new physics parameters of the model associated to the extra abelian symmetry , in particular , the mass of the additional neutral gauge boson @xmath1 and the new gauge coupling constant .
we found that for the lower experimental limit @xmath2 tev constrained by direct research at lhc , the ratio between the @xmath0 and @xmath3 gauge coupling constants is around @xmath4 .
taking into account this limit and the observable relic density of the universe , we search for new constraints on the mass of the dark matter particle of the model .
we found that for a higgsphobic model , the mass of the scalar dark matter must be @xmath5 gev .
we also found different kinematical threshold and resonances that enhance the dispersion of dark matter into ordinary matter for different regions of the space of parameters of the model , obtaining masses as low as @xmath6 gev and as large as @xmath7 gev with not allowed intermediate regions due to resonances . |
wireless relay networks are ubiquitous from classical satellite communications to the increasing use of wireless ad hoc networks where mobiles , tablets , or laptops have the ability to act as wireless relays for other similar devices .
a common coding strategy deployed in relay networks is for the relays to decode the source message , and then forward a function of the message ( e.g. , some parity bits ) to the destination to facilitate decoding . in this context , one can use a _ partial - decode - forward ( pdf ) _ scheme where the relays decode only a fraction of the source message and forward a function of the decoded part , or a _ complete - decode - forward ( cdf ) _ scheme where the relays fully decode the source message and forward a function of it .
another consideration in designing wireless relay networks is whether to use a full - duplex ( communication possible in both directions simultaneously without interference ) or a half - duplex ( communication possible in only one direction at a time ) scheme . the choice of full- versus half - duplex , as well as the choice of cdf or pdf , impacts on the achievable communication rate .
achievable rates have been demonstrated for the full - duplex relay channel using cdf @xcite and pdf @xcite , ( * ? ? ?
16.3 ) , and for the half - duplex relay channel using cdf @xcite and pdf @xcite .
pdf includes cdf and direct transmission as special cases the relay decodes all source messages in the former and decodes nothing in the latter .
for the additive white gaussian noise ( awgn ) full - duplex relay channel , it has been shown @xcite that ( i ) when the source - destination link is better than the source - relay link , the maximum pdf rate can be achieved by not using the relay at all , i.e. , direct transmission ; and ( ii ) when the source - relay link is better than the source - destination link , the maximum pdf rate can be achieved by having the relay decoding all source messages , i.e. , cdf .
while it is not desirable to have the relay decoding only part of the messages in the awgn full - duplex relay channel , pdf can achieve rates strictly higher than direct transmission and cdf in the half - duplex counterpart . in this paper
, we focus on the half - duplex relay channel , and compare the rates achievable by existing coding schemes . for a given half - duplex relay channel ,
let @xmath0 and @xmath1 be the maximum rates achievable ( optimized over channel codes under respective coding schemes ) using the complete - decode - forward and the partial - decode - forward schemes respectively .
for the trivial case where the source - destination link is better than the source - relay link , @xmath1 can be attained by direct transmission ( similar to the full - duplex case ) .
otherwise , we have the following : [ theorem : main ] consider the awgn half - duplex single - relay channel .
if the source - relay link is better than the source - destination link , then @xmath2 this means regardless of channel parameters , partial decoding can achieve at most 12.5% rate gain compared to full decoding .
we further show that in the following three asymptotic cases , pdf and cdf actually achieve the same rates : ( i ) the relay is close to the source , ( ii ) the relay is close to the destination , and ( iii ) the signal - to - noise ratio ( snr ) is low .
we also show that for the case of high snr , cdf achieves rates within a constant bit gap ( which is a function of the channel gains ) of that achievable by pdf .
this gap , when normalized to the cdf rate , diminishes as the snr increases . theoretically , this work gives an upper bound to the performance gain of pdf over cdf .
practically , it shows that without a significant percentage loss in transmission rate , one could choose cdf over pdf , as the former is potentially less complex to implement , and is certainly less complex to design .
[ fig : src ] denotes the single - relay channel consisting of the source ( denoted by node 0 ) , the relay ( node 1 ) , and the destination ( node 2 ) .
we consider the awgn channel where ( i ) the received signal at the relay is given by @xmath3 , and ( ii ) the received signal at the destination is given by @xmath4 . here , @xmath5 is the transmitted signal of node @xmath6 , @xmath7 is the noise at node @xmath8 , and @xmath9 is the channel gain from node @xmath6 to node @xmath8 .
each @xmath7 is an independent zero - mean gaussian random variable with variance @xmath10 = n_j$ ] .
the channel gains are fixed and are made known a priori to all the nodes .
consider a block of @xmath11 channel uses , and let @xmath12 ( and @xmath13 ) denote the transmitted ( and received ) symbol of node @xmath6 at the @xmath14-th channel use .
a rate-@xmath15 block code comprises ( i ) a message set @xmath16 , ( ii ) an encoding function for the source @xmath17 for each message @xmath18 , ( iii ) encoding functions for the relay @xmath19 for all @xmath20 , and ( iv ) a decoding function for the destination @xmath21 . here
, we have used bold - faced symbols to represent vectors of length @xmath11 , e.g. , @xmath22 .
we assume that the message is uniformly distributed in @xmath23 .
the rate @xmath15 is achievable if there exists a sequence of rate-@xmath15 block codes such that @xmath24 as @xmath25 . for the half - duplex channel
, the relay can only transmit or receive ( i.e. , listen ) , but not do both , at any time .
we set @xmath26 if the relay transmits at the @xmath14-th channel use , and @xmath27 otherwise , i.e. , if the relay listens .
note that the definition of block code above is applicable for both full - duplex and half - duplex relay channels with these extra transmit / listen constraints . without
loss of optimality , the source always transmits , and the destination always listens .
we consider a _
fixed - slot _ structure where the transmit / listen mode of the relay for each channel use is fixed prior to the transmissions and is known to all nodes @xcite .
consider @xmath11 channel uses during which the relay listens in @xmath28 channel uses and transmits in @xmath29 channel uses .
we assume the following _ per - symbol _ power constraints @xcite : @xmath30 \leq p_i$ ] for both @xmath31 and for all @xmath32 , where the expectation operation @xmath33 $ ] is taken over the message @xmath34 and the channel noise @xmath35 . as a result
, the source and the relay can not optimize their transmit power for variations in the relay s transmit / listen mode .. ] we denote the snr from node @xmath6 to node @xmath8 by @xmath36 .
we assume that @xmath37 for all @xmath6 and @xmath8 .
we now summarize the encoding and decoding schemes of pdf for the half - duplex relay channel @xcite .
consider @xmath38 blocks , each of @xmath11 channel uses , and @xmath39 source messages @xmath40 , where each @xmath41 .
the source splits each message @xmath42 into two independent parts @xmath43 , where @xmath44 and @xmath45 so that @xmath46 . in block
@xmath47 , the source transmits @xmath48 .
the relay decodes @xmath49 in block @xmath50 , and transmits it in block @xmath51 .
the rate of this code over the entire @xmath38 transmission blocks is @xmath52 bits / channel use .
if the destination can decode all @xmath40 reliably ( i.e. , with diminishing error probability as @xmath11 increases ) , then by choosing a sufficiently large @xmath38 , any rate below and arbitrarily close to @xmath15 is achievable . of the time the relay listens , and in @xmath53 of the time the relay transmits ] as the encoding and decoding operations repeat themselves over the blocks , we focus on encoding in block @xmath50 and decoding over blocks @xmath50 and @xmath51 , depicted in fig .
[ fig : pdf - half ] . during the first @xmath28 channel uses of block @xmath50 ( when the relay listens ) channel uses and
transmits in the remaining channel uses of each block . in general , the channel uses in which the relay listens or transmits can be arbitrarily chosen in each block without affecting the achievable rates as long as @xmath54 is fixed . ] , the source transmits @xmath55 . here
, we have used the superscript @xmath56 to denote the length of the sub - codeword .
assume that the relay has decoded @xmath57 in the previous block , i.e. , block @xmath58 .
during the remaining @xmath29 channel uses of block @xmath50 , the relay transmits an independently generated codeword @xmath59 , while the source splits its transmit power into two parts : ( i ) a portion of @xmath60 of the power is used to transmit the same codeword as the relay , i.e. , @xmath61 , so that this portion of the source s signal and the relay s signal add coherently at the destination twice : @xmath55 in block @xmath50 and @xmath62 in block @xmath51 .
note that these two codewords are independently generated . ] ; and ( ii ) the remaining portion of @xmath63 of the power is dedicated to a different codeword that is to be decoded by the destination , i.e. , @xmath64 . in brief , the source transmits @xmath65 $ ] in block @xmath50 .
the above encoding scheme requires that the relay decodes @xmath49 at the end of block @xmath50 ; the relay can reliably do so if @xmath66 .
the decoding of the message @xmath67 at the destination is performed over blocks @xmath50 and @xmath51 .
assume that the destination has decoded messages @xmath68 .
using the last @xmath29 channel uses of block @xmath50 , the destination can decode @xmath69 if @xmath70 . using the first @xmath28 channel uses of block @xmath50 and the last @xmath29 channel uses of block @xmath51
, the destination can decode @xmath49 if @xmath71 ^ 2 } { ( 1-\beta ) \lambda_{02 } + 1 } \right)$ ] . here
, @xmath72 ^ 2 $ ] is power of the coherently combined signals from the relay and the source , @xmath73 is the power of signals carrying @xmath74 which appears as noise when decoding @xmath49 , and the channel noise power @xmath75 $ ] is one .
combining these rate constraints , pdf @xcite achieves rates up to @xmath76 setting @xmath77 ( which means @xmath78 ; i.e. , the relay decodes all source messages ) , we have cdf @xcite , which achieves rates up to @xmath79 ^ 2 ) \big\}. \label{eq : cdf - rate}\end{aligned}\ ] ] clearly , @xmath80 . furthermore , setting @xmath81 and @xmath82 ( which means @xmath83 ; i.e. , the relay decodes nothing ) , we have direct transmission .
when @xmath84 ( the source - destination link is better than the source - relay link ) , @xmath85 , which is attained only at @xmath86 , i.e. , with direct transmission .
when @xmath87 ( the source - destination link is as good as the source - relay link ) , we again have @xmath85 , which is attained by @xmath82 and any @xmath88 $ ] .
although the relay is used when @xmath89 , pdf still achieves the same rate as that of direct transmission , @xmath90 . for the rest of this paper , we will consider the non - trivial case of @xmath91 ( when the source - relay link is better than the source - destination link ) .
we will show later that @xmath1 can be strictly higher than @xmath0 .
furthermore , it is easy to see that , as all @xmath92 , both pdf and cdf strictly outperform direct transmission .
we define the gap between the pdf and the cdf rates by @xmath93 and the normalized ( to @xmath0 ) gap by @xmath94
first , we have the following lemma : [ lemma : cdf ] let @xmath95 be the @xmath54 that attains @xmath0 .
if @xmath91 , it follows that @xmath96 ^ 2)$ ] .
let @xmath97 and @xmath98 ^ 2)$ ] .
the function @xmath99 is a continuous and increasing in @xmath54 with @xmath100 and @xmath101 ; @xmath102 is a continuous and decreasing function of @xmath54 with @xmath103 and @xmath104 .
so the solution to @xmath105 is at the point where @xmath99 and @xmath102 intersect . from the above lemma
, we can show that [ lemma : cdf - rate ] if @xmath91 , then @xmath106 ^ 2 ) \log ( 1 + \lambda_{01 } ) } { \log ( 1 + [ \sqrt{\lambda_{02 } } + \sqrt{\lambda_{12}}]^2 ) + \log ( 1 + \lambda_{01 } ) - \log ( 1 + \lambda_{02 } ) } .\end{gathered}\ ] ] substituting @xmath95 from lemma [ lemma : cdf ] into gives lemma [ lemma : cdf - rate ] . now , by selecting the @xmath107 $ ] that maximizes the two terms on the right - hand side of ( i.e. , @xmath108 and @xmath77 respectively ) , we can upper bound @xmath1 as @xmath109 ^ 2 ) \big\}\nonumber \\ & \triangleq r_\text{pdf}^\text{ub } , \label{eq : pdf - ub } \end{aligned}\ ] ] where @xmath110 is an upper bound to @xmath1 . using the same arguments in the proof of lemma [ lemma : cdf ] ,
we obtain the following : [ lemma : pdf - ub - a ] let @xmath111 be the @xmath54 that attains @xmath110 . if @xmath91 then @xmath112 ^ 2 ) $ ] .
hence , the upper bound to @xmath1 is given as follows : [ lemma : pdf - ub - b ] if @xmath91 , then @xmath113 ^ 2 ) \log ( 1 + \lambda_{01 } ) } { \log ( 1 + [ \sqrt{\lambda_{02 } } + \sqrt{\lambda_{12}}]^2 ) + \log ( 1 + \lambda_{01 } ) - 2 \log ( 1 + \lambda_{02 } ) } \nonumber \\ & - \frac { \frac{1}{2 } [ \log ( 1 + \lambda_{02})]^2 } { \log ( 1 + [ \sqrt{\lambda_{02 } } + \sqrt{\lambda_{12}}]^2 ) + \log ( 1 + \lambda_{01 } ) - 2 \log ( 1 + \lambda_{02 } ) } .\end{aligned}\ ] ] substituting @xmath111 in lemma [ lemma : pdf - ub - a ] into gives lemma [ lemma : pdf - ub - b ] . we now derive the following upper bound on the normalized gap between the pdf and the cdf rates : @xmath114 first , we define the following : @xmath115 , @xmath116 , @xmath117 , @xmath118 , and @xmath119 ^ 2\right)$ ] . note that while @xmath120 , @xmath121 , @xmath122 , and @xmath14 are completely determined by the channel gains , the source s transmit power , and the receiver noise levels at the relay and the destination , @xmath123 is additionally determined by the relay s transmit power ( in addition to the channel gains , the source s transmit power , and the noise levels ) . with the above definitions , we have the normalized gap upper bounded by @xmath124 to prove , we first show the following lemma : [ lemma : gap ] if @xmath91 , then @xmath125 } .
\label{eq : lemma}\ ] ] we first write @xmath126 where @xmath127 . note that for any fixed channel gains , @xmath128 , @xmath129 , and @xmath130 , the variables @xmath14 , @xmath121 , and @xmath122 are fixed , but the variable @xmath123 can take any value by choosing an appropriate @xmath131 subject to the constraint @xmath132 ^ 2\right ) > \log ( 1 + \lambda_{02 } ) \triangleq u $ ] .
this means @xmath123 can be chosen independent of @xmath14 , @xmath121 , and @xmath122 .
now , for any @xmath14 , @xmath121 , and @xmath122 , we maximize @xmath133 with respect to @xmath123 .
note that as @xmath91 , we must have @xmath134 .
so , for the region of interest , i.e. , where @xmath135 , the function @xmath133 is continuously differentiable . by differentiating @xmath133 with respect to @xmath123 ,
we get @xmath136 ^ 2}$ ] .
the denominator is always positive , and the numerator is a quadratic function with roots @xmath137 and @xmath138 . note that @xmath139 since @xmath140 .
this means @xmath141 so , the maximum of @xmath133 for @xmath142 occurs at @xmath138 .
thus , @xmath143[u + \sqrt{u(u+v)}+v]}.\ ] ] substituting @xmath144 and @xmath145 into the above equation , we get . with the above lemma , we now prove .
first , we define @xmath146 , where @xmath147 because @xmath148 .
the right - hand side of can be written as @xmath149 where @xmath150 $ ] .
this means @xmath151 $ ] .
note that @xmath152 is continuously differentiable for all @xmath147 , and its first derivative is @xmath153 so , @xmath154 .
this gives .
combining and the trivial lower bound @xmath80 , we have theorem [ theorem : main ] .
@xmath155 suppose that @xmath157 .
the optimal @xmath54 for the max - min operation in is close to zero .
hence , @xmath158 ^2)$ ] . for , the optimal @xmath54 is close to zero and the optimal @xmath60 is close to one .
hence , @xmath159 ^2)$ ] .
so the trivial lower bound @xmath80 is ( almost ) tight when @xmath157 .
numerical results show that the normalized gap @xmath156 approaches the upper bound of @xmath160 ( in theorem [ theorem : main ] ) as @xmath161 increases ( for selected @xmath162 and @xmath163 for each @xmath161 ) .
for example , setting @xmath164 , @xmath165 , and @xmath166 , the normalized gap is @xmath167 .
in this section , we consider the free - space path loss model where the channel gain is given by @xmath168 , where @xmath169 is the distance between the transmitter and the receiver , and so @xmath170 .
@xmath171 \big [ \log(1 + [ \sqrt{\lambda_{02 } } + \sqrt{\lambda_{12}}]^2 ) - \log(1+\lambda_{02 } ) \big ] \log(1+\lambda_{02 } ) } { 2\big[\log ( 1 + [ \sqrt{\lambda_{02 } } + \sqrt{\lambda_{12}}]^2 ) + \log ( 1 + \lambda_{01 } ) - 2 \log ( 1 + \lambda_{02})\big ] \big[\log ( 1 + [ \sqrt{\lambda_{02 } } + \sqrt{\lambda_{12}}]^2 ) + \log ( 1 + \lambda_{01 } ) - \log ( 1 + \lambda_{02})\big ] } } \triangleq g^\text{ub}. \label{eq : g-2 } % & = \frac{(w - u)(q - u)u}{2(q+v)(q+w - u ) } , \end{aligned}\ ] ] we fix coordinates of the source at @xmath172 and the destination at @xmath173 , and compute the actual @xmath156 for different relay coordinates @xmath174 .
we assume that @xmath175 , and @xmath176 .
we are interested in networks where @xmath91 , which is equivalent to restricting the relay position to @xmath177 . the result
is shown in fig .
[ fig:3d ] . while an upper bound derived for @xmath156 in the previous section
is @xmath178 for all channel parameters , for this example of equal power and equal noise , the actual gap is only 6.45% or less .
, for varying relay positions ( only positions at @xmath179 are shown due to symmetry ) : we have @xmath180 , width=340 ] now , we present four cases in which @xmath181 or @xmath156 tends to zero , meaning that cdf performs as well as pdf . in the following subsections , unless otherwise stated , the transmit power @xmath128 and @xmath131 , the inter - node distances @xmath182 , @xmath183 , @xmath184 , and the receiver noise @xmath129 , @xmath130 are all positive and finite .
first , we find an upper bound on the gap between the pdf and the cdf rates , given by @xmath185 in . when the relay is close to the source , i.e. , @xmath186 , we have @xmath187 .
since @xmath163 and @xmath161 are finite , we see that both @xmath188 and @xmath189 as @xmath186 .
similarly , when the relay is close to the destination , i.e. , @xmath190 , we have @xmath191 . since @xmath162 and @xmath163 are finite , we also see that @xmath188 and @xmath189 as @xmath190 .
now , we consider fixed node positions , and increase the transmit power of the nodes .
we let @xmath192 , @xmath193 , @xmath194 , @xmath195 for some constants @xmath196 and @xmath197 , and investigate the behavior of @xmath181 and @xmath156 when @xmath198 increases . we further define the following , which are independent of @xmath198 : @xmath199 and @xmath200 ^ 2}{\lambda_{02 } } = \frac{\left[\sqrt{d_{02}^{-2}k_0}+ \sqrt{d_{12}^{-2}k_1}\right]^2}{d_{02}^{-2}k_0}$ ] . for high snr
, we use the approximation @xmath201 , and obtain @xmath202}\nonumber \\ & \approx % \frac { \log c_1 \log c_2 } { 2\log ( c_1c_2 ) } = \frac{1}{2 } \left ( \frac{1}{\log c_1 } + \frac{1}{\log c_2 } \right)^{-1}. \label{eq : high - snr}\end{aligned}\ ] ] so , in the high snr regime where the pdf and the cdf rates are high , the gap between the rates is upper bounded by a constant . this constant , when normalized to the pdf or the cdf rates , approaches zero as the snr increases , i.e. , @xmath203 .
[ remark : large - g ] from , we see that @xmath204 increases as @xmath205 and @xmath206 increase .
this means with a fixed @xmath163 , for sufficiently large @xmath162 and @xmath161 , and a much larger @xmath198 , @xmath185 can be made arbitrarily large .
finally , we investigate the case where the snr tends to zero . we follow the settings ( inter - node distances , transmit power , and receiver noise ) as in the high - snr case above , but with @xmath207 . for low snr , as @xmath207 , both @xmath208 .
so , @xmath209 . using the approximation @xmath210
, we further see that the normalized gap is @xmath211 where @xmath212 , + @xmath213 , and + @xmath214 . as @xmath91
, we have @xmath215 .
so @xmath216 , and it follows that @xmath217 .
consequently , @xmath218 if any of the following is true : ( i ) @xmath219 , i.e. , when the relay is close to the source , @xmath186 ; ( ii ) @xmath220 , i.e. , when the relay is close to the destination , @xmath190 ; or ( iii ) @xmath221 , i.e. , when the relay and the destination are about the same distance from the source , @xmath222 [ by the triangle inequality , we have @xmath223 .
in this letter , we have shown that there is little to be gained by doing partial decoding rather than complete decoding at the relay for the half - duplex awgn single - relay channel .
we have analytically shown that using partial - decode - forward ( pdf ) can increase the rate over complete - decode - forward ( cdf ) by at most one - eighth . we have also identified four scenarios in which the cdf rates asymptotically approach the pdf rates . while a multiplicative - type bound on the gain of pdf over cdf has been obtained in this paper , we note that the absolute gain _ might _ be unbounded in the high snr regime ( we have shown in remark [ remark : large - g ] that our upper bound on the absolute gain is unbounded ) .
hence , one may attempt pdf in the high - snr regime when the absolute gain in transmission rate ( as opposed to the percentage gain ) outweighs the system complexity . the results in this letter
provide a strong motivation to extend the comparison between these two strategies to the multiple - relay channel to investigate if partial - decode - forward is beneficial when there are more relays in the network , in which partial - decode - forward is much more computationally complex than complete - decode - forward .
l. ong , w. wang , and m. motani , `` achievable rates and optimal schedules for half duplex multiple - relay networks , '' in _ proc .
46th allerton conf .
control comput .
( allerton conf . ) _ , monticello , usa , sept .
2326 2008 , pp . 13171324 . | this paper compares the partial - decode - forward and the complete - decode - forward coding strategies for the half - duplex gaussian single - relay channel .
we analytically show that the rate achievable by partial - decode - forward outperforms that of the more straightforward complete - decode - forward by at most 12.5% .
furthermore , in the following asymptotic cases , the gap between the partial - decode - forward and the complete - decode - forward rates diminishes : ( i ) when the relay is close to the source , ( ii ) when the relay is close to the destination , and ( iii ) when the snr is low .
in addition , when the snr increases , this gap , when normalized to the complete - decode - forward rate , also diminishes .
consequently , significant performance improvements are not achieved by optimizing the fraction of data the relay should decode and forward , over simply decoding the entire source message . achievable rate , decode - forward , half duplex , partial decode - forward , relay channel . |
the inclusive radiative decays of the @xmath1 meson play a central role in the search for new physics . while the total rate of @xmath5 is sensitive to new physics in flavor - changing transitions ,
its photon spectrum is almost completely determined by standard model physics and can be employed to extract information on the @xmath1 meson structure and on the @xmath6 quark mass , see for instance @xcite .
the latter are useful in a number of @xmath1 physics applications , like the determination of @xmath7 and of @xmath8 , which are both important inputs for the determination of the ckm unitarity triangle and the study of cp violation in the standard model . after integrating out the heavy degrees of freedom at the electroweak scale and evolving the resulting weak hamiltonian to the @xmath6 quark mass scale ,
inclusive radiative and semileptonic @xmath1 decays are well described by an operator product expansion ( ope ) in inverse powers of the @xmath6 quark mass . in this way one can factorize the long distance dynamics into the matrix elements of a few local operators @xcite . since the wilson coefficients of these operators are perturbative , and the matrix elements of the local operators parameterize the non - perturbative physics , the radiative and semileptonic rates and the moments of their distributions are double series in @xmath9 and @xmath10 , with @xmath11 being the qcd scale .
the lowest order of this expansion corresponds to the decay of a free @xmath6 quark and linear @xmath12 corrections are absent .
the relevant parameters are therefore the heavy quark mass @xmath3 ( and possibly @xmath13 for the charmed decays ) , @xmath9 , and the matrix elements of the local operators : @xmath14 and @xmath15 at @xmath16 , @xmath17 and @xmath18 at @xmath19 . in the case of @xmath20 transitions , the first and second central moments of the photon spectrum , @xmath21 and @xmath22 ,
are proportional to the @xmath6 quark mass and to @xmath14 , respectively , up to subleading corrections in the @xmath23 expansion , and contribute in an important way to the global fits for the extraction of @xmath8 , @xmath3 and the other ope parameters ( see @xcite for recent results ) .
the precision of @xmath5 calculations in this framework is known to be limited in two ways : @xmath24 the dominant contribution to the radiative @xmath6 decay is associated to the electromagnetic dipole operator @xmath25 , with additional operators appearing at @xmath0 .
it turns out that a local ope can be written down for the @xmath26 contribution only , while the other operators are expected to induce unknown @xmath27 contributions , see e.g. @xcite ; @xmath28 even for the dominant @xmath26 component , the experimental cuts employed to isolate the signal introduce a sensitivity to the fermi motion of the @xmath6 quark inside the @xmath1 meson and tend to disrupt the ope .
one can still resum the higher order terms into a non - local distribution function @xcite and since the lowest integer moments of this function are known , one can parameterize it assuming different functional forms @xcite , although alternative approaches are also possible @xcite .
the perturbative corrections to the leading @xmath26 contribution are presently known to complete next - to - next - to - leading order ( nnlo ) , i.e. @xmath29 , while the nnlo contributions involving other operators , quite important for the rate , are not yet complete @xcite , see @xcite for recent updates . as for the non - perturbative corrections to the @xmath26 contribution ,
they are known through @xmath19 @xcite . among the non - perturbative corrections associated to operators other than @xmath26 ,
only one class is known : they are of @xmath30 and @xmath31 and numerically small @xcite .
overall , the known power corrections modify the total @xmath5 rate by just @xmath32 , but they are essential in the calculation of the moments .
in this paper we present the first calculation of the @xmath0 corrections to the @xmath33 terms in @xmath5 .
the @xmath0 perturbative corrections to the @xmath14 coefficient in the total rate are fixed by lorentz invariance @xcite .
their contribution to the photon moments has been computed in the context of a multiscale ope @xcite which applies however to the end - point region only .
we have computed the relevant wilson coefficients at @xmath0 by expanding off - shell amputated green functions around the @xmath6 quark mass shell , and by matching them onto local operators in heavy quark effective theory ( hqet ) .
our results allow for an improved analysis of the radiative moments .
in particular , the inclusion of the @xmath0 perturbative corrections to the variance of the spectrum , permits the extraction of @xmath14 at next - to - leading - order ( nlo ) .
the outline of this paper is as follows . in section 2
we recalculate the wilson coefficients of the matrix elements of dimension five operators , @xmath14 and @xmath15 , at leading order in @xmath34 and introduce our notation .
section 3 is devoted to the calculation of these wilson coefficients at @xmath35 .
it also contains our final results and a first numerical estimate of their importance . in section 4
we summarize and conclude .
at low energy the @xmath36 transition is governed by the effective lagrangian @xmath37 which includes , apart from the qed and qcd interactions of the light quarks @xmath38 and @xmath6 , the renormalization scale dependent effective wilson coefficient @xmath39 and the corresponding electromagnetic operator @xmath26 .
the latter mediates the @xmath36 transition and is given by @xmath40 where @xmath41 is the running @xmath6 quark mass in the @xmath42-scheme , @xmath43/2 $ ] and @xmath44 .
the ellipses in equation ( [ eq::lagrangian ] ) denote contributions from operators that are not relevant for our calculation . using this effective lagrangian
we can derive that part of the differential decay rate that is induced by the self - interference of the electromagnetic dipole operator , @xmath45 here , @xmath46 is the mass of the @xmath1 meson , @xmath47 the momentum of the photon , @xmath48 the hadronic tensor , @xmath49 & \qquad\times\,\langle\bar b(p_b)|\bar b(0)\sigma_{\mu\nu}p_ls(0)|x_s(p_x)\rangle \langle x_s(p_x)|\bar s(0)\sigma_{\alpha\beta}p_rb(0)|\bar b(p_b)\rangle\,,\end{aligned}\ ] ] and @xmath50 the photon tensor , @xmath51 & = q^\alpha q^\nu g^{\mu\beta}-q^\alpha q^\mu g^{\nu\beta } + q^\mu q^\beta g^{\nu\alpha}-q^\beta q^\nu g^{\mu\alpha}\,.\end{aligned}\ ] ] the hadronic tensor @xmath48 itself can be rewritten as @xmath52 which is useful because the time - ordered product can be expanded into a series of local operators that are suppressed by powers of the @xmath6 quark mass . hence we can write @xmath53 where @xmath54 is an operator of dimension @xmath55 that contains @xmath56 derivatives , and @xmath57 is the corresponding wilson coefficient that can be determined in perturbation theory . in the reminder of this section we will review the calculation of @xmath58 for @xmath59 .
ccc ( 0,0)(0,0 ) ( -100,-55)(-40,0 ) ( 100,-55)(40,0 ) ( -40,0)(40,0 ) ( -100,55)(-40,0)25 ( 100,55)(40,0)25 ( -45,-32)[br]@xmath60 ( -18,-35)[br]@xmath61 ( -46,24)[br]@xmath62 ( -42,28)[br]@xmath47 ( 57,22)[br]@xmath63 ( 51,31)[br]@xmath47 ( 72,-25)[br]@xmath64 ( 100,-18)[br]@xmath61 ( -78,-39)[br]@xmath6 ( 78,-39)[bl]@xmath6 ( 0,-8)[bl]@xmath65 ( -35,0)[c ] ' '' '' ( -43,-5)[br]@xmath26 ( 33,0)[c ] ' '' '' ( 43,-5)[bl]@xmath26 & & ( 0,0)(0,0 ) ( -100,-55)(-40,0 ) ( 100,-55)(40,0 ) ( -40,0)(40,0 ) ( -100,55)(-40,0)25 ( 100,55)(40,0)25 ( 0,0)(0,55)35 ( -45,-32)[br]@xmath60 ( -18,-35)[br]@xmath61 ( -46,24)[br]@xmath62 ( -42,28)[br]@xmath47 ( 57,22)[br]@xmath63 ( 51,31)[br]@xmath47 ( 12,18)[br]@xmath66 ( 17,19)[br]@xmath67 ( 72,-25)[br]@xmath64 ( 114,-18)[br]@xmath68 ( -78,-39)[br]@xmath6 ( 78,-39)[bl]@xmath6 ( 0,-8)[bl]@xmath65 ( -35,0)[c ] ' '' '' ( -43,-5)[br]@xmath26 ( 33,0)[c ] ' '' '' ( 43,-5)[bl]@xmath26 + the two feynman diagrams contributing to these wilson coefficients are depicted in fig . [ fig::tree - level ] . the momentum of the incoming @xmath6 quark is @xmath61 , where @xmath69 is the velocity of the @xmath1 meson , and @xmath70 is a residual momentum that accounts for the interaction of the almost on - shell @xmath6 quark with the light degrees of freedom in the @xmath1 meson .
the components of this residual momentum are of @xmath71 , i.e. much smaller than the @xmath6 quark mass .
the same is true for the components of the momentum @xmath67 of the radiated soft gluon that is present in the feynman diagram on the right - hand side . to achieve the anticipated ope ( [ eq::ope ] )
we perform a taylor expansion of the amputated green functions corresponding to the two tree - level diagrams up to @xmath72 for the left diagram and up to @xmath73 and @xmath74 for the right diagram .
this is equivalent to an expansion in inverse powers of @xmath3 , or more precisely , to an expansion in @xmath75 , where @xmath76 . in the algebraic manipulations we refrain from using any on - shell relations for the @xmath6 quark , that is we do not impose any restriction on the residual momenta .
the only on - shell condition we apply is that of the @xmath1 meson , namely @xmath77@xmath78@xmath79 . after extracting the imaginary part of the amputated green functions we replace the residual momenta @xmath70 and @xmath67 with derivatives acting on the @xmath6 and the gluon fields , interpreting the latter as parts of a covariant derivative .
we end up with expressions for the two tree - level diagrams in terms of local operators of the form @xmath80 , where @xmath81 is a matrix - valued function of @xmath82 , @xmath83 and @xmath84 in spinor space . as far as the @xmath65-quark is concerned
, we will use @xmath85 . at this point
it is convenient to introduce hqet @xcite which incorporates the expansion in @xmath86 in a natural manner .
its lagrangian is given by @xcite @xmath87 where @xmath88 and @xcite @xmath89\frac{\alpha_s}{4\pi}+ \dots\,.\ ] ] the numerical values of the color factors are @xmath90 and @xmath91 .
the relation between the @xmath6 quark fields in qcd and in hqet reads @xmath92 at tree - level as well as at @xmath0 .
using ( [ eq::hqetspinor ] ) , together with @xmath93 , in all local operators of dimension 4 and 5 , we end up with the following set of operators , @xmath94 o_1^{\mu } & = \bar b_vid^\mu b_v , & o_3^{\mu\nu } & = \bar b_v\frac{g_s}{2}g^{a\mu}_{\;\;\;\;\alpha}\sigma^{\alpha\nu}t^ab_v\ , , & & & \end{aligned}\ ] ] where @xmath95 $ ] .
we also find operators that include a @xmath96 .
however , we will eventually calculate matrix elements of operators between @xmath1 meson states . from parity considerations it follows then that only the operators given in ( [ eq::operators ] ) give non - vanishing contributions .
hence it is not necessary to calculate the wilson coefficients of those operators that include a @xmath96 .
we should also mention that up to this point of the calculation we did not use the equation of motion for the @xmath97 field that follows from the hqet lagrangian .
for the wilson coefficients of the operators given in ( [ eq::operators ] ) we find , in @xmath98 space - time dimensions , @xmath99 c_{1\mu}^{(0 ) } & = -\frac{1}{2}\,(1-\epsilon)\big[\left(2 v_\mu-\hat q_\mu\right)\delta(1-z ) + \left(v_\mu-\hat q_\mu\right)\delta^\prime(1-z)\big],\nonumber\\[1 mm ] c_{2\mu\nu}^{(0 ) } & = -\frac{1}{4}\,(1-\epsilon)\big [ 2\left(g_{\mu\nu}+v_\mu\hat q_\nu\right)\delta(1-z ) + \left(g_{\mu\nu}+6v_\mu v_\nu-6v_\mu\hat q_\nu\right)\delta^\prime(1-z)\nonumber\\[1 mm ] & \hspace{27mm}+ 2\left(v_\mu v_\nu-2v_\mu\hat q_\nu+\hat q_\mu\hat q_\nu\right ) \delta^{\prime\prime}(1-z)\big],\nonumber\\[1 mm ] c_{3\mu\nu}^{(0 ) } & = -\big[2v_\mu\hat q_\nu-\frac{1+\epsilon}2 g_{\mu\nu}\big]\delta(1-z ) + \big[(3+\epsilon)\hat q_\mu\hat q_\nu-2v_\mu\hat q_\nu + \frac{1+\epsilon}{4}g_{\mu\nu}\big]\delta^\prime(1-z)\,,\end{aligned}\ ] ] where @xmath100 and @xmath101 . in the last step
we have to calculate the forward matrix elements of the four operators given in ( [ eq::operators ] ) between @xmath1 meson states .
since we expressed the leading operator @xmath102 in terms of the @xmath6 quark fields of qcd we know its matrix element exactly , @xmath103 the evaluation of the matrix elements of the other operators involves the equation of motion of the effective theory , and leads to two additional matrix elements @xcite , @xmath104 the velocity dependent @xmath1 meson states used here are related to the momentum dependent ones introduced in ( [ eq::hadronictensor ] ) by @xmath105 . the power correction to
this relation is irrelevant for our calculation . while @xmath106 and @xmath107 are defined in the asymptotic hqet regime , in practical applications one deals with @xmath108 and @xmath109 , defined in terms of the @xmath3-finite qcd states appearing in ( [ eq::hadronictensor ] ) .
now we are in a position to calculate the contribution to the differential decay rate due to the electromagnetic dipole operator only .
we obtain @xmath110\,,\end{aligned}\ ] ] where @xmath111 and @xmath112 c_{\lambda_2}^{(0 ) } & = -9\,\delta(1-z)-3\,\delta^\prime(1-z)\,.\end{aligned}\ ] ] this is in agreement with the well - known results given in @xcite .
the calculation of the wilson coefficients @xmath113 , @xmath114 and @xmath115 is the subject of the next section .
ccc ( 0,0)(0,0 ) ( -100,-55)(-40,0 ) ( 100,-55)(40,0 ) ( -40,0)(40,0 ) ( -100,55)(-40,0)25 ( 100,55)(40,0)25 ( -40,0)(40,-138,0)56 ( -78,-39)[br]@xmath6 ( 78,-39)[bl]@xmath6 ( 12,8)[tl]@xmath65 ( -35,0)[c ] ' '' '' ( -43,-5)[br]@xmath26 ( 33,0)[c ] ' '' '' ( 43,-5)[bl]@xmath26 & & ( 0,0)(0,0 ) ( -100,-55)(-40,0 ) ( 100,-55)(40,0 ) ( -40,0)(40,0 ) ( -100,55)(-40,0)25 ( 100,55)(40,0)25 ( 40,0)(40,-180,-42)56 ( -78,-39)[br]@xmath6 ( 78,-39)[bl]@xmath6 ( -12,8)[tr]@xmath65 ( -35,0)[c ] ' '' '' ( -43,-5)[br]@xmath26 ( 33,0)[c ] ' '' '' ( 43,-5)[bl]@xmath26 + ( 0,0)(0,0 ) ( -100,-55)(-40,0 ) ( 100,-55)(40,0 ) ( -40,0)(40,0 ) ( -100,55)(-40,0)25 ( 100,55)(40,0)25 ( 0,50)(110,-132,-48)59 ( -78,-39)[br]@xmath6 ( 78,-39)[bl]@xmath6 ( 0,8)[tl]@xmath65 ( -35,0)[c ] ' '' '' ( -43,-5)[br]@xmath26 ( 33,0)[c ] ' '' '' ( 43,-5)[bl]@xmath26 & & ( 0,0)(0,0 ) ( -100,-55)(-40,0 ) ( 100,-55)(40,0 ) ( -40,0)(40,0 ) ( -100,55)(-40,0)25 ( 100,55)(40,0)25 ( 0,0)(20,-180,0)54 ( -78,-39)[br]@xmath6 ( 78,-39)[bl]@xmath6 ( 0,8)[tl]@xmath65 ( -35,0)[c ] ' '' '' ( -43,-5)[br]@xmath26 ( 33,0)[c ] ' '' '' ( 43,-5)[bl]@xmath26 + in order to determine the wilson coefficients @xmath113 , @xmath114 and @xmath115 we calculate the amputated green functions corresponding to the feynman diagrams shown in fig . [ fig::oneloop ] .
the momentum assignments of the external lines are exactly the same as in the tree - level calculation performed in the last section . for the green functions that contain a radiated soft gluon we apply the background - field formalism @xcite .
furthermore , we work in the general @xmath116-gauge for the gluon propagator . again
we perform a taylor expansion up to @xmath72 for the diagrams without a radiated soft gluon and up to @xmath73 and @xmath74 for the diagrams with a radiated soft gluon , and refrain from using any on - shell relations for the @xmath6 quark .
we apply integration - by - parts techniques @xcite to reduce all integrals to a few so - called master integrals and solve the latter analytically .
ultraviolet as well as infrared divergences are handled by dimensional regularization .
the ultraviolet divergences can be removed by appropriate renormalization . for the self - mixing of the operator @xmath26 we use the @xmath42-scheme , @xmath117 and for the field renormalization constant of the @xmath6 quark
we apply the on - shell scheme , @xmath118 as far as the renormalization of the background gluon field @xmath119 is concerned , we only have to remember that @xmath120 does not get renormalized . finally , we extract the imaginary part of the amputated green functions , replace the residual momenta @xmath70 and @xmath67 with covariant derivatives @xmath121 and the @xmath6 quark spinors with @xmath97 spinors via ( [ eq::hqetspinor ] ) , and calculate the forward matrix elements of all operators between @xmath1 meson states .
our result is then a linear combination of the three matrix elements introduced in ( [ eq::lo ] ) and ( [ eq::lambdai ] ) with coefficients that are ultraviolet finite but still contain infrared divergences , @xmath122 we have also made explicit the dependence on the gauge parameter @xmath123 .
the infrared divergences are removed in the matching procedure .
we now turn our attention to the right - hand side of the matching equation ( [ eq::ope ] ) . at the one - loop level it schematically looks like @xmath124\,,\ ] ] where @xmath125 and @xmath126 denote the tree - level and one - loop matrix elements of the operator @xmath127 between @xmath1 meson states , respectively , and @xmath128 collects the @xmath129-factors to render this expression ultraviolet finite . in the case at hand
we only have to consider the one - loop matrix elements of the operators given in ( [ eq::operators ] ) since these are the only ones that have non - vanishing wilson coefficients at the tree - level .
the same holds for the tree - level matrix elements that are multiplied by @xmath129-factors .
because of the taylor expansion in the momenta @xmath70 and @xmath67 the one - loop matrix elements of the operators @xmath130 and @xmath131 vanish in dimensional regularization and we only need to compute the one - loop matrix elements of @xmath102 .
the operators that include a @xmath96 can again be discarded due to parity considerations . for the renormalization constants , we use the on - shell scheme for the @xmath6 and @xmath97 spinors , and the @xmath42 scheme for the operator renormalization , @xmath132^{\rm bare } & = z_b^{\rm os}\,c_{b\mu}o_b^\mu\ , , & \left[c_{2\mu\nu}o_2^{\mu\nu}\right]^{\rm bare } &
= z_{b_v}^{\rm os}z_{\rm kin}^{{\ensuremath{\overline{\text{ms}}}},\mu\nu\alpha\beta}\,c_{2\mu\nu}o_{2\alpha\beta}\ , , & & \nonumber\\[1 mm ] \left[c_{1\mu}o_1^\mu\right]^{\rm bare } & = z_{b_v}^{\rm os}\,c_{1\mu}o_1^\mu\ , , & \left[c_{3\mu\nu}o_3^{\mu\nu}\right]^{\rm bare } & = z_{b_v}^{\rm os}z_{\rm chromo}^{{\ensuremath{\overline{\text{ms}}}},\mu\nu\alpha\beta}\,c_{3\mu\nu}o_{3\alpha\beta}\,.\end{aligned}\ ] ] a simple one - loop calculation yields @xmath133
z_{\rm chromo}^{{\ensuremath{\overline{\text{ms}}}},\mu\nu\alpha\beta } & = \frac{c_a}{\epsilon } \left(g^{\mu\alpha}-v^\mu v^\alpha\right)g^{\nu\beta}\,\frac{{\alpha_s}}{4\pi}+\dots\,.\end{aligned}\ ] ] the feynman gauge is obtained by setting @xmath134 . the on - shell renormalization constant @xmath135 can be set equal to 1 since the @xmath0-corrections to the selfenergy of the @xmath97 field depends on no other scale than the renormalization scale @xmath136 .
requiring equality between ( [ eq::lhs ] ) and ( [ eq::rhs ] ) , and solving for @xmath137 , we obtain infrared finite and gauge independent expressions . writing the coefficients as follows ,
@xmath138 c_{\lambda_2}^{(1 ) } & = c_f\left(c_{\lambda_2}^{(1,\mbox{\tiny f } ) } + \delta c_{\lambda_2}^{(1,\mbox{\tiny f})}\right)+c_a\left(c_{\lambda_2}^{(1,\mbox{\tiny a } ) } + \delta c_{\lambda_2}^{(1,\mbox{\tiny a})}\right)\,,\end{aligned}\ ] ] our final results are given by @xmath139_+\nonumber\\[1 mm ] & \hspace{4.5mm}-\!4\left[\frac{\ln(1-z)}{1-z}\right]_++7+z-2z^2 - 2 ( 1+z)\ln(1-z ) - 4\,c_0^{(0)}\ln\frac{\mu}{m_b}\,,\\[3 mm ] c_{\lambda_1}^{(1,\mbox{\tiny f } ) } & = -\frac{2}{3}\left(15 + 2\pi ^2\right)\delta(1-z ) - \frac{4}{3}\left(3-\pi ^2\right)\delta^\prime(1-z)\nonumber\\[1 mm ] & \hspace{4.5mm}-\!\left(\frac{7}{6}-\frac{4\pi^2}{9}\right)\delta^{\prime\prime}(1-z ) + \frac{2}{3}\left[\frac{1}{(1-z)^3}\right]_+-\frac{5}{3}\left[\frac{1}{(1-z)^2}\right]_+\nonumber\\[1 mm ] & \hspace{4.5mm}-\!9\left[\frac{1}{1-z}\right]_++\frac{8}{3}\left[\frac{\ln(1-z)}{(1-z)^3}\right]_+ - 4\left[\frac{\ln(1-z)}{(1-z)^2}\right]_+\nonumber\\[1 mm ] & \hspace{4.5mm}-\!4\left[\frac{\ln(1-z)}{1-z}\right]_++10+\frac{7}{3}\,z-\frac{2}{3}\,(4 + 3z)\ln(1-z ) - 4\,c_{\lambda_1}^{(0)}\ln\frac{\mu}{m_b}\,,\\[3 mm ] c_{\lambda_2}^{(1,\mbox{\tiny f } ) } & = \left(41+\frac{20\pi^2}{3}\right)\delta(1-z ) + 4\left(1+\pi ^2\right)\delta^\prime(1-z)\nonumber\\[1 mm ] & \hspace{4.5mm}-\!9\left[\frac{1}{(1-z)^2}\right]_++15\left[\frac{1}{1-z}\right]_+ - 12\left[\frac{\ln(1-z)}{(1-z)^2}\right]_+\nonumber\\[1 mm ] & \hspace{4.5mm}+\!20\left[\frac{\ln(1-z)}{1-z}\right]_+-6 + 55 z+2(22 - 17 z)\ln(1-z ) - 4\,c_{\lambda_2}^{(0)}\ln\frac{\mu}{m_b}\,,\\[3 mm ] c_{\lambda_2}^{(1,\mbox{\tiny a } ) } & = 2\left(7-\frac{8\pi^2}{3}\right)\delta(1-z)+4\,\delta^\prime(1-z ) - 4\left[\frac{1}{(1-z)^2}\right]_+\nonumber\\[1 mm ] & \hspace{4.5mm}+\!2\left[\frac{1}{1-z}\right]_+ - 16\left[\frac{\ln(1-z)}{1-z}\right]_++2 - 6z+4 ( 1 + 3 z)\ln(1-z)\,.\end{aligned}\ ] ] the contributions @xmath140 and @xmath141 are a consequence of the application of the equation of motion of the effective theory in the evaluation of the matrix elements of the operators given in ( [ eq::operators ] ) .
their explicit expressions read @xmath142 we note that the coefficient @xmath143 agrees with the well - known result of @xcite .
the coefficients @xmath114 and @xmath115 have been calculated here for the first time .
for the first one we confirm the expected relation @xmath144 not surprisingly , @xmath145 diverge more strongly than @xmath146 at the endpoint .
the plus - distributions introduced above follow the prescription @xmath147_+f(z)\nonumber\\[1 mm ] & \qquad = \int_0 ^ 1\!dz\,\frac{\ln^n(1-z)}{(1-z)^m}\left\{f(z ) - \sum_{p=0}^{m-1}(-1)^p\frac{(1-z)^p}{p!}\left[\frac{\partial^pf(z)}{\partial z^p}\right]_{z=1}\right\}\,,\end{aligned}\ ] ] where @xmath148 is an arbitrary test function which is regular at @xmath149 , and @xmath150 , @xmath151 . in case the integration does not include the endpoint , we have ( @xmath152 ) @xmath153_+f(z ) = \int_0^c\!dz\,\frac{\ln^n(1-z)}{(1-z)^m } f(z)\,.\ ] ] we remark that the @xmath136-dependence of @xmath154 has its origin in the @xmath42 renormalization of the electromagnetic dipole operator .
the same is true for the @xmath136-dependence of @xmath155 and @xmath156 .
it reflects the fact that @xmath106 is not renormalized to all orders in perturbation theory . on the other hand , for the coefficient of @xmath107
there is an additional @xmath136-dependence present in @xmath157 .
it originates in the @xmath42 renormalization of @xmath107 .
indeed , using the renormalization group equation @xmath158 where @xmath159 @xcite , we recover the @xmath136-dependence of @xmath141 . in the remainder of this section
we discuss the numerical impact of the new contributions on the total decay rate of @xmath5 as well as on its first and second moments . for the @xmath26 contribution to the total decay rate we find ( using numerical values for @xmath160 and @xmath161 ) @xmath162\nonumber\\[2 mm ] & \hspace{-8mm}-\!\frac{8}{3}\ln^2(1-z_0)-\frac{4}{3}\left(10 - 2z_0-z_0 ^ 2\right)\ln(1-z_0 ) - \frac{4}{9}z_0\left(30 + 3z_0 - 2 z_0 ^ 2\right)\nonumber\\[2 mm ] & \hspace{-8mm}+\!\frac{\lambda_1}{2 m^2_b}\bigg\ { \frac{16}{9}\bigg[4-\pi^2 - 3\ln\frac{\mu}{m_b}\bigg]-\frac{8}{3}\ln^2(1-z_0)-
\frac{4z_0\left(30 - 63z_0 + 31z_0 ^ 2 + 5z_0 ^ 3\right)}{9(1-z_0)^2}\nonumber\\[2 mm ] & \hspace{-8mm}-\!\frac{4\left(30 - 72z_0 + 51z_0 ^ 2 - 2z_0 ^ 3 - 3z_0 ^ 4\right)}{9(1-z_0)^2 } \,\ln(1-z_0)\bigg\}+\frac{\lambda_2(\mu)}{2m^2_b}\bigg\{-\frac{2}{9}\bigg[87 + 32\pi^2\nonumber\\[2 mm ] & \hspace{-8mm}+\!27\ln\frac{\mu}{m_b}\bigg]-\frac{32}{3}\ln^2(1-z_0 ) + \frac{2\left(162 - 244z_0 + 113z_0 ^ 2 - 7z_0 ^ 3\right)}{3(1-z_0)}\,\ln(1-z_0)\nonumber\\[2 mm ] & \hspace{-8mm}+\!\frac{2z_0\left(54 - 49z_0 + 15z_0 ^ 2\right)}{1-z_0}\bigg\}\bigg ) \bigg ] + \dots\,.\end{aligned}\ ] ] the ellipses denote higher order terms in @xmath9 and @xmath10 , and we used @xmath163 . in order to get a rough estimate of the size of the power - corrections at @xmath0 we set @xmath164 and
use the numerical values @xmath165 , @xmath166 , @xmath167 and @xmath168 to obtain @xmath169 , a @xmath170% effect . in the intermediate step
we have singled out the new @xmath171 contributions .
the effect of the new corrections on the rate varies with the cut , from @xmath172 at @xmath173 to @xmath174 at @xmath175gev . for values of @xmath176gev
the corrections to the rate are significant , a @xmath177% effect for @xmath178gev , however , such high values of @xmath179 are well outside the range of applicability of the local ope .
moreover , at high @xmath179 part of the new effect is implicitly contained in the approach of @xcite .
the truncated @xmath55-th moment is defined through @xmath180 after expanding in @xmath9 , the first moment reads @xmath181 & \hspace{-11mm}+\!\frac{2}{27}\,z_0\left(96 - 60z_0 - 22z_0 ^ 2 + 9z_0 ^ 3\right)+\frac{\lambda_1}{2 m^2_b}\bigg\ { \!-\frac{2z_0\left(96 - 120z_0 - 52z_0 ^ 2 + 25 z_0 ^ 3 - 9 z_0 ^ 4\right)}{27(1-z_0)}\nonumber\\[2 mm ] & \hspace{-11mm}+\!\frac{46}{27}-\frac{16\left(4 - 7 z_0+z_0 ^ 3\right)}{9(1-z_0)}\,\ln(1-z_0)\bigg\ } + \frac{\lambda_2(\mu)}{2m^2_b}\bigg\{\!-\frac{1130}{27}-18\ln\frac{\mu}{m_b } + \frac{4}{9}\left(80 - 33z_0\right.\nonumber\\[2 mm ] & \hspace{-11mm}-\!\left.81 z_0 ^ 2 + 25z_0 ^ 3\right)\ln(1-z_0 ) + \frac{2}{27}\,z_0\left(480 + 42z_0 - 425z_0 ^ 2 + 81z_0 ^ 3\right)\bigg\}\bigg)\bigg]+\dots\,,\end{aligned}\ ] ] while the second central moment is given by @xmath182 & \hspace{-28mm}+\!\left.\frac{1}{270}\,(1-z_0)^2\left(61 - 898z_0 - 207z_0 ^ 2 + 144z_0 ^ 3\right)\right ] + \frac{\lambda_1}{8}\bigg[\frac{8}{27}\left(2 - 27z_0 + 7z_0 ^ 3\right)\ln(1-z_0)\nonumber\\[2 mm ] & \hspace{-28mm}+\!\frac{4}{405}\left(169 + 60z_0 - 780z_0 ^ 2 - 385z_0 ^ 3 + 225z_0 ^ 4 - 54z_0 ^ 5\right)\bigg ] + \frac{\lambda_2(\mu)}{8}\bigg[\frac{1}{9}\,\left(43 - 84z_0\right.\nonumber\\[2 mm ] & \hspace{-28mm}+\!\left.258z_0 ^ 2 - 292z_0 ^ 3 + 75z_0 ^ 4\right)\ln(1-z_0 ) - \frac{1}{540}\left(707 - 2580z_0 +
3750z_0 ^ 2 - 13820z_0 ^ 3\right.\nonumber\\[2 mm ] & \hspace{-28mm}+\!\left.14535z_0 ^ 4 - 2592 z_0 ^ 5\right)\bigg]\bigg\ } + \dots\,.\end{aligned}\ ] ] using the same numerical input as above we obtain , for @xmath183 , @xmath184 and @xmath185 , where we have again singled out the contribution of @xmath171 in the intermediate steps .
their effect on the truncated first and second central moment is of around @xmath186 and @xmath187 , respectively . for @xmath188$]gev , the effect on
the first moment varies between @xmath189% and @xmath190% , whereas that on the second central moment varies between @xmath191 and @xmath187 .
it is interesting to compare the @xmath35 coefficients of @xmath106 in ( [ rate ] ) , ( [ mom1 ] ) and ( [ mom2 ] ) with the results of @xcite , where the corrections of @xmath192 with @xmath193 have been computed for the cut rate and the first two moments . expanding our results in @xmath194 and keeping only the leading term we reproduce the results of @xcite .
[ plot : expansion ] summarizes the numerical relevance of the terms suppressed by powers of @xmath195 in the @xmath35 coefficients of @xmath196 for the same numerical input used above . in the cut rate and in the second moment , the leading approximation deviates by roughly + 50% and @xmath197% , respectively , already at @xmath178gev . in the first moment ,
the leading approximation is within roughly 10% of the complete result down to e = 1.8 gev . in conclusion ,
the range of applicability of the leading order approximation in the expansion in @xmath195 is clearly restricted to the region @xmath198gev , where sudakov logarithms become dominant , see for instance @xcite .
[ cols="^,^ " , ] the second moment represents a powerful constraint on the kinetic expectation value @xmath108 . since the @xmath0 correction decreases its coefficient by 5 to 9% in the range of cuts between 0 and 1.8gev , while the @xmath199 corrections are much smaller , we expect to extract a higher value of @xmath4 from radiative moments once the new corrections are included .
indeed , using @xmath34 at a more appropriate scale of order 1 - 2gev , the extracted @xmath4 gets shifted by approximately @xmath200% .
analogously , the small correction to the first moment leads to a roughly 10mev positive shift in @xmath3 .
all the above expressions refer to the on - shell scheme for @xmath3 and @xmath106 , which is inherent in the hqet calculation . in practical applications
however one adopts short - distance definitions of these parameters , as in the kinetic scheme @xcite . in this scheme
the new corrections are identical , but they additionally induce small @xmath201 contributions . therefore , the above rough estimates hold in the kinetic scheme as well . a complete phenomenological analysis of the moments , including all the available contributions , will be presented elsewhere .
we have computed the nlo contributions to the wilson coefficients of dimension five operators relevant for inclusive radiative b decays .
our results allow for a more precise evaluation of the moments of the photon distribution and will improve the determination of @xmath3 and of the kinetic expectation value , @xmath4 , from radiative moments .
we have estimated that the new contributions shift the value of @xmath4 extracted from the radiative moments by approximately @xmath200% and that of @xmath3 by roughly @xmath200mev .
the effect on the @xmath5 rate is below 1% for @xmath202gev .
we have performed the calculation analytically , using an off - shell matching procedure , a method that can be applied to inclusive semileptonic decays as well .
the @xmath203 corrections to the moments of @xmath204 have been computed numerically @xcite , however , the @xmath205 corrections are not yet known .
we also believe that an analytical result might be easier to implement in the fitting codes .
is grateful to g. ridolfi for collaboration at the early stage of this work and to n. uraltsev for many useful discussions .
we also thank m. misiak for relevant communications .
work supported by the eu s marie - curie research training network under contract mrtn - ct-2006 - 035505 ` tools and precision calculations for physics discoveries at colliders ' .
m. antonelli _ et al .
_ , arxiv:0907.5386
. m. artuso _ et al .
_ , eur .
j. c * 57 * ( 2008 ) 309 [ arxiv:0801.1833 [ hep - ph ] ] .
i. i. y. bigi , n. g. uraltsev and a. i. vainshtein , phys .
b * 293 * ( 1992 ) 430 [ erratum - ibid .
b * 297 * ( 1993 ) 477 ] [ arxiv : hep - ph/9207214 ] ; i. i. y. bigi , m. a. shifman , n. g. uraltsev and a. i. vainshtein , phys .
* 71 * ( 1993 ) 496 [ arxiv : hep - ph/9304225 ] .
b. blok , l. koyrakh , m. a. shifman and a. i. vainshtein , phys .
d * 49 * ( 1994 ) 3356 [ erratum - ibid .
d * 50 * ( 1994 ) 3572 ] [ arxiv : hep - ph/9307247 ] ; a. v. manohar and m. b. wise , phys .
d * 49 * ( 1994 ) 1310 [ arxiv : hep - ph/9308246 ] .
e. barberio _ et al .
_ [ heavy flavor averaging group ] , arxiv:0808.1297 , + see http://www.slac.stanford.edu / xorg / hfag/. i. i. y. bigi , m. a. shifman , n. g. uraltsev and a. i. vainshtein , int .
j. mod .
a * 9 * ( 1994 ) 2467 [ arxiv : hep - ph/9312359 ] ; m. neubert , phys .
d * 49 * ( 1994 ) 4623 [ arxiv : hep - ph/9312311 ] .
d. benson , i. i. bigi and n. uraltsev , nucl .
b * 710 * , 371 ( 2005 ) [ arxiv : hep - ph/0410080 ] ; p. gambino , p. giordano , g. ossola and n. uraltsev , jhep * 0710 * , 058 ( 2007 ) [ arxiv:0707.2493 ] .
m. misiak _ et al .
_ , phys .
* 98 * ( 2007 ) 022002 [ arxiv : hep - ph/0609232 ] .
m. misiak , arxiv:0808.3134 ; t. ewerth , arxiv:0909.5027 ; m. misiak , arxiv:0911.1651 .
a. f. falk , m. e. luke and m. j. savage , phys .
d * 49 * ( 1994 ) 3367 [ arxiv : hep - ph/9308288 ] . c. w. bauer , phys .
d * 57 * ( 1998 ) 5611 [ erratum - ibid .
d * 60 * ( 1999 ) 099907 ] [ arxiv : hep - ph/9710513 ] and references therein .
m. neubert , eur .
j. c * 40 * ( 2005 ) 165 [ arxiv : hep - ph/0408179 ] and phys .
d * 72 * ( 2005 ) 074025 [ arxiv : hep - ph/0506245 ] .
h. georgi , phys .
b * 240 * ( 1990 ) 447 ; e. eichten and b. r. hill , phys .
b * 234 * ( 1990 ) 511 .
m. neubert , phys .
245 * , 259 ( 1994 ) [ arxiv : hep - ph/9306320 ] .
a. v. manohar and m. b. wise , camb .
monogr . part .
* 10 * ( 2000 ) 1 .
e. eichten and b. r. hill , phys .
b * 243 * ( 1990 ) 427 .
a. f. falk and m. neubert , phys .
d * 47 * ( 1993 ) 2965 [ arxiv : hep - ph/9209268 ] .
abbott , nucl .
b185 ( 1981 ) 189 ; abbott , acta phys .
b13 ( 1982 ) 33 ; abbott , grisaru , schaefer , nucl .
phys . b229 ( 1983 ) 372 .
a. ali and c. greub , z. phys .
c * 49 * ( 1991 ) 431 ; phys .
b*259 * ( 1991 ) 182 ; phys .
b * 361 * ( 1995 ) 146 [ arxiv : hep - ph/9506374 ] .
a. f. falk , b. grinstein and m. e. luke , nucl .
b * 357 * ( 1991 ) 185 .
k. melnikov and a. mitov , phys .
b * 620 * ( 2005 ) 69 [ arxiv : hep - ph/0505097 ] .
i. i. y. bigi , m. a. shifman , n. uraltsev and a. i. vainshtein , phys .
d * 56 * ( 1997 ) 4017 [ arxiv : hep - ph/9704245 ] ; phys .
d * 52 * ( 1995 ) 196 [ arxiv : hep - ph/9405410 ] .
t. becher , h. boos and e. lunghi , jhep * 0712 * ( 2007 ) 062 [ arxiv:0708.0855 ] . | we compute the @xmath0 corrections to the wilson coefficients of the dimension five operators emerging from the operator product expansion of inclusive radiative @xmath1 decays .
we discuss the impact of the resulting @xmath2 corrections on the extraction of @xmath3 and @xmath4 from the moments of the photon spectrum .
* power suppressed effects in @xmath5 at @xmath0 * 1.5 cm * thorsten ewerth , paolo gambino and soumitra nandi * + _ dip .
fisica teorica , univ .
di torino & infn torino , i-10125 torino , italy _ |
wormhole physics dates back to the formulation of general relativity ( gr ) .
indeed , in 1916 , months after einstein presented his gravitational field equations , karl schwarzschild found the first solution of einstein s equations , which described the gravitational field of a vacuum non - rotating spherically symmetric solution . in the same year , ludwig flamm published a paper in which the geometry of the schwarzschild solution was studied more closely .
he pointed out a `` tunnel - shaped '' nature of space near the schwarzschild radius , this being perhaps the first move towards the modern concept of the `` throat '' in wormholes @xcite . paging through the literature , one finds next that tunnel - like solutions were considered , in 1935 , by einstein and rosen , where they constructed an elementary particle model represented by a bridge connecting two identical sheets @xcite .
they considered the possibility that fundamental particles such as the electron could be represented as microscopic spacetime tunnels that convey fluxes of the electric field .
these tunnels were later denoted the _ einstein - rosen bridge_. in fact , the einstein - rosen bridge is a coordinate artifact arising from choosing a coordinate patch , which is defined to double - cover the asymptotically flat region exterior to the black hole event horizon .
the field had lain dormant for about twenty years when , in 1955 , john wheeler , who was beginning to be interested in topological issues in gr , explored solutions of the coupled einstein maxwell equations , which he denoted _ geons _
( gravitational - electromagnetic entities ) @xcite .
these were considered to be objects of the quantum foam connecting different regions of spacetime at the planck scale .
however , the term ` wormhole ' was only used for the first time in 1957 @xcite , where misner and wheeler presented a tour de force wherein riemannian geometry of manifolds of nontrivial topology was investigated with an ambitious view to explaining all of physics .
the aim was to use the source - free maxwell equations , coupled to einstein gravity , with nontrivial topology , to build models for classical electrical charges and all other particle like entities in classical physics .
subsequently to the geon concept , several wormhole solutions were obtained and discussed within different contexts @xcite .
however , it was only in 1988 that the full - fledged renaissance of wormhole physics took place through the seminal morris - thorne paper @xcite , and the theme is still in full flight .
morris and thorne , considered static and spherically symmetric traversable wormholes , and thoroughly analysed their fundamental properties .
it was found that these traversable wormholes possess a stress - energy tensor that violates the null energy condition ( nec ) , a property that was denoted _
exotic matter_. beside being hypothetical short - cuts in spacetime and consequently useful for inter- and intra - universe travel , they were found to possess other intriguing applications , such as , the usage for time - travel @xcite and investigating the interior of a black hole @xcite , amongst others .
thus , a fundamental ingredient for the morris - thorne wormhole , i.e. , for static and spherically symmetric wormhole solutions , is the violation of the nec @xcite .
exotic matter is particularly troublesome for measurements made by observers traversing through the throat with a radial velocity close to the speed of light , as for sufficiently high velocities , @xmath0 , the observer will measure a negative energy density @xcite .
although classical forms of matter are believed to obey these energy conditions , it is a well - known fact that they are violated by certain quantum fields , such as the casimir effect .
in fact , the recent discovery that the universe is undergoing an accelerated expansion @xcite may be due to an exotic cosmic fluid that lies in the phantom regime .
the realization of this fact has led to the study of wormhole solutions supported by different kinds of phantom fluids ( see , for instance @xcite ) .
indeed the violation of the energy conditions is a subtle issue , as almost all known and physically possible forms of matter satisfy these energy conditions , and we recall that their imposition is one of the necessary assumptions for proving the hawking - penrose singularity theorems @xcite .
thus , one may adopt the approach of minimizing the usage of exotic matter @xcite . in this context
, a plethora of solutions have been investigated , using a wide variety of approaches diffpo , diffpo0,diffpo2,diffpo3,18 .
more specifically , in rotating solutions it was found that the exotic matter lies in specific regions around the throat , so that it is possible for a certain class of infalling observers to move around the throat as to avoid the exotic matter supporting the wormhole @xcite . using the thin shell formalism , solutions where the exotic matter is concentrated at the throat
have also been extensively investigated @xcite . in the context of modified gravity
it was shown that one may impose that the matter threading the wormhole satisfies the energy conditions , so that it is the higher order curvature terms that sustain these exotic geometries @xcite .
astrophysical signatures have also been explored in the literature harko:2008vy . for dynamic wormholes , the nec , or more precisely the averaged null energy condition ,
can be avoided in certain regions hochvisserprl98,hochvisserprd98,kar , kar - sahdev , kim - evolvwh , arellano:2006ex .
a particularly interesting case is that of a wormhole in a time - dependent inflationary background @xcite , in which the primary goal was to use inflation to enlarge an initially small and possibly submicroscopic wormhole .
it is also possible that the wormhole will continue to be enlarged by the subsequent frw phase of expansion .
one could perform a similar analysis to @xcite by replacing the desitter scale factor by an frw scale factor @xcite . in particular , in kar , kar - sahdev specific examples for evolving wormholes that exist only for a finite time were considered , and a special class of scale factors that exhibit ` flashes ' of the wec violation were also analyzed . the present paper investigates the possibility and naturalness of expanding wormholes in higher dimensions which is an important ingredient of the modern theories of fundamental physics , such as string theory , supergravity , kaluza - klein , and others .
one of our motivations for considering wormhole solutions in an expanding cosmological background refers to the inflation theory guth where the quantum fluctuations in the inflaton field are considered as the seed of large scale structures in the universe .
as mentioned above , the non - trivial topological objects such as microscopic wormholes may have been formed during inflation and enlarged to macroscopic ones as the universe expanded @xcite .
we also explore the possibility that these higher - dimensional wormholes satisfy the nec , and we explicitly show that this is indeed the case . this paper is organized in the following way : in section [ secii ] , we present the @xmath1dimensional field equations for the specific case of a spatially - independent curvature scalar . in section [ seciii ] ,
we analyse the two - way traversability conditions of the wormhole structure . in section [ seciv ] , we explore wormhole solutions in different expansionary regimes , and finally in section [ conclusion ] , we conclude .
the action of gr in @xmath2dimensions is written as @xmath3where @xmath4 is the scalar curvature and @xmath5 is the matter lagrangian density ; we have considered @xmath6 . varying this action with respect to the metric
, we obtain the @xmath2-dimensional einstein equations @xmath7 , where @xmath8 , and @xmath9 is the matter stress - energy tensor .
since we are looking for expanding wormhole solutions in a cosmological background , we use the metric@xmath10 , \label{metn}\]]in which @xmath11 is the scale factor and @xmath12 is an unknown dimensionless function , defined as @xmath13 , where @xmath14 denotes the shape function @xcite . note that this metric is a generalization of the friedmann - robertson - walker ( frw ) metric , although being less symmetric than the latter . with this generalization ,
metric ( [ metn ] ) is still isotropic about the center of the symmetry , though not necessarily homogeneous .
when the dimensionless shape function vanishes , @xmath15 the metric ( metn ) reduces to the flat frw metric ; and as @xmath16 it approaches the static wormhole metric . to see that the `` wormhole '' form of the metric is preserved with time , consider an embedding of @xmath17 and @xmath18 slices of the spacetime given by eq .
( [ metn ] ) , in a flat 3-dimensional euclidean space with metric @xmath19 in this context , the metric of the wormhole slice is @xmath20 now , comparing the coefficients of @xmath21 , one has @xmath22 it is important to keep in mind , in particular , when considering derivatives , that eqs .
( [ coef1:phi])-([coef2:phi ] ) do not represent a `` coordinate transformation '' , but rather a `` rescaling '' of the @xmath23 coordinate on each @xmath24 slice @xcite .
with respect to the @xmath25 coordinates , the `` wormhole '' form of the metric will be preserved if the metric on the embedded slice has the form @xmath26 where @xmath27 , i.e. , @xmath28 has a minimum at some @xmath29 .
equation ( [ slice ] ) can be rewritten in the form of eq .
( [ whslice ] ) by using eqs .
( [ coef1:phi])-(coef2:phi ) and @xmath30 the evolving wormhole will have the same overall size and shape relative to the @xmath25 coordinate system , as the initial wormhole had relative to the initial @xmath31 embedding space coordinate system .
this is due to the fact that the embedding space corresponds to @xmath32 coordinates that `` scale '' with time ( each embedding space corresponds to a particular value of @xmath24 ) . following the embedding procedure @xcite , using eqs .
( [ barredslice ] ) and ( [ whslice ] ) , one deduces that @xmath33 which implies @xmath34 therefore , we see that the relation between the embedding space at any time @xmath35 and the initial embedding space at @xmath36 , from eqs .
( [ coef2:phi ] ) and ( [ embed : relation ] ) , is given by the following @xmath37\,.\ ] ] relative to the @xmath25 coordinate system the wormhole will always remain the same size , as the scaling of the embedding space compensates for the evolution of the wormhole .
however , the wormhole will change size relative to the initial @xmath36 embedding space .
writing the analog of the `` flaring out condition '' @xcite for the evolving wormhole we have @xmath38 , at or near the throat . from eqs .
( [ coef1:phi ] ) , ( [ coef2:phi ] ) , ( [ bar : b ] ) , and ( [ barredembedding ] ) , it follows that @xmath39 at or near the throat , where the prime denotes the derivative with respect to @xmath23 .
note that this also implies @xmath40 , at or near the throat .
taking into account eqs .
( [ coef1:phi ] ) , ( [ bar : b ] ) , and @xmath41 , one may rewrite the right - hand - side of eq .
( [ barred : flareout ] ) relative to the barred coordinates as @xmath42 or @xmath43 at or near the throat .
one verifies that using the barred coordinates , the flaring out condition eq .
( [ barred : flareout2 ] ) , has the same form as for the static wormhole .
thus , it can be shown that metric ( [ metn ] ) represents a traversable wormhole provided@xmath44where @xmath45 is the wormhole throat , which represents a minimum radius in the wormhole space - time @xcite .
the second condition is imposed in order to avoid a change in the metric signature .
the third condition is the flaring - out condition and plays a fundamental role in the analysis of the violation of the energy conditions .
note that the comoving radial distance defined by @xmath46should be real and finite everywhere in spite of the fact that the @xmath47-component of the covariant metric diverges at the throat ; the @xmath48 signs denote the upper and lower parts of the wormhole .
the energy - momentum tensor is @xmath49 @xmath50 @xmath51 @xmath52 @xmath53 , so that using the einstein field equations and eq .
( [ metn ] ) , the ( @xmath54)dimensional field equations are satisfied by the following stress - energy profile @xmath55where the overdot denotes a derivative with respect to time .
the ricci scalar will play a fundamental role in our analysis , so we write it down explicitly as @xmath56since the ricci scalar is only a function of time in standard cosmological models , it provides a motivation to use this property as a simplifying assumption in our calculations , in the presence of a wormhole . in other words
, we are looking for classes of solutions corresponding to the choice of a homogeneous ricci scalar , i.e. , @xmath57 , which implies @xmath58 the above differential equation yields the following solution @xmath59where the condition @xmath60 was used to eliminate the integration constant .
we point out that although @xmath61 can , in principle , be a continuous variable , we have used the fact that the space - time is asymptotically frw and applied the normalization @xmath62 for the curvature constant .
it is worthwhile to mention that it is common to consider static wormholes supported by radiation that have a traceless stress - energy tensor @xcite .
in such a case , the ricci scalar vanishes if there is no cosmological constant within the framework of gr .
our assumption leads to the same situation , if the scale factor is assumed to be independent of time , i.e. , for a static case .
as mentioned before , the dimensionless shape function @xmath12 should satisfy the conditions ( [ 3 - 3 ] ) .
it is easy to show that for @xmath63 and @xmath64 these conditions are satisfied whereas for @xmath65 they are not .
therefore we continue our discussions using @xmath63 and @xmath64 which present flat and open universes , respectively . with @xmath12 in hand , given by eq .
( [ 3 ] ) , one can rewrite the field equations for the spatially flat background ( @xmath63 ) as : @xmath66where @xmath67 and @xmath68 are the respective flat background components given by @xmath69respectively . for the specific case of the open background ( @xmath70 )
, we have @xmath71where the @xmath72 and @xmath73 components correspond to the `` open background '' , and are given by @xmath74 since our solutions are in a spherically symmetric cosmological background , the components of the stress - energy tensor should be asymptotically independent of @xmath23 .
it is easy to see this expected behaviour is obeyed by them .
one of the most interesting properties of a wormhole as pointed out by morris and thorne @xcite is its two - way traversability . in this section
, some proofs will be presented to show that the wormholes discussed in this paper are indeed two - way traversable .
consider a radially moving light signal emitted from a co - moving source .
we assume that the signal is emitted at @xmath75 ( @xmath76 is a co - moving coordinate ) and received by a distant co - moving observer at @xmath77 . using the metric ( [ metn ] ) and eq . ( [ 4 - 1 ] ) for a radial beam , we have @xmath78where @xmath79 is the comoving radial distance defined in eq .
( [ 4 - 1 ] ) .
note that @xmath80 and @xmath76 can belong to either side of the throat .
it is obvious that the rhs of eq .
( [ 42 ] ) is independent of time .
therefore , the lhs should also be so and a signal which is emitted in an interval @xmath81 should be received in an interval @xmath82 such that ( @xmath83 ) @xmath84since @xmath82 and @xmath85 are very short time intervals , one deduces that @xmath86where @xmath87 is the scale factor at the time of observation , @xmath88 is the scale factor at the time of emission , and @xmath89 is the cosmological redshift .
this shows that the redshift is the same as the cosmological redshift and no extra redshift is caused by the wormhole .
it remains to examine whether the signal ever reaches the throat in a finite time or not .
this will be addressed below .
since @xmath23 is greater than @xmath90 on both sides of the throat , one can not trivially deduce whether the light signal passes through the throat or not using the @xmath23 coordinate .
therefore , we need to transform from the radial coordinate @xmath23 to the comoving radial coordinate @xmath91 in order to analyse this behavior . using eq .
( [ 4 - 1 ] ) , the analysis is more transparent . consider radial motion so that the geodesic equation reads @xmath92and @xmath93equation ( [ 49 ] ) yields the first integral @xmath94equation ( [ 50 ] )
shows that @xmath95 does not undergo a sign change along the path and neither does it vanish .
this shows that the particle or signal continues its path , passes the throat and goes to the other side of the wormhole and therefore the wormhole is two - way traversable . in order to prove that there is a finite proper distance between a specific point and the throat , eq .
( [ 4 - 1 ] ) should be solved explicitly . for arbitrary @xmath96
, there is no analytical solution for the integral ( [ 4 - 1 ] ) .
therefore , one could obtain @xmath79 in the vicinity of the throat , which is sufficient for our purpose . using the approximate relation @xmath97 , one obtains @xmath98which are clearly finite distances .
therefore , the throat is not located at spatial infinity and a finite time is required to reach it . we should mention that horizons are theoretical constructs that qualitatively have two main specific properties : first , they are one - way membranes , and second , the corresponding redshift ( as observed by a distant observer ) is infinite .
more precise mathematical details can be found in @xcite .
since it is common to consider the singularities of the metric as candidates of being horizons , one might ask whether the coordinate singularity at the throat forms a horizon . based on the above - mentioned general qualitative features of horizons and according to the results obtained in subsections ( [ sub-51 ] ) and ( [ sub-52 ] ) , it is justified that there is no horizon at or around the throat .
the possibility of the existence of a cosmological horizon , however , depends on the behavior of the scale factor @xmath99 and is not essentially affected by the presence or absence of the wormhole .
one of the properties of normal matter is that it satisfies the energy conditions , in particular , the null energy condition ( nec ) and the weak energy condition ( wec ) .
it was mentioned in the introduction that the matter that supports the static wormhole geometry violates the nec and is therefore denoted ` exotic matter ' .
the nec requires that @xmath100 , where @xmath101 is _ any _ null vector . in terms of the energy density , radial pressure and tangential pressure the nec becomes @xcite @xmath102note that the wec , in addition to the conditions considered above , also imposes a positive energy density , @xmath103 . in what follows ,
we investigate the nec for the wormhole solutions in different expansion regimes . in the case of a flat background , one can obtain two different solutions for the scale factor @xmath11 by applying the equation of state @xmath104 , given by @xmath105the specific case of @xmath106 represents the inflationary regime . the case of
@xmath107 represents the radiation dominated and matter dominated expansion regimes , by considering @xmath108 and @xmath109 , respectively .
consider first the inflationary expansion regime , where by using eqs .
( rhoflat)-([bgflat ] ) , one obtains @xmath110it is clear that @xmath111 is always negative while @xmath112 and @xmath113 are always positive .
therefore , the nec is always violated .
but @xmath114 tends to zero as @xmath115 increases and therefore the wormhole matter ranges from an exotic matter regime to normal matter over time . at the throat , assuming @xmath116 , with respect to time , with @xmath117 and @xmath118 for @xmath119 ( solid curve ) and @xmath120 ( dashed curve ) .
the plot shows that although initially there is normal matter at throat , the throat matter tends to an exotic matter regime over time.,width=264 ] we continue our discussions with the second solution for the scale factor , for @xmath107 . in this case
, we have @xmath121for @xmath122 , it is clear that the nec is violated due to @xmath123 . for @xmath124 , we have that @xmath112 and @xmath125 are always positive , while the quantity @xmath111 should be analysed in more detail .
figure [ i1 ] shows the behaviour of @xmath111 with respect to time at the throat @xmath116 for @xmath118 , @xmath117 and @xmath109 and @xmath126 .
it can be seen that although the wormhole matter at the throat initially satisfies the nec , the latter is violated as time passes . in fig .
[ i2 ] , the quantity @xmath111 is plotted against @xmath23 and @xmath115 with @xmath116 , @xmath127 , @xmath117 for @xmath128 and @xmath126 .
the figure shows that the region of exotic matter in the vicinity of the wormhole throat increases as time increases . as we are considering higher - dimensional wormholes in an expanding spacetime , an interesting scenario to examine
is whether these dynamic wormholes could be constructed from normal matter for @xmath129 .
this is indeed the case for the solutions discussed here , by choosing suitable values for the constants .
figure [ ii1 ] plots @xmath111 against @xmath23 and @xmath115 for @xmath130 , @xmath131 , @xmath132 and @xmath133 . as depicted in the figure
, @xmath134 is always positive for this choice of constants and therefore the nec ( and also wec ) is satisfied for the whole wormhole structure . with respect to @xmath23 and @xmath115 for @xmath132 , @xmath116 , @xmath135 and @xmath136 .
this plot shows that for suitable choices of constants , there are wormhole structures constructed from matter that satisfy the null energy condition.,width=302 ] in the case of the open background , by applying @xmath137 , we consider the following analytical solutions @xmath138where the case @xmath106 corresponds to the inflationary regime .
let us first investigate the solution corresponding to the inflationary expansion regime , @xmath139 .
using eqs .
( [ rhoopen])-([bgopen ] ) , we have @xmath140as in the case of the flat background , @xmath111 is always negative , implying the violation of the nec throughout the spacetime , while @xmath112 and @xmath125 are always positive .
however , @xmath111 tends to zero as @xmath115 increases and therefore during the inflationary era the wormhole matter tends from an exotic matter regime to a normal matter one , at temporal infinity .
consider now the second case where @xmath141 and @xmath142 . in this case , one obtains the following relationships @xmath143 \frac{1}{{t}% ^{2 } } , \\ \rho + p_{t } & = & \bigg [ \frac{\left ( n-2\right ) \left
( { \mathit{r}}_{0}^{n}+{% \mathit{r}}_{0}^{n-2}\right ) } { 2{a}_{4}^{2}{r}^{n } } \notag \\ & & + \frac{(n-1)\left ( { a}_{4}^{2}-1\right ) } { { a}_{4}^{2}}\bigg ] \frac{1}{{t}% ^{2}}.\end{aligned}\]]for @xmath144 , it is clear that the nec is violated due to @xmath123 ; note also that @xmath145 .
however , this case @xmath144 should be excluded , as @xmath112 coincides with the background energy density .
this would imply that the energy density of the universe is negative , so it is not physically acceptable . for @xmath146 ,
it is obvious that @xmath147 and @xmath125 are always positive while @xmath111 should be investigated .
figure [ ii2 ] depicts @xmath111 in terms of @xmath23 and @xmath115 for @xmath148 , @xmath149 and @xmath150 .
this figure shows that by choosing suitable constants , there is a wormhole structure constructed from normal matter . with respect to @xmath23 and @xmath115 for @xmath132 , @xmath148 and @xmath149 .
it shows that in the case of an open universe , it is possible to have wormholes constructed from normal matter.,width=302 ]
the present paper investigates the possibility and naturalness of expanding wormholes in higher dimensions which is an important ingredient of modern theories of fundamental physics , for instance , string theory , supergravity and kaluza - klein , amongst others .
one of our motivations for considering wormhole solutions in an expanding cosmological background refers to the inflationary theory where the quantum fluctuations in the inflaton field may have served as the seed for the large scale structures in the universe .
non - trivial topological objects such as microscopic wormholes may have been formed through the quantum foam and enlarged to macroscopic size during inflation and in the subsequent expansion of the universe .
indeed , if most of the wormholes in the quantum foam survived enlargement through inflation , then the universe might be far more inhomogeneous and topologically complicated than we observe .
indeed , postulating higher - dimensional spacetimes is an important ingredient of modern theories of fundamental physics . in this context
, the existence of higher dimensions may help construct wormhole solutions that respect energy conditions .
in particular , in a cosmological set up , microscopic , dynamical wormholes produced in the early universe may be inflated to macroscopic scales and thus be at least in principle astrophysically observable . in this work , by assuming a homogeneous matter field ( i.e. energy density depending only on the time coordinate ) , which holds in the standard cosmology , we arrived at interestingly simple and exact solutions .
more specifically , we considered a particular class of wormhole solutions corresponding to a spatially homogeneous ricci scalar .
the possibility of obtaining solutions with normal and exotic matter was explored and we found new solutions including those that satisfy the nec in specific time intervals .
in particular , in five dimensions , we found solutions that satisfy the nec everywhere . c. w. misner and j. a. wheeler , annals phys . * 2 * , 525 ( 1957 ) .
h. g. ellis , j. math .
phys . 14 ( 1973 ) 104 ; k. bronnikov , acta phys . pol .
b 4 ( 1973 ) 251 ; k. a. bronnikov , v. n. melnikov , g. n. shikin , and k. p. staniukovich , scalar , ann .
phys ( n.y ) 118 ( 1979 ) 84 ; g. clement , gen .
rel . and grav .
16 ( 1984 ) 131 ; g. clement , gen . rel . and grav .
16 ( 1984 ) 477 ; g. clement , gen . rel . and grav .
16 ( 1984 ) 491 .
d. hochberg and m. visser , phys .
d 56 ( 1997 ) 4745 ; d. hochberg and m. visser , phys . rev .
d 58 ( 1998 ) 044021 ; j. l. friedman , k. schliech , and d. m. witt , phys .
71 ( 1998 ) 1486 ; d. ida and s. a. heyward , phys .
a 260 ( 1999 ) 175 ; m. parikh and j. p. van der schaar , arxiv:1406.5163 [ hep - th ] .
s. v. sushkov , phys .
d * 71 * 043520 ( 2005 ) ; f. s. n. lobo , phys .
d * 71 * 084011 ( 2005 ) f. s. n. lobo , phys .
d * 71 * , 124022 ( 2005 ) ; f. s. n. lobo , phys . rev .
d * 73 * , 064028 ( 2006 ) ; f. s. n. lobo , class . quant . grav .
* 23 * , 1525 ( 2006 ) a. debenedictis , r. garattini and f. s. n. lobo , phys
. rev .
d * 78 * , 104003 ( 2008 ) ; f. s. n. lobo , phys .
d * 75 * , 024023 ( 2007 ) .
h. b. g. casimir , proc .
51 ( 1948 ) 793 ; t. a. roman , phys .
d 33 ( 1986 ) 3526 .
m. visser , s. kar and n. dadhich , phys .
* 90 * , 201102 ( 2003 ) b. bhawal and s. kar , phys .
d * 46 * , 2464 ( 1992 ) ; m. h. dehghani and z. dayyani , phys .
d * 79 * , 064010 ( 2009 ) ; s. h. mazharimousavi , m. halilsoy , and z. amirabi , phys .
d * 81 * , 104002 ( 2010 ) ; s. h. mazharimousavi , m. halilsoy , and z. amirabi , class .
gravity * 28 * , 025004 ( 2011 ) ; m. h. dehghani and m. r. mehdizadeh , phys
d * 85 * , 024024 ( 2012 ) .
f. s. n. lobo and m. visser , class .
* 21 * , 5871 ( 2004 ) m. s. r. delgaty and r. b. mann , int . j. mod . phys .
d 4 ( 1995 ) 231 ; j. p. s. lemos , f. s. n. lobo , and s. q. deoliveira , phys .
d 68 ( 2003 ) 064004 ; j. p. s. lemos and f. s. n. lobo , phys . rev .
d 69 ( 2004 ) 104007 .
n. riazi , j. korean astron .
( 1996 ) 283 ; n. riazi , astrophys . space sci . 283 ( 2003 ) 235 ; j. m. maldacena and l. maoz , jhep 0402 ( 2004 ) 053 ; e. bergshoeff , a. collinucci , a. ploegh , s. vandoren , and t. van riet , jhep 0601 ( 2006 ) 061 ; n. arkani - hamed , j. orgera , and j. polchinski , jhep 0712 ( 2007 ) 018 ; v. sushkov and m. kozyrev , phys .
d 84 ( 2011 ) 124026 ; f. s. n. lobo and m. a. oliveira , phys .
d * 81 * , 067501 ( 2010 ) .
s. w. kim and h. lee , phys .
d 63 ( 2001 ) 064014 ; f. parisio , ibid .
63 , 087502 ( 2001 ) ; k. a. bronnikov , phys . rev .
d 63 ( 2001 ) 044005 ; a. v. b. arellano and f. s. n. lobo , class .
quant . grav .
* 23 * , 7229 ( 2006 ) ; h. maeda , t. harada and b. j. carr , phys .
d * 79 * , 044034 ( 2009 ) .
e. teo , phys . rev .
d * 58 * , 024014 ( 1998 ) m. visser , phys .
d 39 ( 1989 ) 3182 ; m. visser , nucl .
b 328 ( 1989 ) 203 ; e. poisson and m. visser , phys . rev .
d * 52 * , 7318 ( 1995 ) ; f. s. n. lobo and p. crawford , class . quant .
* 21 * , 391 ( 2004 ) e. f. eiroa and g. e. romero , gen .
rel . grav .
* 36 * , 651 ( 2004 ) f. s. n. lobo and p. crawford , class .
* 22 * , 4869 ( 2005 ) j. p. s. lemos and f. s.
n. lobo , phys . rev .
d * 78 * , 044030 ( 2008 ) f. s. n. lobo , class .
* 21 * , 4811 ( 2004 ) f. s. n. lobo and a. v. b. arellano , class .
* 24 * , 1069 ( 2007 ) f. s. n. lobo , gen .
* 37 * , 2023 ( 2005 ) .
f. s. n. lobo , phys .
d * 75 * , 064027 ( 2007 ) ; f. s. n. lobo , class .
* 25 * , 175006 ( 2008 ) ; f. s. n. lobo and m. a. oliveira , phys .
d * 80 * , 104012 ( 2009 ) ; n. m. garcia and f. s. n. lobo , phys . rev .
d * 82 * , 104018 ( 2010 ) ; n. montelongo garcia and f. s. n. lobo , class .
* 28 * , 085018 ( 2011 ) ; c. g. boehmer , t. harko and f. s. n. lobo , phys .
d * 85 * , 044033 ( 2012 ) ; s. capozziello , t. harko , t. s. koivisto , f. s. n. lobo and g. j. olmo , phys .
d * 86 * , 127504 ( 2012 ) ; t. harko , f. s. n. lobo , m. k. mak and s. v. sushkov , phys . rev .
d * 87 * , no .
6 , 067504 ( 2013 ) .
t. harko , z. kovacs and f. s. n. lobo , phys .
d * 78 * , 084005 ( 2008 ) ; t. harko , z. kovacs and f. s. n. lobo , phys .
d * 79 * , 064001 ( 2009 ) .
d. hochberg and m. visser , phys .
lett . * 81 * , 746 ( 1998 ) .
d. hochberg and m. visser , phys .
d * 58 * , 044021 ( 1998 ) .
s. kar , phys .
d * 49 * , 862 ( 1994 ) .
s. kar and d. sahdev , phys .
d * 52 * , 2030 ( 1995 ) ; n. dadhich , s. kar , s. mukherjee , and m. visser , phys .
d 65 ( 2002 ) 064004 ; k.a .
bronnikov and s.w .
kim , phys .
d 67 ( 2003 ) 064027 ; h. lu and j. mei , phys .
b 666 ( 2008 ) 511 . | in this work , we consider the possibility of expanding wormholes in higher - dimensions , which is an important ingredient of modern theories of fundamental physics .
an important motivation is that non - trivial topological objects such as microscopic wormholes may have been enlarged to macroscopic sizes in an expanding inflationary cosmological background .
since the ricci scalar is only a function of time in standard cosmological models , we use this property as a simplifying assumption .
more specifically , we consider a particular class of wormhole solutions corresponding to the choice of a spatially homogeneous ricci scalar .
the possibility of obtaining solutions with normal and exotic matter is explored and we find a variety of solutions including those in four dimensions that satisfy the null energy condition ( nec ) in specific time intervals .
in particular , for five dimensions , we find solutions that satisfy the nec throughout the respective evolution . |
heterogeneous and disordered media emerge in several applications in physics , engineering and life sciences .
examples are widespread and concern for instance light propagation through engineered optical materials @xcite or turbid media @xcite , tracer diffusion in biological tissues @xcite , neutron diffusion in pebble - bed reactors @xcite or randomly mixed immiscible materials @xcite , inertial confinement fusion @xcite , and radiation trapping in hot atomic vapours @xcite , only to name a few .
stochastic geometries provide convenient models for representing such configurations , and have been therefore widely studied @xcite , especially in relation to heterogeneous materials @xcite , stochastic or deterministic transport processes @xcite , image analysis @xcite , and stereology @xcite .
a particularly relevant class of random media is provided by the so - called poisson geometries @xcite , which form a prototype process of isotropic stochastic tessellations : a portion of a @xmath0-dimensional space is partitioned by randomly generated @xmath3-dimensional hyper - planes drawn from an underlying poisson process . the resulting random geometry ( i.e. , the collection of random polyhedra determined by the hyper - planes ) satisfies the important property that an arbitrary line thrown within the geometry will be cut by the hyper - planes into exponentially distributed segments @xcite . in some sense ,
the exponential correlation induced by poisson geometries represents perhaps the simplest model of ` disordered ' random fields , whose single free parameter ( i.e. , the average correlation length ) can be deduced from measured data @xcite . following the pioneering works by goudsmit @xcite , miles @xcite and richards @xcite for @xmath4
, the statistical features of the poisson tessellations of the plane have been extensively analysed , and rigorous results have been proven for the limit case of domains having an infinite size : for a review , see , e.g. , @xcite .
an explicit construction amenable to monte carlo simulations for two - dimensional homogeneous and isotropic poisson geometries of finite size has been established in @xcite .
theoretical results for infinite poisson geometries have been later generalized to @xmath1 , which is key for real - world applications but has comparatively received less attention , and higher dimensions by several authors @xcite .
the two - dimensional construction for isotropic poisson geometries has been analogously extended to three - dimensional ( and in principle @xmath0-dimensional ) domains @xcite . in this work
, we will numerically investigate the statistical properties of @xmath0-dimensional isotropic poisson geometries by resorting to monte carlo simulation , with special emphasis on the case @xmath1 .
our aim is two - fold : first , we will focus on finite - size effects and on the convergence towards the limit behaviour of infinite domains . in order to assess the impact of dimensionality on the convergence patterns , comparisons to analogous numerical or exact findings obtained for @xmath5 and @xmath4 ( where available ) will be provided . in so doing ,
we will also present and discuss the simulation results for some physical observables for which exact asymptotic results are not known , yet .
then , we will consider the case of ` coloured ' poisson geometries , where each polyhedron is assigned a label with a given probability .
such models emerge , for instance , in connection to particle transport problems , where the label defines the physical properties of each polyhedron @xcite .
the case of random binary mixtures , where only two labels are allowed , will be examined in detail . in this context
, we will numerically determine the statistical features of the coloured polyhedra , which are obtained by regrouping into clusters the neighbouring volumes by their common label .
attention will be paid in particular to the percolation properties of such binary mixtures for @xmath1 : the percolation threshold at which a cluster will span the entire geometry , the average cluster size and the probability that a polyhedron belongs to the the spanning cluster will be carefully examined and contrasted to the case of percolation on lattices @xcite .
the effect of dimensionality will be again assessed by comparison with the case @xmath4 , for which analogous results were numerically determined in @xcite .
this paper is structured as follows : in sec .
[ construction ] we will recall the explicit construction for @xmath0-dimensional isotropic poisson geometries , with focus on @xmath1 . in sec .
[ uncolored_geo ] we will discuss the statistical properties of poisson geometries , and assess the convergence to the limit case of infinite domains . in sec .
[ colored_geo ] we will extend our analysis to the case of coloured geometries and related percolation properties .
conclusions will be finally drawn in sec .
[ conclusions ] .
for the sake of completeness , in this section we will recall the strategy for the construction of poisson geometries , spatially restricted to a @xmath0-dimensional box .
the case @xmath5 simply stems from the poisson point process on the line @xcite , and will not be detailed here .
the explicit construction of homogeneous and isotropic poisson geometries for the case @xmath4 restricted to a square has been originally proposed by @xcite , based on a poisson point field in an auxiliary parameter space in polar coordinates .
it has been recently shown that this construction can be actually extended to @xmath1 and even higher dimensions @xcite by suitably generalizing the auxiliary parameter space approach of @xcite and using the results of @xcite .
in particular , such @xmath0-dimensional construction satisfies the homogeneity and isotropy properties @xcite .
the method proposed by @xcite is based on a spatial decomposition ( tessellation ) of the @xmath0-hypersphere of radius @xmath6 centered at the origin by generating a random number @xmath7 of @xmath3-hyperplanes with random orientation and position .
any given @xmath0-dimensional subspace included in the @xmath0-hypersphere will therefore undergo the same tessellation procedure , restricted to the region defined by the boundaries of the subspace .
the number @xmath7 of @xmath3-hyperplanes is sampled from a poisson distribution with parameter @xmath8 , with @xmath9 . here
@xmath10 denotes the surface of the @xmath0-dimensional unit sphere ( @xmath11 being the gamma function @xcite ) , @xmath12 denotes the volume of the @xmath0-dimensional unit sphere , and @xmath13 is the arbitrary density of the tessellation , carrying the units of an inverse length .
this normalization of the density @xmath13 corresponds to the convention used in @xcite , and is such that @xmath14 yields the mean number of @xmath3-hyperplanes intersected by an arbitrary segment of length @xmath15 .
is centered in @xmath16 .
the circumscribed sphere centered in @xmath16 has a radius @xmath17 .
the point @xmath18 is defined by @xmath19 , where @xmath20 is uniformly sampled in the interval @xmath21 $ ] and @xmath22 is a random unit vector of components @xmath23 , with @xmath24 , @xmath25 and @xmath26 .
the auxiliary variables @xmath27 and @xmath28 are sampled from two independent uniform distributions in the interval @xmath29 $ ] .
the random plane @xmath30 of equation @xmath31 is orthogonal to the vector @xmath22 and intersects the point @xmath18 . ]
let us now focus on the case @xmath1 .
suppose , for the sake of simplicity , that we want to obtain an isotropic tessellation of a box of side @xmath2 , centered in the origin @xmath16 .
this means that the poisson tessellation is restricted to the region @xmath32 ^ 3 $ ] .
we denote @xmath6 the radius of the sphere circumscribed to the cube . the algorithm proceeds then as follows .
the first step consists in sampling a random number of planes @xmath7 from a poisson distribution of parameter @xmath33 , where the factor @xmath34 stems from @xmath35 .
the second step consists in sampling the random planes that will cut the cube .
this is achieved by choosing a radius @xmath20 uniformly in the interval @xmath21 $ ] and then sampling two other random numbers , denoted @xmath27 and @xmath28 , from two independent uniform distributions in the interval @xmath29 $ ] .
based on these three random parameters , a unit vector @xmath23 is generated ( see fig .
[ fig1 ] ) , with components @xmath36 let now @xmath18 be the point such that @xmath37 .
the random plane @xmath30 will be finally defined by the equation @xmath38 , passing trough @xmath18 and having normal vector @xmath22 . by construction ,
this plane does intersect the circumscribed sphere of radius @xmath6 but not necessarily the cube : the probability that the plane intersects both the sphere and the cube can be deduced from a classical result of integral geometry . for two convex sets @xmath39 and @xmath40 in @xmath41 , with @xmath42 , the probability that a randomly chosen plane meets both @xmath39 and @xmath40 is given by the ratio @xmath43 , @xmath44 being the mean orthogonal @xmath45-projection of @xmath46 onto an isotropic random line @xcite .
the quantity @xmath44 takes also the name of mean caliper diameter of the set @xmath46
@xcite .
the average caliper diameter of a cube of side @xmath2 is @xmath47 , whereas for the sphere the average caliper diameter coincides with its diameter @xmath48 , which yields a probability @xmath49 for the random planes to fall within the cube , the probability that a random line intercepts both square of side @xmath2 and the circumscribed circle of radius @xmath50 is again given by the ratio of the respective mean caliper diameters , which for @xmath4 are simply proportional to the perimeters of each set ( the so - called barbier - crofton theorem ) .
this yields a probability @xmath51 for a random line to fall within the square @xcite . ] .
the tessellation is built by successively generating the @xmath7 random planes .
initially , the stochastic geometry is composed of a single polyhedron , i.e. , the cube .
if the first sampled plane intersects the region @xmath32 ^ 3 $ ] , new polyhedra are generated within the cube and the tessellation is updated .
this procedure is then iterated until @xmath7 random planes have been generated . by construction , the polyhedra defined by the intersection of such random planes are convex . for illustration purposes , some examples of isotropic poisson tessellation of a cube of side @xmath52 obtained by monte carlo simulation are presented in fig .
[ fig2 ] , for different values of the density @xmath13 .
the number of random polyhedra of the tessellation increases with increasing @xmath13 .
the physical observables of interest associated to the stochastic geometries , such as for instance the volume of a polyhedron , its surface , the number of edges , and so on , are clearly random variables , whose statistical distribution we would like to characterize . in the following , we will focus on the case of poisson geometries restricted to a @xmath0-dimensional box of linear size @xmath2 . . for all realizations ,
we have chosen @xmath52 .
the geometry at the top ( a ) has @xmath53 , that at the middle ( b ) has @xmath54 and that at the bottom ( c ) has @xmath55 . for fixed @xmath2 , the average number of random polyhedra composing the geometry increases with increasing @xmath13 . ] with a few remarkable exceptions , the exact distributions for the physical observables are unfortunately unknown @xcite .
a number of exact results have been nonetheless established for the ( typically low - order ) moments of the observables and for their correlations , at least in the limit case of domains having an infinite extension @xcite .
monte carlo simulation offers a unique tool for the numerical exploration of the statistical features of poisson geometries .
in particular , by resorting to the algorithm described above we can @xmath56 investigate the convergence of the moments and distributions of arbitrary physical observables to their known limit behaviour ( if any ) , and @xmath57 numerically explore the scaling of the moments and the distributions for which exact asymptotic results are not yet available .
we will thus address these issues with the help of a monte carlo code developed to this aim . to begin with
, we will first analyse the growth of the number @xmath58 of polyhedra in @xmath0-dimensional poisson geometries as a function of the linear size @xmath2 of the domain , for a given value of the density @xmath13 . in the following
, we will always assume that @xmath54 , unless otherwise specified ( with both @xmath13 and @xmath2 expressed in arbitrary units ) .
the quantity @xmath58 provides a measure of the complexity of the resulting geometries .
the simulation findings for the average number @xmath59 of @xmath0-polyhedra ( at finite @xmath2 ) and the dispersion factor , i.e. , the ratio @xmath60/\langle n_p|l \rangle$ ] , @xmath61 denoting the standard deviation , are illustrated in fig .
[ fig3 ] . for large @xmath2
, we find an asymptotic scaling law @xmath62 : the complexity of the random geometries increases with system size and dimension ( fig .
[ fig3 ] , top ) , as expected on physical grounds .
this means that the computational cost to generate a realization of a poisson geometry is also an increasing function of the system size and of the dimension . as for the dispersion factor , an asymptotic scaling law @xmath60/\langle n_p|l\rangle \sim 1/\sqrt{l}$ ]
is found for large @xmath2 , independent of the dimension ( fig .
[ fig3 ] , bottom ) : for large systems , the distribution of @xmath58 will be then peaked around the average value @xmath59 .
poisson geometries are markovian , which means that in the limit case of infinite domains an arbitrary line will be cut by the @xmath3-surfaces of the @xmath0-polyhedra into segments whose lengths @xmath63 are exponentially distributed , i.e. , @xmath64 with average density @xmath65 .
conversely , the number of intersections @xmath66 of an arbitrary segment of length @xmath15 with the @xmath3-surfaces of the @xmath0-polyhedra in an infinite domain will obey a poisson distribution @xmath67 with mean value @xmath68 .
of @xmath0-polyhedra in @xmath0-dimensional poisson geometries as a function of the linear size @xmath2 of the domain .
the scaling law @xmath69 is displayed for reference with dashed lines .
the dispersion factor @xmath60/\langle n_p|l \rangle$ ] as a function of @xmath2 .
the scaling law @xmath70 is displayed for reference with dashed lines . ] in order to verify that the geometries constructed by resorting to the algorithm described above satisfy the markov properties , we have numerically computed by monte carlo simulation the probability density of the segment lengths and the probability of the number of intersections as a function of the linear size @xmath2 of the domain and for different dimensions @xmath0 . for the former , a poisson geometry
is first generated , and a line is then drawn by uniformly choosing a point in the box and an isotropic direction : this choice corresponds to formally assuming a so - called @xmath71-randomness for the lines @xcite .
the intersections of the line with the polyhedra of the geometry are computed , and the resulting segment lengths are recorded .
the whole procedure is repeated a large number of times in order to get the appropriate statistics .
for the latter , a test segment of unit length is sampled by choosing a point and a direction as before , and the number of intersections with the polyhedra are again determined . of the segment lengths as a function of the linear size @xmath2 of the domain , in dimension @xmath1 .
symbols correspond to the monte carlo simulation results , with lines added to guide the eye : blue triangles denote @xmath72 , red diamonds @xmath73 , green circles @xmath74 , black squares @xmath75 , and purple crosses @xmath76 .
the asymptotic ( i.e. , @xmath77 ) exponential distribution given in eq .
is displayed as a black dashed line for reference .
the inset displays the same data in log - linear scale . ]
the numerical results for @xmath78 at finite @xmath2 are illustrated in figs .
[ fig4 ] and [ fig5 ] . for small @xmath2 ,
finite - size effects are apparent in the segment length density : this is due to the fact that the longest line that can be drawn across a box of linear size @xmath2 is @xmath79 , which thus induces a cut - off on the distribution ( see fig . [ fig4 ] ) . for @xmath80 ,
the finite - size effects due to the cut - off fade away and the probability densities eventually converge to the expected exponential behaviour .
the rate of convergence appears to be weakly dependent on the dimension @xmath0 ( see fig . [ fig5 ] ) .
the case @xmath5 can be treated analytically and might thus provide a rough idea of the approach to the limit case . for any finite @xmath2 ,
the distribution of the segment lengths for @xmath5 is @xmath81 @xmath82 being the marker function of the domain @xmath46 .
the moments of order @xmath83 of the segment length @xmath63 for finite @xmath2 thus yield @xmath84 where @xmath85 is the incomplete gamma function @xcite . in the limit case
@xmath77 , we have @xmath86 , so that for the convergence rate we obtain @xmath87 which for large @xmath80 gives @xmath88 thus , the average segment length ( @xmath89 ) converges exponentially fast to the limit behaviour , whereas the higher moments ( @xmath90 ) converge sub - exponentially with power - law corrections . for @xmath91 ,
the cut - off is less abrupt , but the distributions @xmath78 still show a peak at @xmath92 , and vanish for @xmath93 .
the asymptotic average segment lengths for @xmath77 yield @xmath94 for any @xmath0 : the monte carlo simulation results obtained for a large @xmath95 are compared to the theoretical formulas in tab .
[ tab1 ] .
p. barthelemy , j. bertolotti , and d. s. wiersma , nature * 453 * , 495 ( 2009 ) .
t. svensson , k. vynck , m. grisi , r. savo , m. burresi , and d. s. wiersma , phys .
e * 87 * , 022120 ( 2013 ) .
t. svensson , k. vynck , e. adolfsson , a. farina , a. pifferi , and d. s. wiersma , phys .
e * 89 * , 022141 ( 2014 ) .
a. b. davis and a. marshak , j. quant
. spectr .
transfer * 84 * , 3 - 34 ( 2004 ) .
a. b. kostinskiand r. a. shaw , j. fluid mech .
* 434 * , 389 ( 2001 ) .
f. malvagi , r. n. byrne , g. c. pomraning , and r. c. j. somerville , j. atm .
sci . * 50 * , 2146 - 2158 ( 1992 ) .
v. tuchin , _ tissue optics : light scattering methods and instruments for medical diagnosis _ ( spie press , cardiff , 2007 ) .
e. w. larsen and r. vasques , j. quant .
spectrosc .
transfer * 112 * , 619 ( 2011 ) .
o. zuchuat , r. sanchez , i. zmijarevic , and f. malvagi , j. quant .
transfer * 51 * , 689 - 722 ( 1994 ) .
g. b. zimmerman and m. l. adams , trans .
soc . * 63 * , 287 - 288 ( 1991 ) .
o. haran , d. shvarts , and r. thieberger , phys .
e * 61 * , 6183 - 6189 ( 2000 ) .
n. mercadier , w. guerin , m. chevrollier , and r. kaiser , nature physics * 5 * , 602 ( 2008 ) .
l. a. santal , _ integral geometry and geometric probability _ ( addison - wesley , reading , ma , 1976 )
. s. torquato , _ random heterogeneous materials : microstructure and macroscopic properties _ ( springer - verlag , new york , 2002 )
. s. n. chiu , d. stoyan , w. s. kendall , and j. mecke , _ stochastic geometry and its applications _
( wiley , 2013 ) .
h. solomon , _ geometric probability _
( siam press , philadelphia , pa , 1978 ) .
m. g. kendall and p. a. p.moran , _ geometrical probability _ ( charles griffin and company limited , london , 1963 ) .
d. ren , _ topics in integral geometry _
( world scientific , singapore , 1994 ) .
g. c. pomraning , _ linear kinetic theory and particle transport in stochastic mixtures _ ( world scientific , 1991 ) .
j. serra , _ image analysis and mathematical morphology _
( academic press , london , 1982 ) .
e. underwood , _ quantitative stereology _
( addison - wesley , 1970 ) .
a. yu . ambos and g. a. mikhailov , russ .
j. numer .
modelling * 26 * , 263 - 273 ( 2011 ) . s. goudsmit , rev .
* 17 * , 321 ( 1945 ) .
r. e. miles , proc .
usa * 52 * , 901 - 907 ( 1964 ) .
r. e. miles , proc .
usa * 52 * , 1157 - 1160 ( 1964 ) .
p. i. richards , proc .
usa * 52 * , 1160 - 1164 ( 1964 ) .
p. switzer , ann .
statist . * 36 * , 1859 - 1863 ( 1965 ) .
r. e. miles , adv .
* 1 * , 211 - 237 ( 1969 ) .
r. e. miles , izv .
nauk arm .
ssr ser . mat .
* 5 * , 263 - 285 ( 1970 ) .
r. e. miles , adv .
prob . * 3 * , 1 - 43 ( 1969 ) .
r. e. miles , suppl . adv . appl .
* 4 * 243 - 266 ( 1972 ) .
g. matheron , adv .
* 4 * 508 - 541 ( 1972 ) .
d. stauffer and a. aharony , _ introduction to percolation theory _
( crc press , 1994 ) .
t. lepage , l. delaby , f. malvagi , and a. mazzolo , prog .
* 2 * , 743 - 748 ( 2011 ) .
f. w. j. olver , d. w. lozier , r. f. boisvert , and c. w. clark , _ nist handbook of mathematical functions _
( cambridge university press , cambridge , 2010 ) .
r. coleman , j. appl .
* 6 * , 430 - 441 ( 1969 ) .
k. k. sahu and a. k. lahiri , phil . mag . *
84 * , 1185 - 1196 ( 2004 ) . w. h. press , s. a. teukolsky , w. t. vetterling , and b. p. flannery , _ numerical recipes in c. the art of scientific computing _
( cambridge university press , cambridge , 2002 ) .
k. fischer , b. grtner , and m. kutz , _ proc .
11th european symposium on algorithms ( esa ) _ , 630 - 641 ( 2003 ) .
a. m. becker and r. m. ziff , phys .
e * 80 * , 041101 ( 2009 ) .
r. neher , k. mecke , and h. wagner , j. stat .
p01011 ( 2008 ) .
b. bollobs and o. riordan , probab .
theory relat .
fields * 136 * , 417 - 468 ( 2006 ) .
p. calka , adv .
appl . prob .
* 35 * , 551 - 562 ( 2003 ) .
p. calka , j. stat.phys .
* 132 * , 627 - 647 ( 2008 ) . h. j. hilhorst and p. calka , j. stat.phys . *
132 * , 627 - 647 ( 2008 ) .
m. e. j. newman and r. m. ziff , phys .
lett . * 85 * , 4104 - 4107 ( 2000 ) .
p. grassberger , j. phys .
a * 25 * , 5867 - 5888 ( 1992 ) .
m. d. rintoul and s. torquato , j. phys . a : math .
gen . * 30 * , l585 ( 1997 ) .
j. quintanilla , s. torquato , and r. m. ziff , j. phys .
a : math . gen . * 33 * , l399-l407 ( 2000 ) .
j. c. tanner , j. app .
* 20 * , 778 - 787 ( 1983 ) . | random tessellations of the space represent a class of prototype models of heterogeneous media , which are central in several applications in physics , engineering and life sciences . in this work
, we investigate the statistical properties of @xmath0-dimensional isotropic poisson geometries by resorting to monte carlo simulation , with special emphasis on the case @xmath1 .
we first analyse the behaviour of the key features of these stochastic geometries as a function of the dimension @xmath0 and the linear size @xmath2 of the domain .
then , we consider the case of poisson binary mixtures , where the polyhedra are assigned two ` labels ' with complementary probabilities . for this latter class of random geometries ,
we numerically characterize the percolation threshold , the strength of the percolating cluster and the average cluster size . |
homogeneous cosmological models have successfully explained many important features of our universe .
however , we know that the distribution of matter is not homogeneous and so these models are only an approximation . in the last decade modelling of inhomogeneity in cosmology
has become a topic of substantial interest .
there are several different approaches to the problem . due to the nonlinear nature of einstein equations one should not rely on perturbation theory completely therefore exact models with nonuniform distribution of matter
should be considered as well . among
the most studied exact inhomogeneous solutions of einstein equations belong lemaitre - tolman - bondi ( ltb ) metric @xcite , szekeres solution @xcite , szafron family of solutions @xcite , stephani solution @xcite or lemaitre metric @xcite , which is a generalization of ltb for a fluid with nonzero pressure .
an overview of inhomogeneous cosmological solutions can be found in @xcite .
ltb metric was extensively studied by krasinski and hellaby to model structure formation @xcite - @xcite .
together with bolejko they used the solution to describe formation of voids in the universe @xcite .
ltb metric is a special case of the szekeres solution that was discovered by szekeres @xcite and was developed by hellaby and krasinski in a series of papers @xcite - @xcite .
they gave a geometrical interpretation of the metric functions in all three different types of geometry that are quasi - spherical , quasi - pseudospherical and quasi - planar .
the quasi - spherical case is currently the best understood of all three types .
it has found a cosmological application in the study of bolejko who focused on structure formation @xcite , @xcite and constructed models of a void with an adjourning supercluster .
bolejko also investigated buchert averaging of the quasispherical szekeres metric @xcite , constructed a szekeres - swiss - cheese model that he used to estimate the impact of inhomogeneity on the propagation of light @xcite and cmb observations @xcite .
walters and hellaby showed in a recent paper @xcite how to model inhomogeneity in quasispherical szekeres metric using initial and final data .
they constructed three models where they specified initial and final radial density profile and one extreme value of the density on the final time slice in terms of a deviation function along with the position of the extreme . in this way they set up all 5 degrees of freedom in
the szekeres metric and worked out an algorithm to calculate all other metric functions in the model . for practical purposes
it would be desirable to consider the evolution of inhomogeneities in szekeres spacetime determined only by initial data . in our approach
we specify the curvature function @xmath2 and the radial density profile with one extreme value at the initial time .
thus we set up only 3 degrees of freedom which is however sufficient to model the density contrast .
to make an appropriate choice for the initial density extreme we investigate the shell crossing conditions in terms of the density extremes and show that the resulting constraints are preserved throughout the evolution ( for the assumed value of @xmath2 ) .
the paper is organized as follows . in the next section
we give an overview of the szekeres metric and describe its basic properties . in section 3
we derive conditions that the density extremes have to satisfy in order to avoid a shell crossing singularity . in section 4
we show that if these conditions are met at the initial time , then they are fulfilled at later times if the curvature function vanishes . in section 5
we derive an analytical formula for the density contrast defined as the difference of the extreme values and for the case @xmath3 we show that if the initial radial inhomogeneity is small , then the density contrast is proportional to @xmath4 and the radial dependence of the function does not evolve substantially in time . in section 6 we set up the model by specifying two more functions and show the time evolution of the density contrast in a specific example .
we conclude in section 7 .
the szekeres spacetime is an exact dust solution of einstein equations without any symmetries .
it was found by szekeres @xcite and generalized by szafron @xcite for an energy momentum tensor describing a perfect fluid .
the lt type of szekeres metric can be written as @xcite @xmath5 where @xmath6,\ ] ] @xmath7-\frac{p'}{s}\left(p - p\right)-\frac{q'}{s}\left(q - q\right)\ ] ] and @xmath2 , @xmath8 , @xmath9 , @xmath10 are arbitrary functions of @xmath1 .
the parameter @xmath11 can have only three values @xmath12 and it determines the geometry of the two - spaces of constant @xmath0 and @xmath1 . from einstein equations it follows a dynamical equation for the function @xmath13 @xmath14 and an equation for the density evolution @xmath15 where prime denotes a derivative with respect to @xmath1 , @xmath16 is another arbitrary function and @xmath17 is cosmological constant .
so we have two einstein equations ( [ 11 ] ) , ( [ 12 ] ) and by solving the first one we obtain one more arbitrary function @xmath18 , which enters the solution in the form @xmath19 .
from now on we will use a redefined form of the density @xmath20 and for the sake of simplicity we will drop the bar . as mentioned above
, the metric does not have any symmetries , in other words it has no killing vectors .
nevertheless the three spaces of constant @xmath0 are conformally flat @xcite .
the parametrization of the metric involves 6 arbitrary functions however the number of physical degrees of freedom is 5 , because we can still rescale the radial coordinate ( the metric is invariant with respect to the transformation @xmath21 ) . the sign of the function @xmath2 effects the solution of the einstein equation ( [ 11 ] ) .
it is a dynamical equation for @xmath13 and it looks very similar to the friedman equation , except that here the functions @xmath2 and @xmath16 depend on @xmath1 .
we will assume only the case when @xmath23 .
there are three different types of evolution @xcite . if @xmath24 the evolution is elliptic which means that the universe is first expanding and at some point the expansion stops and the universe is collapsing to a final singularity .
the solution is given in a parametric form @xmath25 for @xmath26 the evolution is hyperbolic in the sense that the sign of the expansion does not change , the universe either expands or collapses depending on initial conditions and the parametric solution to the equation ( [ 11 ] ) is @xmath27 for @xmath3 the evolution is parabolic and is given by equation @xmath28 since @xmath2 is a function of @xmath1 , the universe may have different evolution in different regions .
the parabolic evolution can be on the boundary between two regions , one having elliptic evolution and the other one hyperbolic .
so the sign of @xmath2 determines the sign of scalar curvature of the three - spaces of constant @xmath0 and when @xmath29 , the three - spaces are flat which corresponds to a presence of pure decaying modes as was shown in @xcite .
the function @xmath22 is called the bang time function , because @xmath30 is the moment when big bang happened .
so unlike in homogeneous models , the initial moment of evolution is position dependent .
the interpretation of the other metric functions ( @xmath13 , @xmath16 , @xmath8 , @xmath9 , @xmath10 ) depends on @xmath11 and from now on we will only consider the case @xmath31 which is often called the quasi - spherical case . we can make a coordinate transformation @xcite @xmath32 and rewrite the induced two - metric of the surfaces of constant @xmath0 and @xmath1 into coordinates @xmath33 and @xmath34 . after applying the transformation we get @xmath35 which is a metric on a unit sphere .
the transformation ( [ 16 ] ) is nothing but a stereographic projection and the function @xmath36 describes how the @xmath37 plane is mapped onto a unit sphere .
every sphere is multiplied by @xmath38 and the function is sometimes called the areal radius , because it is actually the radius of a sphere on a comoving coordinate @xmath1 . in order to understand more about the geometrical properties of the metric ( [ 9 ] ) it is important to investigate the function @xmath36 .
we can see that this function appears in the metric and in the equation for the density ( [ 12 ] ) in the form @xmath39 .
particularly , we can investigate when the function is equal to zero and what are its extreme values . using the transformation ( [ 16 ] ) in ( [ 10 ] ) and ( [ 25 ] ) we can write @xmath39 as @xmath40 the equation @xmath41 now becomes @xmath42 and after realizing that @xmath43 , @xmath44 and @xmath45 , the equation ( [ 28 ] ) becomes @xmath46 which is an equation of a plane that goes through the origin of the spherical coordinate system and intersects the @xmath47 sphere in a great circle . now calculating the derivatives of ( [ 31 ] ) with respect to @xmath33 and @xmath34 and putting it equal to zero we can find that there are two extreme values and they are located at opposite sites on the sphere , in other words if one extreme is at the coordinates @xmath48 , the other one is at @xmath49 and they are located symmetrically with respect to the plane @xmath41 , because the coordinates of the extremes are exactly the components of the unit normal to the plane @xmath41 .
the values of the extremes are @xmath50 where max and min refers to maximum and minimum respectively .
so the extremes have opposite values and the function @xmath39 behaves on the sphere like a dipole @xcite .
we are now going to study the formula for the density ( [ 12 ] ) .
we can see that under certain circumstances the density can change sign or possibly diverge if the denominator becomes zero .
those points where this happens are called the shell crossing singularity .
the conditions that the metric functions @xmath16 , @xmath2 , @xmath51 and @xmath13 have to satisfy in order to avoid shell crossing singularity can be found in @xcite . in this section
we will derive the conditions that the extreme values of the density @xmath52 and @xmath53 have to satisfy in order to avoid the shell crossing singularity . we can rewrite ( [ 12 ] ) as @xmath54 where we define @xmath55 and @xmath56 here
@xmath57 is the density that we get from ( [ 12 ] ) if we set @xmath41 .
it is also a radial density in the sense that on a given @xmath1 it is the value of the density around the great circle that lies in the plane that defines the dipole on the sphere .
the index @xmath58 refers to lemaitre - tolman - bondi metric , because its density is given with exactly the same formula as ( [ 35 ] ) .
if we make the choice @xmath59 and use the fact that the function @xmath16 depends only on @xmath1 , we can express it on the initial time slice when @xmath41 as @xmath60 where the index @xmath61 refers to the value of the function at the initial time @xmath62 .
in that case the definitions @xmath63 and @xmath64 can be further rewritten as @xmath65 the initial condition ( [ ic ] ) and formulas ( [ 37 ] ) - ( [ den ] ) will however not be used in the following calculations ( with the exception of equation ( [ 68 ] ) giving critical point position ) and the main results of this section are independent of it , thus giving us generally applicable criteria .
it is good to take a look at the derivative of the density with respect to @xmath39 @xmath66 from the derivative we can see , that if @xmath67 the derivative is positive and the density is growing as @xmath39 increases . on the other hand
if @xmath68 the density will decrease as @xmath39 increases . either way we can see that the density behaves on each sphere also like a dipole in the sense that it has two extreme values , maximum and minimum and they are located at the same position as extreme values of the function @xmath39 . in the case when @xmath68 we can write for the density maximum @xmath69 and for the minimum @xmath70 in the case when @xmath67 the role of @xmath71 interchanges with @xmath72 in the last two equations .
we split the investigation of the density into three parts , in the first part we investigate when both the numerator and the denominator are positive , in the second case we check when they are both negative and in the last case we take a look at the special case when they are both zero
. it will be convenient to split the first part into three more subcases depending on the value of @xmath73 .
for the investigation we will also need a formula for the function @xmath39 given in terms of @xmath57 and @xmath74 which we can derive from ( [ 34 ] ) @xmath75 first case : numerator and denominator of the density are both positive .
+ subcase a : @xmath68 from the positivity of the denominator in ( [ 34 ] ) we get @xmath76 because as follows from ( [ 38 ] ) if @xmath68 we can see that @xmath39 has maximum where the density has minimum . from the dipole property @xmath77 we know that @xmath71 is non - negative and ( [ 42 ] ) implies that @xmath78 and therefore we have to consider @xmath79 .
we can now set conditions for @xmath53 so that the right hand side in ( [ 42 ] ) is nonnegative . inequality ( [ 42 ] ) simplifies to @xmath80 it is reasonable to require @xmath81 , because the minimum should be the smallest value on each sphere ( @xmath82 corresponds to the case when the density is homogeneously distributed on the whole sphere ) . since we consider
@xmath79 it follows that @xmath83 and the right hand side of ( [ 42 ] ) is nonnegative .
it means that besides the obvious condition @xmath81 we do nt have any other restriction for the minimum .
due to the dipole property we can try to derive a condition for the maximum @xmath84 since @xmath52 should be greater than or equal to @xmath57 it implies that @xmath85 in order for the right hand side in ( [ 45 ] ) to be non - negative .
this condition is reasonable since @xmath74 is not true value of the density as follows from its definition ( [ 36 ] ) .
taking this into consideration we can solve the inequality ( [ 45 ] ) and the solution is @xmath86 and this condition is clearly more restrictive than @xmath85 .
for the numerator of ( [ 34 ] ) we need @xmath87 which can be rewritten as @xmath88 it again follows that @xmath78 and @xmath79 .
the right - hand side of ( [ 48 ] ) has to be non - negative and we know that @xmath81 and since @xmath79 we also have @xmath83 so the inequality ( [ 48 ] ) is always true and we do nt get any new condition for @xmath53 .
we now use the property ( [ 44 ] ) and look for a condition for @xmath52 @xmath89 we know that @xmath90 so to make the right hand side non - negative we need @xmath85 . the solution to the inequality ( [ 49 ] )
is @xmath91 the conditions for the denominator and the numerator have to be satisfied simultaneously , so for the maximum we have ( [ 46 ] ) and ( [ 50 ] ) and ( as can be seen from figure 1 ) the condition ( [ 50 ] ) is stronger in the case of @xmath79 and has to be fulfilled in order for the density to be positive .
as far as the minimum is concerned , apart from the obvious condition @xmath81 it is not constrained at all .
first case : numerator and denominator of the density are both positive .
+ subcase b : @xmath67 the subcase b will be investigated similarly .
the only difference is that from ( [ 38 ] ) we know that if @xmath68 the function @xmath39 has maximum where the density has maximum and this modifies ( [ 39 ] ) and ( [ 40 ] ) accordingly . from the denominator of ( [ 34 ] ) we have @xmath92 which again implies @xmath78 and @xmath93 because @xmath94 and it simplifies to @xmath95 the analysis is basically the same , we again need the right - hand side to be non - negative . we know
that @xmath90 and therefore @xmath96 as well , so the inequality ( [ 52 ] ) will be always true as long as @xmath93 and we do nt get any further condition for the maximum . for the minimum we have @xmath97 in order for the right - hand side to be non - negative we need @xmath98 which is a valid condition for the density minimum , because as mentioned above @xmath74 is not true value of the density .
the inequality ( [ 53 ] ) has the solution @xmath99 from the numerator of ( [ 34 ] ) we have the condition @xmath100 it again implies @xmath78 and @xmath93 and it does not give us any constraint for the maximum , because we know that @xmath90 and therefore @xmath96 .
so ( [ 55 ] ) holds as long as @xmath78 and @xmath93 . for the minimum we can write @xmath101 the right - hand side will be positive if @xmath98 and the solution to ( [ 56 ] ) is @xmath102 all together in the case of @xmath67 we have two conditions ( [ 54 ] ) and ( [ 57 ] ) for the minimum and as can be seen from figure 1 the first one is stronger . for the maximum , except for the obvious @xmath90 , we do nt have any constraint so the density will be positive as long as ( [ 54 ] ) holds and @xmath90 . , @xmath103 , @xmath104 , @xmath57 .
] first case : numerator and denominator of the density are both positive .
+ subcase c : @xmath105 before we consider this special case we can first rewrite ( [ 34 ] ) as @xmath106 now using @xmath105 we can see that the formula for the density simplifies @xmath107 in other words the value of the density is independent of @xmath33 and @xmath34 and the density is homogeneously distributed on the whole sphere as follows also from ( [ 38 ] ) because the derivative of the density is now zero .
using the same argument as in subcase a and b we have @xmath78 and therefore @xmath108 . for the extreme values of the density
we obviously need in this case @xmath109 the point where @xmath108 is especially interesting because the shell - crossing conditions for @xmath52 and @xmath53 change here and the location of @xmath52 and @xmath53 interchanges by passing through the plane @xmath41 .
the position of this critical point @xmath110 generally evolves and we can get its value as the solution to the equation @xmath111 where we used the initial condition ( [ ic ] ) and formulas ( [ den ] ) in order to express @xmath57 and @xmath74 . so the time evolution of the critical point will depend on the time evolution of the function @xmath112 .
second case : numerator and denominator of the density are both negative . in order for the denominator of the density ( [ 34 ] ) to be negative ,
we need @xmath113 this condition will not be fulfilled unless @xmath114 because @xmath115 .
the investigation is than similar as in the first case and leads to the same conditions for @xmath53 and @xmath52 as can be easily verified . on the other hand
if @xmath78 and we allow the denominator to be negative on a given @xmath1 , it will be also positive on that @xmath1 for some specific @xmath33 and @xmath34 .
this behaviour may or may not be correct , depending on the behaviour of the numerator and specifically depending on whether or not the numerator changes sign at the same @xmath33 , @xmath34 . for the numerator we have inequality
@xmath116 this can be rewritten as @xmath117 the same argument as in the case of denominator tells us that this will be satisfied only if @xmath114 .
so if the numerator is negative on a given @xmath1 but @xmath78 , there is a region on this @xmath1 where it is positive too . for the positiveness of the denominator in a region
@xmath118 we have @xmath119 and for the numerator the condition is @xmath120 clearly the conditions are not the same unless @xmath108 , so except for this special case , there will always be a region on a given @xmath1 where the numerator and the denominator will have different signs .
the places where the sign changes are in both cases circles on the sphere , that are parallel to the great circle that defines the plane of the dipole .
the order in which the circles go are seen in figure 2 , where the parallel circles are mapped as parallel lines .
so if in the second case @xmath78 there are no conditions for the density extremes that would prevent shell crossing .
third case : numerator and denominator of the density are both zero .
the denominator is zero when @xmath121 but this equation will be satisfied on a given @xmath1 for all @xmath33 and @xmath34 only if both @xmath112 and @xmath39 are zero .
this consequently means that @xmath122 .
the same arguments are valid for the numerator of the density .
as was discussed in @xcite and in the context of ltb metric in @xcite this point will not be a singularity if also @xmath123 .
since @xmath124 and @xmath125 are independent of time it implies that @xmath126 can not evolve and we need stationary configuration in order to avoid shell crossing .
[ tab1 ] @xmath52 + + @xmath127 & @xmath128 + + @xmath129 & @xmath130 + + @xmath82 & @xmath131 + the table 1 summarizes the conditions for the extreme values of the density that we obtained so that shell crossing would be avoided .
from the dipole property ( [ 44 ] ) it follows that there is a constraint between the density extremes @xmath132 which means that knowing one extreme value allows us calculate the other one with a formula @xmath133 that follows from ( [ 69 ] ) . and
for the other extreme we just interchange min for max in the last equation .
consequently it is sufficient to ensure that just one of the extremes satisfies the conditions in table 1 and the constraint ( [ 69 ] ) ensures that the other density extreme has a value that does not break the conditions in table 1 .
in the previous section we derived conditions that @xmath52 and @xmath53 have to satisfy in order to avoid shell crossing singularity . to find out
if a shell crossing occurs at a given time , we need to calculate @xmath52 or @xmath53 at that time and then check if the conditions are met . it would be useful to have conditions in terms of the initial @xmath134 and @xmath135 or @xmath136 that would ensure no shell crossing at any time during the evolution . in this section
we show that in the special case when @xmath3 , if we avoid shell crossing at the initial time , it is guaranteed that no shell crossing occurs during the time evolution .
first we show that if on the initial time slice @xmath137 , then this condition holds at any later time .
we will use the initial condition ( [ ic ] ) and start with the inequality that we want to prove @xmath138 @xmath139 this can be rewritten as @xmath140 and the inequality sign depends on the sign of @xmath126 and here we assumed @xmath78 .
we can see that the time dependence in the last inequality is hidden in the function @xmath141 . the choice @xmath3 has parabolic evolution and the solution to the equation ( [ 11 ] ) is given by ( [ 15 ] ) . if we substitute in ( [ 15 ] ) for @xmath16 from ( [ 37 ] )
we get @xmath142 we can fix the bang time function @xmath51 using ( [ ic ] ) to be @xmath143,\ ] ] where @xmath62 is the initial moment of the evolution . by calculating the radial derivative of ( [ 73 ] ) we can express @xmath141 as @xmath144 after calculating radial derivative of the bang time function ( [ 74 ] ) and substituting it in the last formula we obtain @xmath145 now we define @xmath146 and using this definition we can rewrite ( [ 76 ] ) as @xmath147 from this formula it can be shown that @xmath141 is a monotonic function of @xmath0 . for @xmath148 ,
its value is 1 , which follows from our initial condition @xmath149 and for @xmath150 , @xmath151 so it is increasing if @xmath137 and it is decreasing if @xmath152 . in any case
the function is positive which implies @xmath78 so our assumption of the inequality sign in ( [ 72 ] ) is justified . using the definition ( [ 77 ] ) and substituting for @xmath141 into ( [ 72 ] )
we get @xmath153 since @xmath154 the last inequality will be satisfied as long as @xmath155 , which is equivalent to @xmath137 as can be seen from the definition of @xmath156 .
so we can see exactly what we wanted to prove , if @xmath137 than @xmath138 at any time .
similarly we could prove that if @xmath152 than @xmath157 at any time .
next we will assume that @xmath137 and we will express the shell crossing condition @xmath158 in terms of initial data . for the function
@xmath39 we have formula ( [ 41 ] ) and since the function does not depend on time , we can express it on the initial time slice @xmath159 we plug the last formula into ( [ 40 ] ) and after some calculations we get @xmath160 and we used @xmath161 which follows from its definitions ( [ 35 ] ) and ( [ 36 ] ) .
we will now define @xmath162 and using this definition with ( [ 77 ] ) we can rewrite ( [ 81 ] ) as @xmath163 the inequality @xmath158 can now be rewritten as @xmath164 we need to multiply the last inequality by the denominator , but in order to do that we need to find out its sign . from ( [ 77 ] ) and ( [ 82 ] ) it follows @xmath165 next using the minimal value of @xmath141 to be @xmath166 we can write for the denominator of ( [ 86 ] ) @xmath167 because we assume @xmath137 , i.e @xmath155 .
so the denominator is negative and when we multiply by it in ( [ 86 ] ) , we get after some calculations @xmath168 which is inequality ( [ 72 ] ) written in terms of @xmath156 and we already proved that it holds as long as @xmath155 .
so we can see that if @xmath137 than @xmath158 is fulfilled and the shell crossing will be avoided at any time . similarly it could be shown that in the case @xmath152 , i.e. @xmath169 , if @xmath170 than @xmath171 at any time .
we can also see that if @xmath172 at some point @xmath110 , it means that we have @xmath173 and from ( [ 78 ] ) it follows @xmath174 .
so the position of the critical point @xmath110 as defined in ( [ 68 ] ) is time independent .
we will choose @xmath134 in such a way that @xmath175 .
we already expressed @xmath53 in terms of @xmath57 and @xmath135 in ( [ 83 ] ) .
similarly we can express @xmath52 in terms of @xmath134 and @xmath136 as @xmath176 where we define @xmath177 we are now interested in the time evolution of the difference of the extreme values of the density , so we define @xmath178\ \end{aligned}\ ] ] we can now substitute for @xmath13 from ( [ 73 ] ) and after some more calculations it can be rewritten in the form @xmath179 where we have defined @xmath180 , @xmath181 and we used @xmath182 which can be derived from the constraint ( [ 70 ] ) .
we will now investigate the behaviour of the function @xmath183 .
we set @xmath184 and take a look at how the function @xmath183 behaves in the case when @xmath185 , which corresponds to the situation when the radial derivative of @xmath134 is small .
first we approximate @xmath141 that is given by ( [ 78 ] ) , @xmath186 so in the numerator on the right hand side of ( [ 94 ] ) we get @xmath187 in order to simplify the denominator in ( [ 94 ] ) we set @xmath188 and @xmath189 and we assume that @xmath190 in which case @xmath191 and @xmath192 . in this case
the approximation of the denominator reads @xmath193\approx\nonumber\\ & & \quad\approx\left(c - a\right)\left(d - a\right)+\left(c-1\right)\left(d-1\right)\end{aligned}\ ] ] and for the function @xmath183 we get @xmath194 from ( [ 95 ] ) it follows @xmath195 using ( [ 91 ] ) we can rewrite the denominator in ( [ 99 ] ) @xmath196 so for @xmath183 we can write @xmath197 and for @xmath198 we have @xmath199 from the last equation we can see that if the radial derivative of @xmath200 is small then it does not effect much the shape of the initial difference of the density extremes .
also if at late time @xmath0 the bang time function is small compared to @xmath0 we can see that @xmath198 is proportional to @xmath201 .
we can try to evaluate ( [ 104 ] ) at the initial time @xmath148 @xmath202 so we can see that if a is close to 1 the approximation gives us what we expect .
the szekeres spacetime has 5 degrees of freedom , so to fully specify the model we need to set up 5 functions .
but since we are only interested in the density contrast , it is sufficient to specify just three functions , which is the curvature function @xmath2 that we assume to be zero , the initial radial profile @xmath134 and one extreme value of the density .
we will set up two models @xmath203 and @xmath204 .
for the initial radial profile we choose @xmath205 both of them are peaked at the origin and @xmath206 is the background density at the initial time .
the @xmath134 is chosen in such a way that besides the origin it is less then @xmath207 , therefore for the extreme value we specify the density minimum since the only requirement for it is that it has to be equal to @xmath134 at the origin and less everywhere else .
we choose @xmath135 as @xmath208 the models differ by the size of the radial inhomogeneity that is in the model @xmath204 twice as big as in the model @xmath203 and we want to demonstrate , that the approximation formula will work better for the model @xmath203 , because the function @xmath156 is closer to 1 than in the model @xmath204 .
the functions @xmath134 , @xmath135 , @xmath136 and @xmath207 for the model @xmath203 are shown in figure 3a .
the function @xmath156 for both models is shown in figure 3b .
the initial time is chosen as @xmath209 that approximately corresponds to the time of last scattering .
the final time @xmath210 which is approximately present time .
the initial density contrast @xmath211 and the final density contrast @xmath212 for both models are shown in figures 4 , @xmath213 denotes the background density at the final time .
the impact of the time evolution on the initial shape is minimal as we expected since @xmath156 is close to @xmath214 , on the other hand the magnitude of the density contrast drops significantly because of the factor @xmath215 in ( [ 104 ] ) .
we can also see in the figure 4 that the formula ( [ 104 ] ) approximates the density contrast better for the model @xmath203 , that corresponds to the lower peak , because we chose the radial inhomogeneity lower than in model @xmath204 .
we studied model of inhomogeneity in quasispherical szekeres model .
we set up only 3 of 5 degrees of freedom , which is sufficient for studying the evolution of the density contrast .
the lack of specification of the last two degrees of freedom means that we do not have the detailed information about the density distribution , particularly we do not know the position of the extreme values on the spheres .
we specify the initial radial density profile @xmath134 which is the value of the density around the great circle that lies in the plane that defines the dipole .
next , we choose one extreme value of the density at the initial time , either @xmath135 or @xmath136 , and the last function that we choose is the curvature function @xmath2 . in order to choose an appropriate value for the density extremes
, we investigated the shell crossing conditions in terms of density @xmath52 , @xmath53 , @xmath57 and @xmath74 .
we derived conditions that @xmath52 and @xmath216 have to satisfy in order to avoid shell crossing and we showed that in the special case @xmath3 , if the conditions are fulfilled on the initial time slice , then they will hold at any time .
next , we derived an analytical formula for the density contrast @xmath198 as a function of @xmath0 and @xmath1 . in the special case @xmath3
, we derived an approximation formula that is valid if the initial inhomogeneity is small and we showed that in this approximation the density contrast is proportional to the initial density contrast and depends on time as @xmath215 , so there is a decrease in magnitude during time evolution , however the shape of the function is preserved .
it shows that the dynamics is very simple and close to homogeneous one for small inhomogeneity confirming the expected behaviour . in this sense
one may argue that a small inhomogeneity can be successfully treated in perturbation theory and the influence of nonlinearity is negligible .
lemaitre , g. : lunivers en expansion .
bruxelles * a 53 * , 51 - 85 ( 1933 ) szekeres , p. : quasispherical gravitational collapse .
d * 12 * , 2941 - 8 ( 1975 ) szafron , d.a .
: inhomogeneous cosmologies : new exact solutions and their evolution . j. math
. phys . * 18 * , 1673 ( 1975 ) stephani , h. : ber lsungen der einsteinschen feldgleichungen , die sich in einen fnfdimensionalen flachen raum einbetten lassen .
phys . * 4 * , 137 ( 1967 ) krasinski , a. : inhomogeneous cosmological models
. cup , cambridge ( 1997 ) krasinski , a. , hellaby , ch . :
formation of a galaxy with a central black hole in the lematre - tolman model .
d * 69 * , 043502 ( 2004 ) krasinski , a. , hellaby , ch . : more examples of structure formation in the lematre - tolman model .
d * 69 * , 023502 ( 2004 ) hellaby , ch .
, krasinski , a. : alternative methods of describing structure formation in the lemaitre - tolman model .
d * 73 * , 023518 ( 2006 ) bolejko , k. , krasinski , a. , hellaby , ch . :
formation of voids in the universe within the lemaitre - tolman model .
mon . not .
. soc . * 362 * , 213 - 228 ( 2005 ) hellaby , ch . ,
krasinski , a. : you can not get through szekeres wormholes : regularity , topology , and causality in quasispherical szekeres models .
d * 66 * , 084011 ( 2002 ) hellaby , ch . ,
krasinski , a. : physical and geometrical interpretation of the szekeres models .
d * 77 * , 023529 ( 2008 ) krasinski , a. : geometry and topology of the quasiplane szekeres model . phys . rev .
d * 78 * , 064038 ( 2008 ) ; erratum phys .
d * 85 * , 069903(e ) ( 2012 ) krasinski , a. , bolejko , k. : geometry of the quasihyperbolic szekeres models
d * 86 * , 104036 ( 2012 ) bolejko , k. : structure formation in the quasispherical szekeres model .
d * 73 * , 123508 ( 2006 ) bolejko , k. : evolution of cosmic structures in different environments in the quasispherical szekeres model .
d * 75 * , 043508 ( 2007 ) bolejko , k. : volume averaging in the quasispherical szekeres model .
grav . * 41 * , 1585 - 1593 ( 2009 ) bolejko , k. , celerier , m.n . :
szekeres swiss - cheese model and supernova observations .
d * 82 * , 103510 ( 2010 ) bolejko , k. : the szekeres swiss cheese model and the cmb observations .
. grav . * 41 * , 1737 - 1755 ( 2009 ) walters , a. , hellaby , ch .
: constructing realistic szekeres models from initial and final data .
jcap * 1212 * , 001 ( 2012 ) hellaby , ch .
: the nonsimultaneous nature of the schwarzschild r=0 singularity . j. math
. phys . * 37 * , 2892 - 905 ( 1996 ) berger , b.k . ,
eardley , d.m . ,
olson , d.w .
: note on the spacetimes of szekeres .
d , * 16 * , 3086 - 9 ( 1977 ) hellaby , ch . : a kruskal - like model with finite density . class .
quantum grav .
, * 4 * , 635 - 50 ( 1987 ) sussman , r. , a. , bolejko , k. : a novel approach to the dynamics of szekeres dust models . class .
quantum grav . , * 29 * , 065018 ( 2012 ) | we study the behaviour of the density contrast in quasi - spherical szekeres spacetime and derive its analytical behaviour as a function of @xmath0 and @xmath1 .
we set up the inhomogeneity using initial data in the form of one extreme value of the density and the radial profile .
we derive conditions for density extremes that are necessary for avoiding the shell crossing singularity and show that in the special case of a trivial curvature function , the conditions are preserved by evolution .
we also show that in this special case if the initial inhomogeneity is small , the time evolution does not influence the density contrast , however its magnitude homogeneously decreases . |
throughout this paper a graph @xmath1 is finite , simple , and loopless , and we also usually assume that @xmath1 is connected .
most of our terminology and notation will be standard and can be found in any textbook on graph theory such as @xcite and @xcite . for a positive integer @xmath0 and a graph @xmath1 , the _
@xmath0-colour graph _ of @xmath1 , denoted @xmath3 , is the graph that has the proper @xmath0-vertex - colourings of @xmath1 as its vertex set , and two such colourings are joined by an edge in @xmath3 if they differ in colour on only one vertex of @xmath1 .
we now introduce a subgraph of @xmath3 , called the _ strong @xmath0-colour graph _ of @xmath1 , denoted @xmath2 .
its vertex set contains only proper @xmath0-colourings in which all @xmath0 colours actually appear , and we call such a colouring a _
strong @xmath0-colouring_. questions regarding the connectivity of a @xmath0-colour graph have applications in reassignment problems of the channels used in cellular networks ; see , e.g. , @xcite . for some applications , it is required that all channels in a range are actually used .
such a labelling is sometimes called a `` no - hole '' or `` consecutive '' labelling ; see , e.g. , @xcite . in terms of colourings
, this corresponds to a strong @xmath0-colouring . and
asking questions about the possibility to reassign channels in a cellular network can be done in such a way that all available channels are actually used .
these problems can be expressed in finding paths in the strong @xmath0-colour graph .
questions related to the connectivity of a @xmath0-colour graph have been studied extensively : @xcite . in this note
we initiate similar research on the connectivity of strong @xmath0-colour graphs : given a positive integer @xmath0 and a graph @xmath1 , is @xmath2 connected ? as an example , in figures [ f1 ] and [ f2 ] , we show the strong 3-colour graph of the paths with 4 and 5 vertices , respectively .
one is not connected while the other is connected . .
] . ]
we usually use lower case greek letters @xmath4 to denote specific colourings , and lower case latin @xmath5 to denote specific colours . to avoid trivial cases
, we will always assume that @xmath0 is greater than or equal to the chromatic number of @xmath1 , and the number of vertices of @xmath1 is at least @xmath6 .
for all @xmath7 and @xmath8 , let @xmath9 be a strong @xmath0-vertex - colouring of a complete bipartite graph @xmath10 . each colour that appears in one part of the partition can not be used in the other part .
now we choose one colour from each part and recolour the graph by swapping these two colours on each vertex coloured with one of these colours .
let @xmath11 be the resulting colouring , so @xmath11 is strong as well .
it is easy to see that there is no path in @xmath12 from @xmath9 to @xmath11 .
thus @xmath12 is not connected for all @xmath7 and @xmath8 .
this gives the following lemma .
[ z.8 ] let @xmath1 be a connected , @xmath0-colourable graph such that @xmath2 is connected , with @xmath7
. then @xmath13 , and @xmath1 does not contain a complete bipartite graph as a spanning subgraph .
[ z.6 ] let @xmath1 be a connected , @xmath0-colourable graph such that @xmath2 is connected , with @xmath7 .
suppose the graph @xmath14 is obtained from @xmath1 by adding a new vertex @xmath15 and joining it to @xmath16 vertices in @xmath17 , with @xmath18
. then @xmath19 is connected .
we will show in the next section that the strong @xmath20-colour graph of the @xmath21-vertex path , @xmath22 , is connected if and only if @xmath23 . now add a new vertex and join it to the first and the last vertex of the path , forming to an @xmath24-vertex cycle
. we will show in section [ sec4 ] that the strong @xmath20-colour graph of the @xmath21-vertex cycle , @xmath25 , is not connected for all @xmath21 .
this example shows that the restriction @xmath26 of theorem [ z.6 ] is optimal .
let @xmath27 and @xmath28 be strong @xmath0-colourings of @xmath14 .
we show that there always exists a walk in @xmath29 from @xmath27 to @xmath28 .
we say that a colouring of @xmath14 is _ good _ if all @xmath0 colours appear on @xmath17 .
first suppose that @xmath27 and @xmath28 are good . by ignoring the vertex @xmath15 ,
let @xmath9 and @xmath11 be the strong @xmath0-colourings of @xmath1 obtained from @xmath27 and @xmath28 , respectively .
since @xmath2 is connected , there is a path from @xmath9 to @xmath11 in @xmath2 .
we just follow the recolouring steps of that path to form a walk from @xmath27 to @xmath28 in @xmath29 .
the only extra steps happen when we want to recolour a neighbour @xmath30 of @xmath15 to the same colour as @xmath15 .
since @xmath31 , we can always recolour @xmath15 to a colour different from any of the colours appearing in its neighbourhood and its current colour . after recolouring @xmath15
, we can recolour @xmath30 , and continue the walk .
this walk in @xmath29 finishes in a colouring in which the vertices in @xmath17 have the same colour as they have in @xmath28 . if necessary , we can do one recolouring of @xmath15 to its colour in @xmath28 , completing the walk in @xmath29 from @xmath27 to @xmath28 .
if @xmath27 is not good , then below we show that we can always find a path in @xmath29 from @xmath27 to some good colouring ( and if necessary , we do the same for @xmath28 ) .
together with the method described in the previous paragraph , this completes the proof .
so we now assume that in @xmath27 every vertex of @xmath1 has received one of @xmath32 colours while @xmath15 has the remaining colour .
let @xmath33 be the set of vertices in @xmath17 that do not have a unique colour in @xmath1 for the colouring @xmath27 . since @xmath13 , @xmath33 is not empty . *
case 1*:there is a vertex @xmath34 not adjacent to @xmath15 .
+ by the definition of @xmath33 , there is a vertex @xmath35 such that @xmath36 and @xmath37 have the same colour in @xmath27 .
then we recolour @xmath36 to the same colour as @xmath15 .
the resulting colouring is good . *
case 2*:all vertices in @xmath33 are adjacent to @xmath15 .
+ additionally , define @xmath38 . and @xmath39 .
note that all vertices in @xmath40 have a unique colour in @xmath27 .
* subcase 2.1*:there is a vertex @xmath41 that is not adjacent to some vertex @xmath34 .
+ again , there is a vertex @xmath35 such that @xmath36 and @xmath37 have the same colour in @xmath27 .
then we first recolour @xmath36 to the same colour as @xmath42 , and then recolour @xmath42 to the same colour as @xmath15 .
again , this gives a good colouring .
* subcase 2.2*:every vertex in @xmath40 is adjacent to every vertex in @xmath33 .
+ because of lemma [ z.8 ] , @xmath43 is not empty ( otherwise , the pair @xmath44 would form the parts of a spanning complete bipartite subgraph of @xmath1 ) .
suppose there is some vertex @xmath45 that is not adjacent to some vertex @xmath34 and not adjacent to some vertex @xmath41 .
then we can recolour @xmath36 to the colour of @xmath30 ( this is possible since there is another vertex @xmath35 with the same colour as @xmath36 ) .
then recolour @xmath30 to the same colour as @xmath42 , and lastly @xmath42 to the same colour as @xmath15 .
it is easy to check that the remaining colouring is good .
so we are left with the case that each vertex in @xmath43 is adjacent to every vertex in @xmath33 or to every vertex in @xmath40 .
let @xmath46 be the set of vertices in @xmath43 that are adjacent to every vertex in @xmath33 , and @xmath47 .
then the pair @xmath48 forms the parts of a spanning complete bipartite subgraph of @xmath1 .
because of lemma [ z.8 ] , this contradicts that @xmath2 is connected .
[ z.5 ] let @xmath1 be a connected @xmath0-colourable graph so that @xmath2 is connected , with @xmath7 .
let @xmath49 be a vertex of @xmath1 with neighbourhood @xmath50 .
suppose the graph @xmath14 is obtained from @xmath1 by adding a new vertex @xmath15 and joining @xmath15 to the vertices in @xmath51 for some @xmath52 , @xmath53
. then @xmath19 is connected .
let @xmath27 and @xmath28 be strong @xmath0-colourings of @xmath14 .
we show that there always exists a walk in @xmath29 from @xmath27 to @xmath28 .
we say that a colouring of @xmath14 is _ good _ if @xmath49 and @xmath15 are labelled with the same colour .
first suppose that @xmath27 and @xmath28 are good . by ignoring the vertex @xmath15 ,
let @xmath9 and @xmath11 be the strong @xmath0-colourings of @xmath1 obtained from @xmath27 and @xmath28 , respectively .
since @xmath2 is connected , there is a path from @xmath9 to @xmath11 in @xmath2 .
we just follow the recolouring steps of that path to form a walk from @xmath27 to @xmath28 in @xmath29 .
the only extra step happens when we recolour @xmath49 . in the next step we immediately recolour @xmath15 to the same colour as @xmath49 just received .
it is easy to check that all these recolourings are allowed and give a walk in @xmath29 from @xmath27 to @xmath28 , completing the proof .
assume that @xmath27 is not good .
below we show that we always can find a path in @xmath29 from @xmath27 to some good colouring ( and if necessary , we do the same for @xmath28 ) .
together with the method described in the previous paragraph , this completes the proof .
if there is a vertex @xmath54 with the same colour as @xmath15 , then we just recolour @xmath15 to the same colour as @xmath49 .
this gives a good colouring .
so we now suppose that in @xmath27 every vertex of @xmath1 has received one of @xmath32 colours while @xmath15 has the remaining colour .
remind that @xmath49 and @xmath15 received different colours in @xmath27 .
let @xmath33 be the set of vertices in @xmath17 that did not receive a unique colour in @xmath1 for the colouring @xmath27 .
since @xmath13 , @xmath33 is not empty . *
case 1*:there is a vertex @xmath34 not adjacent to @xmath15 .
+ by the definition of @xmath33 , there is a vertex @xmath35 such that @xmath36 and @xmath37 have the same colour in @xmath27 .
hence we can recolour @xmath36 to the same colour as @xmath15 , and then recolour @xmath15 to the same colour as @xmath49 .
the resulting colouring is good . *
case 2*:all vertices in @xmath33 are adjacent to @xmath15 .
+ additionally , define @xmath38 . and
note that all vertices in @xmath40 have a unique colour in @xmath27 .
* subcase 2.1*:there is a vertex @xmath41 that is not adjacent to some vertex @xmath34 .
+ again , there is a vertex @xmath35 such that @xmath36 and @xmath37 have the same colour in @xmath27 .
then we first recolour @xmath36 to the same colour as @xmath42 .
then recolour @xmath42 to the same colour as @xmath15 , and lastly recolour @xmath15 to the same colour as @xmath49 . again , this gives a good colouring .
* subcase 2.2*:every vertex in @xmath40 is adjacent to every vertex in @xmath33 .
+ because of lemma [ z.8 ] , @xmath43 is not empty ( otherwise , the pair @xmath44 would form the parts of a spanning complete bipartite subgraph of @xmath1 ) .
suppose there is some vertex @xmath45 that is not adjacent to some vertex @xmath34 and not adjacent to some vertex @xmath41 .
then we can recolour @xmath36 to the colour of @xmath30 ( this is possible since there is another vertex @xmath35 with the same colour as @xmath36 ) .
then recolour @xmath30 to the same colour as @xmath42 , @xmath42 to the same colour as @xmath15 , and lastly recolour @xmath15 to the same colour as @xmath49 .
it is easy to check that the remaining colouring is good .
so we are left with the case that each vertex in @xmath43 is adjacent to every vertex in @xmath33 or to every vertex in @xmath40 .
let @xmath46 be the set of vertices in @xmath43 that are adjacent to every vertex in @xmath33 , and @xmath47 .
then the pair @xmath48 forms the parts of a spanning complete bipartite subgraph of @xmath1 .
because of lemma [ z.8 ] , this contradicts that @xmath2 is connected .
it is easy to see that in the normal colour graph @xmath3 there always is a path from any proper @xmath0-vertex - colouring to some strong @xmath0-vertex - colouring .
this shows the following .
[ z.10 ] if @xmath2 is connected , then @xmath3 is also connected .
in this section , we prove that the strong @xmath0-colour graph of a path with @xmath21 vertices , @xmath55 , is connected if and only if @xmath56 , @xmath23 , and @xmath57 .
first , suppose we colour a path @xmath58 , @xmath59 , with two colours .
it is easy to see that there are only two strong 2-vertex - colourings of @xmath58 , and they are not adjacent in @xmath60 .
thus @xmath60 is not connected for all @xmath59 . for @xmath61 ,
we have already seen in figures [ f1 ] and [ f2 ] that @xmath62 is not connected , but @xmath63 is connected .
it is somewhat more work to show that @xmath64 is connected .
[ b.6 ] the strong colour graph @xmath64 is connected . in any strong 4-colouring of @xmath65 , there are only two vertices with the same colour .
let @xmath9 be a strong 4-colouring of @xmath66 .
we call @xmath9 an _ @xmath67-standard colouring _ if @xmath68
. 0.25 mm ( 130,40)(0,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 25,5)(0,0)[t]@xmath72 ( 105,5)(0,0)[t]@xmath73 we will prove the proposition by combining one or more of the following three steps . *
step 1*:there is a path from @xmath74 to any other @xmath67-standard colouring .
+ we will first show that there is a path from @xmath74 to @xmath75 : 0.20 mm ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath69 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath69 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath69 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath69 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath69 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath70 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath70 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath70 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath70 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 by symmetry , there is a path from @xmath74 to @xmath76 as well .
now we consider @xmath67-standard colourings as permutations of @xmath77 .
note that all these permutations can be generated by the transpositions @xmath78 and @xmath79 .
therefore , since we can find a path from @xmath74 to @xmath75 , and from @xmath74 to @xmath76 , there is a path from @xmath74 to any other @xmath67-standard colouring .
* step 2*:there is a path between any two types of standard colourings .
+ first , here is a path from an @xmath67-standard colouring @xmath9 to an @xmath80-standard colouring : 0.20 mm ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath70 next , a path from an @xmath67-standard colouring @xmath9 to an @xmath81-standard colouring : 0.20 mm ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath71 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath71 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath69 ( 105,33)(0,0)[b]@xmath67 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath71 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath69 ( 105,33)(0,0)[b]@xmath67 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath71 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath69 ( 105,33)(0,0)[b]@xmath71 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath69 ( 105,33)(0,0)[b]@xmath71 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath71 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath71 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath69 by symmetry , there is also a path from an @xmath67-standard colouring @xmath9 to an @xmath82-standard colouring . *
step 3*:each colouring has a path to some standard colouring .
+ let @xmath9 be a strong 4-colouring of @xmath65
. then @xmath9 has one of the following forms : 0.20 mm ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath71 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath71 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 the first form already is an @xmath67-standard colouring ; for the second and the third ones we just recolour the vertex @xmath72 to @xmath71 ; while for the fourth and the sixth ones we just recolour the vertex @xmath73 to @xmath69 . finally , for the fifth form the following is a path to a @xmath69-standard colouring : 0.20 mm ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath71 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath71 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 120,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath69 it is straightforward to see that appropriate renaming of the colours and sequence of the paths in steps 13 will transform any strong 4-colouring of @xmath65 into any other strong 4-colouring .
we extend the last result by showing that @xmath83 is connected , for all @xmath84 .
[ b.9 ] for all @xmath84 , @xmath83 is connected .
we will prove this by induction on @xmath0 .
we have already shown the proposition is true for @xmath85 .
let @xmath9 and @xmath11 be strong @xmath0-colourings of @xmath86 , for some @xmath87 .
we can assume that in @xmath9 , the vertex @xmath88 has a unique colour .
otherwise , there is another vertex @xmath89 such that @xmath90 .
then just recolour @xmath89 to a colour different from @xmath91 .
next , we can assume that this unique colour on @xmath88 in @xmath9 is @xmath67 .
if the vertex @xmath88 is the only vertex coloured @xmath67 in @xmath11 as well , then we can just remove @xmath88 .
let @xmath92 and @xmath93 be the strong @xmath0-colourings of @xmath94 obtained from @xmath9 and @xmath11 , respectively .
since , by induction , @xmath95 is connected , there is a path from @xmath92 to @xmath93 in @xmath95 .
using the same steps on @xmath96 gives a path from @xmath9 to @xmath11 in @xmath83 .
so we can assume that in @xmath11 , @xmath88 is not coloured @xmath67 or is not the only vertex coloured @xmath67 .
we distinguish 4 cases .
* case 1*:in @xmath11 , @xmath88 is coloured @xmath67 , but there is a second vertex @xmath89 coloured @xmath67 as well .
+ then just recolour @xmath89 to some different from @xmath67 , and so @xmath88 is now the only vertex coloured @xmath67 .
we are done by the paragraph above .
* case 2*:in @xmath11 , @xmath88 and some other vertex @xmath89 have the same colour @xmath97 , while a third vertex @xmath98 is coloured @xmath67 .
* subcase 2.1*:@xmath88 and @xmath98 are not adjacent .
+ then just recolour @xmath88 to @xmath67 , and we are back to case 1 . *
subcase 2.2*:@xmath88 and @xmath98 are adjacent , i.e. , @xmath99 .
+ call a colouring of @xmath83 _ good _ if we can recolour @xmath89 to a colour which is not one of @xmath100 .
note that @xmath11 is good when @xmath101 , or @xmath102 and @xmath103 .
if @xmath11 is good , we can recolour @xmath89 to the colour @xmath104 for some @xmath105 , to obtain the strong @xmath0-colouring @xmath106 , let @xmath107 be the strong @xmath0-colouring of @xmath96 , obtained from @xmath106 by swapping the colours of @xmath89 and @xmath108 . by ignoring the vertex @xmath88
, we can consider @xmath106 and @xmath107 as strong @xmath109-colourings of @xmath110 .
since @xmath95 is connected , there is a path between these two colourings .
we then apply this path to a path in @xmath83 from @xmath106 to @xmath107 .
next , we will form a path from @xmath107 to a colouring in which vertex @xmath88 is the only vertex coloured @xmath67 . therefore
, we also have a path from @xmath11 to this colouring . in @xmath107 , we first recolour @xmath111 to @xmath112 and then recolour @xmath88 to @xmath67
. finally , recolour vertex @xmath89 , which is previously coloured @xmath67 , to another colour , and we are done .
we now suppose that @xmath11 is not good , i.e. , @xmath102 and @xmath113 .
then there is a path from @xmath11 to a colouring in which @xmath88 is the only vertex coloured @xmath67 .
0.20 mm ( 140,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,35)(0,0)[b]@xmath70 ( 45,35)(0,0)[b]@xmath69 ( 65,35)(0,0)[b]@xmath71 ( 85,35)(0,0)[b]@xmath114 ( 105,35)(0,0)[b]@xmath67 ( 125,35)(0,0)[b]@xmath69 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 140,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,35)(0,0)[b]@xmath70 ( 45,35)(0,0)[b]@xmath67 ( 65,35)(0,0)[b]@xmath71 ( 85,35)(0,0)[b]@xmath114 ( 105,35)(0,0)[b]@xmath67 ( 125,35)(0,0)[b]@xmath69 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 140,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,35)(0,0)[b]@xmath70 ( 45,35)(0,0)[b]@xmath67 ( 65,35)(0,0)[b]@xmath71 ( 85,35)(0,0)[b]@xmath114 ( 105,35)(0,0)[b]@xmath71 ( 125,35)(0,0)[b]@xmath69 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 140,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,35)(0,0)[b]@xmath70 ( 45,35)(0,0)[b]@xmath67 ( 65,35)(0,0)[b]@xmath69 ( 85,35)(0,0)[b]@xmath114 ( 105,35)(0,0)[b]@xmath71 ( 125,35)(0,0)[b]@xmath69 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 140,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,35)(0,0)[b]@xmath70 ( 45,35)(0,0)[b]@xmath67 ( 65,35)(0,0)[b]@xmath69 ( 85,35)(0,0)[b]@xmath114 ( 105,35)(0,0)[b]@xmath71 ( 125,35)(0,0)[b]@xmath67 ( 25,40)(0,15 ) ( 0,20)(1,0)25 ( 140,40)(5,15 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,35)(0,0)[b]@xmath70 ( 45,35)(0,0)[b]@xmath114 ( 65,35)(0,0)[b]@xmath69 ( 85,35)(0,0)[b]@xmath114 ( 105,35)(0,0)[b]@xmath71 ( 125,35)(0,0)[b]@xmath67 * case 3*:in @xmath11 , @xmath89 and @xmath98 are coloured @xmath67 for some @xmath115 .
+ without loss of generality , we may assume that @xmath89 is not adjacent to @xmath88 .
then we recolour @xmath89 to @xmath116 , and we are back to case 2 . *
case 4*:in @xmath11 , @xmath89 and @xmath98 have the same colour @xmath97 , a third vertex @xmath117 is coloured @xmath67 , for some @xmath118 .
+ without loss of generality , we may assume that @xmath89 is not adjacent to @xmath88 .
then we recolour @xmath89 to @xmath116 , and we are back to case 2 . combining it all
, we get the promised result on the strong colour graph of paths .
[ b.10 ] the strong colour graph @xmath55 is connected if and only if @xmath56 , @xmath23 and @xmath57 .
we already have seen that @xmath62 and @xmath60 , @xmath119 , are not connected , while @xmath63 and @xmath83 , @xmath84 , are connected .
applying theorem [ z.6 ] completes the proof .
in this section we want to show that the strong @xmath0-colour graph of a cycle with @xmath21 vertices , @xmath120 , is connected if and only if @xmath84 , @xmath121 and @xmath57 . before we prove the theorem , we prove some tools used in this proof .
to orient a cycle means to orient each edge on the cycle so that a directed cycle is obtained .
if @xmath122 is a cycle , then by @xmath123 we denote the cycle with one of the two possible orientations of @xmath71 .
given a 3-colouring @xmath9 using colours @xmath124 , the weight of an edge @xmath125 oriented from @xmath30 to @xmath49 is @xmath126 the weight @xmath127 of an oriented cycle @xmath123 is the sum of the weights of its oriented edges .
[ lemma5jan ] ( cereceda et al .
@xcite ) let @xmath9 be a 3-colouring of a graph @xmath1 that contains a cycle @xmath122 .
if @xmath128 , then @xmath3 is not connected .
[ a.1.5 ] for all @xmath119 , @xmath25 is not connected . by lemmas [ z.10 ] and [ lemma5jan ] ,
it is enough to find a strong 3-colouring @xmath9 with @xmath129 . if @xmath130 for some positive integer @xmath131 , the pattern 1,2,3,1,2,3, ... ,1,2,3 provides a 3-colouring @xmath9 of @xmath132 with @xmath133 . for @xmath134 , it is easy to see that @xmath135 is a graph with 12 isolated vertices .
if @xmath136 , then we use the pattern 1,2,3,1,2,3, ... ,1,2,3,2 , which gives @xmath137 . finally , if @xmath138 , then we use the pattern 1,2,3,1,2,3, ... ,1,2,3,1,2 , with @xmath139 .
[ a.2.2 ] the strong colour graph @xmath140 is not connected . for any strong 4-colouring of the 5-cycle @xmath141
, there are only two vertices having the same colour .
thus each strong 4-vertex - colouring of @xmath141 can be recoloured only on these two vertices , and each of these two vertices can be recoloured to only one new colour ( since the two different colours of their neighbours are forbidden ) .
this means each colouring has degree two in @xmath140 .
straightforward counting shows that @xmath140 has 120 vertices . but
each colouring in @xmath140 is contained in some cycle of length 20 . to see this , we start with some strong 4-colouring of @xmath141 and recolour : 0.20 mm ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath71 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath71 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath69 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath71 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath69 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath71 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath69 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath71 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath71 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath69 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath71 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath69 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath71 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath71 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath71 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath69 ( 105,33)(0,0)[b]@xmath67 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath69 ( 105,33)(0,0)[b]@xmath67 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath69 ( 105,33)(0,0)[b]@xmath67 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath69 ( 105,33)(0,0)[b]@xmath71 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath69 ( 105,33)(0,0)[b]@xmath71 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 120,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 25,20)(1,0)80 ( 25,20)(65,-5)(105,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 by symmetry , we immediately get that @xmath140 is a disjoint union of six copies of @xmath142 , so it is not connected .
[ a.3 ] the strong colour graph @xmath143 is connected . in any strong 5-colouring of the 6-cycle @xmath144
, there are only two vertices having the same colour .
let @xmath9 be a strong 5-colouring of @xmath145 .
we call @xmath9 an _ @xmath67-standard colouring _ if @xmath146
. 0.25 mm ( 145,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 25,10)(0,0)[t]@xmath72 ( 125,10)(0,0)[t]@xmath147 we will prove the proposition by showing the following three steps . *
step 1*:there is a path from @xmath148 to any other @xmath67-standard colourings .
+ first , we will show that there is a path from @xmath148 to @xmath149 : 0.20 mm ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath114 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath114 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath114 ( 85,33)(0,0)[b]@xmath69 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath114 ( 85,33)(0,0)[b]@xmath69 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath114 ( 85,33)(0,0)[b]@xmath69 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath69 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 by symmetry , there is also a path from @xmath148 to @xmath150 .
next , we show that there is a path from @xmath148 to @xmath151 : 0.20 mm ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath114 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath114 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath69 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath114 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath69 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath114 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath114 ( 125,33)(0,0)[b]@xmath69 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath114 ( 125,33)(0,0)[b]@xmath69 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath114 ( 125,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath114 ( 125,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath114 ( 125,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath114 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath114 ( 125,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath114 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath69 ( 125,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath114 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath69 ( 125,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath114 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath69 ( 125,33)(0,0)[b]@xmath70 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath114 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath69 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath71 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath69 ( 125,33)(0,0)[b]@xmath114 now we consider @xmath67-standard colourings as permutations of @xmath152 .
note that all these permutations can be generated by the transpositions @xmath78 , @xmath153 and @xmath154 .
therefore , since we can find a path from @xmath148 to @xmath149 , from @xmath148 to @xmath150 , and from @xmath148 to @xmath151 , there is a path from @xmath155 to any other @xmath67-standard colourings . *
step 2*:there is a path between any two types of standard colourings .
+ first , here is a path from an @xmath67-standard colouring @xmath9 to an @xmath156-standard colouring : 0.20 mm ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath71 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath71 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath67 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath71 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath71 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath67 ( 125,33)(0,0)[b]@xmath114 next , a path from an @xmath67-standard colouring @xmath9 to an @xmath82-standard colouring : 0.20 mm ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath114 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath114 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath67 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath70 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath114 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath67 by symmetry , there is also a path from an @xmath67-standard colouring @xmath9 to an @xmath157-standard colouring . and finally , a path from an @xmath67-standard colouring @xmath9 to an @xmath81-standard colouring : 0.20 mm ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath71 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath71 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath69 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath71 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath69 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath71 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath69 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath69 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 125,33)(0,0)[b]@xmath114 ( 25,50)(0,5 ) ( 0,20)(1,0)25 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath69 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 125,33)(0,0)[b]@xmath114 * step 3*:each colouring has a path to some standard colouring .
+ let @xmath9 be a strong 5-colouring of @xmath144
. then @xmath9 has one of the following forms : 0.20 mm ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath70 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath67 ( 45,33)(0,0)[b]@xmath69 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 125,33)(0,0)[b]@xmath114 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath71 ( 125,33)(0,0)[b]@xmath114 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 125,33)(0,0)[b]@xmath114 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath67 ( 65,33)(0,0)[b]@xmath70 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath114 ( 125,33)(0,0)[b]@xmath67 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath67 ( 125,33)(0,0)[b]@xmath114 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath67 ( 85,33)(0,0)[b]@xmath71 ( 105,33)(0,0)[b]@xmath114 ( 125,33)(0,0)[b]@xmath67 ( 140,50)(5,5 ) ( 25,20 ) ( 45,20 ) ( 65,20 ) ( 85,20 ) ( 105,20 ) ( 125,20 ) ( 25,20)(1,0)100 ( 25,20)(75,-5)(125,20 ) ( 25,33)(0,0)[b]@xmath69 ( 45,33)(0,0)[b]@xmath70 ( 65,33)(0,0)[b]@xmath71 ( 85,33)(0,0)[b]@xmath67 ( 105,33)(0,0)[b]@xmath114 ( 125,33)(0,0)[b]@xmath67 the first form is already an @xmath67-standard colouring . for the second and the third forms , we just recolour vertex @xmath72 to @xmath70 , and for the seventh and eight forms , we just recolour vertex @xmath158 to @xmath69 .
for all the remaining colourings , we can find a path of length two to some standard colouring .
we will leave checking that to the reader .
it is straightforward to see that appropriate renaming of the colours and sequence of the paths in steps 13 will transform any strong 5-colouring of @xmath159 into any other strong 5-colouring .
[ a.4 ] the strong colour graph @xmath120 is connected if and only if @xmath84 , @xmath121 , and @xmath57 .
we already have seen that @xmath25 , @xmath119 and @xmath140 are disconnected . from theorems [ z.6 ] and [ b.10 ]
we easily obtain that @xmath120 is connected for all @xmath84 , @xmath121 , and @xmath160 .
since @xmath143 is connected , all that is left to prove is that @xmath161 is connected for all @xmath101 .
let @xmath101 and let @xmath9 and @xmath11 be strong @xmath0-colourings of @xmath162 . in @xmath9 , there will be a vertex , say @xmath72 , which has an unique colour , say colour @xmath67 .
we say that a strong @xmath0-colouring of @xmath163 is _ good _ if @xmath72 is the only vertex in @xmath163 , which is coloured @xmath67 .
thus @xmath9 is good .
if @xmath11 is good as well , then remove @xmath72 , and let @xmath92 and @xmath93 be the strong @xmath109-colourings of @xmath110 obtained from @xmath9 and @xmath11 , respectively . since @xmath101 , by proposition [ b.9 ] there is a path in @xmath95 from @xmath92 to @xmath93 .
using the same steps gives a path from @xmath9 to @xmath11 in @xmath163 .
so suppose that @xmath11 is not good .
as we colour the @xmath6 vertices of @xmath163 with @xmath0 colours , there are only two vertices having the same colour .
we distinguish five cases . *
case 1*:in @xmath11 , @xmath72 is coloured @xmath67 , but there is a second vertex @xmath89 coloured @xmath67 as well .
+ then just recolour @xmath89 to another colour .
the resulting colouring is good , and we are done by the paragraph above . *
case 2*:in @xmath11 , @xmath72 and some other vertex @xmath89 have the same colour @xmath97 , while a third vertex @xmath98 with @xmath164 is coloured @xmath67 .
+ then we first recolour @xmath72 to @xmath67 , and then recolour @xmath98 to another colour .
again , this gives a good colouring , so we are done .
* case 3*:in @xmath11 , @xmath72 and some other vertex @xmath89 have the same colour @xmath97 , while a third vertex @xmath98 with @xmath165 is coloured @xmath67 .
+ without loss of generality , assume that @xmath166 .
* subcase 3.1*:we have @xmath167 .
+ now first recolour vertex @xmath89 to colour @xmath168 , then recolour @xmath158 to @xmath169 , @xmath88 to @xmath170 , @xmath171 to @xmath172 , @xmath72 to @xmath67 , and finally recolour @xmath158 to some colour different from @xmath67 .
it is easy to check that the remaining colouring is good . *
subcase 3.2*:we have @xmath173 .
+ recall that @xmath174 .
now first recolour vertex @xmath89 to colour @xmath175 , then recolour @xmath147 to @xmath172 .
now @xmath72 and @xmath147 have the same colour @xmath69 , so we are back to subcase 3.1 . *
case 4*:in @xmath11 , @xmath72 has a unique colour @xmath97 , while there are two vertices @xmath89 and @xmath98 coloured @xmath67 . *
subcase 4.1*:we have @xmath176 .
+ now first recolour vertex @xmath177 to @xmath170 , and then recolour @xmath171 to @xmath172 .
then we are back to subcase 3.2 .
* subcase 4.2*:we have @xmath178 .
+ without loss of generality , assume that @xmath179
. then we can recolour @xmath89 to @xmath172 .
this means that @xmath72 and @xmath89 have the same colour @xmath97 , so we are back in case 2 or 3 . *
case 5*:in @xmath11 , @xmath72 has a unique colour @xmath97 , there is a unique vertex @xmath89 coloured @xmath67 , and two vertices @xmath98 and @xmath117 have the same colour @xmath180 . *
subcase 5.1*:we have @xmath181 .
+ since @xmath182 , we must have @xmath183 or @xmath184 . without loss of generality , assume that @xmath183 .
then recolour @xmath177 to @xmath185 , and next recolour @xmath89 to @xmath172 .
this brings us back to case 3 .
* subcase 5.2*:we have @xmath186 .
+ without loss of generality , assume that @xmath187
. then we can recolour @xmath98 to @xmath172 .
this means that @xmath72 and @xmath98 have the same colour @xmath97 , and we are back in case 2 or 3 .
the aim of this section is to classify the trees @xmath188 for which the strong 3-colour graph @xmath189 is connected . for this we need to consider some special trees .
first , in section [ sec2 ] we saw that the strong @xmath0-colour graph of a complete bipartite graph is not connected , so @xmath190 is disconnected for all @xmath59 . for @xmath191 and @xmath192 ,
let @xmath193 , @xmath194 and @xmath195 be the graphs sketched in figure [ f5 ] , respectively .
0.28 mm ( 60,40)(15,-60 ) ( 15,0)(15,40 ) ( 45,0)(45,40 ) ( 75,0)(75,40 ) ( 15,0)(1,0)60 ( 15,40)(1,0)60 ( 45,0)(0,1)40 ( 15,50)(0,0)[b]@xmath196 ( 45,50)(0,0)[b]@xmath197 ( 75,50)(0,0)[b]@xmath198 ( 15,-10)(0,0)[t]@xmath199 ( 45,-10)(0,0)[t]@xmath200 ( 75,-10)(0,0)[t]@xmath201 ( 70,140)(25,-30 ) ( 25,70 ) ( 45,70 ) ( 65,70 ) ( 70,70 ) ( 75,70 ) ( 95,70 ) ( 65,40 ) ( 65,20 ) ( 65,0 ) ( 65,0)(0,1)40 ( 65,40)(-4,3)40 ( 65,40)(-2,3)20 ( 65,40)(1,1)30 ( 80,35)(0,0)@xmath202 ( 25,80)(0,0)[b]@xmath72 ( 45,80)(0,0)[b]@xmath177 ( 95,80)(0,0)[b]@xmath203 ( 70,140)(25,-30 ) ( 25,70 ) ( 45,70 ) ( 65,70 ) ( 70,70 ) ( 75,70 ) ( 95,70 ) ( 65,40 ) ( 65,20 ) ( 25,-10 ) ( 45,-10 ) ( 65,-10 ) ( 70,-10 ) ( 75,-10 ) ( 95,-10 ) ( 65,20)(0,1)20 ( 65,40)(-4,3)40 ( 65,40)(-2,3)20 ( 65,40)(1,1)30 ( 65,20)(-4,-3)40 ( 65,20)(-2,-3)20 ( 65,20)(1,-1)30 ( 25,80)(0,0)[b]@xmath204 ( 45,80)(0,0)[b]@xmath205 ( 95,78)(0,0)[b]@xmath206 ( 25,-20)(0,0)[t]@xmath207 ( 45,-20)(0,0)[t]@xmath208 ( 95,-20)(0,0)[t]@xmath209 it is straightforward to check that in any strong 3-colouring of @xmath194 we can not recolour the vertex @xmath202 to another colour so that the resulting 3-colouring is strong again .
hence the strong colour graph @xmath210 is disconnected for all @xmath191 .
[ d.5 ] the strong 3-colour graph @xmath211 is connected .
let @xmath9 be a strong 3-colouring of the graph @xmath193 , with vertex set @xmath212 as in figure [ f5 ] .
we call @xmath9 an _ @xmath213-standard colouring _ if @xmath214 and @xmath215 .
easy counting shows that for fixed @xmath216 , there are 15 @xmath213-standard colourings .
( there are 2 choices for each of the other 4 vertices , but one of the resulting 16 3-colourings is not strong . ) as there are 6 choices for pairs @xmath216 from 3 colours , there are a total of 90 strong 3-colourings of @xmath193 .
we will prove the proposition by combining the following two steps .
* step 1*:for given @xmath216 , there is a path containing all @xmath213-standard colourings .
0.20 mm ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath69 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath69 ( 25,-25)(0,0)[b]@xmath70 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath67 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath69 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath70 ( 25,-25)(0,0)[b]@xmath70 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath67 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath69 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath70 ( 25,-25)(0,0)[b]@xmath67 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath67 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath70 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath70 ( 25,-25)(0,0)[b]@xmath67 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath67 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath70 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath70 ( 25,-25)(0,0)[b]@xmath67 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath70 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath69 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath70 ( 25,-25)(0,0)[b]@xmath67 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath70 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath69 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath70 ( 25,-25)(0,0)[b]@xmath70 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath70 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath70 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath70 ( 25,-25)(0,0)[b]@xmath70 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath70 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath70 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath70 ( 25,-25)(0,0)[b]@xmath70 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath67 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath70 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath69 ( 25,-25)(0,0)[b]@xmath70 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath67 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath70 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath69 ( 25,-25)(0,0)[b]@xmath70 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath70 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath69 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath69 ( 25,-25)(0,0)[b]@xmath70 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath70 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath69 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath69 ( 25,-25)(0,0)[b]@xmath67 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath70 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath70 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath69 ( 25,-25)(0,0)[b]@xmath67 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath70 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath70 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath69 ( 25,-25)(0,0)[b]@xmath67 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath67 * step 2*:there is a path containing at least one colouring from each type of standard colourings . 0.20 mm ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath70 ( 45,55)(0,0)[b]@xmath69 ( 65,55)(0,0)[b]@xmath67 ( 25,-25)(0,0)[b]@xmath69 ( 45,-25)(0,0)[b]@xmath70 ( 65,-25)(0,0)[b]@xmath69 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath70 ( 45,55)(0,0)[b]@xmath69 ( 65,55)(0,0)[b]@xmath67 ( 25,-25)(0,0)[b]@xmath69 ( 45,-25)(0,0)[b]@xmath67 ( 65,-25)(0,0)[b]@xmath69 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath70 ( 45,55)(0,0)[b]@xmath69 ( 65,55)(0,0)[b]@xmath67 ( 25,-25)(0,0)[b]@xmath70 ( 45,-25)(0,0)[b]@xmath67 ( 65,-25)(0,0)[b]@xmath69 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath67 ( 45,55)(0,0)[b]@xmath69 ( 65,55)(0,0)[b]@xmath67 ( 25,-25)(0,0)[b]@xmath70 ( 45,-25)(0,0)[b]@xmath67 ( 65,-25)(0,0)[b]@xmath69 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath67 ( 45,55)(0,0)[b]@xmath70 ( 65,55)(0,0)[b]@xmath67 ( 25,-25)(0,0)[b]@xmath70 ( 45,-25)(0,0)[b]@xmath67 ( 65,-25)(0,0)[b]@xmath69 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath67 ( 45,55)(0,0)[b]@xmath70 ( 65,55)(0,0)[b]@xmath69 ( 25,-25)(0,0)[b]@xmath70 ( 45,-25)(0,0)[b]@xmath67 ( 65,-25)(0,0)[b]@xmath69 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath67 ( 45,55)(0,0)[b]@xmath70 ( 65,55)(0,0)[b]@xmath69 ( 25,-25)(0,0)[b]@xmath70 ( 45,-25)(0,0)[b]@xmath67 ( 65,-25)(0,0)[b]@xmath70 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath67 ( 45,55)(0,0)[b]@xmath70 ( 65,55)(0,0)[b]@xmath69 ( 25,-25)(0,0)[b]@xmath70 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath70 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath67 ( 45,55)(0,0)[b]@xmath70 ( 65,55)(0,0)[b]@xmath69 ( 25,-25)(0,0)[b]@xmath67 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath70 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath69 ( 45,55)(0,0)[b]@xmath70 ( 65,55)(0,0)[b]@xmath69 ( 25,-25)(0,0)[b]@xmath67 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath70 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath69 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath69 ( 25,-25)(0,0)[b]@xmath67 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath70 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath69 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath70 ( 25,-25)(0,0)[b]@xmath67 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath70 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath69 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath70 ( 25,-25)(0,0)[b]@xmath67 ( 45,-25)(0,0)[b]@xmath69 ( 65,-25)(0,0)[b]@xmath67 ( 25,120)(0,-30 ) ( 0,20)(1,0)25 ( 80,120)(5,-30 ) ( 25,0)(25,40 ) ( 45,0)(45,40 ) ( 65,0)(65,40 ) ( 25,0)(1,0)40 ( 25,40)(1,0)40 ( 45,0)(0,1)40 ( 25,55)(0,0)[b]@xmath69 ( 45,55)(0,0)[b]@xmath67 ( 65,55)(0,0)[b]@xmath70 ( 25,-25)(0,0)[b]@xmath67 ( 45,-25)(0,0)[b]@xmath70 ( 65,-25)(0,0)[b]@xmath67 these two steps , together with appropriate renaming of the colours , will give all that is needed to transform any strong 3-colouring of @xmath193 into any other strong 3-colouring .
[ d.10 ] let @xmath188 be a tree . then @xmath189 is connected if and only if @xmath188 contains @xmath65 or @xmath193 as a subgraph . since @xmath63 and @xmath211 are connected , one direction is immediately proved by using theorem [ z.6 ] . for the other direction , suppose that @xmath188 does not contain @xmath65 , nor @xmath193 . since @xmath65 is a path with 4 edges , the longest path in @xmath188 can have length at most 3 .
thus @xmath188 has to be one of the following : @xmath217 , @xmath218 , @xmath219 , @xmath220 , or @xmath194 , @xmath191 .
note that @xmath188 can not be @xmath195 for all @xmath192 since then it would contain @xmath193 as a subgraph .
since @xmath217 and @xmath218 have fewer than 3 vertices , @xmath188 can not be one of these two graphs .
we already saw that @xmath221 , @xmath220 , and @xmath210 , @xmath191 , are disconnected .
we can conclude that @xmath189 is not connected , which completes the proof .
we realise that this note contains only some first results on the strong colour graphs . in comparison , there is a growing body of literature on the connectivity of the normal colour graph : @xcite .
an interesting direction of future research would be to investigate how far the theory of strong colour graphs can be reduced to the theory of normal colour graphs .
we have already seen in lemma [ z.10 ] that if the strong colour graph @xmath2 is connected for some @xmath1 and @xmath0 , then so is the normal colour graph @xmath3 . in general , the reverse direction is not true .
for instance , for all @xmath222 , for the complete bipartite graphs @xmath10 we have that for @xmath56 , @xmath223 is connected , whereas @xmath12 is not connected .
in fact , we ve already seen that if @xmath1 has a complete bipartite graph as a spanning subgraph , then @xmath2 is never connected .
for @xmath56 it is not hard to construct other graphs apart from complete bipartite graphs that have this property , but all examples we know of have a fairly special structure ( see for instance the trees @xmath194 in figure [ f5 ] ) .
this makes us raise the following question .
[ q1 ] is it possible to completely describe a class of graphs @xmath224 so that if @xmath1 does not contain a graph from @xmath224 as a spanning subgraph , then @xmath3 is connected if and only if @xmath2 is connected ?
j. billingham , r.a .
leese , and h. rajaniemi , frequency reassignment in cellular phone networks .
_ smith institute study group report _
available from www.smithinst.ac.uk/projects/esgi53/esgi53-motorola/report ( accessed 4 august 2010 ) .
r. jacob , standortplanung mit blick auf online - strategien .
_ graduate thesis , university of wrzburg _ , 1997 .
available from www14.informatik.tu-muenchen.de/personen/jacob/publications.html ( accessed 4 august 2010 ) .
o. marcotte and p. hansen , the height and length of colour switching . in : p.
hansen and o. marcotte ( eds . ) , _ graph colouring and applications , proceedings of the centre de rchrches mathmatiques _ , vol . * 23*. oxford university press , oxford , 1999 . | for a positive integer @xmath0 and a graph @xmath1 , we consider proper vertex - colourings of @xmath1 with @xmath0 colours in which all @xmath0 colours are actually used .
we call such a colouring a strong
. the strong @xmath0-colour graph of @xmath1 , @xmath2 , is the graph that has all the strong @xmath0-colourings of @xmath1 as its vertex set , and two colourings are adjacent in @xmath2 if they differ in colour on only one vertex of @xmath1 . in this paper , we show some results related to the question : for what @xmath1 and @xmath0 is @xmath2 connected ?
* keywords * : strong @xmath0-vertex - colouring , strong @xmath0-colour graph , strong colour graph . |
a main lesson learnt from experimental data collected in ( d)au+au and pb+pb collisions at rhic and the lhc respectively is that bulk particle production in ion - ion collisions is very different from a simple superposition of nucleon - nucleon collisions .
such is evident in terms of the measured charged particle multiplicities , which exhibit a strong deviation from the scaling with the number of nucleon - nucleon collisions : @xmath7 , and also from the non - trivial transverse momentum dependence of nuclear modification factors measured in d+au collisions .
these observations lead to the conclusion that strong coherence effects among the constituent nucleons , or the relevant degrees of freedom at the sub - nucleon level , must be present during the collisions process .
indeed , and regardless of the question whether the cgc is the most suited framework for their description there is broad consensus that coherence effects are essential for the interpretation of present data on heavy ion collisions .
in fact most if not all of the different phenomenological approaches for the description of particle production both in the soft or hard sector. on physical grounds , coherence phenomena are related to the presence of high gluon densities in the wave function of the colliding nuclei at small values of bjorken-@xmath0 proven in a hadronic collision can be obtained using @xmath8 kinematics @xmath9 , with @xmath10 , @xmath11 and @xmath12 being the collision energy and the transverse momentum and rapidity of the produced hadron respectively . ]
while a detailed discussion of the different prescriptions found in the literature to account for coherence effects is beyond the scope of this brief review , one can identify in different models coherence effects at the level of the wave function and also at the level of primary particle production , sketched in fig . 1 ( left ) .
to the first category correspond the nuclear shadowing ( in a partonic language ) or the percolation and string fusion ( in non - perturbative approaches ) . in both cases , when different constituents , whichever the degrees of freedom chosen are , overlap in phase space according to some geometric criterium , recombination of such constituents happen , thus reducing the total number of scattering centers gluons entering the collision process .
similar phase - space arguments motivate the implementation of energy - dependent cut - offs to regulate independent particle production from different sources , normally a working hypothesis in most monte carlo event generators for heavy ion collisions . in fig 1
( right ) we show the collision energy dependence of the transverse momentum cutoff that separates _ hard _ from _ soft _ particle production in several event generators for p+p ( pythia ) and a+a ( hydjety and hijing ) collisions . its strong rise with increasing energy signals the increasing importance of collective effects in particle production processes . also frequent in the literature
are the resummation of multiple scatterings , either in a coherent higher twist formalism or in a incoherent _ glauber like _ formalism . finally , the modification of parton distribution functions to allow for the presence of a normally energy dependent intrinsic transverse momentum constitute other common practice in phenomenological works .
all these ingredients are akin , at least at a conceptual level , to those dynamically built in the cgc , although they are formulated in very different ways .
thus the debate is now which theoretical framework is most suited for their description : whether the cgc ( at its present degree of accuracy ! ) or alternative approaches , typically rooted in the standard collinear factorization framework . .
, title="fig:",scaledwidth=47.0% ] .
, title="fig:",scaledwidth=51.0% ]
briefly , it can be said that the cgc is now entering the next - to - leading ( nlo ) order era .
actually , the kernel of the bk and jimwlk evolution equations are now known to nlo accuracy @xcite or , also , to running coupling accuracy through the resummation of a partial subset of nlo diagrams @xcite . in a similar fashion ,
the following calculations concerning particle production processes have recently become available at nlo accuracy : photon impact factors in deep inelastic scattering @xcite , full nlo @xcite and _ inelastic _ @xcite contributions to hybrid calculation of single inclusive particle production in dilute - dense scattering , running coupling corrections to the @xmath13-factorization formula @xcite and a proof of factorization of multiparticle production processes at nlo @xcite .
furthermore , our knowledge of exclusive particle production and multi - particle correlations ( di - hadrons , hadron - photon etc ) has been advanced significantly through a series of recent works that establish the precise relation between the @xmath14-point functions of the nuclear wave function and the observables of interest @xcite . thus , and despite the intrinsic technical difficulty of higher order calculations in the cgc one should recall that they are performed by expanding on a strong background color field progress on the theoretical side has been steadily delivered over the last years .
notwithstanding the progress brought by these works , the cgc framework is still far from the degree of sophistication and accuracy that characterizes standard pqcd methods based on the collinear factorization formalism and dglap evolution equations .
the reasons for this are manifold : first , the new theory tools discussed above have not been yet fully implemented in a systematic way in phenomenological applications .
current descriptions of data combine nlo with lo theory ingredients in a somehow uncontrolled theoretical scheme .
for instance , the arguably most ambitious effort so far to determine the universal properties of nuclear and proton wave functions at small-@xmath0 in a systematic way is provided by the global aamqs fits to e+p data@xcite and their extension to the nuclear case through monte carlo methods @xcite .
the aamqs fits are rely on the use of the running coupling bk equation to describe the small-@xmath0 evolution of the 2-point function , but then use the lo dipole formalism to calculate the @xmath15-proton cross section .
related issues obscure the successful phenomenological descriptions of data in single @xcite and double @xcite inclusive production in p+p and d+au collisions or the widespread use of @xmath13-factorization in models for total multiplicities , known not to be valid in the case of _ dense - dense _ scattering .
most of the nlo tools being available , the main obstacle to fix such insufficiencies of present phenomenological works is the difficulty of their technical implementation .
nlo and exclusive particle production calculations typically require the knowledge of the @xmath0-dependence of @xmath14-point functions beyond the 2-point function .
hence , the use of the ( running coupling ) bk - equation for the 2-point function , easy to solve numerically , is no longer sufficient and solutions of the full b - jimwlk hierarchy of coupled equations more demanding numerically are needed . a promising analytic method to solve b - jimwlk equations through a gaussian approximation that allows to obtain arbitrary @xmath14-point functions in terms of the 2-point one has been proposed recently @xcite , and its practical implementation would substantially reduce the uncertainties progress related to calculations on di - hadron correlations
. other factor that blurs the predictive power of the cgc framework is the paucity of experimental data on high - energy ( equivalently , small-@xmath0 ) nuclear reactions .
such information is needed to constrain the non - perturbative parameters of the theory , like the initial conditions for the evolution at some initial scale ( @xmath16 tipically taken to be @xmath17 in practical applications ) or the impact parameter dependence of the nuclear unintegrated nuclear distributions .
lacking such information , some degree of modeling is unavoidable .
mean field approaches for the description of the nuclear geometry have been recently superseded by monte carlo methods as the mc - kln or mc - rcbk ones @xcite , where the position of nucleons in the transverse plane are treated as a random variable , thus allowing to account for geometry fluctuations in the collision process . keeping in mind the present limitations of cgc phenomenological works discussed in this section ,
let me now briefly review current cgc predictions for different observables at the lhc in the next sections .
the cgc offers a very economical description of the integrated hadron multiplicities produced in heavy ion collisions .
based solely in dimensional analysis , the number of particles produced per unit of transverse area rises proportional to the saturation scale ( in symmetric collisions ) @xmath18 such approximate expression can be realized either via solutions of the classical equation of motion in the presence of sources ( projectile and target ) or less rigorously for a+a collisions through the use of @xmath13-factorization .
it accounts well for two of the most remarkable features observed in rhic d+au and au+au data as well as in lhc pb+pb data : _
i ) _ approximate factorization of the collision energy and collision centrality of total multiplicities and _ ii ) _ power - law dependence on integrated multiplicities on collision energy . in fig 2
we show the predictions for the rapidity dependence of hadron multiplicities from different cgc models rooted in @xmath13-factorization , where particle production is given by the convolution of the unintegrated gluon distributions of projectile and target .
these models differ mainly in their input for the nuclear unintegrated gluon distributions and in the treatment of the geometry dependence either monte carlo methods or mean field approaches but also in the implementation details like the ir regularization or the rapidity to pseudo - rapidity jacobian or the modeling of the large-@xmath0 component of the corresponding wave functions ( hence the large deviations among them at more forward / backward rapidities ) .
they have been tested against d+au rhic data and then extrapolated to lhc energies without further adjustments .
overall they predict a charged hadron multiplicity @xmath19
an observable that has centered much of the discussion on the relevance of cgc physics in heavy ion collisions is the nuclear modification factor measured in d+au collisions at rhic . at mid - rapidity rhic data show a moderate enhancement of single inclusive particle production tipically atributed to semi - classical multiple scatterings , whereas they are continuously depleted at more forward rapidities .
a good quantitative description of this forward suppression is possible within the cgc @xcite , where it relates to the non - linear evolution of the nuclear ugd .
however , rhic forward data falls very close to the kinematic limit , where large-@xmath0 effects such as energy loss neglected in the in the cgc approach may also be relevant .
indeed , there are alternative explanations of data where energy loss is the main dynamical ingredient ( see e.g @xcite ) .
moreover , the cgc description of the most forward data on neutral pions performed by the star collaboration requires the use of @xmath20-factors smaller than unity , hinting at the relevance of large-@xmath0 effects .
although more exclusive observables like di - hadron or hadron - photon correlations are expected to better discriminate between different approaches , a first test for models of particle production in hic shall come from data on inclusive multiplicities and single particle distributions as they are much easier to obtain experimentally .
generic arguments based on kinematics suggest that one should expect a similar suppression of the nuclear modification factors in lhc p+pb collisions at mid - rapidity as the one observed in forward rhic d+au data , since in both cases the nuclear wave function is probed at similar values of bjorken-@xmath0 . while this argument can be a somewhat misleading ( the proximity of rhic forward data to the kinematic limit limits strongly the transverse momentum of the gluons probed in the nucler ugd ) , it is approximately realized in the different cgc - predictions .
fig 3 show the predictions for @xmath21 at rapidity 0 ( left ) and @xmath22 ( right ) based from the rcbk mc @xcite and ip - sat @xcite models .
they predict a moderate suppression ( similar to the one observed at rhic @xmath23 at moderate to small transverse momentum @xmath24 gev , and a smooth approach to unity at larger transverse momentum .
the error bands in the mc - rcbk calculation originate from the use of different initial conditions to solve the running coupling bk evolution and also from the variation of factorization scales and fragmentation functions in the @xmath13-factorization formula they rely upon . also , it should be note that the mc - rcbk predictions presented here correspond to the _ quenched _ approximation for the nuclear geometry .
the details of this calculation will be presented in @xcite .
remarkably , the cgc predictions for mid - rapidity overlap strongly with those obtained within the collinear factorization framework using the eps09 parametrization for the npdf s ( results from @xcite ) .
one then concludes that only this observable does not suffice to discriminate among different approaches to particle production in high energy heavy ion collisions .
the differences between cgc and collinear factorization formalisms start becoming apparent as one moves to more forward rapidities .
cgc models predict a faster onset of the suppression at higher values of transverse momentum with respect to he eps09 results .
the weak rapidity dependence of eps09 results can be traced back to the flatness of the gluon modification factor at small-@xmath0 in this parametrization .
thus , a rapidity scan of the nuclear modification factors would offer a much larger discriminating power .
cgc predictions at intermediate values of rapidity suffer of a systematic uncertainty : in the cgc there are two distinct but related approaches to hadron production in high energy asymmetric collisions .
particle production processes in the central rapidity region probe the wave functions of both projectile and target at small values of @xmath0 . here , one may employ the @xmath25-factorization formalism where both the projectile and target are characterized in terms of their rcbk evolved unintegrated gluon distributions ( ugds ) . however , at more forward rapidities , the proton is probed at larger values of @xmath0 while the target nucleus is shifted deeper into the small-@xmath0 regime . here
, @xmath13-factorization fails to grasp the dominant contribution to the scattering process .
rather , the _ hybrid _ formalism proposed in ref . @xcite . in the hybrid formalism
the large-@xmath0 degrees of freedom of the proton are described in terms of usual parton distribution functions ( pdfs ) of collinear factorization which satisfy the momentum sum rule exactly and which exhibit a scale dependence given by the dglap evolution equations .
on the other hand , the small-@xmath0 glue of the nucleus is still described in terms of its ugd .
the corresponding limits of applicability of each formalism equivalently the precise value of @xmath0 at which one should switch from one to the other have only been estimated on an empirical basis , and in practice it is taken to be @xmath26 .
the leading order hybrid formalism yield a stronger suppression that the @xmath13-factorization one ( see fig 2 right ) .
recently the hybrid formalism has been improved through the calculation of inelastic contributions and full nlo corrections that may become important at high transverse momentum @xcite .
the inelastic terms were recently implemented in phenomenological work @xcite .
there it was observed that the effect of the inelastic corrections is to slightly increase the value of the nuclear modification factors with respect to the lo result , thus bringing them closer to the results obtained within the @xmath13-factorization framework .
a full phenomenological implementation of the hybrid and @xmath13-factorization at nlo is necessary to better asses this systematic uncertainty . at a strictly qualitative level ,
two generic features of cgc predictions can be highlighted : * : dissapearence of the cronin peak in @xmath21 at central rapidities . * : stronger suppression at forward rapidities : @xmath27 for @xmath28 .
these two features of cgc predictions originate from generic properties of the non - linear small-@xmath0 evolution , regardless of the degree of accuracy of the evolution kernel , nlo or lo etc .
therefore , the persistence of the cronin peak in @xmath21 at the lhc would be very difficult to accommodate in the cgc framework .
similarly the continuos depletion of nuclear modification factors reflects the relative enhancement of non - linear correction to the evolution of nuclear wave function with respect to that of a proton .
correspond to results obtained within the @xmath13-factorization or hybrid formalism respectively . also shown in red the predictions corresponding to the eps09 npdf set within collinear factorization @xcite ( courtesy of p. quiroga - arias ) .
, title="fig:",scaledwidth=49.0% ] correspond to results obtained within the @xmath13-factorization or hybrid formalism respectively . also shown in red the predictions corresponding to the eps09 npdf set within collinear factorization @xcite ( courtesy of p. quiroga - arias ) .
, title="fig:",scaledwidth=49.0% ]
the study of forward di - hadron correlations in d+au azimuthal correlations at rhic has provided the most solid indication for the relevance of cgc effects in data so far .
the disappearance of the away side peak also dubbed _ monojet _ production in more central collisions ( and its persistence in more peripheral collisions and p+p collisions ) can be related to the interplay between the transverse momenta of the produced hadrons and the one acquired during the interaction with the nucleus . in the cgc approach the interaction with the nucleus
is realized in a fully coherent way , and the momentum broadening is parametrically controlled by the @xmath0-dependent saturation scale of the nucleus .
the latter , in turn , is described by means of the rcbk equation .
a first semi - quantitative description of the data was provided in @xcite .
this work neglected some leading in @xmath1 terms which calculation demands knowledge of higher @xmath14-point functions and also the contribution of the gluon channel to the production process .
these caveats were partially fixed in a later work @xcite , where the small momentum imbalance approximation was used . moreover ,
the role of multiparton interactions that may enhance the contribution of the uncorrelated component of the double inclusive cross section ( specially in the forward region ) @xcite has not been fully explored in present cgc calculations .
nevertheless , the description of data is good in these two works is good , see fig 4 .
recently , another description of data based in a higher - twist calculation that also includes nuclear shadowing and cold nuclear matter energy loss has become available @xcite .
a full cgc analyses of data on di - hadron correlations including the missing ingredients in previous cgc analyses is now underway @xcite .
it should serve as the reference to generate precise quantitative predictions for p+pb collisons at the lhc .
so far , qualitative expectations indicate that analogous suppression of azimuthal correlations should be observed in at the lhc .
generically the strength of the decorrelation is expected to be stronger with : _
i ) _ increasing rapidity of the produced pair ; _ ii ) _ increasing collision centrality and _ iii ) _ decreasing transverse momentum of the trigger and associated particle .
it has been recently proposed that hadron - photon @xcite and hadron - dileptons @xcite correlations may exhibit similar azimuthal structure .
these observables offer the advantage that they can be computed in terms of only the rcbk - evolved 2-point function .
although their experimental determination may be more complicated , their measurement at the lhc would provide additional constraints to determine the underlying dynamics of multiparticle production in high - density qcd scattering .
in summary , while it is fair to say that a large number of observables in different systems from e+p to a+a collisions that probe the small-@xmath0 component of the wave function of the projectile or target find their natural interpretation in terms of high gluonic densities and also a good quantitative description in terms of cgc - based calculations , no conclusive claim for the observation saturation physics can be performed yet .
important steps have been taken over the last years in promoting the cgc framework to a predictive and quantitative phenomenological tool .
such has been possible through the systematic implementation of global fit and monte carlo methods and , more importantly , through an intense theoretical work in the determination of higher order corrections to the formalism , including running coupling corrections to non - linear evolution equations and also to particle production processes . nevertheless , this program is far for complete and there is still a large margin for improvement in the cgc phenomenological works
. the p+pb data will provide precious information to sharpen the cgc quantitative tool .
first of all it will provide empiric information needed to constrain the non - perturbative parameters of the theory . next it will allow to test the generic cgc predictions and also whether the present degree of accuracy of the cgc effective theory is sufficient to quantitatively describe data .
i would like to thank the organizers for their invitation to this very interesting conference .
my research is supported by a fellowship from the thorie lhc france initiative funded by the in2p3 .
n. armesto , ( ed . ) , et al .
, heavy ion collisions at the lhc - last call for predictions , j. phys .
g35 ( 2008 ) 054001 . http://arxiv.org/abs/0711.0974 [ ] , http://dx.doi.org/10.1088/0954-3899/35/5/054001 [ ] .
g. a. chirilli , b .- w .
xiao , f. yuan , one - loop factorization for inclusive hadron production in @xmath31 collisions in the saturation formalism , phys.rev.lett . 108
( 2012 ) 122301 .
[ ] , http://dx.doi.org/10.1103/physrevlett.108.122301 [ ] .
t. altinoluk , a. kovner , particle production at high energy and large transverse momentum - the hybrid formalism revisited , phys.rev .
d83 ( 2011 ) 105004 . http://arxiv.org/abs/1102.5327 [ ] , http://dx.doi.org/10.1103/physrevd.83.105004 [ ] .
w. horowitz , y. v. kovchegov , running coupling corrections to high energy inclusive gluon production , nucl.phys .
a849 ( 2011 ) 7297 . http://arxiv.org/abs/1009.0545 [ ] , http://dx.doi.org/10.1016/j.nuclphysa.2010.10.014 [ ] .
f. gelis , t. lappi , r. venugopalan , high energy factorization in nucleus - nucleus collisions , phys .
d78 ( 2008 ) 054019 . http://arxiv.org/abs/0804.2630 [ ] , http://dx.doi.org/10.1103/physrevd.78.054019 [ ] .
f. dominguez , c. marquet , b .- w .
xiao , f. yuan , universality of unintegrated gluon distributions at small x , phys.rev .
d83 ( 2011 ) 105005 . http://arxiv.org/abs/1101.0715 [ ] , http://dx.doi.org/10.1103/physrevd.83.105005 [ ] .
a. dumitru , j. jalilian - marian , e. petreska , two - gluon correlations and initial conditions for small - x evolution , phys.rev . d84 ( 2011 ) 014018 .
http://arxiv.org/abs/1105.4155 [ ] , http://dx.doi.org/10.1103/physrevd.84.014018 [ ] .
j. l. albacete , n. armesto , j. g. milhano , c. a. salgado , non - linear qcd meets data : a global analysis of lepton- proton scattering with running coupling bk evolution , phys .
d80 ( 2009 ) 034031 . http://arxiv.org/abs/0902.1112 [ ] , http://dx.doi.org/10.1103/physrevd.80.034031 [ ] .
j. l. albacete , n. armesto , j. g. milhano , p. quiroga arias , c. a. salgado , aamqs : a non - linear qcd analysis of new hera data at small - x including heavy quarks , eur.phys.j . c71 ( 2011 ) 1705 , * temporary entry * .
http://arxiv.org/abs/1012.4408 [ ] , http://dx.doi.org/10.1140/epjc/s10052-011-1705-3 [ ] .
j. l. albacete , c. marquet , single inclusive hadron production at rhic and the lhc from the color glass condensate , phys .
b687 ( 2010 ) 174179 . http://arxiv.org/abs/1001.1378 [ ] , http://dx.doi.org/10.1016/j.physletb.2010.02.073 [ ] .
j. l. albacete , c. marquet , azimuthal correlations of forward di - hadrons in d+au collisions at rhic in the color glass condensate , phys.rev.lett . 105
( 2010 ) 162301 .
http://arxiv.org/abs/1005.4065 [ ] , http://dx.doi.org/10.1103/physrevlett.105.162301 [ ] .
drescher , y. nara , eccentricity fluctuations from the color glass condensate at rhic and lhc , phys .
c76 ( 2007 ) 041903
. http://arxiv.org/abs/0707.0249 [ ] , http://dx.doi.org/10.1103/physrevc.76.041903 [ ] .
a. h. rezaeian , charged particle multiplicities in pa interactions at the lhc from the color glass condensate , phys.rev .
d85 ( 2012 ) 014028 . http://arxiv.org/abs/1111.2312 [ ] , http://dx.doi.org/10.1103/physrevd.85.014028 [ ] .
a. dumitru , d. e. kharzeev , e. m. levin , y. nara , gluon saturation in @xmath31 collisions at the lhc : kln model predictions for hadron multiplicities , phys.rev .
c85 ( 2012 ) 044920 . http://arxiv.org/abs/1111.3031 [ ] , http://dx.doi.org/10.1103/physrevc.85.044920 [ ] .
p. tribedy , r. venugopalan , qcd saturation at the lhc : comparisons of models to p+p and a+a data and predictions for p+pb collisions , phys.lett .
b710 ( 2012 ) 125133 . http://arxiv.org/abs/1112.2445 [ ] , http://dx.doi.org/10.1016/j.physletb.2012.02.047 [ ] .
b. z. kopeliovich , j. nemchik , i. k. potashnikova , m. b. johnson , i. schmidt , breakdown of qcd factorization at large feynman x , phys .
c72 ( 2005 ) 054606 . http://arxiv.org/abs/hep-ph/0501260 [ ] , http://dx.doi.org/10.1103/physrevc.72.054606 [ ] .
p. quiroga - arias , j. g. milhano , u. a. wiedemann , testing nuclear parton distributions with pa collisions at the tev scale , phys.rev .
c82 ( 2010 ) 034903 . http://arxiv.org/abs/1002.2537 [ ] , http://dx.doi.org/10.1103/physrevc.82.034903 [ ] .
a. dumitru , a. hayashigaki , j. jalilian - marian , the color glass condensate and hadron production in the forward region , nucl .
a765 ( 2006 ) 464482 . http://arxiv.org/abs/hep-ph/0506308 [ ] , http://dx.doi.org/10.1016/j.nuclphysa.2005.11.014 [ ] .
j. jalilian - marian , a. h. rezaeian , hadron production in pa collisions at the lhc from the color glass condensate , phys.rev .
d85 ( 2012 ) 014017 . http://arxiv.org/abs/1110.2810 [ ] , http://dx.doi.org/10.1103/physrevd.85.014017 [ ] .
m. strikman , w. vogelsang , multiple parton interactions and forward double pion production in pp and da scattering , phys .
d83 ( 2011 ) 034029 . http://arxiv.org/abs/1009.6123 [ ] , http://dx.doi.org/10.1103/physrevd.83.034029 [ ] .
kang , i. vitev , h. xing , dihadron momentum imbalance and correlations in d+au collisions , phys.rev .
d85 ( 2012 ) 054024 .
http://arxiv.org/abs/1112.6021 [ ] , http://dx.doi.org/10.1103/physrevd.85.054024 [ ] . j. jalilian - marian , a. h. rezaeian , prompt photon production and photon - hadron correlations at rhic and the lhc from the color glass condensate , phys.rev .
d86 ( 2012 ) 034016 . http://arxiv.org/abs/1204.1319 [ ] , http://dx.doi.org/10.1103/physrevd.86.034016 [ ] .
a. stasto , b .- w .
xiao , d. zaslavsky , drell - yan lepton - pair - jet correlation in pa collisions , phys.rev .
d86 ( 2012 ) 014009 . http://arxiv.org/abs/1204.4861 [ ] , http://dx.doi.org/10.1103/physrevd.86.014009 [ ] . | i present a brief review of present cgc phenomenological applications and of the physics prospects for the forthcoming proton - lead run at the lhc color glass condensate , high - density qcd , lhc besides its crucial role as a reference experiment to calibrate initial state effects in heavy ion collisions , the forthcoming p+pb run at the lhc will provide access to kinematic regions never explored so far in nuclear collisions and thus carries great potential for discovery of new qcd phenomena on its own . in particular , the huge leap forward in collision energy with respect to previous high energy electron - nucleus or proton - nucleus experiments will probe the nuclear wave function at values of bjorken-@xmath0 smaller than ever before .
it is theoretically well established that at small enough values of bjorken-@xmath0 qcd enters a novel regime governed by large gluon densities and non - linear coherence phenomena .
the color glass condensate ( cgc ) effective theory provides a consistent framework to study qcd scattering at small-@xmath0 or high collision energies ( for a review see e.g. @xcite ) .
it is based on three main physical ingredients : first , high gluon densities correspond to strong classical fields , which permit ab - initio first principles calculation of `` wave functions '' at small @xmath0 through classical techniques .
next , quantum corrections are incorporated via non - linear renormalization group equations such as the b - jimwlk hierarchy or , in the large-@xmath1 limit , the bk equation @xcite that describe the evolution of the hadron wave function towards small @xmath0 .
the non - linear , density - dependent terms in the cgc evolution equations are ultimately related to unitarity of the theory and , in the appropriate frame and gauge , can be interpreted as due to gluon recombination processes that tame or _
saturate _ the growth of gluon densities for modes with transverse momenta below a dynamically generated scale known as the saturation scale , @xmath2 .
finally , the presence of strong color fields @xmath3 leads to breakdown of standard perturbative techniques to describe particle production processes based on a series expansion in powers of the strong coupling @xmath4 .
terms of order @xmath5 need to be resummed to all orders .
the cgc provides the tools to perform such resummation although the precise prescription for the resummation may vary from process to process or colliding system while the cgc has been successfully applied to the description of different observables in different collision systems ( from e+p to aa ) , the p+pb run at the lhc will provide an excellent and probably in the near future unique possibility to disentangle the presently inconclusive situation on the role of cgc effects and also to distinguish among different approaches to describe high energy scattering in nuclear reactions . on the one hand , the lhc shall bring us closer to the limit of asymptotically high energy in which the cgc formalism is developed , thus reducing theoretical uncertainties on its applicability .
equivalently , the value of the saturation scale is expected to be a factor @xmath6 times larger than at rhic , so saturation effects should be visible in a larger range of transverse momenta , deeper into the perturbative domain . on the other hand , the much extended reach in the lhc will allow measurements far from the kinematic limit up to very forward rapidities , thus minimizing the role of large-@xmath0 effects which obscured the interpretation of forward rhic data . |
after the 2008 financial crisis , it has become clear that a better understanding of the mechanisms and dynamics underlying the networked worldwide economy is vital @xcite . among the possible channels of interaction among countries
, international trade plays a major role @xcite . combined together
, the worldwide trade relations can be interpreted as the connections of a complex network , the international trade network ( itn ) @xcite , whose understanding and modeling is one of the traditional goals of macroeconomics .
the standard model of non - zero trade flows , inferring the volume of bilateral trade between any two countries from the knowledge of their gross domestic product ( @xmath0 ) and mutual geographic distance ( @xmath1 ) , is the so - called ` gravity model ' of trade @xcite . in its simplest form ,
the gravity model predicts that the volume of trade between countries @xmath2 and @xmath3 is @xmath4 note that the above expression , as well as the rest of this paper , focuses on the undirected version of the network for simplicity . in this representation
, the trade from country @xmath2 to country @xmath3 and the trade rom country @xmath3 to country @xmath2 are combined together .
given the highly symmetric structure of the itn , this simplification retains all the basic network properties of the system @xcite . in its simplest form
, the gravity model is fitted on the non - zero weights observed between all pairs of _ connected _ countries .
this means that the model can predict the pair - specific volume of trade only _
after _ the presence of the trade relation itself has been established @xcite .
this intrinsic limitation is alarming , since almost half of the links in the itn are missing @xcite .
although several improvements and generalizations of the standard gravity model have been proposed to overcome this problem ( see @xcite for excellent reviews ) , so far none of them succeeded in reproducing the observed complex topology and the observed volumes simultaneously .
moreover , the various attempts have not been conceived under a unique theoretical framework and are therefore based on the combination of different mechanisms ( e.g. one for establishing the presence of a trade relation , and one for establishing its intensity ) . in general , the challenge of successfully predicting , via only one mechanism , both trade probabilities and trade volumes remains an open problem . over the past years
, the problem of replicating the observed structure of the itn has been extensively approached using network models @xcite and , more indirectly , maximum - entropy techniques to reconstruct networks from partial information @xcite .
these studies have focused both on the purely binary architecture ( defined solely by the existence of trade exchanges between world countries ) @xcite and on the weighted structure ( when also the magnitude of these interactions is taken into account ) @xcite .
what clearly emerges is that both topological and weighted properties of the network are deeply connected with purely macroeconomic properties ( in particular the gdp ) governing bilateral trade volumes @xcite .
however , it has also been clarified that , while the knowledge of the degree sequence ( i.e. the number of trade partners for each country ) allows to infer the the entire binary structure of the network with great accuracy @xcite , the knowledge of the strength sequence ( i.e. the total volume of trade for each country ) gives a very poor prediction of all network properties @xcite . indeed , the network inferred only from the strength sequence has a trivial topology , being much denser ( if integer link weights are assumed @xcite ) or even fully connected ( if continuous weights are assumed @xcite ) , and in any case much more homogeneous than the empirical one .
this limitation leads back to the main drawback of the gravity model .
indeed , it has been shown that a simplified version of the gravity model ( with @xmath5 and @xmath6 ) can be recovered as a particular case of a maximum - entropy model with given strength sequence ( and continuous link weights ) @xcite .
combined together , the high informativeness of the degree sequence for the binary representation of the itn and the low informativeness of the strength sequence for its weighted representation contradict the naive expectation that , once aggregated at the country level , weighted structural properties ( the strengths ) are _ per se _ more informative than purely binary properties ( the degrees ) .
this puzzle has generated further interest around the challenge of finding a unique mechanism predicting link probabilities and link weights simultaneously . as a step forward in this direction , an improved reconstruction approach @xcite , based on an analytical maximum - likelihood estimation method @xcite , has been recently proposed in order to construct more sophisticated maximum - entropy ensembles of weighted networks .
this approach builds on previous mathematical results @xcite characterizing a network ensemble where both the degree and the strength sequences are constrained .
the graph probability is the so - called generalized bose - fermi distribution @xcite , and the resulting network model goes under the name of enhanced configuration model ( ecm ) @xcite . when used to reconstruct the properties of several empirical networks , the ecm shows a significant improvement with respect to the case where either only the degree sequence ( binary configuration model , bcm for short ) or only the strength sequence ( weighted configuration model , wcm for short ) is constrained .
one therefore expects that combining the knowledge of strengths and degrees is precisely the ingredient required in order to successfully reproduce the itn from purely local information .
indeed , a more recent study has showed that , when applied to international trade data ( both aggregated and commodity - specific ) , the method successfully reproduces the key properties of the itn , across different years and for different levels of aggregation ( i.e. for different commodity - specific layers ) @xcite .
however , in itself the ecm is a network reconstruction method , rather than a genuine model of network formation . to turn it into a proper network model for the itn structure
, it would be necessary to find a macroeconomic interpretation for the underlying variables involved in the method .
this operation would correspond to what has already been separately performed at a purely binary level ( by identifying a strong relationship between the gdp and the variable controlling the degree of a country in the bcm @xcite ) and at a purely weighted level ( by finding a relationship between the gdp and the variable controlling the strength of a country in the wcm @xcite , in the same spirit of the gravity model ) . generalizing the above results to the combination of strengths and degrees
is not obvious , given the different mathematical expressions characterizing the ecm , the bcm and the wcm .
in this paper we show that , indeed , the variables of the ecm are all strongly correlated with the gdp .
this result gives a macroeconomic interpretation of the parameters values satisfying the itn constraints .
reversing the perspective , this result enables us to introduce the first gdp - driven model that successfully reproduces the binary and the weighted properties of the itn simultaneously .
finally , we show that the ecm model can be replaced by a simpler , two - step model that reconciles the binary projection of the ecm model with the topology predicted by the bcm .
these results represent a promising step forward in the formulation of a unified model for the structure of the itn .
since our results are a generalization of previous maximum - entropy approaches to the characterization of the itn , in this section we first briefly review the main results of those approaches , while our new findings are presented in the next section . in so doing , we gradually introduce the mathematical building blocks of our analysis and illustrate our main motivations .
moreover , since previous studies have used different data sets , we also recalculate the quantities of interest on the same data set that we will use later for our own investigation .
this allows us to align the results of previous approaches and properly compare them with our new findings .
we use yearly bilateral data on exports and imports from the united nations commodity trade database ( un comtrade ) @xcite from year 1992 to 2004 .
the sample refers to 13 years , 1992 - 2004 , represented in current u.s .
dollars , and disaggregated over 97 commodity classes . in this paper
we analyze the aggregated level , which results in 13 yearly temporal snapshots of undirected total trade flows .
our network consists of @xmath7 countries , present in the data throughout the considered temporal interval .
+ this data set was the subject of many studies exploring both purely the binary representation , and its full weighted representation @xcite .
another data set which is widely used to represent the itn network is the trade data collected by gleditsch @xcite .
the data contain the detailed list of bilateral import and export volumes , for each country in the period 1950 - 2000 .
if one focuses solely on the binary undirected projection of the itn , then the binary configuration model ( bcm ) represents a very successful maximum - entropy model . in the bcm , the local knowledge of the number of trade partners of each country , i.e. the degree sequence ,
is specified .
it has been shown that higher - order properties of the itn can be simply traced back to the knowledge of the degree sequence @xcite .
this result adds considerable information to the standard results of traditional macroeconomic analyses of international trade .
in particular , it suggests that the degree sequence , which is a purely topological property , needs to be considered as an important target quantity that international trade models , in contrast with the mainstream approaches in economics , should aim at reproducing @xcite .
let us first represent the observed structure of the itn as a weighted undirected network specified by the square matrix @xmath8 , where the specific entry @xmath9 represents the weight of the link between country @xmath2 and country @xmath3 .
then , let us represent the binary projection of the network in terms of the binary adjacency matrix @xmath10 , with entries defined as @xmath11,\:\forall\:{i , j}$ ] .
a maximum - entropy ensemble of networks is a collection of graphs where each graph is assigned a probability of occurrence determined by the choice of some constraints .
the bcm is a maximum - entropy ensemble of binary graphs , each denoted by a generic matrix @xmath12 , where the chosen constraint is the degree sequence . in the canonical formalism @xcite
, the latter can be constrained by writing the following hamiltonian : @xmath13 where the degree sequence is defined as @xmath14,\:\forall\:{i}$ ] . as a result of the constrained maximization of the entropy @xcite
, the probability of a given configuration @xmath12 can be written as @xmath15^{a_{ij}}\left [ \frac{1}{1+z_iz_j}\right]^{1-a_{ij}}\ ] ] where @xmath16 and @xmath17 .
the latter represents the probability of forming a link between nodes @xmath2 and @xmath3 , which is also the expected value @xmath18 according to the maximum - likelihood method proposed in @xcite , the vector of unknowns @xmath19 can be numerically found by solving the system of @xmath20 coupled equations @xmath21 where the expected value of each degree @xmath22 is matched to the observed value @xmath23 in the real network @xmath24 .
the ( unique ) solution will be indicated as @xmath25 . when inserted back into eq.([eq_pzz ] ) , this solution allows to analytically describe the binary ensemble matching the observed constraints .
being the result of the maximization of the entropy , this ensemble represents the least biased estimate of the network structure , based only on the knowledge of the empirical degree sequence . in fig .
[ fig1 ] we plot some higher - order topological properties of the itn as a function of the degree of nodes , for the 2002 snapshot .
these properties are the so - called average nearest neighbor degree and the clustering coefficient . for both quantities ,
we plot the observed values ( red points ) and the corresponding expected values predicted by the bcm ( blue points ) .
the exact expressions for both empirical and expected quantities are provided in the appendix .
we see that the expected values are in very close agreement with the observed properties .
these results replicate recent findings @xcite based on the same un comtrade data .
they show that at a binary level , the degree correlations ( disassortativity ) and clustering structure of the itn are excellently reproduced by the bcm .
as we also confirmed in the present analysis , these results were found to be very robust , as they hold true over time and for various resolutions ( i.e. , for different levels of aggregation of traded commodities ) @xcite .
it should be noted that eq.([eq_pzz ] ) can be thought of as a particular case of the so - called _ fitness model _
@xcite , which is a popular model of binary networks where the connection probability @xmath26 is assumed to be a function of the values of some ` fitness ' characterizing each vertex .
indeed , the variables @xmath25 can be treated as fitness parameters @xcite which control the probability of forming a link . a very interesting correlation between a fitness parameter of a country ( assigned by the model ) and the @xmath0 of the same country
was found @xcite .
this relation is replicated here in fig .
[ fig2 ] , where the rescaled @xmath0 of each country ( @xmath27 ) is compared to the value of the fitness parameter @xmath28 obtained by solving eq.([maxlike ] ) .
the red line is a linear fit of the type @xmath29 the calculated @xmath30 , compared with the @xmath31 ( re - scaled gdp ) for each country for the undirected binary aggregated itn in the 2002 snapshot , with the linear fit @xmath32 ( red line ) . ]
this leads to a more economic interpretation where the fitness parameters can be replaced ( up to a proportionality constant ) with the @xmath0 of countries , and used to reproduce the properties of the network .
this procedure , first adopted in @xcite , can give predictions for the network based only on macroeconomic properties of countries , and reveals the importance of the @xmath0 in to the binary structure of the itn .
importantly , this observation was the first empirical evidence in favour of the fitness model as a powerful network model @xcite .
likewise , other studies have shown that the observed topological properties turn out to be important in explaining macroeconomics dynamics @xcite . despite the importance of the topology ,
the latter is only the backbone over which goods are traded , and the knowledge of the volume of such trade is extremely important . to be able to give predictions about the weight of connections
, one needs to switch from an ensemble of binary graphs to one of weighted graphs .
the simplest weighted counterpart of the bcm is the wcm , which is a maximum - entropy ensemble of weighted networks where the constraint is the strength sequence , i.e. the total trade of each country in the case of the itn .
recent studies have shown that the higher - order binary quantities predicted by the wcm , as well as the corresponding weighted quantities , are very different from the observed counterparts @xcite .
more specifically , the main limitation of the model is that of predicting a largely homogeneous and very dense ( sometimes fully connected ) topology .
roughly speaking , the model excessively ` dilutes ' the total trade of each country by distributing it to almost all other countries .
this failure in correctly replicating the purely topological projection of the real network is the root of the bad agreement between expected and observed higher - order properties .
just like the bcm has been related to the fitness model @xcite , a variant of the wcm has been related to the gravity model @xcite .
the variant is actually a continuous version of the wcm , where the strength sequence is constrained and the weights are real numbers instead of integers .
when applied to the itn , the model gives the following expectation for the weight of the links : @xmath33 where @xmath34 is the total strength in the network , and @xmath31 is the re - scaled @xmath0 as before @xcite .
in essence , the above expression identifies again a relationship between the @xmath0 and the hidden variable ( analogous to the fitness in the binary case ) specifying the strength of a node . equation ( [ eq_ff ] ) coincides with eq.([eq_gravity ] ) where @xmath5 and @xmath6 .
the model therefore corresponds to a particularly simple version of the gravity model .
indeed , the model reproduces reasonably well the observed non - zero weights of the itn @xcite .
however , just like the gravity model , the model predicts a complete graph where @xmath35 , and dramatically fails in reproducing the binary architecture of the network . this can be easily shown by realizing that the continuous nature of edge weights , which can take non - negative real values in the model , implies that there is a zero probability of generating zero weights ( i.e. missing links ) .
we will show the prediction of this model in comparison with our results later on in the paper .
motivated by the challenge to satisfactorily model both the topology and the weights of the itn , the ecm has been recently proposed as an improved model of this network @xcite .
the ecm focuses on weighted networks , and can enforce the degree and strength sequence simultaneously @xcite .
it builds on the so - called generalized bose - fermi distribution that was first introduced as a null model of networks with coupled binary and weighted constraints @xcite . in the ecm
, the degree and strength sequence can be constrained by writing the following hamiltonian : @xmath36\ ] ] where the strength sequence is defined as @xmath37 and the degree sequence as @xmath38,\:\forall\:i$ ] . as a result
, the probability of a given configuration @xmath39 can be written as @xmath40 with @xmath41 and @xmath42 .
the ecm gives the following predictions about the probability of a link ( @xmath43 ) and the expected weight of the link ( @xmath44 ) : @xmath45 according to the maximum - likelihood method proposed in @xcite , the vectors of unknowns @xmath46 and @xmath47 can be numerically found by solving the system of @xmath48 coupled equations @xmath49 and will be indicated as @xmath50 and @xmath51 .
these unknown parameters can be treated as fitness parameters which control the probability of forming a link and the expected weight of that link simultaneously .
the application of the ecm to various real - world networks shows that the model can accurately reproduce the higher - order empirical properties of these networks @xcite .
when applied to the itn in particular , the ecm replicates both binary and weighted empirical properties , for different levels of disaggregation , and for several years ( temporal snapshots ) @xcite . indeed , in fig .
[ fig3 ] we show the higher - order binary quantities ( average nearest neighbor degree and clustering coefficient ) as well as their weighted ones ( average nearest neighbor strength and weighted clustering coefficient ) for the 2002 snapshot of the itn .
we compare the observed values ( red points ) and the corresponding quantities predicted by the ecm ( blue points ) .
the mathematical expressions for all these quantities are provided in the appendix .
we find a very good agreement between data and model , confirming the recent results in @xcite for the data set we are using here .
we also confirmed that these results are robust for several temporal snapshots @xcite .
considering the promising results of the ecm , we now make a step forward and check whether the hidden variables @xmath52 and @xmath53 , which effectively reproduce the observed itn , can be thought of as ` fitness ' parameters having a clear economic interpretation .
this amounts to check whether the relation shown previously in fig .
[ fig2 ] for the purely binary case can be generalized in order to find a macroeconomic interpretation to the abstract fitness parameters in the general weighted case as well .
the calculated @xmath52 ( left panel ) and @xmath54 ( right panel ) compared with the @xmath31 ( rescaled gdp ) for each country for the undirected binary aggregated itn in the 2002 snapshot , with the linear fits ( in log - log scale ) @xmath55 , and @xmath56 ( red lines ) , where a , b , and c are the fitted constant parameters per year . ] in fig .
[ fig4 ] we show the relationship between the two parameters @xmath52 and @xmath53 and the rescaled @xmath0 ( @xmath31 ) for each country of the itn in the 2002 snapshot .
we find strong correlations between these quantities .
the fitness parameter @xmath52 turns out to be in a roughly linear relation with the rescaled gdp @xmath31 , fitted by the curve @xmath57 where @xmath58 is the fitted constant , and @xmath59 ( all the @xmath0s are relative to that specific year ) .
it should be noted that this relation is similar to that found between @xmath30 and @xmath31 in the bcm and shown previously in fig.[fig2 ] , but less accurate .
this observation will be useful later .
by contrast , since the @xmath0 is an unbounded quantity , while the fitness parameter @xmath53 is bounded between 0 and 1 ( this is a mathematical property of the model @xcite ) , the relation between @xmath53 and @xmath31 is necessarily highly nonlinear .
a simple functional form for such a relationship is given by @xmath60 indeed , fig .
[ fig4 ] confirms that the above expression provides a very good fit to the data .
we checked that the above results hold systematically over time , for each snapshot of the itn in our data set .
this implies that , in a given year , we can insert eqs.([eq_xx ] ) and ( [ eq_yy ] ) into eqs.([eq_pp ] ) and ( [ eq_ww ] ) to obtain a gdp - driven model of the itn structure for that year .
such a model highlights that the @xmath0 has a crucial role in shaping both the binary and the weighted properties of the itn .
while this was already expected on the basis of the aforementioned results obtained using the bcm and the wcm ( or the corresponding simplified gravity model ) separately , finding the appropriate way to explicitly combine these results into a unified description of the itn has remained impossible so far .
rather than exploring in more detail the predictions of the gdp - driven model in the form described above , we first make some considerations leading to a simplification of the model itself . at this point
, it should be noted that we arrived at two seemingly conflicting results .
we showed that both the bcm and ecm give a very good prediction for the binary topology of the network .
however , eqs.([eq_pzz ] ) and ( [ eq_pp ] ) , which specify the connection probability @xmath26 in the two models , are significantly different .
the comparable performance of the bcm and the ecm at a binary level ( see figs.[fig1 ] and [ fig3 ] ) makes us expect that , when the specific values @xmath25 and @xmath50 are inserted into eqs.([eq_pzz ] ) and ( [ eq_pp ] ) respectively , the values of the connection probability become comparable in the two models , despite the different mathematical expressions . in fig .
[ fig5 ] we compare the the two probabilities for the itn in the 2002 snapshot .
note that each point refers to the probability of creating a link between a pair of countries , which results in @xmath61 points .
indeed , we can see that the values are scattered along the identity line , confirming the expectation that the connection probability has similar value in the two models .
the above result allows us to make a remarkable simplification . in eqs.([eq_pp ] ) and ( [ eq_ww ] ) , we can replace the expression for @xmath26 provided by the ecm with that provided by the bcm in eq.([eq_pzz ] ) . to avoid confusion ,
we denote the new probability with @xmath62 , where @xmath63 stands for ` two - step ' , for a reason that will be clear immediately .
this results in the following equations for the expected network properties : @xmath64 where the @xmath30 s , and therefore the @xmath65 s , depend only on the degrees through eq.([maxlike ] ) , while the @xmath53 s and the @xmath66 s depend on both strengths and degrees through eqs.([eq_kk ] ) and ( [ eq_ss ] ) . in this simplified model the connection probability , which fully specifies
the topology of the ensemble of networks , no longer depends on the strengths as in the ecm , while the weights still do .
this implies that we can specify the model via a two - step procedure where we first solve the @xmath20 equations determining @xmath62 via the degrees , and then find the remaining variables determining @xmath66 through the ecm .
for this reason , we denote the model as the two - step ( ts ) model .
the probability of forming a link in the ecm @xmath67 compared to the probability in the bcm @xmath68 for the undirected binary aggregated itn in the 2002 snapshot .
the red line describes the identity line . ]
the probability of a configuration @xmath39 reads @xmath69 where @xmath70 is the probability that a link of weight @xmath71 connects the nodes ( countries ) @xmath2 and @xmath3 .
the above probability has the same general expression as in the original ecm @xcite , but here @xmath30 comes from the estimation of the simpler bcm .
it is instructive to rewrite ( [ prob ] ) as @xmath72 to highlight the random processes creating each link . as a first step
, one determines whether a link is created or not with a probability @xmath62 .
if a link ( of unit weight ) is indeed established , a second attempt determines whether the weight of the same link is increased by another unit ( with probability @xmath73 ) or whether the process stops ( with probability @xmath74 ) .
iterating this procedure , the probability that an edge with weight @xmath75 is established between nodes @xmath2 and @xmath3 is given precisely by @xmath76 in eq.([prob2 ] ) .
the expected weight @xmath66 is then correctly retrieved as @xmath77 . using the relations found in eqs.([eq_zz ] ) and ( [ eq_yy ] ) , we can input the @xmath31 as the fitness parameters into eqs.([eq_pts ] ) and ( [ eq_wts ] ) to get the following expressions that mathematically characterize our gdp - driven specification of the ts model : @xmath78 the above equations can be used to reverse the approach used so far : rather than using the @xmath48 free parameters of the ecm ( @xmath46 and @xmath47 ) or of the ts model ( @xmath19 and @xmath47 ) to fit the models on the observed values of the degrees and strengths
, we can now use the knowledge of the gdp of all countries to obtain a model that only depends on the three parameters @xmath79 , @xmath80 , @xmath81 . assigning values to these parameters
can be done using two techniques : maximization of the likelihood function and non - linear curve fitting . since the model is a two - step one , we can first assign a value to the parameter @xmath79 , and only in the second step ( once @xmath79
is set ) we fit the parameters @xmath80 and @xmath81 .
we chose to fix @xmath79 by maximizing the likelihood function @xcite , which results in constraining the expected number of links to the observed number ( @xmath82 ) , as in @xcite .
fixing the values of @xmath80 and @xmath81 is slightly more complicated .
since the model uses the approximated expressions of the ts model , rather than those of the original ecm model , maximizing the likelihood function in the second step no longer yields the desired condition @xmath83 , where @xmath34 is the total strength in the network .
similarly , extracting the parameters from the fit as shown in fig .
[ fig4 ] does not maintain the total strength in the network .
in absence of any a - priori preference , we chose the latter procedure , due to its relative numerical simplicity with respect to the former one . in fig .
[ fig6 ] we show a comparison between the higher - order observed properties of the itn in 2002 and their expected counterparts predicted by the gdp - driven ts model .
again , the mathematical expressions of these properties are provided in the appendix . as a baseline comparison ,
we also show the predictions of the @xmath0-driven wcm model with continuous weights described by eq.([eq_ff ] ) @xcite , which coincides with a simplified version of the gravity model as we mentioned .
we see that the gdp - driven ts model reproduces the empirical trends very well .
of course , as expected , the predictions in fig.[fig6 ] ( which use only three free parameters ) are more noisy than those in fig.[fig3 ] ( which use @xmath48 free parameters ) .
this is due to the fact that eqs.([eq_zz ] ) and ( [ eq_yy ] ) describe fitting curves rather than exact relationships .
importantly , our model performs significantly better than the wcm / gravity model in replicating both binary and weighted properties .
again , the drawback of these models lies in the fact that they predict a fully connected topology and a relatively homogeneous network .
it should also be noted that the plot of average nearest neighbor strength ( @xmath84 ) predicted by our model is slightly shifted with respect to the observed points .
this effect is due to the fact that , as we mentioned , the total strength @xmath34 ( hence the average trend of the @xmath84 ) is only approximately reproduced by our model , as a result of the simplification from the ecm to the ts model . as for all the other results in this paper
, we checked that our findings are robust over the entire time span of our data set .
we therefore conclude that the ecm model , as well as its simplified ts variant , can be successfully turned into a fully gdp - driven model that simultaneously reproduces both the topology and the weights of the itn .
+ the success of the ts model has an important interpretation .
looking back at eqs.([eq_pts ] ) and ( [ twostep1 ] ) , we recall that the effect of the ts approximation is the fact that the connection probability @xmath62 can be estimated separately from the weights @xmath66 , using only the knowledge of the degree sequence if eq.([eq_pts ] ) is used , or the gdp and total number of links if eq.([twostep1 ] ) is used , while discarding that of the strengths . by contrast
, the estimation of the expected weights @xmath66 can not be carried out separately , as it requires that the connection probability @xmath62 appearing in eqs.([eq_wts ] ) and ( [ twostep2 ] ) is estimated first .
this asymmetry of the model means that the topology of the itn can be successfully inferred without any information about the weighted properties , while the weighted structure can not be inferred without topological information .
the expressions defining the ts model provide a mathematical explanation for this otherwise puzzling effect that has already been documented in previous analyses of the itn @xcite .
in this paper we have introduced a novel @xmath0-driven model which successfully reproduces both the binary and weighted properties of the itn .
the model uses the @xmath0 of countries as a sort of macroeconomic fitness , and reveals the existence of strong relations between the @xmath0 and the model parameters controlling the formation and the volume of trade relations . in the light of the limitations of the existing models ( most notably the binary - only nature of the fitness model and the weighted - only nature of the gravity model ) , we believe that our results represent a promising step forward in the development of a unified model of the itn structure .
we have also shown that the full ecm model can be effectively reduced to the simpler ts model .
the success of the latter provides a mathematical explanation of an otherwise puzzling asymmetry , namely the fact that purely topological properties can be successfully predicted without knowing the weights , while weighted network properties can only be predicted if the topological ones are preliminary estimated .
future work should explore how to further improve these results and possibly expand them by introducing additional macroeconomic parameters like geographic distance , thus fully bridging the gap between network - based and gravity - based approaches to the structure of international trade .
aa and dg acknowledge support from the dutch econophysics foundation ( stichting econophysics , leiden , the netherlands ) with funds from beneficiaries of duyfken trading knowledge bv , amsterdam , the netherlands .
this work was also supported by the eu project multiplex ( contract 317532 ) and the netherlands organization for scientific research ( nwo / ocw ) .
in this appendix we give a summarized description of the binary and weighted network quantities which are studied in this paper .
specifically , we first show how the properties are measured over a real network , and then how the expected values under the ecm and the ts model are constructed .
let us note a weighted undirected network as a square matrix @xmath39 , where the specific entry @xmath71 represents the edge weight between country @xmath2 and country @xmath3
. the binary representation of the network is noted by a binary matrix @xmath12 , where the entries are @xmath85,\:\forall\:{i , j}$ ] . + we compute the average nearest neighbor degree as : @xmath86 its calculated as the arithmetic mean of the degrees of the neighbors of a specific node , which is a measure of correlation between the degrees of adjacent nodes .
+ the binary clustering coefficient has the following expression : @xmath87 it is a measure of the tendency to which nodes in a graph form cluster together .
more specifically , it counts how many closed triangles are attached to each node with respect to all the possible triangles .
+ the corresponding weighted properties are the average nearest neighbor strength and the weighted clustering coefficient . the average nearest neighbor strength , defined as : @xmath88 where @xmath89 is the strength ( total flow ) of a country .
the @xmath90 measure the average strength of the neighbors for a specific node @xmath2 . like its binary counterpart
, it gives the magnitude of activity of a specific node neighbors ( weighted activity ) .
the weighted clustering coefficient @xcite defined as : @xmath91 the @xmath92 is a measure of the weight density in the neighborhood of a node .
it classify the tendency of a specific node to cluster in a triangle taking into account also the edge - values .
+ now , the measured properties of the real network need to be compared with the reproduced properties of the different models .
these reproduced properties are the expected values of the maximum entropy ensemble that each model i d generating , and can be calculated analytically .
the expected values can be obtained by simply replacing @xmath93 with the probability @xmath26 for the different models ( @xmath26 is different to each model ) .
this next step is what we will discuss in the next sections .
since the bcm model is only dealing with the binary representation , we will have expected values just for the two binary higher - order properties . while the ecm gives expectations for the weighted counterparts of the binary properties .
+ for the binary higher - order properties , we replace @xmath93 with @xmath26 which is the probability of creating a link , and also the expected value of the edge @xmath94 .
this simple procedure yields the analytic formula of the expected value for the properties .
we compute the expected average nearest neighbor degree as : @xmath95 in the weighted case ( weighted higher - order properties ) , we are left only with the ecm .
the expected average nearest neighbor strength is calculated as : @xmath99 where @xmath100 and we input the @xmath26 of the ecm model as before .
+ in the expected value of the @xmath101 we should be more careful , since it is necessary to calculate the expected product of ( powers of ) distinct matrix entries @xmath102 here again we use the known expressions for the properties and replacing the terms @xmath65 and @xmath108 with the expected values @xmath109 and @xmath44 correspondingly .
however , here the expected values are a function of the @xmath0 of the countries , or more specifically the re - scaled @xmath0 @xmath31 .
the expresions for the higher - order binary properties are as before : @xmath110 and @xmath111 where @xmath112 . for convenience reasons we will write the expression for the weighted clustering coefficient @xmath101 first as a function of the fitness parameters @xmath30 and @xmath53 , and later replaced them with the corresponding @xmath0 terms .
the expected the expected value of the @xmath101 in the two - step case is : @xmath115 as before we observe that @xmath103 t. squartini and d. garlaschelli , _ jan tinbergen s legacy for economic networks : from the gravity model to quantum statistics _ , in econophysics of agent - based models , pp . 161 - 186 ( springer international publishing , 2014 ) .
g. fagiolo , _ the international - trade network : gravity equations and topological properties _ , journal of economic interaction and coordination * 5(1 ) * , 1 - 25 ( 2010 ) .
musmeci , n. , battiston , s. , caldarelli , g. , puliga , m. , gabrielli , a. _ bootstrapping topological properties and systemic risk of complex networks using the fitness model _ , journal of statistical physics , * 151 * , 720 - 734 ( 2013 ) .
caldarelli , g. , chessa , a. , pammolli , f. , gabrielli , a. , and puliga , m _ reconstructing a credit network _ nature physics , * 9 * , 125 - 126 ( 2013 ) .
r. mastrandrea , t. squartini , g. fagiolo , d. garlaschelli , _ enhanced reconstruction of weighted networks from strengths and degrees _ , new j. phys . * 16 * , 043022 ( 2014 ) .
f. picciolo , t. squartini , f. ruzzenenti , r. basosi , d. garlaschelli , _ the role of distances in the world trade web _ , proceedings of the eighth international conference on signal - image technology & internet - based systems ( sitis 2012 ) , pp . 784 - 792 ( edited by ieee ) ( 2013 ) . | recent events such as the global financial crisis have renewed the interest in the topic of economic networks . one of the main channels of shock propagation among countries is the international trade network ( itn ) .
two important models for the itn structure , the classical gravity model of trade ( more popular among economists ) and the fitness model ( more popular among networks scientists ) , are both limited to the characterization of only one representation of the itn .
the gravity model satisfactorily predicts the volume of trade between connected countries , but can not reproduce the observed missing links ( i.e. the topology ) . on the other hand
, the fitness model can successfully replicate the topology of the itn , but can not predict the volumes .
this paper tries to make an important step forward in the unification of those two frameworks , by proposing a new gdp - driven model which can simultaneously reproduce the binary and the weighted properties of the itn .
specifically , we adopt a maximum - entropy approach where both the degree and the strength of each node is preserved .
we then identify strong nonlinear relationships between the gdp and the parameters of the model .
this ultimately results in a weighted generalization of the fitness model of trade , where the gdp plays the role of a ` macroeconomic fitness ' shaping the binary and the weighted structure of the itn simultaneously .
our model mathematically highlights an important asymmetry in the role of binary and weighted network properties , namely the fact that binary properties can be inferred without the knowledge of weighted ones , while the opposite is not true . |
the cepheid period - luminosity ( pl ) relation continues to play a key role in the determination of distances within the local group and to nearby galaxies .
the absolute calibration of the pl relation requires reliable distance measurements for calibrating cepheids as well as accurate corrections for the effects of interstellar reddening and extinction .
galactic cepheids are often heavily obscured , the average colour excess e(_b v _ ) being of order 0.5 mag .
here we propose a new method of accurate colour excess determination that relies on spectroscopically determined stellar parameters .
the traditional methods of establishing the interstellar reddening of cepheids have involved field reddenings derived from photometric and spectroscopic studies of stars surrounding the cepheid , the use of reddening - independent spectroscopic indices , and period - colour relations calibrated from spectroscopic analyses of a number of well - studied cepheids or supergiants .
field reddenings are fairly reliable provided that the surrounding stars are unaffected by circumstellar reddening , although the number of suitable objects has tended to be restricted to cluster cepheids and visual binaries ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
reddening - independent indices include the @xmath8 index @xcite , which measures the depression in stellar spectra caused by the _ g _ band of ch ( @xmath9 ) , and the _ khg _ index @xcite , which combines narrow band photometry of ca ii k ( @xmath10 ) , balmer h@xmath11 ( @xmath12 ) , and the _ g _ band .
narrow band photometry of balmer h@xmath13 ( @xmath14 ) has also been used as a temperature indicator for fgk stars , but has been susceptible to problems arising from colour dependences introduced by mismatches in the effective wavelengths of the narrow and wide h@xmath13 filters employed @xcite .
most cepheid reddenings have been derived from period - colour relations that are tied to specific cepheid or supergiant calibrators , in which specific continuum or spectral line features ( line ratios , for example ) are linked to stellar atmosphere models to infer effective temperatures for the stars ( e.g. , * ? ? ?
* ; * ? ? ?
the direct link between effective temperature and broad band ( _ b
v_)@xmath15 colour has been demonstrated very effectively by @xcite for older , published , stellar atmosphere models , and has been used successfully to test the open cluster calibration of the pl relation @xcite . as argued by @xcite ,
_ b v _ colours for cepheids appear to be more closely indicative of stellar temperatures than indices that include a near - infrared magnitude .
but recent changes to stellar atmosphere models @xcite may affect the effective temperature scale for fgk supergiants significantly enough that a recalibration is necessary .
that is the purpose of the present study .
cepheids also undergo significant changes in effective temperature during their pulsation cycles , which makes it essential to track such changes accurately through close monitoring of individual variables .
photometric monitoring of cepheids has resulted in a large database of published results for many different colour systems @xcite , but the availability of high resolution spectra for such stars throughout their cycles has been extremely limited , until now .
a goal of the present study is therefore to exploit a recently developed collection of high resolution spectra for galactic cepheids in order to present accurate intrinsic parameters for all of the studied objects .
the spectra of the fgk supergiants studied here were obtained using the 1.93 m telescope of the haute - provence observatoire ( france ) equipped with the echelle spectrograph elodie @xcite and retrieved from its on - line archive of spectra @xcite .
the resolving power for the observations was @xmath16 over the wavelength interval 44006800 , with a signal - to - noise ratio for each spectrum of s / n@xmath17100 ( at 5500 ) .
initial processing of the spectra ( image extraction , cosmic ray removal , flatfielding , etc . ) was carried out as described by @xcite .
we also made use of spectra obtained with the ultraviolet - visual echelle spectrograph ( uves ) instrument at the very large telescope ( vlt ) unit 2 kueyen @xcite .
all supergiants were observed in two instrumental modes , @xmath18 ( dic1 ) and @xmath19 ( dic2 ) , in order to provide almost complete coverage of the wavelength interval @xmath20 . the spectral resolution is about 80000 , and for most of the spectra the typical s / n ratio is 300500 in the _ v _ band . for classical cepheids in our sample we have used our previously published results @xcite .
multiphase observations of program cepheids were carried out using the struve 2.1-m reflector and sandiford echelle spectrograph at mcdonald observatory .
the nominal resolving power was 60000 with a total spectral range covering about @xmath21 .
the primary wavelength region was centered at 6200 .
exposure times were selected to ensure an optimal s / n ratio of about 100 .
additional processing of the spectra ( continuum level location , measuring of line depths and equivalent widths ) was carried out using the dech20 software @xcite .
line depths r@xmath22 were measured by means of gaussian fitting .
effective temperatures for our program stars were established from the processed spectra using the method developed by @xcite that is based upon @xmath0line depth relations
. the technique can establish @xmath0 with exceptional precision .
it relies upon the ratio of the central depths of two lines that have very different functional dependences on @xmath0 , and uses tens of pairs of lines for each spectrum .
the method is independent of interstellar reddening , and only marginally dependent on the individual characteristics of stars , such as rotation , microturbulence , metallicity , etc .
briefly , a set of 131 line ratio@xmath0 relations was employed , with the mean random error in a single calibration being 6090k ( 4045k in most cases and 9095k in the least accurate cases ) .
the use of @xmath2370100 calibrations per spectrum reduces the uncertainty in @xmath0 to 1020k for spectra with s / n ratios greater than 100 , and 3050k for spectra with s / n ratios less than 100 .
although the internal error for each @xmath0 determination appears to be small , a systematic shift of the zero - point in the @xmath0 scale may exist .
nevertheless , an uncertainty in zero - point ( if it exists ) can affect the absolute abundances derived for each program star .
it is relatively unimportant for abundance comparisons between stars in the sample . for the majority of supergiants and cepheids we obtain error estimates for @xmath0 that are smaller than 1040 k. the microturbulent velocities , @xmath2 , and surface gravities , @xmath1 , were derived using a modification of the standard analysis proposed by @xcite .
as described there , the microturbulence is determined from the fe ii lines rather than the fe i lines , as in classical abundance analyses .
the surface gravity is established by forcing equality between the total iron abundance obtained from both fe i and fe ii lines . typically with this method
the iron abundance determined from fe i lines shows a strong dependence on equivalent width ( nlte effects ) , so we take as the proper iron abundance the extrapolated total iron abundance at zero equivalent width .
kurucz s width9 code was used with an atmospheric model for each star interpolated from a grid of models calculated with a microturbulent velocity of 4 km s@xmath24 . at some phases
cepheids can have microturbulent velocities deviating significantly from that value ; however , our previous test calculations suggest that changes in the model microturbulence over a range of several km s@xmath24 has an insignificant impact on the resulting element abundances .
typical results obtained for the cepheid @xmath11 cep over its cycle are shown for illustrative purposes in fig . 1 .
final results for our determinations of @xmath0 , @xmath1 , @xmath2 ( in km s@xmath24 ) , and [ fe / h ] for fgk supergiants in our sample are given in table 1 .
typical uncertainties in the cited values are @xmath25 k in @xmath0 , @xmath26 in @xmath1 , and @xmath27 km s@xmath24 in @xmath2 . [
cols="^,^,<,>,>,>,>,^,<,>,>,>",options="header " , ] +
tables 1 and 2 list the results for fgk supergiants and classical cepheids , respectively . for cepheids
the variable star designation and pulsation period @xmath28 are listed in columns ( 1 ) and ( 2 ) respectively , the mean colour excess determined here is listed in column ( 3 ) , the standard error of the mean is listed in column ( 4 ) , and the number of determinations used to calculate the mean is listed in column ( 5 ) .
the mean value of e(_b v _ ) was obtained by averaging the values of e(_b v _ ) over the pulsation cycle .
individual reddenings estimated by @xcite , @xcite , and @xcite are also listed for convenience in table 2 , in columns ( 5 ) , ( 6 ) , and ( 7 ) respectively .
a comparison of the reddenings of @xcite with the present results is shown in fig . 4 , and comparisons with the reddenings of @xcite and @xcite are shown in figs . 5 and 6 .
a statistical comparison of the present cepheid reddenings was made with an overlapping set of space reddenings for 27 stars in common .
the space reddenings are those summarized by @xcite , but with additional data used in a previous analysis of cepheid reddenings @xcite that includes new estimates for previously - unstudied objects @xcite .
four of the objects in the sample display large differences in reddening between the two data sets , in excess of @xmath29e(_b v _ ) = 0.1 , namely v cen , s nor , vy sgr , and wz sgr .
all are southern hemisphere cepheids , noted to be a problem in the previous section , and the first three have only single spectroscopic observations available .
multiple observations ( @xmath30 ) exist for wz sgr , but its space reddening has been noted previously to be a problem @xcite .
for the 23 remaining cepheids the difference in colour excesses @xmath29e(_b v_)(space reddening spectroscopic reddening ) is @xmath31 s.d . , and @xmath32 s.d .
for the 16 cepheids with multiple spectroscopic observations .
it appears that the present compilation of spectroscopic reddenings for cepheids is tied closely to the system of available space reddenings .
in addition , since the average uncertainty for the space reddenings of individual cepheids in the sample is @xmath33 , the precision of the spectroscopic reddenings based upon multiple spectra must be @xmath6 , while the accuracy matches that of the space reddening system .
the reddening system of the present study also appears to be closely matched to older reddening systems for cepheids , as indicated by the trends of figs .
the scatter in the plots of figs .
4 and 5 is suggestive of actual differences that may exist , which may have several explanations .
period - colour relations , for example , do not always account for the intrinsic dispersion of cepheids within the instability strip . small amplitude cepheids tend to lie on the hot edge of the strip , for example @xcite , so their reddenings can be underestimated by period - colour relations .
large amplitude cepheids lie near the centre of the strip ( e.g. , * ? ? ?
* ) , but may not always be representative of the calibrators used .
although similar scatter is seen in the fig .
6 , where the present results are compared with the space reddening system of @xcite , the discrepant points in most cases represent cepheids for which only single spectra are available . here
the problem is not of differences in reddening systems , but problems in the measurements for individual cepheids .
all cepheids undergo period changes @xcite , some undergo random fluctuations in period @xcite , and others display light travel time effects from binarity and other curious changes to the times of light maximum @xcite that affect the accurate determination of light curve phasing , all of which are important factors for reddenings derived from spectroscopic and photometric observations . as noted previously , the problem is particularly acute for the present study when only one spectrum has been obtained ( see * ? ? ?
* ) , which is the case for many of the cepheids in the present sample .
a number of objects in table 2 are also suspected to be type ii cepheids , but their [ fe / h ] values are nearly solar according to the spectroscopic analyses .
another comparison that can be made is with respect to colours derived from model atmospheres .
such a comparison is shown in fig . 7 , where the models ( solid dots ) are @xcite models with the colours given by @xcite .
the models are for @xmath1 = 2.0 , @xmath2 = 2.0 km s@xmath24 , and [ fe / h ] = 0.0 . the same parameters were used to generate intrinsic colours from the calibration determined here . as can be seen
, the comparison is rather good . over the larger part of the range the spectroscopic temperatures used here are 50100 k lower than the temperatures that would be determined using the photometric calibration . another way of stating this is that , at a fixed temperature , the theoretical photometric calibration would yield a
_ b v _ colour excess about 0.05 mag .
smaller than found here .
the difference between spectroscopic and photometric temperatures found here for high luminosity stars is in the same sense and magnitude as that found for spectroscopic versus photometric temperatures of dwarfs and giants @xcite .
it should be emphasized that high - resolution spectroscopic studies of cepheids provide an independent source of reddenings that can be used to refine the existing reddening scale for galactic cepheids and to enhance the precision of the cepheid p
l relation .
the colour excesses derived for cepheids in our sample for which many spectra are available are quite likely more accurate than many previous estimates , and provide a good test of published reddening scales .
additional spectroscopic observations of cepheids in our sample can only strengthen the present results , since they would eliminate the undersampling problem inherent to program obejects with @xmath34 . correcting
that deficiency should improve the precision of the overall sample .
the present study presents newly - derived parameters , namely effective temperature ( @xmath0 ) , surface gravity ( @xmath1 ) , microturbulent velocity ( @xmath2 , in km s@xmath24 ) , and iron abundance ( [ fe / h ] ) , for 74 non - variable fgk supergiants , established from model atmosphere analyses of high resolution spectra for the stars .
colour excesses have been computed for the stars on the basis of a new formulation of the relationship linking such parameters to intrinsic colour , ( _ b v_)@xmath15 .
the formulation has been extended to the 74 fgk supergiants and to a sample of 164 classical cepheid variables to derive new estimates for their colour excesses e(_b v _ ) , presented in tables 1 and 2 .
the reddening estimates are demonstrated to be of extremely high internal precision , and to agree well with the most accurate estimates from the literature .
the cepheid reddenings , in particular , appear to be closely tied to the system of space reddenings that is presently available .
given the large distances of supergiants , the method opens the possibility for large - scale extinction mapping of the galaxy , with a sensitivity of 0.080.2 magnitude .
this work is based on spectra collected with the 1.93-m telescope of the ohp ( france ) and the eso telescopes at the paranal observatory under programme id266.d-5655 .
we thank the referee for useful suggestions that helped to improve the presentation .
andrievsky s. m. , kovtyukh v. v. , luck r. e. , lepine j. r. d. , bersier d. , maciel w. j. , barbuy b. , klochkova v. g. , panchuk v. e. , karpischek r. u. 2002a , a&a , 381 , 32 andrievsky s. m. , bersier d. , kovtyukh v. v. , luck r. e. , maciel w. j. , lepine j. r. d. , beletsky yu .
v. 2002b , a&a , 384 , 140 andrievsky s. m. , kovtyukh v. v. , luck r. e. , lepine j. r. d. , maciel w. j. , beletsky yu .
v. 2002c , a&a , 392 , 491 andrievsky s. m. , luck r. e. , kovtyukh v. v. , 2005 , aj , 130 , 1880 bagnulo s. , jehin e. , ledoux c. , et al . , 2003 , eso messenger , 114 , 10 baranne a. , queloz d. , mayor m. , et al . 1996 , a&as , 119 , 373 berdnikov l. n. , 2007 , _ http://www.sai.msu.ru/groups/ cluster / cep / phe _ berdnikov l. n. , samus n. n. , antipin s. v. , ezhkova o. v. , pastukhova e. n. , turner d. g. 2004 , pasp , 116 , 536 berdnikov l. n. , pastukhova e. n. , gorynya n. a. , zharova a. v. , turner d. g. 2007 , pasp , 119 , 82 bersier d. , 1996 , a&a , 308 , 514 castelli f. , 1999 , a&a , 346 , 564 feltz k. a. jr . , 1972 , pasp , 84 , 497 fernie j. d. , 1987 , aj , 94 , 1003 fernie j. d. , evans n. r. , beattie b. , seager s. , 1995 , ibvs , 4148 fry a. m. , carney b.w . , 1999 , aj , 118 , 1806 galazutdinov g. a. , 1992 , prepr .
sao ras , 92 , 28 gray d. f. , 1992 , the observation and analysis of stellar photospheres , cambridge university press , cambridge katz d. , soubiran c. , cayrel r. , adda m. , cautain r. , 1998 , a&a , 338 , 151 kovtyukh v. v. , 2007 , mnras , 378 , 617 kovtyukh v. v. , andrievsky s. m. , 1999 , a&a , 351 , 597 kovtyukh v. v. , andrievsky s. m. , belik s. i. , luck r. e. , 2005 , aj , 129 , 433 kraft r. p. , 1960
, apj , 131,330 kurucz r. l. , 1992 , in barbuy b. , renzini a. , eds , iau symp .
149 , the stellar population of galaxies , kluwer , dordrecht , p. 225
laney c. d. , caldwell j. a. r. , 2007 , mnras , 377 , 147 luck r. e. , andrievsky s. m. , 2004 , aj , 128 , 343 luck r. e. , heiter u. , 2006 , aj , 131 , 3069 luck r. e. , heiter u. , 2007 , aj , 133 , 2464 luck r. e. , gieren w. , andrievsky s. m. , kovtyukh v. v. , fouqu p. , pont f. , kienzle f. 2003 , a&a , 401 , 939 luck r. e. , kovtyukh v. v. , andrievsky s. m. , 2006 , aj , 132 , 902 mcnamara d. h. , potter d. , 1969 , pasp , 81 , 545 mcnamara d. h. , helm t. m. , wilcken s. k. , 1970 , pasp , 82 , 293 moultaka j. , ilovaisky s. a. , prugniel p. , soubiran c. , 2004 , pasp , 116 , 693 pel j. w. , lub j. , 1978 , in philip a. g. d. , hayes d. s. , eds , iau symp .
80 , the hr diagram : the 100th anniversary of henry norris russell , reidel , dordrecht , p. 229
sasselov d. d. , lester j. b. , 1990 , apj , 360 , 227 schmidt e. g. , taylor d. j. , 1979 , aj , 84 , 1192 soubiran c. , katz d. , cayrel r. , 1998 , a&ass , 133 , 221 spencer jones j. h. , 1989 , mnras , 238 , 269 tammann g. a. , sandage a. , reindl b. , 2003 , a&a 404 , 423 turner d. g. , 1995 , in stobie r. s. , whitelock p. a. , eds , iau colloq . 155 , asp conf .
series 83 , astrophysical applications of stellar pulsation , asp , san francisco , p. 256
turner d. g. , 2001 , odessa astron .
, 14 , 166 turner d. g. , burke j. f. , 2002 , aj , 124 , 2931 turner d. g. , leonard p. j. t. , english d. a. , 1987 , aj , 93 , 368 turner d. g. , abdel - sabour abdel - latif m. , berdnikov l. n. , 2006a , pasp , 118 , 410 turner d. g. , usenko i. a. , kovtyukh v. v. , 2006b , observatory , 126 , 207 turner d. g. , bryukhanov i. s. , balyuk i. i. , et al . , 2007 , pasp , 119 , 1247 turner d. g. , maclellan r. f. , henden a. a. , berdnikov l. n. , 2008 , mnras , submitted | accurate and homogeneous atmospheric parameters ( @xmath0 , @xmath1 , @xmath2 , [ fe / h ] ) are derived for 74 fgk non - variable supergiants from high - resolution , high signal - to - noise ratio , echelle spectra .
extremely high precision for the inferred effective temperatures ( 1040 k ) is achieved by using the line - depth ratio method .
the new data are combined with atmospheric values for 164 classical cepheids , observed at 675 different pulsation phases , taken from our previously published studies .
the derived values are correlated with unreddened _ b v _ colours compiled from the literature for the investigated stars in order to obtain an empirical relationship of the form : @xmath3 + 0.8833(\log g)(\log t_{\rm eff})$ ] .
the expression is used to estimate colour excesses e(_b v _ ) for individual supergiants and classical cepheids , with a precision of @xmath4 mag . for supergiants and cepheids with @xmath5 spectra , reaching @xmath6 mag . for cepheids with @xmath7 spectra , matching uncertainties for the most sophisticated photometric techniques .
the reddening scale is also a close match to the system of space reddenings for cepheids .
the application range is for spectral types f0k0 and luminosity classes i and ii .
[ firstpage ] stars : fundamental parameters stars : colour excesses stars : supergiants stars : classical cepheids . |
the goal of this paper is two - fold ; the first goal is to present an algorithm for developing reduced - order models of the input - output dynamics of _ unstable _ high - dimensional linear state - space systems ( such as linearized navier - stokes equations with actuation and sensing ) , while the second goal is to demonstrate the algorithm by developing estimation - based controllers to stabilize unstable steady states of a two dimensional low - reynolds - number flow past a flat plate at a large angle of attack . development of feedback control strategies based on linearized navier - stokes equations is attractive due to the ready availability of a large class of control techniques , and there has been substantial progress in this direction in the past decade , reviewed in detail by @xcite .
however , many of these techniques are limited to relatively small dimensional systems @xmath0 , while the numerical discretization of fluid flows invariably result in huge dimensional systems , typically @xmath1 .
thus , model reduction has played an important role in making these tools further accessible to fluid flows .
extensive research effort in model reduction has focused on the method of proper orthogonal decomposition ( pod ) and galerkin projection , developed first by @xcite .
the main disadvantage of this technique is that , although the pod modes capture the energetically important structures of the flow , the reduced - order models obtained by the subsequent galerkin projection of the governing equations onto these modes often do not faithfully represent the dynamics .
various modifications to improve this method have been proposed and used for flow control ; refer to the introduction of @xcite for a review of these techniques .
the pod / galerkin methods have been applied for flow control in various contexts , such as bluff - body wake suppression ( @xcite , @xcite , @xcite , @xcite ) , noise reduction in cavity flow ( @xcite , @xcite ) , and drag reduction in turbulent boundary layers ( @xcite , @xcite ) .
another model - reduction technique , based on projection onto the global ( stable or unstable ) eigenmodes of the flow linearized about steady states , has been used by @xcite and @xcite in the context of spatially developing flows such as separated boundary layers . in this paper , we focus on an approximate balanced truncation method developed by @xcite as an approximation to the original method of @xcite .
this technique captures the dynamically important modes of the system , and the non - approximate version provides rigorous bounds for the resulting reduced - order models .
the method , sometimes called balanced pod , was used to obtain models of the linearized channel flow by @xcite and the blasius boundary layer by @xcite , and was shown to accurately capture the control actuation and also to outperform the pod / galerkin models .
the balanced truncation method of @xcite is applicable only to systems linearized about _
stable _ steady states .
an extension to _ unstable _ linear systems was proposed by @xcite , by introducing frequency - domain definitions of controllability and observability gramians .
reduced - order models were obtained by first decoupling the dynamics on the stable and unstable subspaces , and then truncating the relatively uncontrollable and unobservable modes on each of the two subspaces . in this paper , we present an approximation algorithm for balanced truncation of linear unstable systems , which results in models that are equivalent to those of @xcite on the _ stable _ subspace .
the dynamics on the unstable subspace is treated _ exactly _ by a projection onto the global eigenmodes , as in @xcite . as a proof - of - concept study ,
the modeling procedure is applied to the problem of two dimensional low - reynolds - number flow past a flat plate at a large angle of attack .
we develop reduced - order models and design controllers that stabilize the unstable steady states of this flow .
our motivation for the choice of this problem comes from our interest in regulating vortices in separated flows behind low aspect - ratio wings , which is of importance in design of micro air vehicles ( mavs ) .
recently , design of mavs has been inspired from experimental observations in insect and bird flights of a stabilizing leading edge vortex ( see @xcite and @xcite ) , which remains attached throughout the wing stroke and provides enhanced lift .
so , it could be beneficial to design controllers that can manipulate the wake of mavs to enhance lift and achieve better maneuverability in presence of wind gusts .
recent studies in this direction , using open - loop control of the flow past low - aspect - ratio wings using steady or periodic blowing , were performed computationally by @xcite and experimentally by @xcite .
these studies explored different forcing amplitudes and frequencies , locations and directions .
however , the design of feedback controllers remains a challenge , due to the large dimensionality of the problem and complex flow physics .
we present computational tools , that we hope can at least pave a direction and provide techniques towards addressing some of these challenges .
we consider the 2-d flow past a flat plate , actuated by a localized body force close to the leading edge , with two near - wake velocity sensors .
we design a reduced - order compensator and show that it is able to suppress vortex shedding at high angles of attack .
many previous studies have focused on the control of flow past a cylinder , which is qualitatively similar to the flow past a flat plate at large angle of attack , with the natural flow in both the cases being periodic vortex shedding . for the cylinder ,
the flow undergoes a transition from steady state to periodic shedding with increasing reynolds number , while a similar transition occurs in the flat plate with increasing angle of attack .
there has been considerable research effort on suppression of this shedding in cylinder and other bluff body wakes , using passive and active , open - loop and feedback control , as reviewed by @xcite . among those ,
some techniques are based on reduced - order models ; for instance , @xcite developed models using artificial neural networks and a pod basis , @xcite modified the pod / galerkin method to account for actuation by means of cylinder - rotation , while @xcite developed a double pod method to account for changes in the wake structure during transients .
some earlier efforts in the control of a _ flat - plate _ wake include those by @xcite , @xcite who used vortex - based methods to model the flow past a vertical plate ( angle of attack = @xmath2 ) ; vortex - based models form their own class of modeling techniques reviewed recently by @xcite .
lagrangian coherent structures were used by @xcite to enhance mixing in flow past a bluff body with the trailing surface similar to the vertical flat plate .
one of the few efforts towards control of flat plate at an angle of incidence was by @xcite , who used a passive leading - edge suction control along with a potential flow vortex model .
@xcite also used reduced - order vortex models for drag reduction on an elongated d - shaped bluff body .
this paper is organized as follows : in section [ sec : model_red ] , we first briefly describe the balanced truncation method for unstable systems as developed by @xcite , and the approximate balanced truncation procedure called balanced pod of @xcite for large dimensional stable systems
. then , we present an algorithm for approximate balanced truncation of large dimensional _ unstable _
systems , assuming that the dimension of unstable subspace is small and the corresponding global eigenmodes can be computed . in section [ sec : numerics_ibfs ] , we briefly describe the numerical technique of @xcite using a fast immersed boundary method , and present the linearized and adjoint formulations of this numerical method . in section [ section : model_results ] , we present numerical results , using the model example of two - dimensional flow past a flat plate at a large angle of attack and a low reynolds number @xmath3 .
first , we perform a steady - state analysis and compute the branch of steady states in the entire range of angles of attack , @xmath4 .
we also compute the left and right global eigenmodes of the flow linearized about the unstable steady state at @xmath5 .
we present reduced - order models of the linearized dynamics and use linear optimal control techniques to design controllers that stabilize this steady state .
full - state feedback and more practical near - wake velocity - measurement based feedback controllers are derived , implemented in the nonlinear equations , and shown to suppress vortex shedding .
the paper concludes with a brief discussion in section [ sec : discussion ] .
we briefly describe a model reduction procedure using the balanced truncation method for unstable systems developed by @xcite . consider the state - space system @xmath6 where @xmath7 is the state , @xmath8 is the input , and @xmath9 is the output of the system ; the dot over @xmath10 represents differentiation with respect to time .
the eigenvalues of @xmath11 are assumed to be anywhere on the complex plane , except on the imaginary axis .
the standard balanced truncation procedure developed by @xcite , valid only for stable systems , starts with defining the controllability and observability gramians of the system ( [ ss ] ) as follows : @xmath12 where asterisks are used to denote adjoint operators .
a co - ordinate transformation is then obtained such that the gramians ( [ gramians_stable ] ) of the transformed system are equal and diagonal .
the diagonal entries of the transformed gramians , called hankel singular values ( hsvs ) , decrease monotonically and are directly related to the controllability and observability of the corresponding states .
a reduced - order model is obtained by truncating the states with relatively small hsvs , that is , the states which are almost uncontrollable and unobservable . for unstable systems ,
the integrals in ( [ gramians_stable ] ) are unbounded and hence the gramians are ill - defined .
a modified technique was proposed by @xcite based on the following frequency - domain definitions of the gramians : @xmath13 by using parseval s theorem , it can be shown that for stable systems , the frequency - domain definitions ( [ gram_freq ] ) are equivalent to the time - domain definitions ( [ gramians_stable ] ) .
the model - reduction procedure of @xcite begins by first transforming the system ( [ ss ] ) to coordinates in which the stable and unstable dynamics are decoupled .
that is , let @xmath14 be a transformation such that if @xmath15 , the system ( [ ss ] ) transforms to @xmath16 here , @xmath17 and @xmath18 are such that all their eigenvalues are in the right- and left - half complex planes respectively , while @xmath19 and @xmath20 are the corresponding states .
next , denote the controllability and observability gramians corresponding to the set @xmath21 describing the stable dynamics by @xmath22 and @xmath23 respectively .
similarly , denote the gramians corresponding to the set @xmath24 by @xmath25 and @xmath26 .
the gramians of the original system are then related to those corresponding to the subsystems by : @xmath27 a system is said to be balanced if its gramians defined by ( [ gramians_general ] ) are equal and diagonal , in which case the diagonal entries are called the _
generalized _ hankel singular values .
a reduced - order model is obtained by truncating the states with small generalized hsvs . for systems of large dimension @xmath28 such as those encountered here , the gramians ( [ gramians_general ] ) are huge matrices which can not be easily computed or stored .
a computationally tractable procedure was introduced by @xcite for obtaining an approximate balancing transformation .
we first briefly describe this method , valid only for stable systems , and then present an extension to unstable systems .
the procedure consists of computing the impulse responses of the system ( [ ss ] ) and stacking the resulting snapshots of the state @xmath10 as columns of a matrix @xmath29 .
it also requires state - snapshots of the impulse responses of the adjoint system @xmath30 which are stacked as columns of a matrix @xmath31 .
then , the gramians ( [ gramians_stable ] ) can be approximated as @xmath32 the approximate gramians ( [ gramians_approx ] ) are not actually computed due to the large storage cost , but the leading columns ( or modes ) of the transformation that balances these gramians are computed using a cost - efficient algorithm .
it involves computing the singular value decomposition of @xmath33 where @xmath34 is a diagonal matrix of the most significant hsvs greater than a cut - off which is a modeling parameter , while @xmath35 is a diagonal matrix of smaller and zero hsvs .
note that @xmath36 is a small matrix , where @xmath37 and @xmath38 are the number of snapshots of the impulse responses of systems ( [ ss ] ) and ( [ ss_adj ] ) respectively .
for fluid systems that we are interested in , the typical number of snapshots is @xmath39 , thus resulting in a reasonable computational cost , and typically @xmath40 .
the leading columns and rows of the balancing transformation and its inverse are obtained using : @xmath41 where @xmath42 , and the two sets of modes are bi - orthogonal ; that is , @xmath43 .
the reduced - order model of ( [ ss ] ) is then obtained by expressing @xmath44 , @xmath45 , and using the bi - orthogonality of @xmath46 and @xmath47 : @xmath48 when the number of outputs of the system ( rows of @xmath49 ) is large , the algorithm described in section [ sec : approx_baltrunc_stable ] can become intractable .
the reason for this is that it involves one simulation of the adjoint system ( [ ss_adj ] ) for each component of @xmath50 , the dimension of which is the same as the number of outputs .
this number is often large in fluids systems where a good model needs to capture the response of the entire system ( @xmath51 ) to a given input . to overcome this problem ,
@xcite proposed a technique called _ output projection _ , which involves projecting the output @xmath52 of ( [ ss ] ) onto a small number of energetically important modes obtained using proper orthogonal decomposition ( pod ) .
let the columns of @xmath53 consist of the first @xmath54 pod modes of the dataset consisting of outputs obtained from an impulse response of ( [ ss ] ) .
then , for the purpose of obtaining a reduced - order model , the output of ( [ ss ] ) is approximated by @xmath55 where @xmath56 is an orthogonal projection of the output onto the first @xmath54 pod modes .
the resulting output - projected system is optimally close ( in the @xmath57sense ) to the original system , for an output of fixed rank @xmath54 . with this approximation ,
only @xmath54 adjoint simulations are required to approximate the observability gramian ; refer to @xcite for details .
the number of pod modes @xmath54 for output projection is a design parameter and is typically chosen to capture more than @xmath58 of the output energy . in the rest of this paper , the models resulting from this approximation of the output are referred to as _ @xmath59mode output projected models . _ the approximate balancing procedure described in the previous section , which is essentially a snapshot - based method , does not extend to unstable systems since the impulse responses of ( [ ss ] ) and ( [ ss_adj ] ) are unbounded .
we could consider applying the algorithm to the two sub - systems given in ( [ decouple_ss ] ) , but the transformation @xmath14 that decouples ( [ ss ] ) itself is not available .
however , when the dimension of the unstable sub - system is small , we show that it is not necessary to compute the entire transformation @xmath14 and it is still possible to obtain an approximate balancing transformation . here
, we present an algorithm for computing such a transformation and also show that it essentially results in a method that is a slight variant of the technique of @xcite presented in section [ sec : unstable_baltrunc ] .
the idea behind the algorithm is to project the original system ( [ ss ] ) onto the stable subspace of @xmath11 .
then , one obtains a reduced - order model of the projected system using the snapshot - based procedure described in section [ sec : approx_baltrunc_stable ] .
the dynamics projected onto the unstable subspace can be treated exactly on account of its low dimensionality .
we assume that the number of unstable eigenvalues @xmath60 is @xmath61 and can be computed numerically , say using the computational package arpack developed by @xcite .
we further assume that the bases for the right and the left unstable eigenspaces @xmath62 can be computed .
for the algorithm , we need the following projection operator onto the stable subspace of @xmath11 : @xmath63 where @xmath64 and @xmath65 have been scaled such that @xmath66 .
we use the operator @xmath67 to obtain the dynamics of ( [ ss ] ) restricted to the stable subspace of @xmath11 as follows : @xmath68 where @xmath69 the adjoint of ( [ ss_stable ] ) is the same as the dynamics of ( [ ss_adj ] ) restricted to the stable subspace of @xmath70 using @xmath71 , and is given by @xmath72 where @xmath73 .
we compute the state - impulse responses of ( [ ss_stable ] ) and ( [ ss_adj_stable ] ) and stack the resulting snapshots @xmath74 and @xmath75 in matrices @xmath76 and @xmath77 respectively . as in ( [ svd ] ) , we compute the singular valued decomposition of @xmath78 and use the expressions ( [ bal_transf ] ) to obtain the balancing modes @xmath79 and the adjoint modes @xmath80 , where again @xmath81 .
the reduced - order modes are obtained by expressing the state @xmath10 as @xmath82 where @xmath83 and @xmath84 . substituting ( [ modal_exp ] ) in ( [ ss ] ) and pre - multiplying by @xmath85 and @xmath86
, we obtain @xmath87 now , since range@xmath88 span@xmath89 , we can write @xmath90 for some @xmath91 , and using the properties of eigenvectors , we have @xmath92 .
similarly , it can be shown that @xmath93 .
thus , the cross terms in ( [ roma ] ) are zero and the reduced - order model is @xmath94 the procedure described so far to obtain the reduced - order model ( [ rom ] ) is related to the procedure of @xcite described in section [ sec : unstable_baltrunc ] .
it can be shown that the transformation that balances the gramians defined by ( [ gramians_general ] ) results in a system in which the unstable and stable dynamics are decoupled .
furthermore , the resulting stable dynamics are the same as those given by the equations describing the @xmath95-dynamics of ( [ rom ] ) . that is , balancing the stable part of the gramians @xmath96 and @xmath97 defined in ( [ gramians_general ] ) ( balancing @xmath22 and @xmath23 ) is the same as balancing the gramians of the stable subsystem ( [ ss_stable ] )
; a proof is outlined in appendix [ sec : app : baltrunc ] . in our algorithm , the unstable dynamics are not balanced , while they are by @xcite . a disadvantage of zhou s approach is that an unstable mode might be truncated resulting in a model which does not capture the instability , which is undesirable for control purposes . for systems with a large number of outputs , the number of adjoint simulations ( [ ss_adj_stable ] ) can become intractable ;
however , the output projection of section [ sec : outproj ] can readily be extended to unstable systems . instead of projecting the entire output @xmath52 onto pod modes , we first express the state @xmath98 , where @xmath99 and @xmath100 are projections on the unstable and stable subspaces of @xmath11 respectively .
we similarly express the output as @xmath101 .
we then project the component @xmath102 onto a small number of pod modes , of the data set consisting of the outputs from an impulse response of ( [ ss_stable ] ) .
if the pod modes are represented as columns of the matrix @xmath103 , the output of ( [ ss ] ) is approximated by @xmath104 x = cx_u + \theta_s \theta_s^\ast c x_s . \label{yapprox_unstable}\end{aligned}\ ] ] finally , with the state @xmath10 expressed by the modal expansion ( [ modal_exp ] ) , the output of the reduced - order model ( [ rom ] ) is given by @xmath105 the steps involved in obtaining reduced - order models of ( [ ss ] ) , for the case where the output is the entire state ( @xmath51 ) , can now be summarized as follows : 1 .
[ primal_step ] project the original system ( [ ss ] ) onto the subspace spanned by the stable eigenvectors of @xmath11 in the direction of the unstable eigenvectors of @xmath11 to obtain ( [ ss_stable ] ) . compute the ( state ) response to an impulse on each input of ( [ ss_stable ] ) and stack the snapshots in a matrix @xmath76 .
2 . assemble the resulting snapshots , and compute the pod modes @xmath106 of the resulting data - set .
these pod modes are stacked as columns of @xmath107 .
3 . choose the number of pod modes one wants to use to describe the output of ( [ ss_stable ] ) .
for instance , if @xmath108 error is acceptable , and the first @xmath54 pod modes capture @xmath58 of the energy , then the output is the velocity field projected onto the first @xmath54 modes .
thus , the output is represented as @xmath109 .
[ adjoint_step ] project the adjoint system ( [ ss_adj ] ) onto the subspace spanned by the stable eigenvectors of @xmath70 in the direction of the unstable eigenvectors of @xmath70 to obtain ( [ ss_adj_stable ] ) .
compute the ( state ) response of ( [ ss_adj_stable ] ) , starting with each pod mode @xmath106 as the initial condition ( one simulation for each of the first @xmath54 modes ) . stack the snapshots in a matrix @xmath77
5 . compute the singular value decomposition of @xmath110 , where @xmath111 , and @xmath112 rank(@xmath113 ) .
define balancing modes @xmath114 and the corresponding adjoint modes @xmath115 as columns of the matrices @xmath79 and @xmath80 , where @xmath116 7 .
obtain the reduced - order model using ( [ rom ] ) , which can be written as @xmath117 alternatively , the outputs can be considered to be simply the coefficients of the unstable modes @xmath118 and the coefficients of the pod modes @xmath107 of the stable subspace . with this choice , the output can be represented as @xmath119 finally , if the initial state @xmath120 is known , the initial condition of ( [ rom_compact ] ) can be obtained using @xmath121 in the remainder of this paper , we demonstrate the algorithm developed in this section by obtaining reduced order models of the 2-d uniform flow past a flat plate , and develop controllers based on these models to stabilize the unstable steady states that exist at high angles of attack .
the numerical scheme used is a fast immersed boundary method developed by @xcite , and is briefly described here .
the method is then used to develop the linearized and adjoint formulations .
consider the following form of the incompressible navier - stokes equations , based on the continuous analog of the immersed boundary formulation introduced by @xcite : @xmath122 where @xmath123 , @xmath124 and @xmath125 are the appropriately non - dimensionalized fluid velocity , pressure and surface force respectively .
the force @xmath125 acts as a lagrange multiplier that imposes the no - slip boundary condition on the lagrangian points @xmath126 , which arise from the discretization of a body moving with velocity @xmath127 .
we consider the body to be a stationary flat plate at an angle of attack @xmath128 ; that is , here @xmath129 .
the reynolds number is defined as @xmath130 where @xmath131 is the free - stream velocity , @xmath132 is the chord - length and @xmath133 is the kinematic viscosity .
equations ( [ navier1]-[navier3 ] ) are discretized in space using a second - order finite - volume scheme on a staggered grid , and a discrete curl operation @xmath134 is introduced to eliminate the pressure and obtain a semi - discrete formulation in terms of the circulation @xmath135 : @xmath136 the incompressibility condition ( [ navier2 ] ) is implicitly satisfied by an appropriate construction of @xmath49 .
the discrete laplacian is represented by @xmath137 , using the identity @xmath138 ; the constant @xmath139 , where @xmath140 is the uniform grid spacing . the operator @xmath141 smears the dirac delta function of ( [ navier1 ] ) over a few grid points .
the nonlinear term @xmath142 is the spatial discretization of @xmath143 , where @xmath144 is the discrete velocity flux , in turn related to the discrete stream function @xmath145 and circulation @xmath135 as : @xmath146 a uniform grid and a choice of simple boundary conditions result in a _
fast _ algorithm . with a uniform grid ,
the discrete poisson equation ( [ stream ] ) is solved by means of the efficient discrete sine transform .
the boundary conditions specified are dirichlet and neumann for the velocity components normal and tangential to the domain boundaries , which for the flow past a flat plate imply a uniform - flow in the far - field .
these boundary conditions are however valid for only a sufficiently large domain , and are imposed by employing a computationally efficient multi - domain approach .
the domain is considered to be embedded in a series of domains , each twice - as - large as the preceding , with a uniform but a _ coarser _ grid having the same number of grid points .
the poisson equation , with zero boundary conditions , is solved on the largest domain and the stream function is interpolated on the boundary of the smaller domain , which are in turn used to solve the poisson equation on the smaller domain . for the flow past a flat plate considered here ,
the typical size of the largest domain is around 40 chord lengths in each direction .
finally , the time - integration is performed using the implicit crank - nicolson scheme for the viscous terms and the second - order accurate , adams - bashforth scheme for the convective terms . for deriving reduced - order models useful for control design , we first linearize equation ( [ semidisc ] ) about a pre - computed steady state ( @xmath147 , @xmath148 ) .
the linearized equations are the same as equations ( [ semidisc],[constraint ] ) with the nonlinear term @xmath142 replaced by its linearization about the steady state , and is denoted by @xmath149 where the flux @xmath144 is related to @xmath135 by ( [ stream ] ) . thus , the linearized equations are : @xmath150 the boundary conditions for the linearized equations are @xmath151 on the outer boundary of the largest computational domain . in order to derive the reduced - order models using the procedure described earlier , we need to perform adjoint simulations . in order to derive the adjoint equations , we define the following inner - product on the state - space : @xmath152 that is , the inner - product defined on the state - space is the standard @xmath153-inner product weighted with the inverse - laplacian operator . this choice is convenient as it results in the adjoint equations which differ from the linearized equations only in the nonlinear term .
a derivation is outlined in appendix [ sec : app : adjoint ] and the resulting equations are : @xmath154 where the variables @xmath155 , @xmath126 and @xmath156 are the duals of the discrete circulation @xmath135 , stream function @xmath145 and body force @xmath157 , respectively , and @xmath158 is the dual of flux @xmath144 .
the adjoint of the linearized nonlinear term is @xmath159 , which can be shown to be a spatial discretization of @xmath160 .
since equation ( [ adj1 ] ) differs from ( [ linear ] ) only in the last term on the right hand side , the numerical integrator for the adjoint equations can be obtained by a small modification to the linearized equations solver . the nature of the multi - domain scheme used to approximate the boundary conditions of the smallest computational domain , results in a multi - domain discrete laplacian which is not exactly self - adjoint to numerical precision . as a result ,
the adjoint formulation given by ( [ adj1],[adj2 ] ) which also uses the same multi - domain approach , is not precise and results in small , rather insignificant , errors in the computation of the reduced - order models . .
vorticity contours are plotted , with negative contours shown by dashed lines .
the velocity - sensor locations are marked by solid circles.,height=192 ]
we apply the model reduction techniques developed in the previous sections to the uniform flow past a flat plate in two spatial dimensions , at a low reynolds number , @xmath3 .
we obtain reduced - order models of a system actuated by means of a localized body force near the leading edge of the flat plate ; the vorticity contours of the flow field obtained on an impulsive input to the actuator are shown in fig .
[ fig : actuator_sensor ] . using these reduced - order models ,
we develop feedback controllers that stabilize the unstable steady state at high angles of attack .
we first assume full - state feedback , but use output projection described in section [ sec : outproj ] to considerably decrease the number of outputs in order to make the model computation tractable .
later , we relax the full - state feedback assumption , and develop a more practical observer - based controller which uses a few velocity measurements in the near - wake of the flat plate ( shown in fig . [ fig : actuator_sensor ] ) to reconstruct the entire flow . the grid size used is @xmath161 , with the smallest computational domain given by @xmath162\times[-2.5,2.5]$ ] , where lengths are non - dimensionalized by the chord of the flat plate , with its center located at the origin .
we use 5 domains in the multiple - grid scheme , resulting in an effective computational domain @xmath163 times larger the size of the smallest domain ; thus the largest domain is given by @xmath164\times[-40,40]$ ] .
the timestep used for all simulations is @xmath165 . since our approach is to obtain reduced - order models of the flow linearized about a given steady state ,
we first need to compute these steady states .
the model - reduction of unstable systems involves projecting the dynamics onto a stable subspace , for which we also need to compute the right and left eigenvectors of the linearized dynamics .
this section concerns this steady - state analysis , using a `` timestepper - based '' approach as outlined in @xcite and @xcite .
a simple way of computing stable steady states is by simply evolving the time - accurate simulation to stationarity .
however , unstable steady states can not be found in this manner , and stable steady states near a bifurcation point could take very long to converge .
instead , we use a timestepper - based approach which involves writing a computational wrapper around the original computational routine to compute the steady states using a newton iteration .
if the numerical timestepper advances a circulation field @xmath166 at a timestep @xmath167 to a circulation field @xmath168 after @xmath14 timesteps , the steady state is given by the field @xmath147 that satisfies @xmath169 the steady states are given by zeros of @xmath170 , which could , in principle , be solved for using newton s method .
however , the standard newton s method involves computing and inverting jacobian matrices at each iteration , which is computationally infeasible due to the large dimension of fluid systems . instead of computing the jacobian
, we use a krylov - space based iterative solver called generalized minimal residual method ( gmres ) developed by @xcite to compute the newton update ( see @xcite and @xcite for a description of the method ) .
this method requires computation of only jacobian - vector products @xmath171 , which can be approximated using finite differences as @xmath172/\epsilon$ ] , for @xmath173 .
so , the jacobian - vector products can also be computed by invoking the appropriately - initialized timestepper .
a nice feature of gmres is relatively fast convergence to the steady state when the eigenvalues of the jacobian @xmath174 occur in clusters ; see @xcite and @xcite for details . for systems with multiple time - scales , such as navier - stokes ,
most of the eigenvalues of the continuous jacobian lie in the far - left - half of the complex plane .
thus , the corresponding eigenvalues of the discrete jacobian @xmath175 , for a sufficiently large value of @xmath14 , cluster near the origin . and at @xmath3 , showing a transition from a stable equilibrium to periodic vortex shedding at @xmath176 . shown are the force coefficients corresponding to the stable ( @xmath177 ) and unstable ( @xmath178 ) steady states , and the maximum and minimum ( @xmath179 ) , and the mean ( @xmath180 ) values during periodic vortex shedding . also shown
are the vorticity contours ( negative values in dashed lines ) of steady states at @xmath181 and the flow fields corresponding to the maximum and minimum force coefficients at @xmath182 .
, title="fig:",height=134 ] and at @xmath3 , showing a transition from a stable equilibrium to periodic vortex shedding at @xmath176 . shown
are the force coefficients corresponding to the stable ( @xmath177 ) and unstable ( @xmath178 ) steady states , and the maximum and minimum ( @xmath179 ) , and the mean ( @xmath180 ) values during periodic vortex shedding . also shown
are the vorticity contours ( negative values in dashed lines ) of steady states at @xmath181 and the flow fields corresponding to the maximum and minimum force coefficients at @xmath182 .
, title="fig:",height=134 ] and at @xmath3 , showing a transition from a stable equilibrium to periodic vortex shedding at @xmath176 . shown
are the force coefficients corresponding to the stable ( @xmath177 ) and unstable ( @xmath178 ) steady states , and the maximum and minimum ( @xmath179 ) , and the mean ( @xmath180 ) values during periodic vortex shedding . also shown
are the vorticity contours ( negative values in dashed lines ) of steady states at @xmath181 and the flow fields corresponding to the maximum and minimum force coefficients at @xmath182 .
, title="fig:",height=182 ] \(a ) ( b ) + .
( b ) @xmath183 vs. time , with the steady state as an initial condition.,title="fig:",height=144 ] .
( b ) @xmath183 vs. time , with the steady state as an initial condition.,title="fig:",height=144 ] the procedure described above is used to compute the branch of steady states for the angles of attack @xmath184 ; the parameter @xmath14 in ( [ newton ] ) is fixed to 50 timesteps .
the lift and drag coefficients , @xmath183 and @xmath185 , and their ratio @xmath186 with changing @xmath128 are plotted in fig .
[ fig : bifurc ] . as with flow past bluff bodies with increasing reynolds number ( for example
, see @xcite ) , the flow undergoes a hopf bifurcation from a steady flow to periodic vortex shedding as the angle of attack @xmath128 is increased beyond a critical value @xmath187 , which in our computations is @xmath188 .
also plotted in the figure are the maximum , minimum , and mean values of the forces during shedding for @xmath189 .
we see that the ( unstable ) steady state values of the lift coefficient are smaller than the minimum for the periodic shedding till @xmath190 , after which they are slightly higher , but still smaller than the mean lift for the periodic shedding .
the ( unstable ) steady state drag is much lower than the minimum value for periodic shedding .
the ratio @xmath186 of the ( unstable ) steady state is close to the mean value for shedding .
thus , if the large fluctuations in the forces are undesirable at high angles of attack , it would be useful to stabilize the unstable state .
the steady state at @xmath5 is shown in fig .
[ fig : ss](a ) , and a time history of the lift coefficient
@xmath183 with this steady state as an initial condition is shown in fig . [ fig : ss](b ) .
since the steady state is unstable , the numerical perturbations excite the instability , and the flow eventually transitions to periodic vortex shedding .
+ we also compute a basis spanning the right and left unstable eigenspaces ( @xmath64 and @xmath65 ) of the flow linearized about the unstable steady states , which are required in our model reduction procedure , for restricting dynamics onto the stable subspace .
as the flow undergoes a hopf bifurcation , a complex pair of eigenvalues crosses the imaginary axis from the left half of the complex plane ; thus the dimension of the unstable subspace is two . here
, we obtain the basis spanning this subspace by a numerical implementation of the power method ( see page 191 of @xcite ) .
we begin the linearized simulation with a very small random noise to excite the instability . after a sufficiently long time , the stable modes decay , and the dynamics lies close to the unstable subspace .
any two independent snapshots of the remaining dynamics gives a good approximation of the basis spanning the unstable eigenspace .
similarly , a basis spanning the left unstable eigenspace can be computed by initializing the adjoint equations with a small random noise . a basis spanning the right and left unstable eigenspaces of the flow linearized about the steady state at @xmath5 is plotted in fig .
[ fig : evec ] .
these modes are qualitatively similar to the structures during periodic vortex shedding , but have different spatial wavelengths , as reported in earlier studies by @xcite and @xcite .
\(a ) + , projected onto the stable subspace , and ( b , c ) the first- and fifth - most energetic pod modes of the impulse response , restricted to the stable subspace.,title="fig:",height=124 ] + ( b ) ( c ) + , projected onto the stable subspace , and ( b , c ) the first- and fifth - most energetic pod modes of the impulse response , restricted to the stable subspace.,title="fig:",height=124 ] , projected onto the stable subspace , and ( b , c ) the first- and fifth - most energetic pod modes of the impulse response , restricted to the stable subspace.,title="fig:",height=124 ] we now describe the process involved in deriving reduced - order models of the input - output response of ( [ ss ] ) , which in this example are the linearized incompressible navier - stokes equations ( [ linear ] , [ lin_constraint ] ) .
the actuator used is a localized body force close to the leading edge of the flat plate , plotted in fig .
[ fig : actuator_sensor ] .
the models are derived using the procedure outlined in section [ sec : approx_baltrunc_unstable ] . as seen in equation ( [ rom ] ) , the output of the system is considered to be the entire velocity field , observed as a projection onto ( a ) the unstable eigenspace , and ( b ) the span of the leading pod modes of the impulse response restricted to the stable subspace .
the first step in computing the reduced - order models is to project the flow field @xmath191 onto the stable subspace of ( [ linear ] , [ lin_constraint ] ) using the projection operator @xmath67 defined in equation ( [ proj_stable ] ) ; the unstable eigenvectors computed in section [ sec : steady_states ] are used to define @xmath67 numerically .
the vorticity contours of the corresponding flow field @xmath192 are plotted in fig .
[ fig : actuator]a .
the next step is to compute the impulse response of ( [ ss_stable ] ) . instead , for practical reasons , we compute the impulse response of @xmath193 that is , at each timestep of integration
, we project the state @xmath74 onto the stable subspace of @xmath11 .
because the stable subspace is an invariant subspace for the linearized dynamics ( [ linear ] ) , theoretically , the impulse responses of equations ( [ ss_stable ] ) and ( [ ss_stable_numerical ] ) are exactly the same , and they are the same as that obtained by restricting the impulse response of ( [ ss ] ) to its stable subspace . however , due to the ( small ) numerical inaccuracy of the projection @xmath67 ( which is a result of the numerical inaccuracy of the unstable eigenspaces @xmath64 and @xmath65 ) , the dynamics of ( [ ss_stable ] ) is not strictly restricted to the stable subspace and , in the long term , grows without bound in the unstable direction .
next , we compute the pod modes @xmath194 of the impulse response of ( @xmath195 ) , and consider the output of ( [ ss_stable_numerical ] ) to be the state @xmath74 projected onto a certain number of these pod modes . here , 200 snapshots spaced every 50 timesteps were used to compute the pod modes .
the leading 4 and 10 pod modes contain @xmath196 and @xmath197 of the energy respectively and , as it has been observed in previous studies ( see @xcite and @xcite ) , these modes come in pairs in terms of their energy content , a characteristic of traveling structures ; the leading first and third pod modes are shown in fig .
[ fig : actuator ] .
+ the next step is to compute the adjoint snapshots , with the pod modes of the impulse response ( projected onto the stable subspace of the adjoint ) as the initial conditions . as the linearized impulse response
, these simulations are also restricted to the stable subspace .
again , instead of computing the response of ( [ ss_adj_stable ] ) , we compute that of the following system : @xmath198 the snapshots of the impulse responses of systems ( [ ss_stable_numerical ] ) and ( [ ss_adj_stable_numerical ] ) are stacked as columns of @xmath29 and @xmath31 , and using the expressions ( [ svd ] ) and ( [ bal_transf ] ) , we obtain the balancing modes @xmath199 and the adjoint modes @xmath200 .
we used 200 snapshots of the linearized simulation and 200 snapshots of each adjoint simulation , with the spacing between snapshots fixed to 50 timesteps , to compute the balancing transformation .
these number of snapshots and the spacing were sufficient to accurately compute the modes ; further reduction in the spacing did not significantly change the singular values from the svd computation ( [ svd ] ) .
we considered the outputs to be a projection onto 4 , 10 and 20 pod modes ( corresponding to 4 , 10 and 20 mode _ output - projections _ , as introduced in section [ sec : outproj ] ) .
using these modes , we use the expressions in equation ( [ rom_compact],[output_compact2 ] ) to obtain the matrices @xmath201 defining the reduced - order model of the stable - subspace dynamics .
the vorticity contours of the balancing and the adjoint modes , for a 10-mode output projected system , are plotted in fig .
[ fig : modes ] .
the adjoint modes provide a direction for projecting the linearized equations onto the subspace spanned by the balancing modes .
since these modes are quite different from the pod and the balancing modes , the resulting models are also quite different from those obtained using the standard pod - galerkin technique wherein an orthogonal projection is used . since the models obtained using balanced truncation are known to perform better than the pod - galerkin models , as reported by @xcite , the better performance could be attributed to a better choice of projection using the adjoint modes . ) and the diagonal elements of the controllability ( @xmath178 , @xmath202 ) and observability ( @xmath203 , @xmath204 ) gramians of a 25-mode model with a 4 , 10 , and 20-mode output projection , for the unstable steady state at @xmath205.,height=230 ] since the reduced - order models of the stable - subspace dynamics are approximately balanced , the controllability and observability gramians of the @xmath95-dynamics of ( [ rom_compact ] ) , given by expressions ( [ gramians_stable ] ) , are approximately equal and diagonal .
further , their diagonal values are approximately the same as the hankel singular values @xmath206 obtained by the svd ( [ svd ] ) .
the diagonal values of the gramians and the singular values for different output projections are plotted in fig .
[ fig : unstable_hsv ] for a 30-state reduced - order model . with
increasing order of output projection , the hsvs converge to the case with full - state output , and the number of converged hsvs is roughly equal to the order of output projection , as was observed by @xcite .
we see that the diagonal elements of both the gramians are very close to the hsvs for the first 20 modes . for higher modes ,
the observability gramians are inaccurate , which is due to a small inaccuracy of the adjoint formulation mentioned in section [ sec : lin_adjoint ] . for controller design ,
we use models of order @xmath207 , for which these gramians are sufficiently accurate . , @xmath202 )
are compared with predictions of models with 4 ( @xmath203 , @xmath204 ) , 10 ( @xmath178 , @xmath208 ) , and 20 ( @xmath180 , @xmath209 ) modes.,title="fig:",height=163 ] , @xmath202 ) are compared with predictions of models with 4 ( @xmath203 , @xmath204 ) , 10 ( @xmath178 , @xmath208 ) , and 20 ( @xmath180 , @xmath209 ) modes.,title="fig:",height=163 ] finally , to test the accuracy of the reduced - order models , we compare the impulse responses of system ( [ ss_stable_numerical ] ) ( that is , restricted to the stable subspace ) with that of the model ( [ rom_compact ] ) , restricting @xmath210 .
in particular , we compare the outputs of the two systems , which are the projection onto the pod modes ; a representative case in fig .
[ fig : outputs_stable_oproj10 ] shows the results of 4 , 10 and 20 mode models of a system approximated using a 20-mode output projection ( the outputs are projection onto the leading 10 pod modes ) .
the first output , which is a projection onto the first pod mode , is well captured by all the models until @xmath211 , while the 20-mode model performs well for all time . also shown is the eleventh output , which is well captured only by the 20-mode model . as we will see later
, it is important to capture the higher - order outputs for design of observers . .
the state is then multiplied by the gain @xmath212 , computed based on the reduced - order model using lqr , to obtain the control input @xmath123.,height=124 ] .
comparison of the outputs @xmath213 and @xmath214 of a 12-mode reduced - order model ( @xmath215 ) with the projection of data from the linearized simulation ( @xmath216).,title="fig:",height=134 ] .
comparison of the outputs @xmath213 and @xmath214 of a 12-mode reduced - order model ( @xmath215 ) with the projection of data from the linearized simulation ( @xmath216).,title="fig:",height=134 ] the resulting models can now be used along with standard linear control techniques to obtain stabilizing controllers .
we use linear quadratic regulator ( lqr ) to compute the gain @xmath212 so that the eigenvalues of @xmath217 ( where the matrices were defined in ( [ rom_compact ] ) ) are in the left - half of the complex plane , and the input @xmath218 minimizes the cost @xmath219 = \int_0^\infty \ , ( a^\ast q a + u^\ast r u ) \
, dt , \label{lqr_cost}\end{aligned}\ ] ] where @xmath220 and @xmath221 are positive weights computed as follows .
we choose @xmath220 such that the first term in the integrand of ( [ lqr_cost ] ) represents the energy , that is , we use @xmath222 , with @xmath223 defined in ( [ output_compact ] ) .
the weight @xmath221 is chosen to be a multiple of the identity @xmath224 , and typically @xmath132 is chosen to be a large number @xmath225 , to avoid excessively aggressive controllers .
the control implementation steps are sketched in fig .
[ fig : cartoon_fullstate_control ] ; first compute the reduced - order state @xmath226 , using the expression ( [ init_model ] ) , then the control input is given by @xmath227 . here
, we derive the gain @xmath212 based on a 12-mode reduced - order model ( with 2 unstable and 10 stable modes ) , using @xmath228 , and include the same in the original linearized and nonlinear simulations .
the output is approximated using a 4-mode output projection .
the difference between the linear and nonlinear simulations is that , in the latter , the steady state field @xmath120 is subtracted from the state @xmath10 , before projecting onto the modes to compute the reduced - order state @xmath226 .
[ fig : contlin ] compares the model predictions with the projection of data from the simulations of the linearized system ( [ linear ] , [ lin_constraint ] ) , with a control input .
the initial condition used is the flow field obtained from an impulsive input to the actuator .
both the states shown in the figure eventually decay to zero , which implies that the perturbations decay to zero , thus stabilizing the unstable steady state .
more importantly , the model predicts the outputs accurately for the time horizon shown in the plots . vs. time @xmath229 , for full - state feedback control , with control turned on at different times in the base uncontrolled simulation .
the base case with no control ( @xmath178 ) has the unstable steady state as the initial condition , and transitions to periodic vortex shedding .
the control is tested for different initial conditions , corresponding to @xmath230 of the base case , and stabilizes the steady state in all the cases ( @xmath177).,title="fig:",width=432 ] we now use the same controller in the full nonlinear simulations and test the performance of the model for various perturbations of the steady state . a plot of the lift coefficient @xmath183 vs. time @xmath229 , with the control turned on at different times of the base simulation , is shown in fig .
[ fig : cont_fullfback_lift ] .
the initial condition for the base case ( no control ) is the unstable steady state ; eventually , small numerical errors excite the unstable modes and the flow transitions to periodic vortex shedding . in separate simulations ,
control is turned on at times @xmath231 corresponding to the base case . as the figure shows , the control is effective and is able to stabilize the steady state in each case , even when the flow exhibits strong vortex shedding .
we remark that the latter two of these perturbations are large enough to be outside the range of validity of the linearized system , but the control is still effective , implying a large basin of attraction of the stabilized steady state .
we also compare the output of the reduced - order model with the outputs of the nonlinear simulation ; the plots are shown in fig .
[ fig : outputs_control_oproj4 ] .
the models perform well for the initial transients , but for longer times fail to capture the actual dynamics .
this is not surprising as these perturbations are outside the range of validity of the linear models . for control purposes
, it appears to be sufficient to capture the initial transients ( approximately one period ) , during which the instability is suppressed to a great extent .
we remark that one could possibly compute nonlinear models by projecting the full nonlinear equations onto the balancing modes , or enhance the model subspace by adding pod modes of vortex shedding and the shift modes as proposed by @xcite to account for the nonlinear terms . of the uncontrolled case plotted in fig .
[ fig : cont_fullfback_lift ] .
comparison of the outputs @xmath213 and @xmath232 of a 12-mode ( 2 unstable and 10 stable modes ) reduced - order model ( @xmath215 ) with the projection of data from the full nonlinear simulation ( @xmath216).,title="fig:",height=134 ] of the uncontrolled case plotted in fig .
[ fig : cont_fullfback_lift ] .
comparison of the outputs @xmath213 and @xmath232 of a 12-mode ( 2 unstable and 10 stable modes ) reduced - order model ( @xmath215 ) with the projection of data from the full nonlinear simulation ( @xmath216).,title="fig:",height=134 ] finally , we note that the reduced - order model ( [ rom_compact ] ) decouples the dynamics on the stable and unstable subspaces , and also , the dynamics on the unstable subspace can be computed only using the unstable eigen - bases @xmath64 and @xmath65 .
thus , we could derive a control gain @xmath233 , based only on the two - dimensional unstable part of the model , such that the eigenvalues of @xmath234 are in the left half complex plane .
that is , we can obtain a stabilizing controller _ without _ modeling the stable subspace dynamics .
we have performed simulations to test such a controller and found that it also is capable of suppressing the periodic vortex shedding and thus results in a large basin of attraction for the stabilized steady state .
the choice of weight matrices @xmath220 and @xmath221 in the lqr cost ( [ lqr_cost ] ) needs to be different to obtain a comparable performance .
however , as shown in the next section , it is essential to model the stable subspace dynamics to design a practical controller based on an observer that reconstructs the entire flow field using a few sensor measurements .
the full - state feedback control of section [ sec : fullstatefb ] is not practical since it is not possible to measure the entire flow field . here
, we consider a more practical approach of measuring certain flow quantities at a small number of sensor locations .
we assume that we can measure the velocities at the sensors shown in fig .
[ fig : actuator_sensor ] , in the near - wake of the plate .
we remark that , even though these sensors are not experimentally realizable , they serve as a good testing ground for our models . and the sensor measurements @xmath52 are used as inputs to the observer , which reconstructs the reduced - order state @xmath235 .
this state is then multiplied by the gain @xmath212 , to obtain the control input @xmath123 .
both , the controller and observer gains @xmath212 and @xmath236 are computed based on the reduced - order model using lqr and lqg respectively.,height=144 ] using the reduced - order models derived as outlined in section [ sec : models ] , we design observers that dynamically estimate the entire flow field . since the models ( [ rom_compact ] ) have a different output ( projection onto certain modes )
, we modify the output equation so that it appropriately represents the sensor measurements . we replace the output equation of ( [ rom_compact ] ) with @xmath237 where @xmath238 and @xmath145 is the number of sensor measurements .
the matrix @xmath113 is sparse and extracts the values of the output of ( [ rom_compact ] ) at the sensor locations ; thus , each row of @xmath113 is filled with 0s except for the entry corresponding to a sensor measurement , which is 1 . with the output equation ( [ observer_output ] ) , we design an observer using a linear quadratic gaussian ( lqg ) estimation .
this method assumes that the errors in representing the state @xmath226 and and the measurement @xmath52 ( due to the inaccuracies of the model ) are stochastic gaussian processes , and results in an estimate @xmath235 of the state @xmath226 that is optimal in the sense that it minimizes the mean of the squared error ; refer to @xcite for details .
we now discuss briefly our procedure for modeling these noises ; consider the reduced - order model ( [ rom_compact ] ) , but with process noise @xmath239 and sensor noise @xmath50 which enter the dynamics as follows : @xmath240 a key source of the process ( state ) noise @xmath239 arises from model truncation , and second , from ignoring the nonlinear terms in the reduced - order model .
the nonlinearity of the dynamics is important , for instance , when the model is used to suppress vortex shedding .
a source of the sensor noise arises from two sources ; first , the state @xmath10 is approximated as a sum of a finite number of modes ( [ modal_exp ] ) , and second , in the output projection step , the output is considered as a projection of the ( approximated ) state @xmath10 onto a finite number of pod modes ( [ yrom_unstable_outproj ] ) . here ,
we approximate these two noises as gaussian processes whose variances are @xmath241 and @xmath242 gives the expected value . here , @xmath243 is the operator obtained by projecting the nonlinear navier - stokes equations ( [ navier1 ] ) onto the balancing modes @xmath46 , using the adjoint modes @xmath47 .
the state @xmath244 is obtained by projecting the snapshots , obtained from a representative simulation of the full nonlinear system , onto the balancing modes .
the representative simulation we used here is the base case , with no control , shown in fig .
[ fig : cont_fullfback_lift ] , which includes the transient evolution from the steady state to periodic vortex shedding .
the resulting estimator is of the form : @xmath245 where @xmath235 is the estimate of state @xmath226 , @xmath246 is the estimated output , and @xmath236 is the observer gain .
the estimator is then used along with the full - state feedback controller designed in section [ sec : fullstatefb ] to determine the control input ; a schematic is shown in fig .
[ fig : cartoon_obsv_control ] . norm ) obtained from an impulse response ( @xmath177 , @xmath202 ) of ( [ ss_stable_numerical ] ) , and the energy captured by 4 ( @xmath178 , @xmath204 ) , 10 ( @xmath179 , @xmath208 ) , and 20 ( @xmath180 , @xmath247 ) leading pod modes.,height=192 ] , @xmath202 ) , of an impulse response of ( [ ss_stable_numerical ] ) , compared with the reconstruction using 10 ( @xmath179 , @xmath208 ) and 20 ( @xmath178 , @xmath204 ) leading pod modes.,height=259 ] since the observability gramian corresponding to the pair @xmath248 is different from that for the pair @xmath249 , the model ( [ rom_compact ] ) with the output represented by ( [ observer_output ] ) , is not balanced . in principle , it is possible to construct reduced - order models of a system with the sensor measurements as the outputs , using the procedure of section [ sec : approx_baltrunc_unstable ] . since the number of outputs will be typically small , the output projection step would not be required for such models . however ,
if this were done , the cost function ( [ lqr_cost ] ) , based on total kinetic energy in the perturbation , would not be captured well by the model and the model would not be as effective for control design .
vs. time @xmath229 , for estimator - based feedback control , with control turned on at different times in the base uncontrolled simulation .
the base case ( @xmath178 ) is the same as in fig .
[ fig : cont_fullfback_lift ] , and the control is tested for different initial conditions , corresponding to @xmath230 of the base case ( @xmath177 ) . in both the cases , the controller stabilizes the flow to a small neighborhood of the steady state.,title="fig:",width=432 ] , @xmath204 ) by a 22-mode observer quickly converge to the actual states ( @xmath177 , @xmath202 ) .
the initial conditions used are those corresponding to @xmath250 ( top and bottom ) of the uncontrolled case shown in fig .
[ fig : cont_obsvfback_lift].,title="fig:",height=134 ] , @xmath204 ) by a 22-mode observer quickly converge to the actual states ( @xmath177 , @xmath202 ) .
the initial conditions used are those corresponding to @xmath250 ( top and bottom ) of the uncontrolled case shown in fig .
[ fig : cont_obsvfback_lift].,title="fig:",height=134 ] + , @xmath204 ) by a 22-mode observer quickly converge to the actual states ( @xmath177 , @xmath202 ) .
the initial conditions used are those corresponding to @xmath250 ( top and bottom ) of the uncontrolled case shown in fig .
[ fig : cont_obsvfback_lift].,title="fig:",height=134 ] , @xmath204 ) by a 22-mode observer quickly converge to the actual states ( @xmath177 , @xmath202 ) .
the initial conditions used are those corresponding to @xmath250 ( top and bottom ) of the uncontrolled case shown in fig .
[ fig : cont_obsvfback_lift].,title="fig:",height=134 ] the models used in section [ sec : fullstatefb ] for full - state feedback were those of a system whose stable - subspace output was the velocity field projected onto the leading 4 pod modes .
these 4 pod modes capture only about @xmath251 of the energy , but the resulting models were effective in suppressing vortex shedding .
however , for observer design , this representation of the output is inadequate , as the energy content of the flow at the sensor locations is very small , while the pod modes capture the energetically dominant modes .
hence , a greater number of pod modes is required to accurately represent the velocity at the sensor locations .
the temporal evolution of the energy content of the flow , obtained from an impulse response of the system restricted to evolve on the stable subspace , is plotted in figure [ fig : l2norm ] .
also plotted is the energy content of the same flow , but projected onto the leading 4 , 10 and 20 pod modes ; thus , a 4-mode projection leads to noticeable errors , while both 10- and 20-mode projections accurately represent the energy .
the velocity field at the sensor locations , reconstructed by 10 and 20 pod modes , plotted in figure [ fig : sensor_reconst ] , shows that a 10-mode projection does not accurately represent the velocities at the sensor locations . since 20 pod modes are sufficient to represent these velocities ,
we derive models using a _ 20-mode output projection _ , and use the same for observer design .
the models obtained using the modified output ( [ observer_output ] ) are used to design dynamic observers based on the vertical ( @xmath50- ) velocity measurements at the sensor locations .
a 22-mode reduced - order model , with 2 and 20 modes describing the dynamics on the unstable and stable subspaces respectively , is used to design a kalman filter for producing an optimal estimate of the velocity field based on gaussian approximations of error terms ( [ process_noise_model ] , [ sensor_noise_model ] ) .
this estimate is then used along with reduced - order model controller to determine the control input , as shown in fig .
[ fig : cartoon_obsv_control ] . the results of this observer - based controller ( or compensator )
are shown in figs .
[ fig : cont_obsvfback_lift ] , [ fig : states_obsvcontrol_oproj20 ] .
the compensator again stabilizes the unstable operating point , and furthermore , the observer reconstructs the reduced - order model states accurately .
initially , the observer has no information about the states ( the initial condition is zero ) , but it quickly converges to and follows the actual states .
there is a key difference from the full - state feedback control , that the compensator does not exactly stabilize the unstable operating point but converges to its small neighborhood .
the reason is that the observer design is based on the velocities at the sensor locations projected onto the leading 20 pod modes , rather than the exact velocities at these locations .
these small errors enter into the observer s dynamics in the same way that sensor noise enters , resulting in small errors in the state estimate .
we presented an algorithm for developing reduced - order models of the input - output dynamics of high - dimensional linear unstable systems , extending the approximate balanced truncation method developed by @xcite for stable systems .
we assumed that the dimension of the unstable eigenspace is small and the corresponding global eigenmodes can be numerically computed .
the modeling procedure treats the dynamics on the unstable subspace exactly and obtains a reduced - order model of the dynamics on the stable subspace . in a proof - of - concept study ,
the procedure was applied to control the 2-d low - reynolds - number flow past a flat plate at a large angle of attack @xmath128 , where the natural flow state is periodic vortex shedding .
we first performed a continuation study at @xmath3 and computed the branch of steady states with @xmath128 varying from @xmath252 to @xmath2 ; we show that the flow undergoes a hopf bifurcation from steady state to periodic shedding at @xmath253 .
we developed reduced - order models of the linearized dynamics at @xmath5 actuated by a localized body force close to the leading edge of the plate .
the outputs were considered to be the entire flow field , projected onto the unstable eigenmodes and the leading pod modes of the impulse response simulation ( restricted to the stable subspace ) .
we developed stabilizing controllers based on the reduced - order models to stabilize the unstable steady state and showed that the models agreed well with the actual simulations .
we also included the controllers in the full nonlinear simulations , and showed that they had a large - enough basin of attraction to even suppress the vortex shedding .
for such large perturbations , however , the model agreement with the full simulation was good only for short times .
a natural step towards improving these models would be to project the full nonlinear equations onto the balancing modes to obtain nonlinear models .
alternately , the balanced models , which accurately capture the transient dynamics , could be combined with the pod - based models using shift - modes of @xcite which accurately capture vortex shedding and some of the transient dynamics .
an interesting future direction is development of algorithms to compute _ nonlinear _ balanced models , for instance based on the theoretical work of @xcite . instead of computing nonlinear models , here we pursued a step towards more practical controllers by considering an observer - based control design , in which the outputs were modified to be just two near - wake velocity measurements .
the nonlinear terms in the equations , which our models do not capture , were treated as process noise , and the error in modeling the outputs was treated as sensor noise .
we designed a 22-mode reduced - order observer which reconstructed the flow field accurately , and along with the controllers , suppressed vortex shedding and stabilized the flow in a small neighborhood of the unstable steady state .
we remark that the actuator and sensors considered here are not practically realizable , but the methodology presented here can be extended to a more practical actuation such as blowing and suction through the plate and measurements using surface pressure sensors .
furthermore , the choice of sensor locations in this study was ad hoc , and an interesting problem is of finding the optimal sensor locations , for a given actuator .
the controllers present here are designed to operate at a fixed set of parameters ( such as @xmath254 and @xmath128 ) , and it would be interesting to test their performance at off - design parameter values ; the study on the performance of models of the linearized channel flow at off - design @xmath254 by @xcite shows promise in that direction . a motivation for the choice of our model problem was to develop tools towards manipulating wakes of micro - air vehicles .
recently , @xcite performed a numerical study of flow past low - aspect - ratio plates , and a future direction we intend to undertake is to perform a detailed continuation study of the same flow to explore the existence and stabilization of high - lift unstable steady states in this 3-d flow .
the authors would like to thank tim colonius and kunihiko taira for their tremendous help in adopting their immersed boundary solver .
the authors would also like to thank ioannis g. kevrekidis , sung joon moon and liang qiao for their help with the timestepper - based steady - state analysis .
this work was funded by the u. s. air force office of scientific research grant fa9550 - 05 - 1 - 0369 and this support is gratefully acknowledged .
without loss of generality , a transformation @xmath14 ( and its inverse ) that decouples the stable and unstable dynamics of ( [ ss ] ) can be written as : @xmath255 where the columns of @xmath256 and @xmath257 span the unstable and stable _ right _ eigenspaces of @xmath11 , while the columns of @xmath258 and @xmath259 span the unstable and stable _ left _ eigenspaces of @xmath11 .
further , these matrices are scaled such that @xmath260 and @xmath261 .
the transformation ( [ transf ] ) decouples the dynamics of ( [ ss ] ) as given in ( [ decouple_ss ] ) with the various matrices defined as follows : @xmath262 using ( [ transf ] ) in ( [ gramians_general ] ) , the gramians of the original system ( [ ss ] ) are @xmath263 where , @xmath22 and @xmath23 are the gramians corresponding to the system defined by @xmath21 , while @xmath25 and @xmath26 are the gramians corresponding to the system defined by @xmath24 .
let @xmath264 be the transformation that balances the gramians @xmath25 and @xmath26 , while @xmath265 be the transformation that balances @xmath22 and @xmath23 .
then , it can be verified that the transformation that balances the gramians @xmath96 and @xmath97 is given by @xmath266 thus , the balancing transformation consists of two parts @xmath64 and @xmath79 which respectively balance the dynamics on the unstable and stable subspaces of @xmath11 . as per the technique of @xcite , a reduced - order model can be obtained by truncating the columns of @xmath46 that correspond to the relatively uncontrollable and unobservable states .
as we will show now , the algorithm outlined in section [ sec : approx_baltrunc_unstable ] essentially computes the leading columns of @xmath79 ( and the corresponding rows of its inverse ) . we show that the controllability gramian of the stable dynamics of ( [ ss ] ) , which are defined by ( [ ss_stable ] ) , is the same as the `` stable '' part of the gramian defined in ( [ gramians_general_expand ] ) .
note that using ( [ transf ] ) and the definition ( [ proj_stable ] ) , the projection operator @xmath67 can be written as @xmath267 using the definition ( [ gramians_stable ] ) , the controllability gramian of ( [ ss_stable ] ) is @xmath268 which is the same as the stable part of @xmath96 .
similarly , it can be shown that the observability gramian @xmath269 of ( [ ss_stable ] ) is the same as the `` stable '' part of the observability gramian @xmath97 : @xmath270 thus , balancing the gramians @xmath269 and @xmath271 is identical to balancing the parts of the gramians @xmath96 and @xmath97 of the original system ( [ ss ] ) that are related to the dynamics on the stable subspace of @xmath11 .
in this appendix , we derive the adjoint of the linearized semi - discrete equations ( [ linear],[lin_constraint ] ) .
let @xmath272 be the weighting functions corresponding to @xmath273 . then , using the inner product defined in equation ( [ ip ] ) , the weak form of ( [ linear],[lin_constraint ] ) is : @xmath274 integrating by parts with respect to @xmath229 and rearranging terms , @xmath275 for linearization about stable steady states , @xmath276 , as @xmath277 , and if the adjoint equations are integrated backwards in time , @xmath278 .
so , the last term on the left hand side of equation ( [ weak2 ] ) vanishes identically .
if equation ( [ weak2 ] ) is to hold for all values of @xmath135 and @xmath157 , we get the following adjoint equations hold : @xmath279 where @xmath280 and @xmath158 can be thought of as the weighting functions corresponding to the streamfunction @xmath145 and the flux @xmath144 respectively .
now , equations ( [ adjoint1],[adjoint2 ] ) have the same form as ( [ linear],[lin_constraint ] ) except for the nonlinear term .
thus , the same time - integration scheme can be used for both , with the appropriate ( linearized ) nonlinear terms . | we present an estimator - based control design procedure for flow control , using reduced - order models of the governing equations , linearized about a possibly unstable steady state .
the reduced - order models are obtained using an approximate balanced truncation method that retains the most controllable and observable modes of the system .
the original method is valid only for stable linear systems , and in this paper , we present an extension to unstable linear systems .
the dynamics on the unstable subspace are represented by projecting the original equations onto the global unstable eigenmodes , assumed to be small in number .
a snapshot - based algorithm is developed , using approximate balanced truncation , for obtaining a reduced - order model of the dynamics on the stable subspace .
the proposed algorithm is used to study feedback control of two - dimensional flow over a flat plate at a low reynolds number and at large angles of attack , where the natural flow is vortex shedding , though there also exists an unstable steady state . for control design
, we derive reduced - order models valid in the neighborhood of this unstable steady state .
the actuation is modeled as a localized body force near the leading edge of the flat plate , and the sensors are two velocity measurements in the near - wake of the plate .
a reduced - order kalman filter is developed based on these models and is shown to accurately reconstruct the flow field from the sensor measurements , and the resulting estimator - based control is shown to stabilize the unstable steady state . for small perturbations of the steady state , the model accurately predicts the response of the full simulation .
furthermore , the resulting controller is even able to suppress the stable periodic vortex shedding , where the nonlinear effects are strong , thus implying a large domain of attraction of the stabilized steady state . |
collisionless magnetic reconnection drives explosive events in space plasma environments . on the large scale
, reconnection is an mhd - scale process but is facilitated by a compact `` diffusion region '' surrounding the reconnection point , where kinetic physics plays a role . in the hall reconnection
model,@xcite it is thought that the electron diffusion region ( edr ) , where the electrons decouple from field lines , is embedded in the ion s diffusion region @xcite .
the edr is related to a localized dissipation region that allows fast reconnection , and it also adjusts the overall system evolution to the outer ion physics.@xcite owing to its importance in reconnection , the edr is one of the most important targets of nasa s upcoming magnetospheric multiscale ( mms ) mission ( http://mms.space.swri.edu/ ) , which will be the first to probe the electron - scale physics .
the hall reconnection model has been recently challenged by kinetic particle - in - cell ( pic ) simulations that sufficiently resolve electron - scale structures@xcite .
it has been argued that the edr is no longer localized but it unexpectedly stretches in the outflow directions@xcite .
furthermore , it appears to embody a two - scale structure of inner and outer regions @xcite .
minor differences aside , the inner region is a compact region containing the reconnection site .
it features a strong out - of - plane electron current and a dissipative electric field @xmath0 to transport the magnetic fields .
the outer region extends in the outflow direction , accompanied by a fast electron jet , often denoted as the `` super - alfvnic '' jet . since the jet outruns the moving magnetic field@xcite
, the out - of - plane component of @xmath1 has the opposite polarity from that in the inner region .
satellite observations corroborate these structures in near - earth reconnection sites@xcite .
the role of the edr is a subject of recent debate . in principle
, many works have agreed that the reconnection rate is controlled by the inner region or the inner region adjusts its size to the global reconnection rate @xcite .
its long - term , time - dependent behavior remains unclear .
@xcite argued that the ( inner ) edr stretches in the outflow direction .
they further proposed an idea that the secondary island formation maintains fast reconnection by cutting and shortening the elongated edr .
in contrast , @xcite demonstrated that the inner region remains at the finite length and that the reconnection is fast , quasi - steady , and laminar . regarding the outer region
, it does not appear to constrain the reconnection rate @xcite , because reconnection remains fast while the electron jet extends a large distance .
@xcite found that the electron flow essentially consists of the @xmath2 convection and the diamagnetic current for the field reversal .
thus , even though it is frequently mentioned by the outer edr @xcite , it is not clear whether the outer region is dissipative .
furthermore , it is not well understood how the electron jet is terminated .
recently , @xcite proposed a new measure to identify the dissipation region of reconnection . considering the energy dissipation in the rest frame of electron s bulk motion
, they introduced the _ electron - frame dissipation measure _ , @xmath3,\end{aligned}\ ] ] where @xmath4^{-1/2}$ ] is the lorentz factor for the electron velocity and @xmath5 is the charge density .
this measure is a lorentz - invariant , and it is related to the nonideal energy conversion , which is essential for reconnection problems .
their pic simulations demonstrated that @xmath6 identifies a single dissipation region surrounding the reconnection site .
this is substantially different from the previous two - scale picture .
the purpose of this paper is to organize our knowledge on the dissipation region in collisionless magnetic reconnection in line of recent theoretical progress @xcite . using high - resolution two - dimensional pic simulations
, we perform a detailed investigation of internal structures in basic antiparallel configurations .
in particular , the previous two - scale edr are reexamined by using the electron - frame dissipation measure @xmath6 .
the structures are better understood in a hall magnetic geometry .
we further explore the termination region of the electron jet .
this paper is organized as follows .
we briefly review the dissipation measure @xmath6 in sec .
[ sec : theory ] .
we describe our numerical setup in sec . [
sec : setup ] .
we present the simulation results in sec .
[ sec : results ] .
we carefully investigate fine structures in the following subsections .
section [ sec : discussion ] contains discussion and summary .
for better understanding , we review a basic concept in ref . within the nonrelativistic physics .
the essence is to evaluate the ohmic dissipation in the rest frame of electron s bulk motion .
let the prime sign ( @xmath7 ) denote quantities in the electron frame .
a simple algebra yields @xmath8 the last term comes from the convection current ( @xmath9 ) , arising from the motion of the charged frame .
the electric field in the electron frame is given by the nonideal electric field , @xmath10 the energy dissipation in the electron frame yields @xmath11 this is an invariant scalar with respect to the galilean transformation and is equivalent to eq .
[ eq : edr ] in the limit of @xmath12 .
we employ this nonrelativistic formula throughout this paper .
we often discuss the composition of @xmath6 in the following form , @xmath13 the last term deals with the effective energy conversion by the convection current .
hereafter we refer to it by the `` charge term . '' in a neutral plasma , one can obtain @xmath14 by dropping the charge term .
this reduced form is equivalent to the joule dissipation in the plasma rest frame in @xcite , who discussed the energy transfer in mhd and kinetic reconnection systems .
we use a partially - implicit pic code @xcite .
the length , time , and velocity are normalized by the ion inertial length @xmath15 , the ion cyclotron time @xmath16 , and the ion alfvn speed @xmath17 , respectively . here , @xmath18 is the ion plasma frequency , @xmath19 is the reference density , and @xmath20 is the reference magnetic field .
we employ a harris - like configuration , @xmath21 and @xmath22 $ ] , where @xmath23 is the half thickness of the current sheet .
the electron - ion temperature ratio is @xmath24 .
periodic ( @xmath25 ) and reflecting wall ( @xmath26 ) boundaries are used . in the main run ( run 1a ) , the mass ratio is @xmath27 .
the ratio of the electron plasma frequency to the electron cyclotron frequency is @xmath28 . in this case
, the speed of light is @xmath29 .
the domain of @xmath30\times[-19.2 , 19.2]$ ] is resolved by @xmath31 cells , and @xmath32 particles are used .
we carry out several other runs , changing the two key parameters : the mass ratio @xmath33 and the @xmath34 parameter .
all parameters are presented in table . [ table ] .
in all cases , reconnection is triggered by a small flux perturbation .
$ ] , where @xmath36 is the typical amplitude of the perturbed fields .
the initial current is set up accordingly .
.simulation parameters [ cols="^,^,^,^,^,^",options="header " , ] [ table ]
we identify the dominant reconnection site by finding the minimum in @xmath25 of the magnetic flux @xmath37 .
then we discuss the system evolution by using the normalized flux transfer rate or the reconnection rate @xmath38 , @xmath39 where the subscript @xmath40 denotes quantities at @xmath41 upstream of the dominant reconnection site . the normalized reconnection rates @xmath38 ( eq . [ eq : recrate ] ) as a function of a time ( @xmath42 ) . , title="fig : " ] the normalized reconnection rates @xmath38 ( eq . [ eq : recrate ] ) as a function of a time ( @xmath42 ) .
, title="fig : " ] figure [ fig : recrate ] shows the time evolution of @xmath38 .
the rate quickly increases , modestly overshoots at @xmath43 , and then approaches a quasisteady value of @xmath44@xmath45 .
such quasisteady evolution corroborates previous investigations.@xcite the rates are insensitive to the mass ratio in this normalization.@xcite the parameter @xmath34 does not make a significant difference also .
we confirmed that the reconnection rates are independent of domain sizes in the timescale of our interest , by carrying out supplemental small runs .
meanwhile , as will be discussed later , the periodic domain effects start to modulate the outflow structures after @xmath46 in runs 1a and 1b .
magnetic field line structure in run 1a , averaged over @xmath47-@xmath48 .
rear panel : the out - of - plane magnetic field @xmath49 .
front panel : the electron - frame dissipation measure @xmath6 , title="fig : " ] magnetic field line structure in run 1a , averaged over @xmath47-@xmath48 .
rear panel : the out - of - plane magnetic field @xmath49 .
front panel : the electron - frame dissipation measure @xmath6 , title="fig : " ] figure [ fig:3d ] presents a snapshot of run 1a .
physical quantities are averaged over @xmath47@xmath48 . at this stage ,
the reconnection is steadily going on as can be seen in figure [ fig : recrate ] , and the characteristic structures are well developed .
this is the stage of our primary interest .
the gray lines show magnetic field lines in 3d .
magnetic field lines are flipped , because they tend to follow the electron s flow in the @xmath50 direction .
this is a well - known signature of hall reconnection .
for example , @xcite showed a similar field - line structure by using hall mhd simulations .
for example , @xcite found a similar structure in an experiment device .
our fully - kinetic simulation is consistent with these results .
the rear panel in figure [ fig:3d ] shows the out - of - plane magnetic field @xmath49 .
the color indicates the polarity : @xmath51 in red and @xmath52 in blue .
the projection of flipped field lines leads to a characteristic `` quadrupole '' pattern@xcite in @xmath49 .
in addition , as indicated in figure [ fig:3d ] , there are two `` magnetic cavities '' in the outflow region , where @xmath49 becomes weaker .
this signature is often found in pic simulations of hall reconnection , but this has never been discussed before .
the front panel shows the electron - frame dissipation measure @xmath6 ( eq . [ eq : edr1 ] ) .
a central red region of @xmath53 is the dissipation region@xcite .
we will investigate this region in detail later in this paper .
figure [ fig:35 ] show various averaged quantities over @xmath47-@xmath48 in run 1a .
figure [ fig:35](a ) shows the electron outflow velocity @xmath54 .
one can see bi - directional jets from the reconnection site .
they travel much faster than the upstream alfvn speed @xmath55 , which approximates the outflow speed in an mhd scale .
the electron jets are remarkably narrow .
their widths are on an order of the local electron inertial length . outside the separatrices , there are weak reverse flows toward the reconnection site . in figure
[ fig:35](b ) we show the electron current density @xmath56 .
it is distinctly strong in a narrow region near the neutral plane ( @xmath57 ) .
we note that it is remarkably flat in the @xmath25-direction .
thus , @xmath56 appears to be a good marker of the interesting region .
hereafter we call the region an `` electron current layer .
'' figure [ fig:35](c ) shows the vertical electric field @xmath58 . near the reconnection site
, one can see a bipolar signature across the electron current layer : @xmath58 is negative for @xmath59 and positive for @xmath60 .
this is called the hall or polarization electric field @xcite .
this is supported by charge separation between a broader distribution of meandering ions and a narrow distribution of meandering electrons .
the bipolar peaks @xmath61 are typically @xmath62 times stronger than the reconnection electric field @xmath63 . in the inflow region , outside the electron current layer , magnetized electrons tend to travel in the -@xmath64-direction@xcite . in the outflow region ,
the @xmath58 region consists of two parts : the large - scale x - shaped structure along the separatrices and the top / bottom boundaries of the electron current layer@xcite .
we think the latter is related to the electron meandering motion , similarly as the central region .
in addition , as mentioned by a previous work@xcite , one can recognize an inverted signature of @xmath58 in a very vicinity of the neutral plane ( @xmath65 ) inside the electron current layer .
for example , @xmath58 is positive for @xmath59 and negative for @xmath60 around @xmath66 and @xmath67 .
these `` inversion electric field''@xcite are supported by charge separation inside the electron meandering orbit . shown in figure [ fig:35](d )
is the out - of - plane component of the electron ohm s law , @xmath68_y$ ] , in unit of @xmath69 .
one can recognize two distinct regions along the electron jets .
one is the compact region surrounding the reconnection site , where @xmath70 is positive .
the other regions extend in the outflow direction and @xmath70 is negative , because the fast electron jets outrun the field convection .
those two are often referred as the inner and outer edrs in recent literature @xcite . note that @xmath71 just tells us that the ideal assumption breaks down .
it does not always indicate magnetic dissipation .
figure [ fig:35](e ) presents the electron - frame dissipation measure @xmath6 ( eq . [ eq : edr1 ] ) in units of @xmath72 , where @xmath73 is the initial harris current density .
note that the color scale is different from a reduced color scale in fig .
[ fig:3d ] .
one can see that @xmath6 gives a different picture from figure [ fig:35](d ) .
the dissipation region with @xmath53 is located near the reconnection point [ the red region near @xmath74 in fig . [ fig:35](e ) ] . on the other hand
, @xmath6 is weakly negative in the outflow regions , as can be seen in light green in the figure .
there , its amplitude @xmath75 is substantially smaller than near the reconnection point .
figure [ fig:35](f ) presents the last charge term in eq .
[ eq : edr2 ] in the same unit as fig . [ fig:35](d ) .
this term appears only in the close vicinity of the above central region of @xmath53 , and it is an order of magnitude smaller than @xmath6 . ( a ) the out - of - plane component of the electron nonideal condition @xmath76 around the reconnection site , ( b ) the electron - frame dissipation measure @xmath6 , and ( c ) @xmath6 in run 2a .
, title="fig : " ] ( a ) the out - of - plane component of the electron nonideal condition @xmath76 around the reconnection site , ( b ) the electron - frame dissipation measure @xmath6 , and ( c ) @xmath6 in run 2a . , title="fig : " ] here , we focus on the dissipation region .
the panels in figure [ fig : dr ] present the electron nonidealness @xmath70 , the dissipation measure @xmath6 , and the dissipation measure @xmath6 in run 2a at lower mass ratio of @xmath77 .
we find that the lengths of the electron current layer and the dissipation region are nearly stationary in the later phase of simulations .
the panels in figure [ fig : dr ] are taken from the stationary phase .
comparing figs .
[ fig : dr](a ) and ( b ) , one can see that the dissipation region is longer in the outflow direction than the inner region defined by @xmath70@xcite . the comparison of figs .
[ fig : dr](b ) and ( c ) suggests that the size of the dissipation region is controlled by the electron mass . in our ion - based units , the electron inertial length scales with @xmath78 and the electron bounce width scales with @xmath79 ( refs . , ) .
we find that the length and the width of the dissipation region in run 1a is smaller than those in run 2a by a factor of @xmath80-@xmath81 .
this is intermediate between @xmath82 and @xmath83 .
the scaling to the mass ratio needs further investigation , but electron physics appears to be essential for the dissipation region .
composition of the dissipation measure @xmath6 ( a ) along the inflow line ( @xmath84 ) in run 1a , ( b ) along the inflow line ( @xmath85 ) in run 1b , and ( c ) along the outflow line ( @xmath57 ) in run 1a , , title="fig : " ] composition of the dissipation measure @xmath6 ( a ) along the inflow line ( @xmath84 ) in run 1a , ( b ) along the inflow line ( @xmath85 ) in run 1b , and ( c ) along the outflow line ( @xmath57 ) in run 1a , , title="fig : " ] next , we discuss the composition of the dissipation measure @xmath6 . shown in figure [ fig : de ] are contributions from the terms in eq .
[ eq : edr2 ] and the charge separation @xmath86 along the inflow and outflow lines .
the shaded regions indicate the dissipation region by the sign ( @xmath53 ; orange ) and the @xmath87-folding length of @xmath6 ( the inner red region ) , respectively .
the blue dotted lines indicate the electron ideal condition , @xmath68_y=0 $ ] .
the nonideal electric field @xmath70 is positive between the two dotted lines . in the inflow direction in run 1a [ fig .
[ fig : de](a ) ] , @xmath6 is well localized near the neutral plane @xmath57 .
its main contributor is the @xmath64-term ( @xmath88 ) .
the charge term ( @xmath89 ) is responsible for @xmath90 of @xmath6 .
this term has a double peak structure and comes from a double - peak charge distribution ( the red dashed line ) , due to the electron bounce motion .
the charge separation is up to @xmath90 in this case .
the @xmath26-term ( @xmath91 ) has also double peaks , because both @xmath92 and @xmath93 have bipolar signatures . however , since @xmath92 is small , the @xmath26-term is a very minor contributor to @xmath6 .
the @xmath25-term is negligible also .
figure [ fig : de](b ) shows the composition in run 1b at lower ratio of @xmath94 . in this case , the charge term is responsible for -@xmath95 of @xmath6 , because the charge separation is significant , @xmath96 .
a lower @xmath34 allows more significant charge separation .
we confirm this trend at the mass ratio @xmath77 , too . in the outflow direction
[ figure [ fig : de](c ) ] , one can see that the @xmath25-term ( @xmath97 ) is another contributor to @xmath6 . both @xmath6 and @xmath97 look noisy in
the outflow regions , because @xmath98 is noisy .
the gray region is defined by the spatial location of a maximum electron outflow velocity .
let us compare the dissipation region with the conventional inner region@xcite .
in the inflow direction [ fig .
[ fig : de](a ) ] , the inner region has been defined by the electron nonidealness @xmath99 ( the blue dotted lines ) .
both orange and red dissipation regions are thinner than the conventional inner region , bounded by the blue dotted lines [ fig .
[ fig : de](a ) ] .
one case see that the @xmath87-folding region ( red ) is relevant to the electron bouncing region , featuring the charge separation .
on the other hand , there have been several definitions in the outflow direction : the electron outflow velocity ( gray region)@xcite , the sign of the electron nonidealness ( the dotted lines)@xcite , or the out - of - plane current profile@xcite .
note that klimas et al.@xcite employed a different formula @xmath100_y > 0 $ ] , but this is almost equivalent to the popular criteria @xmath101 .
( color online ) the hall structure of the dissipation region .
, title="fig : " ] ( color online ) the hall structure of the dissipation region .
, title="fig : " ] in the outflow direction , the dissipation region is longer than the inner region defined by @xmath70@xcite .
one can intuitively understand this , considering the magnetic geometry @xcite . since the field lines are flipped ( fig .
[ fig:3d ] ) , we consider the curved surface as illustrated in figure [ fig : surface ] .
hereafter we call it the hall surface . as the field lines rotate their directions in the surface
, @xmath102 ( @xmath103 ) is normal to it .
the @xmath102-aligned ( out - of - plane ) component of @xmath1 is responsible for the nonideal field convection in the hall surface . meanwhile , at sufficiently high @xmath34 , one can reduce the dissipation measure to @xmath104 .
this is controlled by the @xmath102-aligned component of @xmath1 .
in other words , @xmath6 automatically takes care of the out - of - plane component of @xmath1 with respect to the hall surface .
thus , as we move to the outflow direction , we should consider @xmath105 as well as @xmath70 .
the positive-@xmath6 dissipation region is always longer than the positive-@xmath70 region@xcite , because @xmath6 further includes contributions from @xmath105 .
according to our preliminary results ( klimas et al . in prep ) , the former is commonly twice longer than the latter , even when the reconnection region is disturbed by secondary magnetic islands . on the other hand , the dissipation region is shorter than the region defined by the peak location of @xmath54@xcite .
we find that @xmath54 is not very useful to define an important region due to the following reasons .
first , it is difficult to identify the peak location , when @xmath54 gradually changes in the outflow direction .
second , the physical meaning of the maximum @xmath54 is unclear in the hall geometry ( fig . [
fig : surface ] ) .
the hall surface is already rotated around there , and so , @xmath54 is highly influenced by the out - of - plane diamagnetic flow@xcite and the angle of the hall surface . practically , we prefer to define the dissipation region by the @xmath87-folding length ( red ) rather than by its sign ( orange ) .
this is because in some cases the slope of @xmath6 is very flat around @xmath106 in the inflow directions . in figures [ fig : de](a ) and
( c ) , the aspect ratios of the dissipation region are @xmath107 ( @xmath108 ) by the @xmath87-folding length and @xmath109 ( @xmath110 ) by the positive dissipation , respectively .
these values reasonably fit to the typical reconnection rate of 0.1 .
next , we visit the outflow regions [ @xmath111 or @xmath112 ; fig . [ fig:35](a ) ] .
one can see the narrow fast electron jets@xcite .
they are flanked by slow uniform flow regions .
the region is denoted as the pedestal@xcite . in the pedestal ,
the @xmath70 component is negligible , suggesting quasi - ideal electron flows .
there is a boundary layer with weak @xmath105 and @xmath93 between the jet and the pedestal .
importantly , as shown in figure [ fig:35](e ) , the dissipation measure is weakly negative @xmath113 in the electron jet . from the viewpoint of the hall surface ( fig .
[ fig : surface ] ) , the out - of - plane component ( @xmath114-aligned component ) of @xmath1 is not exactly zero in a rotated frame @xcite .
the electron jet still outruns the field convection in the outflow direction in the hall surface .
this leads to @xmath115 in the electron jet . on the other hand
, @xmath6 is weakly positive in the pedestal and it often has positive peaks in the boundary layers between the jet and the pedestal .
this is visible in particular in the downstream ( @xmath116 and @xmath117 ) and is more evident at later stages .
the fact @xmath118 tells us that the electron jet region is not dissipative . to understand this , it is useful to discuss the mhd energy balance , as given in previous literature@xcite .
we consider a mass - averaged mhd velocity in an electron - ion plasma , @xmath119 the energy transfer from the fields to plasmas in the mhd frame @xmath120 is similarly given by @xmath121 a small amount of algebra leads to @xmath122 thus @xmath118 indicates @xmath123 , i.e. , the plasma s energy is converted to the field energy even in the mhd frame .
this contradicts with the concept of magnetic dissipation or diffusion , which consumes the field energy in the mhd frame .
no dissipation process takes place here . instead ,
this region is better described as `` anti - dissipative . ''
( color online ) the flip angle [ degeree ] of the hall magnetic field lines ( a ) at @xmath47-@xmath48 and ( b ) at @xmath124-@xmath125 .
( c ) 1d cuts along the two lines .
, title="fig : " ] ( color online ) the flip angle [ degeree ] of the hall magnetic field lines ( a ) at @xmath47-@xmath48 and ( b ) at @xmath124-@xmath125 .
( c ) 1d cuts along the two lines .
, title="fig : " ] figure [ fig : angle](a ) shows a flip angle of hall magnetic field lines .
we calculate it by an arctangent of @xmath49 over @xmath126 , which is a good indicator outside the @xmath127 plane .
the sign of the angle is neglected to better see the structure . as can be seen in figure [ fig:3d ] ,
the magnetic field lines are extremely flipped .
the angle is almost @xmath128 in the electron jet region .
such a strong flipping is found in recent studies ( e.g. , the hodogram analysis by @xcite ) .
the angle changes from zero to @xmath128 from the upstream region to the electron jet region .
further observation tells us that the angle has a two - scale structure in the same @xmath25-location : a moderately flipped pedestal and a highly flipped electron jet .
this is more evident at a later time of @xmath124-@xmath125 [ fig .
[ fig : angle](b ) ] and the 1d cuts along the vertical lines [ fig .
[ fig : angle](c ) ] . in the pedestal region , the angle is @xmath129 .
note that the hall magnetic field @xmath130 is usually comparable to the antiparallel field @xmath131 in many simulations .
we think the @xmath6-pattern and the field line geometry is relevant . in the electron jet , the field lines
are strongly flipped , and @xmath113 implies that the electromagnetic fields gain energy in the electron frame . outside the jet ,
the magnetic field lines are mildly flipped and @xmath132 indicates that plasmas gain energy from the fields .
the electron - field energy transfer @xmath133 shows similar signatures as @xmath6 .
we speculate that the bulk kinetic energy of the fast electron jet is transferred from the jet to the pedestal via the magnetic field lines that thread two regions .
this process needs to be further investigated .
( a ) the stack plot of the electron outflow speed @xmath54 ( thin lines ) and the reconnected magnetic field @xmath134 ( thick lines ) .
( b - e ) averaged properties near the electron jet front at @xmath47-@xmath48 : ( b ) @xmath126 , ( c ) the out - of - plane magnetic field @xmath49 , ( d ) the electron temperature @xmath135 in units of @xmath136 , and ( e ) the electron energy transfer @xmath137 in units of @xmath72 . , title="fig : " ] ( a ) the stack plot of the electron outflow speed @xmath54 ( thin lines ) and the reconnected magnetic field @xmath134 ( thick lines ) .
( b - e ) averaged properties near the electron jet front at @xmath47-@xmath48 : ( b ) @xmath126 , ( c ) the out - of - plane magnetic field @xmath49 , ( d ) the electron temperature @xmath135 in units of @xmath136 , and ( e ) the electron energy transfer @xmath137 in units of @xmath72 . ,
title="fig : " ] we find a new shock - like structure at the fast electron jet front . shown in figure [ fig : eshock](a )
are the time evolution of the electron outflow velocity ( @xmath54 ) and the reconnected magnetic field ( @xmath134 ) . at @xmath138
, one can see that @xmath134 becomes twice stronger ( from @xmath139 to @xmath140 ) across the transition region , @xmath141 . on the other hand
, the fast electron jet suddenly slows down there , and then the electrons are magnetized further downstream .
we also find that the electron temperature and pressure increase across the transition region .
these signatures indicate that it is a shock between the upstream fast electron jet and the downstream magnetized electron flow .
the discontinuity separates an unmagnetized upstream flow and the magnetized downstream flow .
ions are insensitive to the shock .
they are not magnetized on both sides , and the length of the transition region is only one @xmath142 , which is smaller than the ion inertial length based on the local density . in order to distinguish this from full mhd shocks , we hereafter refer to it as an `` electron shock . ''
as time goes on , the electron shock moves outward .
it reaches at @xmath143 at @xmath144 .
after that , the shock propagates backward .
this is the influence of our periodic boundary condition . after a sufficiently long time
, the back pressure from the downstream region is strong enough to push the shock front backward , like a reverse shock reflected by a wall . the shock becomes stronger than the earlier phase , as seen in @xmath134 in figure [ fig : eshock](a ) .
shown in figures [ fig : eshock](b)[fig : eshock](e ) are physical quantities near the electron shock region .
interestingly , @xmath126 changes its sign in the shock transition region of @xmath145 : @xmath146 on the upper side and @xmath147 on the lower side [ fig .
[ fig : eshock](b ) ] , due to a reverse electron current @xmath148 . unlike the dissipation region
, electrons travel in the + @xmath64-direction there , because the fast - traveling electrons are just trapped by @xmath134 . in the upstream region of @xmath149
, @xmath49 sharply changes its polarity near the electron jet , while it is rather uniform outside the jet [ fig .
[ fig : eshock](c ) ] . in the shock - downstream of @xmath150 , @xmath49 suddenly becomes weaker than in the upstream .
the vertical magnetic pressure @xmath151 exhibit similar profiles .
this is identical to the magnetic cavity ( figure [ fig:3d ] ) .
note that the vertical component ( @xmath134 ) is compressed across the shock .
we calculate the electron temperature [ fig .
[ fig : eshock](d ) ] by tracing the diagonal components of the pressure tensor , @xmath152 each component basically shows a similar profile .
one can see that the temperature quickly increases across the shock , due to the fast jet speed .
the energy transfer from the electromagnetic fields to electrons , @xmath137 , is presented in figure [ fig : eshock](e ) .
it is negative @xmath153 around the transition region .
the same signature of @xmath153 was reported by a previous work ( the `` electron dynamo region '' in ref . ) .
these results tell us that the bulk kinetic energy of the electron jet is transferred to the electron heat [ fig .
[ fig : eshock](d ) ] and the magnetic energy [ fig .
[ fig : eshock](e ) ] in the shock downstream . here
, the magnetic energy is stored in the compressed reconnected field @xmath134 , because both @xmath126 and @xmath49 are very weak near the neutral line @xmath57 .
the strong electron heating also explains the magnetic cavity structure .
although the electron pressure substantially increases across the shock , the surrounding tangential fields @xmath154 are not strong enough to confine electrons in the @xmath26-direction in the downstream .
the high - pressure electrons expand the structure outwards along @xmath134 in the @xmath155 directions .
the displacement of the tangential fields lead to the magnetic cavity .
starting with a work by @xcite , the structures of the electron diffusion region has been actively discussed for the past five years .
earlier investigations focused on the violation of the ideal frozen - in condition , @xmath156 .
this lead to the popular two - scale picture@xcite , in which the role of the edrs was not clearly understood . in this work ,
we have explored and reorganized our understanding of fine reconnection structures in line of ref . .
we have confirmed that the electron - frame dissipation measure characterizes the critical region at higher mass ratio of @xmath27 .
since the size of the dissipation region becomes smaller at higher mass ratio , it seems that the electron physics is responsible for the dissipation region .
the relevance to the mass ratio further needs to be investigated in order to extrapolate our results to a realistic ratio of 1836 . in sec .
[ sec : outflow ] , we found that no dissipation takes place in the electron jet region , @xmath123 .
consequently , we do nt think it is appropriate to call it the outer electron `` diffusion '' region or the outer electron `` dissipation '' region . as reconnection proceeds ,
the nondissipative electron jets are elongated , while the central dissipation region remains compact .
this is consistent with the recent consensus that reconnection remains fast @xcite . from the observational viewpoint , the contribution from the charge term in eq .
[ eq : edr2 ] is of strong interest , because the charge density @xmath5 will be difficult to probe . in run 1a , the charge term is an order - of - magnitude smaller than @xmath6 , and it is only localized near the dissipation region [ fig . [ fig:35](f ) ] . in run 1b with @xmath94
, the charge separation appears nonnegligible .
let us estimate the charge separation around the dissipation region .
we assume that the local plasma density @xmath157 is roughly uniform .
electrons are magnetized to the field lines outside the electron current layer , and the electron out - of - plane speed is on an order of the electron alfvn speed @xmath158 .
the hall electric field @xmath58 is approximated by @xmath159 gauss s law in the @xmath26 direction tells us the charge density in the center , @xmath160 where @xmath161 is the typical thickness of the dissipation region . using the result in sec .
[ sec : dr ] , we assume @xmath162 where the index @xmath163 is between @xmath164 . substituting this to eq .
[ eq : gauss ] , we obtain @xmath165 we see that the charge separation is controlled by the @xmath34 parameter and that it is much less sensitive to the mass ratio . in the magnetotail ,
the magnetic field is @xmath166 nt in the lobe ( upstream region ) .
the corresponding electron gyro frequency is @xmath167 hz . the plasma sheet density @xmath168 gives @xmath169 khz .
thus we obtain a typical value of @xmath170 .
in addition , since the upstream plasma occupies the reconnection region , it is useful to compare the upstream conditions .
while our numerical model assumes that the background density is @xmath171 , the lobe density is substantially lower , @xmath172 . in the case of @xmath173 , since an upstream plasma frequency is modified by a factor of @xmath174 , our model with @xmath175 will be relevant .
these estimates suggest that the charge separation effects will be marginally negligible in the magnetotail .
this needs to be verified by further numerical investigations with realistic lobe densities@xcite .
we have found that the fast electron jet terminates at the shock , where the electrons become magnetized .
the magnetic cavity structure can be understood as its consequence .
we expect that shock - like signatures are more prominent at higher mass ratio , because the electron jet speed @xmath176 is much faster than the typical outflow speed @xmath177 . in a larger system or an open - boundary system@xcite
, we expect that the electron shock travels further downstream , unless ( 1 ) an obstacle in the downstream such as a secondary island , ( 2 ) instabilities of the jet , and ( 3 ) 3d effects interrupt the shock propagation .
the electron shock may play a role on particle acceleration as well .
for example , @xcite discussed a two - step scenario of electron acceleration : pre - acceleration near the reconnection site and energetization by @xmath178*b*/curvature drifts near the flux the flux pile - up region .
our results suggest that a significant shock - heating ( and possibly shock - drift type acceleration ) takes place between the two acceleration sites .
the electron shock deserves further investigation in the context of particle acceleration .
we summarize our understanding in figure [ fig : summary ] . as well known , a hall reconnection features the quadrupole out - of - plane magnetic field @xmath49 , which is generated by the hall current circuit @xcite .
there is a narrow channel of @xmath71 near the neutral plane @xcite .
we call this channel the electron current layer , because the electric current @xmath56 is distinctly stronger than in the other regions .
this current layer is beyond the scope of the fluid theory . inside the electron current layer
, there is a single dissipation region surrounding the reconnection point @xcite .
this region is thinner and longer than the conventional inner region .
the fast , narrow electron jet @xcite travels from the dissipation region .
this is a projection of the diamagnetic current@xcite and a part of the hall current circuit@xcite . as discussed in section [ sec : outflow ] , this region is not dissipative .
there is a pedestal region@xcite outside the electron jet .
the electron jet is terminated by an electron shock and the magnetic cavity develops there .
the upcoming mms mission will measure the field and plasma properties at high resolutions in near - earth reconnection sites . due to the limited bandwidth between the satellite and the earth
, it is very important to select the data in high - priority regions .
we have demonstrated that the electron - frame dissipation measure is a good marker of the most important dissipation region .
we have further updated our understanding on fine structures surrounding the dissipation region .
we hope that the measure is helpful to maximize the scientific profit of the mission .
the authors acknowledge keizo fujimoto and nicolas aunai for useful comments .
one of the authors ( s.z . ) acknowledges support from jsps postdoctoral fellowships for research abroad .
this work was supported by nasa s mms mission .
j. birn , j. f. drake , m. a. shay , b. n. rogers , r. e. denton , m. hesse , m. kuznetsova , z. w. ma , a. bhattacharjee , a. otto , and p. l. pritchett , , 3715 , doi:10.1029/1999ja900449 ( 2001 ) . m. a. shay , j. f. drake , b. n. rogers , and r. e. denton , , 3759 , doi:10.1029/1999ja001007 ( 2001 ) . j. f. drake and m. a. shay , in `` reconnection of magnetic fields : magnetohydrodynamics and collisionless theory and observations , '' edited by j. birn , e. r. priest ( cambridge university press , cambridge , 2007 ) , sec . 3.1 .
w. daughton , j. scudder , and h. karimabadi , , 072101 ( 2006 ) .
k. fujimoto , , 072904 ( 2006 ) .
h. karimabadi , w. daughton , j. scudder , , l13104 , doi:10.1029/2007gl030306 ( 2007 ) . m. a. shay , j. f. drake , and m. swisdak , , 155002 ( 2007 ) .
p. l. pritchett , , 3783 , doi:10.1029/1999ja001006 ( 2001 ) .
t. d. phan , j. f. drake , m. a. shay , f. s. mozer , and j. p. eastwood , , 255002 ( 2007 ) .
t. nagai , i. shinohara , m. fujimoto , a. matsuoka , y. saito , and t. mukai , , a04222 , doi:10.1029/2010ja016283 ( 2011 ) .
j. f. drake , m. a. shay , and m. swisdak , , 042306 ( 2008 ) .
a. klimas , m. hesse , and s. zenitani , , 082102 ( 2008 ) .
w. wan and g. lapenta , , 102302 ( 2008 ) .
a. klimas , m. hesse , s. zenitani , and m. kuznetsova , , 112904 ( 2010 ) .
m. hesse , s. zenitani , and a. klimas , , 112102 ( 2008 ) .
s. zenitani , m. hesse , a. klimas , and m. kuznetsova , , 195003 ( 2011 ) .
b. u. . sonnerup , `` magnetic field reconnection , '' in _ solar system plasma physics _ , edited by l. j. lanzerotti , c. f. kennel , and e. n. parker ( north - holland , new york , 1979 ) , vol .
t. terasawa , , 475 , doi:10.1029/gl010i006p00475 ( 1983 ) .
m. a. shay , j. f. drake , r. e. denton , and d. biskamp , , 9165 , doi : 10.1029/97ja03528 ( 1998 ) .
k. arzner , and m. scholer , , 3827 , doi:10.1029/2000ja000179 ( 2001 ) .
p. l. pritchett , , 25961 , doi:10.1029/2001ja000016 ( 2001 ) .
m. hoshino , t. mukai , t. terasawa , and i. shinohara , , 25979 , doi:10.1029/2001ja900052 ( 2001 ) .
a. ishizawa , r. horiuchi , h. ohtani , , 3579 ( 2004 ) .
chen , w. s. daughton , b. lefebvre , and r. b. torbert , , 012904 ( 2011 ) .
d. biskamp and k. schindler , _ plasma phys .
_ * 13 * , 1013 ( 1971 ) .
m. i. sitnov , m. swisdak , and a. v. divin , , a04202 , doi:10.1029/2008ja013980 ( 2009 ) . | it was recently proposed that the electron - frame dissipation measure , the energy transfer from the electromagnetic field to plasmas in the electron s rest frame , identifies the dissipation region of collisionless magnetic reconnection [ zenitani _ et al .
_ phys .
rev .
lett . * 106 * , 195003 ( 2011 ) ] . the measure is further applied to the electron - scale structures of antiparallel reconnection , by using two - dimensional particle - in - cell simulations .
the size of the central dissipation region is controlled by the electron - ion mass ratio , suggesting that electron physics is essential .
a narrow electron jet extends along the outflow direction until it reaches an electron shock .
the jet region appears to be anti - dissipative . at the shock ,
electron heating is relevant to a magnetic cavity signature .
the results are summarized to a unified picture of the single dissipation region in a hall magnetic geometry . |
the existence of anapole moment was theoretically predicted by zeldovich fifty years ago @xcite .
the anapole moment arises in the system where currents or magnetic moments of particles create a ring - like closed distribution of magnetic field lines .
a classic example of such distribution can be a magnetic field inside the toroidal solenoid over the surface of which poloidal currents flow .
a less trivial example is an atomic nucleus in which spins of nucleons compose a ring with forming the corresponding configuration of the magnetic field .
the great interest to the anapole is related to fundamental investigations on the parity nonconservation in atomic - nuclear interactions @xcite . in 1997
the anapole moment of the nucleus @xmath3cs was discovered just in the experiment on parity nonconservation in the atomic transition @xmath4 of cs @xcite .
now the anapole is regarded as a dipole moment of the fundamental family of toroid moments @xcite which naturally arise when currents on a torus surface are described .
toroid moments are now used in problems of current parametrization and of radiation theory @xcite , in investigations of topological effects in quantum mechanics ( the aharonov - bohm effect @xcite ) , in elementary particle physics ( toroid dipole momentum of neutrino @xcite and @xmath5-scattering @xcite for example ) , in the theory of anomalous internal conversion @xcite , in atomic physics ( anapole moments of atoms @xcite , atomic emission in a condensed medium @xcite ) , in the physics of nanostructures ( carbon toroids @xcite ) , in the physics of spin - ordered crystals @xcite etc .
thus the physics of toroid moments is a wide area of scientific activity ranging from elementary particles to condensed matter . at the same time , the nucleus together with the atomic shell remain one of the most suitable quantum object for probing the toroid moments .
the reason is as follows .
the electronic wave functions for @xmath1 and @xmath6 states have large amplitudes at the origin . as a consequence
, the electronic current at such states effectively penetrates the nucleus and interacts with the toroid moment , whose magnetic field does not exceed the boundary of the nucleus .
( this interaction mechanism is a direct analogy of a thought experiment offered in ref .
@xcite for a toroidal solenoid immersed into the electrolyte . ) today the situation is .
the static toroid dipole moment has discovered @xcite . as regards the toroid dipole moment of transition ,
there are evidences @xcite that this moment results in known anomalies in the internal conversion coefficients for @xmath7 transitions in nuclei @xcite and in the process of nuclear excitation by electron transition in the atomic shell @xcite .
thus , more detailed and precise `` conversion '' experiments are needed to detect the toroid dipole moment of transition . in the present work we consider , on the example of the nucleus
atomic shell system , two new effects dealt with the toroid moments .
first of all , we prove that the anomalies mentioned above for @xmath0 transitions result from a `` cyclo - toroid moment '' .
this moment arises when the toroid dipole moments are arranged in a ring .
secondary , we demonstrate a possibility of experimental investigation of the existence of the cyclo - toroid moment by measuring of the energy shift in atomic @xmath1 levels
. the results obtained here would be useful for more fundamental understanding of the nontrivial distribution of the currents inside nucleus and nucleons , and for the processes of the interaction between nucleus and atomic shell .
anomalous internal conversion of @xmath8-rays in deformed nuclei arises because of the penetration of the electron current @xmath9 into the nucleus `` beneath '' the nuclear current @xmath10 , i.e. when the current coordinates satisfy the condition @xmath11 .
the reason which changes the probability of the mentioned atomic - nuclear processes originates from to the additional selection rules @xcite .
nuclear and atomic matrix elements which determine the process of de - excitation ( excitation ) of deformed nuclei through the atomic shell for @xmath11 are different from those for @xmath12 . in some cases
the additional selection rules allow the nuclear transition for @xmath11 and forbid a @xmath8-radiative transition @xcite with nuclear matrix element @xmath13 entering the probability formula for @xmath12 .
( here and further @xmath14 are the standard electric / magnetic multipoles according to the gauge of @xmath15 ( see in ref .
@xcite ) , and @xmath16 are current densities .
we call @xmath16 `` currents '' for short . ) at the beginning of 80s it was found out that the real cause of the mentioned anomaly in the processes of an electric - dipole ( @xmath7 ) interaction of the nucleus with the atomic shell results from the dipole toroid moment of the transition @xcite .
toroid moments are the components of electric moments @xcite .
moreover , the anomaly also exists in conversion magnetic - dipole ( @xmath0 ) transitions @xcite .
let us consider the nature of such @xmath0 anomaly .
the nucleus matrix element corresponding to the anomalous part of the magnetic - dipole process of the nuclear interaction with the atomic shell is as follows @xcite @xmath17 where @xmath18 @xcite , @xmath19 is the energy of a nuclear transition , @xmath20 , @xmath21 are the spherical bessel functions , @xmath22 are the vector spherical harmonics from ref .
@xcite , @xmath23 is the nucleus radius .
we use in this paper the system of units @xmath24 .
the operator structure of @xmath25 is well known . if the nuclear transition current in eq ( [ eq :
n - an ] ) is the sum of convection and spin terms @xmath26 , where @xmath27 and @xmath28 ( we consider one - nucleon current for simplicity ) then it is easy to get the following expression for the anomalous nuclear matrix element @xcite @xmath29 \right ) \varphi_i(r ) .
\label{eq : operat - an}\end{aligned}\ ] ] in formula ( [ eq : operat - an ] ) and in the expression for the transition current the following standard symbols were used : @xmath30 $ ] is the operator of the angular momentum , @xmath31 are the spherical functions @xcite , @xmath32 , @xmath33 and @xmath34 are the charge , the mass , and the wave functions of the nucleon , accordingly , @xmath35 is the operator of the magnetic moment of the nucleon : @xmath36 , where @xmath37 are the pauli spin matrices , @xmath38 is the nuclear magneton , @xmath39 is the empirical magnetic moment of the nucleon @xmath40 for the proton and -1.91 for the neutron .
the spin term of the current @xmath41 in the matrix element @xmath25 in eq ( [ eq : n - an ] ) corresponds to the operator in eq ( [ eq : operat - an ] ) in square brackets under the integral @xmath42 $ ] ( here @xmath43 is the density of the magnetic moment @xmath44 ) .
let us compare this operator with a known density operator of the dipole toroid moment @xmath45/10 $ ] @xcite , that is created by the convection current @xmath46 .
operators in square brackets have different parity and are transformed into each other by exchanging @xmath46 and @xmath43 . taking into account
this fact let us introduce the new operator @xmath47 \label{eq : ct-1}\ ] ] and find out what geometrical pattern corresponds to @xmath48 .
using the spherical functions it is easy to rewrite eq ( [ eq : ct-1 ] ) in the form : @xmath49 then let us remember that the following objects possess the toroid moment : ( a ) a toroidal solenoid with a current @xmath46 ; ( b ) a ring formed by magnetic moments @xmath43 @xcite . for case ( b )
the following relationship is true : @xmath50 , besides that @xmath51 due to helmholtz theorem @xcite . substituting @xmath52 into eq ( [ eq:<ct > ] ) instead of @xmath43
, we get the following relation @xmath53 .
a sub - integral expression can be easily transformed to the vector production ( see for example , in ref .
@xcite ) . as a result
, the operator @xmath48 can be written in as follows : @xmath54 $ ] , in full accordance with similar expressions for the density of a current magnetic moment @xmath55 $ ] and for the toroid moment formed by the ring - like composed magnetic dipoles @xmath56 $ ] @xcite .
expression @xmath57 $ ] results in @xmath58 .
the geometrical image corresponding to the operator @xmath48 , is the ring - like composed dipole toroid moments @xmath59 .
therefore @xmath48 can be called the density of cyclo - toroid moment , and @xmath60 the cyclo - toroid moment .
the properties of a cyclo - toroid are shown by the torus whose surface is formed by either ring - like composed magnetic moments @xmath43 , or by poloidal lines of the magnetic field ( fig .
[ fig1]a ) .
it is also possible to consider the cyclo - toroid as a sequence of toroidal solenoids all together forming again a toroidal solenoid ( such objects were considered for the first time in ref .
@xcite ) . in this case
the equal currents flow in opposite directions on the surface of two toruses ( one of them is embedded into another ) as it is shown in fig .
[ fig1]b .
zeldovich s anapole @xcite in the static case interacts with the external current @xmath61 only .
the energy of this interaction is @xmath62 .
taking also into account the equality @xmath63 , we arrive straightforwardly to the hamiltonian of the interaction of the cyclo - toroid with the external current : @xmath64 that is , the cyclo - toroid interacts locally ( only inside the nucleus ) with the vortex part of penetrating external current .
let us determine now the role of a cyclo - toroid in the anomalous conversion .
consider the internal electron conversion from one of the @xmath1 states of an atom and show first of all that @xmath65 during the @xmath0 transition of an electron to the @xmath66 state of the continuous spectrum .
we write the electron current in a common way @xmath67 , where @xmath68 , @xmath69 and @xmath70 are the dirac matrixes @xcite , @xmath71 is the proton charge , @xmath72 are the electron wave functions .
let us write these wave functions in the form of @xmath73 , where @xmath74 and @xmath75 are correspondingly the large and the small components of radial wave functions , @xmath76 are the spinor spherical harmonics @xcite . for the bound states in the atom
the normalization condition is as follow : @xmath77 .
the wave function of the final state is considered as a superposition of the plane wave and the converging spherical wave . for the states with @xmath78 , @xmath79 and @xmath80
the angular part @xmath81 is very simple , and the components of the bispinor can be written in the following way : @xmath82 , @xmath83 and @xmath84 . within the spherical coordinate system the current is as follows @xmath85 for @xmath86 correspondingly .
we can see that the existence of the @xmath87 transition electron current itself is a totally relativistic effect .
the current disappears ( as it should be ) when the small parts of the dirac wave functions @xmath75 tend to zero .
the interaction of the current with the cyclo - toroid does not occur beyond the nucleus range .
therefore , for the radial components of the @xmath88 states one can use the expressions which are true for small values of arguments : @xmath89 , @xmath90 , and @xmath91 , @xmath92 , where @xmath93 is the bohr radius , @xmath94 and @xmath95 are the coefficients of decomposition .
( here the potential inside the nucleus corresponds to the nuclear charge uniformly distributed over the spherical volume of radius @xmath23 @xcite . ) using the expansion , we can get from eq ( [ eq : j ] ) for @xmath96 the following @xmath97 where the @xmath96 vector components are given in the polar system @xcite and correspond to transitions @xmath98 and @xmath99 accordingly . since @xmath100 and @xmath101 for @xmath102 states , then @xmath103 . as an example we consider the known anomaly in the conversion @xmath0 transition with the energy of @xmath104 kev in @xmath105ta nucleus .
the experimental magnitude of the @xmath106 shell conversion coefficient is 3 times as great as the so - called normal theoretical magnitude @xcite .
the magnitude of the penetration parameter @xmath107 equals to 150 @xcite .
the transition occurs between the states with quantum numbers in the nilsson model being as follows : @xmath108\sigma = 5/2^+5/2[402]\uparrow$ ] and @xmath109\downarrow$ ] @xcite .
the radiation nuclear matrix element of this transition is forbidden by asymptotic quantum numbers , while the anomalous one is allowed @xcite .
however , the selection rules for the operator @xmath110 in eq ( [ eq : operat - an ] ) forbid the anomaly nuclear transition in @xmath105ta @xcite .
therefore we can infer that the anomaly results just from the interaction of the transition cyclo - toroid moment with the vortex part of the penetrating electron current . in order to determine @xmath60
let us use the standard view for the radiation nuclear matrix element @xcite : @xmath111 , where @xmath112 are the klebsch - gordan coefficients .
the reduced matrix element @xmath113 in the nuclear spectroscopy is usually related to the experimentally measured reduced probability of the nuclear transition latexmath:[$b(e
/ ml;j_i\rightarrow{}j_f ) = reduced probability of the @xmath0 transition in the single - particle weisskopf model is @xmath115 @xcite . as a rule ,
the value of the experimental probability is given in weisskopf units @xmath116 .
for the considered transition in @xmath105ta it is @xmath117 @xcite . let us substitute operator in square brackets in eq ( [ eq : operat - an ] ) to operator @xmath118 and then , using the definitions for @xmath119 and @xmath120 , derive the expression for the reduced matrix element of the cyclo - toroid moment of the transition through the parameters measured in the experiment : @xmath121 we introduced here @xmath122 as the `` unit '' for measurement of the cyclo - toroid moment .
for @xmath105ta nucleus the following is true : @xmath123 @xmath124/gev if the radius of a nucleus with atomic number @xmath125 is @xmath126 fm . using eq ( [ eq : ct - reduced ] ) one can estimate the value @xmath127 for the anomalous conversion @xmath0(482 kev ) transition : @xmath128 .
the cyclo - toroid moment is a pseudovector . atomic nuclei ,
like some other objects , for example , nucleons , can have a static cyclo - toroid moment in the ground state .
the value of a @xmath48 for nucleus should not depend on the fermi coupling constant @xmath129 of the weak interaction , unlike the static toroid moment ( the latter as a rule arises due to the parity violation for the nuclear forces and contains @xmath129 @xcite ) .
another positive aspect is the geometric factor similar to the one considered for toroid moments in ref .
the cyclo - toroid moment @xmath60 is proportional to the nucleus volume , that is , to the atomic number @xmath125 . as for the interaction of the cyclo - toroid moment with an atom electron
, this interaction is of the electromagnetic nature and , therefore , is proportional to the fine structure constant @xmath130 . besides
, the additional smallness of @xmath131 is produced by the small component of the dirac wave function @xmath75 .
the interaction of the atomic @xmath132-electron with the cyclo - toroid moment of a nucleus results in the contribution to the atomic level energy already in the first order of the perturbation theory . to demonstrate this , let us estimate the shift of the @xmath133 level in @xmath3cs .
the dependence of the energy on the electron spin projection @xmath134 and on the value of the cyclo - toroid moment @xmath135 ( here and further @xmath135 is measured in @xmath136 units ) can be found from the following formula : @xmath137 that can be obtained from eqs ( [ eq : hint ] ) and ( [ eq : curlj ] ) , if one uses the @xmath138 wave function for the initial and the final state simultaneously , and takes into account the relation @xmath139 , where @xmath140 is the electron mass .
we have @xmath141 ev when @xmath142 , @xmath143 .
hydrogen - like ions in the @xmath2 state are perspective objects for such investigation .
we give here the qualitative estimations for the h - like ions of two heavy nuclei : @xmath144cs@xmath145 ev and @xmath146u@xmath147 ev .
the proton constituted of quarks can also possess a cyclo - toroid moment . in this case
the shift of @xmath148 level of a hydrogen atom as a function of @xmath135 is @xmath149 ev ( the calculation was performed according to formula ( [ eq : deltae ] ) with the following parameters : @xmath150 , @xmath151 if @xmath152 fm , @xmath153 ) . it should be noted that the magnitude of @xmath154 for @xmath155 is at present close to the accuracy of the measurement of the lamb shift for the @xmath148 level of hydrogen @xcite . in conclusion
it is important to note the following . the system depicted in fig .
[ fig1 ] has no charges .
therefore , for all the charge moments there is @xmath156 ( where @xmath157 is the charge density ) .
for all the magnetic moments of the system there is also @xmath158 , since all the elementary magnetic moments are ring - like composed and can be written in the following way : @xmath159 , and @xmath160 . the same is also true for the toroid moment @xmath161 @xcite , because for the object in fig .
[ fig1 ] there is @xmath162 .
thus , the system has no toroid moments . in spite of the fact that the mentioned electromagnetic moments are equal to zero
, the interaction energy of the system with the external current has a nonzero value .
so , the cyclo - toroid moment can be the single non - vanishing moment of the quantum object .
during recent years the author discussed some problems considered in sec . 2 of this paper with dr .
m.a . listengarten . without his benevolent interest , constructive advices and persistent
wish to see work finished , this paper would be hardly written .
this work was partially supported by the istc , project n2651 , and by the leading science school grant n2078.2003.2 . | we offer the hypothesis that atomic nuclei , nucleons , and atoms possess a new type of electromagnetic moment , that we call a `` cyclo - toroid moment '' . in nuclei , this moment arises when the toroid dipole ( anapole ) moments are arrayed in the form of a ring , or , equivalently , when the magnetic moments of the nucleons are arranged in the form of rings which , in turn , constitute the surface of a torus .
we establish theoretically that the cyclo - toroid moment plays a role in the processes of the atomic shell nucleus interaction .
the existence of this moment would explain known anomalies in the internal conversion coefficients for @xmath0 transitions in nuclei .
we show also that the static cyclo - toroid nuclear moment interacts locally inside the nucleus with the vortex part of the atomic electron currents and this leads to an energy shift in atomic @xmath1 states . for the hydrogen atom
the value of this shift may be comparable in order of magnitude to the present accuracy of measurements of the lamb shift for the @xmath2 level . |
clusters of galaxies are efficient tracers of the large - scale structure of the universe . a strong spatial correlation of clusters of galaxies ( bahcall & soneira 1983 ; klypin & kopylov 1983 ) provided some of the first evidence for the existence of organized structure on large scales . to date , much effort has been devoted to determine the correlation function and the power spectrum of clusters of galaxies on large spatial scales ( e.g. postman , huchra & geller 1992 ; peacock & west 1992 ; einasto et al .
1993 ; dalton et al . 1994 ; romer et al . 1994 ; croft et al . 1997 ; einasto et al .
1997a ; retzlaff et al . 1998 ; tadros , efstathiou & dalton 1998 ) .
figure 1 shows the power spectrum of the spatial distribution of the abell - aco clusters as determined by einasto et al .
( 1999 ) , and the power spectrum of the apm clusters as found by tadros , efstathiou & dalton ( 1998 ) . for comparison ,
we show in figure 1 the power spectrum derived from the distribution of galaxies in the apm survey ( baugh & efstathiou 1993 ) .
observations give the power spectrum of clusters for a given number density of clusters , @xmath2 .
the number density of the apm clusters and abell clusters is @xmath3 mpc@xmath4 and @xmath5 mpc@xmath4 , respectively ( dalton et al .
1994 ; einasto et al . 1997b
; retzlaff et al . 1998 ) .
as the observational data on the power spectrum of clusters improve , the need for precise theoretical predictions becomes increasingly important .
the power spectrum of clusters for different cosmological models can be calculated using n - body simulations ( e.g. retzlaff et al .
1998 ; tadros , efstathiou & dalton 1998 ; suhhonenko & gramann 1999 ) . however , this approach is difficult as it requires simulations with a very large dynamical range to identify correctly a sufficient number of clusters .
in this situation it would be useful to have an analytical approach to describe the origin of the power spectrum of clusters of galaxies and to predict it .
the correlation function of clusters of galaxies has been a subject of many attempts at analytical modelling ( e.g. kaiser 1984 ; kashlinsky 1987 ; cole & kaiser 1989 ; mann , heavens , & peacock 1993 ; mo & white 1996 , hereafter mw96 ; catelan et al . 1998 ; kashlinsky 1998 ) . in particular
, using the press - schechter ( ps ) formalism to calculate the correlation function of clusters in lagrangian space and mapping from lagrangian space to eulerian space within the context of the spherical collapse model , mw96 have derived an analytical expression for the correlation function of clusters of mass @xmath6 as @xmath7 where @xmath8 is the correlation function of matter density fluctuations .
the bias parameter @xmath9 is @xmath10 where @xmath11 is the density dispersion on a mass scale @xmath6 and @xmath12 .
this equation has been found to be in good agreement with n - body results by mw96 and by mo et al .
( 1996 ) .
the power spectrum and the correlation function of clusters form a fourier transform pair . therefore , in the linear approximation the power spectrum of clusters of mass @xmath6 can be expressed as @xmath13 where @xmath14 is the power spectrum of matter density fluctuations . in this paper
we develop further the results obtained by mw96 and calculate the power spectrum and the correlation function of galaxy clusters for a given number density of clusters @xmath2 in lagrangian space and in eulerian space ( eq .
[ 15,23 - 24 ] and [ 25 - 27 ] below ) .
observations provide the distribution of clusters in redshift space , which is distorted due to peculiar velocities of clusters . in order to study the power spectrum of clusters in redshift space
, we use the linear approximation derived by kaiser ( 1987 ) . in
3 we examine the cluster power spectrum in real space and in redshift space using n - body simulations with realistic initial power spectra and show that in both cases the ps theory gives an accurate description of the power spectrum of clusters at wavenumbers @xmath15 mpc@xmath16 ( or at wavelengths @xmath0 mpc ) . in
4 we use our model to study the redshift - space power spectrum of clusters in different cold dark matter models .
we study inflationary models with a scale - invariant initial spectrum of density fluctuations and with a steplike initial spectrum derived by starobinsky ( 1992 ) .
5 summarizes the main results .
a hubble constant of @xmath17 km s@xmath16 mpc@xmath16 is used throughout this paper .
an important aspect of the ps model is that , being entirely based on linear theory , suitably extrapolated to the collapse time of spherical perturbations , it is , by definition , local in lagrangian space .
while this lagrangian aspect of the theory does not have immediate implications for the study of the mass function of clusters of galaxies , it is important for the study of their spatial clustering properties .
let us assume that the density contrast @xmath18 on a mass scale @xmath6 is a gaussian random field . here
@xmath19 represents the comoving lagrangian coordinate .
the probability that one would find a density contrast between @xmath20 and @xmath21 is @xmath22 \,\ ,
d\epsilon_m \ , .
\eqno ( 4)\ ] ] the density dispersion @xmath11 can be written as @xmath23 where @xmath24 is the fourier transform of the window function applied to determine the density field . in this paper
we will use the top - hat window function . for the top - hat window ,
the mass @xmath6 is related to the window radius @xmath25 as @xmath26 ( where @xmath27 is the mean background density ) .
press & schechter ( 1974 ) suggested that we can determine the probability that a cluster of mass @xmath6 has formed as @xmath28 \
, , \eqno(6)\ ] ] where @xmath29 is a certain threshold value .
ps used a value @xmath12 , which is motivated by the spherical collapse model .
the number density of clusters with the mass between @xmath6 and @xmath30 is given by ( press & schechter 1974 ) @xmath31 \ , dm \ , .
\eqno(7)\ ] ] equation ( 7 ) involves an additional correction factor 2 to allow all the matter in the universe to form bound structures ( see peacock & heavens 1990 ; bond et al .
1991 for a more detailed discussion of this correction ) .
therefore , the number density of clusters of mass larger than @xmath6 can be expressed as @xmath32 \ , \,\ , { dr \over r^3 } \ , .
\eqno(8)\ ] ] equation ( 8) has been frequently used to determine the cluster abundance in different cosmological models ( e.g. efstathiou et al .
1988 ; white , efstathiou & frenk 1993 ; lacey & cole 1994 ; eke , cole & frenk 1996 ; borgani et al .
1997 ) . for a two - dimensional gaussian distribution the probability density for finding simultaneously @xmath33 on scale @xmath34 and @xmath35 on scale @xmath36 separated by @xmath37 is given by @xmath38 } , \eqno(9)\ ] ] where @xmath39 , @xmath40 and @xmath41 is the correlation coefficient between @xmath33 and @xmath35 .
the @xmath42 is the two - point correlation function of the linear density contrast smoothed on the scales @xmath34 and @xmath36 : @xmath43 where @xmath44 and @xmath45 is the spherical bessel function of the order zero .
similarly to equation ( 6 ) , we can write for the probability for two clusters of mass @xmath34 , @xmath36 to form as @xmath46 to determine the correlation function of clusters of masses [ @xmath34;@xmath47 and [ @xmath36;@xmath48 we can use an expression @xmath49 a similar equation was used by kashlinky ( 1987 , 1998 ) to derive the correlation function of clusters in eulerian space and by catelan et al .
( 1998 ) to describe clustering in lagrangian space . in order to find the correlation function of clusters of mass larger than @xmath6
, we start from approximations ( 6 - 7 ) and ( 11 - 12 ) and write down the probability to find two clusters in small volumes @xmath50 , @xmath51 , one cluster being in the mass range [ @xmath34 , @xmath47 , the other in [ @xmath36 , @xmath48 , as @xmath52 we included in equation ( 13 ) a correction factor 4 to make it consistent with equation ( 7 ) at @xmath53 . the probability to find two clusters of mass larger than @xmath6
can be expressed as @xmath54 \
, dv_1 dv_2 , \eqno ( 14)\ ] ] where @xmath2 is the number density of clusters of mass larger than @xmath6 ( equation 8) .
therefore , the correlation function of clusters can be written in the form @xmath55 equations ( 5,8 - 11 ) and ( 15 ) determine the correlation function of clusters for a given @xmath2 in the ps theory of gravitational clustering .
let us consider the correlation function of clusters in the linear approximation , i.e. neglecting the terms involving @xmath56 in equation ( 9 ) .
in the linear approximation , the probability density @xmath57 \exp \left[- { \epsilon_{m_1}^2 \over 2\sigma^2(m_1)}\right ] \ , \exp \left[- { \epsilon_{m_2}^2 \over 2\sigma^2(m_2)}\right ] .
\eqno(16)\ ] ] substituting approximation ( 16 ) into equation ( 11 ) we find that @xmath58 \ , \exp \left[- { \delta_t^2 \over 2\sigma^2(m_2)}\right ] \eqno ( 17)\ ] ] and @xmath59 .
\eqno(18)\ ] ] for separations @xmath37 much larger that the smoothing lengths , @xmath60 and @xmath61 , the influence of the window functions on the correlations is negligible , and the correlation function @xmath62 ( where @xmath63 is the linear mass autocorrelation function ) . therefore , in the linear approximation the correlation function of clusters of masses [ @xmath34;@xmath47 and [ @xmath36;@xmath48 can be written as @xmath64 where the lagrangian bias factor @xmath65 is @xmath66 mw96 derived this equation to describe clustering in lagrangian space starting from the behaviour of the conditional lagrangian mass function derived by bond et al .
( 1991 ) . by using the spherical collapse model
, mw96 found that in eulerian space the correlation function of clusters of masses [ @xmath34;@xmath47 and [ @xmath36;@xmath48 can be written as @xmath67 were @xmath8 is the correlation function of matter density fluctuations in eulerian space and the eulerian bias factor @xmath68 is @xmath69 catelan et al .
( 1998 ) used a more general method for evolving the spatial distribution of clusters and found a similar result for the linear approximation .
the shift by 1 of the linear bias factor , caused here by the transformation from the lagrangian to the eulerian world , comes from mass conservation in eulerian space .
let us consider the correlation function of clusters for a given number density in the linear ps approximation .
substituting approximation ( 18 ) into equation ( 15 ) we find that in the linear approximation the lagrangian correlation function of clusters for a given @xmath2 can be expressed as @xmath70 where the bias parameter @xmath71 can be written in the form @xmath72
\exp\left[-{\delta_t^2 \over 2\sigma^2(r)}\right ] { dr \over r^3 } .
\eqno(24)\ ] ] similarly to equations ( 21 - 22 ) the eulerian correlation function of clusters for a given number density @xmath2 can be written as @xmath73 where the bias parameter @xmath74 is @xmath75 \exp\left[-{\delta_t^2 \over 2\sigma^2(r)}\right ] { dr \over r^3 } .
\eqno(26)\ ] ] the power spectrum of clusters for a given number density in the linear approximation can be expressed as @xmath76 the cluster bias parameter @xmath74 depends on the minimal mass @xmath6 ( or the window radius @xmath25 ) of clusters and on the power spectrum of density fluctuations , @xmath14 , which determines the function @xmath77 . for a fixed @xmath14 and @xmath2 ,
the minimal mass @xmath6 ( or scale @xmath25 ) can be determined by inverting equation ( 8) . in this approach the power spectrum ( or the correlation function ) of clusters
does not depend on the mean background density @xmath27 ( or the density parameter @xmath78 ) .
observations provide the distribution of clusters in redshift space , which is distorted due to peculiar velocities of clusters . on large scales , where linear theory applies
, the power spectrum of matter density fluctuations in redshift space is given by ( kaiser 1987 ) : @xmath79 p(k ) , \eqno(28)\ ] ] where @xmath80 is the linear velocity growth factor .
in the linear approximation ( 27 ) , relation ( 28 ) takes the form @xmath81 b_{cl}^2 \ , p(k ) .
\eqno(29)\ ] ] equation ( 29 ) determines the power spectrum of clusters for a given @xmath2 in redshift space .
for testing the solutions ( 26 - 27 ) and ( 29 ) , we ran n - body simulations , using a particle - mesh code described by gramann ( 1988 ) .
we investigated the evolution of @xmath82 particles on a @xmath82 grid , with @xmath83 .
the comoving box size was @xmath84 mpc .
the initial density field was taken to be gaussian .
we examined the distribution of clusters in two cosmological models which start from the observed power spectra of the distribution of galaxies and clusters of galaxies . in the model ( 1 ) , the initial linear power spectrum of density fluctuations was chosen in the form @xmath85 at wavelengths @xmath86 mpc . in the model ( 2 ) , we assumed that the initial power spectrum contains a primordial feature at the wavelengths @xmath87 mpc that correspond to the scale of superclusters of galaxies .
suhhonenko & gramann ( 1998 ) studied the mass function , peculiar velocities , the power spectrum and the correlation function of clusters in both models for different values of the density parameter @xmath78 and @xmath88 ( the rms fluctuation on the @xmath89 mpc scale ) .
the results were compared with observations .
they found that in many aspects the initial power spectrum of density fluctuations in the model ( 2 ) fits the observed data better than the simple power law model ( 1 ) . clusters were selected in the simulations as maxima of the density field that was determined on a @xmath82 grid using the cic - scheme .
to determine peculiar velocities of clusters , we determined the peculiar velocity field on a @xmath82 grid using the cic - scheme and found the peculiar velocties at the grid points were clusters had been identified .
the clusters were then ranked according to their density and we selected @xmath90 highest ranked clusters to produce cluster catalogs with a mean intercluster separation @xmath91mpc .
for comparison , the mean separation of the observed apm and abell clusters is @xmath92 and @xmath93 mpc , respectively ( dalton et al .
1994 ; retzlaff et al .
1998 ) .
it is difficult to follow the evolution of rich clusters by using n - body simulations , as it requires simulations with a very large dynamical range to identify correctly a sufficient number of clusters .
we determined clusters as maxima of the density field smoothed on the scale @xmath94 mpc .
this method of identifying clusters is not really identical to that what the ps theory is predicting - the location of collapsed , virialized objects - and this could affect the degree of agreement between the analytic theory and the numerical results .
however , in order to increase the resolution of simulations we must increase the number of test particles and grid points , or we have to follow the evolution of clusters in a smaller box . while the first possibility is technically difficult , in the latter case the number of rich clusters becomes too small to get statistically reliable results .
taking into account the requirements on the number of clusters and on the resolution together with the fact of fixed computer recources we decided to use a box size @xmath84 mpc and a grid size @xmath95 mpc .
figure 2 shows the power spectrum of clusters with a mean separation @xmath96 mpc in our models .
we studied also the power spectrum of clusters with a mean separation @xmath97mpc and @xmath98 mpc , and found similar results . to calculate the power spectrum of clusters in the simulations we determined the density field of clusters on a @xmath99 grid using the cic - scheme and calculated its fourier components , subtracting the shot noise term .
we determined also the poisson error bars for the power spectrum . in a shell of @xmath100-space containing @xmath101 modes ,
the poisson error can be estimated as @xmath102 ( see e.g. peacock & nicholson ( 1991 ) for a more detailed analysis of the poisson errors in the power spectrum ) .
figure 2a shows the results for the model ( 1 ) and figure 2b for the model ( 2 ) . for the model ( 1 )
, the clusters were determined in the simulation for @xmath103 and @xmath104 . for the model ( 2 ) , these parameters were @xmath103 and @xmath105 . for comparison , we examined the power spectrum of clusters in the linear ps approximation ( 27 ) . we inverted equation ( 8) to determine the minimal radius @xmath25 for a given @xmath2 ( or @xmath106 ) and after that we calculated the bias factor @xmath74 using equation ( 26 ) . for a given @xmath88 and @xmath2 , the bias parameter @xmath74 in the model ( 1 ) is similar to that in the model ( 2 ) . for @xmath103 ,
we find that @xmath107 and @xmath108 for separations @xmath109 and @xmath110mpc , respectively . for @xmath104 , @xmath111 and @xmath112 , respectively .
now we can compare the numerical results and the ps theory predictions .
first , consider the power spectrum of clusters at wavenumbers @xmath113 mpc@xmath16 ( or at wavelengths @xmath114 mpc ) . on larger scales ,
the number of modes in the simulations is too small to get statistically reliable results .
figure 2 demonstrates that at wavenumbers @xmath113 mpc@xmath16 , the power spectrum of clusters in the simulations is linearly enhanced with respect to the power spectrum of the matter distribution in both models studied . for @xmath103 ,
the agreement between the results of the n - body simulations and the ps theory predictions is very good .
the mean deviation between the numerical results and the theoretical predictions is about @xmath115% and @xmath116% in the model ( 1 ) and in the model ( 2 ) , respectively . for @xmath117 , we find that the power spectrum of clusters in n - body simulations is somewhat lower than predicted by approximation ( 27 ) .
the power spectrum of clusters is about @xmath118% and @xmath119% of the linear theory predictions in the model ( 1 ) and model ( 2 ) , respectively .
the ps approximation predicts that during the evolution between the @xmath103 and @xmath104 , @xmath120 increases slightly . in simulations we find that during the evolution the power spectrum of clusters decreases
this effect is probably caused by merging of very rich clusters .
further study ( e.g. numerical simulations with a higher dynamical range ) is needed to determine whether this is a real effect in our models , or a numerical effect due to the limited dynamical range of the n - body simulations .
let us now consider the power spectrum of clusters at wavenumbers @xmath121 mpc@xmath16 ( @xmath122 mpc ) . in the model ( 1 ) ,
the numerical results on these scales are in good agreement with the linear approximation ( 27 ) .
but in the model ( 2 ) , the power spectrum of clusters in the simulation is significantly smaller than that predicted by the linear theory . for @xmath105 ,
we find that the power spectrum of clusters is only @xmath123% of the linear theory predictions . therefore , in the model ( 2 ) , we can not use the linear approximation to study the power spectrum of clusters at wavenumbers @xmath124 mpc@xmath16 .
the power spectrum of clusters on these scales is probably determined by non - linear processes in superclusters of galaxies .
figure 3 shows the redshift - space power spectrum of clusters in the simulations and in approximation ( 29 ) . for clarity ,
we did not plot error bars in this figure . in order to study peculiar velocities of galaxy clusters and their distribution in redshift space in models with different @xmath78 , we determined peculiar velocities of clusters in the simulations with @xmath83 and assumed that peculiar velocities of clusters are proportional to the linear growth factor @xmath125 .
let us consider the redshift - space power spectrum of clusters at wavenumbers @xmath113 mpc@xmath16 , where we expect that the linear approximation ( 29 ) applies .
numerical results show that on these scales the power spectrum of clusters in redshift space is linearly enhanced with respect to the power spectrum of clusters in real space . in figure 3 we show the power spectrum of clusters in redshift space for @xmath103 and @xmath83 . in this case , the agreement between the numerical results and approximation ( 29 ) is very good .
the mean deviation between the numerical results and theoretical predictions is about @xmath126% and @xmath116% , in the model ( 1 ) and model ( 2 ) , respectively . for @xmath117
we find that , similarly to real space , the redshift - space power spectrum of clusters in the simulations is somewhat lower than that predicted by the linear theory . for @xmath83 , the power spectrum of clusters is about @xmath118% and @xmath127% of the linear theory predictions in the model ( 1 ) and model ( 2 ) , respectively .
for @xmath128 , we found similar results . within the uncertainties due to various numerical effects in the simulations
, the agreement between the analytic theory and the numerical results is good .
the degree of agreement is similar in real and in redshift space .
now we use approximation ( 29 ) to analyze the redshift - space power spectrum of clusters in different cold dark matter models .
we examine flat cosmological models with the density parameter @xmath129 , the baryonic density @xmath130 and the normalized hubble constant @xmath131 .
these parameters are in agreement with recent nucleosynthesis results ( tytler at al .
1996 ) , with measurements of the density parameter ( e.g. dekel , burstein & white 1996 ; bahcall & fan 1998 ) and with measurements of the hubble constant using various distance indicators ( e.g. tammann 1998 ) . to restore the spatial flatness in the low - density models ,
we assume a contribution from a cosmological constant : @xmath132 .
figure 4 shows the redshift - space power spectrum of clusters in the @xmath1cdm models with a scale - invariant harrison - zeldovich initial spectrum of density fluctuations ( @xmath133 ) . we have used the transfer function derived by bardeen et al .
( 1986 ) and sugiyama ( 1995 ) , and the cobe normalization derived by bunn & white ( 1997 ) . figure 4a demonstrates the power spectrum of the model clusters and the abell - aco clusters with a mean separation @xmath134 mpc .
we show the power spectrum of the abell - aco clusters as found by einasto et al .
this spectrum represents the weighted mean of the power spectra determined by einasto et al .
( 1997a ) and retzlaff et al .
einasto et al .
( 1997a ) determined the power spectrum of the abell - aco clusters from the correlation function of clusters , while retzlaff et al .
( 1998 ) estimated the power spectrum directly ( see einasto et al .
1999 for details ) . the power spectrum of the distribution of the abell - aco clusters peaks at the wavenumber @xmath135 mpc@xmath16 ( or at the wavelength @xmath136 mpc ) . for @xmath137 mpc@xmath16 ,
the cluster power spectrum is well approximated by a power law , @xmath138 , with @xmath139 .
a similar peak in the one - dimensional power spectrum of a deep pencil - beam survey was detected by broadhurst et al .
( 1990 ) and in the two - dimensional power spectrum of the las campanas redshift survey by landy et al .
figure 4b shows the power spectrum of the model clusters with a mean separation @xmath140 mpc . for comparison ,
we show the power spectrum of the observed apm clusters determined by tadros , efstathiou & dalton ( 1998 ) .
they analyzed the redshift survey of 364 clusters described by dalton et al .
the mean intercluster separation of apm clusters is @xmath92 mpc ( dalton et al .
1994 ) .
figure 4 shows that the power spectrum of clusters predicted in the @xmath1cdm models is not consistent with the observed spectra of clusters .
we studied a @xmath141 probability at wavenumbers @xmath142 mpc@xmath16 , where we expect that the linear approximation ( 29 ) applies . for the models presented in figure 4 ,
the probability to fit the observed power spectra of the abell and apm clusters is less than @xmath143 and @xmath144 , respectively .
we also studied a @xmath141 probability for the models , where the amplitude of the power spectrum of clusters is about @xmath119% and @xmath118% of the linear theory predictions and found similar results .
the power spectrum of clusters in the @xmath1cdm models with a scale - invariant initial spectrum is not consistent with the observed spectra of the apm and abell - aco clusters ( see also e.g. tadros , efstathiou & dalton 1998 ; einasto et al
. 1997a ) .
the power spectrum of density fluctuations in the universe depends on the physical processes in the early universe .
the peak in the power spectrum of clusters at wavelength @xmath145 mpc may be generated during the era of radiation domination or earlier .
baryonic acoustic oscillations in adiabatic models may explain the observed power spectrum only if currently favored determinations of cosmological parameters are in substantial error ( e.g. the density parameter @xmath146 ) ( eisenstein et al .
one possible explanation for the observed power spectra of clusters is an inflationary model with a scalar field whose potential @xmath147 has a local steplike feature in the first derivative .
this feature can be produced by fast phase transition in physical field other than an inflaton field .
an exact analytical expression for the scalar ( density ) perturbations generated in this inflationary model was found by starobinsky ( 1992 ) ( see also lesgourgues , polarski & starobinsky 1998 ) .
the initial power spectrum of density fluctuations in this model can be expressed as @xmath148 where function @xmath149 can be written in the form @xmath150 + { 9 \ , p_1 ^ 2 \ , f_2(y ) \over 2 \ ,
y^2 } \left[f_2(y)+f_1(y)\cos{2y } - { 2 \over y } \sin{2y } \right ] \ , .
\eqno(31)\ ] ] here , function @xmath151 , @xmath152 and @xmath153 .
the initial power spectrum in this model depends on two parameters @xmath154 and @xmath155 .
the parameter @xmath154 determines the location of the step and the parameter @xmath155 - the shape of the initial spectrum . for @xmath156
, we recover the scale - invariant harrison - zeldovich spectrum ( @xmath157 ) . at present , the initial spectrum ( 30 - 31 ) is probably the only example of a initial power spectrum with the desired properties , for which a closed analytical form exists .
figure 5 shows the power spectrum of clusters in the @xmath1cdm models with a steplike initial power spectrum . as in figure 4
, we have used the transfer function derived by bardeen et al .
( 1986 ) and sugiyama ( 1995 ) , and the cobe normalization derived by bunn & white ( 1997 ) .
for the models with @xmath158 and @xmath159 , the step parameter was chosen to be @xmath160 mpc@xmath16 and @xmath161 mpc@xmath16 , respectively . in this case , in the models with @xmath158 ,
the power spectrum has a well - defined maximum at the wavenumber @xmath162 mpc@xmath16 and a second maximum at @xmath163 mpc@xmath16 .
in the models with @xmath159 , the picture is inverted .
the power spectrum has a flat upper plateau at wavenumbers @xmath164 mpc@xmath16 , a sharp decrease on smaller scales ( @xmath165 mpc@xmath16 ) and a secondary maximum at @xmath166 mpc@xmath16 . at wavenumbers
@xmath167 mpc@xmath16 , the power spectrum in the @xmath1cdm models with @xmath168 is similar to the power spectrum in the numerical model ( 2 ) , that we studied in 3 ( see figure 2b and figure 3b ) .
figure 5a shows the power spectrum of the model clusters and the abell clusters with a mean separation @xmath134 mpc .
we have studied inflationary models with parameter @xmath169 .
figure 5a shows that the shape of the cluster power spectrum in these models is in good agreement with the observed power spectrum of the abell - aco clusters . however , the amplitude of the observed spectrum of clusters is about @xmath119% lower than predicted in these models by using approximation ( 29 ) .
we found a similar effect in 3 by comparing the analytic and numerical results .
it is possible that the linear approximation ( 29 ) somewhat overestimates the power spectrum of clusters due to dynamical effects that are not taken into account in this approximation .
therefore , the power spectrum of clusters can be expressed as @xmath170 b_{cl}^2 \ ,
p(k ) , \eqno(32)\ ] ] where the factor @xmath171 . for @xmath172 ,
the inflationary models studied in figure 5a , are consistent with the power spectrum of the abell - aco clusters at a confidence level of @xmath173% ( based on a @xmath141 test at @xmath142 mpc@xmath16 ) .
figure 5b demonstrates the power spectrum of the model clusters and the apm clusters with a mean separation @xmath140 mpc .
we have studied the @xmath1cdm model with @xmath174 . for @xmath172 ,
the cluster power spectrum in this model is consistent with the observed power spectrum of the apm clusters at a confidence level of @xmath175% .
( for @xmath176 , at a confidence level of @xmath173% ) .
available data are insufficient to rule out any of the models studied in figure 5 .
note that the power spectra of clusters predicted in the models with @xmath158 and @xmath159 are rather different on large scales , where @xmath164 mpc@xmath16 .
therefore , accurate measurements of the power spectrum of clusters and galaxies on these scales can serve as a discriminating test for these interesting models .
because of the low density of rich clusters of galaxies and their enhanced clustering strength compared to galaxies , rich clusters provide a powerful probe of large - scale structure in the universe . in this paper , we have examined the power spectrum of clusters in the press - schechter ( ps ) theory and in n - body simulations to see how the power spectrum of clusters is related to the power spectrum of matter density fluctuations in the universe .
we have derived an analytical expression to determine the correlation function of clusters for a given number density of clusters in lagrangian space and have examined this expression in the linear approximation . in order to study the power spectrum of clusters in redshift space
, we used the approximation derived by kaiser ( 1987 ) . for testing our analytic results , we used n - body simulations with realistic initial power spectra of density fluctuations .
we investigated the power spectrum of clusters with a mean separation @xmath177 mpc .
the numerical results showed that we can use the linear ps approximation to predict the power spectrum of clusters in real space and in redshift space at wavenumbers @xmath178 mpc . on smaller scales non - linear processes that are not taken into account in the linear ps approximation influence the results .
we also used the ps approximation to analyze the redshift - space power spectrum of clusters in @xmath1cdm models .
we investigated inflationary models with a scale - invariant initial spectrum of density fluctuations and with a steplike initial spectrum derived by starobinsky ( 1992 ) .
the results were compared with observations .
we found that the power spectrum of clusters in the @xmath1cdm models with a scale - invariant initial spectrum is not consistent with the observed spectra of the apm and abell clusters .
we investigated also inflationary models with a steplike initial power spectrum ( 30 - 31 ) that depends on two parameters @xmath154 and @xmath155 .
the parameter @xmath154 determines the location of the step and the parameter @xmath155 - the shape of the initial spectrum . for the models with @xmath158 and @xmath159 ,
the step parameter was chosen to be @xmath160 mpc@xmath16 and @xmath161 mpc@xmath16 , respectively .
we found that the power spectrum of clusters in these @xmath1cdm models is in good agreement with the observed power spectrum of the abell - aco clusters , if the initial parameter @xmath155 is in the range @xmath169 . to describe the power spectrum of the apm clusters
, we can use the @xmath1cdm model with parameter @xmath174 . | we examine the power spectrum of clusters in the press - schechter ( ps ) theory and in n - body simulations to see how the power spectrum of clusters is related to the power spectrum of matter density fluctuations in the universe . an analytic model for the power spectrum of clusters for their given number density
is presented , both for real space and redshift space .
we test this model against results from n - body simulations and find that the agreement between the analytic theory and the numerical results is good for wavelengths @xmath0 mpc . on smaller scales non - linear processes that are not considered in the linear ps approximation influence the result .
we also use our analytic model to study the redshift - space power spectrum of clusters in cold dark matter models with a cosmological constant ( @xmath1cdm ) and with a scale - invariant harrison - zeldovich initial spectrum of density fluctuations .
we find that power spectra of clusters in these models are not consistent with the observed power spectra of the apm and abell - aco clusters .
one possible explanation for the observed power spectra of clusters is an inflationary scenario with a scalar field with the potential that has a localized steplike feature .
we use the ps theory to examine the power spectrum of clusters in this model . |
photogrammetry is a technique to determine the geometric properties of an object from a sequence of photographic images , which has a long history since 1950s .
photogrammetry can be classified into two main categories as aerial photogrammetry and close range photogrammetry ( crp ) based on the camera location during the image acquisition process . in crp
, the camera is closed to the subject and 3d models , measurements and point clouds are the most common outputs .
this technique is used in different application areas , such as modeling and measuring buildings and large engineering structures , accident scenes inspection , mine inspection , quality controlling and film industry .
industrial photogrammetric measurement techniques utilize retro - reflective coded targets ( cts ) encoded with a unique identifier to signalize feature points .
two main categories of cts ; concentric rings [ 1 , 2 ] and dot distributions [ 3 - 5 ] have been introduced so far and two examples are shown in fig .
[ fig : fig1 ] . concentric rings are relatively easy to recognize and decode , but gives a limited number of code possibilities .
therefore , it is not suitable for measuring large scale structures which need several hundreds of markers to cover the whole structure .
dot distribution targets solve this problem by allowing a large number of code possibilities , but have difficulties in recognition due to the complexity .
different dots distribution cts have been introduced so far and some examples are shown in fig . [
fig : fig2 ] .
crp systems work with a large number of cts within an image and 30 to 100 images for one network .
therefore , it is really important to achieve fully automatic target detection to make the system cost and time effective . in this paper , we introduce a fully automatic coded target detection and decoding method based on the ct shown in fig .
[ fig : fig2](b ) .
the coded target detection process starts by converting the captured image to gray - scale and removing the image noise using the gaussian smoothing filter , explained in the eq .
[ eq1 ] . then the adaptive inverse thresholding with a gaussian kernel
is employed to convert the smoothed image to a binary image .
adaptive thresholding defines different threshold values for each pixel in the image by examining the surrounding neighborhood pixels as explained in eq .
[ eq2 ] where @xmath0 is the weighted sum of the neighborhood of @xmath1 and @xmath2 is a user defined constant .
@xmath3 @xmath4 contour detection is applied to the binary image for finding possible circular dots belong to the cts . after detecting contours ,
first we remove very large or small contours which are possibly not circular dots by examining the contour size .
then each contour is approximate to a polygon using douglas - peucker algorithm and analyze the shape to remove the contours which are not convex and have fewer number of sides .
algorithm introduced in [ 6 ] is used to fit each contour to an ellipse and find the center coordinates of circular dots .
finally , these detected circular dots are grouped into possible markers ( cts ) using a nearest neighbor search algorithm based on a k - d tree data structure .
each ct use in this research has eight circular dots where five of them are common for all the cts and define the axis ( orientation ) of the ct .
remaining three circles can be chosen from 22 possibilities and they define the code - word of the ct . the first step of the decoding process is to identify the five common circles from the detected possible cts .
we measure the total distance from each circle to every other circles within a ct and select the circle with the lowest total as center circle @xmath5 .
then we identify the two corner circles @xmath6 and @xmath7 by finding the two circles which make a straight line with @xmath5 .
from all the eight circles , only @xmath6 , @xmath5 and @xmath7 make a straight line within a ct . @xmath6 and @xmath7 can be differentiated by finding the distance to the @xmath5 .
to find the @xmath8 and @xmath9 , we find the two closest circles to the @xmath6 from the remaining circles .
those two circles should reside on the two sides of the line created by @xmath6 and @xmath7 and can be differentiated as @xmath8 and @xmath9 according to the side they reside . after finding the five common circles , the perspective transformation between the original projected ct and the detected one can be calculated .
then the remaining three circles can be projected onto the original marker and the code - word of the ct can be identified using a mask created by original positions of the code circles .
flow chart in fig .
[ fig : fig3 ] summarizes this decoding process .
we test our proposed method with different structures ( fig .
[ fig : fig4 ] ) and achieve nearly 100% detection rate when the cts are nearly orthogonal to the camera view and about 85% for other cases . for every successful detection of a ct , decoding accuracy is almost 100% and most of the errors come from the target detection process .
in an effort to enhance the performance of photogrammetric measurement techniques , in this research , we proposed a fully automated coded target detection and decoding method by using various image processing and vision techniques . as a future work we are planning to integrate this target detection technique with natural feature tracking techniques to develop a more accurate and enhanced photogrammetric system .
this work was supported by the industry core technology program granted financial resource from the ministry of trade , industry & energy , republic of korea ( no . 10043897 ) and samsung heavy industries co. | close - range photogrammetry is widely used in many industries because of the cost effectiveness and efficiency of the technique . in this research
, we introduce an automated coded target detection method which can be used to enhance the efficiency of the photogrammetry .
photogrammetry , coded targets , decoding |
whether a thermodynamic phase transition underlies the transformation of a supercooled liquid into an amorphous solid , or _
glass _ , at the laboratory glass transition temperature @xmath0 is among the central questions addressed by numerous researchers studying the supercooled liquid and glassy states .
the notion of configurational entropy@xcite has played a significant role in attempts to define and understand the thermodynamic nature of the glass transition . in recent times , there have been various attempts to determine the configurational entropy of realistic liquids analytically and by computer simulations @xcite . the purpose of this paper is to compare two such methods that have been studied recently , namely the evaluation of the configurational entropy _ via _ the analysis of local potential energy minima or _ inherent structures _ ( is)@xcite , and by the calculation of _ basin _ free energies by confining the liquid within a localized region of configuration space by the imposition of constraints@xcite . these approaches , and results from their implementation , are described in the following sections .
the model liquid studied is a binary mixture of @xmath1 type @xmath2 and @xmath3 type @xmath4 particles , interacting _ via _ the lennard - jones ( lj ) potential , with parameters @xmath5 , @xmath6 , @xmath7 , and @xmath8 , and @xmath9 , which has been extensively studied as a model glass former@xcite .
results presented in sec .
ii are from molecular dynamics simulations , at a reduced density @xmath10 , which have been described in detail elsewhere@xcite .
since the density is fixed , the dependence on density is not always shown explicitly in the following .
in the inherent structure approach@xcite , one considers the division of configurational space into basins of local potential energy minima . in practice such basins
may be defined as the set of all points in configurational space that maps to the given local minimum under a specified local energy minimization procedure . quite generally , one may then write the total partition function of the system as a sum of restricted partition function integrals over individual basins .
rewriting the partition function in this way introduces an entropy term associated with the number of local potential energy minima . with the expectation that configurations within a given basin are accessible to each other by thermal agitation
while those belonging to distinct minima may not be , the number of distinct potential energy minima can be seen to be a measure of the number of physically distinct configurations or structures the system can adopt , _
i. e. _ a measure of the configurational entropy .
thus , the canonical partition function is re - written as a sum over all local potential energy minima , which introduces a distribution function for the number of minima at a given energy : @xmath11 where @xmath12 is the total potential energy of the system , @xmath13 indexes individual inherent structures , @xmath14 is the potential energy at the minimum , @xmath15 is the basin of inherent structure @xmath13 , @xmath16 is the number density of inherent structures with energy @xmath14 , and the configurational entropy @xmath17 ( note that here @xmath18 is a function of energy ; the equilibrium average of this quantity displayed in fig .
4 as a function of temperature ) .
the probability of finding the system in the basin of an inherent structure of a given energy is given by the above as , @xmath19 the probability distribution @xmath20 can be obtained from computer simulations , and offers a means of obtaining @xmath18 , provided one can estimate @xmath21 ( equivalently the free energy @xmath22 of the system ) and the basin free energy @xmath23
. the free energy at any desired temperature is obtained from thermodynamic integration of pressure and potential energy data from md simulations@xcite .
the absolute free energy @xmath24 of the system at density @xmath25 at a reference temperature @xmath26 is first defined in terms of the ideal gas contribution @xmath27 and the excess free energy @xmath28 obtained by integrating the pressure from simulations : @xmath29 here , @xmath30 is the number of particles , @xmath31 , @xmath32 is the de broglie wavelength , and @xmath33 arises from the mixing entropy .
@xmath34 at a desired temperature may be evaluated by integrating the potential energy , @xmath35 : @xmath36 as observed in @xcite , the @xmath37 dependence of @xmath35 at the studied density is well described by the form @xmath38 , in agreement with predictions for dense liquids@xcite .
a fit of the potential energy data to this form affords a means of extending with confidence the temperature dependence of @xmath35 to @xmath37 values where direct md data is unavailable . the _ basin free energy _
@xmath23 is obtained by a restricted partition function sum over a given inherent structure basin , @xmath39 . for sufficiently low temperatures
, one may expect the basin to be harmonic to a good approximation . in the harmonic approximation , we have @xmath40 where @xmath41 are eigenvalues of the hessian or curvature matrix at the minimum . for individual minima ,
these eigen values are obtained by numerical diagonalization of the hessian .
the basin free energy can then be obtained either as a function of the inherent structure energy ( by averaging free energies within individual energy bins ) or as a function of temperature , by averaging all inherent structures sampled at a given temperature .
@xmath42 is a slowly varying function of temperature ( the temperature dependence is obtained by averaging over @xmath43 , @xmath44 inherent structures at @xmath45 , @xmath46 respectively ) , and is fitted to the form @xmath47 which fits available data quite well . to to if the harmonic approximation to the basin free energy is accurate , inversion of eq .
( [ scdos ] ) , expressing @xmath48 in terms of @xmath49 , @xmath50 ( or @xmath22 ) and @xmath23 , for different temperatures @xmath37 , should result in curves that overlap with each other , as @xmath48 is independent of @xmath37 .
figure 1 shows the result of such inversion , which indicates that below @xmath51 , the various @xmath18 curves do overlap , while they do not at higher @xmath37 .
the procedure applied here is similar to , but improves upon , the procedure of shifting unnormalized @xmath18 curves adopted in @xcite .
thus , fig .
1 indicates that a harmonic approximation to the basin free energy is not valid for temperatures higher than @xmath51 . the temperature dependence of the average inherent structure energy @xmath52 , shown in fig .
2 , is consistent with this conclusion . as discussed in @xcite , a simple expectation for the t - dependence of the average inherent structure energy in the harmonic regime is that @xmath53 .
figure 2 shows that such a t - dependence is indeed valid at low temperatures , but breaks down for @xmath54 .
however , this observation must be viewed in conjunction with two other observations about the topography of the inherent structure basins : ( i ) it has been demonstrated recently @xcite that the separation between ` vibrational ' and ` inter - basin ' relaxation becomes reasonable for temperatures close to and below the mode coupling @xmath55 ( @xmath56 for the model liquid studied here ) .
( ii ) the difference in the potential energy of instantaneous configurations and the corresponding inherent structures is nearly linear with a slope of @xmath57 for temperatures as high as @xmath58 , as shown in fig .
such a linear temperature dependence would normally be associated with harmonic behavior , which in the present case is misleading .
to the total entropy of the liquid @xmath59 as well as the basin entropy @xmath60 are evaluated as a function of density and temperature from the total and basin free energies .
the configurational entropy @xmath61 and the ideal glass transition @xmath62 are then given by , @xmath63 figure 4 shows the configurational entropy so obtained as a function of @xmath37 . by extrapolation ,
based on the assumption that the potential energy varied with temperature as @xmath64 , the ideal glass transition occurs at @xmath65 , in good agreement with estimates in @xcite . to
an alternate approach to defining the basin entropy , which has been explored by speedy@xcite is to impose constraints on a liquid to trap it in one of the basins it samples in equilibrium .
a related approach has also been studied in @xcite . with suitably chosen constraints ,
the calculated properties of the constrained system allow the evaluation of the basin entropy .
a reasonable choice of constraint will restrict the system to a physically meaningful set of configurations related to each other without the need for configurational rearrangement .
further , such a constrained system should behave reversibly . in this work ,
the usefulness of one simple constraint is explored , by calculating the configurational entropy for a set of six temperatures at a fixed density of @xmath66 , and compared with corresponding results from the inherent structure calculations described above .
it is found that the constrained simulations result in comparable numbers for the configurational entropy from the inherent structure results .
ten sample configurations are chosen at each temperature , and the voronoi tessellation is performed for each configuration .
the voronoi cell of each given particle , and the corresponding geometric neighbors , correspond to the _ cage _ a particle experiences at short and intermediate time scales .
a configurational rearrangement of particles in the system will result in a restructuring of the voronoi tessellation as well .
thus , the constraint of restricting particles to their voronoi cells is an _ a priori _ reasonable choice .
hence , a constraint is imposed which confines each particle to its voronoi cell during the monte carlo simulation from which the properties of this constrained system are evaluated .
each monte carlo simulation mentioned below is performed for @xmath67 monte carlo steps .
the constrained system can be studied at any desired temperature ; the temperature of the simulation from which the reference configurations are taken will be referred to as the fictive temperature where there is need to distinguish these two temperatures . in order to estimate the configurational entropy
, we must evaluate the free energy of the constrained system .
this is done by thermodynamic integration@xcite from a reference system where each particle experiences a harmonic potential around the initial configuration ( einstein cystal ) .
considering a potential energy function of the form , @xmath68 where @xmath69 is a tuning parameter that varies between @xmath70 and @xmath71 , @xmath72 is the lennard - jones potential of the unconstrained system , @xmath73 is the constraining potential ( which is zero if the constraint is obeyed and infinity if it is not ) , the corresponding free energy is given by @xmath74 .
\label{cga}\ ] ] the required free energy , @xmath75 is related to that of the einstein crystal ( which may be calculated straightforwardly ) , by @xmath76 where , from differentiating eq .
( [ cga ] ) with respect to @xmath69 , @xmath77 the required average in the above equation is calculated by performing monte carlo simulations for a set of @xmath78 @xmath69 values .
the values of @xmath79 obtained are shown in fig .
5 . the free energy for @xmath80 is @xmath81.\ ] ] all the above calculations are also performed using the inherent structures corresponding to the equilibrated liquid configurations mentioned above . with the free energy of the liquid evaluated as described in the previous section and the free energy of the constrained system obtained as described here , the configurational entropy of the system is given by @xmath82 where @xmath83 is the free energy of the constrained system .
the resulting configurational entropies are shown in fig .
4 . the @xmath18 values from the constrained system simulations are comparable with the inherent structure results , but the agreement is moderate . in particular , the constrained system results vary more weakly with temperature . to verify that the chosen constraint is a reasonable one , the free energies of the constrained system ( for configurations from equilibrium runs at ( fictive ) temperatures @xmath84 ) are obtained independently at @xmath85 and @xmath86 from thermodynamic integration with respect to the einstein crystal .
simulations are also performed for temperatures in between these two values , from which the temperature dependence of the potential energy is obtained . using eq .
( [ efit ] ) , @xmath87 is calculated by integrating from @xmath86 .
the difference of the directly calculated value and the one by integration is found to be @xmath88 for @xmath89 and @xmath90 for @xmath91 . in other words
, the constrained system appears to be reversible within the margin of error represented by these numbers .
however , the discrepancy in the @xmath18 values between the constrained system and the inherent structure estimates is of the same order .
indeed , the discrepancy in the @xmath18 values for @xmath92 and @xmath93 is roughly the same as the discrepancy in the free energies above .
it is likely that the sample of ten configurations used here is too small to obtain more accurate values . to
configurational entropy is obtained for a binary mixture liquid from analysis of inherent structures , and from estimation of the basin free energy _ via _ constrained system simulations . while the harmonic approximation used in the inherent structure approach to evaluate the basin free energy is in general questionable , the difficulty in the constrained system approach is the proper choice of constraint .
the values for the configurational entropy obtained are comparable but show only moderate agreement .
further tests for the accuracy of the constrained system results , and more importantly , exploration of improved constraining methods are desirable for making a more stringent comparison of these two methods of calculating the configurational entropy .
extremely useful discussions with robin speedy are gratefully acknowledged . | computer simulations have been employed in recent years to evaluate the configurational entropy changes in model glass - forming liquids .
we consider two methods , both of which involve the calculation of the ` intra - basin ' entropy as a means for obtaining the configurational entropy .
the first method involves the evaluation of the intra - basin entropy from the vibrational frequencies of inherent structures , by making a harmonic approximation of the local potential energy topography .
the second method employs simulations that confine the liquid within a localized region of configuration space by the imposition of constraints ; apart from the choice of the constraints , no further assumptions are made .
we compare the configurational entropies estimated for a model liquid ( binary mixture of particles interacting _ via _ the lennard - jones potential ) for a range of temperatures , at fixed density . |
galaxies responsible for gravitational lensing have typically been modeled as one - component systems in which the entire mass distribution of the galaxy lens , visible plus dark matter , is fairly well approximated by a singular isothermal sphere ( sis ) : @xmath0 ( see e.g. narayan and bartelmann 1997 ) .
this is probably justified for e / so galaxies where both components have a surface density fairly independent of orientation ( a small ellipticity is required only to understand the presence of four - image systems ) and the high central density represents the high density of the visible matter .
by contrast , in spiral galaxies the disk component has a very different structure from the dark matter halo and the bulge , which results in significantly different projected surface density depending on orientation .
given that the projected surface density of the disk changes by a factor of radius / thickness @xmath1 between a face - on and an edge - on spiral , and that the dark and visible components are known to contribute roughly equally to the total mass within 4 5 disk scale lengths , one would expect significant effects on lensing properties for images within a couple of optical radii .
the identification of the lensing galaxy in the double image system b1600 + 434 as a nearly edge on spiral by jaunsen and hjorth ( 1997 ) gives us a perfect example to demonstrate the importance of including the disk component in modeling lensing by spirals . however , an understanding of the lensing geometry of spirals is important for several reasons besides that of modeling observed systems .
first , recently the gravitational lensing effects on damped lyman alpha systems ( dlas ) have been investigated ( smette , claeskens , and surdej 1997 ; bartelmann and loeb 1996 ) .
at least some dlas are believed to be due to absorption in protogalactic disks or spiral galaxies ( wolfe 1988 , 1997 ) , and if the gas column density falls off exponentially , one is necessarily dealing with lensing within a few exponential scale lengths of the galaxy center , which is where we expect disk lensing to be important . secondly ,
the extra parameters added by introducing a disk component are the physically interesting ones of disk surface density and scale length , so that the observational study of gravitational lensing by spirals will provide new constraints on the distribution of mass in the various components of spirals .
although only a few such systems are currently known , surveys such as the sloan digital sky survey will catalog a large number of quasars , among which there should be a significant number lensed by spiral galaxies . this is why it is also important to understand the effect of disk lensing on statistical properties such as the cross section for multiple - image lensing by spirals . in this paper we first explain our method of modeling spiral galaxy lenses as singular isothermal spheres plus infinitesimally thin disks ( sis+disk ) .
we start with the simple example of a constant density disk , to show the effect of adding a disk .
we show that such a model fits many of the observations on b1600 + 434 , and discuss the results of modeling , more properly , with an exponential disk .
we then discuss the dependence of the lens cross section for multiple lensing of background sources on the inclination of the disk , and discuss also how it would enhance the gravitational lensing effects on dlas .
we finish with the conclusions and implications from our work .
we model a spiral galaxy as consisting of two mass components : a spherical halo and an infinitesimally thin disk .
we assume that the halo has the density profile of a singular isothermal sphere , while the disk is treated as having surface density that is either constant , or falls off exponentially . for simplicity and clarity
we start by treating the case of the constant density disk .
the orientation of the disk defines the coordinate system , such that the x and y directions are along the major and minor axis , respectively , of the projected disk , and the disk center defines the origin .
the model has then 6 parameters : the position of the source ( @xmath2 ) , the inclination of the disk @xmath3 , the velocity dispersion of the sis @xmath4 ( the sis can be equivalently be described by @xmath5 ) , and the central surface density @xmath6 and characteristic length of the disk . for a constant density disk this characteristic length is the radius of the disk , while in the exponential disk case this becomes the scale length , @xmath7 .
thus , measuring the position of the lens , the inclination of the disk , the two image positions ( @xmath8 ) , and the magnification ratio completely determines the system . the inclined circular disk projects onto an ellipse .
we denote the axis ratio of an ellipse in general as @xmath9 , and that of our disk as @xmath3 .
any point @xmath10 external to the ellipse will define a new confocal ellipse with axis ratio @xmath11 . then the reduced deflection angle , modified from schramm ( 1990 ) , is given by @xmath12 here @xmath13 is the usual critical density defined as @xmath14 and @xmath15 is the lens , source , or lens - to - source angular diameter distance .
if the point @xmath10 is inside the disk , then @xmath11 is just replaced with @xmath3 .
in the gravitational lens system b1600 + 434 identified by jackson et al .
( 1995 ) , the lensed quasar is at a redshift @xmath16 , the image separation is 14 , and the intensity ratio is @xmath17 in the radio ( in the optical , image b suffers significant extinction ) . the lens galaxy , identified as a nearly edge on spiral by jaunsen and hjorth ( 1997 ) , has an estimated photometric redshift @xmath18 , and axis ratio @xmath19 , or @xmath20 .
( it is possible that this galaxy is actually an s0 , as suggested to us by c. kochanek , private communication , jannuary 1997 , but this will not effect our analysis as there is still a disk . ) to emphasize that this paper is meant as a qualitative exploration of the sis+disk model , and because @xmath21 is still uncertain , we treat all measurements as exact and include no estimates of errors .
the positions of the images in the lens - centered , axes - defined coordinate system are given in table i. note that the lensing galaxy is just 035 from image b. in all of our calculations , we assume @xmath22 , @xmath23 and @xmath24 . also we calculate angular diameter distances assuming all the matter in the universe is smoothly distributed ( see schneider , ehlers and falco 1992 ) . treating the lens as a singular isothermal sphere requires a velocity dispersion of 200 km / s to account for the image separation .
of course , in this case the lens must be collinear with the images , which is not the case observationally ( one would need to introduce some ellipticity in the dark matter distribution to avoid this ) .
however , we can obtain an estimate of the implied relative magnification and the time delay by shifting the lens to the collinear position along a line perpendicular to @xmath25 .
the implied time delay is @xmath26 days , and the magnification ratio is 3.8 .
( the results are similar if the lens position is shifted keeping the ratio of image - lens distances : @xmath27 , magnification ratio 3.2 . )
if one chooses instead to shift the lens to a position that gives the observed relative magnification , the time delay is @xmath28 days , but then the images would both be at a similar distance from the lens , which is not as observed . if instead we model the system as a sis+constant density disk , we introduce two new parameters , surface density and disk length .
the farther image s light passes through the disk only @xmath29 kpc from its center , so we can start by assuming that the disk length is greater than this and remove one parameter . the equations for a disk with radius greater than @xmath29 become @xmath30 and the lensing potential of the disk in this case is @xmath31 solving for the image positions in table i gives @xmath32 , @xmath33 km / s , and a source position of ( 0.037 , 0.12 ) .
the magnification ratio is 1.3 , and the time delay is now @xmath34 days . see figure 1 .
the addition of the disk component has four main effects .
first it breaks the spherical symmetry , allowing the lens not to be collinear with the images .
second , the velocity dispersion in the halo becomes significantly less , as the disk component contributes substantially to the deflection angle . and
third , in this case the magnification ratio is reduced , though in general it depends strongly on the orientation of the images around the lens axis .
also the time delay is increased , because the potential now has a term quadratic in @xmath35 , and because the geometric time delays do not cancel out as they do in the sis case .
it is interesting to note that the surface density of the disk comes out as a reasonable number , although a priori it could have been anything . the deflection caused by a disk with density @xmath36 is of the same scale as that caused by a halo with @xmath37 km / s .
since we observe roughly this amount of luminous matter in the disk , it is clear that the disk s contribution can not be ignored .
of course a disk with constant surface density out to 9@xmath38 kpc is not a realistic model .
we can realistically assume the disk s surface density falls off as an exponential with scale length @xmath7 . in figure 2
the images from an exponential disk with @xmath39 kpc , @xmath40 and a halo with @xmath41 km / s are shown .
the images are within the position errors for the lens center , and the magnification ratio is roughly correct at 0.97 .
interestingly we find that the magnification ratio is strongly dependent on the relative orientation of the images and the lens axis .
figure 3 shows the magnification ratio as a function of angle from the y - axis .
a magnification ratio of 1.3 requires that the source be about @xmath42 off the y - axis , which also corresponds to the image orientation found in b1600 + 434 . in this preliminary study
we have found a surprisingly large disk - to - halo ratio in order to fit the image positions .
for the exponential disk , the total disk mass is @xmath43 , compared to a halo velocity dispersion of @xmath41 km / s . in comparison , standard lore ( binney and tremaine 1987 ) gives values of @xmath44 for the disk mass and 155 km / s velocity dispersion for the milky way with a disk scale length of 3.5 kpc .
while there are large uncertainties in these parameters for the milky way , our calculated disk mass does seem excessively large . without an understanding of the uncertainties , it is hard to judge how significant these results are . in particular
the inclination we have used of @xmath45 ( based on the axis ratio measured by jaunsen and hjorth 1997 ) could be much higher , which tends to decrease the scale length and hence the total disk mass .
a preliminary inspection of hst archival images of b1600 + 434 reveals a dust lane nearly bifurcating the disk , suggesting an inclination greater than @xmath46 .
however , if the large disk mass did turn out to be significant , it may be evidence that the dark matter halo is flattened ( see for example olling 1996 ; sackett 1996 ; rix 1995 ) , and the non - spherically situated dark matter is contributing to what we have assumed is a visible matter disk .
it is also possible that the dark matter halo may have a core radius , and since the lensing is mostly sensitive to the matter inside the images , the @xmath4 we are measuring is lower than the value one would use to characterize the galaxy .
we are preparing to address these issues in a more detailed study to follow .
the inclination of the disk can have a dramatic effect on the lens cross section for multiple lensing of background sources , as we show in figure 4 . here
we model the disk with constant surface density again , but we choose parameters more representative of an @xmath47-type spiral . in this case
it is possible to obtain a simple expression for the critical inclination , @xmath48 , at which the cross section begins to exceed that of the sis halo alone .
just note that this occurs when the inner caustic due to the disk touches the sis caustic .
if the inclination slightly exceeds @xmath49 , then we can have a source just outside the sis caustic and just inside the disk caustic ( see the @xmath50curve in figure 4 ) .
a source at this location , @xmath51 with @xmath52 , has three images . using the lens equation
@xmath53 we see that , assuming the disk radius is larger than @xmath54 , one image is at @xmath55 because we know that as the source passes through the caustic the images will merge , we know that the other two images must be close to the first one .
thus there are two images at @xmath56 and @xmath57 , where @xmath58 .
for these two images the lens equation requires @xmath59 independently of @xmath60 .
therefore , @xmath61 gives the critical inclination . in figure 4
we show two inclinations for a disk of mass @xmath62 and radius @xmath63 kpc , inside a sis halo with @xmath64 km / s . for this choice of parameters @xmath65 .
thus , at an inclination of @xmath50 the cross section shows a modest increase .
by contrast , at @xmath66 we find a very large increase by a factor of @xmath67 over that of the sis halo alone .
the disk model is not realistic , of course , but the effect is quantitatively similar for an exponential disk . in particular
, @xmath49 is still fairly well approximated by eq .
( 10 ) . whether this increase in cross section significantly increases the fraction of spiral galaxies responsible for multiple image lensing depends on the amount of mass in the disk . in figure 5
one sees that for parameters like those quoted for the milky way the average ( integrated over cos@xmath68 ) increase is small , roughly @xmath69 . while , for a heavy disk like our model of b1600 + 434 , the average cross section increases by @xmath70 .
but since the amplification also increases with disk inclination , the brightest quasars lensed by disk galaxies will probably have highly inclined disks , especially if the disks are massive . thus the fraction of lenses that are spirals , and their relative inclinations , will place strong constraints on the total disk mass .
we shall discuss these issue in more detail in a subsequent publication .
the last problem we study here is how our results would affect the gravitational lensing effects on dlas .
bartelmann and loeb ( 1996 ) and smette ( 1995 ) have pointed out that gravitational lensing could significantly affect the observed properties of dlas for two reasons .
first , magnification bias significantly enhances the probability of observing damped absorption in a qso spectrum .
second , there is a cutoff in the observed column density due to light bending , which makes it far more likely to observe absorption that occurs at large radii .
since the column density of absorbing hi falls off exponentially , this significantly affects the observed column densities . in their studies ,
smette et al . (
1997 ) , bartelmann and loeb ( 1996 ) treated the spiral disk as essentially massless for lensing purposes .
we find that including the disk lensing will significantly enhance both these effects , as we show in figure 6 .
there we show the effects of adding disk lensing on the magnification and impact parameters of two sources .
first , note that there is a significant increase in the magnification of the image that , precisely , occurs on the surface of the disk .
for the example shown , the @xmath71type disk considered above , the point - source magnification increase is @xmath72 for the source at position 1 ( 2 ) .
second , there is an increase in the impact parameter , which would result in further reduction of the observed column density .
these results are typical of what we find at high inclinations of @xmath73 @xmath46 . note that the increased magnification by inclined disks increases the likelihood of lensing in this configuration ; this could obviate the need for rather thick disks ( e.g. , wolfe 1997 ) to account for the metal line systems seen in dlas .
also one must be aware that unresolved multiple images may pass through the same disk at various places complicating the reconstruction of the absorbing system .
a complete analysis is now far more involved , as one must model the disk as an exponential disk , average over the observed magnification , and average over the disk inclination .
also the effects of dust in obscuring images ( as clearly demonstrated in the relative reddening of image b to a in b1600 + 434 ) must be included .
we will discuss the full problem in a subsequent publication .
we have shown that the inclusion of a disk component in a spiral gravitational lens model has significant effects that can not be ignored .
our results imply that a host of issues ranging from the ability of spirals to lens qso s to detailed properties of lens systems should be reconsidered to properly include the lensing effects of their disks .
we have shown , as an example , that a sis+disk model can account for the observations of the b1600 + 434 system with fairly reasonable parameters .
the implications of our results are that a velocity dispersion @xmath4 significantly smaller than the sis prediction of 200 km / s , and a time delay greater than one month , are to be expected . with more data and more detailed analysis
, the modeling of multiply imaged systems lensed by spiral galaxies will constrain the relative amounts of matter in the disk and halo when disk lensing is included . however , because of the dependence on new parameters , further study will be necessary before such systems can give an unambiguous measurement of the hubble parameter @xmath74 .
we have also noted that the disk contribution to lensing will make the amplification and by - pass effects on the observed distribution of dlas in bright qso s even more pronounced at high column densities .
after the first version of this paper was posted on the preprint archive as astro - ph/9701110 , the preprint astro - ph/9702078 was posted by wang and turner discussing critical curves and caustics of inclined , constant density disks .
their equations and figures can be seen to be the same as ours if one is careful to scale appropriately by the hubble constant @xmath74 , and to notice that they have calculated angular diameter distances with all the matter in the universe in clumps while we have assumed it to be distributed smoothly .
this work has been supported by nsf grant phy-9600239 at um - st .
louis , and by nsf grant phy-9402455 and a nasa atp grant at ucsc .
am also acknowledges gaann support .
the authors are grateful for helpful suggestions from the referee , matthias bartelmann ; for help from raja guhathakurta and steve vogt in interpreting the hst archival image of b1600 + 434 ; and correspondence from a.o .
jaunsen , chris kochanek , and tom schramm .
bartelmann , m. & loeb , a. 1996 , apj , 457 , 529 binney , j. & tremaine , s. 1987 , galactic dynamics , ( princeton : princeton univ . press ) , 17 jackson , n. , et al .
1995 , mnras , 274 , l25 jaunsen , a. o. & hjorth , j. 1997 , a&a , 317 , l39 olling , r. p. 1996 , preprint astro - ph/9612018 narayan , r. & bartelmann , m. 1997 , in formation and structure of the universe , ed . a. dekel & j. p. ostriker ( cambridge : cambridge university press ) , in press rix , h .- w .
1996 , in unsolved problems of the milky way , eds .
l. blitz & p. teuben ( dordrecht : kluwer ) , 23 sackett , p. d. , in astrophysical applications of gravitational lensing , eds . c. s. kochanek & j. n. hewitt ( dordrecht : kluwer ) , 165 schneider , p. , ehlers , .j & falco , e. e. 1992 , gravitational lenses , ( springer - verlag : new york ) , 142 schramm , t. 1990 , a&a , 231 , 19 smette , a. , claeskens , j. f. , & surdej , j. 1997 , preprint to appear in new astronomy smette , a. 1995 , in qso absorption lines , ed . g. meylan , ( new york : springer ) , 275 wolfe , a.m. 1997 , to appear in critical dialogues in cosmology , ed .
neil turok ( singapore : world scientific ) .
wolfe , a. m. 1988 , in qso absorption lines : probing the universe , eds .
j. c. blades , d. a. turnshek , & c. a. norman ( cambridge : cambridge univ . press ) , 297 | spheroidal components of spiral galaxies have been considered the only dynamically important component in gravitational lensing studies thus far . here
we point out that including the disk component can have a significant effect , depending on the disk inclination , on a variety of lensing properties that are relevant to present studies and future surveys . as an example , we look at the multiple image system b1600 + 434 , recently identified as being lensed by a spiral galaxy .
we find that including the disk component one can understand the fairly large image separation as being due to the inclination of a typical spiral , rather than the presence of a very massive halo .
the fairly low magnification ratio can also be readily understood if the disk is included .
we also discuss how such lensed systems might allow one to constrain parameters of spiral galaxies such as a disk to halo mass ratio , and disk mass scale length .
another example we consider is the quasar multiple lensing cross section , which we find can increase many fold at high inclination for a typical spiral .
finally , we discuss the changes in the gravitational lensing effects on damped lyman alpha systems ( dlas ) when disk lensing is included . |
mathematical modeling of infectious diseases has become an area of increasing interest in the last decades @xcite .
the development and use of mathematical models are a powerful tool to understand the complex problem of infectious disease transmission .
after the success of the simple sir compartmental model in the description of the basic and common features of the transmission process @xcite , the models have become more complex and specific for the different infectious diseases in order to help in the evaluation and design of control strategies @xcite .
these complexities may include , among others , age structure of the population , immune status of the individuals , structure of the social contacts , spatial heterogeneity , etc . wherever possible , deterministic compartmental models are usually chosen because they are simple to solve numerically ,
the interpretation of results is direct , and the level of complexity may be increased gradually by adding new compartments .
however , from the beginning of infectious disease modeling , the importance of stochastic effects has been mentioned @xcite . in more recent work ,
the significance that a stochastic treatment could have in the transmission of some diseases such as pertussis has been highlighted @xcite .
in particular , the randomness and heterogeneity of contacts are one of the intrinsically stochastic aspects of the contagion process and knowing the accuracy that can be obtained with a deterministic approach to the problem is not a trivial matter . in ref.@xcite
a sirs spatial stochastic model having only local infective contacts is considered on a lattice and a pa scheme to approach the dynamic is developed ( some details are described in section [ sectionda ] ) . in ref.@xcite another model on a lattice is proposed , where local and global contacts are considered with the aim of emulating the more realistic situation of social relations where some `` fixed '' persons are frequently contacted ( the neighbors on the lattice ) and some other unknown people are met by chance ( the global contacts ) . in their model , a parameter @xmath1 characterize the weight of global contacts and , in this way , the influence of the degree of locality or globality of the contacts on the dynamics of the disease transmission is analyzed by changing a single parameter .
other authors have characterized the stochastic fluctuations of this model as a function of @xmath1 for a wide range of model parameters that include several infectious diseases @xcite .
@xcite studied the quasi - stationary state ( qss ) of the model for parameters corresponding to pertussis disease and quantified the relation between the effective transmission of the disease and the correlation of susceptible individuals with their infected neighbors . in the present work we propose a deterministic approximation ( da ) to the stochastic model with local and global contacts ( sm ) mentioned above , where the local contacts on the square lattice are treated here under a pa scheme based on the one used by joo _
we compare the predictions of our proposed da with the results of simulations performed with the stochastic model for parameters corresponding to two infectious diseases : pertussis and rubella .
we make this comparison for different situations of the disease transmission dynamics : the epidemic spread , the endemic state of the disease and dynamic perturbations that may arise from control measures implemented against the disease . in all the cases
studied , the agreement between the results obtained from the sm simulations and da was very good .
we consider an @xmath2 square lattice under periodic conditions where we identify sites with individuals .
each one of the @xmath3 individuals may be in one of the three epidemiological states : @xmath4 , @xmath5 or @xmath6 ( susceptible , infected or recovered ) .
the state of an individual may change through the following processes : infection ( @xmath7 ) , recovering ( @xmath8 ) , death and birth ( @xmath9 , @xmath10 , @xmath11 ) .
infections occur through infective contacts among susceptible and infected individuals .
we define an infective contact ( ic ) as a contact between two individuals such that if one individual is susceptible and the other infected , the former becomes infected .
we assume that an individual at a given site has an ic with a randomly chosen individual on the lattice with probability rate @xmath12 , and with one of their four nearest neighbors with probability rate @xmath13 .
local contacts represent the contacts in the circle of stable relations of an individual , while global random contacts represent people met by chance ( for example , on a bus , at the supermarket , etc . ) . by changing @xmath1
, we may change the degree of locality " of the contacts in the system .
the case @xmath1=1 corresponds to the classical sir model ( uniform mixing ) where an individual may have an ic with any other individual in the system with the same probability rate @xmath14 .
recovery from infection in this model is the same for every site and occurs at a probability rate @xmath15 .
deaths are assumed to be independent of the individual s state and occur at the same probability rate @xmath16 .
when an individual dies at a site , another individual is born simultaneously at this site in order to avoid empty sites during the simulation .
the processes defined above are markovian , so , the defined probability transition rates between states for each site at a given time control the dynamical evolution of the system .
stochastic simulations are performed using gillespie algorithm @xcite .
the algorithm gives a sequence of times and the corresponding states where two consecutive states differ by a single process that occurs at a given site .
the process and the time when it takes place are generated from simple rules and two random numbers ( see ref .
@xcite for details ) . for a detailed description of the model and the implementation of the simulation algorithm ,
see ref.@xcite .
we consider an @xmath2 lattice as in section [ sectionsm ] . at each time @xmath17
, the individuals must be in one of the three states : @xmath4 , @xmath5 or @xmath6 ( susceptible , infected or recovered ) .
we define @xmath18 , @xmath19 and @xmath20 as the total number of individuals in those states .
we denote @xmath21 when the state of the individual @xmath22 is @xmath23 ( for example , @xmath24 ) .
each pair @xmath25 of ( horizontal or vertical ) neighboring individuals can be in six possible states : @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 .
we remark that @xmath27 involves the cases @xmath32 and @xmath33 ; and the same holds for @xmath28 , @xmath30 .
let @xmath34 be the total number of each type pair of neighboring individuals .
our unknowns , arranged in the vector * x*=(@xmath35 ) , are the preceding quantities normalized by @xmath36 .
they are functions of the time @xmath17 and are linked by @xmath37 @xmath38 @xmath39 @xmath40 ( [ l1 ] ) and ( [ l2 ] ) hold because there are @xmath36 individuals and @xmath41 pairs of neighboring individuals . to obtain ( [ l3 ] ) ( and analogously ( [ l4 ] ) ) observe that each individual in susceptible state @xmath4 belongs to four pairs of neighboring individuals of type @xmath26 , @xmath28 or @xmath27 but in their total contribution @xmath42 , the @xmath26 pairs are counted twice . in the da equations that we propose ,
the terms related to the birth - death and recovery rates do not depend on the local structure and are treated as in the classic sir scheme .
but in the case of the transfer rates among classes due to infections , @xmath43 , we make a decomposition @xmath44 being @xmath45 and @xmath46 , @xmath47 two different rates that must be defined for global an local infective contacts respectively .
the decomposition is motivated by the following argument in sm .
consider an individual ( or a pair , according to the case ) in any of the states @xmath4 , @xmath26 , @xmath27 or @xmath28 at time @xmath17 .
it can reach , through an infective contact , the new state @xmath5 , @xmath27 , @xmath29 or @xmath30 respectively .
denote @xmath48 the event the individual ( or the pair ) changes the first state by the second one due to an infective contact in the time interval @xmath49 $ ] "
. then @xmath50 . but @xmath48 is the disjoint union of @xmath51 and @xmath52 , the events corresponding to one global and one local infective contact in @xmath49 $ ] respectively . assuming that @xmath53 and @xmath54 are known , we have ( according to the sm rules ) @xmath55 and @xmath56 . finally , by the total probability law , @xmath57 .
let @xmath25 and @xmath58 be two different pairs of neighboring individuals sharing @xmath59 . the key fact of the pa scheme in ref.@xcite ( with only local infective contacts ) consists in the approximation of the probability that the triplet @xmath60 reaches a given state @xmath61 at time @xmath17 by @xmath62 in order to apply the above approximation to the treatment of the local contacts in da we consider the natural correspondence between fractions ( our unknowns in da ) and probabilities .
for example @xmath63 ( the fraction of susceptible individuals at time @xmath17 in da ) can be regarded as the probability that a given individual @xmath22 is of type @xmath4 at time @xmath17 in sm , and @xmath64 ( the fraction of @xmath27 pairs at time @xmath17 in da ) as the probability that a given pair @xmath25 of neighboring individuals is of type @xmath27 at time @xmath17 in sm .
we present now a characteristic case to show how the preceding observations are used to define the various rates , @xmath43 , contributing to the equations ( we omit the details of the whole construction of the system because the procedure to be applied to the other terms is similar ) .
let @xmath43 be the rate at which @xmath65 grows at the expense of the decrease of @xmath66 . for global contacts @xmath67 .
let @xmath25 be a pair of type @xmath27 at a fixed time @xmath17 such that @xmath68 .
the rate of local infection for this pair is the product of @xmath14 and the fraction of infected individuals in the four nearest neighbors of @xmath59 .
this fraction can be regarded as @xmath69 , where @xmath70 is the probability that a fixed triplet @xmath60 ( @xmath71 being a fixed nearest neighbor of @xmath59 different from @xmath22 ) is in situation @xmath72 , given that @xmath33 .
we have @xmath73 then , using ( [ lebo ] ) and @xmath74 @xmath75 which corresponds to @xmath76 in da . then we take @xmath77 and @xmath78 is incorporated as a substracting term in the fifth equation and as an adding term in the seventh equation of the system ( 6 ) , that we present below .
@xmath79 @xmath80 @xmath81 @xmath82 @xmath83 @xmath84 @xmath85 @xmath86 @xmath87 for better visualization of the dynamics , we presented the initial system of nine differential equations . by using the equalities ( [ l1 ] ) to ( [ l4 ] ) we reduced the system to five ones for the numerical resolution
in this section we compare the results obtained from the stochastic model ( sm ) simulations with the results obtained using the deterministic approximation ( da ) .
we consider l=800 which mimics a city of @xmath36=640,000 inhabitants . for pertussis
we take @xmath88 years ) , @xmath89 days ) and @xmath90 1/day , which are standard parameters for sir description of the disease in the pre - vaccine era @xcite . for rubella
we take @xmath88 years ) , @xmath91 days ) and @xmath92 1/day @xcite .
we consider values of @xmath93 since in ref.@xcite it was observed that for lower values of @xmath1 the probability of establishment and survival of the steady state ( representing the endemic state of pertussis in prevaccine era ) is very low . in the case of rubella , this probability is very low even for @xmath94 . in this work
we are interested in considering values of the parameters that may represent real systems , however , we have included in our study the case @xmath94 for rubella for comparison purposes . in da differential equations
are integrated using euler algorithm with a time step 0.01 days that we have checked to give accurate enough results for the problems studied .
we first study the epidemic spread of the disease for the case @xmath1=0.4 .
we consider the evolution of the system when a single infected individual is introduced in a population of @xmath36 - 1 susceptible individuals . from this initial condition ,
5,000 independent runs were performed for sm with different sets of random numbers .
for da we take * x@xmath95*=@xmath96 as initial condition . in fig.1a
we compare the results for three representative samples of sm simulations and da for pertussis . in fig.1b
the same comparison is shown for rubella . in both cases
da correctly predicts the qualitative spread of the epidemic and also gives a quantitatively good prediction for the maximum fraction of infected individuals reached during the course of the epidemic . in fig.1c and 1d
we compare the result for da with the average of the 5,000 runs of sm simulations .
the discrepancy in the maximum of the fraction of infected individuals in this case is caused because in the average for the sm runs we included the extinctions . at the very beginning of the epidemic spread
there is one infected individual and the total rate of infection is @xmath14 , while the rate of recovery or death of the infected individual is @xmath97 . as @xmath98 for both diseases , da always predicts an initial increase in the fraction of infected individuals . in the sm simulations
there is a positive probability that recovery occurs before infection , and since there is only one infected individual at the beginning , recovery implies extinction . the fraction of extinctions is greater in the case of rubella as @xmath99 is greater as well . at the first steps of the dynamics , the value of @xmath1 is irrelevant as the infected individual can only contact susceptible ones regardless of whether the contacts are local or global
this can be seen in fig.1e and 1f where sm simulations and da predict an exponential increase in the fraction of infected individuals at a rate @xmath100 .
an interesting result predicted by sm simulations ( and very well aproximated by da ) is the approximately exponential growth of the epidemic a few days after its beginning with an exponent that differs from the initial one .
this can be concluded from the linear behavior observed in the curve of fig.1e from 5 to 15 days , and in fig.1f from 10 to 30 days .
exponential fits to the curves in the mentioned ranges give exponents 0.61 and 0.63 for sm and da for pertussis , and 0.267 and 0.272 for sm and da for rubella . in order to characterize this behavior as a function of @xmath1
, we notice that in this exponential regime the time dependence of the fraction of infected individuals may be written as @xmath101 $ ] .
that is to say , the same as in the classical sir model ( with only global contacts ) but with an effective contact rate , @xmath102 , that is different from @xmath14 because of the presence of local contacts .
it is possible to obtain an analytical approximate expression to estimate @xmath102 using da equations as follows . at the beginning of epidemic growth ,
when @xmath103 , we have that @xmath104 , @xmath66 and @xmath105 are also @xmath106 .
so , we drop all the second order terms in @xmath107 and @xmath105 , and also the terms @xmath108 . as @xmath109 ,
we may write the 2nd equation of system ( 6 ) : @xmath110
if @xmath111 presents an exponential behavior , @xmath112 should be proportional to @xmath113 , so , we assume that in this exponential regime : @xmath114 , @xmath115 being a constant . expression ( [ didtap ] ) may now be written as : @xmath116 where @xmath117 plays the same rol as @xmath14 in the exponential growth when only global contacts are present .
so , we define @xmath118 .
taking @xmath114 and replacing it in the 2nd and 5th equations of system ( 6 ) , we may obtain the value of @xmath115 and compute : @xmath119 an interesting point here is that , at this level of approximation , da predicts that @xmath120 only depends on @xmath1 and not on the other parameters characterizing the disease transmission . in fig.2
we plot this relation for da ( from eq.([betaexp ] ) ) and for sm ( obtained from exponential fits for different @xmath1-values ) for pertussis and rubella .
the agreement between da and sm is good , but it becomes worse for lower values of @xmath1 .
however , even for @xmath94 ( when the agreement of sm and da values for @xmath102 is not so good ) @xmath121 accurately verifies the da prediction of being dependent only on @xmath1 for pertussis and rubella .
comparison of sm and da predictions for the exponential growth regime in an epidemic as a function of @xmath1 .
the computation of @xmath120 is performed through expression [ betaexp ] for da .
for sm , @xmath122 is obtained by fitting the exponential regime of the curve obtained for the average fraction of infected individuals as a function of time to @xmath123 $ ] for each case , where @xmath71 is a constant.,width=257,height=188 ] for the sm presented in section [ sectionsm ] and a finite value of @xmath36 , the only fixed point ( stationary equilibrium ) corresponds to the case where all the individuals on the lattice are susceptible . for the stochastic nature of the system ,
sooner or later a fluctuation leads the number of infected individuals to zero and there is no process that produces new infected people if there are none .
but for large values of @xmath36 ( as the one taken in the present study ) the system may fluctuate for a long time around a quasi - stationary state ( qss ) before extinction .
the definition and properties of such a state have been addressed in other contexts from a mathematical point of view @xcite or with empirical approaches @xcite . in the present work
we adopt the empirical strategy developed in ref.@xcite where the qss of the same sm defined in section [ sectionsm ] is studied .
we define a time window ( @xmath124 ) and generate ( for each system studied ) more than 20,000 independent samples keeping those that do not become extinct before @xmath125 .
we compute the averages of susceptible and infected individuals over the surviving samples and observe that in the ( @xmath124 ) interval they remain constant in time within a given precision .
we take @xmath126=20,000 days and @xmath125=40000 days ; with these values the considered averages are independent of the initial conditions ( as @xmath126 is large enough so that correlations in the system may be established ) and enough statistics is obtained to compute averages with a relative precision of 0.01 .
( for details of the definition of the qss and characterization of the fluctuations for the pertussis case see ref.@xcite . ) in the case of da , stationary values are defined as the values of the variables that make all the derivatives in equations ( 6 ) zero .
they were obtained numerically solving the equations from a given initial condition , * x@xmath95 * , until * x*@xmath127 is constant in time with a precision 0.00001 . for * x@xmath95 * we take for @xmath128 and @xmath104 the values corresponding to the stationary solution of the sir model and complete the remainding values of * x@xmath95 * as in the uncorrelated case .
for example , @xmath66 is initialized as @xmath129 because it is twice the probability that a given pair is of type @xmath27 whose value is @xmath130 in the uncorrelated case . in fig.[figstat]a and b
we compare the da stationary values obtained for the fractions of susceptible and infected individuals for pertussis and rubella with the values of these observables averaged over the samples in the qss of sm .
results for both diseases show an excellent agreement between sm and da . in fig.[figstat]c
we make the same comparison for the fraction of @xmath27 pairs on the lattice .
the agreement is remarkable considering that this is a correlated quantity directly involved in the da to sm .
we here study the ability of da to describe the system response to a sudden change in the globality of the contacts represented in the model by the parameter @xmath1 .
we simulate disease transmission at the endemic state in a city where , at a given time , infective contacts among people change and become more local
. such a change could occur because of several circumstances that affect social activities .
for example , in argentina , during the 2009 pandemic flu , the authorities took preventive measures that included suspension of classes at school in july .
the population was advised to stay at home whenever possible and to avoid crowded places .
these measures aimed at reducing the possibility of an epidemic outbreak during the winter .
our purpose here is to simulate the consequence that this sudden reduction in globality of the contacts may have on the disease transmission of other infectious diseases that are at the endemic state .
we perform simulations with sm for parameters corresponding to pertussis in the prevaccine era and @xmath131 .
once the qss is stablished , at a given time ( t=20,000 days ) we change the value of @xmath1 to 0.4 , which corresponds to a 20% reduction in the globality of contacts . in fig.4a
we compare the dynamic evolution of the system for three samples obtained from sm simulations and da result . to obtain the da curve we take the * x * corresponding to the stationary state for @xmath1=0.5 as initial condition and at @xmath132 we change the value of @xmath1 to 0.4 .
we observe that da captures the fall down in the fraction of infected individuals after lowering @xmath1 and the pronounced peaks observed in the following years .
the perturbation at @xmath132 days puts in phase the oscillations for the different samples that become out of phase again from the third peak .
da also accurately reproduces the time elapsed between peaks .
oscilations are exponentially damped out in da until the new stationary value is reached . in the case of sm
, fluctuations may be very different from sample to sample @xcite , but as we have seen in section [ subsecsb ] , once the qss is reached , the fraction of infected individuals oscilates around a value very similar to that predicted by da .
we repeated this study for the case that the system is initially at the qss corresponding to @xmath133 and at @xmath132 @xmath1 is changed to 0.4 . in this case , although the final state is the same , the change in the globality of contacts is 33% . in fig.4b
it can be seen that after the perturbation , the system reaches a deep minimum followed by pronounced maxima much higher than in the previous case when the change in @xmath1 was 20% .
again , da gives a good description of the observed effects . in this case , most of the samples obtained from the sm simulations become extinct when the system is at the first and pronounced minimum .
this is not surprising and could be inferred from the da result that predicts the number of infected individuals to be below 15 in the time interval ( 20,400 , 20,800 days ) for the population considered here .
extinctions in the sm simulations are thus expected , given the fluctuations in the number of infected individuals observed in the dynamic behavior of the system .
it is worth mentioning that a dynamic behavior similar to that observed in fig.4 has been found in a recent study of pertussis transmission performed with an age - structured deterministic model with nine epidemiological classes @xcite . in that work
, the authors proposed that a sudden change in contact rates among individuals could cause a dynamic effect with the presence of deep valleys followed by sharp maxima such as those recently observed in some us states .
here we obtained a similar behavior for the disease transmission as a consequence of having reduced the degree of globality of the contacts in a much more simple epidemiological model .
in this work we propose a deterministic approach ( da ) for an sir - type epidemiological model that includes global and local contacts on a square lattice .
when comparing our da with simulations performed with the stochastic version of the model ( sm ) , we obtained a very good agreement for qualitatively different scenarios of disease transmission and parameters corresponding to pertussis and rubella in the prevaccine era .
it would be very valuable to increase the complexity of the model used in the present work in order to study pertussis transmission in the vaccine era .
the introduction of pertussis vaccination in the model forces one to include new epidemiological classes that account for individuals that have partially ( or totally ) lost their immunity @xcite as it is well known that immunity to pertussis is not lifelong @xcite .
we expect the good description obtained with da for sm to hold if a model with additional epidemiological classes is considered , since the da to sm involves bassically the description of local contacts that has shown to work very well in different dynamical situations .
stochastic simulations for a more complex model could be much more time - consuming and da might be the only possible approach to the problem .
00 r.m . anderson and r.m .
`` infectious diseases of humans : dynamics and control '' ( oxford university press , oxford , 1991 ) .
keeling and p. rohani .
`` modeling infectious diseases in humans and animals '' ( princeton university press , princeton , 2008 ) .
hans heesterbeek , _ et al .
_ science * 347 * , aaa4339 ( 2015 ) .
doi:10.1126/science.aaa4339 w.o .
kermack and a.g .
mckendrick . proc .
a * 115 * , 700 ( 1927 ) .
biosciences * 145 * , 89 ( 1997 ) ; _ ibid . _ * 158 * , 47 ( 1999 ) .
granich , ch.f .
gilks , ch .
dye , k.m . de cock and b.g . williams .
lancet * 373 * , 48 ( 2009 ) g.fabricius , p.bergero , m.ormazabal , a.maltz and d. hozbor .
epidemiology and infection * 141 * , 718 ( 2013 ) .
third berkeley symp . on math .
statist . and prob .
* 4 * ( univ .
of calif . press , 1956 ) , 81 .
p. rohani , d. earn and b.t .
science * 286 * , 968 ( 1999 ) .
j. joo and j. lebowitz .
rev e * 70 * 036114 ( 2004 ) .
j. verdasca , m.m .
telo da gama , a. nunes , n.r .
bernardino , j.m .
pacheco and m.c .
journal of theoretical biology * 233 * , 553 ( 2005 ) .
m simes , m.m .
telo da gama and a. nunes .
interface * 5 * , 555 ( 2008 ) .
m. dottori and g. fabricius .
physica a * 434 * , 25 ( 2015 ) .
j. of comput .
* 22 * , 403 ( 1976 ) .
g. rozhnova and a. nunes .
j. r. soc .
interface * 9 * , 2959 ( 2012 ) .
m. j. keeling , p. rohani and b. t. grenfell .
physica d * 148 * , 317 ( 2001 ) .
darroch and e. seneta .
journal of applied probability * 4 * , 192 ( 1967 ) i. n@xmath134ssel
. j. r. statist .
b * 61 * , 309 ( 1999 ) m. martins de oliveira and r. dickman . phys
e * 71 * , 016129 ( 2005 ) .
j. blanchet , p.w .
glynn and s. zheng .
evolve - a bridge between probability , set oriented numerics , and evolutionary computation ii ( advances in intelligent systems and computing vol . *
175 * ) , 19 ( 2012 ) .
p. pesco , p.bergero , g. fabricius and d. hozbor .
epidemics * 7 * , 13 ( 2014 ) .
a.m. wendelboe , a. van rie , s. salmaso and j.a .
the pediatric infect .
journal * 24 * , s58 ( 2005 ) . | an epidemic model with births and deaths is considered on a two dimensional @xmath0 lattice .
each individual can have global infective contacts according to the standard sir model rules or local infective contacts with its nearest neighbors .
we propose a deterministic approach to this model and verified that there is a good agreement with the stochastic simulations for different situations of the disease transmission and parameters corresponding to pertussis and rubella in the prevaccine era .
x_s keywords : epidemics ; sir ; lattice ; deterministic model ; pair approximation ; pertussis . |
considerable experimental effort is directed towards creating dense samples of cold molecules for precision measurements @xcite , cold chemistry experiments @xcite , quantum information processing @xcite , degenerate quantum gases with dipolar interactions @xcite , etc . established techniques for generating velocity
controlled or trapped cold molecules include the deceleration using time - varying electric @xcite , magnetic @xcite and optical fields @xcite and the velocity filtering of polar molecules out of an effusive source using static or time - varying electric fields @xcite .
alternative routes to producing cold molecules have been demonstrated utilizing the kinematics in elastically or reactively colliding molecular beams @xcite .
a more general and conceptionaly simple approach to producing slow beams of cold molecules is translating a supersonic jet to low longitudinal velocities by means of a rapidly counter - rotating nozzle .
this technique was demonstrated by gupta and hershbach @xcite and recently improved in our group @xcite . using this technique
we have demonstrated the production of dense beams of various atomic and molecular species with tunable velocity ranging from thousand of m / s down to @xmath3 m/s .
since the molecules are not confined to any external potential this technique suffers from the drawback that beam density rapidly decay due to transverse beam expansion during beam propagation from the nozzle to the interaction region .
this effect can be partly compensated by installing additional guiding elements such as electrostatic quadrupole guides @xcite .
however , such extended objects can not easily be brought close to either the nozzle or the interaction region .
besides , electrostatic quadrupole or higher multipole guides produce confining potentials only for molecules in low - field - seeking rotational states . as an alternative guide geometry
we use a thin charged wire in this work , which is spanned through the whole molecular beam apparatus from the rotating nozzle up to the quadrupole mass spectrometer ( qms ) that we use as a detector .
this concept was proposed by sekatskii @xcite and experimentally demonstrated by loesch @xcite .
recently , guiding of polar molecules in surface - based electrostatic potentials has been demonstrated @xcite . in our arrangement
we capture the trappable molecules already in the supersonic expansion region and guide them all the way close to the detector . by comparison with classical trajectory simulations
we find that nd@xmath0 as well as chf@xmath0 molecules are guided in high - field - seeking states that feature both a linear and a quadratic stark effect .
an enhancement of the beam density on the beam axis due to the guiding effect of up to a factor 7 is measured for decelerated beams with velocities @xmath4 m/s .
the guiding concept is characterized in terms of optimum geometries and in terms of the applicability to slow molecules in the ground state .
the experimental arrangement is identical to the one reported earlier @xcite with the difference that the quadrupole guide placed between skimmer and qms detector is replaced by a thin wire , as schematically shown in fig .
[ fig : setup ] .
we use a gold - plated beryllium - copper ( becu ) wire with a radius @xmath5 m .
the total length of the wire from the jet expansion region up to the suspension in front of the detector amounts to @xmath6 cm .
the wire is attached at one end to a @xmath7-translation stage placed inside the source chamber such that the wire can be aligned transversally with respect to the beam axis to pass close by the nozzle ( @xmath8 m in diameter ) and concentrically through the skimmer that has an aperture of 1 mm in diameter .
further downstream the wire passes through an intermediate chamber that serves as a differential pumping section and is suspended inside the detector chamber by a perpendicularly spanned second isolated copper ( cu ) wire @xmath9 mm in front of the crossed beam ionizer of the qms .
the becu wire is held under tension by a weight of @xmath10 g such that the wire curvature at the point of suspension can be assumed to follow the radius of the cu wire ( @xmath11 m ) .
owing to the resulting kink in the electric field a large fraction of the guided molecules are output coupled and enter the detection region of the ionizer of the qms .
alternatively , we have installed a home - built detection unit consisting of a heated filament radially displaced from the wire at a distance of @xmath12 mm and a channeltron detector on the opposite side at a distance of @xmath13 mm from the wire . in this way
, the molecules can be ionized directly in the guiding potential of the wire . both the filament and the channeltron were biased to constant voltages with respect to the wire in order to ensure that the detection efficiency is independent of the wire voltage .
the measured signals followed the same trends as the signals recorded using the qms .
however , a high background signal level from ionized background gas and fluctuating signal offsets caused the results to be less indicative than those measured with the qms .
the guide wire is electrically connected to a high voltage feedthrough inside the source chamber .
the maximum applicable voltage is limited to @xmath14kv with respect to the surrounding vacuum chambers by sparkover from the wire to either the skimmer or to the bored titanium ferrule that forms the nozzle at the tip of the rotor arm . sparking to the skimmer
can be prevented by setting the skimmer to high voltage .
however , the guiding efficiency was found to diminish considerably in this case .
therefore all measurements presented here are performed with the skimmer set to ground potential and with a minimum distance @xmath15 between the center of the nozzle orifice and the wire center @xmath16 m to avoid sparking .
this , in turn , means that the nozzle is displaced away from the beam axis by the same distance since the wire is coaxially with the beam axis . as a result ,
the peak density of the transmitted beam without guide voltage is reduced by about a factor 4 as compared to the situation when no wire is installed and when the nozzle position is optimized ( @xmath17 ) .
molecules in rotational states @xmath18 ( quadratic stark effect ) and @xmath19 ( linear stark effect ) .
a wire diameter of @xmath11 m and an applied voltage @xmath20kv is assumed.,title="fig:",scaledwidth=53.0% ] in the present study we use nd@xmath0 as test molecules to demonstrate the potential of the charged wire setup for guiding slow polar molecules .
since the energy difference between the vibronic ground state @xmath21 and the lowest rotationally excited state @xmath22 is about @xmath23@xmath24 mostly these two states are thermally populated at the estimated low rotational temperatures in the jet of @xmath25k .
nd@xmath0 is particularly well suited for guiding and deceleration experiments using electric fields due to the strong linear stark effect of both the high - field - seeking state @xmath26 and the low - field - seeking state @xmath27 correlating to @xmath22 rotational level which arises from the small inversion splitting in the absence of electric fields @xcite .
the rovibronic ground state @xmath28 is high - field - seeking and features quadratic stark effect .
when placing a nd@xmath0 molecule in the electric field of a cylindrical capacitor at a distance @xmath29 from the center , @xmath30 created by the charged wire of radius @xmath31 inside a vacuum apparatus that is assumed to have cylindrical symmetry with inner radius @xmath32 mm we obtain a transverse trapping potential @xmath33 that scales as @xmath34 for the linear stark state @xmath19 and as @xmath35 for the ground state @xmath28 , see fig .
[ fig : potential ] .
the shaded area in the center of the figure indicates the range excluded by the wire .
while the potential @xmath33 is attractive for both states in the shown range of distances @xmath33 is much deeper ( @xmath36k@xmath37k@xmath38 ) and more extended for the @xmath19-state as compared to the ground state @xmath39 .
thus , more efficient guiding of molecules in the state @xmath19 may be expected .
moreover , molecules in @xmath19 follow stable elliptical trajectories in analogy to the kepler problem , whereas the @xmath40-potential experienced by the ground state does not sustain stable trajectories . in this case
the trajectories either spiral down to hit the wire after a few revolutions or quickly escape away from the wire , as discussed in detail by denschlag _
et al . _
therefore a guiding effect may be expected only for a certain range of velocities in which the molecules are transiently bound to the wire and reach the detector before crashing into the wire or being expelled away from it .
this situation is discussed in more detail at the end of the following section .
note that in the experiment performed by loesch @xcite using alkali - halide molecules even the rotational ground states were subjected to a @xmath41-potential due to `` brute force '' orientation as a result of much larger dipole moments and higher fields .
molecules detected behind the wire guide for wire voltage on ( 2kv ) and off ( 0v ) .
the dashed line represents a model of the free jet expansion based on gaussian transverse velocity distributions .
( b ) relative enhancement of the peak density due to guiding by the charged wire .
the solid lines show the result of trajectory simulations ( see text).,title="fig:",scaledwidth=45.0% ] the guiding efficiency characterized in this section is determined by comparing the detected beam density when switching on the wire voltage @xmath42 against the density of the unguided beam when the wire potential is set to ground ( @xmath43 ) .
such a measurement using nd@xmath0 ( @xmath44 ) seeded in krypton ( @xmath45 ) for variable beam velocities is depicted in fig .
[ fig : velocity ] as open ( @xmath43 ) and as filled ( @xmath46kv ) symbols .
the drop of the absolute density of molecules ( fig .
[ fig : velocity ] ( a ) ) with decreasing beam velocity is due to the transverse and longitudinal dispersion of the beam .
this behavior is well reproduced by simple considerations based on the expansion of a bunch of molecules that has gaussian velocity distributions in longitudinal and transverse directions ( dotted line ) @xcite .
clearly , this drop can be mostly compensated by the wire guide when applying high voltage @xmath42 to the wire ( filled symbols ) .
note that guiding is particularly efficient at low velocities in proportion to the unguided beam .
the relative density increase due to guiding we call enhancement , which is illustrated in fig .
[ fig : velocity ] ( b ) .
thus , at beam velocities around @xmath47 m/s we measure an increased beam density by up to a factor 7 .
a second measurement was done after the wire was replaced and newly aligned with respect to the nozzle , skimmer , and qms detector .
the two measurements are in good agreement at high beam velocities , whereas at low velocities there are slight deviations probably due to a slightly different distance @xmath15 between the nozzle and the wire .
the solid lines depict the result of classical trajectory simulations which account for the thermal population of rotational levels @xmath21 and @xmath22 at @xmath48k as well as for nuclear spin statistics @xcite .
initial values for the spatial and velocity coordinates are determined by a monte - carlo method according to gaussian velocity distributions and a longitudinal spatial slit opening function of the arrangement determined from fits to the experimental time of flight measurements of the unguided beam at various beam velocities @xcite .
one data point reflects the average of 5000 trajectories bound to the wire potential for each of the 10 rotational states correlating to @xmath21 and @xmath22 .
the two solid lines result from the same simulation and merely reflect the uncertainty in the effective cross section of the ionizer of the qms .
the upper and lower lines correspond to estimated circular cross sections with radius @xmath49 mm and @xmath50 mm , respectively .
irrespective of this uncertainty , the simulation tends to overestimate the guiding effect .
possible detrimental effects in the experiment that are not accounted for in the simulation include modulations of the guide potential along the beam axis due to drastically changing outer radius @xmath51 of the assumed cylindrical capacitor configuration in particular when passing through the skimmer as well as the unknown efficiency of the output coupling process at the wire bend in front of the detector .
according to the simulations , the enhancement of the guiding effect will rise steeply to exceed a factor of 10 as the speed of the molecules is further reduced below @xmath52 m/s .
unfortunately we did not reach this velocity range with the present setup due to the limited maximum rotor frequency ( @xmath53hz ) @xcite .
molecules in the high- field - seeking states @xmath18 and @xmath19 due to guiding by the charged wire as a function of wire voltage for a beam velocity @xmath54 m/s .
( b ) enhancement of the measured nd@xmath0 peak density for various beam velocities ( symbols ) .
the calculations show the average of trajectory simulations for the low energy stark states correlating to @xmath21 and @xmath22 and for two effective detector cross sections with radius @xmath49 mm and @xmath50 mm ( upper and lower lines , respectively).,title="fig:",scaledwidth=45.0% ] the measured enhancement of the molecule density due to guiding as a function of the wire voltage @xmath42 for different beam velocities @xmath55 is depicted as symbols in fig . [ fig : voltage ] ( b ) . as @xmath42 increases
, the number of guided molecules grows monotonically , leading to a relative enhancement factor of up to 7 at @xmath54 m/s . at high voltages
the enhancement slightly saturates .
this is due to the dominating contribution of the linear stark state @xmath19 which saturates at voltages @xmath56kv , shown in fig .
[ fig : voltage ] ( a ) . in this voltage range
nearly all molecules in the guidable @xmath19-state are actually captured by the confining wire potential .
the quadratic stark state @xmath28 contributes much less in spite of its larger relative population as a consequence of the much weaker guide potential ( see fig .
[ fig : potential ] ) . the solid lines in fig .
[ fig : voltage ] ( b ) depict the results of trajectory simulations for @xmath54 m/s when averaging over all stark states correlating to @xmath21 and @xmath22 .
again , @xmath49 mm and @xmath50 mm are assumed ( upper and lower lines , respectively ) .
although the trend of a monotonically increasing guiding efficiency as a function of @xmath42 is well reproduced , saturation sets in earlier in the simulation than in the experiment .
this is presumably due to imperfections in the experimental setup as mentioned above .
peak density due to guiding by the charged wire as a function of the distance @xmath15 between the center of the nozzle orifice and the center of the wire .
the beam velocity and wire voltage are fixed to @xmath57 m/s and @xmath20kv , respectively.,title="fig:",scaledwidth=45.0% ] the dependence of the enhancement factor on the distance @xmath15 between the center of the charged wire and the nozzle orifice is shown in fig .
[ fig : distance ] . in this measurement ,
the wire voltage is set to a moderate value @xmath20kv to avoid sparking , the beam velocity is held constant at @xmath57 m/s , and the distance between the wire and nozzle is varied from @xmath58 m down to the minimum distance of about @xmath59 m .
this is achieved by shifting the whole baseplate that supports the rotating nozzle setup with respect to the position of the charged wire which is kept fixed . in this way
, the geometry of the guiding field and of the detector is maintained unchanged during the measurement .
note that the nozzle to wire distance is obtained by viewing the nozzle position through a telescope along the wire axis .
therefore the value of @xmath15 must be regarded as an estimate with an uncertainty of about @xmath11 m . clearly , as the nozzle is brought as close as possible to the charged wire the guiding efficiency slightly increases up to a factor of about 2 .
this is due to the increased depth of the guide potential close to the wire surface such that a larger fraction of molecules with non - vanishing transverse velocity components still remain bound to the wire potential . at the same time , the absolute signal of the unguided beam is increased by a factor of @xmath60 since the nozzle is moved towards the beam axis , as mentioned above .
the trajectory simulations for @xmath49 mm and @xmath50 mm ( solid lines ) reproduce this trend , but slightly overestimate the enhancement . by further decreasing the distance ,
the enhancement could be increased by an additional factor of about 2 according to the simulation .
peak density due to guiding by the charged wire as a function of the wire radius for @xmath61 m and @xmath62 m ( upper and lower lines , respectively ) . the beam velocity and the wire voltage
are fixed to @xmath57 m/s and @xmath20v , respectively.,title="fig:",scaledwidth=45.0% ] since the electric field which is generated by applying a certain voltage @xmath42 to the wire depends on the wire radius @xmath31 according to eq .
[ eq : efield ] , the choice of @xmath31 crucially influences the guide efficiency . in order to determine the optimum value of @xmath31 for the given conditions in our experiment we have performed simulations of the density enhancement for nd@xmath0 at @xmath20kv and @xmath57 m/s for various values of @xmath31 ranging from @xmath63 m down to about @xmath64 m ( fig .
[ fig : radius ] ) .
the two lines for either @xmath28 and @xmath19-states represent the results for @xmath61 m and @xmath62 m . as @xmath31
is reduced down to @xmath65 m the guiding efficiency first rises in spite of the decreasing electric field ( eq .
[ eq : efield ] ) due to an increased number of trajectories that circle around the wire without hitting it . a clear maximum occurs around @xmath66 m corresponding to an optimum guiding efficiency for molecules in the @xmath19-state . for smaller values of @xmath67 m ,
the guiding efficiency goes down again due to a sharp drop of the electric field .
however , for practical reasons such as stability against tensile stress and electric sparkover we chose the becu wire with @xmath11 m diameter .
molecules in the ground state @xmath28 as a function of velocity down to @xmath68 m/s ( solid line ) .
the dashed line indicates the number of molecules that pitch into the wire on their way from the nozzle to the detector in proportion to the number of detected molecules.,title="fig:",scaledwidth=49.0% ] the simulations presented so far show that nd@xmath0 in both the linear @xmath19 and the quadratic stark state @xmath28 can be guided along the charged wire in spite of the trajectories in a @xmath40-potential being unstable .
therefore the question arises , under which conditions this instability becomes apparent as reduced guiding efficiency . in order to illustrate this effect we simulate the enhancement of the density of nd@xmath0 molecules in the ground state @xmath28 for beam velocities reaching down to @xmath68 m/s ( fig .
[ fig : slow ] ) . the wire voltage @xmath20kv , the nozzle position @xmath69 m and the wire radius
are held constant .
as the velocity is decreased from @xmath70 m/s down to @xmath71 m/s the density enhancement first rises similarly to the state - averaged density shown in fig .
[ fig : velocity ] .
however , at lower beam velocities @xmath72 m/s the guiding effect sharply breaks down due to an increasing fraction of molecules that pitch into the wire ( dashed line ) .
possibly the position of the maximum shifts down to lower velocities as the wire radius is further reduced .
thus we conclude that while the charged wire may be useful for guiding beams of ground state molecules at moderate velocities in the range of hundreds of m / s , is is inapplicable for molecules decelerated well below @xmath73 m/s .
in conclusion , we have demonstrated that a simple thin charged wire placed along the axis of a beam of decelerated polar @xmath74 molecules can be used for enhancing the molecule density even at the exit of the wire guide by up to a factor 7 .
@xmath74 and @xmath75 , that was also used , show similar guiding behavior for comparable experimental parameters .
classical trajectory simulations are found to be in reasonable quantitative agreement with the measurements and to reproduce well the systematic trends .
while rotational states with linear stark effect are guided along the wire in stable kepler orbits the quadratic ground state is only transiently enhanced at moderate beam velocities that ensue short flight times .
possible technical improvements of the present setup include a new nozzle design with a thinned end cap of the titanium ferrule in order to reduce the distance between the nozzle orifice and the wire surface down to @xmath76 m . in this way both the guiding efficiency with respect to @xmath43 is enhanced and the absolute beam intensity is increased due to better beam alignment . furthermore , the wire voltage could be increased up to @xmath77kv by insulating the tip of the rotor against the vacuum apparatus to suppress sparking . in total , higher peak densities by about a factor @xmath78 as compared to the reported guided beam densities should be attainable .
several applications of the presented concept are conceivable . since only polar molecules experience the electrostatic guiding force in contrast to the rare - gas atoms in the carrier gas of the seeded expansion such a guiding element can be used to separate the molecules out of the carrier gas beam , in particular if an additional bend section is incorporated .
the implementation of a wire guide into novel intense sources of cold molecules that rely on buffer gas cooling in combination with expansion cooling may be promising @xcite .
such an approach may be particularly advantageous in applications where decelerated cold molecules are to be injected into micro - structured devices such as micro - cavities or atom chips . | we demonstrate experimentally the guiding of cold and slow nd@xmath0 molecules along a thin charged wire over a distance of @xmath1 m through an entire molecular beam apparatus .
trajectory simulations confirm that both linear and quadratic high - field - seeking stark states can be efficiently guided from the beam source up to the detector .
a density enhancement up to a factor 7 is reached for decelerated beams with velocities ranging down to @xmath2 m/s generated by the rotating nozzle technique . |
@xcite presented a search for infrared counterparts to galactic supernova remnants ( snrs ) that were sampled by the galactic legacy infrared mid - plane survey extraordinaire ( glimpse ) legacy survey @xcite .
the glimpse survey imaged the inner galactic plane specifically the galactic latitude and longitude ranges of @xmath10@xmath11@xmath10@xmath121@xmath13 and 10@xmath13@xmath14@xmath15@xmath1665@xmath13 , respectively at four wavelength bands ( 3.6 , 4.5 , 5.8 and 8.0 @xmath17 m ) with the infrared array camera ( irac ) @xcite aboard the
_ spitzer space telescope _
the results presented by @xcite were very intriguing : of the 95 snrs known ( at the time ) that were sampled by glimpse , eighteen snrs were detected ( many for the first time ) and the infrared emission from these sources typically coincided ( at least partially ) with known radio structures . @xcite also presented simple models to generate template colors in the irac bands for the primary emission mechanisms expected from snrs .
these primary emission mechanisms include the following : ( i ) _ molecular _ line colors that were estimated using a three temperature component h@xmath18 excitation model based on numerous h@xmath18 lines detected toward ic 443 @xcite , ( ii ) _ ionized _ gas line colors that were estimated based on hydrogen recombination as well as the dominant fine - structure lines of iron and argon ( assuming solar abundances ) and ( iii ) interstellar medium ( ism ) polycyclic aromatic hydrocarbon ( pah ) colors that were estimated using a spectrum of the reflection nebula ngc 7023 . in their analysis
, @xcite found that the spectra of nine snrs , three snrs and four snrs were dominated by molecular lines , fine - structure lines and pahs , respectively .
the spectra of the remaining snrs indicated a mixture of ionic and molecular shocks , suggesting a combination of shock types .
interestingly , many of the snrs detected with irac also appear to be interacting with relatively dense gas .
evidence for such interactions is the excess emission detected from these snrs at a wavelength of 4.5 @xmath17 m : such an excess is likely produced by h@xmath18 and co line emission and such line emission is associated with molecular shocks .
furthermore , follow - up spectroscopy observations conducted with the infrared spectrograph ( irs ) aboard _ spitzer _ @xcite revealed h@xmath18 emission from all of the snrs with irac by the survey conducted by @xcite , suggesting that each of these snrs are interacting with molecular clouds @xcite . we shall refer to these snrs as molecular snrs for the remainder of the paper .
several molecular snrs are detected in the x - ray ( see @xcite for a review of x - ray emission from snrs ) and all of them belong to the class of sources known as mixed - morphology snrs ( mm snrs ) .
the distinguishing characteristics of mm snrs are a center - filled thermal x - ray morphology paired with a well - defined radio shell @xcite .
examples of well - known mm snrs include w28 , w44 and 3c 391 : each of these snrs was observed in the infrared with the long wavelength spectrometer ( lws ; * ? ? ?
* ) aboard the _ infrared space observatory _
( iso ) @xcite and h@xmath18 emission was detected from all three snrs @xcite .
many of the molecular snrs detected by @xcite have not been well - studied in the x - ray , probably because the galactic column densities toward these sources are elevated ( @xmath19@xmath20 @xmath3 10@xmath21 @xmath22 ) , causing them to appear to be faint in the x - ray . in this paper
we describe the x - ray properties of five molecular snrs detected by @xcite : kes 17 ( [email protected] ) , [email protected] , [email protected] , ctb 37a ( [email protected] ) and [email protected] . for this study
, we use data from archival observations made with @xmath23 , @xmath24 and @xmath25 .
these snrs had been studied almost exclusively in the radio before the infrared observations made with _
spitzer _ : they have not been detected in the optical most likely because of the high extinction toward these sources .
we note that three of these snrs ( kes 17 , [email protected] and ctb 37a ) have been the subjects of pointed _ chandra _ , _ xmm - newton _ and _ suzaku _ observations that have been described and analyzed in separate publications @xcite .
serendipitous observations made of [email protected] , ctb 37a and [email protected] with _
asca _ have also been described by @xcite . in this paper , we present new analyses of the _ asca _ observations of kes 17 and [email protected] ( including the first reported detection of x - ray emission from this snr ) and a spatially - resolved spectroscopic analysis of the _ chandra _ observation of ctb 37a .
we will compare the results of our analyses with the results for all of the snrs with the previously - published papers mentioned above .
the organization of this paper may be described as follows : the observations as conducted by each observatory and the subsequent data analyses are described in section [ obssection ] . in section [ propsxrayimagespecsection ]
we present a description of the five snrs in our sample and the x - ray results both imaging and spectroscopic for each snr .
a summary of the plasma conditions of these sources as inferred from these x - ray observations is presented in section [ plasmasection ] . in section [ overionizationsection ]
we discuss the phenomenon of overionization in mixed - morphology snrs and comment on its applicability to the snrs in our sample . in section [ comparisonsection ]
we compare the properties of the snrs in our sample and discuss the implications for snr evolution .
finally , the conclusions reached by this work are presented in section [ conclusionsection ] . in table
[ snrpropstable ] we list general properties including ra ( j2000.0 ) , dec ( j2000.0 ) , angular diameter , distance , physical size , column density @xmath19@xmath20 and optical extinction @xmath26@xmath27 of the five snrs considered in the present paper .
all five of the snrs studied in the present paper were sampled ( either directly by pointed observations or indirectly by pointed observations that targeted other sources but included the snr within the field of view ) by the advanced satellite for cosmology and astrophysics ( _ asca _
@xcite ) . the instruments carried by _
asca _ included two solid - state imaging spectrometers denoted as sis0 and sis1 @xcite and two gas imaging spectrometers denoted as gis2 and gis3 @xcite : these instruments were located at the focal planes of four thin - foil x - ray telescopes ( xrt @xcite ) .
a single gis unit sampled a field of view @xmath350@xmath28 in diameter while a single sis unit sampled a field of view approximately 44@xmath28@xmath2944@xmath28 in size .
the nominal full - width at half - maximum ( fwhm ) angular resolution of both the gis and the sis at 1 kev was approximately 1 arcminute .
data for all of the _ asca _ observations considered in the present paper were extracted from the archive located at the high energy astrophysics science archive research center ( heasarc ) and reduced using standard tools in the ftools package . however , only the observations of the kes 17 , [email protected] and [email protected] yielded enough counts from the particular snrs for detailed spectral fitting . in the cases of the snrs ctb 37a and
[email protected] which appear adjacent to each other in the sky , these particular sources were located too close to the edges of the gis fields of view to extract useful spectra . in addition , as noted by @xcite , contaminating stray flux from bright x - ray sources located out of the field of view and south of these two snrs made it difficult to isolate their x - ray emission ( see section [ xmmg348section ] for more details , particularly about our attempt to extract x - ray spectra for [email protected] from the _ xmm - newton _ observation , which also contained contaminating stray x - ray flux ) .
the angular extents of the three remaining snrs for which we extracted _ asca _ spectra ( that is , kes 17 , [email protected] and [email protected] ) were compact enough to fit within the fields of view of the gis2 and the gis3 : in the case of kes 17 , the snr was also compact enough to fit entirely within the fields of view of the sis0 and the sis1 .
therefore , we also extracted sis spectra of this snr as well .
unfortunately the signal - to - noise of the extracted sis1 source spectrum of kes 17 was too poor for any detailed spectral fitting , so we do not consider it further in this paper .
the data reduction was accomplished using another program available from the heasarc site , xselect ( version 2.4b ) : standard rev2 screening was applied to both the gis and sis datasets .
we used xselect to extract gis2 and gis3 spectra for kes 17 , [email protected] and [email protected] : in the case of kes 17 , we also used xselect to extract an sis0 spectrum as well . for spectral analysis ,
the standard gis2 and gis3 response matrix files ( rmfs ) were used while the ftool sisrmg was used to generate an rmf for the sis0 spectrum of kes 17 .
lastly , the ftool ascaarf was used to create ancillary response files ( arfs ) for each extracted gis and sis spectra . in table
[ ascaobssummarytable ] , we present a summary of the _ asca _ observations made of the snrs that we consider in this paper . while all five of the snrs considered in the present paper were sampled ( either directly by pointed observations or indirectly by pointed observations that targeted other sources but included the snr within the field of view ) by the _ xmm - newton _
observatory @xcite , in the cases of two snrs [email protected] and [email protected] the source was located so far from the centers of the fields of view and the exposure times of the observations were so short that no useful spectral data could be extracted for these snrs by these observations .
we therefore only consider the observations made by _ xmm - newton _ of the other three snrs in our study , namely kes 17 , ctb 37a and [email protected] .
the main science instruments aboard _ xmm - newton _ are the pn - ccd camera @xcite and the multi - object spectrometer cameras @xcite : we will refer to these instruments as the pn , mos1 and mos2 for the remainder of the paper and note that both of these instruments comprise the european photon imaging camera ( epic ) .
the fields of view of the pn , mos1 and mos2 cameras are approximately 30 arcminutes and the nominal angular fwhm at 1 kev of all three cameras is @xmath36@xmath30 .
data for the observations was downloaded from the on - line _ xmm - newton _ data archive .
the observations were reduced using standard tools in the science analysis system ( sas @xcite ) software package version 11.0.0 .
the tool emchain was used to process data from the observations made by the mos1 and the mos2 cameras and the tool mos - filter was used to filter out background flares during the observation and to identify good time intervals .
similarly , the tool epchain was used to process data from the observations made by the pn camera and the tool pn - filter was used to identify background flares during the observation and to identify good time intervals . to extract mos1 , mos2 and pn spectra for both source and background regions ,
the tool evselect was used and the extracted spectra were grouped to a minimum of 15 counts per channel .
subsequently , the tool backscale was run to compute the proper value for the backscal keyword in the extracted spectra .
finally , the tools rmfgen and arfgen were used to generate the rmfs and arfs ( respectively ) for the spectra which are necessary for spectral fitting . in table
[ xmmobssummarytable ] , we present a summary of the _ xmm - newton _ observations made of the snrs that we consider in this paper .
one of the snrs studied in the present paper
ctb 37a was the target of a pointed observation made with the _ chandra _ x - ray observatory @xcite : the corresponding obsid of this observation is 6721 and data for this observation was downloaded from the on - line _ chandra _ data archive .
the observation of ctb 37a was conducted in very faint mode on 2006 october 7 with the advanced ccd imaging spectrometer ( acis ) at a focal plane temperature of @xmath1120@xmath13 c. the entire angular extent of the x - ray emitting plasma associated with snr fit within the field of view of the acis - i array and the geometric center of this angular extent aligned roughly with the aimpoint of the observation .
properties of the acis chips ( including the acis - i array ) have been described by @xcite : to summarize , the acis - i array is composed of four front - illuminated ccd chips arranged in a 2@xmath292 array . each chip has a field of view of 8.3@xmath31 @xmath29 8.3@xmath31 : thus , the entire array has a field of view of 17@xmath31 @xmath29 17@xmath31 .
the chips are sensitive to photons with energies ranging from 0.4 to 10.0 kev and the full - width at half - maximum ( fwhm ) angular resolution of each chip at 1 kev is 1@xmath32 .
the maximum effective collecting area of each chip is 525 @xmath33 : this maximum corresponds to photons with energies of 1.5 kev .
finally , the spectral resolution of each chip at 1 kev is 56 .
the dataset from this observation was reduced using the _ chandra _ interactive analysis of observations ( ciao ) software package version 4.3 ( caldb version 4.4.1 ) .
standard processing was applied to this dataset : to generate a new event file , the ciao tool @xmath34 was used .
this tool applies the latest temperature - dependent charge transfer inefficiency ( cti ) correction , the latest time - dependent gain adjustment as well as the latest gain map : @xmath34 also flags bad pixels and applies background cleaning for datasets created in very faint mode , which is the case of the observation considered here .
the new event file was further filtered for bad grades ( that is , events which had grade values of 1 , 5 or 7 were excluded ) and proper status columns : in addition , the good time interval supplied by the pipeline was also applied .
finally , the event file was also screened for background flare activity using the ciao tool chips : background flares did not affect the observation at a significant level .
the resulting total effective exposure time of the observation was 19.7 ks : in table [ chandraobstable ] we present a summary of the parameters of this observation .
the ciao tool specextract was used to extract spectra from different regions of the x - ray emitting plasma associated with ctb 37a ( along with selected background regions located outside of the x - ray emitting plasma ) .
in addition to extracting spectra for source and background regions , specextract also creates weighted arfs and rmfs by initiating the ciao tools mkwarf and mkrmf , respectively .
finally , specextract also initiates the ciao tool dmgroup to group the source and background spectra : in the case of our spectral analysis , the spectra were all grouped to a minimum of 15 counts per channel .
we used the ciao wavelet detection routine @xmath35 @xcite to identify unrelated discrete x - ray sources detected along the line of sight toward ctb 37a by the _
chandra _ observation : such sources may contribute undesired contaminating flux that interferes with spectral analyses of the snr itself . when regions of spectral extraction were defined , care was taken to exclude flux from the positions of these detected sources so that the analyzed spectra more faithfully represented the true emission from the snr . the locations of these source and background regions as well as the analysis of these spectra are presented in section [ ctb37axraysubsection ] .
we used _ spitzer _
irac and mips images from @xcite and @xcite , respectively , of the five snrs considered in this paper to compare the infrared morphologies of these sources with their x - ray and radio morphologies . for details on the irac and mips observations of these snrs and the subsequent data processing and analyses of the images ,
the reader is referred to those papers . as part of our study
, we also included radio maps of each snr that were made at a frequency of 843 mhz by the molonglo observatory synthesis telescope ( most , * ? ? ?
* ; * ? ? ?
* ) : these maps have been consolidated into the most supernova remnant catalogue ( msc ) .
we obtained publicly - available reduced radio maps for each snr from the mgps web page for the purposes of the present research .
the msc contains known southern galactic snrs located within the area 245@xmath13 @xmath36 @xmath37 @xmath36 355@xmath13 , @xmath10@xmath11@xmath10 @xmath36 1.5@xmath13 that were observed with the most during a survey conducted at the frequency of 843 mhz of the southern galactic plane @xcite .
the angular resolution attained by these observations was 43@xmath30 @xmath29 43@xmath30 csc @xmath38 , the sensitivity was approximately 1 - 2 mjy / beam ( 1 @xmath39 ) and the positional accuracy was 1 - 2@xmath30 . for more information about most , mgps and these observations
the reader is referred to the above references .
we now present an analysis of the x - ray properties of each snr .
all spectra were analyzed using the _ xspec _ software package @xcite version 12.6.0q .
kes 17 ( [email protected] ) was first identified as a discrete radio source by the survey conducted by @xcite at a frequency of 408 mhz with the molonglo cross telescope .
it was later classified as a radio snr by @xcite based on its measured non - thermal radio spectral index ( @xmath40=-0.77 to be defined as @xmath41@xmath42 @xmath43 @xmath44@xmath45 . ] ) .
an 843 mhz map of this snr presented by @xcite based on observations made with the most reveal an incomplete shell - like morphology with bright southern and northwestern rims : a gap in the radio shell is seen toward the northeastern boundary . @xcite measured the angular extent of kes 17 to be 8 arcminutes and reported an integrated flux density at 843 mhz for the snr of 18 jy .
@xcite reported on the detection of hydroxyl ( oh ) maser emission at 1720 mhz from kes 17 based on observations made with the parkes 64-meter telescope .
based on hi absorption measurements along the line of sight to this source , @xcite estimated a minimum distance to kes 17 of 9.7 kpc : we will adopt this distance to this snr for the remainder of this paper .
previously published analyses of x - ray observations of kes 17 have been presented by @xcite and @xcite ( who examined _ xmm - newton _ and _ suzaku _ observations , respectively ) : we compare our analyses with the results from those two papers in section [ kes17discusssubsection ] . finally , @xcite reported the detection of @xmath46-ray emission from kes 17 based on observations made by the large area telescope ( lat ) aboard the @xmath47 @xmath46-ray space telescope @xcite . in figure [ kes17_asca_gis2gis3_most ] we present an _
gis2 image of kes 17 over the energy range from 0.7 to 10 kev .
kes 17 was offset by approximately seven arcminutes from the center of the field of view of the gis detectors when the observation was conducted .
radio emission ( as detected by the most at 843 mhz ) is plotted as contours overlaying the x - ray emission . the observed emission
is broadly contained within the radio shell but the poor angular resolution of the x - ray observation prevents a detailed description of the morphology of the x - ray emission . in figure [ kes17_asca_sis_most ]
we present an _ asca _ sis0 image of kes 17 ( again , the radio emission from the snr is shown with contours ) : in the gis and sis images , it is clear to see that the x - ray emission is contained within the radio shell and that the bright radio rims ( as seen toward the southern and western portions of the snr ) are x - ray faint .
we extracted gis2 , gis3 and sis0 spectra for kes 17 using the extraction regions indicated with blue circles in figures [ kes17_asca_gis2gis3_most ] and [ kes17_asca_sis_most ] .
for the gis2 and gis3 , we used a source region approximately 5.5 arcmin radius and a background region approximately 9 arcmin in radius from a portion of the field of view away from the location of kes 17 itself . for the sis0
, we extracted a source spectrum from a region approximately 5.5 arcmin in radius : because this source region filled virtually the entire field of view of the sis0 chip , we extracted a background spectrum from a sis0 blank sky observation .
we fit the extracted source spectra with the thermal models apec ( astrophysical plasma emission code
see * ? ? ?
* ) and nei ( non - equilibrium ionization see * ? ? ?
* ) : satisfactory fits were obtained using these models . because a statistically acceptable fit was obtained with the apec model ( which describes a plasma in collisional ionization equilibrium
cie ) and because the ionization timescale parameter @xmath48 derived for the nei fit is rather high ( @xmath48 @xmath3 10@xmath49 s @xmath50 with lower limits of approximately @xmath48 @xmath3 10@xmath51 - 10@xmath52 s @xmath50 ) we conclude that the x - ray - emitting plasma associated with kes 17 is in cie . for the spectral fits for all datasets considered in this paper , the photoelectric absorption model phabs
will be used to account for galactic photoelectric absorption along the lines of sight to the snrs .
the quality of the fits ( as measured by the @xmath53@xmath54 statistic ) were improved in both cases if we allowed the abundance parameter to vary : in each case , the derived abundance was sub - solar ( approximately 0.2 ) . to investigate the possibility of variations in the abundances of particular elements
, we also fitted the extracted spectra with the thermal model vapec this model has as free parameters the abundances of particular elements .
we generated two acceptable fits with the vapec model : in the first fit , we found marginal evidence for enhancement in the abundance of magnesium ( mg=2.38@xmath55 all quoted errors in the present paper for fit parameters correspond to 90% confidence levels ) while the abundances of other elements remained frozen to solar values . in the second fit
, we found evidence for subsolar abundances for silicon and sulfur ( si=0.47@xmath56 and s=0.21(@xmath120.56 ) ) while the abundances of the other elements were left frozen to solar values .
for all of our fits , both the column density @xmath19@xmath20 and the temperature @xmath2 parameters remained broadly consistent ( approximately @xmath19@xmath20 @xmath3 3 - 4@xmath2910@xmath21 @xmath22 and @xmath2 @xmath3 0.5 - 0.7 kev ) . in figures
[ kes17ascaspectra ] we present the extracted gis2 , gis3 and sis0 spectra of kes 17 as fit with the phabs@xmath29apec model with variable abundance and in figure [ kes17ascaspectralplots ] we present confidence contour plots ( namely @xmath19@xmath20 versus @xmath2 and @xmath2 versus abundance ) for this fit . in table
[ gississpectrasummarykes17 ] we present a summary of all of the spectral fits to the extracted _ asca _ spectra of kes 17 .
confidence contour plots for the @xmath19@xmath20 , @xmath2 and the abundance parameters for the phabs@xmath29apec fit are presented in figure [ kes17ascaspectralplots ] . in figure [ kes17_mos1_most ]
we present an _
xmm - newton _ mos1 image of kes 17 : once again , radio emission from the snr is plotted with contours . with an angular resolution of approximately 6@xmath30 , the _ xmm - newton _ observation of kes 17
has yielded the highest angular resolution image ever produced of this snr .
inspection of this image clearly shows that the diffuse x - ray emitting plasma is enclosed by the radio shell and that the bright radio rims are x - ray dim .
such a contrasting x - ray and radio morphology of this snr strongly indicates that it belongs to the well - known class of snrs known as the mixed - morphology snrs @xcite .
we extracted mos1 , mos2 and pn spectra of kes 17 using a source extraction region approximately 4 arcminutes in radius and centered on the nominal center of the x - ray - emitting plasma : a background spectrum was extracted from an annulus surrounding the source extraction region and with a thickness of approximately 1 arcminute ( the locations of these extraction regions are indicated in figure [ kes17_mos1_most ] ) .
we then fit the extracted spectra with the same thermal models used in analyzing the extracted _ asca _ spectra : once again , statistically - acceptable fits were derived with the apec and nei models . in agreement with the analysis of the _ asca _ spectra , the derivation of statistically - acceptable fits with the apec model and a high derived ionization timescale with the nei model ( @xmath48 = 3.70@xmath57@xmath2910@xmath52 s @xmath50 )
imply that the x - ray emitting plasma is close to cie .
in contrast to our fits derived in analyzing the _ asca _ spectra , thawing the elemental abundance as a free parameter in the apec model still returns a value of the abundance which is broadly consistent with a solar value ( 0.53@xmath58 ) and did not improve the quality of the fit . extending our prior analysis of the _ asca _ spectra of kes 17 , in analyzing the extracted _ xmm - newton _ spectra we also searched for variations in the abundances of particular elements by using the vapec model .
we again first fit the spectra with the magnesium abundance thawed and the abundances of the other elements frozen to solar values .
separately , we fit the spectra again with the vapec model but this time with the abundances of silicon and sulfur thawed and the abundances of the other elements frozen to solar values . in the case of the latter fit ,
the abundances of silicon and sulfur were both broadly consistent with solar values ( si=0.79@xmath59 and s=0.71@xmath60 ) but in the former case we again find marginal evidence for an overabundance of magnesium ( mg=2.75@xmath61 ) .
we also note that for all of our fits with thermal models , the values for both @xmath19@xmath20 and @xmath2 remain broadly similar ( that is , @xmath19@xmath20@xmath33@xmath2910@xmath21 @xmath22 and @[email protected] - 1.10 kev ) .
finally , we fit the extracted spectra using the vapec model with the abundances of magnesium , silicon and sulfur all thawed : this new fit did not yield a significant improvement in the value of @xmath53@xmath62 .
if the detection of an overabundance of magnesium in the x - ray emitting plasma associated with kes 17 is verified , this snr may be the newest member of the class of mixed - morphology snrs which feature ejecta - dominated emission : another example of such a source is the mixed morphology snr ctb 1 , which was shown to feature an overabundance of oxygen in spectra extracted from a _
observation of the snr @xcite . in figure [ kes17_mos1mos2pn_spectra ]
we present the mos1 , mos2 and pn spectra of kes 17 as fitted with the phabs@xmath29apec model as well as a confidence contour plot for column density @xmath19@xmath20 and temperature @xmath2 for this fit . in table
[ xmmspectrasummarykes17 ] we provide a summary of the fits to the extracted mos1 , mos2 and pn spectra of kes 17 . in figures
[ kes17xmmimage ] and [ kes17rayinfrared ] we present multi - wavelength images x - ray , radio and infrared ( namely _ spitzer _ irac and mips images ) of this snr .
we first describe prior analyses made of the x - ray properties of kes 17 .
@xcite presented a study of a pointed _ xmm - newton _ observation of kes 17 that included an analysis of mos1 , mos2 and pn spectra of this snr : we have analyzed this same observation in the present paper .
those authors divided the snr into three regions and extracted spectra for each region while we have considered just the spectra of the entire diffuse emission .
@xcite fit the extracted spectra using the absorption model wabs @xcite coupled with the non - equilibrium ionization model with constant temperature pshock@xcite and a power law model with a photon index @xmath63 ( defined such that @xmath64@xmath65 ) .
in addition , @xcite allowed the abundance of the pshock model to vary .
the values for @xmath19@xmath20 , @xmath2 , abundances , @xmath48 , and @xmath63 obtained in the fits presented by @xcite show some variation , which was interpreted by those authors to indicate spectral variations from one portion of the snr to the other .
we adopted mean values for these parameters and fit our extracted mos1 , mos2 and pn spectra with these mean values in a wabs@xmath29(pshock+power law ) model : we present the results of the fit using these parameters for table [ xmmspectrasummarykes17 ] .
we note that we obtain a marginally statistically acceptable fit to the mos1+mos2+pn spectra using this model with these fit parameters ( @xmath53@xmath62 = 1.21 ) .
the values for the parameters of the thermal model and the column density derived by @xcite are comparable to those that we obtained .
like @xcite , we conclude that the x - ray emitting plasma associated with kes 17 is in cie ( as indicated by their fitted value for @xmath48 ) , but unlike @xcite , we do not require an additional power law component to obtain an acceptable fit to the extracted spectra . when we added the power law , the value of @xmath66 was not improved ( as listed in table [ xmmspectrasummarykes17 ] ) .
the differences between our results and those published by @xcite may originate from the choice of background subtraction ( we have chosen an annular region for the extraction of a background spectrum while @xcite extracted background spectra from a region where no x - ray emission was detected " ) .
after careful background subtraction , the detection of emission above 5 kev was marginal in the spectra .
@xcite presented an analysis of observations made of kes 17 with the x - ray imaging spectrometers ( xiss ) aboard _
suzaku_. those authors fitted the extracted spectra using the wabs absorption model coupled with the variable abundance thermal model vmekal @xcite : the abundances of magnesium , silicon and sulfur were thawed by those authors during the fitting while the abundances of the other elements were frozen to solar values .
in addition , @xcite used an additional power law component in their spectral fitting like @xcite . to compare our results with those presented by @xcite
, we fitted our extracted mos1+mos2+pn spectra with the fit parameters published in that paper .
we present the results of this fitting in table [ xmmspectrasummarykes17 ] : we obtain a statistically acceptable fit using these fit parameters ( @xmath53@xmath67@xmath42 = 1.11 ) and again the values for the parameters @xmath2 and @xmath19@xmath20 derived by @xcite are similar to those obtained by @xcite and the present work .
@xcite also concluded that the x - ray emitting plasma associated with kes 17 is in cie .
we also attempted to include a power law component in the spectral fitting : although the improvement of the fit was not significant , the temperature did become lower ( from 0.90 to 0.84 kev ) .
the temperatures that we obtained were comparable to the temperatures published in those two papers . since the qualities of the fits of the thermal model with or without a power law in the @xcite and @xcite were not compared , the essentiality of the power law is not clear in the x - ray energy range .
inspection of our high energy _ xmm - newton _ image of kes 17 ( see panel ( c ) of figure [ kes17xmmimage ] ) does suggest that some high energy x - ray emission is indeed associated with kes 17 and the origin of this emission is uncertain .
possible origins for this particular emission include high energy thermal emission from the snr or synchrotron emission from an as yet undiscovered neutron star associated with kes 17 . finally , we comment that @xcite , @xcite and our present work all agree that the contrasting center - filled x - ray morphology and shell - like radio morphology of kes 17 motivate its classification as a mixed - morphology snr .
we will return to the study of the mos1+mos2+pn spectra of kes 17 when we discuss overionization in mixed - morphology snrs in section [ overionizationsection ] .
[email protected] was first discovered during surveys made with the parkes 64-meter radio telescope at 5000 mhz @xcite and the molonglo radio telescope at 408 mhz @xcite .
it was first classified as a snr by @xcite based on its nonthermal spectral index ( @[email protected] ) .
those authors also measured an angular diameter of 3.9 arcminutes for [email protected] and extrapolated a flux density at 1 ghz of 3.7 jy from measured flux densities at 408 mhz and 5000 mhz . using the flux density at 1 ghz , the measured angular extent and the surface - brightness diameter ( @xmath68-d ) relation for snrs derived by @xcite , @xmath43 @xmath69@xmath70 when @[email protected] which is approximately the same index as the one measured for [email protected] ] @xcite calculated a distance of 12.5 kpc to this snr and a corresponding diameter of 14 pc : we adopt this distance to [email protected] in this paper .
we also note that @xcite placed a lower limit of 6.6 kpc on the distance to this snr based on its hi absorption profile .
additional radio observations made of a field including this snr at the same two frequencies as @xcite and @xcite were described by @xcite : the values for the angular extent , flux densities and spectral index measured by those authors for the snr were in good agreement with the values given by @xcite .
a shell - like radio morphology for this snr was first suggested by @xcite : in addition , @xcite described observations made of [email protected] at the frequency of 843 mhz using the most : they also reported a shell - like morphology for the source and measured a flux density of 2.9 jy at 843 mhz for this snr .
finally , @xcite compared most observations made at this frequency with observations made at 8.3 @xmath17 m with the midcourse space experiment ( msx , * ? ? ? * ) and reported that emission from [email protected] at this wavelength was not detected by msx down to a limiting flux of 1.0@xmath2910@xmath71 w m@xmath72 sr@xmath73 = 2.35@xmath2910@xmath74 ergs @xmath22 sec@xmath73 .
@xcite included [email protected] among a set of galactic snrs that were observed with the intent of detecting maser emission from the oh satellite line at 1720.5 mhz with the parkes 64-meter radio telescope , but no such emission was detected from [email protected] .
there have been no previously - published analyses of x - ray emission from this snr . in figure [ g311ascaimage ]
we present a co - added _ asca _ gis2+gis3 image of [email protected] : the snr was offset by approximately 7 arcminutes from the nominal center of the field of view during the observation .
radio emission ( as detected at the frequency of 843 mhz ) is plotted as contours overlaying the x - ray emission .
the work we describe here represents the first detection and analysis of x - ray emission from [email protected] .
the x - ray emission in fact is faint and coupled with the poor angular resolution of the gis instruments prevents us from making very definitive statements about the x - ray morphology . to the limit of the existing data
it appears that the emission is centrally - concentrated , in contrast to the shell - like radio morphology of the snr .
as shown below , our spectral analysis strongly indicates that the x - ray emission from [email protected] is thermal in origin .
a centrally - concentrated thermal x - ray morphology coupled with a shell - like radio morphology strongly suggests that [email protected] is in fact an mm snr , though additional x - ray observations with improved sensitivity and angular resolution are needed to investigate the true nature of the x - ray morphology of [email protected] more thoroughly .
gis2 and gis3 spectra were extracted for [email protected] using a circular aperture 5 arcminutes in radius and centered at a position corresponding to the apparent middle of the radio shell of this snr .
the corresponding background gis2 and gis3 spectra required for fitting were generated using the ftool mkgisbgd .
the number of detected counts for both spectra were very modest ( 75 and 111 counts for the gis2 and gis3 after background subtraction , respectively ) , so we only used simple models ( namely a power law model and the thermal plasma apec model ) to fit the spectra coupled with the phabs model to account for photoelectric absorption along the line of sight .
a summary of these fits is presented in table [ gisspectrasummaryg311 ] : we find that both the power law model and the apec model adequately fit the extracted spectra , with each fit featuring corresponding values for @xmath53@xmath62 of @xmath31.1 .
in addition , the values for the column density @xmath19@xmath20 returned by the two fits are comparable ( ranging from @xmath19@xmath20 @xmath3 2.7@xmath2910@xmath21 @xmath22 to @xmath19@xmath20 @xmath3 3.3@xmath2910@xmath21 @xmath22 ) .
the value for @xmath2 as returned by the fit with the phabs@xmath29apec model is elevated but plausible for an snr ( that is , @[email protected]@xmath75 kev ) while the value for the photon index @xmath63 returned in the phabs@xmath29power law fit is steep enough ( @xmath63=3.10@xmath76 ) to also suggest a thermal origin of the x - ray emission rather than ( for example ) synchrotron radiation . in figure [ g311ascaspec ]
we present both the _ asca _ gis2 and gis3 spectra as fit by the phabs@xmath29apec model described above ( and summarized in table [ gisspectrasummaryg311 ] ) and a confidence contour plot for this particular fit for the parameters _ kt _ and @xmath19@xmath20 .
we conclude that the _ asca _ spectra of [email protected] favor a thermal origin for the observed x - ray emission from this source : follow - up observations made with higher spectral resolution with sufficient counts are required to confirm this interpretation .
our analysis of the _ asca _ observations of g311.5 - 0.3 represents the first detection of x - ray emission from this snr . as mentioned above
, the distance to this snr is unknown : to help constrain the distance to this snr more rigorously , we have examined co data obtained from observations made toward [email protected] .
data from a co observation of this source ( corresponding to a velocity of 39.7 km s@xmath73 ) is presented in figure [ g311co ] : we note the presence of a giant molecular cloud ( gmc ) detected in proximity to this snr .
@xcite have argued that [email protected] is interacting with molecular clouds based on the detection of h@xmath18 emission at infrared wavelengths .
based on this argument , we claim that [email protected] is associated with the gmc seen in projection toward [email protected] : the particular velocity of 39.7 km s@xmath73 for the gmc corresponds to a distance of 12.5 kpc .
this distance matches our adopted distance to this snr . the fitted column density toward this source as determined from our x - ray spectral fitting ( that is , @xmath19@xmath20 = 2.7 @xmath29 10@xmath21 @xmath22 ) and our calculated x - ray luminosities
have been determined using this adopted distance .
we also note that the distance toward [email protected] as adopted by @xcite is also based on this high column density as suggested by the velocity of the gmc detected in co : the uncertainty of the distance estimate is not known . in figure
[ g311xrayinfrared ] we present a _ spitzer _
three - color irac image of [email protected] with x - ray contours ( corresponding to the emission detected by the
gis2+gis3 detectors ) overlaid .
the nearly - complete infrared emission with colors consistent with molecular shocks ( see * ? ? ? * ) and the center - filled thermal x - ray emission support the classification of this source as an mm snr . based on its center - filled thermal x - ray morphology and its shell - like radio morphology
, we suggest that [email protected] belongs to the class of mm snrs .
as noted above , the infrared properties of this snr can be modeled with shocked molecular emission and we find evidence ( based on a co observation ) of a cloud seen in projection toward [email protected] : the large distance implied to this snr based on the high column density ( as derived from the spectral fitting ) matches the implied distance to this cloud .
the infrared colors of [email protected] imply that the snr is a molecular snr @xcite and irs follow - up spectroscopy confirmed that the _ spitzer _ irac 4.5@xmath17 m band emission shown in figure [ g311xrayinfrared ] is h@xmath18 emission @xcite which likely originates from the interaction between the snr and dense clouds .
our _ asca _ analysis of [email protected] revealed for the first time x - ray emission from this snr .
the x - ray morphology of the emission is center - filled lying within the radio and infrared shells of emission and the x - ray emission is thermal with a temperature @xmath2 @xmath3 0.98 kev . for these reasons , we suggest that [email protected] is an mm snr , but because the number of photons in the _ asca _ observation of [email protected] is limited , future observations of this snr with improved angular resolution and flux sensitivity are required to confirm this classification .
such observations may also help strengthen the correlation between the phenomenon of mm snrs and snrs interacting with adjacent molecular clouds .
like [email protected] , [email protected] was first detected as a radio source and classified as an snr based on surveys made with the parkes 64-meter radio telescope at 5000 mhz and the molonglo radio telescope at 408 mhz @xcite . these observations along with more recent radio observations made of the snr by @xcite using the very large array ( vla ) at 1.465 ghz as well as by @xcite using the most at 843 mhz
reveal the snr to be shell - like with a diameter of approximately 8.2 arcminutes .
@xcite measured a flux density of 8.1 jy at 1.465 ghz and also estimated a spectral index of @xmath40 = -0.5 for this snr .
those authors also commented that the shell of radio emission from the snr appeared to be flattened with a noticeable compression along the eastern and northwestern borders .
based on observations made with the vla of the oh ground state transition at 1720 mhz , @xcite detected five masers along the southern rim of the snr .
@xcite also reports the detection of oh maser emission at 1720 mhz from [email protected] as observed with the parkes 64-meter telescope .
interestingly , the radio emission from the snr is strongest from the northern , eastern and western portions of the shell but is noticeably weaker along the southern rim , where the masers are detected .
@xcite estimated the distance to [email protected] to be @xmath39 kpc based on the @xmath68-@xmath69 relationship for snrs that was presented by @xcite while @xcite provided lower and upper bounds on the distance to this snr of 5.5 kpc and 11 kpc respectively based on the estimates of the rotational velocities of the molecular clouds argued to be interacting with [email protected] based on detected maser emission .
we adopt a distance of 11 kpc to this snr for the remainder of this paper : we note that a larger distance estimate to [email protected] is supported by the enhanced column density seen toward this snr based on the results of our spectral fitting ( see section [ g346resultssection ] ) .
previously - published x - ray studies of [email protected] have been presented by @xcite ( who analyzed an _ asca _ observation of [email protected] ) , @xcite and @xcite ( who both analyzed a _ suzaku _
observation of this snr ) .
we compare our results in the present paper with the results published in those papers in section [ g346discusssubsection ] . in figure [ g346ascaimage ]
we present a co - added _ asca _ gis2+gis3 image of [email protected] : the center of the x - ray emission from the snr was offset by approximately 14 arcminutes from the nominal center of the field of view during the observation .
we have overlaid radio contours at a frequency of 843 mhz to illustrate the radio morphology of the shell : similar to the situation with [email protected] described previously , while [email protected] is only modestly detected by _ asca _ , the x - ray morphology appears to be center - filled while the radio emission is shell - like .
again , similar to the situation with [email protected] , unfortunately the poor angular resolution of the gis2 and gis3 instruments and the modest detection of the snr prevent us from making more conclusive statements about the general x - ray morphology of [email protected] . as shown below
, the x - ray emission detected from this snr is predominately thermal in origin .
the contrasting x - ray and radio morphologies of [email protected] coupled with the thermal nature of its x - ray emission motivate the classification of this source as a mixed - morphology snr ( which also parallels our results with [email protected] ) .
as mentioned previously , the detection of maser emission from this snr strongly suggests that [email protected] is currently interacting with a molecular cloud : interactions with molecular clouds is another commonly - seen characteristic of mm snrs .
gis2 and gis3 spectra were extracted for [email protected] using a circular aperture 5 arcminutes in radius and centered at a position corresponding to the apparent middle of the radio shell of this snr .
the region of spectral extraction is indicated in figure [ g346ascaimage ] : similar to our work in analyzing the gis2 and gis3 spectra of [email protected] , we used the ftool mkgisbgd to generate background spectra needed for spectral fitting .
inspection of the extracted spectra revealed that there was a significantly greater amount of emission at higher energies compared to the extracted spectra of kes 17 and [email protected] .
note that our re - analysis of _ asca _ data of g346.6 - 0.2 produces higher quality spectra than those presented by @xcite .
we first fitted the extracted spectra with a simple power law model : the photon index of the power law fit was @xmath63 @xmath3 2.5 but the fit was not statistically acceptable ( @xmath53@xmath62=1.34 ) .
we then fitted the spectra with the plasma model apec ( with the abundance frozen to solar ) : the temperature of the fit was @xmath2 @xmath3 2.5 kev but once again the fit was not acceptable ( @xmath53@xmath62=1.45 ) .
we then fit the spectrum with the apec model again , this time with the abundance parameter thawed : the fit returned by this model was significantly improved ( @xmath53@xmath62=1.22 ) but still not statistically acceptable .
the upper limit on the fitted elemental abundance was 0.4 ( relative to solar ) while the fitted temperature was @xmath2 @xmath3 3.4 kev , which is higher than those typically measured for mmsnr snrs .
if this estimated temperature for the snr is indeed true , it may indicate that [email protected] is a particularly young snr .
we also attempted to fit the spectra with the vapec model : we obtained a statistically - acceptable fit with the vapec model ( @xmath53@xmath62=1.14 ) though the only elemental abundance that was found to exceed solar was neon , and in this case the fitted abundance was very high ( with a lower limit of 26 times solar ) .
however , because the low signal - to - noise of the extracted gis spectra made it too difficult to search for variations in the abundances of specific elements in a rigorous manner , we discarded this phabs@xmath29vapec fit .
next , we fit the spectrum with the non - equilibrium ionization plasma model nei ( with the abundance frozen to solar ) : an acceptable fit was obtained ( @xmath53@xmath62 of the fit was 1.07 ) but again the fitted temperature ( @xmath2 = 2.80@xmath77 kev ) was high for an snr while the ionization timescale ( @xmath48 = 7.27@xmath78@xmath2910@xmath79 @xmath50 s ) was lower than other mm snrs . this result indicated that the x - ray - emitting plasma associated with [email protected] is not in cie and motivated the use of an nei model with thawed abundance coupled with a power law to fit properly model the high energy x - ray emission . due to the low signal - to - noise in the spectrum that would make calculating the error bounds of many individual parameters difficult , we attempted fits where we froze values of @xmath48 and @xmath63 to physically realistic parameters ( specifically 10@xmath52 @xmath50 s for @xmath48 to reflect the x - ray - emitting plasma being outside of cie and 0.5 and 2.0 for @xmath63 to reflect values obtained in fits of components of hard emission observed toward other galactic snrs ) .
we attempted to analyze diagnostics of the nei model using line ratios such as he@xmath40 over ly@xmath40 for mg , si or s , but none of them show clear lines so such analysis was not possible . in considering the fits with high @xmath48 ,
all appear to be statistically acceptable with approximately the same values of @xmath53@xmath42@xmath67 : we believe that the fits made using adopted values for @xmath63 and @xmath48 of 0.5 and 10@xmath52 @xmath50 s respectively might be the most physically reasonable because these values for these fit parameters may be the most physically accurate .
we present a summary of these fits in table [ gisspectrasummaryg346 ] . in figure [ g346spectra ] we present the _
asca _ gis2 and gis3 spectra as fit by the phabs@xmath29(nei+power law ) model described above where we have frozen @xmath48 and @xmath63 to 8@xmath2910@xmath51 @xmath50 s and 0.5 , respectively .
the results of all of the fitting ( including values for fit parameters ) are listed in table [ gisspectrasummaryg346 ] .
finally , in figure [ g346xrayinfraredradioimage ] we present a multi - wavelength image x - ray , radio and infrared ( namely _ spitzer _ mips ) of [email protected] . in a previous analysis of extracted gis spectra of this snr , @xcite presented statistically - acceptable fits using a thermal component ( wabs@xmath29mekal , with fitted values of @xmath19@xmath20=2.0@xmath80@xmath2910@xmath21 @xmath22 and @xmath2 = 1.6@xmath81 kev ) or a power law component ( wabs@xmath29power law , with fitted values of @xmath19@xmath20=2.6@xmath82@xmath2910@xmath21 @xmath22 and @xmath63=3.7(@xmath831.7 ) .
we find that we can obtain statistically acceptable fits to our extracted spectra using these models within the quoted error bounds .
more recently , @xcite described an analysis of an observation made of [email protected] with @xmath84 : those authors presented spectral fits to the diffuse emission using several different thermal models ( including vmekal , vnei and the constant temperature plane parallel shock plasma model with variable elemental abundances vpshock see @xcite ) coupled with a power law model ( the fitted values of the photon indices were @[email protected] - 0.6 ) . the thermal models all featured abundances of several elements ( namely magnesium , silicon , sulfur , and iron ) which were all sub - solar while the abundance of calcium exceeded solar .
we have fitted our extracted spectra with the models and the associated fit parameters published by @xcite and we present our results in table [ gisspectrasummaryg346 ] .
we obtain statistically acceptable fits to our extracted spectra using these models with these associated parameters : like @xcite , we need a combination of a thermal component and a power law component to obtain a statistically acceptable fit with physically realistic parameters .
unfortunately our extracted gis spectra lack sufficient signal - to - noise for a detailed analysis of the elemental abundances of the x - ray - emitting plasma as conducted by @xcite . using _ asca _
data , we confirm that that the x - ray emitting plasma associated with [email protected] is thermal and that the x - ray morphology is center - filled within a radio shell .
these facts motivate the classification of [email protected] as an mmsnr .
@xcite present a more detailed analysis of the _ suzaku _ observations of [email protected] and argue that the x - ray spectrum of this snr is best fit with a recombining plasma model , while such spectral analysis was not possible with the _ asca _ data due to the limited number of counts and the poorer spectral resolution .
based on the measured elemental abundances in their fits to the extracted spectra , @xcite argued that the x - ray emission from [email protected] is ejecta - dominated and by analyzing the relative abundances of the elements that a type ia supernova explosion produced this snr .
we note , however , that the h@xmath18 emission detected from this snr by @xcite indicates the presence of a fast j - shock ( with a velocity of approximately 150 km s@xmath73 ) : the spectral lines associated with shocked h@xmath18 suggest a high preshock density , such as the elevated preshock density associated with molecular clouds .
we argue that the presence of this shocked h@xmath18 indicates an interaction between this snr and the adjacent molecular cloud ( as supported by the detection of maser emission from the perimeter of this snr by @xcite ) .
we therefore argue that [email protected] had a massive stellar progenitor and that the type of supernova that created this snr was a type ib / ic / ii supernova ( that is , the supernova types that are all associated with massive stellar progenitors ) . such massive stellar progenitors are usually associated with mm snrs .
xmm - newton _ and _ chandra _ observations of [email protected] would be critical in confirming or refuting these conclusions .
the ctb 37 radio complex was first cataloged as a radio source by a 960 mhz survey conducted by the owens valley radio observatory @xcite .
this complex was resolved into two discrete radio sources
ctb 37a and ctb 37b by radio observations made at multiple frequencies by @xcite , who measured non - thermal radio spectral indices for both objects ( @xmath40=-0.55 in each case ) and thus classified the two sources as snrs .
subsequent radio observations of ctb 37a @xcite revealed this snr to feature a shell - like morphology with bright northern and eastern rims . @xcite estimated an integrated flux density for ctb 37a at 1410 mhz of approximately 76 jy . based on most observations of this snr , @xcite commented on the shell - like morphology of ctb 37a by noting that this source possessed one of the highest mean surface brightnesses of any snr in the most catalog . @xcite
also estimated the angular extent of ctb 37a to be 19@xmath2916 arcminutes and commented on the presence of faint emission associated with this snr extending into the southwest : such an extension of emission is consistent with the expansion into a cavity .
oh masers observed at a frequency of 1720 mhz have been detected along the northern , western and southern boundaries of this snr as well as toward its center @xcite , though a bimodal distribution of measured velocities was observed for these masers ( specifically , masers with associated velocities of @xmath165 km s@xmath73 were observed toward the center of the snr as well as the eastern and southern boundaries while masers with associated velocities of @xmath124 km s@xmath73 were observed toward the northern boundary ) .
@xcite estimated a distance of [email protected] kpc to ctb 37a based on its observed hi absorption profile while @xcite derived a distance to the snr of 11.3 kpc based on a putative interaction between this snr and adjacent molecular clouds , as observed by emission from the carbon monoxide ( co ) @xmath86=1 to 0 transition at approximately 115 ghz ) .
more recently , @xcite bounded the distance to ctb 37a between 6.3 and 9.5 kpc based on the absorption profile along the line of sight toward the snr .
those authors concluded that ctb 37a lies in front of the far side of the molecular cloud seen toward this snr . in the present paper , we adopt a distance of 8 kpc toward this snr , approximately the mean of the lower and upper distance bounds proposed by @xcite .
we also note that @xcite detected @xmath46-ray emission from ctb 37a using the cherenkov telescopes of the high energy stereoscopic system ( hess see @xcite ) : specifically , those authors associated the @xmath46-ray source hess j1713@xmath1385 with ctb 37a ( in particular with the region of hard emission seen toward the northwestern rim of the snr ) . based on fits to the spectrum of hess j1713@xmath1385 , @xcite speculated that the observed emission had a hadronic origin ( that is , it is produced by the decay of pions that are created by collisions between cosmic - ray protons and protons in an adjacent molecular cloud ) . regarding previous x - ray observations of ctb 37a , published work based on observations made with _ asca _
@xcite , _ chandra _ , _ xmm _ ( both presented by @xcite ) and finally _ suzaku _
@xcite have all appeared previously in the literature : we compare our results with the results presented in those papers in section [ ctb37adiscusssubsection ] . in figure [ ctb37a_mos1figure ] we present an _
xmm - newton _ mos1 image of ctb 37a over the energy range from 0.7 to 10.0 kev : contours depicting the detected radio emission at 843 mhz as detected by most are overlaid .
the x - ray emission detected by _ xmm - newton _ is indeed extended and appears to lie interior to the radio shell : a similar description of the x - ray morphology of ctb 37a was also presented by @xcite and @xcite . in figure [ ctb37a_xmmspitzerfigure ]
we present a _ spitzer _
irac 5.8 @xmath17 m image of ctb 37a with _ xmm - newton _ mos1 x - ray emission overlaid with contours .
we have extracted mos1 , mos2 and pn spectra for the entire diffuse emission ( except for a region to the northwest which is known to feature a significantly harder spectrum than the diffuse emission from the rest of the snr see the description below ) .
this extraction region was approximately four arcminutes in diameter and was centered on the nominal center of the observed diffuse x - ray emission . for the purposes of spectral fitting , a background spectrum was extracted from an annular region extending approximately 0.5 arcminutes in radius beyond the region of source extraction and with the same center .
these regions of extraction are depicted in figure [ ctb37a_mos1figure ] .
we also attempted to fit extracted mos1 , mos2 and pn spectra of the northwest region known to feature hard emission , but our extracted spectra lacked a sufficient signal - to - noise ratio suitable for detailed spectral analysis .
we will discuss the properties of this source in more detail when we discuss an analysis of its _ chandra _ acis - i spectrum below .
the mos1 , mos2 and pn spectra of the diffuse emission were fitted with the apec and nei models : statistically acceptable fits were derived with both models . because the apec model returned a statistically acceptable fit and because the ionization timescale returned by the nei fit was elevated ( @xmath48 @xmath3 10@xmath49 s @xmath50 with a lower bound of @xmath3 10@xmath52 s @xmath50 ) , we conclude that the x - ray emitting plasma is close to thermal equilibrium .
we also allowed the abundances of the apec and nei models to vary to see if the quality of the fits could be improved .
in fact , small improvements in the quality of the fit ( as measured by the @xmath53@xmath62 parameter ) were realized as sub - solar values for the abundances were derived ( approximately 0.3 - 0.4 ) .
we also tried fitting the spectra with the vapec and vnei models : only in the cases where silicon was varied were modest improvements in the value of @xmath53@xmath62 attained . in those cases ,
the abundance of silicon was sub - solar ( approximately 0.5 - 0.6 ) . in table
[ xmmspectrasummaryctb37a ] we present a summary of the parameters for the different model fits to the mos1 , mos2 and pn spectra of ctb 37a . in figure [ ctb37axmmspectralplots ]
we present a plot of the mos1 , mos2 and pn spectra of ctb 37a as fit with the phabs@xmath29apec model with variable abundances .
we also present in that figure confidence contour plots that depict @xmath19@xmath20 versus @xmath2 as well as @xmath2 versus abundance . in figure [ chandramostctb37a ] we present an exposure - corrected _ chandra _ acis - i image of ctb 37a over the energy range from 0.5 through 10.0 kev :
contours depicting the detected radio emission at 843 mhz as detected by most are overlaid .
similar to the morphology seen in figure [ ctb37a_mos1figure ] where the x - ray emission detected by the _ xmm - newton _ mos1 is detected , the x - ray emission from ctb 37a as detected by the _ chandra _
acis - i appears to lie interior to the well - defined radio shell : the emission is brightest toward the eastern half of the snr . in figure [ chandractb37athreecolor ]
we present a three - color exposure - corrected _ chandra _ acis - i image of ctb 37a with the soft ( 0.5 - 1.5 kev ) , medium ( 1.5 - 2.5 kev ) and hard ( 2.5 - 10.0 kev ) emission depicted in red , green and blue , respectively .
we have extracted spectra from three particular regions of ctb 37a : we denote these regions as the northwest region , the northeast region and the southeast region , respectively , and in figure [ ctb37a_chandra_specregions ] we show these regions of spectral extraction ( along with the accompanying regions where background spectra were extracted ) . to the best of our knowledge ,
this is the first spatially - resolved spectroscopic study of the x - ray emission from this snr . through this analysis ( as described below )
, we have identified spectral variations between the northeastern and southeastern regions of the snr .
we first discuss our analysis of the northwest region : inspection of figure [ ctb37a_chandra_specregions ] clearly indicates that this source emits primarily hard x - rays : we obtained an adequate fit to this spectra using a simple power law with @xmath63 = 1.18@xmath87 and @xmath19@xmath20 = 4.08@xmath88@xmath2910@xmath21 @xmath22 .
these fit parameters are consistent with those obtained by @xcite who also extracted and fit the spectrum of this source ( which they denoted as cxou j171419.8@xmath1383023 we shall use this designation for the remainder of the paper ) with a phabs@xmath29power law model . @xcite also speculated that this hard x - ray source is associated with the @xmath46-ray source hess j1714@xmath1385 and regarding the true nature of this source considered such scenarios as the high energy emission originating from an interaction between the snr and the adjacent molecular cloud complex or from a pulsar wind nebula , with the latter scenario seeming more plausible . in the cases of the spectra of the northeast and southeast regions
, we obtained acceptable fits using the thermal models apec and nei .
statistically - acceptable fits were obtained with both models : because acceptable fits were obtained with the apec model and because the ionization timescales derived by the nei models were high ( @xmath48 @xmath3 10@xmath52 s @xmath50 with lower limits of 10@xmath51 s @xmath50 or higher ) , we conclude that the x - ray emitting plasma in both regions is close to ionization equilibrium . the fitted column densities for both regions
are comparable ( @xmath19@xmath20 @xmath3 3@xmath2910@xmath21 @xmath22 ) .
this fitted value is lower than the column density derived toward cxou j171419.8@xmath1383023 but within the error bounds the estimated column densities for all three regions are comparable .
interestingly , a difference in temperatures is seen , with the fitted temperatures of the northeast region ( @xmath2 = 0.55 - 0.57 kev ) being noticeably higher than the southeast region ( @xmath2 = 0.78 - 0.83 kev ) .
again , compared to the fit to the extended thermal emission from ctb 37a as derived by @xcite , our derived column densities for both regions are comparable and our derived temperature for the northeast region is quite similar but our temperature for the southeast region is noticeably lower .
we also find differences in the fitted values for the abundances of the two regions : while superior fits to the spectra of the northeast region are obtained with both the apec and nei models when the value of the abundance is sub - solar ( @xmath30.4 ) , the derived values of the abundances obtained from the fits to the spectra of the southeast region with both the apec and nei models are consistent with a solar abundance .
the detection of a reduced value of the abundance is consistent with the result seen previously for the extracted _ xmm - newton _ spectra of ctb 37a . in table
[ acisspectrasummaryctb37a ] we present a summary of the fit parameters for the spectra of these three regions . in figure [ ctb37anwrimspectralplots ]
we present the extracted _ chandra _ acis - i spectrum of cxou j171418.8@xmath1383023 as fit with the phabs@xmath29power law model and a confidence contour plot for this fit for the parameters @xmath19@xmath20 and @xmath63 . in figure [ ctb37aneregionspectralplots ]
we present the extracted _ chandra _ acis - i spectrum of the northeast region as fit with the phabs@xmath29apec model with a thawed abundance as well as confidence contour plots for the parameters @xmath19@xmath20 and @xmath2 as well as abundance and @xmath2 .
lastly , in figure [ ctb37aserimspectralplots ] we present the extracted _ chandra _ acis - i spectrum of the southeast region as fitted with the phabs@xmath29apec model as well as a confidence contour plot for the parameters @xmath19@xmath20 and @xmath2
. we comment on the discrete x - ray source located at ra ( j2000.0 ) 17@xmath89 14@xmath90 28.6@xmath91 , dec ( j2000.0 ) @xmath138@xmath13 36 01 : it is seen in both the _ xmm _ mos1 and the _ chandra _ acis - i images of figure [ chandramostctb37a ] . in the latter image
, the source appears to be blue , indicative of a hard spectrum for this source : it was also detected by @xcite and cataloged by those authors as cxou j171428.5@xmath1383601 .
we will adopt that designation for the source for the remainder of the paper .
it is possible that this source may be the neutron star produced by the supernova explosion that created the snr ctb 37a itself . to investigate this possibility , we extracted an acis - i spectrum for this source and fitted it with a power law coupled with the photoelectric absorption model phabs : we obtained a statistically - acceptable fit ( @xmath53@xmath62 = 0.73 ) using this model .
the photon index of the fit is flat ( @xmath63 @xmath12 1.3 ) consistent with a spectral index expected for a pulsar @xcite and the fitted column density ( @xmath19@xmath20 @xmath12 4.6@xmath2910@xmath21 @xmath22 ) is consistent with the values for @xmath19@xmath20 derived from thermal fits to extracted spectra for portions of the x - ray emitting plasma of ctb 37a .
we present in figure [ ctb37ahardsourceplots ] the extracted spectrum for cxou j171428.5@xmath1383601 and the confidence contours for the phabs@xmath29power law fit described here : it is possible that this source is the neutron star associated with ctb 37a but additional observations are required to determine its true identity . for the purposes of completeness , in figure [ ctb37aascaimage ] we present _ asca _ gis and sis images of ctb 37a and the adjacent snr [email protected] ( the latter object is seen as a radio filament in these images and is discussed in more detail in section [ xmmg348section ] ) .
the image shows some hints of emission from both of these snrs , but because the signal - to - noise ratios of these images are poor , we have instead concentrated on the more sensitive _ xmm - newton _ and _ chandra _ observations of these snrs .
previous x - ray spectral analyses of ctb 37a have been published in the literature based on data from _ asca _ gis observations @xcite , _ chandra _ and _ xmm _ @xcite and finally _ suzaku _ @xcite . in the case of the _ asca _ observations ,
@xcite fit the extracted combined gis spectrum ( where the authors had merged the gis2 and gis3 spectra to boost the signal - to - noise ) with a wabs@xmath29mekal model with elemental abundances frozen to solar , a wabs@xmath29mekal model with silicon abundance thawed and other elemental abundances thawed and a wabs@xmath29power law model . because the signal - to - noise of the fitted spectrum was poor , the fit parameters were poorly constrained : fitted values for @xmath19@xmath20 , @xmath2 and @xmath63 ranged from 1.7@xmath2910@xmath21 to 2.4@xmath2910@xmath21 , 2.2 kev to 2.8 kev and 2.4 to 3.6 , respectively .
we attribute the differences between these fit results and the fit results presented in this paper to the poor signal - to - noise of the merged gis spectrum and significant contaminating flux from bright sources located outside the field of view .
@xcite presented spectral fits to the extracted _ chandra _ and _ xmm - newton _ spectra of the entire diffuse emission associated with ctb 37a .
those authors derived a column density @xmath19@xmath20 = 3.15@xmath92 @xmath29 10@xmath21 @xmath22 and a temperature @xmath2 = [email protected] kev .
we find that while our derived value for @xmath19@xmath20 ( @xmath33@xmath2910@xmath21 @xmath22 ) is comparable to the value derived by @xcite , our derived values for the temperature ( @xmath2 @xmath3 0.5 - 0.7 kev ) are slightly lower .
we argue that our fit values are broadly consistent with the results published by @xcite : we also note that @xcite applied the fit parameters derived by @xcite to their extracted merged gis spectrum and were able to obtain a statistically - acceptable fit .
@xcite presented a spectral analysis of the whole x - ray emission detected from ctb 37a using data obtained from a pointed observation made of the snr by _ suzaku_. those authors fit the extracted xis spectra using a wabs@xmath29(vmekal+power law ) model : their derived fit parameters were @xmath19@xmath20@xmath33@xmath2910@xmath21 @xmath22 , @[email protected] kev and @xmath63=1.6 : they obtained a statistically - acceptable fit with all elemental abundances frozen to solar values .
we have fit our extracted _ xmm - newton _ spectra with this model and present the results of this fit in table [ xmmspectrasummaryctb37a ] : we obtain a statistically - acceptable fit ( with a reduced @xmath53@xmath67 value of 1.13 ) using the parameters derived by @xcite . regarding the additional power law component present in the @xcite fit : we do not require such a component to adequately fit the mos1+mos2+pn spectrum of ctb 37a .
we attribute the differences between the spectral fit that we derived and the spectral fit derived by @xcite to different selections of regions to extract background spectra . @xcite
chose a background region which had the lowest x - ray flux available ( see figure 1 in their paper ) , not from a region accurately reflecting the galactic x - ray background behind the snr .
we argue that our choice of region for extraction of a background spectrum is more appropriate .
[email protected] was first identified as a jet "-
like feature extending from the eastern edge of the bright radio rim of ctb 37a in early radio observations made by @xcite .
while those authors favored the scenario where the jet was a radio feature generated by ctb 37a , subsequent work by @xcite instead argued that the jet was a separate snr seen in projection toward ctb 37a . the classification of [email protected] as a distinct snr was confirmed by @xcite based on radio observations made at the frequencies of 333 mhz , 1443 mhz and 4835 mhz .
the high angular resolution radio maps provided by those authors of [email protected] reveal that the morphology of the snr may be described as a concave rim pointed downward with an angular extent of approximately 3.5 arcminutes .
a similar morphology for the snr is described by @xcite based on their observations made with the most of this snr at a frequency of 843 mhz .
@xcite estimated the flux density and the spectral index of the snr to be 6.2 jy at 1443 mhz and @[email protected] , respectively : however , those authors stressed that their estimated flux density at 1443 mhz ( and at the other frequencies as well ) must be interpreted as a lower limit because observations of this source at frequencies greater than 1000 mhz ( including their own observation at 4835 mhz ) resolve out much of the flux density of the source .
for the same reason , the calculated spectral index should be treated as a lower limit as well ( that is , @xmath40@[email protected]@xmath850.1 ) .
@xcite described observations made with the green bank telescope ( gbt ) of a sample of galactic snrs ( including [email protected] ) at the frequencies corresponding to all four ground - state transitions of the oh radical ( namely the frequencies of 1612.231 , 1665.4018 , 1667.359 and 1720.530 mhz ) .
those authors identified two masers that were spatially coincident with the snr .
@xcite favored a distance of 13.5 kpc to this snr based on a putative association between a weak co concentration observed toward the west of the radio rim .
however , @xcite instead argued for an upper limit on the distance to [email protected] of 6.3 kpc based on the hi absorption profile observed along the line of sight toward this snr .
we therefore adopt a distance of 6.3 kpc to [email protected] for the remainder of this paper . a clear detection of x - ray emission from ( in particular , from the luminous southern radio rim )
has proven to be challenging with the datasets which are currently available for this snr . in figure [ ctb37aascaimage ]
we present _ asca _ gis and sis images of [email protected] : emission from this snr is not clearly seen in either of these images . as noted previously , @xcite stated that the _ asca _ datasets are badly confused by contaminating stray flux originating from a bright source located just south of the field of view .
for this reason , we were not able to detect clearly x - ray emission from the radio rim using the _ asca _ gis data . while the pointed _ chandra _ observation of ctb 37a which was conducted with the acis - i array also included the radio rim of , unfortunately much of this rim fell into a gap between chips in the acis - i array ( see figures [ chandractb37athreecolor ] and [ ctb37a_chandra_specregions ] ) .
this situation coupled with the short exposure time of the observation prevented the clear detection of x - ray emission from the radio - luminous limb of this snr as well .
the pointed _ xmm - newton _ observation made of ctb 37a also sampled [email protected] and we were able to extract mos1+mos2+pn spectra for the luminous radio rim ( though the short exposure time of the observation and the contaminating stray flux severely limited the quality of the data ) .
an elliptical region with radii of 3.52 arcminutes @xmath29 2.19 arcminutes was used to extract x - ray spectra while a region with the same linear dimensions and position angle located approximately 5 arcminutes to the south was used to extract a background spectrum .
the locations of these extraction regions are shown in figure [ ctb37a_mos1figure ] : we also note that in figure [ ctb37a_xmmspitzerfigure ] we depict x - ray emission as detected by mos1 overlaid in contours on a @xmath93 irac 5.8@xmath17 m of a field containing both ctb 37a and [email protected] ( infrared emission from the latter snr is visible as the concave - downward arc visible at the eastern edge of the image ) .
after background subtraction , we estimate the total number of counts in the extracted mos1 , mos2 and pn spectra of [email protected] over the energy range of 0.5 to 10.0 kev to be 68 , 26 and 31 , respectively .
these numbers of counts are unfortunately not sufficient for a sophisticated spectral analysis .
we placed an upper limit on x - ray flux and luminosity of [email protected] over this energy range using the following method .
we used the standard program colden to determine that the nominal galactic column density @xmath19@xmath20 seen toward this snr is @xmath31.7@xmath2910@xmath21 @xmath22 . adopting half of this value as the column seen toward [email protected] ( because the value returned by colden corresponds to the value seen all the way through the galactic disk ) and assuming a raymond - smith thermal plasma model @xcite with solar abundances and a temperature of @xmath2 @xmath3 0.5 kev
, we used the portable , interactive multi - mission simulator ( pimms ) version 4.2 to estimate a corresponding unabsorbed flux of @xmath36.9@xmath2910@xmath74 ergs @xmath22 sec@xmath73 for [email protected] . assuming a distance to [email protected] of 6.3 kpc
, we estimate an upper limit for the unabsorbed x - ray luminosity of the luminous radio rim to be @xmath94@xmath9 @xmath33@xmath2910@xmath95 ergs sec@xmath73 ( over the energy range of 0.5 - 10.0 kev ) .
the true x - ray luminosity of [email protected] may be far below this value .
unfortunately , we still have not clearly detected x - ray emission from this snr : new x - ray observations conducted with greater sensitivity may be more fruitful in detecting such emission .
we now discuss the properties of each snr in turn . we derive and comment upon such physical properties of the snrs as age ( @xmath6 ) , hydrogen nuclei density @xmath4@xmath20 , electron density @xmath4@xmath5 ( we will assume that @xmath4@xmath5 = 1.2 @xmath4@xmath20 ) , pressure @xmath96 and swept - up x - ray - emitting mass @xmath8@xmath9 for each snr .
we pattern the discussion in the present paper with the discussion we presented in a previous work @xcite where we derived and commented upon these properties for two other galactic mm snrs , hb 21 ( [email protected] ) and ctb 1 ( [email protected] ) .
we use the following equations for our analysis .
values for @xmath4@xmath5 and @xmath4@xmath20 may be calculated from the value obtained for the normalization of thermal components in fits to the extracted x - ray spectra of the snrs based on the equation @xmath97 where @xmath98 is the distance to the snr in cm and @xmath99 = @xmath100 @xmath98@xmath99 is the total volume of the snr ( we will adopt uniform densities for @xmath4@xmath5 and @xmath4@xmath20 expressed here in @xmath50 and spherical volumes for the snrs ) .
we estimated the remnant age t@xmath7 assuming the snrs are in the snowplow stage of evolution ( see * ? ? ?
* ) using the following relation @xmath101 where @xmath99@xmath102 is the shock velocity of the snr .
values of @xmath99@xmath102 were measured from ionic lines detected in the infrared spectra by @xcite and @xcite of the five snrs considered in this paper ( all of these snrs are in the radiative phase of evolution ) .
we also consider the ratio @xmath96/@xmath103 ( where @xmath103 is boltzmann s constant ) and calculate this ratio using the equation @xmath104
lastly @xmath8@xmath9 is calculated from the equation @xmath105 where @xmath106 is the volume filling factor of the x - ray emitting plasma ( assumed to be unity in this paper ) and @xmath107@xmath20 is the mass of a hydrogen nucleus .
we discuss these calculated properties for each snr below : in table [ comparisonthreesnrstable ] we present our calculated values for each property in the cases of the snrs kes 17 , [email protected] and ctb 37a . for comparison , we include other published values for these same properties of these snrs ( we compare our results with these other publications in the text below ) . in table [ snrxraystable ] we present a summary of the properties of the diffuse emission of the snrs considered in this paper ( including the values of @xmath99@xmath102 used in our calculations )
. we will discuss table [ snrxraystable ] in more detail in section [ comparisonsection ] .
_ kes 17 : _ based on the derived parameters from our x - ray spectral fits ( namely the phabs@xmath29apec fits to the extracted mos1+mos2+pn spectra of kes 17 ) , we have applied equations [ normeqn ] , [ radageeqn ] , [ pressureeqn ] and [ masseqn ] to calculate an electron density of the x - ray - emitting plasma to be @xmath4@xmath5 = 0.42 @xmath50 , a hydrogen density @xmath4@xmath20 = 0.35 @xmath50 , an age @xmath6@xmath7=2.3@xmath2910@xmath108 , a pressure @xmath96/@xmath103 = 5.6@xmath2910@xmath109 k @xmath50 and a swept - up mass @xmath8@xmath9 of 52 @xmath8@xmath110 .
_ [email protected] : _ through inspection of our spectral fit values and using the equations above , we calculate the following values for the properties of [email protected] : @xmath4@xmath5=0.20 @xmath50 , @xmath4@xmath20=0.17 @xmath50 , @xmath6@xmath7=2.5 - 4.2@xmath2910@xmath108 yr , @xmath96/@xmath103 = 4.4@xmath2910@xmath109 k and @xmath8@xmath9 = 13 @xmath8@xmath110 . _ [email protected] : _ using the values of the fit parameters that we derived for the gis spectra of [email protected] using the phabs@xmath29(nei + power law ) model ( specifically where we froze the values of @xmath48 = 10@xmath52 @xmath50 s and @xmath63=0.5 ) , we calculate @xmath4@xmath5 = 0.24 @xmath50 , @xmath4@xmath20 = 0.20 @xmath50 , @xmath6@xmath7 = 4.7@xmath2910@xmath108 yr , @xmath96/@xmath103 = 1.7@xmath2910@xmath111 k @xmath50 and @xmath8@xmath9 = 84 @xmath8@xmath110 .
we note that this is the only snr considered in this paper which features x - ray spectra that indicate that the snr is far from ionization equilibrium . _
ctb 37a : _ based on the fit parameters derived through analysis of the extracted mos1+mos2+pn spectra for the whole snr , we estimate the following properties for ctb 37a : @xmath4@xmath5 = 0.77 @xmath50 , @xmath4@xmath20 = 0.64 @xmath50 , @xmath6@xmath7 = 3.2 - 4.2@xmath2910@xmath108 yr , @xmath96/@xmath103 = 8.2@xmath2910@xmath109 k @xmath50 and @xmath8@xmath9 = 76 @xmath8@xmath110 . all of these parameters are calculated based on spectral analysis of the x - ray bright portion of the northern half of the snr .
we emphasize that the angular extent of this snr as defined by the extended radio emission includes a southwestern portion ( a possible blow - out " region ) : no x - ray emission has been detected from this southwestern portion .
this region may in fact feature x - ray emission which has remained undetected up to this point by pointed x - ray observations of the snr .
we therefore view our estimate of the x - ray emitting mass associated with this snr as a lower limit .
the phenomenon of overionization has been observed in the x - ray - emitting plasmas associated with numerous galactic mm snrs ( as well as other galactic snrs ) .
overionization occurs when the electron temperature ( @xmath112@xmath5 ) is lower than the ionization temperature ( @xmath112@xmath113 ) of the plasma : in such cases , the recombination process is dominant .
examples of mm snrs that feature overionized x - ray - emitting plasmas are ic 443 ( based on observations made by _ asca _ , _ xmm - newton _ and _ suzaku _ see @xcite , @xcite and @xcite ) , w28 , w44 , w49b and [email protected] ( all based on observations made by _ suzaku _ see @xcite , @xcite , @xcite , @xcite , respectively ) .
reviews of the phenomenon of overionized x - ray - emitting plasmas associated with snrs is provided by @xcite .
observationally , overionization of the x - ray - emitting plasma associated with an snr is manifested by significant residuals that appear when spectra of the plasmas are fitted with cie and nei models .
for example , in the case of w44 , @xcite found significant residuals at @xmath31.5 kev and @xmath3 2 kev , which are attributed to the mg - ly@xmath40 line and the si - ly@xmath40 line , respectively . in the cases of w44 @xcite ,
ic 443 @xcite and g359.1 - 0.5 @xcite , residuals that emerged after fitting with cie models are seen at 2.7 and 3.5 kev : lines at these energies are associated with h - like si ( 2.67 kev ) and h - like s ( 3.48 kev ) , respectively .
finally , in the case of w49b , @xcite identified residuals at the energies of 5.6 , 6.2 , 6.7 , 6.9 , 7.9 , 8.5 kev : these lines are attributed to cr k@xmath40 , mn@xmath40 and four fe lines ( of he@xmath114 , ly@xmath40 , he@xmath115 and he@xmath46 ) , respectively . we have examined the x - ray spectra of the snrs in our sample to search for evidence of overionized plasmas associated with these sources . unfortunately , in the case of ctb 37a , the extracted _ chandra _ spectra do not have sufficiently high signal - to - noise to reveal significant residuals ( if present ) : therefore , there is no clear evidence that the x - ray emitting plasma associated with this snr is overionized .
recently , @xcite analyzed _ suzaku _ spectra of [email protected] and argued that the plasma associated with this snr is overionized .
unfortunately the signal - to - noise of our _ asca _ spectra of [email protected] is insufficient for a study of the overionization of this snr .
inspection of the _ asca _ spectra reveals a hint of residual dips at 1.5 kev ( note the residuals between 1.5 - 2.5 kev as shown in figure [ g346spectra ] ) .
the residual at 1.5 kev ( which is consistent with the mg ly@xmath40 line ) could be evidence of over - ionization as seen in w44 @xcite but the scatter is too large to quantify the overionization .
lastly , the signal - to - noise of our _ asca _ spectra of [email protected] is too poor to identify any lines , while [email protected] was not clearly detected by any of the x - ray observations considered here .
examination of the mos1+mos2+pn spectra of kes 17 ( as shown in figure [ kes17_mos1mos2pn_spectra ] ) reveals evidence for residuals at 1.5 kev ( a positive dip ) and 1.9 kev ( a negative dip ) .
based on the presence of these residuals , we have investigated whether or not the x - ray - emitting plasma associated with kes 17 is indeed overionized .
we conducted a spectral analysis patterned after the analysis presented by @xcite where those authors determined that the x - ray - emitting plasma associated with [email protected] is overionized . using the spex software package @xcite , those authors fitted the extracted _ suzaku _ xis spectra of the snr using the neij model .
as described by @xcite , this model describes an x - ray emitting plasma where initially the ion temperature @xmath2@xmath116 and the original electron temperature @xmath2@xmath117 were the same ( that is , @xmath2@xmath116 = @xmath2@xmath117 ) and the plasma itself was in collisional ionization equilibrium . over time , the electron temperature drops to @xmath2@xmath118 due to rapid electron cooling .
such a model can reflect accurately a rapid electron cooling as predicted in models of snr evolution .
in addition , the neij model was multiplied by the absm model to account for interstellar absorption ( also consistent with the work of @xcite ) .
we first used the spex ( version 2.03.03 ) tool trafo ( version 1.02 ) to convert the extracted _ xmm - newton _ spectral files for the whole snr into .spo files suitable for analysis by the software package .
we were unable to generate a statistically - acceptable fit with the neij model that is superior ( in other words , with an improved value for @xmath53@xmath67@xmath42 ) to the fits obtained with the thermal models described above .
the residuals of the fits obtained with the neij model originate from inconsistencies between the different datasets and because of these limitations the overionization model can not improve the fit .
thus , the poor signal - to - noise of the datasets prevent us from testing sufficiently whether or not the x - ray - emitting plasma is overionized .
a deeper x - ray observation is required to investigate more rigorously the possibility that kes 17 belongs to the class of overionized mm snrs .
in table [ snrxraystable ] we present a summary of the properties of the diffuse emission associated with the four snrs examined in this paper that were detected in the x - ray .
we find that the x - ray properties of three of these snrs kes 17 , [email protected] and ctb 37a are similar to those of such archetypal mm snrs as 3c 391 and w44 @xcite . specifically , the x - ray temperatures of these three snrs range from @[email protected] kev to @xmath2@xmath31 kev , the abundances of the diffuse emission associated with each snr are approximately solar and all of the snrs are middle - aged ( that is , the ages of all of the snrs are @xmath6@xmath7@xmath310@xmath108 yr old ) .
lastly , all three of these snrs are known to produce infrared emission that originates from shocked molecular hydrogen ( that is , warm h@xmath18 ) : this indicates that these snrs are interacting with dense ( 10@xmath119 - 10@xmath120 @xmath50 ) molecular clouds @xcite .
therefore , we conclude these three particular snrs in our sample are similar to the archetypal mm snrs listed above and others ( see * ? ? ?
the x - ray properties of [email protected] differ somewhat from both those of these three mm snrs as well as those of the archetypical mm snrs . while [email protected] does feature a center - filled thermal x - ray morphology with a contrasting well - defined shell - like radio morphology that typifies mm snrs ,
the x - ray spectrum of this snr is harder with a higher temperature and the presence of a second harder spectral component is implied .
the morphology shows center - filled x - ray emission within a well - defined radio shell like those in other mm snrs .
however , the x - ray spectrum is harder than these other mm snrs ( with a higher temperature , @[email protected] - 3.4 kev ) and the fitted ionization timescale is rather lower .
we do note however that an age estimate based on the observed infrared line emission from this snr recall that the shock velocity used in our analysis is based on detected ionic infrared emission is similar to the ages estimated for kes 17 , [email protected] , ctb 37a and the archetypal mm snrs mentioned above .
we do not have a clear idea why [email protected] shows harder x - ray emission , but perhaps it is somehow related to a particular property of this snr : the detected h@xmath18 emission from [email protected] differs from the h@xmath18 emission detected from the other mm snrs in that the detected emission from [email protected] requires a high velocity j - shock component or a high - velocity c - shock component coupled with a slow - velocity c - shock component to be fit properly .
in contrast , the emission detected from the other snrs requires two slow c - shock components .
it is possible that the x - ray emission from [email protected] is associated with the forward shock of this snr , which is necessarily a j - shock because c - shocks do not shock - heat the plasmas of snrs .
in addition , or perhaps as an alternate explanation , c - shocks are associated with lower shock velocities : this might indicate that [email protected] is dynamically a younger snr .
therefore , this would be an indirect connection between the shock and the x - ray emission seen from the interior of this snr .
more x - ray observations of this intriguing snr are clearly warranted .
observations made with the _ fermi _ telescope have detected many mm snrs at @xmath46-ray energies , including w28 , w44 , w49b , w51c and ic 443 @xcite , among others ( also see * ? ? ?
* ; * ? ? ?
the @xmath46-ray emission detected from these sources may be explained as hadronic particle acceleration by the mm snrs expanding into a dense environment .
of the five snrs considered in this paper , two kes 17 and ctb 37a have been detected at @xmath46-ray energies : kes 17 was detected by _
@xcite , while ctb 37a was detected with very high energy ( vhe ) @xmath46-ray telescopes @xcite and _ fermi _ @xcite .
our present research has helped to confirm that kes 17 and ctb 37a are mm snrs that feature center - filled x - ray emission .
we argue that additional @xmath46-ray observations of mm snrs are needed to investigate the inter - relationship between x - ray and @xmath46-ray emission from these particular sources and mechanisms of particle acceleration .
we note that infrared observations of the five snrs considered in this paper have revealed that the infrared emission detected from these snrs is radiative and produced by shocks with velocities @xmath99@xmath102@xmath370 to 150 km s@xmath73 .
it appears that each of these snrs is currently in the snowplow stage of evolution and the x - ray emission detected from kes 17 , [email protected] , [email protected] and ctb 37a is fossil radiation that was produced during an earlier evolutionary stage of the snrs . the x - ray emission from several mm snrs as observed by _ suzaku _
often exhibits multiple ionization states which come from elemental differences in ionization and recombination timescales ( e.g. * ? ? ?
the rarefaction of the recombining plasma is likely due to continuous fast cooling of the plasma due to radiative shocks encountering dense clouds ( as inferred by the characteristics of the detected infrared emission from these snrs ) .
our sample of four x - ray - detected mm snrs in this paper help confirm that a strong correlation exists between infrared shocked h@xmath18 emitting snrs and mm snrs .
we also note that _ chandra _ observations of other mm snrs such as w44 and 3c 391 have confirmed the presence of uniform temperature profiles in x - ray emission from these sources @xcite .
in addition , wide ranges of plasma conditions and metal enrichment have also been revealed by _ asca _ and _ chandra _ x - ray observations of mm snrs , such as an over - ionized plasma in ic 443 @xcite , a suggested abundance gradient at the bright projected center of w44 @xcite and iron - rich ejecta in the mm snrs 3c 397 and sgr a east @xcite .
finally , evidence for x - ray emission from an ejecta - dominated plasma associated with ic 443 as detected by _ xmm - newton _ was presented by @xcite .
simulations have shown that the center - filled thermal x - ray emission as observed in mm snrs can be produced through anisotropic thermal conduction .
specifically , snrs expanding into denser environments tend to be smaller , making it easier for thermal conduction to dictate large changes in the temperatures of their expanding hot gas bubbles .
therefore , thermal conduction plays a prominent role in determining the observed x - ray morphology of mm snrs @xcite .
magnetohydrodynamic ( mhd ) modeling of snrs expanding through an inhomogeneous ism also confirms that x - ray emission detected from mm snrs can be reproduced when both thermal conduction and the reverse shock of the snr are included in the modeling @xcite .
after the reverse shock has reached the center of the snr , a maximum in the x - ray emission is seen toward the center of the snr and the morphology is centrally brightened : such a morphology is a defining characteristic of mm snrs .
therefore , it is evident that thermal conduction plays a crucial role in producing snrs of this class .
follow - up x - ray observations of the mm snrs in our sample are vital to investigate further their properties . while these snrs have already been observed with such current x - ray observatories as _ xmm - newton _ , _ chandra _ and _ suzaku
_ , we argue that a deeper observing campaign of these snrs ( as well as other known mm snrs ) is crucial and will permit detailed comparisons between these snrs to be made .
analysis of the x - ray spectra obtained from these observations will allow us to investigate the line emission of elements present in the x - ray emitting plasma of these sources , specifically their ionization states and temperatures .
such analysis may also reveal the presence of enhanced elemental abundances in the ejecta emission : this phenomenon has been seen in other mm snrs .
finally , @xmath46-ray observations of the mm snrs in our sample which have not yet been detected at such high energies such as [email protected] and [email protected] as well as other mm snrs are clearly warranted .
the detection of @xmath46-ray emission from such sources will provide exciting opportunities to advance our understanding of particle acceleration by these sources .
the conclusions of this paper may be summarized as follows : \1 ) using archival x - ray observations made with _
asca _ , _ xmm - newton _ and _ chandra _ , we have conducted a spectroscopic analysis of the x - ray properties of five snrs ( kes 17 , [email protected] , [email protected] , ctb 37a and [email protected] ) detected in the infrared by the _
glimpse survey .
four of the five snrs ( namely kes 17 , [email protected] , [email protected] and ctb 37a ) are clearly detected by these observations and we have placed an upper limit on the x - ray luminosity of [email protected] ) .
our work represents the first published detection of x - ray emission from [email protected] .
\2 ) the four x - ray detected snrs all feature a center - filled x - ray morphology with a contrasting shell - like radio morphology : therefore , these sources may be classified as mm snrs .
the x - ray emission from each snr appears to be thermal in origin : however , when the spectral properties of the x - ray - emitting plasmas of the snrs are compared , significant variations are seen from one snr to the other .
for example , the plasma associated with kes 17 may be overabundant in magnesium , suggesting that the plasma may be ejecta - dominated .
in addition , evidence exists for the presence of hard components in the x - ray emission from [email protected] .
we also note that we have identified a hard discrete x - ray source seen in projection toward ctb 37a ( and distinct from the extended emission with a hard spectrum seen toward the northwest rim of the snr ) and suggest that it may be a neutron star associated with this snr .
\3 ) we have discussed the plasma conditions of the four x - ray detected snrs and estimated such properties of the plasma as @xmath4@xmath5 , @xmath6@xmath7 and @xmath8@xmath9 : these values range from @xmath30.4 - 0.8 @xmath50 , @xmath32.3 - 4.2@xmath2910@xmath108 yr and @xmath313 - 88 @xmath8@xmath110 , respectively .
these values are similar to those observed for other mm snrs .
\4 ) we discuss the implications of our results for studies of snr evolution . because all four of the x - ray detected snrs feature lines produced by shocked @xmath121@xmath18 in their infrared spectra , we argue that each of them is interacting with a nearby molecular cloud .
we believe that our results help strengthen the link between mm snrs and interactions between snrs and molecular clouds : this may help to explain the origin of the center - filled thermal x - ray morphologies of these sources .
we thank the anonymous referee for many useful comments that helped to improve greatly the quality of this paper . t.g.p .
also thanks koji mukai for assistance with analyzing the _ asca _ datasets considered in this paper and nicholas lee for assistance with analyzing the _ chandra _ observation of ctb 37a . finally , t.g.p .
also thanks jelle de plaa for assistance with using the spex software package .
this research has made use nasa s astrophysics data system as well as data obtained from the high energy astrophysics science archive research center ( heasarc ) , provided by nasa s goddard space flight center .
abdo , a. a. , ackermann , m. , ajello , m. , baldini , l. , ballet , j. , barbiellini , g. , baring , m. g. , bastieri , d. et al .
2009 , , 706 , l1 abdo , a. a. , ackermann , m. , ajello , m. , baldini , l. , ballet , j. , barbiellini , g. , bastieri , d. , baughmann , b. et al . 2010a , , 712 , 459 abdo , a. a. , ackermann , m. , ajello , m. , allafort , a. , baldini , l. , ballet , j. , barbiellini , g. , bastieri , d. et al .
2010b , , 718 , 348 abdo , a. a. , ackermann , m. , ajello , m. , baldini , l. , ballet , j. , barbiellini , g. , baring , m. g. , bastieri , d. et al .
2010c , science , 327 , 1103 abdo , a. a. , ackermann , m. , ajello , m. , baldini , l. , ballet , j. , barbiellini , g. , bastieri , d. , behtol , k. et al . 2010d , , 722 , 1303 aharonian , f. , akhperjanian , a. g. , barres de almeida , u. , bazer - bachi , a. r. , behera , b. , beilicke , m. , benbow , w. , bernlhr , k. et al .
2008 , , 490 , 685 andersen , m. , rho , j. , reach , w. t. , hewitt , j. w. & bernard , j. p. , 2011
, , 742 , 7 arnaud , k. a. 1996 , astronomical data analysis software and systems v , eds .
jacoby , g. and barnes , j. , p. 17
, asp conference series vol .
101 atwood , w. b. et al .
2009 , , 697 , 1071 benjamin , r. a. , churchwell , e. , babler , b. l. , bania , r. m. , clemens , d. p. , cohen , m. , dickey , j. m. , indebetouw , r. et al .
2003 , , 115 , 953 bernlhr , k. et al .
2003 , aph , 20 , 111 borkowski , k. j. , lyerly , w. j. & reynolds , s. p. 2001
, , 548 , 820 brandt , t. & _ fermi _ lat collaboration , 2012 , american astronomical society meeting abstracts , 219 , # 213.03 brogan , c. l. , frail , d. a. , goss , w. m. & troland , t. h. 2000 , , 537 , 875 burke , b. e. , mountain , r. w. , harrison , d. c. , bautz , m. w. , doty , j. p. , ricker , g. r. & daniels , p. j. 1991 , ieee trans . , electron devices , 38 , 1069 castro , d. & slane , p. , 2010 , 717 , 372 caswell , j. l. , murray , j. d. , roger , r. s. , cole , d. j. & cooke , d. j. 1975 , , 45 , 239 caswell , j. l. & barnes , p. j. 1985 , , 216 , 753 chen , y. et al .
2004 , apj , 616 , 885 clark , d. h. , caswell , j. l. & green , a. j. 1973 , nature , 246 , 28 clark , d. h. , caswell , j. l. & green , a. j. 1975 , aujpa , 37 , 1 clegg , p. e. et al .
1996 , , 315 , l38 cohen , m. & green , a. j. 2001 , , 325 , 531 combi , j. a. , albacete colombo , j. f. , snchez - ayaso , e. , romero , g. e. , mart , j. , luque - escamilla , p. l. , muoz - arjonilla , a. j. , snchez - sutil , j. r. & lpez - santiago , j. 2010 , , 523 , 76 downes , a. 1984 , , 210 , 845 dubner , g. m. , moffett , d. a. , goss , w. m. & winkler , p. f. 1993 , , 105 , 2251 fazio , g. g. et al .
2004 , apjs , 154 , 10 foster , a. r. , ji , l. , smith , r. k. , & brickhouse , n. s. , 2012 , , 756 , 128 frail , d. a. , goss , w. m. , reynoso , e. m. , giacani , e. b. , green , a. j. & otrupcek , r. 1996 , , 111 , 1651 freeman , p. e. , kashyap , v. , rosner , r. & lamb , d. q. 2002 , , 138 , 185 fruscione , a. et al .
2006 , proc .
spie , 6270 , 60 gabriel , c. , et al .
2004 , astronomical data analysis software and systems ( adass ) , xiii , 314 , 759 garmire , g. p. , bautz , m. w. , ford , p. g. , nousek , j. a. & ricker , g. r. jr .
, 2003 , proc .
spie , 4851 , 28 gk , f. & sezer , a. , 2012 , , 423 , 1215 green , a. j. , frail , d. a. , goss , w. m. & otrupcek , r. 1997 , , 114 , 2058 green , d. a. 2009a , ` a catalogue of galactic supernova remnants ( 2009 march version ) ' , astrophysics group , cavendish laboratory , cambridge , united kingdom ( available at `` http://www.mrao.cam.ac.uk/surveys/snrs/ '' ) .
green , d. a. 2009b , basi , 37,45 hewitt , j. w. , yusef - zadeh , f. & wardle , m. 2008 , , 683 , 189 hewitt , j. w. , rho , j. , andersen , m. & reach , w. t. , 2009 , , 694 , 1266 hewitt , j. w. & _ fermi _ lat collaboration , 2012 , american astronomical society meeting abstracts , 219 , # 239.23 houck , j. r. et al .
2004 , , 154 , 18 huang , y .- l . &
thaddeus , p. 1985 , , 295 , l13 jansen , f. et al .
2001 , , 365 , l1 kaastra , j. s. , mewe , r. & nieuwenhuijzen , h , 1996 , uv and x - ray spectroscopy of astrophysical and laboratory plasmas , ed .
k. yamashita and t. watanabe ( universal academy press , tokyo ) , p. 411
kawasaki , m. t. , ozaki , m. , nagase , f. , masai , k. , ishida , m. & petre , r. et al .
2002 , , 572 , 897 kassim , n. e. , baum , s. a. & weiler , k. w. 1991 , , 374 , 212 kessler , m. f. et al .
1996 , , 315 , l27 kesteven , m. j. l. 1968a , aujph , 21 , 369 kesteven , m. j. l. 1968b , aujph , 21 , 739 koralesky , b. , frail , d. a. , goss , w. m. , claussen , m. j. & green , a. j. 1998 , , 116 , 1323 liedahl , d. a. , osterheld , a. l. & goldstein , w. h. 1995 , , 438 , l115 maeda , y. , baganoff , f. k. , fiegelson , e. d. , morris , m. , bautz , m. w. , brandt , w. n. , burrows , d. n. , doty , j. p. et al .
2002 , , 570 , 671 makishima , k. et al .
1996 , , 48 , 171 mazzotta , p. , mazzitelli , g. , colafrancesco , s. & vittorio , n. 1998 , , 133 , 403 mewe , r. , gronenschild , e. h. b. m. & van den oord , g. h. j. 1985 , , 62 , 197 mewe , r. , lemen , j. r. , van den oord , g. h. j. 1986 , , 65 , 511 mill , j. , oneil , r. r. , price , s. d. , romick , g. j. , uy , o. m. & gaposchkin , e. m. 1994 , j. spacecraft rockets , 31 , 900 mills , b. y. 1981 , pasa , 4 , 156 milne , d. k. & hill , e. r. 1969 , aujph , 22 , 211 milne , d. k. & dickel , j. r. 1975 , aujph , 28 , 209 milne , d. k. , goss , w. m. , haynes , r. f. , wellington , k. j. , caswell , j. l. & skellern , d. j. 1979 , 188 , 437 morrison , r. & mccammon , d. 1983 , , 270 , 119 ohashi , t. et al .
1996 , pasj , 48 , 157 ohnishi , t. , koyama , k. , tsuru , t. g. , masai , k. , yamaguchi , h. & ozawa , m. 2011 , , 63 , 527 orlando , s. , bocchino , f. , miceli , m. , reale , f. & peres , g. 2009 , memorie della societa astronomica italiana suppl . , 13 , 97 ozawa , m. , koyama , k. , yamaguchi , h. , masai , k. & tamagawa , t. , 2009 , , 706 , l71 pannuti , t. g. , rho , j. , borkowski , k. j. and cameron , p. b. 2010 , , 140 , 1787 raymond , j. c. & smith , b. w. 1977 , , 35 , 419 reach , w. t. , & rho , j. , 2000 , apj , 544 , 843 reach , w. t. , rho , j. , tappe , a. , pannuti , t. g. , brogan , c. l. , churchwell , e. b. , meade , m. r. , babler , b. et al .
2006 , , 131 , 1479 retallack , d. s. 1980 , ph .
d thesis , university of sydney reynoso , e. m. & mangum , j. g. 2000 , , 545 , 874 rho , j. , petre , r. , schlegel , e. m. , & j. hester , 1994 , , 430,757 rho j. & petre , r. , 1996 , , 467 , 698 rho , j. & petre , r. 1998 , , 503 , l167 rho , j. , jarrett , t.j . , cutri , r. m. & reach , w. t. 2001 , , 547 , 885 rho , j. & borkowski , k. j. 2002 , , 575 , 201 robertson , j. g. 1991 , aujph , 44 , 729 safi - harb , s. , dubner , g. , petre , r. , holt , s. s. & durouchoux , p. 2005
, , 618 , 321 sawada , m. & koyama , k. , 2012 , pasj , 64 , 81 serlemitsos , p. j. et al .
1995 , , 47 , 105 sezer , a. , gk , f. , hudaverdi , m. & ercan , e. n. 2011a , , 417 , 1387 sezer , a. , gk , f. , hudaverdi , m. , kimura , m. & ercan , e. n. 2011b , , 415 , 301 shaver , p. a. & goss , w. m. 1970a , aujpha , 14 , 1 shaver , p. a. & goss , w. m. 1970b , aujpha , 14 , 77 shaver , p. a. & goss , w. m. 1970c , aujpha , 14 , 133 shelton , r. l. , kuntz , k. d. , & petre , r. 2004 , , 615 , 275 strder , l. et al .
2001 , , 365 , l18 tanaka , y. , inoue , h. & holt , s. s. 1994 , , 46 , l37 tian , w. w. & leahy , d. a. 2012 , , 421 , 2593 tilley , d.a . ,
balsara , d.s .
& howk , j.c .
2006 , , 371 , 1106 troja , e. , bocchino , f. , miceli , m. , & reale , f. 2008 , , 485 , 777 turner , m. j. l. et al .
2001 , , 365 , l27 uchida , h. , tsunemi , h. , katsuda , s. , mori , k. petre , r. & yamaguchi , h. 2012 , , 64 , 61 uchiyama , y. , blandford , r. d. , funk , s. , tajima , h. & tanaka , t. 2010 , , 723 , l122 vink , j. 2012 , a&ar , 20 , 49 weisskopf , m. c. , brinkman , b. , canizares , c. , garmire , g. , murray , s. & van speybroeck , l. p. 2002
, , 114 , 1 weisskopf , m. c. , karovska , m. , pavlov , g. g. , zavlin , v. e. & clarke , t. 2007 , ap&ss , 308 , 151 werner , m. w. et al .
2004 , , 154 , 1 whiteoak , j. b. z. & green , a. j. 1996 , , 118 , 329 wilson , r. w. 1963 , , 68 , 181 wilson , r. w. & bolton , j. g. 1960 , , 72 , 331 wu , e. m. h. , hui , c. y. , tam , p. h. t. , huang , r. h. h. , kong , a. k. h. & cheng , k. s. 2011 , , 740 , l12 yamauchi , s. , ueno , m. , koyama , k. & bamba , a. 2008 , pasj , 60 , 1143 yamauchi , s. , nobukawa , m. , koyama , k. & yonemori , m. 2013 , , 65 , 6 yamaguchi , h. , ozawa , m. , koyama , k. , masai , k. , hiraga , j. s. , ozaki , m. , yonetoku , d , 2009 , , 705 , 6 lccccccc [ snrpropertiestable ] ra ( j2000.0 ) & 13 05 59 & 14 05 38 & 17 10 19 & 17 14 06 & 17 15 26 & ( 1 ) + dec ( j2000.0 ) & @xmath162 42 & @xmath161 58 & @xmath140 11 & @xmath138 32 & @xmath138 28 & ( 1 ) + angular diameter ( arcmin ) & 8 & 5 & 8 & 15@xmath122 & 5@xmath123 & ( 1 ) + distance ( kpc ) & @xmath1249.7 & 12.5 & 11 & 6.3 - 9.5@xmath125 & @xmath1266.3 & ( 2 ) , ( 3 ) , ( 4 ) , ( 5 ) + physical size ( pc ) & @xmath12423 & 18 & 26 & 21 & @xmath12610 & ( 6 ) + @xmath19@xmath20 ( 10@xmath21 @xmath22 ) & @xmath33.9 & @xmath33.0 & @xmath31.7 & @xmath33.4 & & ( 6 ) + @xmath26@xmath27 ( mag ) & 19.3 & 11.4 & 14.3 & 15 & 15 & ( 6 ) lcccccccccc 57013000 & kes 17 & 1999 february 12 & 13 06 44.2 & @xmath162 40 47 & 24656 & 6.5@xmath2910@xmath72 & 24656 & 7.8@xmath2910@xmath72 & 23328 & 5.4@xmath2910@xmath72 + 56047000 & [email protected] & 1998 march 2 & 14 06 18.7 & @xmath161 53 02 & 17858 & 4.2@xmath2910@xmath127 & 17870 & 6.2@xmath2910@xmath127 & & + 54004090 & [email protected] & 1996
september 3 & 17 09 34.3 & @xmath139 59 38 & 8986 & 7.4@xmath2910@xmath72 & 8986 & 6.3@xmath2910@xmath72 & & ccccccccccc 0303100201 & kes 17 & 2005 august 25 & 13 05 40.7 & @xmath162 41 36 & 20176 & 1.3@xmath2910@xmath72 & 19895 & 1.1@xmath2910@xmath72 & 16151 & 9.7@xmath2910@xmath127 + 0306510101 & ctb 37a & 2006 march 1 & 17 14 41.0 & @xmath138 30 53 & 3090 & 7.7@xmath2910@xmath72 & 2857 & 6.5@xmath2910@xmath72 & 939 & 1.1@xmath2910@xmath127 + 0306510101 & [email protected] & 2006 march 1 & 17 14 41.0 & @xmath138 30 53 & 3090 & @xmath122.2@xmath2910@xmath72 & 2857 & @xmath129.1@xmath2910@xmath127 & 939 & @xmath123.3@xmath2910@xmath72 lccccccccc phabs@xmath29apec & 3.48@xmath128 & 0.70@xmath129 & & 3.58@xmath2910@xmath72 & 1.0 ( frozen ) & 1.18 ( 355.08/302 ) & 2.00@xmath2910@xmath130 & 7.96@xmath2910@xmath131 & 8.96@xmath2910@xmath132 + phabs@xmath29apec & 3.56@xmath133 & 0.63@xmath134 & & 1.30@xmath2910@xmath73 & 0.20@xmath135 & 1.12 ( 338.07/301 ) & 2.01@xmath2910@xmath130 & 7.09@xmath2910@xmath131 & 7.98@xmath2910@xmath132 + phabs@xmath29nei & 3.72@xmath136 & 0.64@xmath137 & 9.60 ( @xmath830.5 ) & 5.34@xmath2910@xmath72 & 1.0 ( frozen ) & 1.20 ( 359.74/301 ) & 1.96@xmath2910@xmath130 & 1.15@xmath2910@xmath138 & 1.29@xmath2910@xmath139 + phabs@xmath29nei & 3.58@xmath140 & [email protected] & 9.63 ( @xmath830.02 ) & 1.26@xmath2910@xmath73 & 0.22@xmath141 & 1.13 ( 340.38/300 ) & 2.00@xmath2910@xmath130 & 7.53@xmath2910@xmath131 & 8.48@xmath2910@xmath132 + phabs@xmath29vapec & 4.22@xmath142 & 0.60@xmath143 & & 6.41@xmath2910@xmath72 & mg=2.38@xmath55 & 1.17 ( 352.69/301 ) & 1.89@xmath2910@xmath130 & 1.43@xmath2910@xmath138 & 1.61@xmath2910@xmath139 + phabs@xmath29vapec & 4.27@xmath144 & 0.56@xmath145 & & 1.13@xmath2910@xmath73 & si=0.47@xmath56 , & 1.11 ( 334.12/300 ) & 1.99@xmath2910@xmath130 & 2.32@xmath2910@xmath138 & 2.61@xmath2910@xmath139 + & & & & & s=0.21(@xmath120.56 ) lccccccccc phabs@xmath29apec & 3.15@xmath146 & 0.92@xmath147 & & 1.67@xmath2910@xmath72 & 1.0 ( frozen ) & 1.10 ( 308.77/281 ) & 1.81@xmath2910@xmath130 & 3.35@xmath2910@xmath131 & 3.77@xmath2910@xmath132 + phabs@xmath29apec & 3.29@xmath148 & 0.82@xmath149 & & 3.17@xmath2910@xmath72 & 0.53@xmath58 & 1.09 ( 306.48/280 ) & 1.76@xmath2910@xmath130 & 3.95@xmath2910@xmath131 & 4.45@xmath2910@xmath132 + phabs@xmath29nei & 3.14@xmath150 & 1.10@xmath151 & 3.70@xmath57 & 1.32@xmath2910@xmath72 & 1.0 ( frozen ) & 1.08 ( 303.10/286 ) & 1.95@xmath2910@xmath130 & 2.56@xmath2910@xmath131 & 2.88@xmath2910@xmath132 + phabs@xmath29vapec & 3.69@xmath152 & 0.79@xmath153 & & 2.31@xmath2910@xmath72 & mg=2.75@xmath61 & 1.08 ( 303.11/280 ) & 1.77@xmath2910@xmath130 & 5.28@xmath2910@xmath131 & 5.95@xmath2910@xmath132 + phabs@xmath29vapec & 3.47@xmath154 & 0.80@xmath155 & & 2.56@xmath2910@xmath72 & si=0.79@xmath59 , & 1.08 ( 314.50/291 ) & 1.77@xmath2910@xmath130 & 5.43@xmath2910@xmath131 & 6.11@xmath2910@xmath132 + & & & & & s=0.71@xmath60 + phabs@xmath29vapec & 3.46@xmath156 & 0.90@xmath157 & & 1.82@xmath2910@xmath72 & mg=2.76@xmath158 , & 1.09 ( 302.70/278 ) & 1.81@xmath2910@xmath130 & 5.34@xmath2910@xmath131 & 6.02@xmath2910@xmath132 + & & & & & si=1.08@xmath159 , + & & & & & s=0.88@xmath160 + phabs@xmath29(vapec & 3.23@xmath161 & 0.84@xmath162 & & 3.01@xmath2910@xmath127 , & mg=9.49(@xmath831.65 ) , & 1.07 ( 292.50/274 ) & 2.05@xmath2910@xmath130 & 5.40@xmath2910@xmath131 & 6.09@xmath2910@xmath132 + + power law ) & & @xmath63=3.98@xmath163 & & 8.01@xmath2910@xmath127 & si=5.17(@xmath830.85 ) , + & & & & & s=4.88(@xmath830.44 ) + reference fits + wabs@xmath29(pshock & 3.09 ( frozen ) & 0.75 ( frozen ) & 21 ( frozen ) & 1.92@xmath2910@xmath72 , & 0.8 ( frozen ) & 1.21 ( 338.16/280 ) & 2.23@xmath2910@xmath130 & 3.83@xmath2910@xmath131 & 4.31@xmath2910@xmath132 + + power law ) & & @xmath63=2.4 ( frozen ) & & 9.12@xmath2910@xmath164 + wabs@xmath29(vmekal & 3.77 ( frozen ) & 0.78 ( frozen ) & & 2.66@xmath2910@xmath72 , & mg = 1.1 ( frozen ) , & 1.11 ( 312.68/252 ) & 1.93@xmath2910@xmath130 & 5.77@xmath2910@xmath131 & 6.50@xmath2910@xmath132 + + power law ) & & @xmath63=1.5 ( frozen ) & & 5.32@xmath2910@xmath165 & si = 0.97 ( frozen ) , + & & & & & s = 0.50 ( frozen ) + wabs@xmath29(pshock & 3.67@xmath166 & 0.75@xmath167 & 1667 ( @xmath8310 ) & 3.05@xmath2910@xmath72 , & 0.61@xmath168 & 1.11 ( 305.83/275 ) & 1.88@xmath2910@xmath130 & 4.62@xmath2910@xmath131 & 5.76@xmath2910@xmath132 + + power law ) & & @xmath63=1.58@xmath169 & & 1.17@xmath2910@xmath164 + wabs@xmath29(vmekal & 3.73@xmath170 & 0.79@xmath171 & & 1.33@xmath2910@xmath72 , & mg=4.96(@xmath831.34 ) , & 1.08 ( 298.93/276 ) & 1.92@xmath2910@xmath130 & 5.61@xmath2910@xmath131 & 6.12@xmath2910@xmath132 + + power law ) & & @xmath63=4.16@xmath172 & & 8.08@xmath2910@xmath127 & si=2.18 ( @xmath830.64 ) , + & & & & & s=1.95 ( @xmath830.32 ) + lcccccccc phabs@xmath29power law & 3.31@xmath173 & & 4.82@xmath174 & 4.49@xmath2910@xmath127 & 0.79 ( 5.56/7 ) & 2.25@xmath2910@xmath74 & 6.95@xmath2910@xmath130 & 1.30@xmath2910@xmath132 + phabs@xmath29apec & 2.69@xmath175 & 0.98@xmath75 & & 1.58@xmath2910@xmath127 & 0.57 ( 3.97/7 ) & 2.23@xmath2910@xmath74 & 3.00@xmath2910@xmath130 & 5.61@xmath2910@xmath176 lcccccccccc phabs@xmath29power law & 1.54@xmath177 & & 2.54@xmath178 & & 7.68@xmath2910@xmath127 & & 1.34 ( 66.78/50 ) & 9.07@xmath2910@xmath130 & 2.35@xmath2910@xmath138 & 3.40@xmath2910@xmath139 + phabs@xmath29apec & 1.39@xmath179 & 2.49@xmath180 & & & 1.34@xmath2910@xmath72 & 1.0 ( frozen ) & 1.45 ( 72.25/50 ) & 8.43@xmath2910@xmath130 & 1.80@xmath2910@xmath138 & 2.61@xmath2910@xmath139 + phabs@xmath29apec & 1.08@xmath181 & 3.44@xmath182 & & & 1.46@xmath2910@xmath72 & @xmath120.4 & 1.22 ( 59.88/49 ) & 8.68@xmath2910@xmath130 & 1.52@xmath2910@xmath138 & 2.20@xmath2910@xmath139 + phabs@xmath29nei & 2.13@xmath183 & 2.80@xmath77 & & 7.27@xmath78 & 1.77@xmath2910@xmath72 & 1.0 ( frozen ) & 1.07 ( 52.52/49 ) & 8.66@xmath2910@xmath130 & 1.77@xmath2910@xmath138 & 2.56@xmath2910@xmath139 + phabs@xmath29nei & 1.97@xmath184 & 2.82@xmath185 & & 9.84@xmath186 & 1.82@xmath2910@xmath72 & 0.49@xmath187 & 1.07 ( 51.38/48 ) & 8.62@xmath2910@xmath130 & 1.06@xmath2910@xmath138 & 1.53@xmath2910@xmath139 + phabs@xmath29(nei+power law ) & 1.77@xmath188 & 1.57@xmath189 & 0.5 ( frozen ) & 80 ( frozen ) & 1.81@xmath2910@xmath72 , & 0.25@xmath190 & 1.17 ( 63.06/54 ) & 1.43@xmath2910@xmath131 & 3.59@xmath2910@xmath131 & 5.20@xmath2910@xmath132 + & & & & & 3.43@xmath2910@xmath164 + phabs@xmath29(nei+power law ) & 1.76@xmath191 & 0.97@xmath192 & 2.0 ( frozen ) & 80 ( frozen ) & 2.42@xmath2910@xmath127 , & 2.67 ( @xmath830.56 ) & 1.16 ( 62.37/54 ) & 1.10@xmath2910@xmath131 & 4.46@xmath2910@xmath131 & 6.46@xmath2910@xmath132 + & & & & & 4.13@xmath2910@xmath127 + phabs@xmath29(nei+power law ) & 1.74@xmath193 & 1.50@xmath194 & 0.5 ( frozen ) & 100 ( frozen ) & 1.88@xmath2910@xmath72 , & 0.23@xmath195 & 1.19 ( 64.21/54 ) & 1.45@xmath2910@xmath131 & 3.37@xmath2910@xmath131 & 4.88@xmath2910@xmath132 + & & & & & 3.57@xmath2910@xmath164 + phabs@xmath29(nei+power law ) & 1.79@xmath196 & 0.87@xmath197 & 2.0 ( frozen ) & 100 ( frozen ) & 5.46@xmath2910@xmath127 , & 1.32 ( @xmath830.42 ) & 1.16 ( 62.82/54 ) & 1.09@xmath2910@xmath131 & 4.50@xmath2910@xmath131 & 6.52@xmath2910@xmath132 + & & & & & 4.09@xmath2910@xmath127 + reference fits + wabs@xmath29(vmekal & 2.0 ( frozen ) & 0.97 ( frozen ) & 0.5 ( frozen ) & & 2.29@xmath2910@xmath72 , & varied & 1.16 ( 65.02/56 ) & 1.63@xmath2910@xmath131 & 4.91@xmath2910@xmath131 & 7.11@xmath2910@xmath132 + + power law ) & & & & & 4.5@xmath2910@xmath164 + wabs@xmath29(vnei & 2.1 ( frozen ) & 1.22 ( frozen ) & 0.5 ( frozen ) & 292 ( frozen ) & 2.29@xmath2910@xmath72 , & varied & 1.16 ( 65.05/56 ) & 1.63@xmath2910@xmath131 & 4.91@xmath2910@xmath131 & 7.11@xmath2910@xmath132 + + power law ) & & & & & 4.5@xmath2910@xmath164 + wabs@xmath29(vpshock & 2.1 ( frozen ) & 1.3 ( frozen ) & 0.6 ( frozen ) & 692 ( frozen ) & 2.29@xmath2910@xmath72 , & varied & 1.15 ( 64.50/56 ) & 1.63@xmath2910@xmath131 & 4.91@xmath2910@xmath131 & 7.11@xmath2910@xmath132 + + power law ) & & & & & 4.5@xmath2910@xmath164 + lccccccccc phabs@xmath29apec & 3.15@xmath198 & 0.73@xmath199 & & 3.37@xmath2910@xmath72 & 1.0 ( frozen ) & 1.11 ( 207.36/192 ) & 2.32@xmath2910@xmath130 & 7.43@xmath2910@xmath131 & 5.69@xmath2910@xmath132 + phabs@xmath29apec & 3.26@xmath200 & 0.64@xmath201 & & 1.03@xmath2910@xmath73 & 0.29@xmath202 & 1.08 ( 200.04/186 ) & 2.28@xmath2910@xmath130 & 7.73@xmath2910@xmath131 & 5.92@xmath2910@xmath132 + phabs@xmath29nei & 3.53@xmath203 & 0.62@xmath204 & 357 ( @xmath833 ) & 6.29@xmath2910@xmath72 & 1.0 ( frozen ) & 1.11 ( 207.13/186 ) & 2.24@xmath2910@xmath130 & 1.35@xmath2910@xmath138 & 1.03@xmath2910@xmath139 + phabs@xmath29nei & [email protected] & 0.64@xmath205 & 6.73 ( @xmath830.52 ) & 1.05@xmath2910@xmath73 & 0.35@xmath206 & 1.09 ( 201.19/185 ) & 2.29@xmath2910@xmath130 & 9.37@xmath2910@xmath131 & 7.18@xmath2910@xmath132 + phabs@xmath29vapec & 3.71@xmath207 & 0.55@xmath208 & & 9.29@xmath2910@xmath72 & si=0.48@xmath209 & 1.05 ( 198.98/186 ) & 2.24@xmath2910@xmath130 & 1.91@xmath2910@xmath138 & 1.46@xmath2910@xmath139 + phabs@xmath29vnei & 3.69@xmath210 & 0.57@xmath211 & 43 ( @xmath831 ) & 9.36@xmath2910@xmath72 & si=0.55@xmath212 & 1.07 ( 198.82/185 ) & 2.25@xmath2910@xmath130 & 1.95@xmath2910@xmath138 & 1.49@xmath2910@xmath139 + reference fits & & & + phabs@xmath29(vmekal & 2.90 ( frozen ) & 0.63 ( frozen ) & & 3.14@xmath2910@xmath72 , & 1.0 ( frozen ) & 1.13 ( 214.80/190 ) & 3.20@xmath2910@xmath130 & 7.41@xmath2910@xmath131 & 5.68@xmath2910@xmath132 + + power law ) & & and @xmath63=1.6 ( frozen ) & & 3.40@xmath2910@xmath164 + lcccccccccc + phabs@xmath29power law & 4.08@xmath88 & & 1.18@xmath87 & & 3.97@xmath2910@xmath164 & & 0.84 ( 91.97/109 ) & 2.99@xmath2910@xmath130 & 4.49@xmath2910@xmath130 & 3.43@xmath2910@xmath176 + + phabs@xmath29apec & 3.07@xmath213 & 0.80@xmath214 & & & 1.31@xmath2910@xmath72 & 1.0 ( frozen ) & 1.09 ( 144.42/132 ) & 1.12@xmath2910@xmath130 & 1.88@xmath2910@xmath131 & 1.44@xmath2910@xmath132 + phabs@xmath29apec & [email protected] & 0.78@xmath215 & & & 2.30@xmath2910@xmath72 & 0.45@xmath216 & 1.06 ( 138.81/131 ) & 1.13@xmath2910@xmath130 & 1.64@xmath2910@xmath131 & 1.26@xmath2910@xmath132 + phabs@xmath29nei & 3.30@xmath217 & 0.83@xmath218 & & 4.15 ( @xmath832.55 ) & 1.53@xmath2910@xmath72 & 1.0 ( frozen ) & 1.07 ( 140.38/131 ) & 1.16@xmath2910@xmath130 & 2.42@xmath2910@xmath131 & 1.85@xmath2910@xmath132 + phabs@xmath29nei & 3.24@xmath219 & 0.80@xmath220 & & 2.44 ( @xmath830.25 ) & 2.63@xmath2910@xmath72 & 0.43@xmath221 & 1.02 ( 132.56/130 ) & 1.16@xmath2910@xmath130 & 2.14@xmath2910@xmath131 & 1.64@xmath2910@xmath132 + + phabs@xmath29apec & 3.03@xmath222 & [email protected] & & & 1.79@xmath2910@xmath72 & 1.0 ( frozen ) & 0.95 ( 95.32/100 ) & 7.11@xmath2910@xmath74 & 1.55@xmath2910@xmath131 & 1.19@xmath2910@xmath132 + phabs@xmath29apec & 3.02@xmath223 & [email protected] & & & 2.30@xmath2910@xmath72 & 0.73@xmath224 & 0.96 ( 95.02/99 ) & 7.12@xmath2910@xmath74 & 1.51@xmath2910@xmath131 & 1.16@xmath2910@xmath132 + phabs@xmath29nei & 3.31@xmath225 & 0.55@xmath226 & & 4.23(@xmath830.86 ) & 2.76@xmath2910@xmath72 & 1.0 ( frozen ) & 0.94 ( 93.08/99 ) & 7.36@xmath2910@xmath74 & 2.51@xmath2910@xmath131 & 1.92@xmath2910@xmath132 + phabs@xmath29nei & 3.32@xmath227 & [email protected] & & 3.34(@xmath830.52 ) & 2.79@xmath2910@xmath72 & 0.86 ( @xmath830.34 ) & 0.95 ( 93.04/98 ) & 7.37@xmath2910@xmath74 & 2.55@xmath2910@xmath131 & 1.95@xmath2910@xmath132 [ acisspectrasummaryctb37a ] lcccc kes 17 ( [email protected] ) & @xmath4@xmath5 = 0.42 @xmath50 & @xmath4@xmath5 = 0.9 - 2.3 @xmath50 & ( 1 ) & ( 1 ) calculated age from @xmath6 = @xmath48/@xmath4@xmath5 . + & @xmath48 @xmath3 10@xmath228 s @xmath50 & @xmath48 = 1 - 3@xmath2910@xmath228 s @xmath50 + & @xmath6@xmath7 = 2.3@xmath2910@xmath108 yr & @xmath6 = 2.8 - 6.4@xmath2910@xmath108 yr + & @xmath8@xmath9 = 52 @xmath8@xmath110 & @xmath8@xmath9 = 8 - 15 @xmath8@xmath110 + & & @xmath4@[email protected] @xmath106@xmath229 & ( 2 ) & ( 2 ) assumed distance of 10 kpc + & & @xmath48@xmath310@xmath228 s @xmath50 & & and adopted @xmath106 as filling factor .
+ & & @[email protected]@xmath2910@xmath108@xmath106@xmath230 yr + & & @xmath8@xmath9@xmath3380@xmath106@xmath230 @xmath8@xmath110 + [email protected] & @xmath4@xmath5 = 0.24 @xmath50 & @xmath4@xmath5 = 0.82 @xmath50 & ( 3 ) & ( 3 ) assumed distance of 8.3 kpc + & @xmath48 @xmath3 10@xmath52 s @xmath50 & @xmath48 = 2.92@xmath2910@xmath52 s @xmath50 & & and calculated age from @xmath6=@xmath48/@xmath4@xmath5 .
+ & @xmath6@xmath7 = 4.7@xmath2910@xmath108 yr & @xmath6 = 1.1@xmath2910@xmath108 yr + & @xmath8@xmath9 = 84 @xmath8@xmath110 & & & + ctb 37a ( [email protected] ) & @xmath4@xmath5 = 0.77 @xmath50 & @xmath4@xmath5 = 1 @xmath50 & ( 4 ) & ( 3 ) assumed distance of 11.3 kpc + & @xmath48 @xmath23110@xmath228 s @xmath50 & @xmath48 = @xmath23110@xmath228 s @xmath50 & & and a filling factor @xmath106=1 .
age + & @xmath6@xmath7 = 3.2 - 4.2@xmath2910@xmath108 yr & @xmath6 = 3@xmath2910@xmath108 yr & & calculated from @xmath6=@xmath48/@xmath4@xmath5 .
+ & @xmath8@xmath9=76 @xmath8@xmath110 & @xmath8@xmath9 = 530 @xmath8@xmath110 + lcccc radius ( pc ) & @xmath12411 & 9 & 13 & 10 + @xmath99@xmath102 ( km s@xmath73 , radiative ) & 150 & 100 , 170 & 70 - 100 & 75 - 100 + @xmath6@xmath7 ( 10@xmath108 yr ) & 2.3 & 2.5 - 4.2 & 3.3 - 4.7 & 3.2 - 4.2 + x - ray mechanism & thermal & likely thermal & thermal & thermal + @xmath2 ( kev ) & 0.79 - 1.10 & 0.98 & 2.49 - 3.44 & 0.55 - 0.83 + @xmath48 ( @xmath50 s ) & 3.7@xmath2910@xmath52 & @xmath3@xmath8310@xmath228 & 7.3@xmath2910@xmath79 & @xmath8310@xmath228 + @xmath4@xmath5 ( @xmath50 ) & 0.42 & @xmath830.40 & 0.25 & 0.77 + @xmath8@xmath9 ( @xmath8@xmath110 ) & 52 & 13 & 88 & 76 + morphology & center - filled & center - filled & center - filled & center - filled + snr classification & mm snr & mm snr & mm snr & mm snr + evidence of molecular cloud interaction & yes ( h@xmath18 ) & yes ( h@xmath18 ) & yes ( h@xmath18 ) & yes ( h@xmath18 ) + @xmath46-ray detection & yes & no & no & yes [ snrxraystable ] | we present a study of the x - ray properties of five galactic supernova remnants ( snrs ) kes 17 ( [email protected] ) , [email protected] , [email protected] , ctb 37a ( [email protected] ) and [email protected] that were detected in the infrared by @xcite in an analysis of data from the galactic legacy infrared mid - plane survey extraordinaire ( glimpse ) that was conducted by the _ spitzer _ space telescope .
we present and analyze archival _ asca _ observations of kes 17 , [email protected] and [email protected] , archival _ xmm - newton _ observations of kes 17 , ctb 37a and [email protected] and an archival _ chandra _ observation of ctb 37a .
all of the snrs are clearly detected in the x - ray possibly except for [email protected] .
our study reveals that the four detected snrs all feature center - filled x - ray morphologies and that the observed emission from these sources is thermal in all cases .
we argue that these snrs should be classified as mixed - morphology snrs ( mm snrs ) : our study strengthens the correlation between mm snrs and snrs interacting with molecular clouds and suggests that the origin of mixed - morphology snrs may be due to the interactions between these snrs and adjacent clouds .
our _ asca _ analysis of [email protected] reveals for the first time x - ray emission from this snr : the x - ray emission is center - filled within the radio and infrared shells and thermal in nature ( @xmath2 @xmath3 0.98 kev ) , thus motivating its classification as an mm snr .
we find considerable spectral variations in the properties associated with the plasmas of the other x - ray - detected snrs , such as a possible overabundance of magnesium in the plasma of kes 17 .
our new results also include the first detailed spatially - resolved spectroscopic study of ctb 37a using _ chandra _ as well as a spectroscopic study of the discrete x - ray source cxou j171428.5@xmath1383601 , which may be a neutron star associated with ctb 37a .
finally , we also estimate such properties as electron density @xmath4@xmath5 , radiative age @xmath6@xmath7 and swept - up mass @xmath8@xmath9 for each of the four x - ray - detected snrs .
each of these values are comparable to archetypal mixed - morphology snrs like 3c 391 and w44 . in an analysis of the spectrum of kes 17
, we did not find evidence of over - ionization unlike other archetypal mixed - morphology snrs like w28 and w44 . |
globular star clusters ( gcs ) are pressure - supported collections of between @xmath4 and @xmath5 stars which are usually associated with galaxies , although there is evidence that some clusters of galaxies contain a population of `` free '' gcs which are associated with the cluster potential as a whole and not any individual galaxy ( west et al . 1995 ) .
individual galactic gcs are made up of stars with a single overall chemical abundance suggesting that they formed in a single star formation event .
gcs typically have integrated magnitudes of @xmath6 making gc systems visible out to redshifts of @xmath7 , the approximate distance to the great wall galaxies .
the shape of the gc luminosity function ( lf ) has been assumed to be universal , so the gc lf has been used as a distance indicator ( see harris 1991 for a review ) .
however , recent work has suggested that the shape of the gc lf may depend on the metallicity of the gc system ( ashman et al .
1995 ) and the details of stellar evaporation from individual gcs ( okazaki & tosa et al .
1995 ) . because of the wide - spread use of gcs to determine distances to external galaxies , it is important to determine if gcs truly are the same from one galaxy to the next .
this is best determined by studying the physical structures and stellar populations of gcs in nearby galaxies .
the nearest large gc system outside the milky way galaxy is that of the andromeda galaxy ( @xmath8 m31 @xmath8 ngc 224 ) .
m31 is located at a distance of 725 kpc ( @xmath9 , van den bergh 1991 ) so individual stars in m31 s gcs can be easily resolved with the _ hubble space telescope _ ( _ hst _ ) and large ground - based telescopes at sites with sub - arcsecond seeing .
m31 s low inclination ( @xmath10 , hodge 1992 ) means that many of its gcs are not superimposed against the disk of m31 , making identification of gcs , and photometry of their stars , relatively straightforward .
m31 offers a unique laboratory to study the outer regions of gcs . because of their small fields - of - view it is not practical to used ccds to obtain star counts in the outer regions of galactic gcs . beyond distances of approximately half the tidal radius , @xmath11 ,
the projected stellar densities of the clusters are overwhelmed by random fluctuations in the background stellar number density ( innanen et al .
. however , m31 s gcs have sufficiently small angular sizes ( @xmath12 to @xmath13 ) that both the cluster and the background can be imaged on a single large format ccd image .
this eliminates the need to match photometric zero - points between the cluster and the background , which makes possible a more precise subtraction of the background light from the cluster light .
the first study of the internal structures of m31 gcs was undertaken by battistini et al .
( 1982 ) , who estimated core radii for several clusters .
pritchet & van den bergh ( 1984 ) found that the surface brightness profile for g1 ( @xmath8 mayall ii ; the g - numbers used in this paper are from sargent et al .
[ 1977 ] ) had an excess of light at large radii compared to the best - fitting seeing - convolved analytical king ( 1966 ) model .
they did find that g1 was well fit by empirical king ( 1962 ) models with core radii of @xmath14 pc .
crampton et al .
( 1985 ) used seeing - convolved king ( 1962 ) models to derive core radii for nearly 500 m31 gcs .
their values , however , are systematically @xmath1550% larger than those of battistini et al .
( 1982 ) , despite the better seeing conditions of the crampton et al .
( 1985 ) data set .
bendinelli et al .
( 1990 ) used ground - based data to produce seeing - deconvolved radial profiles for six bright gcs in m31 , but seeing and pixel scale limitations restricted them to resolutions of @xmath1503 , insufficient to resolve the cores of the clusters .
still , their data suggest that m31 gcs have king - like profiles similar to those of galactic gcs .
the core structures of some of m31 s gcs have been studied using the pre - refurbished _
bendinelli et al . (
1993 ) detected a power - law density cusp in g105 using _ hst _
s faint object camera ( foc ) images and a variety of image restoration and seeing deconvolution techniques .
in addition , fusi pecci et al .
( 1994 ) used similar methods to obtain half - width at half - maxima ( hwhm ) and half - light radii ( @xmath16 ) for thirteen m31 gcs from foc images .
they found hwhms similar to the core radii of galactic gcs .
their data , however , could not be used to find the tidal radii of these clusters because the point - spread function ( psf ) before the refurbishment of the _ hst _ overfilled the foc s field of view .
cohen & freeman ( 1991 ) derived tidal radii for thirty m31 gcs by fitting seeing - convolved king models .
although their fits to individual clusters were quite uncertain they did find a mean tidal radius for the m31 clusters after adjustment for differences in galactic masses and rotation velocities which was very similar to that of the milky way clusters .
gcs do not exist in isolation but sit in the tidal field of a galaxy .
any stars that evaporate from a gc by the galaxy will have velocity vectors similar to the velocity vector of the gc .
this can result in the gc being surrounded by an extended halo of unbounded stars which move in approximately the same direction as the gc and have approximately the same velocity .
this idea has been explored numerically by oh & lin ( 1992 ) who predicted that gcs could be surrounded by extended halos of escaped stars which can persist for up to a hubble time . the size and shape of such a halo can be influenced by the tidal field of the parent galaxy .
evidence for extended halos has been observed in some galactic gcs by grillmair et al .
in addition , grillmair et al .
( 1996 , hereafter referred to as gaf ) have observed an excess of resolved and unresolved stars beyond the formal michie king ( mk ) tidal radii of several gcs in m31 , as would be expected if extended halos were present .
there has been some interest in determining the ellipticities of m31 gcs .
pritchet & van den bergh ( 1984 ) measured an ellipticity of @xmath17 for the region of g1 with @xmath18 .
spassova et al .
( 1988 ) measured ellipticities for approximately two dozen gcs while a study by lupton ( 1989 ) suggests that the mean ellipticity measured in the inner 7 to 14 pc of a m31 gc ( @xmath19 ) is indistinguishable from the mean ellipticity of galactic gcs .
in the outer 14 to 21 pc , however , the mean ellipticity of an m31 cluster is @xmath20 , greater than that of the galactic gcs ( @xmath21 ) and similar to gcs in the large magellanic cloud ( @xmath22 ) .
baev et al .
( 1997 ) found systematic differences between the shapes of m31 s disk and halo gcs .
they found that the disk gcs are triaxial ellipsoids while the halo gcs are oblate or prolate spheroids , but cautioned that this is a preliminary result since the sample of halo gcs in their study is small compared to the sample of disk gcs . the first color magnitude diagrams ( cmds ) for gcs in m31 were for g1 by heasley et al .
( 1988 ) and g219 ( @xmath8 mayall iv ) by christian & heasley ( 1991 ) .
couture et al .
( 1995 ) undertook a systematic study of five m31 gcs ( g11 , g319 , g323 , g327 @xmath8 mayall vi , and g352 @xmath8 mayall v ) with a range of iron abundances .
unfortunately none of these ground - based studies were able to reach the level of the horizontal branch ( hb ) at @xmath23 .
the first cmds constructed from _
data were for g1 ( rich et al .
1996 ) ; g58 , g105 , g108 , and g219 ( ajhar et al .
1996 ) ; and g280 , g351 , and bo468 ( fusi pecci et al .
these cmds were able to resolve stars one to two magnitudes below the red portion of the hb . in order to study the stellar populations and structures of gcs in m31 we obtained deep _
hst _ wide - field / planetary camera-2 ( wfpc2 ) images of two halo gcs ( g302 and g312 ) located @xmath24 and @xmath25 respectively from the center of m31 along the southeast minor axis . our data is unique in that we have centered the gcs on the wf3 ccd in order to take advantage of the wfc s field of view to search for extended halos around each gc .
we obtained deep _ v_- and _ i_-band images of two bright gcs , g302 ( @xmath26 ) and g312 ( @xmath27 ) in the halo of m31 .
these data were obtained using the _ hst _ in cycle 5 program # 5609 .
we used the wfpc2 operating at @xmath28c with a gain setting of 7 @xmath29/adu .
table [ table : obs_log ] lists the exposures obtained in each filter .
exposure times were based on the central surface brightnesses of the two gcs as determined from ground - based observations ( holland et al . 1993 ,
holland 1997 ) and chosen to avoid saturating the cores of the gcs .
the total exposure times were 4320 seconds in the f555w ( wfpc2 broadband _ v _ ) filter and 4060 seconds in the f814w ( wfpc2 broadband _ i _ ) filter for each field .
the data was pre - processed through the standard stsci pipeline for wfpc2 data .
known bad pixels were masked , and geometric corrections were applied using standard techniques .
g302 is located @xmath30 from the center of m31 ( a projected distance of @xmath31 kpc ) approximately along the southeast minor axis while g312 is located @xmath32 from the center of m31 ( @xmath33 kpc ) approximately along the southeast minor axis .
each gc was centered on the wf3 ccd .
we chose to locate the gcs on the wf3 ccd , instead of taking advantage of the higher spatial resolution offered by the smaller pixels of the pc ccd , for three reasons : first , the tidal radii of g302 and g312 , as determined by fitting mk models ( michie 1963 , king 1966 ) to the surface brightness profiles of g302 and g312 derived from hrcam observations taken at the cfht , are @xmath34 ( holland et al .
1993 , holland 1997 ) .
therefore , each gc would occupy a significant fraction of the total field of view of the pc , leaving insufficient area to make reliable determinations of the spatial variations in the background stellar density .
second , grillmair et al .
( 1995 ) found evidence for extended stellar halos , extending to @xmath15 2 to 3 times the tidal radius , in several galactic gcs .
recently , gaf and holland ( 1997 ) found weak evidence for stellar density enhancements out to @xmath35 in some m31 gcs including g302 and g312 . by centering the gcs on the wf3 ccd we ensured that as much as possible of any extended halos would be located on the same ccd as the gc itself .
this enabled us to avoid systematic effects that may occur if surface brightness and stellar number density data are transferred between the pc and wfc ccds , with their different pixel scales and sensitivities .
third , by locating the gcs at the center of the wf3 ccd we were able to use the wf2 and wf4 ccds to probe the stellar populations in the halo of m31 .
these results are published in holland et al .
( 1996 , hereafter referred to as hfr ) .
having two background fields for each gc field enables us to differentiate between stellar number - density enhancements that are due to statistical fluctuations in the star counts from m31 halo stars , and those that are due to physical structures , such as extended halos of unbound stars , that are related to the gcs .
photometry was performed on stars imaged by the wfpc2 s wide field ccds using the daophot ii / allframe software ( stetson 1987 , stetson 1994 ) .
all nineteen images for each field were re - registered to a common coordinate system and median - combined .
this eliminated cosmic ray hits and increased the signal - to - noise ratio ( s / n ) in our data allowing fainter stars to be detected . in order to eliminate large - scale gradients in the images due to unresolved light from the gcs ( the gradient in the light from the halo of m31 was @xmath36 1% in both the g302 and g312 fields )
a square median - filter was run over the combined image to produce a map of the large - scale gradients .
this smoothed image was then subtracted from the combined image and a constant sky added back to produce a median - subtracted image .
the size of the square median - filter was @xmath155 times the stellar full - width at half - maximum ( fwhm ) to avoid smoothing out structure in the stellar images ( e.g. stetson & harris 1988 ) . subtracting the median - filtered image resulted in a reasonably clean subtraction of the integrated light from each gc to within @xmath15 25 pixels ( @xmath37 25 @xmath37 9 pc ) of the center of each gc .
this process is the digital equivalent of unsharp masking .
the daophot find routine , with a find threshold of @xmath38 , was used to detect peaks on the median - subtracted image .
we experimented with different find thresholds and found that @xmath38 excluded most of the mis - identifications of noise spikes and cosmic - ray events as stars at faint magnitudes while including almost all of the stars to the photometric limit of the data .
find was able to reliably detect stars to within @xmath15 25 pixels ( @xmath15 25 ) of the center of each gc .
artificial star tests showed that detections closer to the centers of the gcs than this were usually spurious .
the resulting list of stellar candidates was used as input to allframe .
allframe does simultaneous psf fitting on all the _ original _ images ( _ not _ the combined median - subtracted image ) to preserve the photometric properties of the data .
the psfs used ( stetson 1996 ) were moffatians ( moffat 1969 ) with @xmath39 and a look - up table of residuals .
the psfs varied quadratically over each ccd to account for variations in the form of the psf from one part of the ccd to another .
only stars which appeared in at least seven frames for each filter were considered to be real stars .
tests with different minimum numbers of frames showed that this requirement provides a clean cmd without significantly affecting the limiting magnitude .
aperture corrections and photometric calibrations were performed in exactly the same manner as described in hfr .
we adopted a distance modulus of @xmath40 , a reddening of @xmath41 , and interstellar extinctions from tables 12a and 12b of holtzman et al .
these extinction values are dependent on the stellar spectrum . for a k5 giant
the extinctions were @xmath42 and @xmath43 .
see hfr for a discussion of why these values were adopted .
table [ table : apcor ] lists the adopted aperture corrections for the wfpc2 ccds .
we were unable to obtain reliable aperture corrections for the pc images of the g312 field due to a lack of bright stars in this field .
photometry in this field exhibited an unusually high degree of scatter , although the stellar images appeared to be normal to the eye .
as a result , we have not included this data in any of the analysis presented here .
allframe s @xmath44 statistic represents the ratio of the observed pixel - to - pixel scatter in the psf fitting residuals to the expected amount of scatter given the noise properties of the data .
values of @xmath44 which are significantly greater than unity indicate that the image was not well fit by the stellar psf and thus may not be a star . in light of this
we discarded any stars with a final @xmath44 value of greater than 2 . in order to eliminate stars with poorly determined magnitudes we used an interactive approach based on the uncertainty in the calibrated magnitude .
a plot of uncertainty vs.magnitude was made and a locus defined which corresponded to the expected photometric scatter at a given magnitude .
any stars which were judged to have a significantly greater uncertainty than the typical value for that magnitude were discarded .
visual examinations of the cmds and the wf images suggested that we did not remove any legitimate stars from our sample .
we wish to stress that this culling was done _
solely _ on the basis of the photometric uncertainty of each star , not on the star s location on the cmd or the ccd .
mean photometric uncertainties for our entire culled data set are listed in table 3 of hfr .
the mean photometric uncertainties , based on the formal photometric uncertainties returned by daophot ii / allframe , for the gc stars are listed in table [ table : mag_errs ] of this paper .
the complete set of calibrated stellar photometry for g302 and the surrounding fields is listed in table [ table : g302_sample_phot ] .
the photometry for g312 and its surrounding fields ( with the exception of the pc field ) is listed in table [ table : g312_sample_phot ] .
column ( 1 ) gives the ccd the star was observed on .
column ( 2 ) gives an identification number for each star .
the identification numbers are only unique for stars on a given ccd .
columns ( 3 ) and ( 4 ) give the location of the star on the ccd in pixel coordinates . column
( 5 ) is the _
v_-band magnitude , column ( 6 ) the 1-@xmath45 uncertainty in the _
v_-band magnitude , and column ( 7 ) is the value of the daophot @xmath44 statistics .
columns ( 8) to ( 10 ) are the same as columns ( 5 ) to ( 7 ) except they give values for the _ i_-band .
tables [ table : g302_gc_sample_phot ] and [ table : g312_gc_sample_phot ] list the photometry for stars with @xmath46 of the center of g302 and g312 .
the data is the same as for tables [ table : g302_sample_phot ] and [ table : g312_sample_phot ] except an extra column , giving the distance of the star from the center of the gc , has been added .
only the first ten lines of tables [ table : g302_sample_phot ] , [ table : g312_sample_phot ] , [ table : g302_gc_sample_phot ] , and [ table : g312_gc_sample_phot ] are published here .
the full tables can be found on the aas cd - rom volume 9 .
a total of 12,289 stars were found in the three g302 wfc fields while 3,355 stars were found in the three g312 wfc fields .
there are 626 stars within @xmath47 of the center of g302 and 325 stars within @xmath47 of the center of g312 . in order to determine the effects of crowding , as well as the efficiencies of our star - finding and photometry techniques , on our photometry
, we undertook a series of artificial star tests .
we used the following procedure to generate and recover artificial stars : 1 . for each field we generated a series of artificial stars with @xmath48 colors based on the observed cmd and lf of the m31 halo stars .
the artificial stars were added to each of the original images for a field .
the total number of artificial stars added in each field was @xmath49% of the total number of stars in that field .
3 . the entire photometry process described in [ section : data ] ( find / allframe / culling / matching ) was performed on this new set of images and the resulting photometry was matched against the original list of artificial stars for that field .
this process was repeated until we had recovered approximately the same number of artificial stars as there were real stars found on that ccd .
this process was performed for each ccd field .
figure [ figure : art_shift ] shows the relationship between the input magnitudes of the artificial stars and their recovered magnitudes for a typical set of stars , in this case using the wf2 field near the gc g312 .
artificial stars with magnitude differences of more than @xmath50 mag have been omitted .
this figure plots the magnitude difference as a function of the recovered , not the input , magnitudes of the artificial stars .
this was done because the observed magnitudes of the real stars are the recovered magnitudes , not the actual ( or input ) magnitudes .
therefore , figure [ figure : art_shift ] can be used to directly determine the uncertainty in a star s observed magnitude .
smirnov ( ks ) tests show that the distributions of recovered magnitudes for the artificial red - giant branch ( rgb ) and hb stars are the same at the 98.892% confidence level in the _ v_-band and at the 99.999% confidence level in the _ i_-band .
the apparent excess scatter in the hb artificial stars is merely an artifact of the large number of artificial hb stars .
table [ table : art_shift_rgb ] lists the systematic shifts in @xmath51 and @xmath52 as well as the degree of scatter as a function of magnitude for the artificial red - giant branch ( rgb ) stars of g312 .
table [ table : art_shift_hb ] lists these quantities for the g312 artificial hb stars . in these tables
the notation @xmath53 $ ] indicates the median value of the individual @xmath54 values .
the columns are ( 1 ) the boundaries of the bin , ( 2 ) the median recovered magnitude , @xmath55 $ ] ; ( 3 ) the median photometric uncertainty in the recovered magnitudes ( from daophot ) , @xmath56 $ ] ; ( 4 ) the median magnitude difference , @xmath57 \equiv [ v_{\rm recovered } - v_{\rm input}]$ ] ; ( 5 ) the amount of scatter present in the artificial stars in that bin , @xmath58|\bigr]$ ] ; columns ( 6 ) to ( 9 ) are the same quantities for the @xmath52 color ; and ( 10 ) is the number of artificial stars in that bin .
these quantities are identical to those described by stetson & harris ( 1988 ) except for @xmath59 , which we defined in non - parametric terms instead of in gaussian terms .
our reduction process has a slight tendency to underestimate the brightnesses of stars near the photometric limit of our data ( see column 4 of table [ table : art_shift_rgb ] ) .
this systematic shift is less than the internal scatter in the photometry and is smaller than @xmath60 mag for rgb stars at the level of the hb . therefore , we conclude that the systematic uncertainties in our photometric reductions do not affect the morphology of the cmds above @xmath61 . figure [ figure : art_rgb ] shows a cmd created from a set of artificial rgb stars added to the g312 wf3 image .
the hb stars have been omitted in order to show the increase in the scatter in the recovered colors at fainter magnitudes .
the amount of scatter seen at a particular magnitude is comparable to the amount of scatter seen in the cmd of g312 , which is believed to be made up of stars with a single metallicity .
if the stellar populations in m31 gcs are analogous to the stellar populations in galactic gcs , then the observed spread in g312 s rgb is primarily due to photometric uncertainties .
figure [ figure : art_rgb ] shows that photometric uncertainties are sufficient to explain all the observed width of the rgb of g312 .
table [ table : art_shift_hb ] and figure [ figure : art_hb ] show that hb stars , because of the large color spread in the hb , can suffer badly from systematic differences between their input and recovered magnitudes .
blue hb stars are scattered by as much as 0.5 mag . in @xmath52 , while red hb stars have a tendency to be recovered at fainter magnitudes and redder colors than they really are .
this tends to move red hb stars into the lower half of the rgb resulting in our star counts underestimating of the total number of red hb stars in the gc .
the artificial star tests show that stars from the red and blue sides of the hb can be scattered into the rr lyrae gap .
this making it difficult to distinguish rr lyrae candidates from stars which have simply scattered into the rr lyrae gap .
there are four possible sources of contamination in the cmds of the gcs .
the first three galactic halo stars , background galaxies , and m31 disk stars are discussed in 3.1 of hfr . by scaling those results to the area on the sky covered by each gc
we estimate that at most @xmath62 of the objects with @xmath63 in each of figures [ figure : g302_cmd ] ( g302 ) and [ figure : g312_cmd ] ( g312 ) are due to contamination from these three sources .
the fourth source of contamination is the halo of m31 itself . to estimate the number of halo stars that should be visible in the cmd for each
gc we computed the stellar surface densities in the wf2 and wf4 ccds for each field .
these ccds were centered @xmath64 from the center of the wf3 ccd ( and thus the gcs ) .
since g302 and g312 have tidal radii of @xmath65 the wf2 and wf4 ccds will not suffer from contamination from gc stars .
the mean background for m31 halo stars with @xmath63 near g302 is @xmath66 stars/ , corresponding to @xmath67 halo stars in the rgb for g302 . near g312
the background is @xmath68 stars/ , corresponding to @xmath69 stars in the rgb of g312 .
these estimates are reasonably consistent with the numbers of objects in the upper halves of figures [ figure : g302_cmd ] and [ figure : g312_cmd ] that do not fall on either the rgb or the asymptotic - giant branch ( agb ) of each gc . on the basis of these calculations
we expect that contaminating objects make up @xmath70% of the objects in the g302 cmd .
similarly contaminating objects make up @xmath71% of the objects in the g312 cmd .
these numbers are consistent with the number of discrepant objects in each cmd so we conclude that the morphologies of the cmds in figures [ figure : g302_cmd ] and [ figure : g312_cmd ] are real and not due to contamination . figure [ figure : g302_cmd ] shows the cmd for the gc g302 .
stars located between @xmath72 from the center of the cluster are shown .
the @xmath73 cut - off approximately corresponds to the tidal radius of the gc while the @xmath74 cut - off is the point where extreme crowding conditions make it impossible to obtain reliable stellar photometry . to estimate the metallicity of g302 we took the dereddened fiducial rgb sequences for several galactic gcs ( from da costa & armandroff 1990 , hereafter referred to as dca ) , shifted them to the distance of m31 ( @xmath9 ) , and overlaid them onto figure [ figure : g302_cmd ] .
the lack of a turn - over at the top of the g302 rgb is consistent with a metal - poor population with an iron abundance of @xmath1}}\simeq -1.5 $ ] to @xmath75 .
we interpolated between the dca fiducial sequences to obtain @xmath1}}= -1.85 \pm 0.12 $ ] for g302 , where the uncertainty is one - third of the spread in the rgb and corresponds to approximately one standard deviation if we assume the spread in the rgb is due to normally distributed errors in the photometry .
our estimate of the iron abundance for g302 agrees with spectroscopic estimates by huchra et al .
( 1991 , hereafter referred to as hbk ) , @xmath1}}= -1.76 \pm 0.18 $ ] , and de freitas pacheco ( 1997 ) , @xmath1}}= -1.80 \pm 0.2 $ ] .
agb stars in g302 can be seen to the blue of the low - metallicity fiducial sequences in figure [ figure : g302_cmd ] .
unfortunately the agb and the rgb blend together so it is not possible to unambiguously distinguish agb stars from rgb stars .
this is due to photometric scatter ( @xmath76 at @xmath77 ) and contamination from rgb stars in the halo of m31 .
no stars brighter than the tip of the rgb are present in the data .
the lack of such super - luminous stars supports the hypothesis that gcs in m31 have ages similar to those of the galactic gcs .
spectral line blanketing in cool giants can cause very red giants to appear fainter than the bluer , less evolved , rgb stars .
this can result in some massive evolved stars appearing to have @xmath78 colors considerably redder than the rest of the gc stars .
our artificial star tests suggest that we are able to detect any stars above the hb so we are confident that we have not missed any super - luminous agb stars in the outer regions of g302 .
however , since these stars , being more evolved , will have slightly more massive progenitors than the rest of the rgb and agb stars , it is possible that mass segregation will have concentrated them in the inner @xmath74 of the gc making them undetectable with our data .
to date no super - luminous agb stars have been detected in any m31 gc ( e.g. see rich et al . 1996 , fusi pecci et al .
1996 , jablonka et al .
1997 for cmds obtained using _ hst _ observations ) .
this strongly supports the idea that m31 gcs have comparable ages to those of galactic gcs .
we wish to emphasize , however , that although we see no evidence in our data for g302 having an age that is any different from the ages of the galactic gcs , we can not rule out the presence of super - luminous stars in the inner regions of g302 .
further evidence that g302 is an old gc with an age similar to that of the galactic gcs comes from de freitas pacheco ( 1997 ) . that study used integrated spectra and single - population stellar populations models to estimate the ages of twelve m31 gcs including g302 .
the result was a mean age of @xmath79 gyr for their sample of twelve m31 gcs and an age of @xmath80 gyr for g302 .
the weighted mean magnitude of the stars in the rr lyrae gap is @xmath81 ( standard deviation ) . using this , our derived iron abundance of @xmath1}}= -1.85 $ ] , and @xmath82}}+ 0.98 \label{equation : feh_mv_cd96}\ ] ] ( chaboyer et al .
1996 ) , we get a distance modulus of @xmath83 for g302 . in order to identify the outer limiting radius of g302 , and to search for any radial dependence in the morphology of the various branches of the cmd , we constructed cmds in four annuli centered on the cluster .
these cmds are shown in figure [ figure : g302_cmd_rings ] .
the fiducial sequences plotted in figure [ figure : g302_cmd_rings ] are those of ngc 6397 ( @xmath1}}= -1.91 $ ] ) and 47 tuc ( @xmath1}}= -0.71 $ ] ) taken from dca . for @xmath84
the ngc 6397 fiducial provides a good fit to the g302 rgb .
the gc appears to end at a radius somewhere between @xmath85 and @xmath86 , in agreement with what is found by fitting mk models ( see [ section : structure ] ) . beyond @xmath86
the rgb of the metal - rich halo population is clearly visible .
the apparent increase in the number of blue hb stars in the inner regions of g302 is an illusion .
a ks test shows that the radial distribution of blue hb stars is the same as that of the rgb stars at the 99.91% confidence level .
figure [ figure : g312_cmd ] shows the cmd for the gc g312 . as with g302 only stars between @xmath72 are shown since star counts and the surface brightness profiles suggest that the mk tidal radius of g312 occurs at @xmath65 ( holland et al .
1993 , holland 1997 ) .
interpolating between the fiducial giant branches suggests that g312 has an iron abundance of @xmath1}}= -0.56 \pm 0.03 $ ] , comparable to the mean iron abundance of stars in the halo of m31 .
this is an agreement with the spectroscopic iron abundance determined by hbk of @xmath1}}= -0.70 \pm 0.35 $ ] .
the quoted uncertainty in our iron abundance estimate is the formal uncertainty in the interpolation based on the observed width of the rgb .
it does not include uncertainties in the iron abundances of the fiducial rgb sequences or uncertainties in the distance to g312 . in light of this the quoted uncertainty
should be considered a lower limit on the true uncertainty in our estimate of @xmath1}}$ ] .
the small numbers of rgb stars falling under the metal - poor fiducial sequences are consistent with the expected degree of contamination ( see [ section : contamination ] ) from disk and metal - poor halo stars in m31 ( see 3.2 of hfr for details on the metallicity distribution of halo stars in m31 ) .
the hb of g312 is a red clump located slightly to the blue of the rgb .
no hb stars bluer than @xmath87 are present so it is not possible to determine the magnitude of the hb in the rr lyrae gap .
therefore we estimated @xmath88 , by taking the weighted mean magnitude of the red hb stars , to be @xmath89 . in order to get @xmath88
we added a correction to @xmath90 ( see e.g. lee et al .
1994 ) of @xmath91 ( see sarajedini et al . 1995 , ajhar et al .
this gives @xmath92 for g312 .
using the iron abundance we derived from the shape of the rgb the chaboyer et al .
( 1996 ) relation ( eq .
[ equation : feh_mv_cd96 ] ) gives a distance modulus of @xmath93 .
since so few halo stars are expected in the upper rgb of g312 , the observed width of the rgb provides a good estimate of the true photometric uncertainties in our observations .
photometric and spectroscopic evidence strongly suggests that all stars in a single galactic gc have the same metallicity and there is no evidence to suggest that this should be different for gcs in m31 .
the mean width of the rgb is @xmath94 , consistent with what is predicted by the artificial star simulations ( see [ section : art_stars ] ) .
therefore , we believe that the rgb width in figure [ figure : g312_cmd ] is typical of a single - metallicity population for our wf observations . this figure can be compared to figures 4 and 5 of hfr to show that the large range of metallicities seen in the halo of m31 is not merely an artifact of photometric uncertainties along the halo rgb .
figure [ figure : g312_cmd_rings ] shows no radial dependence in the morphology of the g312 cmd .
figure [ figure : ivi_cmds ] shows the cmds for g302 and g312 after converting the calibrated magnitudes to absolute _
i_-band magnitudes .
this was done to determine the color of the rgb for each gc at @xmath95 , hereafter referred to as @xmath96 .
dca found a strong relationship between iron abundance and @xmath96 for galactic gcs with @xmath97}}\le -0.7 $ ] .
figure [ figure : vi - feh]a shows the relationship between @xmath96 and @xmath1}}_{\rm s}$ ] for g302 and g312 and several other m31 gcs with published cmds .
we have adjusted the published data to the distance modulus used in this paper ( @xmath9 ) .
the @xmath1}}_{\rm s}$ ] values are the spectroscopically determined iron abundances of hbk , so the @xmath96 color and iron abundance are determined completely independently of each other for each gc . color , iron abundance , and reddening data for each gc is listed in table [ table : vi - feh ] along with the source of the cmd ; the projected distance of each gc from the center of m31 ; and @xmath98 , the projected distance of each gc from the major axis of m31 with positive @xmath98 being on the northwest side of the major axis of m31 . since several of the m31 gcs in figure [ figure : vi - feh]a have iron abundances at least as high as that of 47 tuc we extended the dca relationship to higher iron abundances .
extending this relationship to gcs more metal - rich than 47 tuc ( @xmath1}}= -0.71 $ ] ) is somewhat difficult since the rgb in the @xmath99 plane becomes asymptotically flat as the iron abundance increases .
this makes it difficult to define @xmath100 since , as @xmath1}}$ ] approaches the solar value , this point occurs in the horizontal portion of the rgb . in order to extend the dca relationship to higher iron abundances we determined @xmath96 for the metal - rich galactic gc ngc 6553 using the @xmath101 cmd of ortolani et al .
( 1990 ) . we assumed a distance modulus of @xmath102 ( guarnieri et al .
1995 ) , a reddening of @xmath103 ( bico & alloin 1986 ) , and an iron abundance of @xmath1}}= -0.29 $ ] ( zinn & west 1984 ) . using this data we found @xmath104 for ngc 6553 .
this point lies just below the region where the rgb becomes horizontal so there is a significant uncertainty in the value of @xmath96 for ngc 6553 solely due to the width of the rgb in @xmath52 at @xmath95 .
we extrapolated the dca relation to @xmath1}}= -0.29 $ ] simply by connecting the @xmath1}}= -0.71 $ ] end of the dca relation to the location of ngc 6552 .
the resulting relation is @xmath105}}= -15.16 + 17.0(v - i)_{0,-3 } - 4.9(v - i)_{0,-3}^2\ ] ] for @xmath97}}\le -0.7 $ ] ( from dca ) and @xmath105}}= -1.36 + 0.44(v - i)_{0,-3}\ ] ] for @xmath106}}\le -0.29 $ ] ( our extension ) .
we wish to stress that our extension to high iron abundances is an estimate and is intended only to provide a reasonable estimate of what happens to the dca relation at high iron abundances .
if the spectroscopic iron abundances are adopted then nine of the twelve m31 gcs are redder than predicted by the extended dca relation .
this discrepancy is present at all iron abundances but is most noticeable for the metal - rich gcs .
the individual uncertainties in @xmath1}}_{\rm s}$ ] for all the gcs except g1 and g327 are large enough to account for the difference between their locations in figure [ figure : vi - feh]a and the predictions of the extended dca relation .
a recent paper by de freitas pacheco ( 1997 ) gives iron abundances for twelve gcs in the m31 system .
most of their iron abundances are higher than the hbk values by @xmath107 dex .
an examination of the cmds for each gc shows that several gcs have rgbs that are significantly redder and flatter than would be expected from their spectroscopic iron abundances .
this is particularly noticeable for the more metal - rich gcs . to test whether errors in the spectroscopically determined iron abundances are sufficient to account for the discrepancies between the gcs and the extended dca relation in figure [ figure : vi - feh]a , we replotted the gcs using iron abundances determined from the shapes of the rgbs
this data is shown in figure [ figure : vi - feh]b .
all gcs now fall within their uncertainties of the extended dca relationship with the exception of g108 , which has a @xmath96 value nearly @xmath108-@xmath45 redder than its predicted value .
alternately , the iron abundance ajhar et al . (
1996 ) derived from g108 s cmd could be in error .
however , their figure 20 shows the upper portion rgb of g108 to be not quite as flat than that of 47 tuc , indicating that g108 has slightly less iron than 47 tuc does .
it would be difficult , using their data , to fit a fiducial to the g108 rgb that is flatter ( and thus more iron rich ) than the 47 tuc fiducial .
therefore , we believe that the estimated iron abundance of g108 is correct .
one possible explanation for the red color of g108 is that the in m31 extends @xmath109 kpc beyond the outer edge of the optical disk , which is past the location of g108 .
cuillandre et al .
( 1997 ) found dust beyond the optical limit of the m31 disk which correlates with the distribution of beyond the disk .
since g108 is located within the extended disk of around m31 it is possible that g108 s @xmath96 color may be due to internal reddening in the extended disk of m31 . using the maps of emerson ( 1974 ) , and the relationships between column density and reddening given in cuillandre et al .
( 1997 ) , we find an excess reddening of @xmath110 .
however , an examination of the cmd for g108 ( figure 20 of ajhar et al .
1996 ) suggests the internal reddening from m31 does not exceed @xmath111
. this discrepancy may be due to the patchy nature of the distribution in the outer regions of m31 s extended disk .
brinks & bajaja ( 1986 ) found many small regions in the distribution in m31 where the column densities were significantly less than the surroundings .
these holes were typically a few arcminutes in diameter , smaller than the beam - size used to construct the emerson ( 1974 ) maps .
it is possible that g108 ( which has an angular diameter of @xmath112 ) is located in a hole in the distribution that is smaller than the resolution of the maps .
if this is the case , the emerson ( 1974 ) maps will lead to our overestimating the reddening toward g108 and our estimate of the excess reddening towards g108 would not be inconsistent with the published cmd for that gc .
the raw , @xmath113 , and completeness - corrected , @xmath114 , @xmath51-band stellar lfs for g302 and g312 are presented in table [ table : lfs ] .
the raw lfs include all stars within @xmath115 of the center of each gc .
the background lf for each gc was obtained by taking the lf for the m31 halo near each gc ( from hfr ) and scaling them to the area covered by each gc .
the scaled background lf was subtracted from the raw lf before the effects of incompleteness were considered . subtracting the raw background lf assumes that the completeness corrections in the wf2 and wf4 fields are the same as in the wf3 fields .
an examination of figure [ figure : lfs ] shows that the completeness corrections are small for stars with @xmath116 .
therefore the raw background luminosity function instead of the completeness - corrected background luminosity function will not alter the results .
the raw and completeness - corrected lfs are related by @xmath117 where @xmath118 is the finding - probability matrix .
@xmath118 describes the probability of a star with a true magnitude of @xmath119 being recovered with a magnitude of @xmath120 .
our artificial star tests suggest that completeness corrections need to be applied to the raw gc and background lfs .
these corrections are small for @xmath116 , but the artificial star tests suggest that bin - jumping becomes significant at approximately the level of the hb and needs to be compensated for in order to accurately reconstruct the lf .
given @xmath118 there is no ideal method for reconstructing a stellar lf from incomplete star counts . for a review of the problem see fahlman ( 1993 ) .
ideally the true lf , @xmath114 , can be obtained from the observed lf , @xmath113 , using @xmath121 as described in drukier et al .
( 1988 ) and mighell ( 1990 ) . unfortunately
, many faint artificial stars are lost in the noise and thus not recovered at all .
this results in some artificial stars bin - jumping off the faint end of the finding - probability matrix so there is insufficient information to reliably invert @xmath118 .
an alternative to computing @xmath122 is to use a model lf , @xmath123 , and solve the forward equation @xmath124 .
the estimated lf can then be adjusted until @xmath125 is minimized .
this approach will not accurately reconstruct the faint end of the lf since @xmath118 does not contain sufficient information at faint magnitudes , due to bin - jumping by faint stars , to do this .
however , unlike the inversion method , it will work if @xmath118 is nearly singular and thus allows us to estimate the completeness - corrected lf in the face of incomplete information .
the drawbacks to the forward approach are that there is no obvious way to compute the uncertainties in the resulting lf and that some _ a priori _ information is needed about the true lf
. this forward method will give different results depending on the initial guess at @xmath114 and the scheme used to adjust @xmath126 during the iteration .
we found that , for our data , the lf is not dependent on the initial guess at the form of the lf when @xmath116 ( approximately the level of the hb ) .
we adjusted @xmath123 using @xmath127 and iterated @xmath124 until @xmath128 was minimized .
figure [ figure : lfs ] shows the observed and completeness - corrected , background - subtracted lfs for g302 and g312 .
these lfs are reliable for @xmath116 . in figure
[ figure : lfs_comp]a we compare the cumulative lf for g302 with that of the metal - poor ( @xmath1}}= -1.65 $ ] ) galactic gc m13 ( simoda & kimura 1968 ) . in figure
[ figure : lfs_comp]b we compare the cumulative gc for g312 to a theoretical lf for an old ( @xmath129 gyr ) gc with @xmath1}}= -0.47 $ ] and @xmath130}}= + 0.23 $ ] from bergbusch & vandenberg ( 1992 ) .
stellar evolution models do not show any relationship between the slope of the lf of the upper rgb and either the iron abundance or the age of the stars ( bergbusch & vandenberg 1992 ) so figure [ figure : lfs_comp ] can not be used to determine the ages or the iron abundances of g302 or g312 .
there is a general agreement between the shape of the g312 lf and the shape of the theoretical metal - rich lf , as well as between the shape of the g302 lf and the metal - poor comparison lf , which suggests that our star counts are reasonably complete down to the level of the hb ( @xmath23 ) .
we determined the tidal radii by fitting mk models to the background - subtracted surface brightness profiles of each gc .
the unresolved background light was measured by masking out a circle of radius 300 pixels ( @xmath131 ) centered on the gc and fitting a plane to the remaining area on the wf3 ccd .
we found that subtracting a constant sky level resulted in residual variations of @xmath132% across the wf3 ccd for both the g302 and the g312 fields .
we found no evidence for intensity or color gradients in the unresolved background light .
surface brightness profiles for each gc were obtained using the iraf implementation of the stsdas task ellipse .
this task uses jedrzejewski s ( 1987 ) isophote - fitting algorithm to fit elliptical isophotes to the observed isophotes on the background - subtracted images .
this algorithm determines the centers , ellipticities , orientations , and intensities for a series of ellipses logarithmically spaced between @xmath15 01 ( @xmath15 1 pixel ) and @xmath15 200 ( @xmath15 200 pixels ) .
the spacing between successive ellipses varies from significantly less than the size of the wf3 seeing disc ( fwhm @xmath133 ) near the center of each gc to @xmath1510 times the size of the seeing disc beyond the tidal radius of the gc .
the properties of the ellipses are not independent of each other for radii of less than approximately eight times the size of the seeing disc ( e.g. schweizer 1979 ) .
since the spacing between ellipses is logarithmic the degree of correlation will vary with the distance from the center of the gc .
there is an additional correlation caused by the overlap in the subsets of pixels being used to compute successive best - fitting ellipses . in light of these effects
the computed uncertainties in the surface brightnesses , ellipticities , and orientations of the individual fitted ellipses should be treated as guides to the reliability of the individual values relative to the values for the other fitted ellipses for that gc and _ not _ taken as statistically rigorous estimates . in the inner regions of each gc ( @xmath1340 ) the large gradient in surface brightness , and the small number of pixels , make the fitted ellipses less reliable than those fit further out . finally , we extracted the one - dimensional surface brightness profile for each gc along the effective radius axis of each gc .
, of an ellipse with semi - major and semi - minor axes @xmath135 and @xmath136 respectively is defined by @xmath137 .
the effective radius axis is either of the two axes along which the perimeter of the ellipse intersects the perimeter of a circle with radius @xmath138 . ] .
we used cern s minuit function minimization package to fit single - mass one - dimensional seeing - convolved mk models to the surface brightness profiles for each gc .
the fitting was done by simultaneously adjusting all the free parameters until the @xmath139 statistic between the model and the weighted surface brightness profile was minimized . after some experimentation we adopted a weighting scheme of @xmath140 where @xmath141 is the rms scatter in the isophotal intensity about the @xmath142 fitted isophotal ellipse .
we tested this weighting scheme by fitting mk models to a series of surface brightness profiles created by drawing points from a mk distribution and found good agreement between the input and the fitted models .
the exact form of the weighting did not significantly affect the values of the best - fitting model parameters provided that the weights assigned to each data point were reasonably indicative of the reliability of each point .
each mk model was convolved with the psf for the wf3 ccd at the center of the gc .
we determined the central potential , @xmath143 , king core radius , @xmath144 , tidal radius , @xmath11 , concentration , @xmath145 , anisotropy radius , @xmath146 , and half - mass radius , @xmath16 , for g302 and g312 in the _ v_- and _ i_-bands .
table [ table:1d_models ] gives the parameters of the best - fitting mk models for each gc .
v_- and _ i_-band surface brightness profiles , along with the best - fitting mk models , are shown in figures [ figure : g302_sdp ] and [ figure : g312_sdp ] .
the profiles for both gcs are well fit by isotropic mk models out to @xmath147 . beyond this distance from the center of the gc
there is excess light over what is predicted by isotropic ( @xmath148 ) mk models . in order to test if the excess light is due to velocity anisotropy in the gcs we fit anisotropic mk models to each gc .
the results are listed in table [ table:1d_models ] . in all cases
the anisotropic models yield slightly larger @xmath149 values than the corresponding isotropic models , suggesting that the isotropic models are formally better fits to the data than are the anisotropic models .
furthermore , the best - fitting anisotropy radii are greater than the fitted tidal radii for each gc , which is physically meaningless .
therefore , we conclude that the overabundance of light at large radii is not due to velocity anisotropy .
this supports the hypothesis that these gcs are surrounded by extended halos of tidally stripped stars .
minuit returns formal uncertainties of @xmath15 5 to 10% on the mk parameters .
however , minuit was generally unable to compute a fully accurate covariance matrix for the fitted parameters so minuit s uncertainty estimates are not reliable .
we estimated the uncertainties in our fits by generating a series of artificial gcs and fitting seeing - convolved mk models to them in exactly the same way that we did to our data .
we found that minuit tended to underestimate the uncertainties in our fits by a factor of between 1.5 and 2 .
this suggests the true 1-@xmath45 uncertainties in the parameters are between @xmath1510 and 15% of the best - fit values of the parameters .
the mean color for @xmath150 for g312 is @xmath151 while for g302 @xmath152 .
these colors are consistent with g302 having a metal - poor stellar population and g312 having a metal - rich stellar population . figure [ figure : colours ] shows the dereddened @xmath52 color profiles of g302 and g312 .
these profiles were obtained by subtracting the isophotal surface brightnesses in @xmath51 and @xmath153 then dereddening the resulting profiles .
the error bars represent the 1-@xmath45 standard error in the color .
the inner portion of the g302 profile is truncated at @xmath154 ( @xmath8 1 pixel ) since the innermost pixels ( @xmath155 pixel ) of our images of g302 were slightly saturated .
individual stars can be resolved to within @xmath156 of the center of the gc and the isophotes become less certain when the light becomes resolved into individual stars .
artificial star tests suggest that the large scatter in the color profiles in the outer regions of g302 and g312 represent stochastic star placement within the gcs , not intrinsic color gradients . in the inner @xmath157
the color gradients are @xmath158 for g302 and @xmath159 for g312 . to determine if these color gradients are due to chance we performed 10,000 bootstrap resamplings of the color data for each gc and computed the color gradients for each resampled data set .
this analysis gave a 2.1255% probability that the observed color gradient ( for @xmath150 ) in g302 is due to chance and a 0.0005% probability that the observed color gradient ( for @xmath150 ) in g312 is due to chance .
figures [ figure : g302_shape ] and [ figure : g312_shape ] show the ellipticity and orientation profiles for g302 and g312 respectively .
the error bars in figures [ figure : g302_shape ] and [ figure : g312_shape ] should be treated as guides to the reliability of the individual measurements and _ not _ taken as statistically rigorous estimates of the uncertainties in each point . in order to estimate the overall projected ellipticity of each gc we determined the weighted mean ellipticities , @xmath160 and position angles , @xmath161 ( in degrees east of north on the sky ) using the values for the individual isophotes between @xmath162 and @xmath163 .
the results are presented in table [ table : shapes ] .
the uncertainties are standard errors in the mean and @xmath3 is the number of isophotes used to compute @xmath160 and @xmath161 .
the projected major axis of g302 points approximately towards the center of m31 .
the ellipticity of each gc is the same in the _ i_-band as it is in the _ v_-band but the projected major axis of g302 in the _
v_-band is located @xmath164 west of the projected major axis in the _ i_-band .
this shift in orientation between the two colors is visible at all radii in g302 and is likely due to a small number of giant stars dominating the light from the gc .
our data show that the projected major axis of g302 in the _ v_-band is oriented @xmath165 farther west from the _ v_-band orientation obtained by lupton ( 1989 ) ( @xmath166 ) , but that lupton s orientation is only @xmath167 east of our _ i_-band orientation .
this supports the idea that the difference between the _ i_- and _ v_-band orientations that we observed is due to the different numbers of resolved stars in the different bands and data sets .
our ellipticity ( @xmath168 ) is somewhat higher than that found by lupton ( 1989 ) ( @xmath169 .
figure [ figure : g302_shape ] shows an apparent decrease in the ellipticity of g302 between @xmath170 and @xmath171 .
the slope of @xmath172 is @xmath173 . in order to determine if this slope represents a real decrease in ellipticity with radius we performed 10,000 bootstrap resamplings of the ellipticity data between @xmath174 and computed the slope for each set of bootstrapped ellipticities .
we found that there was a 22.34% chance of observing a slope at least as great as the slope seen in the actual data simply by chance .
therefore , we conclude that there is no evidence for a change in ellipticity with radius between @xmath175 and @xmath163 in g302 .
it is not possible to say if the apparent increase in ellipticity between @xmath176 and @xmath177 is real or due to the presence of a small number of unresolved bright stars located within a few core radii of the center of g302 .
since the stellar fwhm is @xmath178 , the chance location of a single unresolved bright star within @xmath179 of the center of g302 along the projected minor axis of the gc could reduce the ellipticity of the fitted isophotes .
g312 has a projected ellipticity of @xmath180 but there is considerably more scatter in the fitted position angles ( the standard deviation is @xmath181 ) than in the g302 position angles ( @xmath182 ) .
this suggests that the observed projected ellipticity of g312 is primarily due to the stochastic placement of stars within g312 , variations in the stellar background of the m31 halo , and the intrinsic precision of the ellipse algorithm . running the ellipse task on a series of circular artificial mk gc models
which were convolved with the wf3 psf suggests that seeing alone can produce ellipticities of between 0.01 and 0.02 at radii of between approximately @xmath183 and @xmath184 .
this agrees with the fluctuations in ellipticity seen in figure [ figure : g312_shape ] at these radii .
while seeing has very little effect on the observed ellipticity beyond @xmath185 , artificial star tests indicate that stochastic star placement can introduce an uncertainty in the ellipticity of between @xmath186 and @xmath187 . in light of this
we believe there is no evidence for g312 being elliptical in the plane of the sky .
it is , however , possible that g312 is elongated along the line of sight .
gaf have observed an excess of resolved and unresolved stars beyond the formal mk tidal radii of several gcs in m31 , as would be expected if the stars which evaporated from these gcs remained in extended halos around those gcs .
another test for extended halos of unbound stars is to identify an asymmetrical overdensity of stars in two dimensions beyond the formal tidal radius .
this has been done for some galactic globular clusters ( see grillmair et al .
1995 ) using multiple schmidt photographic plates for each gc .
however , it has not been attempted for the m31 gcs , where the gc , and the background , can be imaged on a single ccd .
this eliminates possible systematic effects arising from comparing star counts across multiple fields . to search for extended halos around g302 and g312 we looked at the two - dimensional distribution of stars beyond the tidal radii of each gc .
the total background stellar densities around each gc were small ( @xmath188 for g302 and @xmath189 for g312 ) .
we rebinned the stellar positions into `` super - pixels '' so there were enough stars in each bin to ensure that the poisson fluctuations in a bin would be small compared to the total number of stars in that bin .
in order to compute the optimum binning size we assumed that all the stars on the wf3 ccd were part of the gc and that the gc could be approximated by a gaussian distribution of stars . it can be shown ( heald 1984 ) that for any gaussian distribution the bin size , @xmath190 , which maximizes the signal - to - noise in each bin , is given by @xmath191 where @xmath45 is the standard deviation of the gaussian and @xmath3 is the size of the sample .
this technique provides a balance between the need to keep the bin sizes small compared to the width of the gaussian , and the need to have a large number of data - points in each bin to reduce poisson noise in the bin .
since we are searching for overdensities of stars beyond the mk tidal radii of the gcs the assumption that a gc can be approximated by a gaussian will result in an underestimate of the bin size needed to maximize the signal - to - noise .
we found that bin sizes of 32 pixels ( @xmath37 32 ) , approximately two to three times the computed optimal size for gaussians , provided the best signal - to - noise for our data .
we smoothed the resulting binned data by convolving it with a unit gaussian with a dispersion equal to the bin size ( @xmath192 ) .
the resulting stellar number density distribution for g302 is shown in figure [ figure : g302_tail ] and the stellar number density distribution for g312 is shown in figure [ figure : g312_tail ] .
figure [ figure : g302_tail ] shows an asymmetric overdensity of stars around g302 extending to at least twice the mk tidal radius .
the coherence of the isodensity contours which lie beyond the tidal radius , but inside the background contour , suggests that the observed overdensities are not an artifact of the binning , smoothing , or contouring processes .
the surface brightness profile of g302 ( see figure [ figure : g302_sdp ] ) light beyond the tidal radius ( @xmath193 ) of the gc is in agreement with the asymmetric halo of stars seen in figure [ figure : g302_tail ] .
the surface brightness profile of g312 , however , is consistent with there being no light beyond the tidal radius ( @xmath194 ) .
this is reflected in the lack of an extended halo of stars in figure [ figure : g312_tail ] .
the fact that an overdensity of stars is seen beyond the tidal radius of g302 , but not beyond the tidal radius of g312 , suggests that the overdensity is real and not an artifact of the analysis . in order to determine if the observed halo surrounding g302 is real we applied our contouring procedure to the star counts from the wf2 and wf4 images in the g302 and g312 fields .
figure [ figure : bkgd_tail ] shows that no structures of comparable size to the halo around g302 are seen in any of the four background fields . to test whether a random distribution of stars could give rise to coherent structures which could be mistaken for an extended halo we constructed a series of random star fields with number densities comparable to those found in the g302 and g312 background fields .
none of these artificial star fields showed evidence for coherent sub - structure .
we also used mk surface density profiles to place artificial stars onto randomly generated star fields to see if the presence of a gc would bias our binning / smoothing / contouring process in favor of finding structure beyond the tidal radius when none was really present .
none of these monte - carlo images showed any evidence for extended halos . in order to quantify the orientation of the halo around g302 we computed the second moment of the distribution of stars beyond the fitted tidal radius . only stars with @xmath195 from the center of the gc
were used to ensure that neither the gc or the edges of the ccd biased our sample .
we computed the statistic @xmath196 where @xmath3 is the number of stars beyond the tidal radius and @xmath197 is the distance from the @xmath142 star to an arbitrary axis of symmetry for the gc .
this statistic corresponds to the root - mean - square distance between the stars and the axis of symmetry so the value of @xmath198 will be at its minimum when the assumed axis of symmetry coincides with the true major axis of the distribution .
the @xmath198 values for g302 and g312 , assuming a range of angles of symmetry , are shown in figure [ figure : zeta ] .
a second statistic , @xmath199 , gives a measure of the degree of symmetry in the distribution of stars about the center of the gc .
the value of @xmath200 for a distribution depends on the shape of the distribution and the number of stars present . for a given distribution , the greater
the value of @xmath200 measured the greater the deviation from a circular distribution of stars . for a perfectly circular distribution
, @xmath198 would be the same for all angles resulting in @xmath201 . in order to determine the probability of getting the observed value of @xmath200 by chance from a uniform circular halo of stars around a gc we generated 10,000 sets of stellar coordinates and measured @xmath200 for each set .
the stellar coordinates were randomly drawn from a uniform distribution of stars with the same size and number density as the field around the gc .
this data formed the cumulative probability distributions shown in figure [ figure : eta_prob ] . since the stellar number density in the g302 field is different from the stellar number density in the g312 field we computed separate probability distributions for the g302 field and the g312 field
for g302 this analysis yields @xmath202 , which occurs at a position angle for the major axis of the extended halo of @xmath203 east of north .
the cumulative probability distribution for @xmath200 in the g302 fields ( see figure [ figure : eta_prob]a ) shows that there is only a 1.84% chance of obtaining a value of @xmath204 from a uniform distribution of stars .
in contrast , the wf2 and wf4 fields near g302 yielded @xmath200 values of 3.5530 and 4.6618 corresponding to 38.47% and 20.62% chances respectively of occurring by chance .
the orientation of the extended halo around g302 is @xmath205 to the west of the projected major axis of g302 as observed in the _ v_-band and @xmath206 west of the projected major axis for the gc as observed in the _ i_-band . in light of our inability to find coherent overdensities in any of the background images or simulated images , the results of our moment analysis , and the observed overdensity of stars beyond the tidal radius of the best - fitting mk model , we believe that the extended asymmetric halo around g302 seen in figure [ figure : g302_tail ] is a real feature of g302 .
since the probability of observing a given value of @xmath200 by chance depends on the number density of stars on the image we performed a separate series of 10,000 monte - carlo simulations for the g312 data ( see figure [ figure : eta_prob]b ) .
the resulting probability distribution indicated that the observed value of @xmath207 for the stars beyond the tidal radius g312 in the wf3 field had a 9.96% chance of occurring in a randomly distributed set of stars .
the wf2 and wf4 fields around g312 have @xmath200 values of 9.6342 and 9.1801 respectively , corresponding to 27.65% and 30.73% probabilities of occurring at random .
this , and the lack of a significant excess of stars beyond the tidal radius of the best - fitting mk models , leads us to believe that g312 does not exhibit any evidence for having an extended halo of stars .
we are not , however , able to rule out the possibility that such a halo does exist and is aligned along the line of sight .
figures [ figure : g302_sdp ] and [ figure : g302_tail ] show that there is an excess of light and stars beyond the tidal radius of the best - fitting mk model for g302 . by computing the amount of mass beyond the tidal radius we can estimate the mass - loss rate for g302 .
we converted our _ v_-band surface brightness profile for g302 to a projected mass profile using a mean mass - to - light ratio of @xmath208 solar units ( from dubath & grillmair 1997 ) .
the total mass inside the mk tidal radius ( @xmath209 ) was @xmath210 .
we estimated the total mass of g302 , by applying the mass - to - light ratio to the total integrated _ v_-band magnitude of g302 , to be @xmath211 .
this gives a total mass of @xmath212 beyond the formal tidal radius of g302 .
mass - segregation within a gc can result in @xmath213 increasing with radius so the total mass beyond the tidal radius that we derive here will be a lower limit on the true mass in the extended halo around g302 .
if we assume an age of @xmath129 gyr and a constant rate of mass loss for g302 then the mass - loss rate required for the observed amount of mass to escape beyond the mk tidal radius is @xmath214/gyr .
the relaxation time at the half - mass radius for a gc is given by ( spitzer 1987 ) @xmath215 where @xmath216 is the total mass of the gc in solar masses , @xmath16 is the half - mass radius of the gc in pc , @xmath217 is the mean stellar mass in the gc in solar masses , and @xmath218 is the estimated number of stars in the gc .
following djorgovski ( 1993 ) we adopted @xmath219 .
the half - mass radius obtained by fitting isotropic mk models to g302 is @xmath220 pc which gives a half - mass relaxation time of @xmath221 gyr .
this yields a projected escape rate of @xmath222 per half - mass relaxation time .
converting observed surface brightnesses to a deprojected mass profile involves solving an abel integral which contains the radial derivative of the surface brightness distribution .
since the surface brightness data for g302 contains noise this inversion is inherently unstable .
therefore , we chose to project the theoretical escape rates of oh & lin ( 1992 ) into the plane of the sky before comparing them to our observed escape rate .
the oh & lin ( 1992 ) evaporation rates for isotropic gcs with ages of @xmath223 in the galactic potential then become @xmath224 to @xmath225 per relaxation time .
this is comparable to the escape rate inferred from our data .
this result further strengthens our claim that we have detected an extended halo of unbound stars around g302 . using the methods described above we find that g312 has a mass of @xmath226 beyond its fitted isotropic mk tidal radius ( @xmath227 ) .
the large uncertainty in the amount of mass beyond the tidal radius is due to the large uncertainties in the surface brightness profile at large radii .
figure [ figure : g312_tail ] does not show any evidence for an extended halo of stars around g312 so it is likely that the light seen beyond the fitted tidal radius in figure [ figure : g312_sdp ] is due to low - level variations in the unresolved background light .
alternately , it is possible that the use of multi - mass mk models would result in a slightly larger tidal radius .
if we assume that the observed excess of light in figure [ figure : g312_sdp ] is due to an extended halo of stars , then the mass - loss rate required is @xmath228/gyr , approximately 10% of the mass - loss rate from g302 .
g312 has a half - mass relaxation time of @xmath229 gyr , so the escape rate per relaxation time is @xmath230 .
this is consistent with oh & lin s ( 1992 ) theoretical values .
however , the large uncertainties in our photometry near the tidal radius of g312 ( and thus the large uncertainty in our observed escape rate ) suggest that the observed escape rate is consistent with there being no observable mass - loss from g312 .
hbk measured a heliocentric radial velocity of @xmath231 km@xmath232s@xmath233 for g302 .
correcting for rubin & ford s ( 1970 ) 21 cm velocity for m31 ( @xmath234 km@xmath232s@xmath233 ) gives g302 a radial velocity of @xmath235 km@xmath232s@xmath233 relative to the center of m31 .
this is one of the largest radial velocities observed for a gc in m31 , suggesting that almost all of g302 s space velocity is along the line - of - sight .
this can only occur when g302 is near its perigalactic passage .
since very little of g302 s space velocity is in the plane of the sky , and g302 is near its perigalacticon , g302 must be at approximately the same distance from us as is the center of m31 .
this means that the distance measured on the sky between g302 and the center of m31 is the true separation ; so the perigalactic distance for g302 is @xmath236 corresponding to @xmath237 kpc assuming a distance modulus of @xmath9 for m31 . in order to constrain the orbit of g302 we used a three - component model for the m31 potential .
the disk was modeled by a miyamoto & nagai ( 1975 ) potential , the bulge by a hernquist ( 1990 ) potential , and the halo by a spherical logarithmic potential : @xmath238 @xmath239 @xmath240 we adopted @xmath241 , @xmath242 , @xmath243 kpc , @xmath244 kpc , @xmath245 kpc , @xmath246 kpc , and @xmath247 km@xmath232s@xmath233 .
these parameters were chosen by scaling the galactic values to the observed mass and diameter of m31 . using this model , and the observed radial velocity and perigalactic distance of g302
, we found that the orbit of g302 has an eccentricity of @xmath248 where @xmath249 is the apogalactic distance ( in kpc ) for g302 .
this value assumes that all of the space velocity of g302 is along the line - of - sight , and gives g302 an apogalactic distance of @xmath250 or @xmath251 kpc .
if we allow 40% of g302 s space velocity to be in the plane of the sky then the orbital parameters become @xmath252 and @xmath253 .
since g302 s radial velocity is faster than 95% of the radial velocities of other m31 gcs it is unlikely that more than @xmath254% of g302 s space velocity is tangential to the line - of - sight .
if this is the case , the orbital eccentricity for g302 is @xmath255 and the apogalactic distance is @xmath251 kpc .
the period of g302 s orbit was defined to be twice the time required for the gc to move from its perigalacticon to its apogalacticon .
if all of the gc s space velocity is along the line - of - sight then the orbital period is @xmath256 gyr . if 40% of the space velocity is tangential to the line - of - sight then @xmath257 gyr
since this work was completed new parallax measurements from the _ hipparcos _ satellite have been published which have led to a revised estimate of the distance to m31 .
the new distance modulus is @xmath258 ( feast & catchpole 1997 ) , which corresponds to a distance of @xmath259 kpc , a 25% increase over the distance that we adopted for m31 .
if this distance modulus is used then our results change in the following ways : 1 .
adjusting the fiducial rgb sequences in [ section : cmds_g302 ] and [ section : cmds_g312 ] yields iron abundances of @xmath1}}= -2.2 $ ] for g302 and @xmath1}}= -0.7 $ ] for g312 .
the new g312 iron abundance is consistent with the hbk spectroscopic iron abundance but the new g302 value is @xmath260-@xmath45 lower than the hbk or the de freitas pacheco ( 1997 ) value .
2 . chaboyer et al . (
1997 ) have derived a new relation , @xmath261}}+ 1.9 ) + ( 0.41 \pm 0.08)$ ] , between @xmath1}}$ ] and @xmath262 which uses the _ hipparcos _ parallaxes .
this relation , and the iron abundances and hb magnitudes derived in [ section : cmds_g302 ] and [ section : cmds_g312 ] , gives distance moduli of @xmath263 for g302 and @xmath264 for g312 .
these distance moduli would place g302 @xmath265 kpc in front of m31 if m31 is 900 kpc away .
this is inconsistent with the hbk radial velocity of g302 and the orbital parameters determined in [ section : orbit_g302 ] .
the distance modulus for g312 , and the feast & catchpole ( 1997 ) distance to m31 , puts this gc @xmath266 kpc in front of m31 .
3 . adopting a distance modulus of @xmath258 would result in the metal - rich gcs in figure [ figure : vi - feh ] being in better agreement with the dca relation and our extension to it .
however , the metal - poor gcs would have bluer @xmath96 colors which would result in them falling to the left of the dca relation .
this would mean that the spectroscopic iron abundances of the metal - poor m31 gcs have been systematically overestimated .
since iron abundances are more difficult to determine in spectra of metal - rich gcs it is more likely that the @xmath1}}$ ] values for the metal - rich gcs have been underestimated than the @xmath1}}$ ] values for the metal - poor gcs have been underestimated . in light of these results
we find no reason to prefer the @xmath258 distance modulus over the @xmath267 value .
we have used _ hst _ wfpc2 photometry to construct deep ( @xmath0 ) cmds for two gcs in the halo of m31 . both gcs appear to have a single old population of stars similar to what is found in galactic gcs .
the shape of the rgb for g302 gives an iron abundance of @xmath1}}= -1.85 \pm 0.12 $ ] , in agreement with the published values obtained using spectroscopy .
for g312 we obtain @xmath1}}= -0.56 \pm 0.03 $ ] , which is somewhat more metal - rich than the spectroscopic value . neither gc shows any indication that there is a second parameter acting upon their hb morphologies .
both gcs have mk tidal radii of @xmath268 , core radii of @xmath269 , central concentrations of @xmath270 , and half - mass radii of @xmath271 .
there is no evidence for velocity anisotropy in either g302 or g312 .
g302 has a color of @xmath272 while the color of g312 is @xmath273 .
g302 has a projected ellipticity of @xmath274 with the major axis oriented approximately towards the center of m31 .
this gc has an excess of light beyond its formal tidal radius which is not consistent with either an isotropic or an anisotropic mk model .
the two - dimensional distribution of stars around g302 is consistent with the presence of an extended halo extending to two to three times the formal tidal radius from the cluster .
g312 , on the other hand , has an ellipticity of @xmath275 .
neither the integrated light , nor the star counts , show any evidence for an extended halo .
it is possible that such a tail does exist for g312 , but is oriented along the line of sight .
we have estimated the projected mass - loss rate from g302 to be @xmath276 per gyr which corresponds to a projected escape rate of @xmath277 per half - mass relaxation time .
the projected escape rate from g312 is @xmath230 per half - mass relaxation time .
these are consistent with the escape rates predicted by oh & lin ( 1992 ) although the large photometric uncertainties near the tidal radius of g312 makes the escape rate for this gc much less reliable than that for g302 .
this research is based on observations made with the nasa / esa _ hubble space telescope _ obtained at the space telescope science institute .
stsci is operated by the association of universities for research in astronomy inc . under nasa contract
nas 5 - 26555 .
support for this research was provided by operating grants to ggf and hbr from the natural science and engineering research council of canada .
sh would like to thank peter stetson for kindly making a copy of the allframe software , as well as _
v_- and _ i_-band psfs for the wfc , available .
we would also like to thank the anonymous referee who made several useful comments .
rlcrcc g302 & nov . 5 & f555w & 8 @xmath278 500 & & & 2 @xmath278 160 & & f814w & 7 @xmath278 500 & & & 1 @xmath278 400 & & & 1 @xmath278 160 g312 & oct . 31 & f555w & 8 @xmath278 500 & & & 2 @xmath278 160 & & f814w & 7 @xmath278 500 & & & 1 @xmath278 400 & & & 1 @xmath278 160 rrrccr g302 & pc & f555w & @xmath279 & 0.992 & 12 & & f814w & @xmath280 & 1.009 & 29 & wf2 & f555w & @xmath281 & 1.065 & 25 & & f814w & @xmath282 & 1.017 & 80 & wf3 & f555w & @xmath283 & 1.003 & 87 & & f814w & @xmath284 & 1.018 & 217 & wf4 & f555w & @xmath285 & 1.011 & 90 & & f814w & @xmath286 & 1.005 & 252 g312 & pc & f555w & & & & & f814w & & & & wf2 & f555w & @xmath287 & 1.008 & 186 & & f814w & @xmath288 & 1.034 & 119 & wf3 & f555w & @xmath289 & 1.018 & 101 & & f814w & @xmath290 & 1.015 & 125 & wf4 & f555w & @xmath291 & 1.001 & 85 & & f814w & @xmath292 & 1.027 & 89 lrrrrrrrrr wf3 & 892 & 64.328 & 212.348 & 17.032 & 0.082 & 1.000 & 16.294 & 0.071 & 1.000 wf2 & 3440 & 466.457 & 631.376 & 17.628 & 0.084 & 0.744 & 16.365 & 0.077 & 0.727 wf3 & 2583 & 437.139 & 413.836 & 18.195 & 0.065 & 1.341 & 17.356 & 0.059 & 1.551 wf3 & 2606 & 439.952 & 415.701 & 18.454 & 0.063 & 1.462 & 17.642 & 0.064 & 1.830 wf3 & 2489 & 438.016 & 410.038 & 18.854 & 0.095 & 1.700 & 17.995 & 0.057 & 1.240 wf4 & 4532 & 300.305 & 699.237 & 19.152 & 0.031 & 0.739 & 17.418 & 0.030 & 0.877 wf3 & 2474 & 441.016 & 410.137 & 19.183 & 0.111 & 1.823 & 17.977 & 0.073 & 1.819 wf2 & 150 & 321.521 & 80.650 & 19.311 & 0.036 & 0.657 & 16.489 & 0.089 & 0.941 wf2 & 2719 & 310.787 & 523.757 & 19.669 & 0.070 & 1.499 & 18.432 & 0.063 & 1.959 wf3 & 2554 & 442.354 & 412.389 & 19.741 & 0.101 & 1.325 & 18.639 & 0.076 & 1.507 lrrrrrrrrr wf3 & 503 & 689.043 & 346.030 & 17.369 & 0.050 & 0.827 & 16.723 & 0.025 & 0.758 wf2 & 1198 & 252.106 & 634.779 & 17.758 & 0.057 & 0.745 & 16.875 & 0.055 & 0.830 wf4 & 537 & 650.877 & 360.807 & 17.934 & 0.051 & 0.982 & 17.217 & 0.026 & 0.801 wf3 & 802 & 417.826 & 417.893 & 17.983 & 0.060 & 1.509 & 17.020 & 0.061 & 1.704 wf4 & 869 & 310.743 & 507.150 & 18.569 & 0.028 & 0.630 & 17.479 & 0.026 & 0.789 wf3 & 811 & 419.875 & 419.020 & 18.662 & 0.078 & 1.841 & 17.422 & 0.053 & 1.494 wf3 & 1654 & 229.986 & 766.329 & 18.743 & 0.035 & 0.907 & 18.116 & 0.019 & 0.601 wf3 & 412 & 167.937 & 301.246 & 18.896 & 0.034 & 0.877 & 16.521 & 0.030 & 0.756 wf3 & 770 & 415.495 & 417.557 & 19.926 & 0.055 & 1.014 & 18.819 & 0.065 & 1.407 wf3 & 1451 & 676.379 & 661.313 & 20.225 & 0.038 & 0.948 & 18.614 & 0.025 & 0.736 lrrrrrrrrrr wf3 & 2583 & 437.139 & 413.836 & @xmath293 & 18.195 & 0.065 & 1.341 & 17.356 & 0.059 & 1.551 wf3 & 2606 & 439.952 & 415.701 & @xmath294 & 18.454 & 0.063 & 1.462 & 17.642 & 0.064 & 1.830 wf3 & 2489 & 438.016 & 410.038 & @xmath295 & 18.854 & 0.095 & 1.700 & 17.995 & 0.057 & 1.240 wf3 & 2474 & 441.016 & 410.137 & @xmath296 & 19.183 & 0.111 & 1.823 & 17.977 & 0.073 & 1.819 wf3 & 2554 & 442.354 & 412.389 & @xmath294 & 19.741 & 0.101 & 1.325 & 18.639 & 0.076 & 1.507 wf3 & 2532 & 429.789 & 412.006 & @xmath297 & 20.376 & 0.079 & 1.847 & 19.356 & 0.067 & 1.961 wf3 & 2420 & 440.762 & 402.723 & @xmath298 & 20.812 & 0.071 & 1.629 & 19.648 & 0.063 & 1.828 wf3 & 2553 & 428.372 & 414.756 & @xmath299 & 20.947 & 0.058 & 1.355 & 19.735 & 0.043 & 1.256 wf3 & 2460 & 418.943 & 405.789 & @xmath300 & 21.148 & 0.027 & 0.679 & 19.992 & 0.029 & 0.922 wf3 & 2456 & 427.305 & 405.329 & @xmath301 & 21.241 & 0.048 & 1.071 & 19.853 & 0.046 & 1.304 lrrrrrrrrrr wf3 & 802 & 417.826 & 417.893 & @xmath302 & 17.983 & 0.060 & 1.509 & 17.020 & 0.061 & 1.704 wf3 & 811 & 419.875 & 419.020 & @xmath303 & 18.662 & 0.078 & 1.841 & 17.422 & 0.053 & 1.494 wf3 & 770 & 415.495 & 417.557 & @xmath304 & 19.926 & 0.055 & 1.014 & 18.819 & 0.065 & 1.407 wf3 & 866 & 417.573 & 427.651 & @xmath305 & 22.019 & 0.086 & 1.941 & 20.250 & 0.045 & 1.217 wf3 & 783 & 429.730 & 414.665 & @xmath306 & 22.086 & 0.056 & 1.330 & 20.499 & 0.050 & 1.467 wf3 & 803 & 405.224 & 418.070 & @xmath307 & 22.180 & 0.053 & 1.224 & 20.275 & 0.048 & 1.413 wf3 & 841 & 396.894 & 424.414 & @xmath308 & 22.310 & 0.073 & 1.267 & 20.965 & 0.059 & 1.244 wf3 & 871 & 412.104 & 427.591 & @xmath309 & 22.355 & 0.072 & 1.544 & 21.183 & 0.066 & 1.812 wf3 & 915 & 418.244 & 434.008 & @xmath310 & 22.645 & 0.079 & 1.777 & 21.687 & 0.068 & 1.789 wf3 & 796 & 434.824 & 416.729 & @xmath311 & 22.676 & 0.041 & 0.920 & 21.468 & 0.035 & 0.953 ccccccc 20.25 & & & 20.25 & 0.031 & 20.75 & & & 20.75 & 0.026 & 0.023 21.25 & & & 21.25 & 0.031 & 0.028 21.75 & 0.038 & & 21.75 & 0.036 & 0.033 22.25 & 0.037 & & 22.25 & 0.038 & 0.039 22.75 & 0.043 & 0.046 & 22.75 & 0.046 & 0.043 23.25 & 0.049 & 0.049 & 23.25 & 0.051 & 0.049 23.75 & 0.055 & 0.051 & 23.75 & 0.067 & 0.063 24.25 & 0.065 & 0.062 & 24.25 & 0.081 & 0.077 24.75 & 0.079 & 0.077 & 24.75 & 0.118 & 0.099 25.25 & 0.094 & 0.090 & 25.25 & 0.145 & 0.153 25.75 & 0.129 & 0.121 & 25.75 & 0.217 & 0.222 26.25 & 0.176 & 0.182 & 26.25 & 0.333 & 0.314 26.75 & 0.262 & 0.271 & 26.75 & 0.570 & 0.457 27.25 & 0.356 & 0.310 & 27.25 & & 27.75 & & & 27.75 & & 28.25 & & & 28.25 & & rrrrrrrrrr 22.523.0 & 22.947 & 0.058 & @xmath312 & 0.039 & 1.536 & 0.074 & @xmath313 & 0.049 & 2 23.023.5 & 23.361 & 0.065 & @xmath314 & 0.061 & 1.382 & 0.082 & @xmath315 & 0.028 & 19 23.524.0 & 23.800 & 0.064 & @xmath316 & 0.056 & 1.237 & 0.089 & @xmath317 & 0.062 & 31 24.024.5 & 24.262 & 0.081 & @xmath318 & 0.055 & 1.081 & 0.102 & @xmath319 & 0.064 & 55 24.525.0 & 24.781 & 0.087 & @xmath320 & 0.085 & 1.007 & 0.115 & @xmath321 & 0.087 & 78 25.025.5 & 25.251 & 0.104 & @xmath322 & 0.098 & 0.918 & 0.138 & @xmath316 & 0.094 & 110 25.526.0 & 25.752 & 0.128 & @xmath323 & 0.176 & 0.890 & 0.177 & @xmath324 & 0.187 & 116 26.026.5 & 26.235 & 0.179 & @xmath325 & 0.277 & 0.881 & 0.244 & @xmath326 & 0.192 & 88 26.527.0 & 26.654 & 0.232 & @xmath327 & 0.412 & 0.873 & 0.319 & @xmath328 & 0.184 & 33 27.027.5 & 27.076 & 0.308 & @xmath329 & 0.052 & 0.826 & 0.415 & @xmath330 & 0.169 & 3 rrrrrrrrrr 24.024.5 & 24.384 & 0.073 & @xmath331 & 0.100 & 0.623 & 0.106 & @xmath332 & 0.159 & 12 24.525.0 & 24.911 & 0.086 & @xmath333 & 0.123 & 0.429 & 0.134 & @xmath334 & 0.181 & 94 25.025.5 & 25.182 & 0.102 & @xmath335 & 0.090 & 0.537 & 0.153 & @xmath336 & 0.136 & 823 25.526.0 & 25.615 & 0.214 & @xmath337 & 0.200 & 0.945 & 0.236 & @xmath338 & 0.293 & 12 26.026.5 & 26.054 & 0.246 & @xmath339 & 0.000 & 1.569 & 0.274 & @xmath340 & 0.000 & 1 lccccrrc g1 & @xmath341 & @xmath342 & @xmath343 & @xmath344 & 1523 & @xmath345 & 4 g11 & @xmath346 & @xmath347 & @xmath348 & @xmath344 & 757 & @xmath349 & 2 g58 & @xmath350 & @xmath351 & @xmath351 & @xmath352 & 282 & @xmath353 & 1 g105 & @xmath354 & @xmath355 & @xmath355 & @xmath356 & 648 & @xmath357 & 1 g108 & @xmath358 & @xmath359 & @xmath360 & @xmath361 & 208 & @xmath362 & 1 g219 & @xmath363 & @xmath364 & @xmath365 & @xmath356 & 872 & @xmath366 & 1 g302 & @xmath367 & @xmath368 & @xmath369 & @xmath344 & 321 & @xmath370 & 3 g312 & @xmath371 & @xmath372 & @xmath373 & @xmath344 & 498 & @xmath374 & 3 g319 & @xmath375 & @xmath376 & @xmath377 & @xmath344 & 721 & @xmath378 & 2 g323 & @xmath379 & @xmath380 & @xmath381 & @xmath344 & 538 & @xmath382 & 2 g327 & @xmath383 & @xmath384 & @xmath385 & @xmath344 & 997 & @xmath386 & 2 g352 & @xmath387 & @xmath388 & @xmath389 & @xmath344 & 871 & @xmath390 & 2 rrrrr 21.75 & 3 & 2.719 & 0 & 0.000 22.25 & 17 & 16.184 & 0 & 0.000 22.75 & 26 & 22.139 & 2 & 0.725 23.25 & 32 & 20.397 & 19 & 16.635 23.75 & 46 & 34.960 & 12 & 4.148 24.25 & 58 & 61.228 & 25 & 26.724 24.75 & 125 & 166.622 & 48 & 57.009 25.25 & 116 & 43.772 & 85 & 139.459 25.75 & 62 & 0.000 & 36 & 61.921 26.25 & 50 & 142.707 & 26 & 157.769 26.75 & 25 & 0.000 & 13 & 313.156 27.25 & 4 & 0.000 & 3 & 216.386 ccrrrrrrrr g302 & @xmath51 & 7.56 & @xmath391 & @xmath392 & 1.70 & @xmath393 & @xmath394 & 1.827 & 49 & & 7.61 & @xmath391 & @xmath395 & 1.71 & @xmath396 & @xmath397 & 1.855 & 48 & @xmath153 & 7.65 & @xmath391 & @xmath398 & 1.73 & @xmath393 & @xmath399 & 0.846 & 51 & & 7.47 & @xmath400 & @xmath401 & 1.67 & @xmath402 & @xmath403 & 0.906 & 48 g312 & @xmath51 & 7.57 & @xmath404 & @xmath405 & 1.70 & @xmath393 & @xmath406 & 0.839 & 53 & & 7.48 & @xmath391 & @xmath407 & 1.67 & @xmath408 & @xmath403 & 0.882 & 52 & @xmath153 & 7.46 & @xmath400 & @xmath409 & 1.67 & @xmath393 & @xmath403 & 0.458 & 53 & & 7.47 & @xmath400 & @xmath410 & 1.67 & @xmath411 & @xmath403 & 0.467 & 52 | we present deep ( @xmath0 ) _ v_- and _ i_-band stellar photometry of g302 and g312 , two globular star clusters in the halo of m31 .
these data were obtained using the _ hubble space telescope _ s wide field / planetary camera 2 .
we find iron abundances of @xmath1}}= -1.85 \pm 0.12 $ ] for g302 and @xmath1}}= -0.56 \pm 0.03 $ ] for g312 , consistent with spectroscopic measurements . the color
magnitude diagrams for each cluster show no evidence for an intermediate - aged population of stars , or a second parameter effect in the morphology of the horizontal branch .
g302 shows no evidence for a color gradient but the inner regions of g312 are bluer than the outer regions .
g312 shows no evidence of ellipticity or an extended halo of unbound stars .
g302 has a projected ellipticity of @xmath2 with the projected major axis oriented towards the center of m31 .
g302 also shows evidence of an extended asymmetric stellar halo extending to at least twice the fitted michie
king tidal radius .
the amount of mass beyond the tidal radius of g302 is consistent with the stellar escape rates which have been predicted by @xmath3-body simulations of globular clusters in the galactic tidal field . |
it is well known that the quantum fluctuations of the electromagnetic ( em ) field coupled to electric charges generate an observable interaction energy @xcite .
the fluctuations which mediate the self - interaction of electrons bound to atomic nuclei give rise to the lamb shift of atomic levels ; the fluctuations which mediate the interaction between nearby molecules generate van - der - waals energies ; and finally the fluctuations between macroscopic dielectrics generate the casimir energy .
direct observation of these energies is possible by spectroscopy , atomic interferometry or nanomechanical means @xcite .
+ less well - known is the fact that other observable quantities , functions of the em field , can be influenced by quantum fluctuations under certain symmetry conditions .
that is , when the time - space symmetries of the medium to which the fluctuations couple are compatible with the symmetries of some observable operator , the expectation value of that operator in the vacuum state of the system medium - em field may take a non - zero value .
this is the case of the linear momentum of the em field when quantum fluctuations couple to a medium in which both parity ( p ) and time - reversal ( t ) symmetries are broken .
in particular , the constitutive equations of a magneto - electric medium contain a non - reciprocal ( _ nr _ ) em susceptibility @xcite , @xmath0 , which results from broken p and t. generally @xmath0 is an antisymmetric t , p - odd tensor which generates an electric polarization in response to a magnetic field , @xmath1 , and conversely , a magnetization as a response to an electric field , @xmath2 . as a result of the matter - field coupling and momentum conservation
, linear momentum can be transferred from the em vacuum to matter during the process that controls the break down of the symmetries .
a nonzero tensor @xmath3 is found in any medium in crossed external electric and magnetic fields , in a moving dielectric medium and in a chiral medium exposed to a magnetic field @xcite . in this article
we concentrate on the latter case . + it is a generic phenomenon in field theory that the breakdown of a symmetry is accompanied by a non - zero vacuum expectation value ( vev ) of some physical observable associated to the symmetry . in our case
the p and t symmetries happen to be broken explicitly by the presence of a chiral molecule and the action of an external magnetic field , @xmath4 .
correspondingly , a non - zero vev of the em momentum shows up in the direction along which the symmetries are broken , @xmath4 .
the question arises whether it could be possible to take advantage of this phenomenon for practical purposes . to this end
we will show that , due to the conservation of total linear momentum , there exists necessarily a transfer of kinetic momentum to the chiral molecule of equal magnitude and opposite sign to the vev of the em momentum .
+ also , in the context of high energy physics , it is known that the electro - weak interaction violates the p and t symmetries @xcite , and cp ( i.e. , t ) is also expected to be naturally broken in qcd . in the latter case , the existence of a light pseudo - scalar particle , the axion , has been postulated as a solution of the so - called _ strong cp problem _
@xcite , and indirect observations of the axion through its coupling to the em field have been suggested @xcite and put into practice @xcite . in the pvlas experiment
@xcite signatures of the axion - photon interaction are investigated through the anomalous rotation of the polarization direction of optical light in the presence of an intense magnetic field , similar to the vacuum birefringence effect expected from the polarization of the qed vacuum @xcite .
nonetheless the rotation of the polarization axis of light may be caused by the break down of the p or t symmetries separately , while the effect described in the present article needs the simultaneous breakdown of both symmetries and is independent of polarization .
+ already in classical electrodynamics , a kinetic momentum is acquired by matter under external time - dependent electric and magnetic fields perpendicular to each other @xcite .
conservation of total momentum implies that the classical em momentum and the kinetic momentum must be equal in magnitude and opposite in sign .
several expressions can be found in the literature for this classical em momentum , all derived from some kind of phenomenological assumptions .
this controversy is known as the abraham - minkowski problem . according to abraham s prescription , the density of em momentum is @xmath5 , while minkowski s formula reads @xmath6 .
more recently , the microscopical approach of nelson @xcite , which appeals to the lorentz force over bound and free charges and currents for the coupling between matter and radiation , predicts a momentum density of an em wave in a material medium , @xmath7 .
+ starting with a classical lagrangian that reproduces maxwell s equations and using the constitutive equations of a homogeneous and isotropic medium of density @xmath8 moving at velocity @xmath9 , feigel @xcite found a total conserved momentum density @xmath10 , from which the kinetic momentum density of matter is found , @xmath11 @xcite .
however , following an approach close to nelson s , van tiggelen has arrived to another conservation law @xcite . on the one hand , the combination of maxwell s equations and the constitutive equations for a moving medium of mass density @xmath12 yields , @xmath13 , with @xmath14 and @xmath15 the lorentz force density , @xmath16 being the density of charge . on the other hand , newton s law reads @xmath17 .
lastly , combining both equations and integrating in space one obtains the total conserved momentum @xmath18 .
the equation of motion which derives from this approach is @xcite @xmath19 .
+ in addition to the classical em momentum coming from external crossed fields , em quantum fluctuations can generate extra terms when the constitutive equations contain non - reciprocal susceptibilities , @xmath0 @xcite .
the basic argument is that a non - zero @xmath0 generates what is called spectral non - reciprocity .
that is , there exists a difference in the frequency of normal modes propagating in opposite directions along the axis where the p symmetry is broken , @xmath20 .
this implies that the addition of the linear momenta propagated by these modes do not cancel out .
the total momentum of the normal modes is the momentum of the vacuum fluctuations and therefore it is referred to as casimir momentum .
+ as pointed out in refs.@xcite feigel s approach suffers from a number of problems . in the first place ,
since @xmath21 is proportional to the energy current , one expects that in thermodynamic equilibrium @xmath22 , and so no variation of the kinetic momentum can be obtained from variations of @xmath23 .
second , the final result obtained by feigel for the casimir momentum density , @xmath24 , seems to lack a reference frame since this equation is not lorentz or even galilean invariant @xcite .
+ in addition , the derivations of the above expressions for the conserved linear momentum are purely phenomenological and lack an explicit quantum interaction hamiltonian between matter and radiation .
also , a macroscopic approach can easily lead to an erroneous prediction .
for instance , it is shown in ref.@xcite that for a chiral distribution of electrically polarizable molecules in an external magnetic field , @xmath25 should hold locally , for a non - magnetic medium , and the energy flow @xmath26 vanishes . ] although the medium contains a nonzero nonreciprocal effective ( _ eff _ ) response which would make @xmath27 .
another problem that should be resolved in a quantum microscopic treatment is the uv divergence of the expression for the quantum em momentum density obtained in a homogeneous medium @xcite .
this line of investigation was started by kawka , van - tiggelen and rikken in refs.@xcite , where they computed the casimir momentum of an atom in crossed external electric and magnetic fields .
the divergences were found there to disappear by mass renormalization and the leading quantum contribution was found to be a factor @xmath28 smaller than the classical em momentum . + in ref.@xcite we reported the quantum computation of the casimir momentum for a chiral molecule in a uniform external magnetic field . in our model chirality
breaks the parity symmetry and the magnetic field breaks time reversal , so it is symmetry allowed to have a non - vanishing value for the momentum of the electromagnetic field . in this case
no classical contribution to the abraham force exists since there is no external electric field . in this article
we perform a semiclassical computation of the casimir momentum for the same model and compare it with the quantum result .
we explain the physical mechanisms which mediate the transfer of linear momentum from the em vacuum to the chiral molecule and we analyse the transfer of energy . + the article is organized as follows . in section [ sec2 ]
we present the model .
next , we perform a semiclassical computation of the casimir momentum in section [ sec3 ] . in section [ qapproach ]
we review the quantum computation of ref.@xcite , paying special attention to the physical mechanisms involved in the transfer of momentum . in section [ sec5 ]
we explain the exchange of energy between molecule , quantum vacuum and magnetic field . in section [ sec6 ]
we summarize our conclusions .
we propose the simplest model for a chiral molecule that exhibits all necessary features to leading order in perturbation theory : broken mirror symmetry , zeeman splitting of energy levels and coupling to the quantum vacuum , and we neglect relativistic effects . in our model the optical activity of the molecule is determined by a single chromophoric electron within a chiral object which is further simplified to be a two - particle system in which the chromophoric electron of charge @xmath29 and mass @xmath30 is bound to a nucleus of effective charge @xmath31 and mass @xmath32 . the binding interaction is modeled by a harmonic oscillator potential , @xmath33 , to which we add a term @xmath34 to break the mirror symmetry perturbatively in first order .
the coordinates @xmath35 , @xmath36 , @xmath37 are those of the relative position vector , @xmath38 , and @xmath39 with @xmath40 .
the center of mass position vector is @xmath41 .
the conjugate momentum of @xmath42 is @xmath43 , while the conjugate momentum of @xmath44 is the total conjugate momentum , @xmath45 .
@xmath46 was first introduced by condon _
et al_. @xcite to explain the rotatory power of chiral compounds with a single oscillator model .
both the anisotropy in @xmath47 and the chiral potential @xmath46 are determined by the coulomb interaction of the two - body system with the rest of atoms within the molecule .
in particular , the parameter @xmath48 is the sum of all third order coefficients of the expansion of the coulomb interaction of the chromophoric group with the surrounding charges around their mean distance [ see eq.(42 ) of ref.@xcite ] .
it is a pseudo - scalar which does not necessarily vanish for a chiral environment . in principle , all parameters of this model can be computed _ ab initio _ from the experimental data @xcite .
+ when an external uniform and constant magnetic field @xmath4 is applied the total hamiltonian of the system reads , @xmath49 , with @xmath50^{2}+v^{ho}+v_{c},\label{h0}\\ h_{em}&=\sum_{\mathbf{k},\mathbf{\epsilon}}\hbar\omega_{\mathbf{k}}(a^{\dagger}_{\mathbf{k}\mathbf{\epsilon}}a_{\mathbf{k}\mathbf{\epsilon}}+\frac{1}{2 } ) + \frac{1}{2\mu_{0}}\int\textrm{d}^{3}r\mathbf{b}_{0}^{2},\label{hem}\\ w&=\sum_{i = e , n}\frac{-q_{i}}{m_{i}}[\mathbf{p}_{i}-q_{i}\mathbf{a}_{0}(\mathbf{r}_{i})]\cdot\mathbf{a}(\mathbf{r}_{i } ) + \frac{q^{2}_{i}}{2m_{i}}\mathbf{a}^{2}(\mathbf{r}_{i}),\label{elw}\end{aligned}\ ] ] where @xmath51 is the minimal coupling interaction potential . in the vector potential
we have separated the contribution of the external classical field , @xmath52 , from the one of the quantum field operator , @xmath53 .
note that , having incorporated the internal electrostatic interaction of the two - body system within @xmath47 and the electrostatic interaction of the system with the surrounding within @xmath46 , the em vector potential in @xmath51 is just transverse . in the sum ,
@xmath54 and @xmath55 are the creation and annihilation operators of photons with momentum @xmath56 , frequency @xmath57 and polarization vector @xmath58 respectively .
the magnetostatic energy is a constant irrelevant to us that we will discard .
+ in the absence of coupling to the vacuum field , the system with hamiltonian @xmath59 possesses a conserved pseudo - momentum , @xmath60 , which satisfies @xmath61=\mathbf{0}$ ] and has continuous eigenvalues @xmath62 . here
@xmath63 is the kinetic momentum of the center of mass , @xmath64 , which relates to @xmath65 through @xmath66 $ ] .
the unitary operator u@xmath67}$ ] maps the hamiltonian @xmath59 into @xmath68u@xmath69u , which conveniently separates the motion of the center of mass from the relative motion @xcite , @xmath70 in this equation @xmath71 is the zeeman potential with @xmath72 .
terms of order @xmath73 and @xmath74 are cast in @xmath75 . in the following ,
both @xmath46 and @xmath76 will be considered first order perturbations to the harmonic oscillator potential and higher order perturbative terms like those in @xmath77 will be neglected .
+ the ground state of the hamiltonian @xmath78 is , up to order @xmath48b@xmath79 in stationary perturbation theory , @xmath80 correspondingly , the ground state of @xmath59 is @xmath81u@xmath82 , with a pseudo - momentum @xmath83 to be fixed . the fact that @xmath84 implies that @xmath83 is the kinetic momentum of the oscillator , i.e. , no other contribution exists to the pseudo - momentum @xmath85 of the bare molecule in its ground state . in the above equation
the states @xmath86 refer to the eigenstates of the harmonic oscillator hamiltonian .
the dimensionless parameters are , @xmath87 with @xmath88 and @xmath89 , @xmath90 , @xmath91 . here , the indices @xmath92 take on the three spatial directions , @xmath93 .
the @xmath94 factors are assumed to be small quantities which quantify the anisotropy of the oscillator .
they _ all _ have to be nonzero for the optical activity of the molecule to survive rotational averaging .
the dimensionless parameters , @xmath95 , @xmath96 and @xmath97 , are the expansion parameters of our perturbative calculations .
we will restrict to the lowest order in them all .
in ref.@xcite van tiggelen has derived an expression for the casimir momentum using a green s function formalism which is compatible with nelson s prescription . that formalism was latter applied in ref.@xcite to compute the casimir momentum in a configuration of magnetic dipoles disposed in a twisted h configuration . + in our case we are interested in the casimir momentum in the presence of a single molecule . as noticed in ref.@xcite , the casimir momentum which derives from abraham and nelson s prescriptions coincide approximately for a diluted medium ( same applies to the forces generated by time varying fields ) . for the case of a single particle
the so - called abraham momentum reads @xmath98 , @xmath99 being the electric dipole moment of the molecule . here
we apply a linear response formalism upon @xmath98 following a similar semiclassical treatment to the one for the lamb shift of a single dipole or the one for the van der waals forces between atoms @xcite .
this is equivalent to the green s function formalism of refs.@xcite considering a single magnetochiral scatterer .
+ we aim to compute the vacuum expectation value @xmath100 .
we start the calculation by decomposing both the electric dipole operator and the magnetic field into induced ( _ ind _ ) and free ( _ fr _ ) components .
the induced operators relate to the free operators through the linear susceptibilities , i.e. , polarizabilities and green s functions .
any space - time dependent free operator @xmath101 can be written in fourier space as a sum of positive frequency and negative frequency components , @xmath102 $ ] , where @xmath103 and @xmath104 are the @xmath105-mode creation and annihilation operators of @xmath106 respectively . in the following
we will be interested in the em field and dipole moment operators evaluated at the position of the molecule , @xmath107 . therefore , unless necessary , we will omit the dependence of the operators on the position . in the case of the electric field of a plane wave propagating in free space in the direction @xmath108
, we find @xmath109 , with @xmath110 , and likewise for the magnetic field .
therefore , together with the dependence on @xmath105 we will add the dependence on @xmath111 to the electric field annihilation / creations operator of a plane wave , @xmath112 _idem _ for the magnetic field .
+ since the relation between induced and free operators is linear and the action of annihilation operators on the vacuum vanishes , it follows from @xmath100 that @xmath113 in ref.@xcite we have obtained the constitutive equations of our magneto - chiral model by computing the response of the molecule to a monochromatic em plane wave of frequency @xmath105 and wave vector @xmath111 . up to electric quadrupole contributions
, the induced electric dipole moment operator reads , @xmath114,\label{dind } % \textrm{m}^{ind}_{i}(\omega)&=&\alpha_{m}\delta_{ij}\textrm{b}_{fr}^{j}(\mathbf{k},\omega ) + \zeta\epsilon_{ijk}\textrm{b}^{j}_{0}\dot{\textrm{b}}_{fr}^{k}(\mathbf{k},\omega ) % + \beta\delta_{ij}\dot{\textrm{e}}_{fr}^{j}(\mathbf{k},\omega)\nonumber\\&-&\gamma\epsilon_{ijk}\textrm{b}^{j}_{0}\textrm{e}_{fr}^{k}(\mathbf{k},\omega),\label{mind}\\ % \langle\theta_{ij}\rangle&=&i\xi(\textrm{e}^{fr}_{i}(\mathbf{k},\omega)\textrm{b}^{0}_{j } % + \textrm{e}^{fr}_{j}(\mathbf{k},\omega ) \textrm{b}^{0}_{i})+i\varsigma(\mathbf{b}_{0}\cdot\mathbf{e}_{fr}(\mathbf{k},\omega))\delta_{ij}.\nonumber\end{aligned}\ ] ] where a rotational average is implicit in this equation .
the factor @xmath115 is the ordinary electric polarizability , @xmath116 describes the faraday effect , @xmath117 is the molecular rotatory factor responsible for the natural optical activity , and @xmath118 and @xmath119 give rise to the magnetochiral anisotropy .
the expressions of all these factors have been summarized in [ appendb ] .
it has been shown in ref.@xcite that , once the rotational power and the refractive index of a compound are given i.e . , the parameters @xmath115 and @xmath117 , all other parameters of our model can be deduced .
+ defining the effective electric polarizability and the crossed magneto - electric polarizability tensors respectively as , @xmath120\delta^{ij}+\textrm{k}^{i}\textrm{b}_{0}^{j}-i\omega\chi\epsilon_{ikj}\textrm{b}^{k}_{0 } , \nonumber\\ \alpha_{em}^{ij}(\omega)&=i\omega\beta\delta^{ij}+\gamma\epsilon_{ikj}\textrm{b}^{k}_{0},\end{aligned}\ ] ] we can write eq.([dind ] ) as , @xmath121 using maxwell s equation for the free fields , @xmath122 , we can write @xmath123 as a response to the free electric field alone , @xmath124\mathbf{k}.\end{aligned}\ ] ] from here it is obvious that the non - reciprocal response comes from the fourth term on the r.h.s . , which depends on the relative direction of the wave vector , @xmath111 , with respect to the external magnetic field
. we will denote this non - reciprocal ( _ nr _ ) polarizability by @xmath125 .
it is at the origin of the magnetochiral birefringence @xcite .
the induced magnetic field can be written as a linear response to the free dipole located at @xmath107 , @xmath126,\label{bind}\end{aligned}\ ] ] where @xmath127 is the @xmath105-mode of the free magnetic dipole moment operator and the green functions @xmath128 and @xmath129 relates to the green function of maxwell s equation for the em vector potential in free space , @xmath130\mathbb{g}^{(0)}(\mathbf{r}-\mathbf{r}';\omega ) = \delta^{(3)}(\mathbf{r}-\mathbf{r}')\mathbb{i},\ ] ] @xmath131 when substituting the induced operators in eq.([pcaso ] ) as functions of the free operators we find the vacuum expectation values of bilinear operators .
those expectation values ( dipole and field quadratic fluctuations ) relate to the imaginary part of their respective linear response functions through the fluctuation - dissipation theorem @xcite , @xmath132 @xmath133_{ij}\delta(\omega-\omega')\nonumber\\&= \frac{\hbar}{\pi}\epsilon_{ikj}\textrm{b}^{k}_{0}\im{\{\gamma\}}\delta(\omega-\omega')\nonumber\\&+\textrm{reciprocal terms}\label{13},\end{aligned}\ ] ] where it is understood that the dipole moment operators act upon the location of the only molecule at @xmath107 . in the last equation above we use the result obtained in ref.@xcite , @xmath134 , and we restrict to nonreciprocal terms for simplicity . using the linear response relations of eqs.([dresp ] ) and
( [ bind ] ) in eq.([pcaso ] ) and applying the expectation values of eq.([13 ] ) we end up with the relation , @xmath135,\label{cc}\end{aligned}\ ] ] where @xmath136 and @xmath137 are the fourier transforms of @xmath138 and @xmath139 in @xmath140-space respectively , @xmath141 which are necessary to account for the dependence of @xmath142 on the wave vector .
symmetry considerations imply that only the nonreciprocal terms of @xmath142 and @xmath143 in @xmath144 survive the angular integration . after integrating over @xmath140
we arrive at @xmath145 this expression is nothing but feigel s formula @xcite simplified to the case of a diluted medium in the one - particle limit @xmath146 , @xmath147 being the total volume and @xmath8 the numerical density of molecules @xcite .
hence , following croze s argument @xcite we can write the above integral as the sum over the momenta of dressed normal modes , @xmath148 , with @xmath149 being the index of refraction , @xmath150 with @xmath151 the nonreciprocal part of the refractive index due to magnetochiral birefringence . at leading order in @xmath8
it reads @xcite , @xmath152 . in the one - particle limit , @xmath146 , and
turning the summation into an integral in the continuum limit , eqs.([ec1 ] ) and ( [ ec2 ] ) coincide . from the semiclassical result it follows that @xmath153 is the momentum of radiative modes only .
this result would reinforce the interpretation that the casimir momentum originates in the non - reciprocity of the spectrum of normal modes propagating in the effective medium , i.e. @xmath154 , with @xmath155 the effective frequencies of photons propagating in opposite directions with bare momentum vectors @xmath156 . to conclude
, we use the explicit formulas for @xmath118 and @xmath119 in [ appendb ] and integrate over frequencies , @xmath157 where @xmath158 .
it is remarkable that the frequency integral does not present any uv divergence , unlike the expression for @xmath153 in ref.@xcite for a magneto - electric medium in crossed fields .
we will see later on that this semiclassical result is included in the microscopic qed result , but it enters as a small correction to the leading term .
the system of the hamiltonian @xmath159 possesses a conserved pseudo - momentum @xcite , @xmath160 which satisfies @xmath161=\mathbf{0}$ ] .
its eigenvalues are therefore good quantum numbers .
the terms in @xmath162 can be also arranged as , @xmath163 where , beside the kinetic momentum , @xmath164 is the abraham momentum and we define the casimir momentum operator , @xmath165 , as the momentum operator of the vacuum field .
@xmath153 is composed of a longitudinal casimir momentum , @xmath166 $ ] , and a transverse casimir momentum , @xmath167 @xcite .
the latter is just the sum of the momenta @xmath168 of radiative photons .
the longitudinal momentum is more subtle .
it stems from the transverse electromagnetic gauge field coupled to electric charges .
note that in the coulomb gauge @xmath169 is fully transverse .
@xmath170 is referred to as `` longitudinal '' component since it can be written as the integral over the vector product of the longitudinal coulomb electric field and the magnetic field generated by the charges , @xmath171 @xcite .
+ under the action of a time - varying magnetic field , @xmath172 , the time derivative of the expectation value of @xmath162 in the ground state , @xmath173 , vanishes , @xmath174\rangle+e\frac{\partial\mathbf{b}_{0}(t)}{\partial t}\wedge\langle\mathbf{r}\rangle=\mathbf{0}.\nonumber\ ] ] this follows from @xmath161=\mathbf{0}$ ] and @xmath175 for the chiral but unpolarized ground state . @xcite . ]
the latter ensures also that the variation of the abraham momentum vanishes in the ground state . from the expectation value of eq.([p ] ) it follows that , for an arbitrary variation of the magnetic field , the variation of the kinetic momentum of the chiral oscillator is equivalent in magnitude and opposite in sign to the variation of the casimir momentum of the vacuum field , @xmath176 let us consider the molecule initially at rest in its ground state at zero magnetic field . during the switching of the magnetic field eq.([dps ] ) implies that @xmath177 at any time .
this relation makes the casimir momentum an observable quantity . in particular
this equality holds well after the switching has ended and the magnetic field achieves a stationary value @xmath4 . in the following
we evaluate the casimir momentum in the asymptotically stationary situation in which the molecule is in its ground state at constant kinetic momentum , @xmath83 , constant magnetic field , @xmath4 and once coupled to the em vacuum .
we denote by @xmath178 this asymptotic state of the molecule , which we will compute applying up to second order perturbation theory to @xmath179 with the interaction potential @xmath51 .
we define @xmath180 , and calculate separately the transverse momentum , @xmath181 , and the longitudinal momentum , @xmath182 .
+ for the computation of @xmath183 at @xmath184 and lowest order in the fine structure constant , @xmath185 , we need to compute @xmath178 applying second order perturbation theory to @xmath179 .
note that this implies applying up to fourth - order perturbation theory in @xmath186 to the ground state of the harmonic oscillator hamiltonian , @xmath187 .
we use the u - transformed states and the u - transformed potential , with u@xmath188}$ ] , @xmath189 to arrive at , @xmath190 where @xmath191 , @xmath192 and @xmath193 are the energies of the intermediate states , @xmath194 ( _ idem _ for the prime states ) .
the atomic states @xmath195 are eigenstates of @xmath78 and may have a priori any pseudo - momentum @xmath62 .
the em states , @xmath196 , are 1-photon states with momentum @xmath56 and polarization vector @xmath58 .
zeros in the denominator are avoided in the summation .
writing the em quantum field in eq.([tildew ] ) as usual @xcite , @xmath197,\ ] ] with @xmath147 a generic volume ; passing the sums over @xmath62 , @xmath198 , @xmath111 , @xmath199 and @xmath200 in eq.([pcos ] ) to continuum integrals and by summing over polarization states we arrive at , @xmath201\nonumber\\ & \times\frac{\cdot(\mathbb{i}-\frac{\mathbf{k}\otimes\mathbf{k}}{k^{2}})\cdot } { ( \hbar^{2}k^{2}/2m+\hbar ck-\hbar\mathbf{k}\cdot\mathbf{\mathbf{q}}_{0}/m - e_{0}+h_{0})^{2}}\nonumber\\ & \times\bigl[e^{-i\frac{m_{e}}{m}\mathbf{k}\cdot\mathbf{r}}(\mathbf{p}/m_{n}+\mathbf{q}_{0}/m-\frac{e}{2m_{n}}\mathbf{b}_{0}\wedge\mathbf{r})\nonumber\\ & + e^{i\frac{m_{n}}{m}\mathbf{k}\cdot\mathbf{r}}(\mathbf{p}/m_{e}-\mathbf{q}_{0}/m+\frac{e}{2m_{e}}\mathbf{b}_{0}\wedge\mathbf{r})\bigr]|\tilde{\omega}_{0}\rangle . \nonumber\end{aligned}\ ] ] in this equation we can distinguish four terms . in two of them
the exponentials to the r.h.s . and l.h.s .
of the fraction compensate .
they correspond to the feynman diagrams in which a photon is created and annihilated at the position of one of the particles , i.e. , either at the nucleus or at the electron position [ fig.[fig1]i(@xmath202 ) and fig.[fig1]i(@xmath203 ) , respectively ] . in the other two terms
the exponentials amount to @xmath204 .
they correspond to the feynman diagrams in which the virtual photons are created and annihilated in different particles [ fig.[fig1]ii(@xmath202 ) and fig.[fig1]ii(@xmath203 ) ] . in the latter the complex exponentials evaluated in the states of the harmonic oscillator yield an effective cut - off for the momentum integrals at @xmath205 , making their contribution negligible w.r.t .
the other diagrams . among the diagrams of fig.[fig1]i
the dominant one is i(@xmath203 ) in which virtual photons are created and annihilated at the electron position , as for the non - relativistic calculation of the lamb shift @xcite .
it reads , [ fig1 ] . for the case of @xmath206
there are two intermediate atomic states , @xmath207 and @xmath208 , and diagram i@xmath209 is negligible . for the case of @xmath210
both diagrams contribute at the same order , and there is only one intermediate atomic state .
+ ii.feynman diagrams of the subdominant processes contributing to @xmath211 .
virtual photons of momentum @xmath111 and polarization @xmath58 fly between the electron and the nucleus , yielding an effective wavelength cut - off of the order of the interparticle distance , @xmath212.,title="fig:",width=340,height=366 ] @xmath213 for the above integral not to vanish , the product of the quantum operators in the integrand must be even in @xmath111 .
we observe that when moving the complex exponential on the l.h.s . to the r.h.s . of the fraction ,
respecting the canonical commutation relations the momentum @xmath214 in @xmath78 gets shifted , @xmath215 , @xmath216^{-1}.\nonumber\end{aligned}\ ] ] as a result , the recoil kinetic energy in the denominators of eq.([pperpo ] ) becomes @xmath217 and an additional term @xmath218 shows up there . because the internal velocity of the electron is generally much greater than the center of mass velocity , @xmath219 , the term @xmath220 will be a higher order correction .
it is easy to see that the contribution of all the terms linear in @xmath83 is of the order of @xmath221 @xcite , which we interpret as a radiative correction to the total rest mass , @xmath222 .
in contrast , the term @xmath218 generates a doppler shift on the photon frequency which depends on the relative direction between @xmath111 and the current of the chomophoric electron , @xmath223 .
it is the break down of time reversal and mirror symmetry that makes @xmath214 take a non - vanishing transient value we mean the quantum amplitude of @xmath214 evaluated between the ground state and an intermediate state , say @xmath224 , @xmath225 .
the quantum amplitude @xmath226 in the expression for @xmath227 must be intended this way .
] in the direction parallel to the external magnetic field .
it is this term which generates the spectral non - reciprocity for photons propagating in opposite directions , @xmath228 , needed for the transfer of ( transverse ) linear momentum from the vacuum field to matter .
+ the denominator of eq.([pperpo ] ) must be expanded up to first order in the doppler shift term . at this order @xmath229
is the non - relativistic doppler shift , which does not clash with our non - relativistic approach @xcite .
this yields an even factor of @xmath111 in the integrand , necessary for the integral not to vanish .
the term @xmath230 generates a vanishing contribution . disregarding the mass renormalization terms , we find , @xmath231 at the same time , since both the chiral interaction @xmath46 and the zeeman potential @xmath76 featuring in @xmath78 are treated as perturbations to the harmonic oscillator , the denominator is to be expanded up to order @xmath232 and only terms at @xmath233b@xmath234 must be retained .
the calculation includes a number of finite integrals of the form , @xmath235\nonumber\\ & \simeq\frac{2m^{2}_{e}}{\hbar^{4}(e_{1}-e_{2})}\bigl[\log{(e_{1}/e_{2})}+\mathcal{o}(\hbar\omega_{0}/m_{e}c^{2})\nonumber\\&+\mathcal{o}[(\hbar\omega_{0}/m_{e}c^{2})^{2}]+ ... \bigr ] , \label{integral}\end{aligned}\ ] ] where @xmath236 and @xmath237 are the energies of the electronic transitions between intermediate states , which are of the order of @xmath238 .
additional integrals involving the product of three and four fractions appear also in eq.([pperpotro ] ) which yield terms of the same orders .
the final expression can be greatly simplified assuming small anisotropy factors , @xmath239 . averaging over the molecule s orientations ( @xmath240 ) @xcite and expanding the result up to leading order in the anisotropy factors we obtain , @xmath241\nonumber\\ & \times\frac{ce^{3}\mathbf{b}_{0}}{93312\pi^{2}c\epsilon_{0}m_{e}^{2}\omega_{x}\omega_{y}\omega_{z } } \eta^{zy}\eta^{xz}\eta^{yx}\nonumber\\&\simeq \frac{-1.06ce^{3}\mathbf{b}_{0}}{144\pi^{2}c\epsilon_{0}m_{e}^{2}\omega_{x}\omega_{y}\omega_{z } } \eta^{zy}\eta^{xz}\eta^{yx}.\label{pperptot}\end{aligned}\ ] ]
it is clear that @xmath242 is the expression to be compared with the semiclassical result of eq.([pcascalss ] ) , since from eq.([ec2 ] ) we read that the semiclassical @xmath153 is the momentum of the radiative ( i.e. , transverse ) modes only .
first we observe that @xmath242 is a factor @xmath243 times larger than the semiclassical @xmath153 . from eq.([integral ] ) we have that the semiclassical result is included in our quantum calculation ( there are no other terms of the same order in the diagrams neglected ) , but it enters at a second order correction , @xmath244 $ ] , to the leading term .
we conclude that the reason for this discrepancy is the failure of the semiclassical approach to account for the doppler effect due to the relative motion of the internal charges in order to generate the necessary spectral non - reciprocity .
instead , spectral non - reciprocity comes in the semiclassical approach from the effective magnetochiral birefringence .
we also note that the result of eq.([pperptot ] ) assumes an integration over @xmath245 up to infinity in eq.([pperpotro ] ) , which violates the non - relativistic limit .
however , had we assumed a cut - off for @xmath246 of the order of the inverse of the compton electronic wavelength , @xmath247 , the expression on the r.h.s .
of eq.([pperptot ] ) would change by only a factor of order unity . for the computation of @xmath248 the details of the calculation
have been already published in refs.@xcite and the supplemental material there . here
we concentrate on the underlying mechanism which originates the transfer of momentum from the vacuum field to the molecule .
the calculation at @xmath233b@xmath234 implies up to third - order perturbation theory in @xmath186 to the ground state of the harmonic oscillator hamiltonian , @xmath187 , @xmath249 where @xmath191 and @xmath250 .
the @xmath83 dependent terms are shown in ref.@xcite to give rise to mass renormalization factors .
as for the rest we have , @xmath251-[m_{e}\rightarrow m_{n}],\end{aligned}\ ] ] where @xmath252 $ ] means that the same expression within the square brackets must be evaluated exchanging @xmath30 with @xmath253 and subtracted .
in contrast to @xmath206 , no doppler shift term enters the calculation and no spectral non - reciprocity exists .
nonetheless the longitudinal momentum is also generated by the non - vanishing transient currents parallel to the external magnetic field . in this case , however , the currents are due to the internal motion of both the chromophoric electron and the nucleus , @xmath254 .
they are the sources of the magnetic fields at some point @xmath255 measured w.r.t .
the center of mass , @xmath256 each of these magnetic fields , combined with the coulomb field ( @xmath257 ) of the respective charges , @xmath258 for the nucleus and @xmath259 for the electron , @xmath260 give rise to the longitudinal casimir momentum of eq.([plong ] ) , @xmath261 in order to obtain an equivalence between equations ( [ plong ] ) and ( [ plongy ] ) , one of the @xmath245 factors in the fourier transform of the magnetic fields in eq.([bfield ] ) must be substituted by @xmath262 , where the momentum shift comes from the sum over the intermediate states which enter the transient currents .
it is clear that the only difference between the two terms of eqs.([plong],[plongy ] ) is in the sign of the charges which enter the electrostatic fields and the masses which determine their respective currents . as a result
, the addition of both terms must be proportional to @xmath263 , with @xmath264 a real functional of the masses .
the result is , @xmath265 where @xmath266 is the three - dimensional levi - civita tensor , the indices @xmath92 take on the axis labels @xmath93 and @xmath267 is a unitary vector along the @xmath268 axis . after averaging over the molecule orientations we end up with
, @xmath269 as for the transverse casimir momentum , the above calculation holds if we take the upper limit of integration of eq.([plong ] ) at infinity .
again , for a cut - off of the order of @xmath247 , additional terms of the same order as those in eq.([pkinparx ] ) would have been obtained . finally , adding up the transverse and the longitudinal contributions we obtain , @xmath270}{144\pi^{2}c\epsilon_{0}\mu\mu^{*}\omega_{x}\omega_{y}\omega_{z } } \eta^{zy}\eta^{xz}\eta^{yx}.\ ] ] this formula is a simple expression for @xmath271 in terms of the chiral parameter @xmath48 , the magnetic field and the natural frequencies of the oscillator .
we can write it also in terms of the fine structure constant , @xmath272 , the static optical rotatory power , @xmath273 , and the static electric polarizability of the molecule , @xmath274 . by comparing the above expression with the formulas for @xmath117 and @xmath115 in [ appendb ] we obtain , @xmath275e\mathbf{b}_{0}.\ ] ] in ref.@xcite
we speculate that , apart from constants of order unity , this expression is model - independent . for a given set of natural frequencies ,
all the atomic lengths in the problem are determined by quantum mechanics .
in particular , @xmath276 is a length , a fraction of the electronic compton wavelength , that we identify with a chiral length , @xmath277
. therefore , we can write @xmath278b@xmath279 and thus interpret it as the leading qed correction to the classical abraham momentum .
in the precedent sections we have found that the chiral molecule acquires a kinetic momentum during the switching of the external magnetic field as a result of its interaction with the vacuum field .
on the other hand we have shown that , as a result of the conservation of the total momentum @xmath162 , there exists a transfer of linear momentum from the vacuum field to the molecule .
the question arises whether the resultant variation of kinetic energy is provided by the vacuum field or by an external source . in the following
we prove that this variation of kinetic energy is part of the energy provided or removed by the source of the external magnetic field to generate a vacuum correction to the molecular magnetization . that magnetization energy can be considered as part of the magnetic lamb energy. + to this aim we make use of the hellmann - feynman - pauli theorem @xcite . according to this theorem
, the variation of the total energy of a system in its ground state with respect to an adiabatic parameter , @xmath280 , can be computed from the expectation value of the functional derivative of the hamiltonian w.r.t .
that parameter .
the variation of the total energy is the work done over the system of interest by an external source responsible for the adiabatic variation of the parameter .
if the parameter enters an interaction potential , @xmath281 , that work reads @xmath282 for instance , this approach is used to compute intermolecular forces by considering their center of mass position vectors as adiabatic parameters @xcite .
+ in our case , the adiabatic parameter to be varied is the magnetic field , @xmath4 , and the variation of energy to be computed is a magnetic energy , @xmath283 . since @xmath4 enters @xmath76 and @xmath51 linearly , and @xmath77 quadratically in eq.([effective ] ) , in @xmath77 , @xmath284 , does not contribute to @xmath283 . ]
the variation of the total energy of the system molecule - em vacuum during the adiabatic switching of the magnetic field from @xmath285 up to its final value @xmath4 is , @xmath286,\label{work}\end{aligned}\ ] ] where in the last equality @xmath287 $ ] includes only the @xmath288-dependent terms of @xmath289 .
we note that @xmath283 is the work done by the external source that generates the uniform magnetic field @xmath4 in the space occupied by the chiral molecule in the presence of the quantum vacuum .
the expression within the square brackets in eq.([work ] ) is indeed the total magnetization of the molecule in its ground state , @xmath290 , with a minus sign in front .
hence we can write , @xmath291 .
+ up to our level of approximation @xmath283 must include terms of up to order two in @xmath48 , @xmath4 and @xmath272 .
it follows that perturbation theory has to be applied up to order six in @xmath292 .
we first observe that at leading order in @xmath77 and quadratic in @xmath76 , and at zero order in @xmath46 and @xmath289 , @xmath283 reduces to the magnetization energy of the diamagnetic molecule in the absence of vacuum field , @xmath293 , @xmath294 being the static magnetic polarizability given in [ appendb ] . at higher order in perturbation theory , including the chiral potential and the vacuum field , we find a number of additional terms in @xmath283 .
most of these terms amount to variations in the energy of internal atomic levels , which are not relevant to us .
we are however interested in those terms of @xmath283 which are of the order of the variation of the kinetic energy of the chiral group , @xmath295^{2}/2m-\mathbf{q}^{2}_{0}/2m$ ] , for an initial kinetic momentum @xmath83 at zero magnetic field . in the following we
search for these terms by inspection and show that altogether they add up to give @xmath296 . as a result
, we can say that the variation of the kinetic energy is supplied by the source of the external magnetic field to produce a vacuum correction in the magnetization of the chiral molecule .
+ first we note that keeping finite the momentum of the molecule , @xmath297 , and bearing in mind that during the adiabatic switching of the magnetic field @xmath297 equals @xmath298 , @xmath283 contains terms quadratic in @xmath48 , @xmath4 and @xmath272 from the application of perturbation theory to @xmath299 up to order one in @xmath46 and @xmath76 , and up to order two in @xmath289 , as for the computation of the casimir momentum . following the same steps as for the calculation of @xmath211 , we start with the dressed ground state , @xmath300 , and apply to it second order perturbation theory with the interaction potential @xmath289 .
the calculation is therefore similar to that of the ordinary lamb shift except for the fact that the ground state and the intermediate states in our case are dressed up to first order in @xmath48 and @xmath288 .
in addition to the diagram of fig.[fig1]i(@xmath203 ) for the self - energy of the electron we must add the one of fig.[fig1]i(@xmath202 ) for the nucleus . we will denote the resultant energy by @xmath301 and refer to it as magnetic lamb energy , @xmath302 let us reorganize first the terms in @xmath289 as @xmath303 this shows that the longitudinal casimir momentum appears naturally coupled to the canonical momentum of the center of mass . in the following
we disregard the terms of @xmath304 associated to internal energies and restrict ourselves to those of the order of @xmath305 . in the first place , from the combination of the factors @xmath306 $ ] and @xmath307 in @xmath308 we have , @xmath309\nonumber\\ & -[m_{e}\rightarrow m_{n}].\end{aligned}\ ] ]
since the ground state @xmath310 is an eigenstate of the center of mass canonical momentum , @xmath311 , the above equation can be written as @xmath312\nonumber\\ & -[m_{e}\rightarrow m_{n}].\end{aligned}\ ] ] the expression under the sum is exactly the one for the longitudinal casimir momentum in eq.([plong ] ) with a minus sign in front .
therefore we find , @xmath313 this relation explains the subscript @xmath314 in @xmath315 .
+ likewise , from the combination of two factors @xmath316 in @xmath308 we have , @xmath317 next we note in the denominator a doppler shift term due to the motion of the center of mass , @xmath318 . expanding the fraction up to first order in this term and discarding the zero - order terms
we find , @xmath319 lastly , by expanding the denominator up to first order in @xmath218 and comparing with eq.([pperpotro ] ) we end up with @xmath320 we have found in sec.[qapproach ] that @xmath321 in stationary conditions .
therefore , combining the results of equations ( [ eparal ] ) and ( [ eperp ] ) and considering eq.([work ] ) we can write the magnetic lamb energy associated to these two terms as , @xmath322 which equals @xmath295^{2}/2m-\mathbf{q}^{2}_{0}/2m$ ] as anticipated . from this result
we see that the variation of the kinetic energy of the chiral group , induced by the transfer of linear momentum from the vacuum to the molecule , has its origin in the magnetic lamb energy ( i.e. , that part of the interaction energy that is induced by both the magnetic field and the quantum vacuum ) .
this kinetic energy is the magnetic energy associated with the vacuum correction to the magnetization of the molecule , @xmath323 , @xmath324 .
is a shortcut to write the vacuum correction to the magnetization , and so is the expression in the integrand of eq.([work ] ) for writing the total magnetization . strictly speaking it
should be written @xmath325 . ]
the sign of this energy depends on the magnitudes of the initial momentum and final magnetic field , @xmath326 and @xmath327 , as well as on their relative orientation .
a positive sign means that the energy is provided by the external source which generates the magnetic field .
a negative sign means that part of the initial kinetic energy of the molecule is removed by the external source during the magnetization process .
we have derived an expression for the casimir momentum transferred from the em vacuum to a chiral molecule during the switching of an external magnetic field .
we have modeled the chiral molecule using the single quantum oscillator model of condon _ et al .
we have applied both a semiclassical approach and a fully quantum non - relativistic approach .
+ the quantum approach reveals that two distinct mechanisms operate in the transfer of momentum , they both based on the production of transient internal currents in the direction parallel to the external magnetic field .
the transverse casimir momentum [ eq.([pperpotro ] ) ] has its origin in the spectral non - reciprocity generated by a doppler shift in the frequency of the vacuum photons .
this shift is due to the internal momentum of the chromophoric electron , @xmath214 .
the break down of time reversal and mirror symmetry makes @xmath214 take a non - vanishing transient value parallel to the external magnetic field . on the contrary
, the longitudinal casimir momentum [ eq.([plong ] ) ] has its origin in the combination of the electrostatic field of the internal charges and the magnetic field generated by their transient currents parallel to the external magnetic field [ eq.([plongy ] ) ] .
therefore , we have found that the longitudinal casimir momentum is the momentum of the source em field while the transverse momentum is the momentum of the sourceless vacuum field . in particular
, the finding @xmath328 proves that by modifying the symmetries of the space - time it is possible to vary the vacuum expectation value of observable quantities .
+ the quantum result in eq.([main1 ] ) is linear in @xmath4 and proportional to the fine structure constant , @xmath272 , and to the molecular rotatory power .
it is conjectured that this result is universal and model - independent , up to numerical prefactors of order unity which might include relativistic corrections .
+ the semiclassical approach only applies to the calculation of the transverse casimir momentum and relies on the non - reciprocity of the spectrum of effective normal modes .
however , it fails to incorporate the net effect of the doppler shift which comes in into the microscopical quantum approach and which is seen to dominate largely .
the semiclassical result of eq.([pcascalss ] ) appears as a second order correction to the leading order term of the quantum result .
+ we have proved that the variation of the kinetic energy of the chromophoric group has its origin in the magnetic lamb energy [ eq.([ener ] ) ] .
this kinetic energy is found to be part of the magnetic energy associated with the vacuum correction to the magnetization of the molecule .
when positive , this energy is provided by the external source which generates the uniform magnetic field . when negative , it is part of the initial kinetic energy of the molecule which is removed by the external source during the magnetization process .
+ this work was supported by the anr contract photonimpuls anr-09-blan-0088 - 01 .
at leading order in the anisotropy factors , the polarizabilities that enter eq.([dind ] ) are @xcite , @xmath329 where @xmath330 and @xmath331 are dimensionless functions of the anisotropy factors , @xmath332 , @xmath333 .
100 donaire m , van tiggelen b a and rikken g l j a 2013 _ phys .
lett . _ * 111 * 143602 .
casimir h b g 1948 _ kon .
proc . _ * 51 * 793 .
lamoreaux s k 1997 _ phys .
lett . _ * 78 * 5 .
milonni p w 1994 _ the quantum vacuum _
( academic press , san diego ) .
grisenti r e _ et al .
_ 1999 _ phys .
lett . _ * 83 * 1755 .
_ 2007 _ europhys .
lett . _ * 77 * 54001 .
bressi g , carugno g , onofrio r and ruoso g 2002 _ phys .
lett . _ * 88 * 041804 .
munday j n , capasso f and parsegian v a 2009 _ nature _ * 457 * 170 .
barron l d and vrbancich j 1984 _ mol .
phys . _ * 51 * 715 .
van tiggelen b a 2008 _ eur .
j. _ d**47 * * 261 .
wu c s , ambler e , hayward r w , hoppes d d and hudson r p 1957 _ phys .
rev . _ * 104 * 254 .
alavi - harati a _ et al .
_ 1999 _ phys .
_ * 83 * 22 .
peccei r d and quinn h r 1977 _ phys .
lett . _ * 38 * 1440 .
maiani l , petronzio r and zavattini e 1986 _ phys .
_ b**175 * * 359 .
bakalov d _
_ 1998 _ hyperfine interact .
_ * 114 * 103 .
klein j j and nigam b p 1964 _ phys .
_ * 135 * b1279 .
rikken g l j a and rizzo c 2000 _ phys .
_ a**63 * * 012107 .
rikken g l j a and van tiggelen b a 2011 _ phys .
lett . _ * 107 * 170401 .
rikken g l j a and van tiggelen b a 2012 _ phys .
lett . _ * 108 * 230402 .
nelson d f 1991 _ phys .
_ a**44 * * 3985 .
feigel a 2004 _ phys .
lett . _ * 92 * 020404 .
van tiggelen b a , kawka s and rikken g l j a 2012 _ eur .
j. _ d**66 * * 272 .
van tiggelen b a , rikken g l j a and krsti v 2006 _ phys .
lett . _ * 96 * 130402 ; 2005 _ the feigel process : lorentz - invariance , regularization , and experimental feasability _ , report no .
ao4532/18806/05/nl / mv , advanced concepts team , european space agency , madrid , spain .
croze o a 2012 _ proc .
_ a**468 * * 429 .
kawka s and van tiggelen b a 2010 _ europhys .
lett . _ * 89 * 11002 .
condon e u 1937 _ rev .
phys . _ * 9 * 432 .
condon e u , altar w and eyring h 1937 _ j. chem .
* 5 * 753 .
herold h , ruder h and wunner g 1981 _ j. phys .
_ b**14 * * 751 .
dippel o , schmelcher p and cederbaum l s 1994 _ phys .
_ a**49 * * 4415 .
babington j and van tiggelen b a 2011 _ eur .
_ * 93 * 41002 .
buhmann s y , butcher d t and scheel s 2012 _ new j. phys . _
* 14 * 083034 .
milonni p w 2007 _ phys .
_ * 76 * c167 .
craig d p and thirunamachandran t 1998 _ molecular quantum electrodynamics _
( dover ed .
, new york ) .
craig d p and thirunamachandran t 2002 _ j. mol .
* 591 * 19 - 25 .
donaire m , rikken g
l j a and van tiggelen b a 2014 _ eur .
_ d**68**:33 .
cohen - tannoudji c , dupont - roc j and grynberg g 2004 _ photons and atoms .
introduction to quantum electrodynamics _
( wiley - vch verlag gmbh @xmath334 co. kgaa , weinheim ) @xmath335 i.b.5 .
loudon r 1983 _ the quantum theory of light _
( oxford university press , second edition ) .
wilkens m 1993 _ phys .
a * 47 * 671 ; 1994 _ phys .
rev . _ a * 48 * 570 . pines d and nozieres p 1966 _ the theory of quantum liquids _
( new york : benjamin ) pp .
295 - 297 .
zhang j z h 1999 _ theory and applications of quantum molecular dynamics _
( world scientific , singapore ) @xmath335 1.3 . | + in a recent publication @xcite we have shown using a qed approach that , in the presence of a magnetic field , the quantum vacuum coupled to a chiral molecule provides a kinetic momentum directed along the magnetic field . here
we explain the physical mechanisms which operate in the transfer of momentum from the vacuum to the molecule .
we show that the variation of the molecular kinetic energy originates from the magnetic energy associated with the vacuum correction to the magnetization of the molecule .
we carry out a semiclassical calculation of the vacuum momentum and compare the result with the qed calculation . |
during a first - order phase transition , the system has to overcome a free energy barrier in order to transform to the thermodynamically stable phase .
when this process is driven only by thermal fluctuations , it is called _ homogeneous nucleation_. more often , nucleation is facilitated by external surfaces and the corresponding process is known as _
heterogeneous nucleation_. nucleation is an important process for its many practical applications , as it is ubiquitous in various types of technologically important materials such as metals , semiconductors , pharmaceuticals , and foods .
nucleation has been thoroughly studied ( for recent reviews see for example refs .
@xcite ) , but it is still not entirely understood , due to the experimental difficulty in accessing the small size of the crystalline nuclei involved in the transition ( which is of the order of 10 - 1000 molecules in usual conditions ) .
the difficulty in observing the nucleation process at the microscopic level has profound repercussions not only on our theoretical understanding , but also on the practical task of estimating nucleation rates , where considerable disagreements between experimental measurements and numerical predictions are often reported even for the simplest systems .
one notable example is the condensation of argon , a noble gas , where reports of theoretical and experimentally measured nucleation rates can differ even by 26 orders of magnitude @xcite . a molecular system whose crystallization represents the most important physical transformation on earth is water .
understanding nucleation of water is of fundamental importance in many industrial and natural processes .
for example , the nucleation of ice nuclei in atmospheric clouds is an important factor in establishing earth s radiation budget .
our lack of understanding of the freezing mechanisms of water in clouds has a negative impact on many climate models : ice crystals control the amount of light reflected by clouds , which is one of the largest unknown parameters that affect atmospheric warming @xcite differently from atomic and molecular systems , colloidal systems are characterized by lengthscales ( of the order of @xmath0 @xmath1 m ) and timescales ( of the order of seconds ) that can be accessed at a single particle level with microscopy techniques .
moreover , the interactions between colloids are mediated by fast degrees of freedom ( like the solvent ) ; these effective interactions @xcite can be controlled to a high degree by changing a few external parameters , like the quality and concentration of the dissolved salts .
these considerations give to colloidal suspensions the quality of ideal systems , where our theoretical understanding of crystallization can be benchmarked against accurate experimental results .
for a recent review of nucleation in colloidal systems see ref .
it is surprising to observe that even for the simplest colloidal system , i.e. colloidal hard spheres @xcite , discrepancies between the observed and predicted nucleation rates can be as big as ten orders of magnitude @xcite . while numerical simulations have found a rapid increase of the nucleation rate with increasing colloid volume fraction @xmath2 , growing by more than 15 orders of magnitude from @xmath3 to @xmath4 @xcite , experiments on both polymethyl methacrylate and polystyrene microgel systems , instead report nucleation rates that are much less sensitive to volume fraction @xcite .
the simplest and most general understanding of nucleation is embodied in classical nucleation theory ( cnt ) @xcite .
focusing here on the liquid - to - solid transition , the process starts from small crystalline nuclei that spontaneously form in the supercooled liquid from thermal fluctuations . at any temperature below the melting temperature
, there is a characteristic size at which these nuclei are equally likely to dissolve or to grow until the liquid - to - solid transition is complete .
this size is called _ critical nucleus size_. classical nucleation theory rests on the assumption that the critical nucleus is amenable to a thermodynamic description , and that the process is effectively one - dimensional , meaning that it can be described by just one reaction coordinate , with all order parameters involved in the transition proceeding simultaneously .
the idea of describing crystal nucleation as a one - dimensional process is borrowed from first - order phase transitions between phases with the same symmetry , most notably the gas - liquid nucleation , or its reverse process , cavitation . but the liquid - to - solid phase transition is different , as both the translational and orientational symmetries of the liquid are broken with crystallization .
this is seen by expanding the density of the crystal in its fourier components @xcite : @xmath5 where @xmath6 is the density of the liquid phase , @xmath7 is the fractional density change at freezing , and @xmath8 are the fourier components associated with the crystalline periodic structure ( @xmath9 for the liquid phase ) .
the liquid - to - solid transition is thus associated with an infinite number of order parameters ( @xmath10 , @xmath8 ) . by assuming the crystal to be harmonic at melting
, all high order order parameters can be related to @xmath11 . in this effective case ,
the transition is thus represented by two order parameters , @xmath10 and @xmath11 , which are associated with translational and orientational order respectively .
we note , however , that the latter still includes translational order due to the constraint on the wave vector , @xmath12 , which is a feature of the density functional theory of crystallization .
the fact that nucleation is inherently multi - dimensional is obvious in all those systems where the breaking of positional and orientational order occurs at different temperatures , the most famous example being the freezing of two dimensional hard disks . here , between the liquid and solid phases
, an intermediate phase appears , called the hexatic phase , characterized by quasi - long range orientational order , but short - range positional order . in one of the most important cases of algorithmic development in recent years @xcite
, it was shown that the transition between the liquid and hexatic phase is first - order , while the hexatic - to - solid is a second order transition . but a decoupling of translational and orientational order can occur even when both symmetries break down simultaneously as in the case of crystallization in three dimensions .
note that translational ordering automatically accompanies orientational ordering .
in particular , in the next sections we show examples where two - step ordering can be identified prior to the transition : in the first step crystallization involves an increase of orientational order even in the liquid state , only later followed by a change of density ( positional order ) . this decoupling can be intuitively understood on the basis that liquids are difficult to compress , and some order has to appear for realizing efficient packing before density can increase . in this article
, we will consider several examples of two - step crystallization .
the term `` two - step '' was originally applied to crystallization processes which involved an intermediate step corresponding to metastable phases where nucleation occurs at a higher rate .
it has then been extended to describe crystallization processes that are continuous , but where nucleation occurs in preordered regions of the melt
. we will start with considering how a metastable gas - liquid phase separation alters the nucleation pathway by adding an intermediate step , i.e. the formation of a dense liquid droplet , whose high density considerably increases the crystallization rate .
we will then consider examples where the nucleation pathway occurs through crystallization precursors spontaneously formed in melts . in this case
we show that the liquid - to - solid transition is effectively driven by structural fluctuations , and the nuclei reach their bulk density only when their size considerably exceeds the critical nucleus size . in the case of spherically symmetric interaction potentials , we show that crystallization occurs in regions of high crystal - like bond - orientational order , and that these regions effectively select the polymorph that will be nucleated from them .
connection are made also with vitrification phenomena for hard spheres , in which we associate crystalline orientational order with slow regions in the system , and identify five - fold symmetric regions as inhibitors of crystallization .
it is worth stressing that such bond orientational ordering does not accompany the density change .
we will then focus on the special case of water , where the presence of directional hydrogen bonds considerably alters the pathway to nucleation .
all these indicate that a liquid is not a homogeneous isotropic state at the microscopic level , but generally tends to have more local structural order to lower its free energy .
this tendency , which increases for a lower temperature , has an impact on the structural and dynamical properties of a supercooled liquid state , which further influences crystal nucleation and vitrification @xcite . in this article
, we consider only a ( nearly ) single - component system , where density and bond orientational order are two important order parameters to describe the liquid state .
however , for a multi - component system , we need an additional order parameter , local composition , to describe it . here
we do not discuss this important but difficult case .
we are going to show that the multiple - order - parameter description is crucial to understand the low - temperature phenomena seen in a supercooled liquid state .
the simplest and most general understanding of nucleation is embodied in classical nucleation theory ( cnt ) @xcite . focusing here on the liquid - to - solid transition , the process starts from small crystalline nuclei that spontaneously form in the supercooled liquid from thermal fluctuations . a critical nucleus size , at which these nuclei are equally likely to dissolve or to grow , can be easily obtained from the gibbs - thomson equation .
the chemical potential of a solid particle inside a nucleus of size @xmath13 is given by @xmath14 where @xmath15 is the bulk chemical potential of the solid phase , @xmath16 is the interfacial tension , and @xmath17 is the bulk density of the solid . at the critical nucleus size @xmath18 , the crystalline nucleus is in equilibrium with the liquid phase , @xmath19 , from which we obtain @xmath20 below the critical nucleus size
, the clusters made of @xmath21 particles are thermally equilibrated at all times , meaning that they are evolving in a free energy landscape where the nucleation rate density can be written as @xmath22 where @xmath23 is a kinetic factor , and @xmath24 is the height of the free energy barrier separating the liquid from the solid phase .
here @xmath25 is the number of particles in the critical nucleus . according to cnt
, the kinetic term takes the form @xmath26 where @xmath27 is the density , @xmath28 is the molecular mass , @xmath29 is the attachment rate of particles to the critical nucleus , and @xmath30 is the curvature of the barrier at the critical nucleus size @xmath25 .
the term @xmath31 is also known as the zeldovich factor .
the free energy barrier @xmath32 is instead derived from a competition between a bulk term which accounts for the free energy gain of transforming to the stable phase , and a surface term which expresses the free energy penalty in creating a liquid / solid interface @xmath33 where @xmath34 is a numerical factor that depends on the shape of the nucleus ( @xmath35 for a spherical nucleus ) , @xmath36 is the liquid - solid chemical potential difference , and @xmath16 is the liquid - crystal surface tension . here
it is worth mentioning that the transport crucial for crystal nucleation and growth is not controlled by viscosity , but by translational diffusion ( see , e.g. , ref .
this is important below the melting point , where the stokes - einstein relation rather significantly breaks down . in the case of heterogeneous nucleation
, the free energy barrier is decreased by a geometrical factor @xmath37 where @xmath38 is the contact angle of the crystal nucleus with the external surface .
classical nucleation theory provides a successful framework to understand and analyze data from nucleation for a large variety of processes .
but , on quantitative level , its predictions are hard to test , as small changes in the experimental conditions can significantly impact the yield rates of crystallization . in the introduction section
, we highlighted the large discrepancies in the nucleation rates obtained experimentally and theoretically for arguably the simplest system that exhibits a liquid - to - solid transition , i.e. colloidal hard spheres .
several attempts have been made at solving this inconsistency , but no consensus has been found yet .
approximations in the calculations of nucleation rates can be found both for experiments and simulations , and explanations are being proposed which take into account the limitations of both methods .
in one recent proposal @xcite , the effects of the gravitational fields on the experimental nucleation rates are considered .
the difficulty of exactly density matching colloidal samples implies that gravitational lengths are not negligibly short compared to the size of colloidal systems .
the coupling between the gravitational force with long range hydrodynamic effects can then cause an increase of density fluctuations in the sample , followed by an increase of nucleation rates compared to the gravity - free case .
a different proposal @xcite instead focuses directly on hydrodynamic interactions , which are often neglected in simulations .
it is argued that hydrodynamic interactions can effectively increase nucleation rates to account for the discrepancy between experiments and simulations . to this day ,
consensus on the resolution of the discrepancy in nucleation rates has not been found .
classical nucleation theory rests on several approximations .
the most important one is the so - called _ capillarity approximation _ , which is the assumption that small crystalline nuclei ( or the order of 100 of particles ) are still amenable to a thermodynamic description , as they retain the same properties of the bulk solid .
this assumption has been challenged in many studies @xcite . , and the average density at the center of mass is plotted against the radius @xmath13 of the nucleus . @xmath39 and @xmath17 are the equilibrium densities of the fluid and of the crystal respectively , while @xmath18 is the critical nucleus radius.,width=302 ] as an example we consider here a hard sphere system at reduced pressure @xmath40 , where @xmath41 , and @xmath42 is the diameter of the spheres ( see ref .
@xcite ) . at these conditions
the critical nucleus size consists approximately of @xmath43 spheres , and direct crystallization events can be observed in simulations . by gathering statistics on many independent crystallization events , in fig .
[ fig : density_nucleus ] we plot the average density at the center of mass of the nucleus as a function if its radius @xmath13 .
density is normalized with the density of the fluid @xmath6 and of the solid @xmath17 at their equilibrium values , so that it is @xmath44 in the bulk @xmath45 solid , and @xmath46 in the metastable liquid phase .
the radius is instead scaled with the critical radius @xmath18 . from the figure
we can see that at equilibrium , @xmath47 , the density at the center of the nucleus is far below the equilibrium density of the crystal @xmath48 .
another parameter that is not well described by the capillarity approximation is the surface tension @xmath16 .
several studies have found that the surface free energy cost at equilibrium is not the relevant free energy penalty for nucleation , both in metals and charged colloids @xcite . among the many theoretical treatments that go beyond the capillary approximation
see for example ref .
@xcite : these include modifications of cnt ( as for example introducing a lenghtscale - dependent surface tension , i.e. tolman length ) @xcite , and new applications like density functional theory @xcite and kinetic nucleation theory @xcite .
it is worth noting that in most descriptions of nucleation , the melt is described as structure - less and homogeneous .
several studies have shown instead that the supercooled state contains a lot of structural order aside from crystalline order @xcite .
for hard spheres , for example , the competition between crystalline order and icosahedral ( or five - fold symmetric ) order plays a big role in the nucleation process , and on its avoidance @xcite . for systems with a metastable critical point ,
e.g. with short - range attractions , spatial heterogeneity due to critical fluctuations also plays a big role @xcite .
cnt rests on the assumption that the nucleation process can be described by just one reaction coordinate , and that all order parameters involved in the transition proceed simultaneously .
several works have addressed this assumption in systems where different reaction coordinates can be followed simultaneously during nucleation , and have demonstrated that a single - order parameter description fails at capturing important aspects of the transition @xcite . in the following we will focus on this particular aspect of the nucleation process , and show how new insights can be obtained with a multi - dimensional ( or , multi - order - parameter ) representation of the transition . in colloidal systems , unlike metals and noble gases , polydispersity plays a big role in the crystallization process . formally speaking
, a continuous polydispersity introduces a new order parameter for each specie in the mixture @xcite . at high polydispersities , crystallization requires fractionation , and generally destabilizes the crystal . at low polydispersities , it is possible to treat the system like an effective one - component system , but care has to be taken in describing the nucleation process , as several thermodynamic quantities are very sensitive to polydispersity , such as surface tension @xcite .
it is important to note that kinetic effects can have a profound impact on the crystallization process .
kinetic effects become especially important at high supercooling , where diffusion becomes the rate limiting process for both nucleation and crystal growth . according to classical nucleation theory , the kinetic term in eq .
( [ eqn : k ] ) is controlled by a thermally activated process @xcite , usually thermal diffusion ( at low temperatures diffusivity and viscosity can have different activation energies , i.e. stokes - einstein violation ) @xmath49 where @xmath50 is the atomic jump distance in the liquid . in the crystallization of a lennard - jones system @xcite , it was shown that the crystallization rate predicted by eq .
( [ eqn : k2 ] ) is two orders of magnitude smaller than the measured one .
this suggests that , for simple monoatomic systems , the kinetic factor is not controlled by thermally activated processes .
this behaviour is consistent with the growth of the lennard - jones crystal @xcite where the crystallization rates were found to be independent of diffusion , but determined by the ideal gas thermal velocity @xmath51 .
this means that , at high supercooling , crystal growth can proceed despite diffusive processes becoming negligible , even below the glass transition temperature . in a series of experiments @xcite on molecular liquids ( both organic and inorganic glass formers )
, it was discovered that the crystal growth rate can display an unusual enhancement below the glass transition temperature .
the growth rate becomes orders of magnitude faster than expected for diffusion - controlled growth , in conflict with the established view that the kinetic barrier for crystallization is similar to that for diffusion .
simulations of hard spheres have also shown crystallization without diffusion @xcite , that the static properties of the crystals are not much affected by dynamics @xcite , and that the dynamics in a crystallizing glass becomes intermittent ( avalanche like ) @xcite .
there have been several attempts at describing the process by which particles move from the liquid to the crystal without assuming an activated process analogous to that which governs self - diffusion .
these include the substitution of diffusion with secondary relaxations @xcite ( the so called @xmath52 processes ) , or by taking into account the extensional stress around crystals growing inside a glass of lower density @xcite . despite these attempts
, a clear explanation is still lacking .
for example , simulations of binary metallic glasses have instead confirmed diffusional growth of the interface @xcite .
these simulations have noticed a connection between the attachment rate of liquid particles to the crystal surface with the width of the surface , which suggests that the structural properties of the melt , and its preordering in advance of the front propagation , might play a bigger role than dynamics .
we will investigate the role of structural preordering of the fluid in detail in section [ sec : precursors ] . despite its limitations , cnt is still our best effective theory of the nucleation process . rather than focusing on the underlying symmetries ( translational and orientational order ) cnt uses the size of the nucleus as its reaction coordinate , and gives a unified description of nucleation in a big variety of systems .
but the simplifications that are adopted ( capillarity approximation , and reduction to a one - dimensional description ) hide a lot of the phenomenology of the nucleation process , and in the next sections we unveil some of the steps that have been made in understanding nucleation from a microscopic perspective .
one of the first demonstrations of a non - classical pathway to nucleation was given by ten wolde and frenkel @xcite .
they considered how the crystallization pathway is influenced by the presence of a metastable gas - liquid transition , whose order parameter is a scalar density field .
the model considered was an adapted lennard - jones potential , with a hard - repulsion at contact , and short - range attraction .
in soft matter systems , reducing the range of the attractions pushes the gas - liquid critical point to lower temperatures , where it eventually becomes metastable with respect to crystallization @xcite .
the parameters were chosen in order to have a critical point lying approximately 20% below the equilibrium crystallization curve , a condition that is found to be common in globular protein solutions . according to the study , a novel crystallization pathway was found in proximity of the gas - liquid critical point . instead of direct nucleation from the gas phase ( one - step crystallization ) ,
the system was found to crystallize inside disordered liquid regions of higher density that form due to critical density fluctuations .
this process is referred as two - step crystallization : the first step involves the formation of dense liquid regions , and the second step is the nucleation of the crystal phase inside these dense regions . in the original work @xcite the formation of dense precursors
was attributed to the long - range density fluctuations that originate from the presence of the metastable critical point .
it is thus expected that the nucleation rate would have maximum around the critical point .
a recent study @xcite has instead found that that the nucleation rate monotonically increases going from the supercritical region towards the two - phase coexistence region .
the two - step mechanism is thus not strictly due to critical fluctuations , but rather to the formation of a dense liquid phase that is thermodynamically stabilized below the critical point .
away from the metastable critical point , the system was found to crystallize classically in one step , where densification and structural ordering happen simultaneously .
thus it was demonstrated how a very simple model ( spheres with short range attractions ) can display very complex crystallization behavior depending on the thermodynamic parameters @xcite . in experiments , two - step pathways
have been first considered in protein crystallization @xcite , most notably lysozyme , but were quickly observed also in colloidal systems @xcite .
numerous examples of two - step crystallization from solutions has been reported in the last years , and several reviews have been written on the subject @xcite .
probably the best evidence for two - step nucleation pathways came from theoretical investigations , where microscopic information is readily available .
the main theoretical tool for the study of crystallization in simple model systems are density functional theory ( dft ) @xcite and molecular simulations .
two - step crystallization pathways have been suggested to occur also outside the region of stability of the dense fluid phase . in this case , crystallization starts inside metastable liquid drops ( also called pre - nucleation clusters @xcite ) whose lifetime has to be long enough to promote nucleation .
the first example came from experimental studies of lysozyme crystals , which were found to crystallize more easily just outside the metastable gas - liquid phase separation transition @xcite .
similar processes have been found in both small molecules , like glycine @xcite and calcium carbonate @xcite , and colloidal systems @xcite .
two - step crystallization in absence of a metastable liquid phase was also confirmed theoretically in spin models @xcite and density functional theory @xcite .
even the presence of metastable liquid phases might not be a necessary condition , as several studies have reported two - step nucleation pathways in systems where a dense fluid phase does not exist .
hard spheres , due to the lack of attractive interactions , are an example of such systems : the stable phases are a fluid phase and close - packed crystalline phases , such as hcp , fcc and the phase obtained by random stacking of hcp and fcc layers .
the large size of colloidal particles means that the lengthscales and timescales can be ( relatively ) easily observed experimentally , and many classical experiments have examined the process of crystal nucleation in microscopic detail @xcite . also simulations have studied in detail the composition , shape and free energy barrier to nucleation @xcite .
leveraging the sensibility of light scattering experiments to density fluctuations , dense amorphous precursors were reported in colloidal hard spheres @xcite , pmma particles @xcite , and microgel colloids @xcite .
these experiments were also corroborated by molecular dynamics simulations @xcite . whether dense precursors can be formed in systems without a metastable fluid - fluid demixing transition ( like hard spheres )
is not so clear physically and has been debated in the literature . in ref .
@xcite we argued that orientational order foreshadows translational order in hard - spheres , i.e. that the nucleation process is accompanied by a faster increase of bond - orientational order than translational order .
this also implies that precursor regions , defined as the regions of the melt from which the critical nucleus will emerge from , are better characterized by bond orientational order than density . in the next section
we will go through the different arguments in favor of this scenario , and in section [ sec : whosfirst ] we will comment briefly over the comparison with other scenarios .
and @xmath27 fluctuations in hard spheres at reduced pressure @xmath40 .
( a ) snapshot of particles with the @xmath53 highest value of @xmath54 .
red particles have higher density than the median density ; green particles have lower density than the median density .
( b ) snapshot of particles with the @xmath53 highest value of @xmath27 .
red particles have higher @xmath54 than the median @xmath54 ; green particles have lower @xmath54 than the median @xmath54 .
adapted from ref .
, width=321 ] in the previous section we focused on pathways where the metastability of a liquid phase induces the formation of dense drops that promotes crystallization
. we will show here that , in absence of a metastable gas - liquid transition , the opposite scenario is instead favored , with structural order fluctuations preceding the increase of density .
this can be intuitively understood on the basis that the melt is difficult to compress , and some structural order has to appear due to packing effects before density can increase . in order to study this process from a microscopic perspective
, one should follow the nucleation process from the nucleation event to the growth of the crystalline nucleus over its critical nucleus size . by using local order parameters , describing the degree of translational and orientational order around individual particles , it is then possible to construct a statistical map of the microscopic pathway to crystallization in a multiple - order - parameter space . while most of our knowledge comes from molecular simulations on model systems @xcite ,
recent confocal microscope experiments have also started to shed light on this process in colloidal systems @xcite .
translational order can be computed from two - body correlation functions , such as the pair distribution function or the two - body excess entropy @xcite , but it is found that all these measures correlate well with the local density ( as obtained from the voronoi diagram of particle configurations ) . it is thus common to use the local density as a measure of translational order . unlike translational order , which is obtained from two - body correlation functions ,
orientational order is obtained by considering many - body correlations . for spherical particles , an adequate measure of orientational order
was introduced by steinhardt et al .
@xcite , and later popularized by frenkel and co - workers in the study of crystallization @xcite .
the order parameter most often used for hard spheres is @xmath55 , where the subscript @xmath56 denotes the use of spherical harmonics of degree @xmath56 , and it is computed for each particle in the system by considering only the position of nearest neighbors ( @xmath57 for hard spheres ) .
lechner and dellago have also proposed to coarse - grain orientational order parameters , significantly reducing fluctuations by including the orientational order up to the second shell of nearest neighbors @xcite .
we use here capital letter to denote coarse grained versions of orientational order parameters : for example , @xmath54 is the coarse - grained version of @xmath55 . for the exact definitions of these order parameters we refer to ref .
@xcite . here
we note that a combination of more than two bond orientational orders is usually necessarily to specify a particular rotational symmetry . for details
, please refer to ref .
@xcite . in the following we focus on the hard - sphere system , as a model of fluids that crystallize in close - packed structures .
we compare the bond orientational field and the density field according to five different criteria . for the supercooled melt at @xmath40 .
the dashed white arrow is the steepest descent path from the maximum of the distribution .
adapted from ref .
@xcite.,width=302 ] correlation length of the fluctuations : : in a two - step scenario , one expects the crystalline nucleus to appear inside a region where the relevant order parameter is higher than the rest of the melt .
the size of this region should also increase with supercooling .
it is well known that the length - scale of density fluctuations barely changes with supercooling , while refs .
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) showed indeed that the correlation length of bond orientational order increases with supercooling .
for example , figs .
[ fig : fluctuations](a ) and ( b ) show that while density fluctuations are almost uncorrelated , bond orientational order is much more correlated .
direct observation of nucleation events have shown that the nuclei do appear to form inside regions of high bond orientational order @xcite .
lifetime of fluctuations : : in order to promote crystallization , the fluctuations need to be sustained enough for a nucleus to appear inside them and eventually reach the critical size .
the lifetime of density fluctuations is given by the decay of the intermediate scattering function , which is used as a definition of the structural relaxation time , @xmath58 .
several studies have shown that structural fluctuations can last much longer than @xmath58 ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this is particularly the case for two - dimensional liquids @xcite .
landau free energy : : the landau free energy in the two - dimensional space of bond orientational order and density @xcite is obtained by computing the joint probability of fluctuations in @xmath54 and @xmath27 , @xmath59 .
figure [ fig : map ] shows the result for hard spheres at reduced pressure @xmath40 , taken from ref .
@xcite , and displays a good decoupling between @xmath54 and @xmath27 , expressing the fact that @xmath54 and @xmath27 capture independently the fluctuations in bond orientational order and density @xcite .
a cubic fit to the free energy shows that the dominant cubic term is of the form @xmath60 ( see ref .
@xcite for a possible form of the free energy functional ) .
because the interaction is quadratic in @xmath27 and linear in @xmath54 , the system can increase its orientational order without increasing the translational order , but the contrary is not true .
this constrains the fluctuations towards a stronger increase in its orientational order .
the analysis also shows a weak linear coupling between @xmath27 and @xmath54 , which is visible in the small positive tilt of the free energy contour in fig .
[ fig : map ] .
this linear term also indicates that regions of high orientational order will , on average , have also higher density than the melt ( see fig .
[ fig : fluctuations](a ) ) .
but the opposite is not true : due to the stronger @xmath60 term in the free energy expansion , regions of high density are not , on average , characterized by higher @xmath54 ( see fig . [
fig : fluctuations](b ) ) . + in black circles ) at the center of a crystalline nucleus ( its center of mass ) as a function of the scaled nucleus radius , @xmath61 , where @xmath62 is the critical nucleus size.,width=321 ] radial profile of crystalline nuclei : : it is possible to compute the average radial profile for both @xmath27 and @xmath54 in crystal nuclei of a certain size .
the results show that , going from the fluid phase towards the center of the nucleus , the system first develops orientational order , only later followed by an increase of density @xcite .
similar results have been obtained with density functional theory @xcite . at the critical nucleus size ,
the center of the nucleus is closer to the @xmath54 of the bulk crystal , yet the density still being just @xmath63 of the bulk value for a critical nucleus of size between 80 - 90 particles . extracting the values of the order parameters at the center of the nucleus as a function of nucleus size , fig .
[ fig : radial ] shows that the density extrapolates to the fluid density at zero size , while @xmath54 seems to form at values higher than the bulk fluid .
this clearly suggests that indeed the nuclei appear from regions of high bond orientational order , but whose density is still comparable to the fluid s density .
this means that crystal nucleation takes place heterogeneously in space . + for crystalline particles ( symbols and solid line ) and non - crystalline particles ( dashed line ) in a metastable liquid state at @xmath64.,width=264 ] translational vs orientational order curves : : ref .
@xcite introduced the idea of plotting the average density for a metastable liquid state as a function of orientational order for different subsets of particles , as for example liquid and solid particles .
the figure is reproduced in fig .
[ fig : curves ] , showing that : i ) the solid branch shows a characteristic plateau in @xmath27 after which it becomes the stable branch at that state point ; ii ) the solid branch looses stability at volume fractions of approximately @xmath65 , thus close to the glass transition .
the presence of the distinct plateau in @xmath27 clearly shows that the transition occurs at constant density and involves instead an increase of bond orientational order .
@xcite showed that the liquid and solid branch start crossing approximately at the melting pressure , below which the two curves do not cross .
similar calculations have shown how polydispersity destabilizes the solid branch @xcite , and have been found to apply in colloidal experiments @xcite .
evidence of bond orientational foreshadowing of crystallization has been suggested for colloidal systems both in simulations @xcite and experiments @xcite .
interestingly , the role of structural fluctuations has been investigated also in a variety of different systems , like metallic melts @xcite , anisotropic particles @xcite , and polymers @xcite .
charged colloids are good model systems for metals and alloys , and a parallel between the two systems , in which the development of short - range order is highlighted , can be found in ref .
recent advances in x - ray spectroscopy @xcite hold the potential of computing bond orientational order from scattering experiments , and could open a new era in the investigation of local structural information in disordered systems .
the idea of studying crystallization under external fields by studying the effects of the field on the bond orientational order is also a new direction for future research , which has been considered for example in ref .
@xcite for the case of crystallization in sheared suspensions .
there has been some debate in the literature as to whether the nuclei appear in dense - precursors @xcite or in bond - orientational - ordered precursors @xcite .
a very recent @xcite re - analysis of the simulation trajectories of ref .
@xcite ( `` density - first '' case ) showed a simultaneous increase of density and bond - orientational order leading up to nucleation . as we observed in fig .
[ fig : map ] there is a ( weak ) linear coupling between bond - orientational order and density which ensures that fluctuations toward high bond - orientational order will on average lead to a higher density than the melt .
this also means , as found in ref .
@xcite , that following crystal nucleation in time will lead to a simultaneous increase of density and bond - orientational order , as was observed in ref .
@xcite .
this consideration may also have an implication on light scattering results of crystal nucleation @xcite .
light scattering measurements are basically sensitive to the development of translational order , but not to that of bond orientational order . thus , the nucleation process can only be followed as soon as the scattering from the [ 111 ] plane is observed , but still before the formation of any higher order peaks .
at these resolutions , the distance between [ 111 ] planes is observed to be considerably shorter than the nearest neighbour spacing in the remaining melt , which signals that densification is also occurring during the formation of 2d sheets of hexagonal order .
the fact that both order parameters are changing during the transition , does not mean that their role is the same .
positional order and orientational order are characterized by their spatial correlation functions .
it has been shown @xcite , in both mono- and poly - disperse systems , that with supercooling ( compression over the freezing pressure ) hard - spheres rapidly develop orientational order with negligible signs of increase in translational order : while the correlation length of positional correlation stays constant with supercooling , the bond - orientational correlation length steadily increases , almost doubling its size with respect to fluid state below the freezing pressure .
so there is a clear decoupling between orientational and positional order , even if it is not strong enough to sustain an intermediate phase between the fluid and crystal ( analogous to the hexatic phase in two - dimensional hard - disks @xcite ) .
instead , the simple first - order nature of the fluid - to - solid transition , dictates a discontinuous change of positional and orientational order , and instead of two - step crystallization ( like what happens near a metastable critical point ) it is more appropriate to speak of precursors mediated crystallization in hard sphere like systems . as we have observed in fig .
[ fig : map ] , the two order parameters do not have symmetric roles in the free energy , and while higher bond - orientational order corresponds to higher density , the opposite is not true . as we discussed in section
[ sec : two - step ] ( fig .
[ fig : map ] ) , the linear coupling between the two order parameter is weaker than the term @xmath60 , and higher density does not imply higher bond - orientational order .
in fact , there are many local configurations which increase the density locally , but do not increase crystalline bond - orientational order .
the most notable of these structures are icosahedral clusters , which correspond to the highest local packing of hard - spheres , but that are incompatible with crystalline bond - orientational order .
while at the fluid - to - solid transition , both order parameters become long - ranged , the transition state ( the critical nucleus ) has a more developed bond - orientational order , while the density is still comparatively low compared to the bulk crystal density ( fig .
[ fig : radial ] ) .
confocal experiments have the great potential of following the early stages of the nucleation process , and they do confirm a more rapid development of orientational order over densification @xcite .
a natural extension of the ideas presented in the previous section is their application to polymorph selection , which has attracted a lot of investigations .
the observation that structuring foreshadows densification and crystallization leads to the idea that these pre - structured regions ( precursors ) could play a role in determining which polymorph is being nucleated .
this idea requires the precursors to have some underlying symmetry that is common with the crystalline structures that are nucleated from them .
such behavior was observed for hard spheres , as reported in ref .
@xcite .
the first studies of polymorph selection from precursor regions were done with the gaussian core model ( gcm ) , which is a good model for the effective interaction between the centers of mass of polymers dispersed in a good solvent @xcite .
the model presents two stable crystalline phases : the fcc at low pressures , and the bcc phase at high pressures .
the density of the bcc phase is lower than the fcc phase , and it becomes lower than the density of the fluid phase above a certain pressure , after which re - entrant melting is observed .
nucleation in the gcm model was first considered in ref .
@xcite , where the pre - structured cloud surrounding the nuclei was taken into account to obtain a better description of the transition . ref .
@xcite then proposed a link between the precursor regions and the crystal phase which was nucleated from them .
contrary to thermodynamic predictions , it noted that the gcm model has a kinetic preference for the bcc phase , even in regions where the stable phase is fcc .
@xcite then found that the nuclei have a mixed nature , not consisting of a single polymorph , and that the kinetic pathway selected during nucleation persists even when the nucleus is many times above its critical size .
another example comes from hard spheres . here
the relevant crystalline structures are the close - packed fcc and hcp crystals . despite a negligible bulk free energy difference ( of the order of @xmath66 of the thermal energy in favor of fcc @xcite )
, simulations @xcite and experiments @xcite have both found that nucleation predominantly forms fcc .
this preference for the fcc crystal form is indeed found in the precursor regions @xcite , which show a clear preference for fcc - like local environments over the hcp - like ones .
this somehow highlights the idea that the kinetic pathway by which nuclei appear plays a fundamental role in polymorph selection .
the melt is not simply a structure - less background where nucleation can occur homogeneously , but it is instead composed of local structures with different symmetries .
symmetries that are compatible with long - range crystalline order serve as precursors , and their relative abundance dictates the likelihood of formation of a specific polymorph .
structures found in the supercooled melt are most often incompatible with crystalline order .
the most notable of these structures is the icosahedral - packing which , since the pioneering work of frank @xcite , is the archetypal model of amorphous structures . while precursor regions are linked to nucleation , five - fold symmetric structures are linked to its avoidance . in fig .
[ fig : curves ] we showed that , for hard - spheres at high density , the population of solid particles is outgrown by particles with icosahedral order @xcite .
even a small fraction of icosahedral particles strongly suppresses the crystallization process @xcite .
the effect is enhanced by polydispersity in particle sizes : for entropic reasons , polydispersity increases the amount of icosahedral structures , where often the central particle is smaller than its neighbors . at the same time
, size asymmetry suppresses crystalline structures @xcite .
frustration effects of local structures of five - fold symmetry on crystallization were also discussed in two - dimensional single - component spin liquids @xcite .
the effects of polydispersity described here apply to close - packed crystalline structures .
it was found that open crystalline structures ( as the structure of ice in water ) are instead comparatively more stable to disorder .
angular disorder , in particular , was shown to have little effect on the crystallizability of systems with directional interactions @xcite .
the effects of external fields on precursor regions were also considered @xcite .
external walls were considered experimentally and numerically in vertically vibrated quasi - two - dimensional granular liquids and two- and three - dimensional colloidal systems respectively @xcite , showing that the walls enhance structural order in the fluid , in a manner that is consistent with the bulk behavior , under an influence of wall - induced short - range translational ordering ( i.e. , layering ) .
this not only leads to the enhancement of glassy slow dynamics but also may induce heterogeneous nucleation near walls .
@xcite instead considered the effect of rough walls on colloidal suspensions under the effects of gravity , where the suppression of bond orientational order and layering next to the walls results in a strong suppression of nucleation .
also shear was shown to suppress precursor regions @xcite , while at the same time increasing the rate of growth of crystals .
the interplay between these two effects is at the origin of the non - monotonous crystallization rate as a function of the shear rate @xcite . up to now
we have addressed the role of structural heterogeneity in the melt .
this not only promotes crystal nucleation , but also causes heterogeneous dynamics , which may be related to the so - called `` dynamic heterogeneities '' , widely observed in glass - forming liquids .
it is natural to expect that more structurally ordered regions are more stable and have slower dynamics than less ordered regions .
the lifetime of these heterogeneous regions , @xmath67 , is comparable to / much longer than structural relation times for three-/two - dimensional systems respectively @xcite . for hard - sphere systems ( both mono and poly - disperse systems )
, the study of dynamic heterogeneities in supercooled melts has showed that there is a close match between slow regions and regions with high bond - orientational order @xcite .
this is suggestive of a possible general link between the glass transition and the growing length - scale of structural order , such as bond orientational order , at least for ( nearly ) single - component systems .
a link between static and dynamic properties was found to apply also to system of monodisperse ellipsoids , where lenghtscales corresponding to glassy dynamics , heterogeneous dynamics , and structural order , where all found to diverge at a common volume fraction for both translational and orientational degrees of freedom @xcite .
we mention that in polydisperse emulsions a link between dynamic heterogeneities and local order was found , but the order corresponds to the development of five - fold symmetric structures @xcite .
further work will be needed to rationalize all these results in a single framework .
once a glass is formed , devitrification is still possible , but the mechanism is still not fully understood ( see sec .
[ sec : extending ] ) , despite the importance that nucleation from the glassy state has in many systems that have industrial applications , like pharmaceuticals and metallic glasses .
recent investigations @xcite suggest that the devitrification phenomena is linked to an intermittent dynamic in the melt . during a series of discontinuous events , called _ avalanches _ , a small fraction of particles undergoes rapid motion , which is followed by an increase of the crystalline order in the melt .
a topic of practical importance in many applications is related to how the freezing transition is avoided @xcite , and what are the factors that determine the glass - forming ability of a material . in fig .
[ fig : curves ] we saw that , at high density , the formation of crystalline particles is preempted by other structures that have a high local density .
these structures were identified as having five - fold symmetric symmetry , like icosahedral clusters @xcite .
it was also shown @xcite that , by increasing polydispersity , these clusters are comparatively more stable than crystalline environments .
it is thus clear that five - fold symmetric structures play an important role in the avoidance of the crystallization transition @xcite .
see ref .
@xcite for an in - depth review of this topic .
it further suggests an intrinsic link between crystallization and glass transition : for example , suppressing the formation of crystal precursors may lead to higher activation energy for crystal nucleation and thus to higher glass - forming ability .
this topic is beyond the scope of this article and will be discussed elsewhere @xcite . in the above
, we consider precursors formed in systems interacting with isotropic potentials . in this case ,
local breakdown of rotational symmetry in a liquid state is only weakly coupled to local density , which , in the initial stage of crystal nucleation , results in a stronger increase of orientational ordering over the translational one . however , for a system interaction with a directional potential such as a tetrahedral liquid , the situation is very different since the symmetry selection by directional bonding automatically accompanies a local density change .
the typical example can be seen in tetrahedral liquids such as water , si , and ge . , defined as the radial distance between the @xmath68-th and @xmath21-th nearest neighbour to the central atom . for water
( panel a ) we consider the oxygen atoms and @xmath69 , while for hard spheres ( panel b ) @xmath70 .
( a ) tip4p/2005 water model at ambient pressure .
different curves correspond to temperatures ranging from @xmath71 down to @xmath72 , with steps of @xmath73 .
the blue @xmath74 curve refers to hexagonal @xmath74 at @xmath72 and ambient pressure .
( b ) hard sphere model , comparing the probability distribution of @xmath75 between the supercooled melt at @xmath40 ( red curve ) and the fcc solid ( blue curve ) . reproduced from ref . @xcite .
, width=302 ] in fig .
[ fig : firstsecondshell ] we compare the development of translational order in a popular molecular model of water , tip4p/2005 @xcite , with the hard sphere system .
locally , translational order is a measure of the radial order between pairs of atoms . in fig .
[ fig : firstsecondshell ] we plot the probability distribution of the distance between the first and second shell of nearest neighbours , @xmath76 , which is a measure of local translational order . in the case of water ( fig .
[ fig : firstsecondshell](a ) ) , the probability distribution of @xmath75 shows an exponential decay at high temperatures , which is the expected distribution in a disordered environment .
but as the system is supercooled , the distribution function shows the development of states with a finite value of @xmath75 , which signals the formation of the second shell . at @xmath72
the second shell is almost fully developed , and @xmath76 is close to the one measured in the ordered phase ( the hexagonal @xmath74 ice ) .
hard spheres , on the other hand , always have an exponential decay of @xmath76 , up to the limit of homogeneous nucleation , fig .
[ fig : firstsecondshell](b ) , so that there is little development of translational order prior to crystallization . the different crystallization pathway , which involves the formation of locally favoured structures with a high degree of translational order , is at the origin of thermodynamic and dynamic anomalies seen in such water - type liquids @xcite , as will be shown in the next section .
water exhibits a surprising array of unusual properties that mark it as `` the most studied and least understood of all known liquids '' @xcite .
departures from the behaviour of simple liquids are often described as water anomalies , the most famous being the density maximum at about @xmath77c .
most of these unique properties derive from hydrogen bonding , which is a strong directional interaction favoring local tetrahedral arrangement of water in both its liquid and crystalline phases @xcite .
hints about the importance of structural ordering were first detected in the supercooled liquid region . here , the liquid - liquid critical point ( llcp ) scenario hypothesizes the existence of a second critical point in the metastable supercooled phase @xcite , below which liquid water would phase separate into two different liquids , low - density liquid ( ldl ) and high - density liquid ( hdl ) ( the scenario has been intensely debated over the last years @xcite , but recent results seem to have confirmed it @xcite ) . also glassy water has been found to exist in different states , called low - density ( lda ) , high - density ( hda ) and very high - density ( vhda ) @xcite amorphous ices , which can interconvert with each other by the application of pressure . but also at ordinary conditions , many physical quantities exhibit behaviours suggestive of the presence of different states , such as infrared and raman spectra @xcite , x - ray absorption spectroscopy , x - ray emission spectroscopy and x - ray small angle scattering @xcite ( but these results are still debated @xcite ) , and simulations @xcite .
these phenomena directly point to the ability of the water molecule to sustain different local environments @xcite .
this features have attracted considerable attention since the proposal of a mixture model by rntgen @xcite and others ( see , e.g. , refs .
these models regarded water as a mixture of two rather distinct structures , whose entropy difference is not so significant . on a similar basis , but considering the formation of unique locally favored tetrahedral structures of low entropy in more disordered normal - liquid structures of high entropy , a two - order - paremter model was proposed @xcite . in this model ,
the state of water is characterized by two order parameters , density and the number density of locally favored structures , @xmath78 . in this framework
, the liquid - liquid transition can be viewed as a gas - liquid - like transition of the order parameter @xmath78 @xcite .
it was shown that modeling water as a mixture of two states has proved extremely successful in describing the anomalies of water using a restricted number of fitting parameters @xcite , but their basic assumption , i.e. , the existence of structurally different components still lacked a microscopic justification . here
it may be worth pointing out that it may not be appropriate to use the terms ldl and hdl to express fluctuations in water , although this is popular .
for example , if we consider a system near its gas - liquid critical point , we characterize fluctuations in terms of the order parameter , density , and we do not refer low and high density regions to gas - like and liquid - like regions respectively .
now we re - examine water s unique properties from a microscopic viewpoint by focusing on its nucleation pathway , which we saw involved strong development of local translational order prior to crystallization @xcite .
we focus here on a popular molecular model for water , tip4p/2005 , which was especially designed to reproduce water s thermodynamic anomalies @xcite . in the case of water , and tetrahedral materials in general ,
orientational order is best captured by an order parameter called @xmath79 which was introduced in ref .
@xcite and shown to match the orientational symmetry of all relevant crystalline structures with local tetrahedral coordination . a good translational order parameter for water
should then take into account the structure of water up to the second nearest - neighbors shell , defined in terms of the network of hydrogen bonds .
common choices include the local structure index ( lsi ) @xcite , the tetrahedral order parameter @xcite , @xmath80 ( the average density of fifth - nearest neighbor ) @xcite .
recently we introduced a new structural parameter , @xmath81 parameter @xcite : for each water molecule , it is defined as the difference between the radial distance of the closest oxygen in the second shell that is hydrogen bonded to the central water molecule , with the radial distance of the farthest oxygen atom in the first shell of neighbors .
this @xmath81 parameter includes information on hydrogen bonding network in addition to the distance measure ( see ref .
@xcite on the details ) .
the probability distribution of the @xmath81 parameter shows two distinct populations at supercooled conditions @xcite .
the first population , with a high value of @xmath81 , is characterized by an open tetrahedral structure , while the second population , with a small value of @xmath81 , is characterized by a collapsed second nearest - neighbor shell , with substantial shell interpenetration .
supercooled water structures can thus be divided in two different states that differ for the structure of their second nearest - neighbor shell @xcite .
locally favored states are represented by a gaussian population centered around a finite positive value of @xmath81 , while the disordered state by a gaussian population centered around a null value of @xmath81 , with abundant number of states with shell interpenetration ( @xmath82 ) .
this bimodal distribution of the @xmath81 parameter is in agreement with a two - order - parameter model of water @xcite , in which water consists of two types of structures , normal liquid structures and locally favored structures and the fraction of the latter is the key order parameter @xmath78 specifying the state of water .
state ( @xmath78 ) as a function of temperature for all simulated pressures .
the symbols are the values obtained by the decomposition of the order parameter distribution , @xmath83 , at the corresponding state point .
continuous lines are fits according to the two - state model .
( b ) temperature dependence of density for several pressures .
continuous lines are simulation results , while symbols are obtained from the two - state model . reproduced from ref .
@xcite.,width=302 ] the advantage of the @xmath81 parameter is that it allows a quantitative estimation of the fraction of low - energy ( high @xmath81 ) local structures directly from computer simulation trajectories .
this fraction is reported as @xmath78 in fig .
[ fig : fig6](a ) ( symbols ) .
the values of @xmath78 can be used for a two - state modeling of supercooled water , in which the free energy is written as a regular mixture of two states , the high @xmath81 and low @xmath81 states @xcite .
the prediction of the two state model is plotted with continuous lines in fig .
[ fig : fig6](a ) . once the free energy of the model is obtained in terms of @xmath78 , it is possible to derive from it all thermodynamic anomalies of water @xcite and compare them with the ones measured in simulations .
for example , figure [ fig : fig6](b ) shows the comparison between simulations ( lines ) and two - state model predictions ( symbols ) for the density anomaly of a popular molecular model of water , tip4p/2005 water @xcite .
the two - state model shows excellent agreement with the measured anomalies , proving that a microscopic two - state description of water s phase behavior is possible .
if there is a cooperative formation of such locally favored structure , the presence of two liquid phases , which are the gas and liquid state of locally favored structures , are expected at a certain condition @xcite , but whose direct observation is preempted by crystallization . we have thus seen that structures with high values of @xmath81 are directly responsible for the anomalous behavior of water , and a description based on a two - state model is able to capture the intensity of the anomalies in a quantitative way .
the parameter @xmath81 is an expression of local translational order , but structures with high values of @xmath81 are not crystalline structures .
an analysis of the structures with high values of @xmath81 in supercooled water reveals in fact that a large fraction of second nearest neighbors participates in five - membered rings of hydrogen bonded molecules , and that this fraction increases with decreasing temperature and pressure @xcite .
this five - membered rings are incompatible with the stable crystalline structures ( both ice ih and ic only have six - membered rings ) .
as the temperature decreases , the lifetime of hydrogen bonds increases , and the opening of five - membered rings to form six - membered rings becomes increasingly more rare .
this partly explains why water has such a large metastability gap , in which , in absence of impurities , it can persist in its liquid form down to very low temperatures . the existence of the homogeneous crystallization line , where the relaxation time of the liquid becomes shorter than the homogeneous crystallization time , implies that , at the best of today s knowledge , deeply supercooled liquid states will never be reached at bulk conditions , as already is found in simulations of coarse - grained water models @xcite .
thus , the experimentally inaccessible supercooled liquid region is now widely known as `` no - man s land '' @xcite , although there have been efforts to overcome the difficulty by either strong spatial confinement @xcite , mixing an anti - freezing component @xcite , or rapid cooling @xcite .
homogeneous nucleation of ices has attracted considerable attention because of its importance in nature . however , accessing it by numerical simulations has been very challenging particularly for realistic water models because of a large number of possible configurations of a disordered hydrogen - bond network , which makes the potential energy landscape complex .
the first homogeneous nucleation event in water at constant volume was reported for tip4p @xcite .
it was found that ice nucleation occurs when a sufficient number of relatively long - lived hydrogen bonds develop spontaneously at the same location to form a fairly compact initial nucleus . at a high density
, crystallization takes place more easily @xcite .
ice nucleation was also reported by metadynamics simulations @xcite and replica - exchange umbrella sampling @xcite in the isothermal - isobaric ensemble ; however , the results may depend on the choice of order parameters .
simulations using a coarse - grained mw water model @xcite were also performed and shown to be very useful because of the relatively low computational cost @xcite .
for example , moore and molinero showed that a sharp increase in the fraction of four - coordinated molecules in supercooled liquid water is responsible for water s anomalous thermodynamics and the rapid increase in the rate of ice formation @xcite .
there are also many interests on how ice polymorphs are selected upon ice nucleation @xcite .
besides this problem of polymorph selection , the estimation of the barrier height and critical nucleus size is also an important issue and recently there have been many efforts toward this direction @xcite .
a link between the so - called widom line @xcite , or the thermodynamic anomalies associated with the liquid - liquid transition , and the ice nucleation process was also suggested @xcite .
when we study crystal nucleation by numerical simulations , the choice of order parameters relevant for describing the miscroscopic pathway of crystal nucleation is crucial . in previous researches of ice nucleation ,
the bond orientational order parameter @xmath54 and the hydrogen - bond network topology characterized by six - member rings are often used as key parameters .
these choices are based on the assumption that the stable ice forms at ambient conditions are either hexagonal ice ih or cubic ice ic . however , we recently showed @xcite that these order parameters may not be sufficient for describing the crystal nucleation pathway of water - type liquids @xcite .
below we discuss this problem . in order to crystallize ,
water structures need to acquire both translational and orientational order . here
we will briefly account for a recent scenario in which crystallization of ice near the homogeneous nucleation line , and in absence of impurities , is influenced by a new metastable crystalline phase .
the difference in the locally favored structure of supercooled liquid water and the stable ice ( ice i ) forms is a source of frustration of the crystallization process . to overcome frustration effects
, the pathway to crystallization can occur through intermediate steps , in line with ostwald s step rule of phases .
in fact , in many molecular and soft - matter systems , crystallization does not occur directly in the stable crystalline phase , but instead involves one or more intermediate steps where the melt crystallizes first in metastable phases .
these metastable crystals are structurally more similar to the melt than the stable phase .
this structural similarity leads to a significant reduction of the interfacial energy , although the bulk free energy of metastable states is only intermediate between the one of the melt and of the stable phase .
this idea was recently put forward in ref .
@xcite , where a novel metastable phase was identified , called ice 0 , with a tetragonal unit cell with 12 molecules , whose thermodynamic and structural properties are intermediate between the melt and the solid crystalline phase .
the thermodynamic stability of the new phase was originally demonstrated for mw and tip4p/2005 models of water @xcite , and additional density functional calculations have confirmed that , among the known metastable ice phases , ice 0 has the closest free energy to that of stable ice i @xcite .
k and ambient pressure .
the color code is the following : ice ic ( yellow ) , ice ih ( green ) , clathrate ( orange ) , and ice 0 ( magenta ) . in ( a)-(c ) we also plot particles for which the number of connections is equal or bigger than 6 . and color them according to their total coherence as shown in the color bar .
( a ) a pre - critical nuclei ; ( b ) a nucleus of critical size ( @xmath84 ) ; ( c ) the same nucleus at post - critical sizes .
( d ) the same as ( c ) but where only particles with 12 or more connections are plotted . reproduced from ref .
@xcite.,width=302 ] a typical crystallization trajectory of mw water @xcite at @xmath85 k and ambient pressure is shown in fig .
[ fig : fig7 ] ( see ref .
@xcite on the details ) .
the orientational order parameter @xmath79 is used to detect precursor regions , where crystallization is likely to occur .
precursors are colored in pink in the figure , and the color intensity increases with the amount of @xmath79 order ; the stable ic ( cubic ice ) and ih ( hexagonal ice ) are colored in yellow and green respectively .
precursor regions are not randomly distributed throughout the system , instead they display a considerable amount of spatial correlation , which increases with supercooling .
furthermore , fig .
[ fig : fig7](a ) shows that the nucleation event happens inside the precursor regions .
this suggests that crystal nucleation is triggered by precursors in the supercooled liquid , which have a local symmetry consistent with that of the crystal to be formed .
the idea that polymorph selection starts already in the supercooled liquid state is essentially the same as the one discussed above for hard spheres @xcite . in fig .
[ fig : fig7](b ) we can see that an ice ic core has formed , and finally in fig .
[ fig : fig7](c ) , a post - critical nucleus is shown . in fig .
[ fig : fig7](d ) , where only crystalline particles ( 12 or more connections ) are shown , we can see that indeed the surface is in contact with precursor regions . between the dipole moment of a water molecule and the dipole moment of its hydrogen - bonded nearest neighbours ( panel a ) , and second nearest neighbours ( panel b ) .
different curves correspond to tip4p/2005 water , at @xmath86 k and @xmath87 bar , in the supercooled liquid ( blue line ) , ice @xmath88 ( red line ) , ice @xmath74 ( green line ) , and ice @xmath46 ( magenta line ) phases .
see supplementary information of ref .
@xcite for more details . reproduced from ref .
@xcite.,width=283 ] the idea to associate precursors with a metastable bulk phase ( ice 0 ) is based on the structural similarity between supercooled water and the ice 0 phase .
the similarity can be seen not only in local orientational symmetry , but also in the distribution of hydrogen - bonded ring structures : both liquid water and ice 0 have almost identical broad distributions of 5 , 6 , 7-membered rings peaked at 6 , whereas ice i is composed of 6-member rings only .
the presence of 5 and 7 member rings acts as the source of frustration against direct crystallization into ice i , but not for crystallization into ice 0 .
another element of structural similarity is the dipole orientation between neighbouring water molecules , where supercooled liquid water and ice 0 share great similarities .
[ fig : dipoledipole ] plots the distribution function of the average angle @xmath38 between the dipole moment of a water molecule and the dipole moment of its hydrogen - bonded nearest neighbours ( panel a ) , and second nearest neighbours ( panel b ) . in both cases , it shows the remarkable similarity between the dipole - dipole local environment of supercooled liquid water and the ice @xmath46 phase , while the stable phases ice @xmath74 and @xmath88 show a very different local distribution of dipole moments . as ostwald
s step rule of phases would suggest , nucleation could start in the intermediate ice 0 form , and then converting to the stable crystalline ices when nuclei are sufficiently big .
-@xmath89 phase diagram of mw water .
continuous lines indicate coexistence between the liquid phase and different crystal structures : ice ih / ic ( blue ) , ice 0 ( red ) and clathrate cs - ii ( green ) . dashed lines indicate constant chemical potential differences between the liquid and ice ih / ic ( @xmath90 , in blue ) , and the liquid and ice 0 ( @xmath91 , in red ) .
the green dashed - dotted line is the ic / cs - ii coexistence line .
the red open circles indicate state points where homogeneous nucleation is observed in simulations . adapted from ref .
@xcite.,width=302 ] for example , in fig .
[ fig : fig8 ] we report the phase diagram of the mw model of water ( a popular coarse - grained model of water ) .
both melting lines of the stable ice i phase ( in blue ) and of the metastable ice 0 phase ( in red ) are depicted .
the homogeneous nucleation line , where the liquid phase looses its metastability , is represented with open ( red ) dots and it is seen coinciding with a line of constant thermodynamic driving force with respect to ice 0 ( i.e. , the homogeneous nucleation line is the locus of constant chemical potential difference between the melt and ice 0 ) , rather than the stable phase ice i. in this scenario , it is shown that the presence of a metastable crystalline phase can fundamentally alter the pathway to nucleation in water .
in the above , we show that local structural ordering in a liquid state plays a significant role in both its static and dynamical behavior and non - equilibrium transitions taking place there , such as crystallization and vitrification .
conventional liquid - state theories are based on a physical picture that the liquid state is spatially homogeneous and disordered as in the gas state .
this is of course true on a macroscopic scale , however , may not necessarily be the case on a microscopic scale , particularly at low temperatures .
this local structural ordering can not always be detected by the density order parameter and we need additional order parameters such as bond orientational order parameters .
this indicates that the liquid state can not be specified by the density field alone .
such ordering must be driven by the free energy of the system , and structural ordering should be a consequence of these low free - energy structures .
the many - body nature of this structural ordering is a direct consequence of rotational symmetry selection either by excluded volume effects under dense packing or by directional interactions that often have three - body nature .
as seen in the case of water , structural ordering can play a role also above the melting transition , the density anomaly at @xmath77c being the most famous example of this .
for ordinary fluids , dominated by repulsive interactions ( such as the hard sphere model ) structural order seems to play an increasing role below the melting transition , and a liquid - state theory would require the addition of these features to effectively describe the supercooled state . on noting that crystallization and vitrification take place in a supercooled state
, such an extension may be essential for the description of these phenomena , as shown here ( see also ref .
@xcite ) . here
we note that the former - type ordering can be induced by excluded volume effects under dense packing for systems having rather hard repulsive interactions .
for very soft systems , however , excluded volume effects may not be strong enough for symmetry selection due to large thermal symmetry fluctuations .
so crystallization and vitrification phenomena may have different characters between soft and hard systems .
as reviewed above , pre - ordering in a liquid plays a crucial role in all kinds of phenomena taking place at rather low temperatures , such as crystallization , thermodynamic and dynamic anomalies of water - type liquids , liquid - liquid transition , and vitrification @xcite .
pre - ordered regions in a supercooled state act as precursors for crystal nucleation since the local symmetry of these regions should be similar to that of crystals because both symmetries are selected by the same free energy .
this matching of the symmetry leads to a large reduction of the liquid / crystal interfacial energy . when there are more than two polymorphs , a polymorph whose symmetry is similar to precursors should be statistically favored upon nucleation .
this should favor sequential ordering , known as the ostwald step rule , in which such a metastable polymorph is first formed , and then transforms to another polymorph having a lower free energy .
however , it is worth noting that the selection rule explained above , which is based on the symmetry matching and the resulting low interfacial energy cost , is different from the original ostwald step rule , which is based on the liquid / crystal free energy difference .
thermodynamic and dynamic anomalies of water - type liquids can be explained by a two - state model that assumes that the anomalies are caused by the formation of locally favored structures , which have static and dynamic properties different from normal liquid structures .
although we did not discuss liquid - liquid transition in this article , it can be viewed as cooperative local structural ordering and the two liquid states can be regarded as the gas and liquid states of locally favored structures @xcite .
finally , vitrification can be viewed as failure of crystallization because of frustration on crystalline preordering or competing ordering @xcite . in this article
, we mainly focus on ( nearly ) single - component systems . for multi - component systems , which are a quite important class of materials ,
we should consider compositional order in addition to density and structural order .
even for binary mixtures , thus , we need at least three order parameters to describe their liquid state .
this is quite important to understand crystallization and vitrification in multi - component systems .
this important and complex problem , in which one has to deal with a larger number of order parameters and with their couplings , is open for future investigation .
in the present focus article we examined nucleation as a multi - dimensional process in which the transition from the liquid to the solid phase is driven by fluctuations in two ( or more ) order parameters . after a short introduction to classical nucleation theory , we showed that the first non - classical pathways to crystallization were discovered in the context of two - step crystallization in models for proteins or colloids with short - range attractions . in two - step crystallization scenarios ,
nucleation is aided by the the formation of metastable liquid droplets that considerably increase the crystallization rate .
this pathway was found to apply broadly for a variety of systems , from small molecules , to colloidal solutions , and protein solutions .
even more surprisingly , the pathway of densification prior to nucleation was proposed to apply to systems outside the gas - liquid binodal , meaning that nucleation can occur even outside the region of metastability of the liquid droplets
. this may be worth further investigation . in the absence of a metastable liquid - gas phase separation , recent simulation and experimental work on colloidal systems
have shown that , for spherically symmetric interactions , the precursor field is not density but bond orientational order .
nucleation occurs first in regions with high bond orientational order , which then compactify as crystallization proceeds .
this is quite natural , on noting that an increase in orientational order is necessary for compactification of a dense system .
moreover , with a high degree of orientational symmetry , the precursor regions can select the type of polymorph which will nucleate from it .
several simulation results have shown that the polymorph that has been selected during nucleation can persist to sizes far bigger than the critical size .
the nucleation process , i.e. the first stage of formation of tiny crystal nuclei , has a fundamental impact on the final state , or the macroscale structure .
we also touched upon the relation between precursor regions and heterogeneous dynamics in supercooled liquids , and the effect of external fields ( walls and shear ) on these regions .
we then considered the case of water .
water has many unique thermodynamic and dynamic properties .
the study of the nucleation pathway for both the translational and orientational field has shed new light on these properties .
the translational order parameter field was described in terms of a new structural parameter ( @xmath81 ) which was used to divide the population of water structures in two states . a low - energy , high volume , ordered state ( at high values of @xmath81 ) , and a high - energy , low volume and disordered state ( at small values of @xmath81 ) .
the two - state model that derives from this microscopic description is able to quantitatively capture the anomalous behavior of water .
the orientational order of water is instead captured by another order parameter , @xmath79 .
we have shown that the development of orientational order , and the formation of small crystalline nuclei could be influenced by a new metastable phase , ice 0 , whose structure is intermediate between that of the stable ice forms and of the supercooled liquid state . finally , we discuss the general implications of this type of approach in our understanding of liquid - associated phenomena and mention important open problems .
we hope that a liquid - state theory constructed on the basis of multiple order parameters and their static and dynamical couplings will be developed in the near future , and contribute to a better understanding of these fundamental open problems .
this work was partially supported by grants - in - aid for scientific research ( s ) ( 21224011 ) and specially promoted research ( 25000002 ) from the japan society for the promotion of science ( jsps ) .
j.r . acknowledges support from the royal society through a university research fellowship .
241ifxundefined [ 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 * * , ( ) link:\doibase 10.1039/c2sm26473k [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1063/1.3506838 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1038/srep00505 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1063/1.4769981 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) \doibase http://dx.doi.org/10.1063/1.4794573 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop ( ) link:\doibase 10.1103/physrevb.28.784 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1063/1.2977970 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * ( ) @noop ( ) @noop * * , ( ) @noop ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1146/annurev.pc.34.100183.003113 [ * * , ( ) ] link:\doibase 10.1038/35065704 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/15/45/r01 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1063/1.4775738 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1209/epl / i2000 - 00276 - 4 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1063/1.4802992 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1073/pnas.0507870102 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) | nucleation is an activated process in which the system has to overcome a free energy barrier in order for a first - order phase transition between the metastable and the stable phases to take place . in the liquid - to - solid transition
the process occurs between phases of different symmetry , and it is thus inherently a multi - dimensional process , in which all symmetries are broken at the transition . in this focus article , we consider some recent studies which highlight the multi - dimensional nature of the nucleation process . even for a single - component system ,
the formation of solid crystals from the metastable melt involves fluctuations of two ( or more ) order parameters , often associated with the decoupling of positional and orientational symmetry breaking . in other words ,
we need at least two order parameters to describe the free - energy of a system including its liquid and crystalline states .
this decoupling occurs naturally for asymmetric particles or directional interactions , focusing here on the case of water , but we will show that it also affects spherically symmetric interacting particles , such as the hard - sphere system .
we will show how the treatment of nucleation as a multi - dimensional process has shed new light on the process of polymorph selection , on the effect of external fields on the nucleation process , and on glass - forming ability . |
recent progress in nanotechnology has allowed to prepare quasi one - dimensional ( 1d ) semiconductor systems containing high - mobility charge carriers , which exhibit ballistic behavior in the regime when the decoherence time is longer than the characteristic time needed for an electron or a hole to pass through the structure . in this case
the transport of charge carriers in such systems is of coherent nature .
since the ballistic transport is not accompanied by the joule losses , the conductance of the quasi 1d semiconductor systems in a single mode regime at small drain - source biases depends only on the transmission coefficient describing the elastic scattering in the ballistic region @xcite .
the latter is determined by the geometry of the system and can depend on such parameters , as the fermi energy of the carriers and external electric or magnetic fields . in some cases
, the scattering becomes spin - dependent .
this happens if the ballistic region contains confined electrons possessing non - compensated spin which interacts with the spins of the conducting electrons .
a standard textbook example of this phenomenon is kondo related phenomena in transport through individual quantum dots connected to one dimensional leads @xcite .
another example is the formation of the `` 0.7 anomaly '' in the ballistic conductance of an individual quantum point contact ( qpc ) split off from the first step in the quantum conductance staircase @xcite .
although the exact mechanism of the formation of the `` 0.7 anomaly '' is still a matter of debates @xcite , several experimental observations has indicated the importance of the spin component for the behavior of this feature .
first , an electron g - factor is found to raise from 0.4 to 1.3 as the number of occupied 1d subbands decreases @xcite .
second , the height of the feature attains to a value of @xmath0 in a strong external magnetic field @xcite .
these results have defined the spontaneous spin polarization of a 1d gas in a zero magnetic field as one of the possible mechanisms of the appearence of the feature @xcite .
the localized spins affect the propagating carriers via exchange interaction of either ferromagnetic or antiferromagnetic type .
since it is defined by the mutual orientation of spins , the transmission coefficient through the qpc becomes spin - dependent .
if the qpc contains a single localized electron , the eigenstates of the system consisting of localized and transmitted electrons , are the essence of singlet and triplet states @xcite . if the energy of the triplet state is lower than the one of the singlet state , the potential barrier for the carriers in the singlet configuration becomes higher than one for the triplet state .
therefore , at small concentrations of carriers , the ingoing electron in the triplet configuration passes by the ring region freely , while the carriers in the singlet configuration are reflected . in zero magnetic field
the probability of realization of the triplet configuration equals to @xmath1 against @xmath2 in the case of the singlet one ; thus the full conductance in the considered regime equals @xmath3 @xcite .
in contrast , if the singlet configuration is energetically preferable , the conductance equals to @xmath4 @xcite . if the uncompensated spin of electrons localized in the qpc @xmath5 , the quantization pattern becomes more complicated , and the value of the fractional plateau becomes @xmath6 for the ferromagnetic interaction and @xmath7 for the antiferromagnetic interaction @xcite .
spin related effects become even more pronounced in the ballistic transport through non - single connected objects , such as aharonov - bohm ( ab ) rings .
spin - orbit interaction of the rashba type in the ab ring induces the aharonov - casher ( ac ) and berry phase shifts between the spin waves propagating in the clockwise and anticlockwise directions , which results in the large conductance modulations due to the interference of the spin wavefunctions @xcite .
experimental observation of the ac oscillations in the gate - controlled ab rings has been reported for both electrons @xcite and holes @xcite .
moreover , formation of localized states with uncompensated spin can be also expected to affect the transport properties of the rings .
it was shown that the insert of a quantum point contact ( qpc ) in one of the ring s arms , changes the conductance pattern of the ab ring significantly due to the exchange interaction between the electron localized inside the qpc and a propagating electron .
it manifests itself in the formation of `` 0.7 feature '' on the quantum conductance staircase of the whole ring structure @xcite . in the current work ,
we analyse theoretically spin - dependent transport in a double - slit ab ring with confined delocalized spontaneously spin polarized electrons which interact with conductance electrons passing through the ring in the ballistic regime .
the localized spins affect the propagating carriers via exchange interaction of either ferromagnetic or antiferromagnetic type .
since it is defined by the mutual orientation of spins , the transmission coefficient through the ring becomes spin - dependent , which is reflected in the dramatic changes of the patterns of the conductance as a function of the fermi energy and external magnetic field .
the paper is organized as follows . in section
ii we describe the model of the device we consider in our work .
in section iii we present the formalism for the calculation of the conductance in the ballistic regime and in section iv give the describe the obtained results .
section v contains conclusions of the work .
a sketch of the device we consider is presented in fig .
1 . it represents a quantum ab ring containing spin - polarized interacting electron gas connected to two symmetrically placed leads which serve as a source and a drain of electrons from the two opposite sides . both the ring and the leads
are considered to be 1d in order to make semi - analytical treatment possible .
this approximation is legitimate as long as the fermi energy and the leads cross - section area are small enough and the condition that only lowest subband of the dimensional quantization is occupied holds : @xmath8 .
the ring contains spontaneously spin polarized electron gas with uncompensated spin @xmath9 .
[ figsketch ] and @xmath10 correspond to the transmission and reflection amplitudes . due to the exchange interaction between propagating electrons and spin- polarized electrons in the ring
these amplitudes may become spin- dependent as discussed in the main text .
, title="fig : " ] the ab ring and the leads are connected by means of two identical qpcs , scattering on which is considered to be of elastic nature . besides , scattering on them is presumed spin - independent , thus we assume that the spin of a carrier is conserved while passing through the qpcs .
thus , the propagating electrons are supposed to interact with the localized spin only in the region of the ring .
the main parameter of a qpc is the amplitude of elastic backscattering of a carrier propagating inside the lead : @xmath11 , @xmath12 , determined by the system geometry ( a qpc becomes transparent if @xmath13 ) . in our research
we consider zero temperature , hence the charge carriers ( electrons ) have a step - like distribution in the leads .
the drain - source voltage @xmath14 is also considered to be small enough , @xmath15 .
thus only those electrons whose energy lies in the vicinity of the fermi energy can take part in the transport .
the radius of the ab ring @xmath16 is taken much smaller than the inelastic scattering length to ensure the validity of ballistic transport approximation and use of well known landauer - buttiker approach to calculate the conductance .
@xcite an external magnetic field @xmath10 is applied perpendicularly to the plane of the ab ring .
this field influences both the spatial and spin coordinates of the electrons propagating inside the ab ring and the leads and thereby defines the aharonov - bohm phase shift and the zeeman splitting .
the latter we neglect for simplicity of derivation and transparency of results .
our main goal will be to calculate the dependence of the conductance of the system on chamical potential ( fermi energy ) and external magnetic field in various regimes .
let us first consider the case when ab ring contains a single localized electron .
the full hamiltonian of the system in the basis of uncoupled states dorresponding to the spins of condictance ( _ e _ ) and localized ( _ s _ ) electrons ( @xmath17 , @xmath18 , @xmath19 , @xmath20 ) can be written in the form : @xmath21 where interaction of the localized and propagating electrons inside the qpc is modelled within the framework of the heisenberg exchange hamiltonian : @xmath22 here @xmath23 characterizes the coulomb interaction between the moving and localized electrons plus the effect of the applied bias voltage ; @xmath24 corresponds to the exchange interaction ; @xmath25 ; the spin operators @xmath26 and @xmath27 correspond to the propagating and localized electrons , respectively .
hamiltonian ( [ egfullham ] ) can be easily diagonalized by the canonical transformation , @xcite and we obtain : @xmath28 where we used the notation : @xmath29 .
hamiltonian @xmath30 corresponds to the new eigensates of the coupled system : @xmath31 , @xmath32 , @xmath33},\\ \label{eqk2 } k_{s}=\sqrt{\frac{2m}{\hbar^2}\left[\mu - v_{dir}-3v_{ex}\right]}.\end{aligned}\ ] ] in a zero magnetic field , the singlet and triplet state energies are split by the value of @xmath34 . in the case of ferromagnetic coupling ( @xmath35 ) ,
the triplet configuration has lower energy and thus corresponds to the ground state ; the opposite is valid for the antiferromagnetic coupling ( @xmath36 ) . due to the processes of exchange interaction between the localized and freely propagating electrons ,
the latter can either conserve their spin projection or undergo a spin - flip .
the probabilities of these processes depend on the mutual orientation of the two spins as well as on the exchange matrix element , @xmath24 .
if the spins of the localized and propagating electrons are parallel , only spin - conservative processes are allowed .
on the other hand , if the spins are antiparallel , a spin - flip process becomes possible . using the formalism of ref.@xcite , we can write the expression for the ballistic conductance of the ab ring with exchange interaction ( at finite temperature ) : @xmath37 \left(\frac{\partial n_{k\uparrow}}{\partial\mu}\right)\\ \nonumber + \left [ p_{s\downarrow}|a_{e\downarrow s\downarrow\rightarrow e\downarrow s\downarrow}(k)|^2 + p_{s\uparrow } \left ( \right.\right .
% \nonumber \left.\left .
\right ) \right ] \left(\frac{\partial n_{k\downarrow}}{\partial\mu}\right)\\ \nonumber + \left [ p_{s\downarrow } -p_{s\uparrow } \right ] % \\
% \times \left [ n_{k\uparrow}(\mu)\left(\frac{\partial n_{k\downarrow}}{\partial\mu}\right ) -n_{k\downarrow}(\mu)\left(\frac{\partial n_{k\uparrow}}{\partial\mu}\right ) \right ] kdk \}. \label{eqcond}\end{aligned}\ ] ] in this formula we have introduced a number of parameters . the mean number of electrons with a definite spin projection and wavevector @xmath38 is the same if we do not account for the zeeman effect and reads : @xmath39 here @xmath40 is the chemical potential in the quantum wire .
it is shifted by the value @xmath41 in the right lead ; @xmath42 and @xmath43 are probabilities of the localized electron findings in spin - up and spin - down states , given by the formulae : @xmath44 and in the case of absent zeeman splitting , @xmath45 .
parameters @xmath46 ( namely , @xmath47 , @xmath48 , @xmath49 ; @xmath50 , @xmath51 , @xmath52 ) are the amplitudes of transmission through the ab ring ; the subscripts denote the spin states before and after the scattering event . to find the transmittion amplitudes ( using other scattering amplitudes )
, we will address the scattering matrix approach discussed in works .
@xcite the amplitudes of the transmitted and reflected waves can be found from the conditions of the conservation of the flux at the contacts connecting the ab ring and the leads .
as we account for the exchange interaction inside the ring , we should consider separately the cases when the spins of the propagating and localized electrons are aligned parallel or anti- parallel . in the further discussion we present the equations for the simplest case of the localized spin @xmath53 only .
the generalization for the case @xmath5 is straightforward and can be done according to the lines presented in ref ..
first , let us consider the case in which the spins of both the propagating and the localised electrons are parallel and thus electrons are in the triplet configuration .
note , that as exchange interaction conserves the total spin projection of the interacting electrons , in this case the spin- flip processes in the ring are impossible .
moreover , as we consider the case of spin- conserving contacts between the ring and the leads , the orientations @xmath54 and @xmath55 can be treated separately .
the problem thus effectively reduces to one for the transport of the spinless particles . introducing the spin - independent scattering matrix for the contacts @xmath56
one can obtain the following system of linear algebraic equations connecting the transmission and reflection amplitudes , @xmath57 and @xmath58 with the amplitudes of the waves propagating in the ring @xmath59,@xmath60 ( @xmath61 ) and shown in fig . 1 : @xmath62 the qpcs connecting the ingoing and outgoing leads to the ring
are considered to be identical .
the scattering matrix is characterized by the parameters @xmath63 which have the following physical meaning .
the parameters @xmath64 and @xmath65 are reflection and transmission amplitudes of the qpcs inside the ab ring ; @xmath11 is the reflection amplitude from the lead to itself and @xmath66 is the transmission amplitude from a lead to the ab ring or from the ab ring to a lead .
the scattering amplitudes @xmath64 , @xmath65 , @xmath11 , and @xmath66 are assumed to be real numbers .
these parameters depend on the properties of the junction , in particular on the band mismatch between the leads and the ab ring which can be electrically induced by the vertical gate voltage in the ring region .
the condition of flux conservation resulting in the hermicity of the scattering matrix allows to reduce the number of its independent elements . according to buttiker @xcite ,
the following parametrization can be used : @xmath67 where @xmath68 .
therefore , the effect of the qpcs on the scattering of a particle in the ab ring is defined by single parameter : @xmath69 $ ] .
the case @xmath70 corresponds to the fully transmitting contact , the case @xmath71 to the fully reflecting one .
the terms @xmath72 are the phase shifts between the clockwise and anticlockwise traveling electron waves , correspondingly : @xmath73,\ ] ] where @xmath74 is the magnetic flux through the ring , @xmath16 is the radius of the ab ring , and @xmath75 is the electron charge .
the wavevector @xmath76 correspond to triplet configuration of the interacting spins and is defined by eq .
( [ eqk1 ] ) . in the absence of an external magnetic field ,
the phase shift is equal for the electrons moving in both the clockwise and anticlockwise directions .
now , let us consider the situation when the spins of the localised and propagating electrons are untiparallel , and the spin flip process can occur .
for instance , consider the case when propagating electron is initially in the spin - up state and localized electron is in spin - down state ( the opposite case is fully equivalent ) .
the amplitudes of scattering then become spin dependent and instead of the single transmission amplitude one needs to introduce two amplitudes corresponding to the cases of the conservation of the spin of the propagating electron and its spin flip , @xmath48 and @xmath49 respectively ( the same is valid for the amplitudes of reflection @xmath77 and @xmath78 ) . therefore , the system of the equations for the amplitudes becomes more complicated and reads : @xmath79 @xmath80 where @xmath81,\ ] ] and @xmath82 are given by eqs .
( [ eqk1 ] ) , ( [ eqk2 ] ) .
it should be noted , that as initial configuration @xmath83 is not an eigenstate of the exchange hamiltonian ( [ egdirexham ] ) , both singlet and triplet configurations corresponding to different wavenumbers of the propagating electrons are possible inside the ring .
together , eqs .
( [ eqcond ] ) , ( [ eqscatmatr1 ] ) and ( [ eqscatmatr2 ] ) allow for calculation the conductance in the ballistic regime .
in the calculations presented herein below we used the following set of parameters : @xmath84 @xmath85 , @xmath86 , @xmath87 . the radius of the ring was taken equal to @xmath88 @xmath89 .
we also considered the cases of different values of the uncompensated spin in the ring @xmath90 ( @xmath91 ) .
the temperature was put zero .
results of modeling are presented in figs . 2 - 5 .
figures 2 illustrate the influence of the many - body interaction on the ab ring conductance for the cases of zero and non - zero magnetic fields .
+ [ figgmu ] both direct and exchange interactions of the moving and localized electrons form an effective potential barrier for the carriers entering into the ring , thus determining the conductance . in general
, there are two such sub - barriers in the ring region , for triplet and singlet spin orientations corresponding to the energies : @xmath92 and @xmath93 .
figure 2a corresponds to the case of ferromagnetic interaction ( @xmath94 ) .
the behavior of @xmath95 has a clear explanation .
let us first consider the case when magnetic field is absent shown by the red line .
the conductance @xmath95 is zero below the @xmath96 ( the lowest step of the barrier ) since independently on the mutual orientation of the spins of propagating and localized electrons the energy of a particle is not enough to overcome even the lowest barrier and all carriers are reflected .
further , if the chemical potential lies in the range @xmath97 $ ] shown by two vertical liens on the plot the carriers in triplet configuration can enter the ring while the carriers in the singlet configuration can not .
as there are three states corresponding to the triplet configuration and one state corresponding to the singlet configuration and contacts are considered to be fully transparent , the conductance in these regime equals to @xmath98 .
if @xmath99 , the electrons in both singlet and triplet configurations can enter the ring and conductance reaches the value of the elementary conductance quantum , @xmath100 .
the conductance pattern is thus equivalent to those revealed by an individual qpc with localized spin studied in refs.@xcite .
[ figgb ] if one applies an external magnetic field inducing different ab phaseshifts for the electrons moving clockwise and anticlockwise and interference is not always constructive , the conductance as a function of the chemical potential reveals oscillations instead of the plateaus .
the amplitude of these oscillations is @xmath101 if chemical potential lies in the range @xmath97 $ ] and only triplets are transmitted and increases to @xmath102 if chemical potential is increased above the value @xmath103 if transport of singlets also becomes possible .
[ figgeps ] figure 2b corresponds to the case of antiferromanetic interaction ( @xmath104 ) .
now , the singlet configuration becomes energetically preferable .
the conductance @xmath95 is zero if chemical potential lies below the @xmath105 further , if chemical potential lies in the region @xmath106 $ ] the conductance becomes equal to @xmath107 since only electrons in singlet configuration can enter the ring .
finally , at @xmath108 the ring becomes transparent for any spin orientation and the value @xmath109 is recovered .
[ figgj ] figures 3 illustrate the dependence of the conductance on the external magnetic field for different values of the chemical potential and @xmath70 .
an oscillatory behavior revealing ab effect is observed .
this happens since the conductance of the ring is governed by the phase factors which become different for the clockwise ( @xmath110 ) and anticlockwise ( @xmath111 ) propagating waves .
it should be noted , that figs .
3a , b are in good correspondence with figs .
the amplitude of the ab oscillations is defined by the value of the chemical potential .
it changes from @xmath101 to @xmath102 when @xmath40 increases above @xmath93 for the ferromagnetic case and changes from @xmath112 to @xmath102 when @xmath40 increases above @xmath113 for the antiferromagnetic case .
we have also analyzed the change of the conductance pattern if transparency of the contacts described by the parameter @xmath66 is changed .
the results are shown in figs . 4 for ferromagnetic case only ( aniferromagnetic case is similar ) .
the deviation from the case of fully transparent contacts corresponding to @xmath70 is revealed by the onset of the oscillatory pattern in the dependence of the conductance on the chemical potential even at zero magnetic field .
if @xmath114 $ ] , then only electrons in the triplet configuration can pass and the amplitude of the oscillations is @xmath101 .
when @xmath40 exceeds the value of @xmath113 , the ring becomes accessible for both singlet and triplet configurations .
the values of the wavenumber for the ballistic electron and thus conditions of the constructive interference in these two configurations are different . if the deviation from the condition of full transparency is small ( e.g. @xmath115 ) , this results in the onset of the broad oscillations with the amplitude a bit smaller then @xmath102 in the region @xmath108 , as shown in fig .
however , if one decreases the transparency further , the resonances become sharper , and conductance pattern at @xmath108 consists of the series of well resolved peaks of the height @xmath101 and @xmath112 corresponding to resonant transmission of triplets and singlets , respectively . in the dependence of the conductance on magnetic field shown at the insets of the figures ,
the decrease of the transparency of the contacts leads to the appearance of the higher harmonics connected with the increased probabilities of the round trips inside the ring and characteristic to the transition from aharonov- bohm to aharonov - altshuler - spivak oscillations , typical for the weak localization regime .
@xcite finally , we have analyzed how the increase of the localized spin affects the conductance patterns of the ab rings with exchange interaction .
the results are shown in figs .
5 . in the case of transparent contacts , the increase of the spin leads to the decrease of the value of the sub- step from @xmath101 to @xmath116(2e^2/h)$ ] in case of the ferromagnetic interaction and
increase from @xmath112 to @xmath117(2e^2/h)$ ] for antiferromagnetic interaction .
this is in agreement with the picture of the fractional quantization of the ballistic conductance presented in ref.@xcite .
the application of the external magnetic field leads to the onset of the oscillatory behavior , similar to those observed for @xmath53 .
in conclusion , we have considered ballistic conductance of an aharonov - bohm ring containing localized uncompensated spin accounting for the exchange interaction between the spins of propagating and localized electrons . we have shown that exchange interaction drastically modifies the conductance as function of magnetic field and chemical potential .
the obtained results are in agreement with the concept of the fractional quantization of the ballistic conductance proposed earlier . +
the work was supported by fp7 irses project `` spinmet '' .
thanks the university of iceland for hospitality .
i.g.s . acknowledges support of the eimskip fund . | we investigate spin - dependent conductance through a quantum aharonov - bohm ring containing localized electrons which interact with the propagating flow of electrons via exchange interaction of the ferromagnetic or antiferromagnetic type .
we analyze the conductance oscillations as a function of both the chemical potential determined by the concentration of the carriers and external magnetic field .
it is demonstrated that the amplitude of the conductance oscillations in the ballistic regime is governed by the value of the non - compensated spin localized in the ring .
the results are in agreement with the concept of fractional quantization of the ballistic conductance , proposed by us earlier [ phys .
rev .
b 71,113311 ( 2005 ) ] . |
in the year 2000 the standard model was believed to be consistent with all known experimental facts ( except the possible existence of a neutrino mass ) .
yet there are many aspects of the standard model which are assumed but not explained : the wave nature of matter , the quantization of the photon ( @xmath0 ) , and newton s second law of motion .
in addition gravity is not usually considered to be part of the standard model , but a separate interaction . with the clear measurement of dark energy by the wmap collaboration@xcite
, it became important to know what makes up the majority of our universe .
the explanation apparently does not lie within the standard model .
this paper investigates the possibility that dark energy is caused by the wave nature of matter .
the resulting model apparently succeeds in explaining dark energy .
it also appears to naturally explain the quantization of the photon and newton s second law , as well as provide an electromagnetic model of gravity . according to the standard model
@xcite electrons are believed to be point objects when examined on scales as small as @xmath1 m .
recent limits @xcite show that quarks are also not composites of other objects when examined on similar scales .
however , we know that quarks have _ external _ structure
. they are bound together by color flux tubes@xcite to form hadrons , so that free quarks can not exist .
we here discuss the postulate that the electron is similarly bound to another object by a line of magnetic flux a quantized vortex in the electromagnetic field .
we suggest that this line may be key to understanding the full meaning of spin . since we model the structure of the electron , it is possible to attempt a calculation of the properties of the electron in terms of planck s constant @xmath2 , the electron charge @xmath3 , and the distribution of matter in the universe .
this is a _ classical _ model in the sense that we attempt to follow the detailed motion of vortex - lines as a function of position in space and time .
the detailed motions of lines are then averaged to find the behavior of waves and particles normally discussed in quantum mechanics .
we do not follow the copenhagen interpretation@xcite of quantum mechanics , but rather take the view , in sympathy with einstein@xcite , that it should be possible _ in principle _ ( though it would take a rather large computer ) to predict the future course of the universe given _ perfect _ knowledge of the positions and velocities of all lines ( and other objects hadrons ... ) at a given time .
thus , in this view , indeterminacy is not intrinsic but results from measurement uncertainties .
some authors@xcite would call the positions and velocities of these lines _
hidden variables_. these quantities are difficult , but not impossible , to measure .
of course , all measurement uncertainties are subject to limitations from the heisenberg uncertainty principle , but discussion of the motions of the lines is useful because it can provide underlying structure to relate and interpret more measureable quantities .
this model is a classical model of a quantized superconducting object .
hence it is intrinsically a quantized model in which e.g. newtonian gravity is obtained as an average over the line - pair distribution .
the individual quantum fluctuations are automatically built in .
the classical solution for the motion of electrons and lines appears much like in standard quantum theory .
this model is based on the quantization of magnetic flux in free space ( discussed in section [ ssec - fq ] , and which we claim was first derived by dirac@xcite ) , which follows from the requirement that wave functions be single valued .
we take this result from quantum mechanics and use it as a general principle which reaches its most general form in equation [ eq - quantization ] , which implies the relativistic london equations [ eq - london ] for our superconducting fields .
we make a major assumption in section [ ssec - electroncharge ] about the origin of electric charge . aside from this quantization condition and our assumption about electric charge , this classical model of a quantized object is based on the principle of least action , or hamilton s principle .
the principle of least action leads to the microscopic maxwell equations which we take to be valid at arbitrarily small length scales : @xmath4 @xmath5 @xmath6 @xmath7 we generally follow the conventions of jackson@xcite .
the electric and magnetic induction fields can be defined in terms of the vector and scalar potentials @xmath8 @xmath9 which we presume to satisfy the lorentz gauge condition @xmath10 so that the four - vector potential and current density take their standard form @xmath11 @xmath12 for @xmath13 .
however , we deviate from the standard assumptions by asserting that gauge invariance is broken and the electromagnetic four - vector potential @xmath14 , at least sometimes , is directly measureable .
gauge invariance is already asserted @xcite to be broken in superconductors , and this fact is quite relevant to our present dicussion .
we prove below ( section [ ssec - static ] ) that in a very important special case , our line of flux , the vector potential is unique .
our default reference frame uses flat minkowski space referred to the local rest frame of the universe .
this is defined to be the frame in the region of our solar systsem in which the 2.73k black - body radiation is isotropic@xcite .
we claim that this is a classical model even though it starts with the dirac quantization condition a quantum mechanical result which implies the structure of our line of flux .
since electron wave functions in standard quantum mechanics have infinite extent , the only way in which they can be single - valued in the presence of magnetic flux is if the lines of flux are quantized and filamentary .
this implies that the magnetic flux in the lines must be contained ( equation [ eq - ampcirc ] ) . since normal charged particles are not sufficient to contain this flux in general , the most natural assumption we can make is that the flux is contained by a superconducting line , with charge carriers of a new type .
if we take this model seriously , it pays great dividends .
we presume that one end of a line is an electron , the other a positron ( or ends in a proton ) .
thus space - time is crisscrossed by many such lines ( in pairs - section [ ssec - overview ] ) .
then , in turn , all fundamental quantization conditions including the dirac quantization condition , the quantization of the photon , and the quantization of matter waves can be understood from the fact that space - time is a multiply connected superconductor ( equation ( [ eq - multisup ] ) ) .
this , in turn , implies quantum mechanics itself , when the interaction of a bare electron with the flux lines is taken into account ( section [ sec - quantmech ] ) , and appropriate averages are taken .
thus , the dirac quantization condition , our starting point , is understood at a deeper level and , in fact , implied by our classical model in a self - consistent way .
the classical model implies quantum mechanics , but the reverse is not true since the _ classical _ model is at the deeper level .
( we , of course , take the value of @xmath2 from experiment and we define _ classical _ wave functions which are _ not _ continuous , before smoothing to find the quantum mechanical wave functions . )
we suggest that at the birth of quantum mechanics the problem which the founding fathers solved was not due to the inadequacy of classical mechanics .
they had the wrong model of the electron .
it s not a point particle , but rather the end of a string .
since our method is totally new , it seems appropriate to summarize the highlights of our model to orient the reader to let the reader know what we will claim : 1 .
our model , many superconducting line pairs crisscrossing in space - time , naturally explains six dominant aspects of everyday physics which are either unexplained or questionable in the standard model : * quantization of the photon ( section [ ssec - quantphoton ] ) and matter waves ( section [ sec - quantmech ] ) : @xmath0 * wave - particle duality ( section [ sec - wpduality ] ) and the wave nature of matter ( section [ sec - quantmech ] ) * gravity ( section [ sec - gravity ] ) * newton s @xmath15 law of motion ( section [ sec - newton ] ) : @xmath16 * mass of the electron ( section [ ssec - masselectron ] ) * spin of the electron ( section [ ssec - spinelectron ] ) and spin of the photon ( section [ ssec - spinphoton ] ) 2 . our model gives a natural explanation of dark energy ( section [ ssec - darkenergy ] ) .
the line pairs are under tension or negative pressure ( section [ ssec - tensiondarkenergy ] ) .
we have modeled the propagation of the photon as a helical wave on the two lines of a pair ( section [ ssssec - helicalphoton ] ) . after applying the photon quantization condition and making a binding energy correction ( section [ ssec - energyphoton ] )
, we calculate the spin of the photon ( section [ ssec - spinphoton ] ) to be @xmath17 with no free parameters .
our natural model of the bare electron ( end of a line of flux ) has the pole field of a dirac monopole plus a line field ( inside a superconducting sheath ) .
after considering interaction with the surrounding line pairs , the motion of this bare electron is _ zitterbewegung _ ( section [ ssec - motionbpp ] ) , very rapid motion near the speed of light in rapidly changing directions .
the line pairs completely screen the pole field due to the quantization of magnetic flux so that the magnetic induction field reaching macroscopic distances is zero .
5 . in order to agree with experiment we must add the electric charge of an electron by hand ( section [ ssec - electroncharge ] ) to our model of the bare electron .
this is consistent with the principle of least action and does not change the character of the zitterbewegung motion ( appendix [ ssec - motionelectron ] ) .
the unnatural part of this addition is that the electric charge and effective magnetic charge of the electron must have their relative separation _ fixed _ at a very small distance parallel to the line .
this direction is the direction of the spin of the electron , and the magnitude of the spin is calculated to be @xmath18 .
6 . using a toy version of our model ( which we hope to be accurate within a factor of 10 ) , we have calculated formulas for newton s constant @xmath19 ( equation ( [ eq - g ] ) ) and the mass of the electron ( equation ( [ eq - me ] ) ) , containing some parameters whose values we do not currently know .
we have also calculated the dark energy density of the universe .
we then combine measurements of the dark energy density , @xmath19 and @xmath20 to calculate the line pair density near the earth ( @xmath21 ) ( equation ( [ eq - fulldensity ] ) ) , which is consistent with known constraints , and a factor of about @xmath22 larger than would be expected without inflation .
our model predicts that the gravitational force will reverse under conditions of very high density .
at least heuristically , this could explain inflation and agn jets .
we offer a new explanation of the pauli exclusion principle . we need to do this since our model ties a string on every electron in the universe , and therefore electrons are _ distinguishable_. the explanation occurs naturally in the model .
we offer a new explanation of the quantum hall effects simply by suggesting that the flux lines seen in these effects are real .
this model amounts to a classical field theory of interacting strings ( or lines ) .
the assumption that lines will condense into pairs avoids having to solve the general case .
we presume that intercommutation is not an issue since all excitations are quantized and excitations with a small wavelengh must have a high energy .
since this model implies quantum mechaincs and @xmath16 for the electron , differences from the standard model should be hard to find .
( but see section [ sec - relevantexperiments ] . ) in order to restrict the scope of our discussion , weak and strong interactions are essentially not discussed .
we use the word @xmath23 in the title of this paper with some trepidation , because it is an anathema to most physicists today .
if the reader prefers he / she can substitute a euphemism such as @xmath24 .
but we have apparently succeeded in describing the photon propagating as a quantized electromagnetic wave on one of our line pairs which was the original function of the ether .
we have also apparently succeeded in modeling dark energy , a uniform energy density in space which is widely believed to be inconsistent with the standard model .
this real , measureable energy density is not consistent with the idea of @xmath24 . in section [ ssec - overview ]
we address einstein s reasons for rejecting the ether concept . but ( using the language of string theory )
, we may have found the vacuum selection principle"@xcite relevant to our universe .
it makes sense that only electromagnetic strings exist outside nuclei , because the nuclear force has a short range .
our selection principle is that every neutron , or electron - proton combination , is the source of a pair of flux lines , probably stretching out to great distances in all directions .
our model gives a reasonable description of dark energy , gravity and the generation of mass without scalar fields .
newton s second law of motion , @xmath25 , is valid only in _ inertial _ frames .
the fact that inertial frames at a given point all move with constant velocity relative to each other implies that acceleration has absolute significance in our universe , unlike velocity for which only relative motion has significance . with respect to what must acceleration be defined ?
mach pointed out@xcite that the mass of the earth and the other celestial bodies must determine the inertial frames .
inertial frames near the earth do not seem to accelerate with respect to the average mass of the universe .
this paper was stimulated by the fact that somehow information on the the average local rest frame of the universe must be present _ at _ each object , to act as its absolute reference for acceleration .
this information could not be carried by a smooth gravitational field .
if we were far from any stars , the newtonian gravitational field would add to zero in a symmetric universe .
somehow the gravitational field of the universe must have _
structure_. it must be quantized .
following dirac@xcite we consider a region of space containing a wave function of a particle of charge @xmath26 .
if a nodal line appears in that wave function , this nodal line can contain magnetic flux as long as the wave function remains single valued@xcite in the presence of the flux .
this requires that @xmath27 where the path of the integral includes the line .
this implies the quantization of magnetic flux in the line @xmath28 where @xmath29 is the quantum of flux@xcite and @xmath30 is an integer .
this condition is usually@xcite taken to imply that charge is quantized if magnetic monopoles exist .
we reason that electron wave functions fill the universe and therefore we take this condition to imply the quantization of magnetic flux in free space@xcite ( with _ q = e _ ) , which , as we will see , leads in a fairly natural way to the quantization of not only electric charge , but also spin angular momentum even though true monopoles do not exist , at least in this model .
( the experimental situation on this point is discussed in section [ ssec - existing ] . )
figure [ fig - line ] shows the building block of our model a line of flux@xcite static , for the moment .
we only consider the case @xmath31 in equation ( [ eq - fq ] ) .
here we consider the case @xmath32 appropriate if the wave function is that of an electron ( charge @xmath3 ) .
magnetic induction field enters through one pole ( labelled @xmath33 ) , continues through the line and exits through the other pole ( @xmath34 ) .
the total flux exiting @xmath34 is @xmath35 , where @xmath36 is the strength of a dirac monopole@xcite .
hence the line of flux is a dirac string .
dirac introduced a singularity to exclude the line from physical reality so that the pole is a monpole . here
we _ include _ the line in our picture of reality and attempt to interpret one end of the line as an _ electron _ and the other as a _
positron_. since we assert that gauge invariance is broken , the position of the line is observable@xcite . note that the total magnetic flux through the positive pole is zero , since we must include the line field as well as the pole field .
we will speak of the total configuration as a _ vortex_. it consists of one line field and two pole fields , one of each sign .
the _ line _ is a _ string _ and we will use both terms , depending on the context .
figure [ fig - vortex ] shows the vector potential associated with a static vortex / line of flux and indicates why we call it a vortex .
the differential dipole magnetic moment of an element of the line is given@xcite by @xmath37 where @xmath38 denotes a point on the line and @xmath39 points in the direction of the magnetic induction field along the line ( which we call @xmath40 ) so that the total field of the vortex at the field point @xmath41 is given by @xmath42 taking the curl of @xmath43 we find the magnetic induction field of the vortex @xmath44 to 2.6 in to 2.0 in to 1.5 in where @xmath45 and @xmath46 are the positions of the negative and positive poles at the ends of the line .
the two - dimensional dirac delta function operates in the plane perpendicular to @xmath47 .
the line is presumed to have no sharp kinks , but it can bend smoothly in an arbitrary manner . in the particular case
that the static line is straight , the vector potential from the vortex at the field point @xmath41 is given by @xmath48}\ { \hat{\phi } } \label{eq - astline}\ ] ] where @xmath49 and @xmath50 are the polar angles from the positive and negative poles to the point @xmath41 respectively , as shown in figure [ fig - geom ] , where @xmath51 is the cylindrical radius to point @xmath41 from the line ( or extension of the line ) , and @xmath52 is the azimuthal unit vector .
lamb@xcite has summarized the classic work on vortices in the 1800 s , principally by helmholtz@xcite and lord kelvin@xcite .
most of this early work was stimulated by the circulation theorem of helmholtz , which showed that in a perfect ( nonviscous and incompressible ) fluid , the circulation of a vortex @xmath53 is an absolute constant , where * v * is the fluid velocity and the path of integration includes the vortex and moves with it . hence vortices retain individual identity so that kelvin proposed@xcite that vortex atoms " might be the building blocks of matter ( 30 years before the discovery of the electron ) .
the quantized vortices discussed in this paper bear strong resemblance to the classic vortices with the identification that the vector potential @xmath54 corresponds to the classical fluid velocity @xmath55 and that the magnetic induction field @xmath56 corresponds to the classic vorticity ( @xmath57 ) .
the circulation around one of our vortices @xmath58 ( far from the ends to avoid flux returning from the poles ) is quantized , implying that they also retain a separate identity and are good candidates for building blocks .
the classic vortices could not end , so that they either had to form rings or end on the walls of the container .
our vortices , in the terminology of the classic works , actually do not end either .
since the fields of our vortices satisfy the maxwell equation @xmath59 everywhere , technically our vortices do not end at the poles .
their cross sections just become very large at the poles , and the magnetic field extends to infinity in all directions .
there are two major differences , however , between our quantized vortices and the classic vortices .
since our vortices are formed in the electromagnetic field , they have no rest mass and do not sense pressure in the normal way , except at the poles .
( see section [ sec - pauli ] . )
consequently we must develop new techniques to follow the motions of our vortices . since we wish to consider the energies of the lines of flux ,
the singular behavior of the magnetic induction field ( equation ( [ eq - singular ] ) ) in the filamentary model is unacceptable .
we need to integrate @xmath60 .
consequently we develop a model of a line of flux with a finite radius @xmath61 , called the _ line radius _ , which we take to be constant . for the moment
we consider only a straight line far from its ends .
we simply use a model of the abrikosov vortex @xcite which has been successful@xcite in modeling the vortices in type ii superconductors @xmath62 where @xmath63 is the hyperbolic bessel function of the second kind@xcite of order @xmath64 .
the line radius plays the role of the penetration depth in a type ii superconductor .
since the magnetic flux is confined , there must be a current distribution associated with the line . in the static case , from equation ( [ eq - ampcirc ] )
we find @xmath65 where @xmath66 is the hyperbolic bessel function of the second kind of order 1 .
we assume that this current density is _
superconducting_. it is associated with the magnetic field at zero voltage , as in a type ii superconductor . we assume the superconducting region around the line extends to a distance @xmath67 from the centerline , where @xmath68 , so that the flux is contained to any required accuracy
. we will try to measure @xmath69 .
we assume that the current density satisfies the london equations@xcite since , if a coherence length is relevant to this situation , we take it to be negligible .
we treat this current density as a _ field _ , so we can take its charge carriers to be massless ( section [ ssec - superconfield ] ) . hence the current density is a new concept since it can not be due to the motion of known charged parcicles !
these assumptions seem required by our main point , that , following ( [ eq - fq ] ) , magnetic flux is quantized in free space .
we will see that every wave function contains such lines , and therefore all wave functions superconduct .
this last statement is not surprising ; it is just diamagnetism@xcite .
hence the surprising assumptions which we have been led to make in this section have one reasonable consequence .
we need to see what else is implied . with this magnetic induction field , or magnetic flux density ,
the total flux through the line at a given point on the centerline is @xmath70 as required .
the self - energy per unit length of the line is given by @xmath71 where @xmath72 is the tension in the line .
the differential magnetic moment of an element of the line is given by @xmath73 also as required .
following the expected behavior@xcite in superconductor we assume a trial solution for the vector potential given by @xmath74 taking the curl of @xmath75 , we indeed find equation ( [ eq - bline ] ) .
but our trial solution has the wrong behavior for large @xmath51 ( see equation ( [ eq - astline ] ) , for points far from the ends ) since @xmath76 for @xmath77 . in order to obtain the proper solution
we must add a field with zero curl ( except at @xmath78 ) @xmath79 which gives the correct long - distance field of the line , so that the total field is given by @xmath80 the curl of ( [ eq - aunique ] ) gives the correct magnetic induction field of a static line ( [ eq - bline ] ) at all points@xcite ( not near the ends of the line ) .
hence ( [ eq - aunique ] ) gives the _ unique _ solution for the vector potential of the line ( for n=1 , not near the ends ) .
it is determined by the azimuthal symmetry of the line and the quantization condition ( [ eq - fq ] ) . actually equation ( [ eq - singv ] )
is a special case .
the requirement that a particle s wave function be single valued at a fixed time is stated more generally by @xmath81 where @xmath82 is the total momentum of the particle , and where we use a notation common in the case that the wave function represents superconducting pairs of electrons . if we extrapolate from the case of superconducting pairs of electrons @xmath83 ( where @xmath84 is the number of superconducting pairs per unit volume ) , to our line current density in terms of the penetration depth@xcite @xmath85 ( which has been identified with the line radius ) we find @xmath86 and hence @xmath87 from ( [ eq - aunique ] ) @xmath88 we can find the generalization of @xmath89 to our case @xmath90 which we call the _
phase field_@xcite of the vortex , because , in the case of our vortex , it does not behave like a momentum .
it is independent of both @xmath91 and @xmath84 .
we associate this phase field with the vortex supercurrent wave function .
we say that this new type of current is due to _ aon _ superconductivity , in order to distinguish it from superconduction by cooper pairs .
we will see that this new phenomenon is completely characterized by the line radius , which we call _
evidently , in our model of the core with a finite size , the phase field still has a filamentary curl , and in the general static case it is given by ( see equation ( [ eq - agenline ] ) ) @xmath92 which is just the law of biot and savart . a supercurrent ( i.e. aon ) wave function will be single valued as long as @xmath93 throughout the region of interest . in general , using the line of reasoning for equation ( [ eq - singular ] ) , the curl of the phase field of a single static vortex not including the ends ( where @xmath89 is not smooth ) is given by @xmath94 the line field itself is quantized precisely so that its supercurrent wave function encircling the centerline is single - valued .
since the curl of the phase field is filamentary , there is no region of space in which the wave function of one line is not single - valued and therefore two lines ( otherwise isolated ) can move through each other driven by their magnetic interaction with no extra currents ( quantization currents ) needed to make their wavefunctions single - valued@xcite .
we will see ( section [ ssec - effectivemagcharge ] ) that this is _ not _ true for the ends of a line .
the london equations@xcite may be expressed very simply in terms of the phase field @xmath95 @xmath96 where @xmath97 is the 0-component of the 4-vector @xmath98 and @xmath99 is the 4-velocity ( extrapolating from the case of superconducting pairs of electrons ) .
the relativistic form of the london equations @xmath100 just implies that @xmath101 the phase field can be expressed as the gradient of a scalar ( @xmath17 times the phase ) .
evidently the center of the vortex is outside the superconducting region , as in a type ii superconductor .
this implies a phase transition in the center of the vortex .
if we allow time variation , the requirement that a wave function be single - valued can be written in all generality @xmath102 so that @xmath103 the london equations imply that this requirement is automatically satisfied @xmath104 _ if _ the path and surface are totally in superconductor .
( we use stokes theorem in four dimensions and @xmath105 is an element of the area enclosed by the path . )
we consider the interaction of two static lines ( far from their ends ) , first in the case in which the magnetic induction fields of both lines are aligned with the @xmath106-axis .
the magnetic energy per unit length is @xmath107 @xmath108 @xmath109 where @xmath110 is the interaction energy .
this interaction energy is easily calculated @xmath111 ( where @xmath112 locates the center of line @xmath113 relative to line @xmath114 ) using the expression obtainable from jackson@xcite @xmath115i_m(\rho_</a)k_m(\rho_>/a ) \label{eq - k0fact}\ ] ] for writing @xmath116 in completely factorized form . here
@xmath117(@xmath118 ) is the greater ( lesser ) of @xmath51,@xmath119 .
if @xmath120 and @xmath121 we obtain @xmath122\ \ .\ ] ] recognizing a wronskian determinant@xcite @xmath123\ = \ + { 1 \over { u^\prime } } \label{eq - wronskian}\ ] ] we find that the interaction energy is given by @xmath124 and the energy per unit length of two aligned lines is @xmath125 \ = \ { { g^2}\over{a^2}}\ \bigl [ 1+{{\rho^\prime}\over a}k_1({\rho^\prime}/a ) \bigr ] \ \ .\ ] ] since @xmath126 as @xmath127 , the total energy of two aligned lines is four times as great as the energy of one line ( see equation ( [ eq - t ] ) ) , as expected .
hence two lines of like orientation _ repel _ each other . from the above discussion , two lines with opposite orientation
will attract each other .
the interaction energy simply reverses sign .
@xmath128 \
= \ { { g^2}\over{a^2}}\ \bigl [ 1-{{\rho^\prime}\over a}k_1({\rho^\prime}/a ) \bigr ] \ \ .
\label{eq - twolinepot}\ ] ] the minimum of the twoline potential energy occurs at @xmath129 , and has the value @xmath64 .
the two line self energies are cancelled by the interaction energy . since the lines are fields , the cancellation is complete except , of course , for excitations of the twoline system .
we note that the annihilation of two oppositely directed vortices has been observed@xcite in type ii superconductors .
at the birth of the universe we envision a very large concentration of energy in the universe in accord with standard ideas of the hot big bang model@xcite .
we envision that the energy is originally in the form of continuous quantized vortices ( lines ) which have no ends . when sufficient energy is available locally , some lines break creating electron - positron pairs .
as the universe continues to expand , oppositely directed lines are attracted to each other and eventually bind .
initially excitations of the twoline systems are very large and lines frequently switch partners . but as the universe cools excitations become smaller and smaller , and in our current universe the line pairs are assumed to be condensed with an energy very close to zero ( equation ( [ eq - twolinepot ] ) ) . by this assumption
we avoid having to solve the general case . in order to limit the scope of our discussion
, weak and strong interactions will not be discussed more than is absolutely necessary . presumably cooling processes occurred with these interactions , which could be similar to the processes which we will discuss .
but , first we need to mention how protons and neutrons fit into our picture . noting the decays @xmath130 @xmath131 ( the latter occurring only in nuclei ) we presume that if a negative electric charge ( electron or antiproton ) is at one end of a line , any positive charge ( positron or proton ) can be at the other end . from ( [ eq - pton ] )
we infer that the proton must have the same type of line as the positron ( defined by positive charge ) , and from ( [ eq - ntop ] ) we infer that the neutron has _ both _ types of line , i.e. contains the end of a line pair . thus in today s universe , essentially all lines are bound in pairs , with the upper limit on the number of unbound lines set by the neutrality of matter .
the stability of these line pairs is discussed in section [ ssec - tensionpair ] . from ( [ eq - ntop ] )
we presume that one line pair can end at a hydrogen atom or at a neutron .
the ends of the lines near the earth are bound together in atoms . in stars
significant ionization occurs , but from the net neutrality of plasma matter in stars and energy considerations , it is reasonable to believe that the two ends of a line pair do not become greatly separated .
the threshold for stringbreaking ( @xmath132 ) is barely approached in the cores of the most massive stars@xcite .
our picture offers hope of explaining the evolution of the universe from the big bang to the present day@xcite , without adding any scalar fields .
the earliest epoch is thought to be dominated by strings which may correspond to our lines .
the evolution of this model seems natural because the strings are still here !
we suggest that inflation was due to the reversal of the force of gravity under the very high densities of the early universe ( which we will discuss ) , when our line pairs were radiation dominated and pressures were naturally positive . in recent times , when the line pairs are cold , their tension ( negative pressure ) dominates and they become the source of dark energy which may explain the recent acceleration of the universe .
photons are the differential excitations of a twoline system , and gravitational waves are common mode excitations of a twoline system ( differs from the standard model ) .
the higgs field of the standard model is replaced by the field of line pairs .
the concept of an ether was rejected in the early part of the @xmath133 century for many reasons .
we quote einsteins@xcite reasons for rejecting the ether concept , along with our reply to each point for why we now think the concept is useful : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ if light was to be interpreted as undulatory motion in an elastic body ( ether ) , this had to be a medium which permeates everything ; because of the transversality of the lightwaves in the main similar to a solid body , yet incompressible so that longitudinal waves did not exist . " : in this model the ether does permeate essentially everything but it is not a solid body , but rather a mesh of lines .
light waves are only transverse because they are waves on a string ; they are not bound longitudinally . this ether had to lead a ghostly existence alongside the rest of matter , inasmuch as it seemed to offer no resistance whatever to the motion of ponderable `` bodies . '' : the ether is ghostly " because it is made of magnetic fields . it does offer resistance to the motion of massive bodies , because we postulate that it is the origin of mass . it offers resistance to acceleration , but does not offer resistance to velocity in accord with newtons @xmath15 law , and special relativity . in order to explain the refraction - indices of transparent bodies as well as the processes of emission and absorption of radiation , one would have had to assume complicated reciprocal actions between the two types of matter , something which was not even seriously tried , let alone achieved . " : there is only one type of matter .
electrons are literally connected to the ether in this model .
if an electron accelerates it causes a wave on a string . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ it should be noted that there is a great similarity between maxwell s _ molecular vortices_@xcite with which he constructed his model of the ether and the quantized vortices considered here .
there is a big difference , however .
following the prejudices of his day , maxwell attempted to explain electromagnetism in terms of mechanics .
we , however , attempt to explain mechanics in terms of his electromagnetism .
it is necessary to estimate the average length of a line pair , @xmath134 , in the universe today . in the early universe
there were two competing processes that determined the length of the average line .
during inflation the velocity of separation of the two ends of a line exceeded the velocity of light , perhaps greatly@xcite .
this expansion competed with the breaking of lines forming an electron - positron pair at each break . because of
these two processes there will be a statistical distribution of line lengths about some mean , @xmath134 . in order to avoid dependence on a specific model of inflation
we simply take , as a first guess , @xmath135 @xmath134 to be the distance to the horizon ( where @xmath136 is the time of the big bang ) . within a minute@xcite after the big bang both rapid inflation and string - breaking are expected to have stopped , so that the lines ( now in pairs ) have simply expanded with the universe since that time . thus we might expect our estimate for @xmath134 to be close in the absence of inflation . in this sense we view it as a _
minimum _ value .
the value of @xmath134 is a major uncertainly in these considerations and we will attempt to measure it . according to the picture
we have developed , essentially all lines are paired and , at least near the earth , they start and end at atoms .
there is on average throughout the universe about one hydrogen atom ( or neutron ) per cubic meter@xcite , so this minimum average length of line pairs per unit volume is @xmath137 we associate each atom with one - half of @xmath134 since a line pair has two ends .
we assume all points in the universe are equivalent and we eliminate edge effects by taking the universe to be very large . taking the time since the big bang to be @xmath138 years we have @xmath139 and @xmath140 we refer to this quantity as the line - pair density of the ether .
if the number of line pairs per unit area per unit solid angle is denoted by @xmath141 , it can be shown that @xmath142 if the distribution of line pairs is isotropic , and that @xmath143 where the solid angle of line pairs only changes from @xmath64 to @xmath144 to avoid double counting .
( we also refer to @xmath145 and @xmath146 as line pair densities . )
thus within the first bohr orbit of a hydrogen atom , with the above minimum value for @xmath134 @xmath147 there are about @xmath148 line pairs .
however , within the radius of the proton ( @xmath149 as determined by electron scattering@xcite ) there are only @xmath150 about @xmath151 line pairs .
( in section [ ssec - estimatelpd ] we will attempt to measure @xmath152 and tentatively show that it should be increased over our assumed value ( [ eq - lave ] ) by an estimated factor of @xmath22 , presumably as a result of inflation .
but since this inflation factor is quite uncertain , we prefer to show these minimum estimates here . )
we note that the net flux of line pairs is always zero because no net direction is defined .
if , however , the ether could be completely polarized ( elimintate lines with * b * pointing from right to left ) , the magnetic induction field would be tremendous ( with the above value of @xmath152 ) .
@xmath153 @xmath154 @xmath155 this field indicates how cold ( condensed ) our universe has become .
magnetic induction fields of this strength are thought to exist@xcite in neutron stars .
a condensed line pair ( far from its ends ) has little energy density and no net magnetic field .
how can it be detected ?
it can respond to a test particle ( electron ) which does have an electromagnetic field , and cause it to have mass . the quantum nature of this field of line pairs should be emphasized .
essentially every one of these line pairs ends at a particular atom , or at least near a particular location , someplace in the universe .
since the universe is homogeneous and isotropic on the large scale , the field of line pairs , reflecting the universe as a whole , should also be homogeneous and isotropic as well .
but , for instance , if a star approaches the field point under our consideration , the line - pair field will vary both in time ( as the star approaches ) , and direction ( strongest toward the star ) .
the star may be assumed to have one line pair per nucleon stretching off to great distances . as a first approximation
, we assume the line pairs stretch out isotropically from a star .
( see section [ sec - rotations ] for a discussion of rotation . )
the density of line pairs is obviously quantized , but with the above value for @xmath145 , the density of line pairs is sufficiently large for most purposes that we can average over the quantum fluctuations of this distribution , and treat it as a smooth field .
our superconducting fields are _ aon _ superconductors .
they become polarized in an electric or magnetic field but do not conduct electrons or cooper pairs since the electric and magnetic fields of the electron can not exist inside the superconductor .
( we will show ( [ eq - poleforce ] ) that , at least near the earth , the pole field of the electron dominates at close range and repels superconductor . ) hence the line pairs do not short out all voltage differences in the universe .
the density of line pairs in the ether is very large and we must remember that these pairs are _
superconducting_. hence the line pairs certainly have effects on each other .
the line pairs cross and hence form closed superconducting circuits which cause the quantization of magnetic flux .
we will show that these circuits also cause the quantization of the photon .
_ because _ of these quantization effects it is reasonable to presume that cold , condensed line pairs can move freely through each other , i.e. intercommute freely , once this quantization is taken into account .
we have seen that quantized lines can move through each other ( section [ ssec - phase ] ) affected only by their magnetic interaction .
excitation energies are quantized and low energy excitations ( such as photons ) have very long wavelengths , which do not couple well to objects as small as another line pair ( smaller than an atom ) . with this presumption
we avoid a problem which would otherwise cause great difficulty@xcite , especially with the above density of line pairs ( equations ( [ eq - dldvnum ] ) and ( [ eq - dndanum ] ) ) .
we presume that intercommutation was important only in the early universe , and perhaps in the interior of stars .
we will show that _ all _ excitations of a line pair obey the photon quantization conditions .
when these conditions are satisfied , as indicated above , the line pair can _ not _ be considered isolated .
supercurrent flows around closed circuits as necessary to maintain the quantization conditions .
we call these supercurrents _ quantization currents_. we will also show that all line pairs in the universe ( at the same time ) have the same tension if they do not have an excitation .
in section [ sec - introduction ] we defined the _ microscopic _
maxwell equations , and until now we have had no need to deal with macroscopic electromagnetic quantities , which jackson@xcite informs us involve a spatial average over a typical length of order @xmath156 , appropriate for standard atomic matter . in our case by _
microscopic _ we mean on a length scale @xmath157 , where @xmath158 is the typical separation of line pairs .
we take _ macroscopic _ to mean on a length scale @xmath159 . after inflation
is taken into account we will find that @xmath158 is very small .
( see table [ table - lpdatearth ] . )
now we will need to consider macroscopic electromagnetic fields which we define as @xmath160 and @xmath161 in order to distinguish them from the microscopic fields @xmath162 and @xmath56 .
the standard macroscopic fields are : @xmath161 - magnetic induction , @xmath163 - magnetic field , @xmath164 - magnetization , @xmath160 - electric field , @xmath165 - displacement vector , and @xmath166 - polarization .
the fields @xmath163 , @xmath164 , @xmath165 , and @xmath166 are only defined as macroscopic fields .
an electric field will certainly polarize the ether since dipole fields will be set up on the exterior of every line pair , ensuring that the net electric field falls to zero several penetration depths inside a line pair .
hence , including this polarization @xmath166 , the ether has a displacement vector given by @xmath167 where @xmath168 is the dielectric constant of the ether .
a similar expression holds for the magnetic permeability @xmath169 @xmath170 we can achieve an exact analogy to the electric polarization by rearranging ( [ eq - bmac ] ) to find @xmath171 since a superconductor is a perfect diamagnet , the magnitization @xmath164 is opposite to the applied field @xmath56 in order to cancel @xmath56 well within the superconductor .
this follows because a magnetic dipole @xmath172 produces a @xmath56-field parallel to @xmath164 , but an electric dipole @xmath173 produces an @xmath162-field _ opposite _ to @xmath166 .
hence with the negative sign in ( [ eq - h ] ) we have an exact analogy with equation ( [ eq - d ] ) . electric field
* e * is excluded from each superconducting line pair in ( [ eq - d ] ) , and @xmath56 excluded from the same line pairs in ( [ eq - h ] ) .
hence @xmath174 so @xmath175 for the superconducting ether
. we can achieve a fairly realistic calculation of the dielectric constant of the ether by considering the case of normal matter where atoms are modeled by superconducting spheres of radius @xmath176 , as originally proposed by mosotti@xcite .
( he called them perfect conductors before superconductivity was known to exist . ) in this case the local field at the atom in question is given by @xmath177 and the atomic electric dipole moment is given by @xmath178 where @xmath179 if @xmath30 is the number of such atoms per unit volume then @xmath180 where @xmath181 is the fractional concentration of superconductor .
we can approximate a line pair by a row of superconducting spheres , so we will simply take @xmath182 to be an expression for the dielectric constant of the ether , where @xmath183 is the fractional concentration of superconducting fields .
this expression approaches @xmath184 as @xmath185 as it must for a uniform superconductor .
there is a major difference , however , between atomic spheres , which are subject to the pauli exclusion principle ( section [ sec - pauli ] ) and line pairs , which can move through each other .
because line pairs can move through each other , they can have significant overlap . as
a result @xmath186 is lower than it would be if the line pairs were impenetrable .
but if the density of line pairs becomes _ very _ large , e.g. near a black hole , we expect @xmath186 to approach infinity as the ether becomes a continuous superconductor .
we presume that equation ( [ eq - epsminus1 ] ) is completely general and will apply it even if the ether is not homogeneous and isotropic .
this equation is particularly useful because of its simplicity . from the macroscopic maxwell equation@xcite @xmath187 we see that the charge on the bare positron is _ not _
@xmath188 but @xmath189 after correction for the dielectric constant of the ether .
this implies that the flux quantum is actually @xmath190 where @xmath191 if we use the correct value of the bare electron s charge .
standard methods of measuring the flux quantum@xcite use the measured charge of the electron , so will obtain @xmath192 .
since the density of the ether ( [ eq - dndanum ] ) is large , @xmath186 could be large .
we assume that @xmath193 and @xmath194 are fundamental constants but that @xmath188 is _ not _ , because the density of the ether changes as the universe expands .
hence in all formulas to this point in this paper we should make the replacement @xmath195 in particular the tension of a line becomes @xmath196 the phase field ( [ eq - ps ] ) involves the product @xmath197 , so does not change , and the fundamental quantization condition ( [ eq - quantization ] ) is not affected .
hence we recognize that the large magnetic field in ( [ eq - bunbalanced ] ) could be too large if @xmath198 .
but it is the field which would be measured using current techniques .
the classical solution for the motion of a filimentary light string has been illustrated by goddard , goldstone , rebbi , and thorn@xcite . following this reference we take the action for a light string proportional to the area of the surface swept out by the string between fixed times .
if the string is parameterized as a function of @xmath199 , determining the position along the string , and @xmath200 determining the position in time , then a general position along the string is given by @xmath201 the action is given by @xmath202 where @xmath203 @xmath204 @xmath205 and @xmath72 is the tension ( [ eq - t0 ] ) , presumed constant .
for some purposes it is convenient to consider this parameterization arbitrary@xcite .
it is easy to see that the action is independent of the particular parameterization . in order to consider a detailed solution , it is convenient to take @xmath206 . in this case
@xmath207 ( since we choose @xmath208 , @xmath209 is _ not _ a 4-vector . ) if @xmath210 ( a function of @xmath199 ) represents length along the string @xmath211 is a unit vector .
if @xmath212 @xmath213 and @xmath214 the generalized momenta are @xmath215 and @xmath216 hence , from the variation of the action , the general equation of motion is @xmath217 with the boundary conditions @xmath218 where @xmath219 is the variation in @xmath220 , which is presumed to vanish at @xmath221 and @xmath222 .
the boundary conditions are satisfied if @xmath223 the ends of the line are held fixed . for some purposes
this might be a good approximation since we will see that the ends of many lines are bound in matter .
but in general we can not require the ends of the lines to be fixed . in the more useful case that the ends of a line are free , the boundary conditions are satisfied if @xmath224 all four of these boundary conditions at each end point can be satisfied if @xmath225 and @xmath226 .
this implies that the effective tension at the end of the line is zero and the end moves with the speed of light transverse to the line .
we will interpret this motion as the _
zitterbewegung_@xcite of a bare electron familiar from the theory of the dirac equation .
for the moment we simply assume that the boundary conditions are satisfied so that we can proceed to discuss the motion of a line .
we give a fuller discussion of the boundary conditions in section [ ssec - motionbpp ] .
it probably comes as no surprise that any wave traveling at the speed of light is an exact solution to the equation of motion of the string ( [ eq - eqofmotion ] ) .
we presume that the string stretches over space in the @xmath106-direction so that it is convenient to take @xmath227 as the parameterization along the string
. then a solution of general shape is given by @xmath228 where @xmath229 is a smooth function of arbitrary shape .
we have @xmath230 @xmath231 ( where @xmath232 is the derivative of @xmath229 with respect to its argument ) @xmath233 @xmath234 @xmath235 @xmath236 so that @xmath237 \biggr)\ ] ] @xmath238 note that the wave is only a function of its argument @xmath239 so that , when acting on @xmath232 @xmath240 and @xmath241 the equation of motion ( [ eq - eqofmotion ] ) becomes @xmath242 and we have a solution
. note that in our parameterization the wave is transverse to the @xmath243-direction , the average direction of the string , but it is _ not _ transverse to the instantaneous direction of the string as assumed in reference@xcite , because the string bends .
indeed @xmath244 is always negative or zero .
the only solution with @xmath245 everywhere requires @xmath246 , i.e. no wave .
the authors of reference@xcite go on to discuss the quantum dynamics of the light string and claim that in order for a reasonable solution to exist , the dimension of spacetime must be 26 .
this claim does not apply to our model because it is a classical model though we will construct from it the elements of quantum mechanics .
we work in four spacetime dimensions , but do not exclude the existence of hidden , compactified dimensions . using the above methods
, it can be shown that a slightly more general wave traveling at the speed of light @xmath247 is also a solution of ( [ eq - eqofmotion ] ) , where both @xmath248 and @xmath229 are smooth functions of @xmath249 with arbitrary shape . since the electromagnetic string can interact with itself
, we must consider interaction even when considering motion of a single line .
it has been shown by barut and bornzin@xcite that the equation of motion for the filamentary dirac string with interactions is given by @xmath250 where now @xmath251 is the mechanical string lagrangian density given by equation ( [ eq - lagrangiandensity ] ) and the total lagrangian density including interactions is given by @xmath252 where @xmath253 is the dual of the electromagnetic field tensor at point @xmath254 .
@xmath255 dirac@xcite showed that in the filamentary model , but otherwise in the general case , that the dual field of the line is given by @xmath256 and therefore the electromagnetic field tensor of the line is given by @xmath257 dirac s pole current formalism can be used ( section [ ssec - diracpolecurrent ] ) to find the field at the ends of the line . using maxwell s equations
the line current density is given by @xmath258 we define a superconducting field as a region of spacetime within which the relevant phase field obeys the equation @xmath259 for all closed paths enclosing areas that are totally within that region , so that the corresponding wave function is single - valued . in this case
we see from ( [ eq - quantization0 ] ) that the london equations ( [ eq - london ] ) are exact .
if the superconducting field is multiply connected , we must allow flux quanta to be outside the superconducting region but enclosed by the path @xmath260 here the phase field corresponds to the momentum field of cooper pairs ( [ eq-4ps ] ) in a type ii superconductor . in the case of a superconducting line
we extrapolate all results to the limit @xmath261 and @xmath262 such that the line radius @xmath61 ( [ eq - lambdal ] ) remains finite , at its characteristic constant value .
the phase field must also satisfy the general relationship between @xmath99 and @xmath14 implied by maxwell s equations with sources @xmath263 in terms of the four - vector potential we obtain @xmath264 this is the usual equation relating the electromagnetic potential to its source current . in a superconducting field the london equations ( [ eq - london ] ) give @xmath265 or , with the use of @xmath266 we obtain @xmath267 implying that the supercurrents respond to the electromagnetic field at the same point in spacetime . using the equation of continuity ( which follows from maxwell s equations ) @xmath268 we obtain @xmath269 \ + \ { { 4\pi } \over c}\ j^\nu\ = 0\ ] ] so that @xmath270 this equation depends only on the london equations and maxwell equations , so it is _ always _ true for a supercurrent obeying ( [ eq-4supercurrent ] ) .
the general solution of ( [ eq-4potto4current ] ) in a superconducting field is simply @xmath271 where we have used ( [ eq-4ps ] ) and ( [ eq-4supercurrent ] ) . using ( [ eq - lorentzgauge ] ) , ( [ eq - london ] ) , ( [ eq - continuity ] ) and
( [ eq - daljej ] ) we find
@xmath272 so we indeed have the solution . this general solution ( [ eq - generalsol ] )
is useful because @xmath273 can often be determined . in a simply connected superconducting field
far from other objects @xmath274 as is well known for the 3-vector @xmath89 from static situations@xcite , so that @xmath275 barut and bornzin@xcite claim that the relation between the four - vector potential and the source current density uses the standard retarded green function .
this is _ not _ the case in a superconducting field because the electromagnetic field generates supercurrents ( [ eq - emgensupercurrent ] ) , and this is not taken into account in the standard formalism .
thus in a superconducting field the electromagnetic field does _ not _ propagate outward from the source supercurrents like a wave .
the 4-vector potential is proportional to the supercurrent density at the _ same _ position in spacetime , with the possible addition of a field with zero curl which can not affect the electromagnetic field .
the principle of causality holds that no information can travel faster than the speed of light .
this principle is _ not _ required by special relativity since it is violated by an object that moves with a finite size that does not propagate as a wave .
we have seen that in a superconducting line the electromagnetic field does not propagate like a wave .
motion as an object of finite size satisfies ( [ eq-4potto4current ] ) and ( [ eq - generalsol ] ) because @xmath276 , @xmath277 and @xmath278 move together .
so we assert that the entire cross section of a line moves together as a unit , continuing to satisfy the quantization condition ( [ eq - multisup ] ) for a multiply connected region .
this is a violation of the principle of causality on the microscopic scale . using the phase field
we can illustrate this violation directly . for a path which includes the centerline of a superconducting line
, we obtain from ( [ eq - multisup ] ) in the transverse rest frame of an element of the line centered at the point @xmath279 @xmath280 if the line contains one quantum of flux ( taking @xmath281 for convenience ) . here
the @xmath282 refers to the transverse rest frame at @xmath279 , moving at @xmath283 ( see ( [ eq - beta ] ) ) relative to the lab ( default rest frame , see section [ sec - introduction ] ) . because of equation ( [ eq - intpseh ] ) we can no longer have @xmath284 in the presence of the centerline .
but the london equations ( [ eq - london ] ) must hold outside of the centerline , so within the line @xmath285 we presume that @xmath286 , since @xmath287 would be equal to @xmath64 if not for the flux quantum at the core of the line .
we now consider motion in the lab but restrict consideration to nonrelativistic motion transverse to the average direction of the line ( @xmath288 ) see ( [ eq - beta ] ) .
we then use the fact that an appropriate vector field can be specified by its divergence and curl@xcite to find@xcite that the phase field is completely specified by its behavior on the centerline @xmath289 evaluating ( [ eq - pstx ] ) we simply obtain ( [ eq - psx ] ) , the static result . but this is valid for the moving line as long as relativistic effects of the transverse motion can be neglected .
this indicates that the whole transverse cross section of the line moves together as an object with finite size .
this type of motion is required by the quantization condition ( [ eq - intpseh ] ) and by the nature of the superconducting line .
this type of motion is theoretically convenient because the lagrangian density for the filamentary line with interactions ( [ eq - lagrangiandwint ] ) can be used to derive the equations of motion for the superconducting line with only minor modifications .
there is no need to consider added degrees of freedom for the transverse size of the line .
the state of the entire line can still be completely specified as a function of only two parameters , which can still be taken as @xmath200 and @xmath199 .
now @xmath290 refers to the position in time and space of a point on the centerline of the line . because this model is _ acausal
_ , it is also _ nonlocal _ ; the response at a point is affected by a field over an extended region of spacetime .
( see section [ ssec - selfintline ] . )
the nonlocality of this model makes even self interactions finite , and indeed self interactions are an important part of the model . in order to consider the superconducting line , we generalize the filamentary model of ( [ eq - filemdual]-[eq - fil4current ] ) .
for instance the electromagnetic field ( [ eq - filem ] ) is a rank 2 tensor in 4 dimensions , because @xmath291 ( and @xmath292 ) are position 4-vectors , the @xmath293-function is lorentz invariant , and the integral formalism shown is independent of the type of parameterization as long as @xmath199 covers the full length of the line , and @xmath200 covers the full time under consideration .
we must modify the @xmath293-function to give the line a finite transverse size .
prescriptions for doing this are frame dependent , so we must start in the transverse rest frame of an element of the line . then we can obtain more general results via lorentz transformation to the lab . we evaluate ( [ eq - filem ] ) in the transverse rest frame of an element of the superconducting line by making the replacement @xmath294 where @xmath295 approaches a dirac @xmath293-function in the dimensions transverse to the line as @xmath296 . here
@xmath297 is the perpendicular vector from the centerline to point @xmath298 and this perpendicular intersects the centerline at @xmath299 .
( if there is more than one such point , we must add the contributions in order to find the total field . ) in the transverse rest frame @xmath300 @xmath301 where both of these quantities are 4-vectors because @xmath302 and @xmath303 represent the proper time and a length element transverse to @xmath304 in the rest frame of an element of the line .
hence the electromagnetic field in the transverse rest frame of this element of the string is given by @xmath305 @xmath306 @xmath307 @xmath308 where the matrix contains components of the vector @xmath309 giving the direction of the centerline in the transverse rest frame of the line at the point @xmath310 .
so the electromagnetic field in this frame is given by @xmath311
@xmath312 this looks like the point at which we started ( [ eq - bline ] ) , but now the unit vector @xmath313 does not necessarily point along the @xmath243 axis and its direction can change as a function of @xmath200 and @xmath199 . by performing a tranverse lorentz transformation from the rest frame of the above element of the line to the lab we find the electromagnetic field of the line ( for the line ( [ eq - yzline ] ) moving in the @xmath314 plane ) in the lab @xmath315 @xmath316 where @xmath317 is the result of lorentz transformation of @xmath318 , selecting events simultaneous in the lab to describe the moving field . the extra @xmath319 factors arising from the transformation of the field enable @xmath317 to take the direction of the field when expressed in lab coordinates .
in fact the net result of the lorentz transformation , expressed in lab coordinates is the change @xmath320 .
since @xmath318 and @xmath317 are unit vectors , they are , in fact , equal . the directions of the line in its rest frame and the lab are the same . here
we use @xmath321 @xmath322 and @xmath323 is the pancaking cylindrical radius of the line expressed in lab coordinates .
by our use of a lorentz transformation we are ignoring the fact that the rest frame of the element of the line is an accelerating reference frame . according to meissner , thorne , and wheeler@xcite this approximation , using a comoving inertial frame , is permissible as long as inertial forces are small and typical accelerations are much less than @xmath324 where @xmath325 is the amplitude of a typical oscillation .
we will assume that these conditions hold true except near the ends of a line .
we wish to use maxwell s equations to find the supercurrent in the lab .
maxwell s equations act at the _ field _
point @xmath41 and therefore we do not consider the explicit space and time dependence of the string coordinate @xmath38 , which acts effectively as a source point .
neither do we consider the explicit space and time dependence of @xmath319 and @xmath326 which is determined by @xmath38 .
however , the field at * x * can change with time due to motion of the string . since the electromagnetic field does not change ( is _ quantized _ ) in the rest frame of the string , all time dependence is due to motion .
so at a fixed field point @xmath41 , the change of the magnetic induction field with time is given by @xmath327 @xmath328 so for the line ( [ eq - yzline ] ) moving in the @xmath314 plane @xmath329 here @xmath304 is the velocity of the line at the field point @xmath41 which is the same as the velocity of the point on the centerline @xmath38 associated with @xmath41 , since the cross section of the line moves together .
for similar reasons the change of the electric field with time is given by @xmath330 and if the line satisfies ( [ eq - yzline ] ) @xmath331 using maxwell s equations we find @xmath332 where @xmath176 is the instantaneous radius of curvature at @xmath333 .
this current satisfies the equation of continuity ( [ eq - continuity ] ) exactly .
details of these calculations are included in appendix [ ssec - instradius ] , where the derivatives of @xmath40 are explicitly evaluated .
it is also shown in appendix [ ssec - calcemfield ] that if we apply equation ( [ eq - generalsol ] ) and attempt to derive the electromagnetic field ( [ eq - efieldline ] ) and ( [ eq - bfieldline ] ) from ( [ eq - supercurrentyz ] ) we do not get precise agreement . we get the expected result plus remainder terms proportional to @xmath334 and @xmath335 which are small for large radii of curvature .
we take this to imply that our simple procedure , which involves giving the delta function a finite size and using a lorentz transfomation from a comoving inertial frame , does not provide an exact solution for the electromagnetic field .
it will be sufficient , however , for our present purposes . performing an inverse lorentz transformation we find the supercurrent in the rest frame of an element of the line : @xmath336
the parameters @xmath319 and @xmath326 enter this expression because we have defined @xmath176 in the lab .
( see appendix [ ssec - instradius ] ) the charge density is zero in the rest frame .
it is shown in appendix [ ssec - energyofbending ] that the energy of the finite size line is independent of the instantaneous radius of curvature @xmath176 at a given point .
this fact is important because we do not include any term in the lagrangian density for the energy of bending .
so in the absence of interactions , general wave solutions of the type ( [ eq - yzline ] ) and ( [ eq - circularline ] ) traveling in the positive @xmath106-direction could occur with the superconducting line . replacing the argument @xmath337 with @xmath338 produces solutions traveling in the negative @xmath106-direction .
following the sign conventions of bjorken and drell@xcite ( which are consistent with jackson ) the action ( except for weak and hadronic modes ) is @xmath339 where no mass term appears in the lagrangian because we assert that mass is not fundamental . here @xmath340 is the mechanical string lagrangian density ( [ eq - lagrangiandensity ] ) for line @xmath341 , @xmath342 is the electromagnetic current due to electric charges , and @xmath343 is the electromagnetic current due to superconducting lines . applying the principle of least action to the superconducting line with finite size , using ( [ eq - deltatrans ] ) in ( [ eq - fil4current ] ) , it can be shown that the equation of motion for a given line , instead of ( [ eq - filmotionwint ] ) , is given in the case of interactions by @xmath344 where @xmath345 is the average current over the cross section of the line weighted by the shape of the line field .
this averaging must be performed in the transverse rest frame of the line .
but the resulting current is a 4-vector , and can be transformed to any frame ( within the approximations which have been discussed@xcite ) . indeed , since the transverse rest frame varies with the line element in question for solutions of the type ( [ eq - yzline ] ) , it is most convenient for these solutions to think of the quantities @xmath346 , @xmath347 , and @xmath348 transformed from the rest frame to the lab
. then if we wish to make a change of parameterization to the particular parameterization of section [ ssec - neglecting ] , we can do so .
( it is only important that the parameterization be the same on both sides all equations . ) for a given line we work in the large radius of curvature approximation , @xmath349 , where @xmath350 is the radius of curvature of the centerline
so that all points which make a significant contribution to the average are perpendicular to the line at a single unique point ( always true for points closer to the line than @xmath350 ) .
to evaluate the interaction of the line with itself , we use equation ( [ eq - supercurrentyz * ] ) in the interaction term and obtain ( with @xmath351 ) @xmath352 @xmath353 where we have used the fact that the radius of curvature of the centerline can be related to the shape of the trajectory ( see appendix [ ssec - radiusandshape ] ) according to the equation @xmath354 in order to learn how to transform the radius of curvature from the lab to the line s rest frame @xmath355 because of azimuthal symmetry the first term in ( [ eq - selfintline ] ) gives a result of zero . unless the radius of curvature is infinite , the second term in ( [ eq - selfintline ] ) is easily seen to be finite but not equal to zero .
in fact if @xmath356 , we obtain @xmath357 hence we have evaluated the self interaction of the superconducting line .
it is finite and calculable in spite of the fact that the field diverges along the centerline .
the averaging required by the principle of least action leads to a finite result . because the self - interaction term in equation ( [ eq - eqofmotionwint ] ) is not zero , and traveling waves are an exact solution of this equation without interactions , there are _ no _ traveling wave solutions propagating at the velocity of light for the motion of a single isolated superconducting line once self - interactions are included .
wave motion of a single superconducting line may still be possible when the photon quantization condition is satisfied
but then the line is not isolated , and we are currently unable to describe the motion .
now we are ready to define what we mean by the wave function of the electron in this model which plays such a prominent role in the dirac quantization condition ( [ eq - singv ] ) .
we define the electron wave function to be the change in the ether caused by the bare electron , i.e. by the end of the appropriate line .
this change will be the response of the ether to the bare electron s electromagnetic field .
( see section [ ssec - masselectron ] . )
since the bare electron s electromagnetic field has infinite extent , the wave function of the electron also has infinite extent as in standard quantum theory , and consistent with our reasoning in section [ ssec - fq ] . on the microscopic scale , this wave function is not continuous ( differing from the standard model ) .
it is a mesh of line pairs .
each line pair is continuous .
we should realize that the cross sections of our line pairs are small on the atomic scale and mostly fit into the holes in matter .
the outer radius of a line pair ( section [ ssec - static ] ) must be much smaller than the bohr radius of the hydrogen atom : @xmath358 .
hence there are many holes in the wave function which constitute a larger version of dirac s nodal lines " .
if three of the line pairs form a complete circuit around such a hole , and if magnetic field links the circuit , the circuit will superconduct in response to the vector potential ( figure [ fig - fluxlhalf ] ) , causing quantization of magnetic flux ( [ eq - fq ] ) through the circuit .
if the applied flux through the complete circuit is less than half a flux quantum , the net flux will be brought to zero .
if the applied flux through the circuit is greater than half a flux quantum , experimental evidence@xcite indicates that the supercurrent in the complete circuit will _ increase _ the flux as necessary to match the nearest flux quantum ( figure [ fig - fluxghalf ] ) .
we will see that the photon can be considered as an excitation of the wave function of the electron . to 2.6 in to 2.6 in
we seek the types of waves which can propagate at the speed of light on the two lines of an isolated pair .
we attempt to find them using the known solutions ( [ eq - yzline ] ) and ( [ eq - circularline ] ) without the interaction terms . from the principle of least action for the twoline system
we find @xmath359 @xmath360 \delta { y^\mu_1}\ ] ] @xmath361 \delta { y^\mu_2 } \biggr \ } \ \ \ .\ ] ] due to the interaction terms the variations in @xmath362 and @xmath363 are coupled so we can not set the coefficients of @xmath364 and @xmath365 separately to zero , without first knowing the solutions to the equations of motion .
to handle this situation , we define the 4-vectors @xmath366 as the average and half the difference between the wave solutions of line 1 and line 2 .
then we find @xmath367 we express the mechanical string lagrangian densities ( [ eq - lagrangiandensity ] ) for line 1 and line 2 in terms of @xmath368 and @xmath325 : @xmath369 @xmath370 to find the common mode solutions we set @xmath371 where @xmath372 is a constant 4-vector . note that @xmath325 has no time component and no @xmath106-component by construction . thus @xmath373 and the twoline system exhibits no excitation .
it moves together as one system . since the common mode solutions exhibit
no excitations , and hence no net electromagnetic field at a given @xmath374 , we interpret them as gravitational waves .
( see section [ sec - generalrel ] . ) since @xmath325 is constant , @xmath375 and @xmath376 , hence the mechanical lagrangian density of the system becomes @xmath377 just twice the lagrangian density ( [ eq - lagrangiandensity ] ) .
since for common mode solutions @xmath378 the variation of the action becomes @xmath359 @xmath379\ ] ] @xmath380 and the equation of motion becomes @xmath381 @xmath382 where @xmath383 and @xmath368 must satisfy equations ( [ eq - yd ] ) and ( [ eq - ydif ] ) , and the left hand side of the equation refers to @xmath368 .
we label the 8 interaction terms @xmath384 where @xmath385 or 2 corresponding to the line where the average is performed , @xmath386 or 2 corresponding to the line field which is averaged , and @xmath387 or 0 corresponding to the term in the current ( containing either a @xmath66 or a @xmath63 , see ( [ eq - supercurrentyz * ] ) ) . for the common mode solutions , line 1 and line 2
have the _ same _ instantaneous rest frame , which simplifies the averaging calculation . as in section [ ssec - selfintline ] , we find that @xmath388 . in order to obtain wave solutions with the remaining interaction terms we specify the constant 4-vector to be in the @xmath389-direction @xmath390 ( labeling the constant length also @xmath372 ) if the motion is in the @xmath314 plane . note that @xmath391 since the fields are in opposite directions .
( see figure [ fig - commonmodegeom ] . ) using the large radius of curvature approximation and equations ( [ eq - k0fact ] ) , ( [ eq - wronskian ] ) and the derivative of ( [ eq - wronskian ] ) we find @xmath392 where the normalized separation of the two lines in the @xmath389 direction is given by @xmath393 ( we quote the covariant components @xmath394 , so the interaction is _
attractive_. ) as expected , this interaction term is zero if @xmath395 and attractive for all @xmath396 greater than 0 . from rotational symmetry
we see from figure [ fig - commonmodegeom ] that @xmath397 to 1.8 in plane.,title="fig : " ] after taking advantage of the fact that for this common mode geometry the instantaneous rest frames and instantaneous radii of curvature of the two lines are the same if the separation is perpendicular to the motion , we find that @xmath398 if @xmath399 .
since @xmath400 for common mode solutions , we find that @xmath401 and the sum of all interaction terms for this common mode geometry is zero .
hence , before considering the quantization conditions , common mode solutions gravitational waves of types ( [ eq - yzline ] ) can propagate freely at the speed of light if @xmath402 , the direction of the 3-vector associated with @xmath372 , is perpendicular to the motion .
it is shown in appendix [ sec - generalmotionincommonmode ] that the orientation of @xmath402 does not affect common mode motion .
so common mode solutions of type ( [ eq - circularline ] ) can also propagate at the speed of light .
we will see in section [ ssec - generalityquant ] that the photon quantization conditions must still be satisfied , even in this case .
but the quantization currents have no role beyond their role in keeping the separation @xmath372 constant .
( see section [ ssec - tensionpair ] . ) since the orientation of @xmath402 does not matter in the common mode , and the sum of the interaction terms is zero in common mode , electromagnetic waves with @xmath402 spinning ( helical polarization ) can propagate at the speed of light on top of gravitational waves as long as the photon quantization conditions are satisfied . to find differential mode solutions we set @xmath403 without variation , and therefore @xmath404 we consider the case where @xmath325 has only transverse components ( unlike appendix [ sec - generalmotionincommonmode ] ) so @xmath325 , @xmath405 and @xmath406 are all orthogonal to @xmath368 , @xmath407 and @xmath408 .
the total mechanical lagrangian density becomes @xmath409 if we simply define @xmath410 the mechanical lagrangian density becomes @xmath411 with the standard solutions of the form ( [ eq - yzline ] ) and ( [ eq - circularline ] ) without interaction .
we refer to this differential solution as @xmath412 ( rather than @xmath362 ) to indicate generality .
both @xmath362 and @xmath363 are solutions to this mechanical lagrangian without interaction .
so for the differential mode @xmath413 we require @xmath414 and therefore @xmath415 the variation of the action becomes @xmath359 @xmath416 ( + \delta d^\mu)\ ] ] @xmath417(- \delta d^\mu ) \biggr \}\ ] ] @xmath418\ ] ] @xmath419 and we see that the differential mode solutions involve the _ difference _ between the interaction terms of the two lines .
the equation of motion for differential mode solutions becomes : @xmath381 @xmath420 where @xmath421 and @xmath422 must satisfy ( [ eq - y1dy2d ] ) and the left hand side of the equation refers to @xmath412 ( hence either @xmath421 or @xmath422 ) . _
a. radial excitation linear polarization _ [ ssssec - radialexcitation ] now we let @xmath423 in order to look for differential solutions of the type ( [ eq - yzline ] ) .
we consider the moment in space and time shown in figure [ fig - radialexcitation ] when the two lines reach maximum separation . at this moment
the two lines will be at rest in the lab , hence they have common mode geometry in their joint rest frame .
accordingly we can use ( [ eq - i121 ] ) and ( [ eq - i110pi120 ] ) to calculate the interaction terms for differential mode waves with radial excitation @xmath424 @xmath425 @xmath426 @xmath427 at the distance @xmath428 of maximum separation ( with @xmath429 ) .
to 2.6 in we keep the sign conventions of the common mode interaction terms for clarity .
since we must take the _ difference _ of these terms , it is very easy to see that the differential mode interaction does not vanish for the case of radial excitation .
although we have only considered the point of maximum separation , clearly the nonvanishing of the interaction terms is general throughout the motion with radial excitation .
the interaction terms ( [ eq - i121m211 ] ) and ( [ eq - i110p120m210m220 ] ) point in the @xmath430 direction , but this is no longer true away from maximum separation since @xmath431 and @xmath432 are _ not _ in opposite directions .
since motion with radial excitation is in the same direction as nonzero interaction terms , we can not model differential mode waves with radial excitation propagating at the speed of light .
if the photon quantization conditions are satisfied , such motion could still occur , but a model would require knowledge of the quantization currents which we currently do not understand .
_ b. model of the photon
helical polarization _ [ ssssec - helicalphoton ] to model the photon we use solutions of the type ( [ eq - circularline ] ) with sinusoidal functions @xmath248 and @xmath229 , i.e. one fourier component , where @xmath433 is assumed constant .
@xmath434 @xmath435 here @xmath436 is the direction in the rotating frame ( which is the rest frame of _ both _ lines along their entire length ) , which is initially vertical ( @xmath437 direction ) in the lab , @xmath438 and the wave has a frequency @xmath439 . with this choice
we can achieve lack of radial excitation since @xmath440 ( see figure [ fig - helicalphoton ] ) .
the whole system rotates with positive helicity .
the velocities are given by to 2.6 in , @xmath441.,title="fig : " ] @xmath442 @xmath443 and the magnetic field directions are given by the following expressions : @xmath444 @xmath445 since both lines are at rest in the rotating frame , we calculate the currents in this frame directly .
we use ( [ eq - eline*]-[eq - bline * ] ) for the electromagnetic fields of both lines .
then from appendix [ ssec - generalcurl ] and [ ssec - helicalmodegeom ] we find the following expressions : @xmath446 @xmath447 @xmath448 where @xmath449 is the instantaneous radius of curvature of the centerline , which is constant for both lines .
hence we find : @xmath450\ ] ] @xmath451\ ] ] these currents satisfy the equation of continuity ( [ eq - continuity ] ) exactly , but we still find a small inconsistency if we try to use ( [ eq - generalsol ] ) to recalculate the electromagnetic field ( see appendix [ ssec - calcemfield ] ) .
we find the interaction terms in helical mode @xmath452\ ] ] where we take into account that the radii of curvature for line 1 and line 2 point in opposite directions .
( hence @xmath453 and @xmath454 are different functions of position though the magnitude of @xmath455 is the same on both centerlines . )
we make a correction for the fact that @xmath456 and @xmath457 have different axes by changing @xmath458 into @xmath459 where @xmath460 is the cylindrical radius perpendicular to the @xmath243-axis , and symmetric for both bessel functions . we obtain @xmath461 where @xmath462 and in a similar fashion @xmath463 so the sum of the interaction terms ( including the negative signs for differential mode ) @xmath464 @xmath465 is in the radial @xmath430 direction and attractive for the helical differential mode of the photon . since the motion of the helical photon is orthogonal to the radial direction , it can propagate freely at the velocity of light .
the quantization currents must play a particularly simple role .
the left hand side of ( [ eq - eqofmdiff ] ) is zero in the radial direction , but the right hand side of ( [ eq - eqofmdiff ] ) , given by ( [ eq - inttermsdiff ] ) , is not zero in the radial direction . the quantization currents ( which have not been considered so far ) must freeze the radial separation of the two lines of the photon at @xmath466 .
this allows us to isolate the effects of the quantization currents , which we do not fully understand , and model the helical photon . in figure [ fig - wfelectron ]
we consider the wave function of an electron with 4-momentum @xmath467 .
the wave function is a mesh of line pairs , each of which ( from quantum mechanics , as we will see in section [ sec - quantmech ] ) oscillates at frequency @xmath468 and with a wavelength @xmath469 if @xmath470 points in the direction of the line pair and @xmath471 in the more general case , where @xmath472 is the direction of the line pair .
we use the absolute value sign to indicate that no positive sense for @xmath472 is defined .
hence @xmath473 is the phase field of this wave function . to 1.5 in
we have pictured the mesh as real and drawn it with regular spacing for clarity . in general , of couse , a wave function is complex . in reality
neither the spacing nor the directions will be regular .
closed superconducting circuits will be formed with many shapes ( triangles , rectangles , pentagons , ... ) and sizes , and in many different orientations . in each circuit
@xmath474 since @xmath475 is the phase field for this ( single - valued ) wave function .
( we are using a plane - wave electron for simplicity .
the fourier composition of wave functions will be considered in section [ sec - quantmech ] . )
we see that the phase field of an electron wave is more than just the electron s 4-momentum .
it represents the entire wave function of a plane - wave electron .
( it is probably useful to think of it as a differential form@xcite . ) but now consider figure [ fig - photonfixedt ] .
a photon is propagating on one of the line pairs , hence on only one leg of one or more closed circuits . to 1.5 in for _ this _
line pair , the number of oscillations per unit length has increased and so the phase field is @xmath476 the sum of the 4-momenta of the electron and the photon . hence , since this photon is propagating along only one of the line pairs , we must have ( from ( [ eq - multisup ] ) ) @xmath477 so that the photon s momentum and energy are quantized . for the electron s wave function to be single valued in the presence of the photon , ( [ eq - multisup ] ) must hold .
thus ( [ eq - multisup ] ) must hold for _ all _ quanta , not just flux quanta .
consider a circuit at _ fixed time _
( figure [ fig - photonfixedt ] ) . with @xmath478 ( and hence the direction of the relevant line pair )
defined to be in the @xmath243-direction , equation ( [ eq - pepg ] ) becomes @xmath479 so that @xmath480 we presume that the length of one photon is the wavelength @xmath481 , since it is the smallest length satisfying periodic boundary conditions bringing the two lines back to their quiescent condition .
consider a circuit at _ fixed z _
( figure [ fig - photonfixedz ] , with @xmath478 still in the @xmath243-direction ) to 1.5 in .,title="fig : " ] @xmath482 where we presume the photon passes in a time interval of one period @xmath483 ( the smallest time duration satisfying periodic boundary conditions ) . so @xmath484 and the photon must be _ quantized _ if it is to propagate on a line pair through the superconducting ether . to consider whether or not a photon can exist we must integrate over the relevant line pair for at least a wavelength @xmath481 or a period of time @xmath72 and close the integral on nearby line pairs without excitation .
if the quantization condition ( [ eq - pepg ] ) is satisfied , apparently the constant radial separation of the two lines of the helical photon is maintained at @xmath466 .
this explains how a sine wave can start and stop instantly .
the two lines of the photon are in general _ not isolated _ if the quantization condition ( [ eq - pepg ] ) is satisfied .
the situation is similar to figure [ fig - fluxghalf ] , where currents in the ether maintain a flux quantum .
we observe from equations ( [ eq - pepg ] ) and ( [ eq - multisup ] ) that the quantization of the photon really does not depend on the photon being an excitation of any particular electron s wave function .
it is a property of the mesh of superconducting lines that comprises the ether and forms the wave functions of all electrons .
even though we will show that the transverse motion of the helical photon is nonrelativistic , we attempt to make no approximations ( beyond the large radius of curvature approximation , which is well satisfied ) , because we will obtain precise results . we perform separate local lorentz transforms ( using ( [ eq - beta1 ] ) or ( [ eq - beta2 ] ) as appropriate ) on the electromagnetic field ( [ eq - eline*]-[eq - bline * ] ) in the rotating frame to find the electromagnetic fields of line 1 and line 2 in the lab : @xmath485 @xmath486 @xmath487 @xmath488 @xmath489 the velocity of the entire cross section of each line is the velocity of its centerline in accord with section [ ssec - causality ] .
we note that the transverse components of * e * and * b * are equal for both lines and the transverse fields of the two lines add but the net field in the @xmath243-direction is zero .
the poynting vector of the helical photon , expressing the flow of energy per unit area per unit time , is given by @xmath490 @xmath491 - { \hat x}^\ast \bigl [ b^{\ast}(1)^2 - b^{\ast}(2)^2 \bigr ] \biggr \}\ ] ] which shows that some energy is directed along the @xmath243-axis and some energy circulates about the @xmath243-axis . at fixed @xmath106 ,
the energy flow per unit time is given by @xmath492 with @xmath493 and is totally in the @xmath243-direction .
( as for ( [ eq - tiltbessel ] ) we take into account the tilt of both bessel functions with respect to the @xmath243-axis . )
we wish to find the energy in one photon , @xmath494 .
it should be given by the energy passing a fixed @xmath106 in one period .
@xmath495 @xmath496 if we want to compare with measured photon energies we must make a _ binding energy correction _ , because when the line pair transfers the photon energy to an atom , the photon must provide the energy necessary to separate the lines , because one line will go to an electron and the other to the nucleus .
so we see that the overlap term containing @xmath497 clearly does not belong in the energy transferred to an atom .
we assert that the corrected energy @xmath498 transferred to an atom is just ( [ eq - egwoverlap ] ) without the overlap term , because it gives the sum of the electromagnetic energies transferred by each line separately .
we presume that in the process of the lines separating , angular momentum about the common mode centerline is conserved , because the force between the lines is central between the centers .
so we can check our method by calculating the angular momentum about the common mode center i.e .
the spin of the photon .
this binding energy correction is approximately a factor of two decrease in the energy . using similar reasoning we can also neglect the effects of the photon quantization currents on the energy of the traveling photon because this energy is not transferred to the absorbing atom .
quantization current effects are also part of the photon binding energy . using the above correction
we find @xmath499 so @xmath500 we will show that when the large radius of curvature approximation is satisfied @xmath501 is very close to @xmath114 .
_ hence the quantization of the helical photon implies that the amplitude of the photon wave is an absolute constant ! _ if we assume that @xmath502 near the earth ( as we will infer from section [ ssec - parameters ] ) , we have @xmath503 since the line radius @xmath61 would give a finite size to the electron , from section [ sec - introduction ] we have @xmath504 .
so for a green photon ( @xmath505 ) we have @xmath506 @xmath507 in addition we find @xmath508 so @xmath509 for helical geometry the large radius of curvature approximation @xmath510 is well justified , for a visible photon . in figures [ fig - currentsfortogether ] and [ fig - currentsforapart ]
, current distributions are shown with which the quantization condition could be maintained . in figure
[ fig - currentsfortogether ] currents running on the outside surfaces of the line pairs tend to decrease the radial separation of the central pair , and in figure [ fig - currentsforapart ] currents are shown which tend to separate the central pair .
the dominant contributions are likely to come from the currents running on the outer surfaces of the central pair .
currents in the same direction attract .
currents in the opposite direction repel . hence in their simplest configuration
, the photon quantization currents could form a loop around the outer surfaces of the central line pair for the length of the photon ( one wavelength ) .
the angular momentum about a given point on the common mode centerline of the pair is given by @xmath511 so at fixed @xmath106 the angular momentum per unit length is @xmath512 perhaps the easiest was to evaluate ( [ eq - dlemdz ] ) is to realize that in taking moments of momentum , the entire momentum of each line can be considered to act at its centerline ( center of momentum ) .
so , for example , for a term involving @xmath513 we have @xmath514 @xmath515 \times ( -{\hat x}^\ast ) \ ] ] @xmath516 \times ( -{\hat x}^\ast ) \ ] ] @xmath517 \times ( -{\hat x}^\ast ) \ ] ] @xmath518 therefore the spin of the photon is given by the angular momentum in one wavelength : @xmath519 @xmath520 assuming the binding energy correction is valid , this result is exact except for the the large radius of curvature approximation ( [ eq - largeroc ] ) .
figures [ fig - wfelectron ] , [ fig - photonfixedt ] , and [ fig - photonfixedz ] are simplifications .
we will see ( section [ ssec - motionbpp ] ) that the typical spacing of line pairs is less than @xmath521 , whereas the wavelength of a green photon is @xmath522 .
thus the quantization of the photon involves many closed circuits . from figures [ fig - currentsfortogether ] and [ fig - currentsforapart ] , the currents which maintain the quantization of the helical photon , i.e. the constancy of the photon amplitude , can be complex and , as with figure [ fig - fluxghalf ]
, we do not have a detailed understanding , beyond the quantization condition itself , of what makes the currents flow as they do .
but apparently such flow is required for single - valued wave functions , and so it happens .
we have made progress with our classical picture but are now meeting difficulty in understanding the quantization condition itself . to 1.5 in to 1.5 in an atom
generates each photon on the relevant line pair , which acts as a transmission line .
once the photon exists and satisfies the quantization condition ( [ eq - pepg ] ) , apparently it is maintained by the quantization currents as it transits the universe until it is absorbed by another atom , eventually .
apparently these photon quantization currents persist forever , like currents in a superconducting loop , which is apparently what they are .
we see from the fact that ( [ eq - pepg ] ) and ( [ eq - multisup ] ) are identical , that this quantization condition is a completely general result applying to the phase field of any particle traveling in the superconducting ether .
the ether is a multiply connected superconductor . in particular this quantization condition must hold for any excitation of a line pair .
it is _ not _ restricted to particles traveling at the speed of light , and must hold for matter waves and gravitational waves , as well as electromagnetic waves .
we can now identify a matter wave as a gravitational ( i.e. common mode ) wave which is stimulated by a particle , so that it does not travel at the velocity of light , but rather travels with the particle .
it is the particle s wave function . in the case that there are no excitations of the line pairs under consideration we have @xmath523 for any closed loop at a fixed time .
if there were a change in the radial separation of any line pair in the circuit , this change would necessarily be quantized , contrary to our assumption .
hence the radial separation in quiescent common mode and the tension of a line pair @xmath524\ + \ \tau_{qc } \label{eq - tensiontau0}\ ] ] must be constant _ for the whole universe _ at a given time . here
@xmath525 is the tension ( energy per longitudinal length ) associated with the quantization currents for this line pair .
it is convenient to think of the quantization currents in this case as flowing in opposite directions on the outer surfaces of each pair .
these currents are likely to be the dominant quantization currents .
we assume that they also persist forever .
the radial separation , the line - pair tension and the quantization currents , however , must change with time as the universe expands , reflecting the conservation of energy . apparently this line pair tension and radial separation is a measure of world time@xcite .
we use dirac s pole current formalism@xcite to find the electromagnetic field of the poles at the ends of a line . with our line of finite size
the dual field of the line ( [ eq - filemdual ] ) becomes @xmath526 with from ( [ eq - deltatrans ] ) @xmath527 the pole current is defined by the extension of maxwell s equations if magnetic poles were to really exist : @xmath528 @xmath529 @xmath530 after an integration by parts we obtain the pole current @xmath531 if we presume that @xmath221 and @xmath222 are sufficiently distant that they are of no concern , the pole current becomes @xmath532 @xmath533 we get our expectation for a pole current from each end of the line of finite size _ except _ for the signs of both poles .
the reason is that we actually cut off the integral over the line too soon .
the dual field actually extends to infinity beyond the ends of our line .
so by cutting off the dual field at the mechanical end of the line , we are _ absorbing _ the dual field field at the end of the line .
this is the reason for the incorrect signs .
contrary to our view dirac , excluded the line from physical reality because he wanted the pole field to appear as a monopole , in the space _ external _ to the line .
therefore we take our dual field as the _ negative _ of dirac s dual field . in our model maxwell s equations
are satisfied everywhere , and the pole currrent is actually @xmath64 .
@xmath534 but we can use dirac s pole current method to find the effective pole current outside the ends of our lines just by reversing the signs of the pole currents : @xmath535 the field which appears to be absorbed in ( [ eq - dualcurrent ] ) , because we cut off the integral too soon , actually exits the line and spreads out in all directions .
we assume , without justification , that there is simply a sharp , clean break at the end of the line the superconducting field and the supercurrent simply stop .
the implications of this simple phenomenological assumption are rich and interesting . using maxwell s equations and ( [ eq - generalsol ] )
we can show that in the superconducting field : @xmath536 @xmath537 therefore the electromagnetic field at the break at the end of a line should act as a source for the electromagnetic field beyond the line , which satisfies the homogeneous wave equation .
we will see that the bare positron - pole ( using positive charge for convenience ) bounces around ( _ zitterbeweging _ ) so rapidly that a proper treatment of the shape of the field of the bare positron - pole is unnecessary , as long as the total outward flux is correct . in the limit @xmath296
we simply find @xmath538 the current of point poles .
therefore we can use the standard lienard - wiechert potentials@xcite with the substitutions @xmath539 to find the electromagnetic field of the bare positron - pole @xmath540_{ret } + \ { g_0 } \
\biggl [ { { { \hat r } \times \bigl \ { ( { \hat r}-{\bm \beta } ) \times { \dot { \bm \beta } } \bigr \ } } \over { \kappa^3 r } } \biggr ] _ { ret } \label{eq - bppbfield}\ ] ] @xmath541_{ret } \label{eq - bppefield}\ ] ] with @xmath542 and where the expressions in brackets are evaluated at the retarded time @xmath543 and @xmath544 @xmath545 ( the bare positron - pole will acquire it s electric charge in section [ ssec - electroncharge ] and then be called the bare positron . ) the field of the bare electron - pole is similar , coming from the negative pole in ( [ eq - poles ] ) the line field is , of course , present , but it is quantized and shielded by a superconducting sheath , so we can often neglect its interaction ( section [ ssec - intercommutation ] ) with the surrounding line pairs . in section
[ ssec - neglecting ] , we saw that without considering interactions , the end of a line moves at the speed of light transverse to the line .
it is interesting to note that , if the bare positron - pole moves at the speed of light , it s electromagnetic field ( [ eq - bppbfield]-[eq - bppefield ] ) is well behaved .
in fact , we can see that in the absence of acceleration ( @xmath546 ) , the electromagnetic field of the bare positron - pole is actually @xmath64 at the speed of light . in this universe
, however , electromagnetic fields are always present .
we will see that in the presence of external fields the speed of the bare positron - pole is less than the speed of light but , unless the bare positron - pole is in a _ very _ strong field , its speed will be close to the speed of light .
so the natural state of the end of a line is to move near the speed of light transverse to the line .
this is in great contrast to our concept of a point particle whose natural state is to remain at rest .
evidently , this is because the end of a line carries no inertia . to solve the equations of motion for the bare positron - pole , we need to find the average of the line s dual field ( [ eq - lineemdual ] ) over the cross section of the line at the end points . in the rest frame of the bare positron - pole
we find @xmath547 since @xmath548 is the proper time , the vector @xmath549 is a 4-vector .
after transformation to the lab , the transformed vector is @xmath550 remembering our definition of @xmath200 ( see ( [ eq - taudef ] ) ) .
hence in the lab we find @xmath551 in order to deal with the end of a line we _ pick _ the directions defined by @xmath552 and @xmath199 to be transverse to the motion .
( so they are no longer necessarily in the @xmath243 direction . ) then using the fact that @xmath553 and @xmath554 are both 4-vectors , we find @xmath555 and so @xmath556 in the presence of interactions , the boundary conditions ( [ eq - bcendofline ] ) which must be satisfied at the ends of a line become @xmath557 and since @xmath219 is in general not zero , we obtain @xmath558 or at each end point @xmath559 where @xmath560 is given by ( [ eq - pldymup ] ) .
the self interaction term is simply twice the mechanical term , so the equation of motion for _ either _ end point becomes @xmath561 where @xmath562 is a very strong field ( 3/2 times the average of the self field ) .
we define @xmath563 as the direction of the component of the electric field which is perpendicular to the line ( @xmath317 ) @xmath564 and we also define @xmath565 we assume that the velocity of the bare positron - pole is perpendicular to @xmath563 @xmath566 so the equations of motion become @xmath567 @xmath568 @xmath569 @xmath570 usually we can neglect the averaging requirement for the external electromagnetic field , if it does not vary too rapidly over the cross section of the line .
we use a simplified notation without averaging , but , in principle , we always have to take the average in the rest frame of the bare positron - pole and transform the resulting field to the lab ( as we did in section [ ssec - selfintbareelectronpole ] ) .
we first consider the case where the electric field is zero . in this case
we find from ( [ eq - bpplmotion ] ) @xmath571
so it is obvious in this case that @xmath572 must be negative , i.e. the line orients both of its ends so that the external magnetic field is opposite the magnetic field of the line , as we would expect from energy considerations . from ( [ eq - bpptmotion ] ) and ( [ eq - bpppmotion ] ) we find that @xmath244 is proportional to @xmath573 .
@xmath574 hence we can solve both ( [ eq - bpptmotion ] ) and ( [ eq - bpppmotion ] ) by presuming that the ends of the line both align @xmath317 opposite the magnetic field so that @xmath575 .
next we consider the case where the electric field and the magnetic field are perpendicular as is often the case with electromagnetic waves and is true of the field of another moving pole ( [ eq - bppbfield]-[eq - bppefield ] ) . in this case
we can solve the equations of motion with the following presumptions : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \1 ) at each end of the line , @xmath317 aligns opposite the external magnetic field .
\2 ) the external electric field defines the direction of @xmath563 , perpendicular to @xmath317 ( @xmath576 ) .
\3 ) then @xmath226 and the velocity of the bare positron - pole is in the direction @xmath577 ( so @xmath578 ) perpendicular to the line , and the value of @xmath579 can be found from ( [ eq - bpplmotion ] ) , \4 ) equation ( [ eq - bpp3motion ] ) is satisfied because @xmath580 and @xmath581 are both zero . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ so we have the solution for the motion of the bare positron - pole in the case that the external electric and magnetic induction fields are perpendicular .
the bare positron - pole carries no inertia so it responds to the field conditions instantaneously , as it moves and the external fields change . in the general case where @xmath582 , the end of the line must orient itself such that the @xmath563 component of the lorentz force on the moving pole is @xmath64 .
this requires a nonzero value of @xmath580 . if @xmath583 , it is easy to find the appropriate orientation of @xmath317 if @xmath584 is small .
the general case requires the solution of a transcendental equation .
( see appendix [ ssec - orientation ] . ) in all cases we use ( [ eq - bpplmotion ] ) to find the velocity of the moving pole : @xmath585 and @xmath586 where the @xmath587 sign takes the same value in both equations .
we see from ( [ eq-1ogamp ] ) that we must have the @xmath588 sign because @xmath572 and @xmath589 are independent , i.e. @xmath572 could be zero .
hence the velocity of the bare positron - pole is given by @xmath590 @xmath591 using the minimal value of the line - pair density of the ether given by ( [ eq - dldvnum ] ) we can estimate @xmath158 , the typical distance between line pairs @xmath592 we find @xmath593 and hence @xmath594 is the typical time between collisions of the bare positron - pole with line pairs at the minimal density of ( [ eq - dldvnum ] ) . during each collision the external electromagnetic field changes radically in both magnitude and direction and the bare positron - pole travels near the velocity of light in a new direction .
the main causes of the external field are _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \(1 ) reflection of the pole field ( [ eq - bppbfield]-[eq - bppefield ] ) by the superconducting line pairs ( see ( [ eq - pealemac ] ) and ( [ eq - memalbmac ] ) ) , and \(2 ) currents in closed superconducting loops , stimulated by the pole magnetic field , bringing the magnetic flux linking each loop to zero because of flux quantization _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ thus the motion of the bare positron - pole is _ zitterbewegung _ , jittery random - walk motion with a very short time between collisions .
even though the bare positron - pole is traveling near the velocity of light , for many purposes it can be considered at rest in a frame traveling at a much smaller velocity .
contrary to the general rule ( section [ ssec - generalityquant ] ) this zitterbewegung motion is _ not _ quantized , because the end of a line ( electron ) is not part of a closed loop . at this point
we are forced to make an assumption which does not follow from the principle of least action applied to the superconducting line . in order to agree with experiment
we need to essentially put the appropriate electric charges at the ends of each line .
we will try to do this so that the electron receives the correct charge and also the correct spin .
we can do this while avoiding questions about how we attach the charge by assuming that in exiting the end of a line the electromagnetic field undergoes a phase transition .
we are implying that the core of a line is a third phase , in addition to the superconducting field and normal space outside the line . in the rest frame of the bare positron - pole ( still using posititve charge for convenience ) , this phase transition is given by @xmath595 where , outside the line , @xmath596 is the expected magnetic induction field of a point pole .
the magnetic induction @xmath56 is constant across the phase boundary , so that the maxwell equations are satisfied in this phase transition .
the conservation of charge is not violated since an electron and positron are assumed to be pair produced when a line breaks .
assumption ( [ eq - phasetrans ] ) is equivalent to the transformation on the charges @xmath597 and so is our model for the origin of electric charge .
we can take our new assumption about charges conveniently into account with the substitution @xmath598 in equation ( [ eq - bppbfield ] )
. then if @xmath599 is equation ( [ eq - bppbfield ] ) and @xmath600 is equation ( [ eq - bppefield ] ) with the replacement @xmath601 , we can express the total electromagnetic field ( _ except _ the line field ) of the bare positron as a complex vector in the form @xmath602 @xmath603 where @xmath604 is just ( [ eq - bppbfield ] ) with the substitution @xmath605 .
it will be convenient to use this complex notation .
the real field comes from the magnetic charge , and the imaginary field ( which , of course , is actually physical ) comes from the electric charge .
thus we are saying that the bare positron is almost a _
dyon_@xcite , but with a _ spin - dependent _ magnetic charge . in section [ ssec - motionbpp ] the motion of the bare positron - pole
was discussed entirely on the basis of the principle of least action applied to the superconducting line , before making our assumption on the origin of electric charge ( [ eq - origincharge ] ) .
we note that the equations of motion ( [ eq - motionep ] ) or ( [ eq - motionepparts ] ) do _ not _ involve the bare positron - pole s pole field if we evaluate the self field just inside the end of the line .
they do involve the pole field if we evaluate the self field just outside the line .
our assumption ( [ eq - origincharge ] ) makes a difference only outside the line . with or without this assumption
the dominant external electromagnetic field will be due to the reflection ( including the role of current loops ) of the electromagnetic field of the bare positron .
we , therefore , presume that the character of the motion of the bare positron ( _ zitterbewegung _ ) is unchanged by our assumption on the origin of electric charge ( [ eq - origincharge ] ) .
we prove that this is true in appendix [ ssec - motionelectron ] .
however , by adding an electric charge , we have added angular momentum ( spin ) to the end of each line
. this will affect the orientation of each end of the line .
so far we do _ not _ have a model of the magnetic moment of the electron or positron , so we expect our results for the _ orientation _ of the end of each line to be _
nevertheless we presume that the microscopic motion of the bare positron is still zitterbewegung . from section [ ssec - motionbpp ]
we expect rapid changes in the external electromagnetic field felt by the bare positron on a time scale of about @xmath606 or less .
( with a higher line - pair density of the ether , this time would be shorter . ) no measurement device in human possession can measure this time dependence , so we consider the electromagnetic field averaged over a time interval of @xmath607 during which there are at least @xmath608 collisions , enough to form a reliable measurement of the average field . note that although the bare positron is rapidly accelerating , it _ does not radiate _ because the magnetic induction field is blocked by the closed loops maintaining flux quantization .
hence we can see immediately that @xmath609 the time - avearaged magnetic induction field is zero at macroscopic distances from the positron .
the magnetic induction field is reflected back toward the bare positron with no energy loss ( by superconducting loops ) , where it is reflected again by other current loops , and again , and again until the sense of direction of the field is lost , because the current loops are in random orientations .
hence , effectively , all the magnetic induction field from the bare positron is contained in a very small region near the moving positron , and so in this region , it is _ very _ intense .
we work in the effective rest frame of the bare positron
the frame in which it is at rest over @xmath610 and hence @xmath611 if we average over the bare positron s spin and are far from any masses the only unique vector is in the direction @xmath612 which we presume is reasonably constant over @xmath610 .
hence in finding the average electromagnetic field of the positron we can see that all terms transverse to @xmath613 average to zero .
hence @xmath614 the average microscopic electric field near the bare positron is given by @xmath615 the macroscopic electric field of the positron is @xmath616 as expected .
thus the measured charge of the positron is @xmath617 when the dielectric constant of the ether is taken into account .
hence the measured charge of the electron is not constant .
it should vary with the dielectric constant of the ether as the universe expands .
a portion of the cosmological red shift could be due to the change of this dielectric constant with time .
the expected change with time is in the observed direction .
ignore the positron s line field _ , and ignore , for the moment , magnetic screening due to flux quantization , we can find the effective magnetic charge on the positron ( working with positive charge ) using@xcite the modified maxwell equation @xmath618 the magnetic field at macroscopic distances from the positron will be @xmath619 but the magnetic induction field will be @xmath620 determined by the strength of the effective free pole ( magnetic charge ) . both of these fields will be blocked by screening currents due to flux quantization .
we do not yet have a model for these screening currents , but we can at least take these screening currents into account through the use of a _ screening function _
@xmath621 which cuts off the magnetic field of the positron as it would be , on average , cut off by screening currents .
@xmath622 the actual shape of the magnetic induction field of a positron will vary from this smooth model according to the actual positions of the closed circuit current loops relative to the particular positron in question .
equation ( [ eq - bmacscreenedpole ] ) just represents the average shape of the magnetic induction field coming from the positron itself .
reflected fields are not included .
it is well known@xcite that the electromagnetic angular momentum of a point monopole @xmath623 in the presence of a point charge @xmath188 is @xmath624 where the angular momentum vector points in the direction from the point charge to the point pole ( if both are positive ) .
since our model of the positron is almost a dyon , we can calculate the spin of the positron if we assume that the line is filamentary , and that it points in the @xmath243-direction , where it can make no contribution to the angular momentum .
we obtain @xmath625 this calculation assumes that the line radius is small enough that the point charge and point pole approximations can be used .
we also assume that the screening currents are sufficiently far away that they do not affect the calculation .
the calculation is independent of the separation between the pole and the charge , but there must be some separation to determine the _ direction _ of the spin . since there is no natural separation between the charge and the pole in our assumption ( [ eq - origincharge ] ) , we must look to experiment to gain insight . for a positive charge
we assume that the spin is in the same direction as the magnetic moment of the _ line _ ( @xmath317 ) .
( we do not yet have a model of the magnetic moment of the positron . )
figure 14 shows schematically the small separation between pole and charge required to have the spin point in the assumed direction for a positron .
figure [ fig - electronspin ] also shows the situation for an electron . for both particles the assumed spin direction is _
opposite _ to the line . to 1.8 in it is interesting to note that the microscopic calculation of the spin ( [ eq - leg]-[eq - spinpositron ] ) gives the same result as the macroscopic calculation @xmath626 the microscopic calculation is probably the most relevant since the dominant contributions to either integral ( [ eq - spinpositron ] ) or ( [ eq - legmac ] ) come from regions very near the pole and charge .
the separation between the pole and the charge is the only length involved and defines the importance of contributions to the integrals . in the final integration over the @xmath106-direction , half of the integral is contributed by the region between the pole and charge . since
this separation is apparently _ very _ small ( see equation [ eq - rasmall ] ) , the approximation that the screening fields do not affect the integral is probably valid . from sections [ ssec - motionbpp ] and [ ssec - electroncharge ] we see that the electromagnetic field of the bare positron ( again considering positive charge ) does not carry inertia , i. e. mass .
in the average rest frame of the bare positron , however , the superconducting ether will certainly respond to the average electromagnetic field of the bare positron .
this response implies a potential energy , which we interpret as mass .
the wave function of the positron is the response of the ether to the electromagnetic field of the bare postron@xcite @xmath627 where the sum is over all line pairs which respond to the bare positron .
so the mass of the positron is just the potential energy of its wave function in its rest frame .
this response is a discontinuous function of position ( as in section [ ssec - wfelectron ] ) .
it is only defined at equilibrium positions ( in absence of distortion ) of line pairs .
the wave function is a vector field , but it always points toward ( or away from ) the average position of the bare positron , so we can take it as a scalar without loss of information .
so the mass of the positron is given by @xmath628 \label{eq - mesum}\ ] ] where the integrals have the same end points in space , but differing paths .
the points in space @xmath629 and @xmath630 are chosen to be on line pair @xmath341 but beyond the response of line pair @xmath341 to the bare positron .
the first integral is over the actual path of the distorted line pair @xmath341 using the actual tension @xmath631 which could be different from @xmath632 in general , because this line pair is part of a matter wave , a quantized excitation of the ether .
( but see below ( [ eq - tensiontau0 ] ) and ( [ eq - metau0 ] ) where we will find , in fact , that @xmath633 ) .
the second integral is over the straight line path ( in the @xmath243-direction ) which the line pair would have in absence of the positron . since in this case the line pair is assumed not to be in an excited state , the tension is @xmath632 ( [ eq - tensiontau0 ] ) .
the mass of the positron ( or electron ) in this model is clearly calculable . at small distances
there is no convergence problem since we sum over a finite number of line pairs . at large distances
we can approximate the sum ( [ eq - mesum ] ) by an integral .
all integrals clearly converge since we will see that the electromagnetic force causing the response of a line pair falls off like @xmath634 where @xmath635 is the magnitude of the separation from the bare positron s average position .
the zitterbewegung motion of the bare positron is far too fast to affect the ether s response . with typical times of @xmath606 ,
very large energies would be required for such a response , because all responses are quantized .
the ether responds to the _ average _ electromagnetic field of the bare positron , as it undergoes its zitterbewegung motion in its wave function .
the quantized changes occur over long times and can therefore have low energy .
we calculate the response of a line pair to the average electromagnetic field of the bare positron in its average rest frame , where the calculation is static .
if @xmath173 is the electric dipole moment per unit length of the line pair in question , the electric force per unit length on the line pair is @xmath636 but since the dipole moment per unit length is induced by the electric field of the positron itself we have @xmath637 where @xmath638 is the polarizability per unit length of the line pair , which in general depends on @xmath639 , the angle between the electric field and the line pair@xcite .
( see figure [ fig - lpstars12 ] . ) to 1.5 in hence the electric force per unit length on the line pair @xmath640 is proportional to the gradient of the square of the electric field
independent of the sign of the charge of the electron or positron .
the component of the opposite force along the line pair is the gravitational force on the electron or positron .
it is independent of the sign of the charge because it is _
induced_. hence the electric force which the positron exerts on a line pair is given by @xmath641 and is always _ attractive _ , as required for the force of gravity in our region of the universe .
( we assume that we can neglect the second term in this case because @xmath642 , but see section [ ssec-1or3 ] ) the effective magnetic charge of the positron will also cause a force on our line pair given by@xcite @xmath643_i\ = \ \bigl [ ( { \bf m } \times \nabla)\ \times\ { \bf b}^{mac } \bigr ] _ i\ = \
\sum_k m_k \nabla_i b_k^{mac}\ ] ] where @xmath172 is the magnetic moment per unit length of the line pair .
the induced magnetic moment is _ opposite _ to the applied magnetic induction field @xmath644 where the polarizability is the same as in the electric case .
hence the magnetic force per unit length on a line pair is @xmath645 again independent of the sign of the charge .
using ( [ eq - bmacscreenedpole ] ) the magnetic force per unit length is @xmath646 _ repulsive _ and independent of the dielectric constant of the ether .
the screening function @xmath621 ( e.g. @xmath647 ) must be positive and strongly increasing as a function of @xmath635 .
it can also vary slowly with the position of the positron ( e.g. @xmath648 above could vary ) due to extra line pairs from a passing star .
because of this screening , we assume the electric force is dominant , except _ very _ close to the positron , and in the case that the dielectric constant of the ether is _ very _ large .
thus the total force per unit length on a line pair is given by @xmath649 @xmath650 in figure [ fig - lpstars12 ] we consider a line pair stretching out from star 1 and passing through our positron s wave function on the way toward star 2 ( which could be _ very _ far away , e.g. over the horizon ) . in figure [ fig - tensionbalance ]
we see a closeup of how the force per unit length @xmath651 changes the tension , which is considered to point along the line pair in the direction from star 1 to star 2 ( except in figure [ fig - tensionbalance](a ) ) . to 1.5 in ( a ) considering the tension always to point from left to right.,title="fig : " ] since the forces must balance ( in the static configuration ) @xmath652
so @xmath653 and @xmath654 where * x * is any point on the distorted line pair and @xmath655 is the initial tension vector at point @xmath629 .
hence @xmath656 the tension @xmath657 varies such that the effective tension , _ including _ the effects of the force due to the positron , is constant .
the transverse components change the direction of the tension , while the longitudinal components can change the magnitude .
it is convenient and reasonable to _ define _ the tension to include the longitudinal force of the positron .
this force is effectively part of the tension of the responding line pair .
then the magnitude of the tension is always @xmath632 in accord with ( [ eq - tensiontau0 ] ) , and so that the matter wave is a gravitational / common - mode wave .
hence we replace @xmath631 by @xmath632 in ( [ eq - mesum ] ) and obtain for the mass of the positron @xmath658 where @xmath659 is the length difference of line pair @xmath341 with the positron and without it .
the response of a line pair extends far beyond the point of application of the force @xmath660 , just as the distortion of a bow string extends far beyond the point at which it is plucked .
just how far the response extends should depend on the surrounding matter density because other charged particles will exert forces on this same line pair .
these forces , which are exerted on the line pair under consideration by other charged particles , can also be considered using ( [ eq - taux ] ) .
thus the calculation is a many body problem , by its nature , because everything is connected .
appropriate assumptions must be made to find a numerical solution .
this work has not yet been done .
dependence on the surrounding matter density has been observed .
the mass of the electron changes from it s vacuum " value when the electron is in a solid@xcite .
this change of mass in a solid is thought to be a collective effect .
we are suggesting that mass may be a collective effect even in vacuum " . in order to gain insight into this mass calculation we construct a toy model in which we assume ( figure [ fig - lpstars12toy ] ) that the line pair in question is fixed at a distance @xmath433 from the positron .
we neglect the curvature of the line pair in evaluating the total force and consider the total force to act at the point of closest approach .
we assume the electric force dominates so that ( taking only the first term ) the components of the force per unit length which the positron exerts on the line pair are given by to 1.5 in @xmath661 @xmath662 where we simply take @xmath663 @xcite .
the total force perpendicular to the line is @xmath664 where @xmath368 is the distance of closest approach of the line to the positron .
the total longitudinal force ( along the line ) is zero but for future reference we calculate the force that the positron exerts on the line pair in the directions of star 1 and star 2 separately : @xmath665 @xmath666 if @xmath667 is the displacement of the line pair at closest approach , from figure [ fig - lpstars12toy ] @xmath668 @xmath669 and since @xmath670 we have @xmath671 since the line pair tension must balance @xmath672 @xmath673 we find @xmath674 @xmath675 so the mass of the positron or electron in this toy model is given by @xmath676 so , replacing the sum by an intrgral @xmath677 we have ( since @xmath678 ) @xmath679 @xmath680 at this stage the only quantities for which we know a numerical value in this equation are @xmath188 and @xmath681 , but the formula illustrates what quantities are important in determining the mass of the electron in this model
. we can cross check this formula with others which we will obtain .
we expect order of magnitude accuracy if we were to evaluate @xmath682 and @xmath433 with care using appropriate averaging , probably with a monte carlo .
the cutoff @xmath682 should be evaluated so the integral in ( [ eq - sumtoint ] ) approximates the required sum .
it depends , at least in principle , on @xmath145 and on the function being integrated ( here @xmath683 ) . near the earth
we will see ( section [ ssec - parameters ] ) that the dielectric constant of the ether is estimated to be very close to @xmath114 . in this case
we have @xmath684 and @xmath685 , the strength of the dirac monopole .
hence as @xmath686 the force @xmath651 on a line pair is dominated by the pole force , and hence is _
repulsive_. so we expect the effective value of @xmath682 to be the size of a typical deflection with the full pole strength . outside
@xmath682 ( or @xmath687 , reflecting radial symmetry of the configuration with many line pairs ) line pairs will be concentrated as the pole is rapidly screened and @xmath651 becomes attractive due to the electric force . since @xmath682 should depend only on the size of a typical deflection at full pole strength , we will simply take @xmath682 to be a constant ( at least in vacuum near the earth , relevant to the calculation of @xmath19 in this toy model ) .
we discuss the response of our macroscopic electron in its mass shell ( wave function ) to an external electromagnetic field and the possible effects of its string in appendix [ sec - response ] .
we have seen in section [ ssec - masselectron ] that the mass of the electron is proportional to @xmath145 the line pair density of the ether .
we assumed that the line pair density is isotropic reflecting the large scale structure of the universe , for which ( from ( [ eq - dndadomega ] ) ) @xmath688 where the subscript references the _ background _ ether line pairs for which star 1 and star 2 are _ very _ far away ( perhaps over the horizon ) .
in contrast to the background ether , the line pair density from a star is certainly _ not _ isotropic . since we can approximately ( better than @xmath689 accuracy ) associate a line pair coming from the star for each neutron mass ( @xmath690 )
@xmath691 where the delta function in solid angle ensures that all line pairs point in the radial direction from the center of the star ( a simplification ) and where @xmath692 where @xmath693 is the mass of the star and @xmath176 is the distance from the center of the star . in table
[ table - lpdatsurface ] we list some line - pair densities at the surfaces of known astronomical objects , where sagittarius @xmath694 is a star at the center of our galaxy which is thought to be a supermassive black hole@xcite . also listed is the star at the center of the m87 galaxy which appears to be an even more massive black hole@xcite .
for the two black holes we take @xmath695 the radius at the surface to be the schwarzschild radius .
we also quote @xmath158 values at the surface of the object using ( [ eq - ltyp ] ) .
bject = @xmath696 = @xmath697 = @xmath698 @xmath699 + earth @xmath700 @xmath701 @xmath702 @xmath703 + sun
@xmath704 @xmath705 @xmath706 @xmath707 + sagittarius @xmath708 @xmath709 @xmath710 @xmath711 @xmath712 + m87 @xmath713 @xmath714 @xmath715 @xmath716 + [ table - lpdatsurface ] bject = @xmath696 = @xmath717 = @xmath698 @xmath699 + sun @xmath704 @xmath718 @xmath719 @xmath720 + central bulge @xmath721 @xmath722 @xmath723 @xmath724 + background ether - minimum @xmath725 @xmath726 + background ether - measured @xmath727 @xmath728 + [ table - lpdatearth ] we note that the line - pair density at the surface of the earth , from the earth alone , is 11 orders of magnitude greater than our previously assumed minimum line - pair density of the background ether ( [ eq - dndanum ] ) ! assuming masses do not vary significantly with altitude above the surface of the earth@xcite , we infer that the line - pair density of the background ether must be 13 orders of magnitude or more higher than our minimum estimate ( [ eq - dndanum ] ) .
this fact has significant implications for the discussion of inflation . in table
[ table - lpdatearth ] we list the line - pair densities at the earth due to our sun and the central bulge of our galaxy@xcite .
the mass contained in this central bulge is only perhaps a factor of 2 less than the visible mass of our galaxy inside the orbit of the sun .
we see that the contribution of the central bulge is much smaller than the contribution of the sun , which is much smaller than the contribution of the earth itself .
we also list our new estimate for the minimum line - pair density of the background ether ( @xmath729 and our rough measurement ( [ eq - fulldensity ] ) . we have painted a very static picture in discussing equation ( [ eq - dndastar ] ) .
this static picture should be viewed as a statement of the average behavior line pairs emerging from a star .
all changes from this picture are quantized , but changes will certainly occur .
line pairs are attracted by charges of both signs .
so if a star is moving through a field of line pairs , the line pairs will tend to stick to the charges , and stretch as the star continues to move . when the line pairs have stretched sufficiently that their tension force is strong enough to free them ( perhaps due to a changing angle ) ,
they will contract ( being perfectly elastic ) and move to the next row of charges .
the line pairs will tend to form straight lines , however , because in that way they minimize their energy .
most line pairs are part of the background ether
so end , and hence are bound strongly , in stars which are _ very _ far away . assuming the universe is still expanding , star 1 and star 2 , at the ends of a typical line pair , are moving apart as the line - pair tension @xmath632 continues to decrease very slowly with time throughout the universe .
two line pairs can bind to form a 4-line system , but the binding is weak compared to the binding for a pair . in figures [ fig - lpstars12 ] and [ fig - lpstars12toy ]
we identify the force of gravity from star 1 via our line pair @xmath341 on our standard positron to be ( from ( [ eq - fz1 ] ) ) @xmath730 in a similar way we identify @xmath731 the force from star 2 via the same line pair on our positron .
the situation is intrinsically 3-body and equations ( [ eq - fg1j ] ) and ( [ eq - fg2j ] ) represent the only self consistent way to isolate @xmath732 and @xmath733 respectively , as is required to describe the force of gravity .
( in principle we should include @xmath734 in the force from both stars , but this force adds to zero when we consider all line pairs near the positron . ) both stars exert attractive forces on the positron via this line pair @xmath341 .
after all , line pair @xmath341 is _ part of both stars_. the force of gravity propagates instantaneously because both stars are already at the position of the positron .
the quantum nature of these forces is illustrated by the fact that many line pairs from star 1 will come through the positron s wave function if it is sufficiently near star 1 .
however , the nearest line pair coming from star 2 could be light years away if star 2 is over the horizon . thus star 2
will not exert significant force on the positron , similar to many other stars in random directions .
thus the total force of gravity due to all line pairs from star 1 on the positron is @xmath735 @xmath736 @xmath737 we observe that this force is not obviously proportional to the mass of the positron ( [ eq - me ] ) .
we examine the possibility that this force obeys newton s law of universal gravitation , setting @xmath738 and we find that @xmath739 so @xmath740 ( remember that we are only using a toy model here . )
the value of @xmath682 , like @xmath186 , must be constant within our solar sysem .
the value of @xmath682 has an additional restriction
. it must be the same whether the positron is attracted by the sun , or the earth , or a baseball .
we note that the force of gravity on a single electron or positron has not been well measured@xcite .
it is a very difficult measurement so we need a consistency check for this determination of @xmath19 .
we consider a single electron which combines with a single proton to form a hydrogen atom the smallest bit of neutral matter , for which newton s law of universal gravitation is well established in our solar system . we are trying to stay out of the nucleus
so we simply assume that @xmath741 the gravitational force on the proton from star 1 is a factor of @xmath183 times the gravitational force on the electron .
so negelcting binding effects , the force on the hydrogen atom is just @xmath742 times the force on the electron .
@xmath743 normal matter , neglecting binding effects , is just made up of proton - electron combinations , perhaps in the form of neutrons , so the total gravitational force on a mass @xmath744 ( _ assumed _ proportional to mass ) is just @xmath745 so @xmath746 and so we also obtain @xmath747 if we compare our two values of @xmath19 we find @xmath748 hence our two values of @xmath19 agree using a value of @xmath183 which indicates that the gravitational force exerted by star 1 is proportional to the mass of the object on which it acts .
this proportionality is shown more generally by ( [ eq - proportionaltom ] ) .
we have not discussed the gravitational force which our positron exerts on star 1 .
it is zero unless the positron s line goes through the matter of star 1 .
newton s third law is apparently violated by the quantization of the gravitational force , but it should be valid on average .
( the electromagnetic reaction forces to both gravitational forces , as we have defined them , do satisfy newton s third law , as we have shown . )
our calculation of @xmath19 only comes from a toy model , but it shows some important features . using the cgs values @xmath749 @xmath750
we obtain @xmath751 @xmath752 so apparently @xmath69 , the outer radius of a line pair , is very small . if @xmath682 is of order @xmath158 , @xmath753 ! since the gravitational field emerging from a star in this model
is proportional to the number of line pairs emerging ( hence the number of nucleons in the star ) , the gravitational force in this model is _ not _ rigorously proportional to active mass ( the source of the field ) .
we have simply _ assumed _ that the gravitational force is proportional to passive mass because we can not model the nucleus . for a rigorous treatment to an accuracy
better than 1% and a discussion of experimental tests of our model of gravity , see appendix [ sec - rigorousgravity ] . in the presence of star 1 ,
the second term in ( [ eq - ftotal ] ) will lead to a gravitational force on our standard positron and we need to either consider it , or show that it is small . from ( [ eq - epsminus1 ] )
we find @xmath755 where @xmath183 is the fractional concentration of superconducting fields , which implies that @xmath756 near a star , @xmath183 will have two contrubution
one from the background ether and the other from the star . for the star we have @xmath757 where @xmath176 is the distance from the center of the star , and
@xmath758 can not be greater than @xmath114 .
since the superconducting fields can overlap , the fractional concentration _ free _ of superconductor is just the product of the respective concentrations without superconductor @xmath759 implying that @xmath760 hence , adding up the contributions of all relevant line pairs , the total force of gravity on the positron due to the gradient of @xmath186 is @xmath761 also attractive . in calculating ( [ eq - fge ] )
we only included line pairs coming from star 1 , to calculate the total gravitational force from star 1 .
the total force from the background ether was zero because it is isotropic .
but here star 1 causes the dielectric constant of the ether to be _
anisotropic_. so all line pairs , even from the background ether , will exert a force on the positron in the direction of @xmath762 .
the ratio of the magnitudes of the two forces of gravity is @xmath763 for most circumstances @xmath764 and we can neglect ( [ eq - fgeor3 ] ) with respect to ( [ eq - fge ] ) .
however , in extreme circumstances where the fractional concentration of superconductor becomes very close to 1 , the @xmath765 term implies that @xmath766 could dominate .
we rewrite ( [ eq - g ] ) to show the dependence of @xmath19 on @xmath186 , the dielectric constant of the ether : @xmath767 actually @xmath20 ( [ eq - me ] ) and by assumption ( [ eq - fgp ] ) @xmath690 both should vary with @xmath186 .
so we consider the quantity @xmath768 proportional to the gravitational force on representative masses , to most effectively represent the weakening of the gravitational force as @xmath186 changes .
if @xmath682 is constant , @xmath769 will weaken significantly in regions where @xmath186 is important , because @xmath186 becomes large near @xmath770 .
thus the gravitational force should weaken compared to expectations when the density of line pairs becomes high , for instance near a black hole where the gravitational force is expected to be strongest . since @xmath769 varies as @xmath771
, the effects of black holes could be considerably weaker than expected .
this could contribute to the fact that ligo@xcite has not yet seen gravitational waves .
on the other hand , the masses of black holes could be considerably larger than measured while assuming constant @xmath19 .
as shown by ( [ eq - fgeepsofge ] ) in the limit @xmath770 , the @xmath754 gravitational force ( [ eq - fgeor3 ] ) could dominate ( [ eq - fge ] ) .
but in this limit both electric terms in ( [ eq - ftotal ] ) should be dominated by the repulsive magnetic terms , which are screened but independent of @xmath186 .
we will suggest in section [ sec - rotations ] that the reversal of the gravitational force has indeed been seen , in a region which should have a _ very _ high density of line pairs .
if this reversal has , indeed , been seen , a reversal should certainly have occurred in the high densities of the early universe .
since we have so far been unable to derive newton s second law directly from the properties of the ether , we use einstein s principle of equivalence of gravitation and inertia@xcite ( extended slightly ) and consider a positron with acceleration @xmath772 .
this acceleration is equivalent in all respects to a gravitational field which points _ opposite _ to @xmath772 .
( think of a scale in an elevator . )
@xmath773 the latter form applies if the gravitational field is due to star 1 .
what we interpret as the gravitational force on the positron from star 1 , we have seen is actually a force of the line pairs from star 1 on the positron .
@xmath774 hence in the equivalent situation of acceleration @xmath772 , line pairs exert a force on the positron given by @xmath775 the acceleration must be due to some applied force , @xmath776 .
the forces on the positron must balance : @xmath777 @xmath778 hence the ether exerts a force equal to @xmath779 on an accelerating mass , leading to newton s second law of motion . for a deeper understanding of newton s second law , including how it can be made consistent with the discussion of section [ ssec - calcg ] , see appendix [ sec - understandingnewton ] .
as a result of the above discussion we see that the force exerted by the positron on the line pairs of the ether is given by @xmath780 this force is exerted on the line pairs that go through the wave function of the positron .
the line pairs that go through the positron s wave function respond by stretching and deforming ( not by accelerating like a massive particle ) .
by our use of maxwell s equations we have assumed special relativity from the beginning of this paper .
we point out how special relativity fits naturally into our construct .
special relativity applies to the situation of a particle in the background ether , which is homogeneous and isotropic because it is far from any masses .
if a photon passes this particle , its velocity is @xmath781 , because of the characteristics of the line pair ( without rest mass ) on which it is propagating
independent of the motion of the source .
this observation is sufficient@xcite to build the entire special theory of relativity .
our concept of mass ( [ eq - metau0 ] ) fits very naturally into the overall construct , as potential energy of the ether . for a discussion of the full meaning of lorentz invariance ,
see appendix [ sec - understandingnewton ] .
it should be possible to relate the quantized line - pair density of the ether , approximated by @xmath782 , to the metric tensor of general relativity in static situations .
this work has not been done but has promise of uniting general relativity with quantum theory .
the fact that @xmath19 is not constant must be taken into account .
we have not yet done a serious job of studying the dynamics of the ether , e.g. using ( [ eq - fonpairs ] ) .
such a study will be needed for a complete model .
we note that there is an essential difference between line pairs in the ether and the structure of space - time in general relativity .
line pairs are quantized objects which can _ move _ , and are not fixed in space - time .
we also note here that the concept of gravitational waves used in this paper is quite different from the gravitational waves of the standard model@xcite .
our concept of gravitational waves is appropriate to our quantized model .
in this model , wave - particle duality is not a problem it is natural .
the particle is the bare electron .
the wave is the response of the ether to the bare particle@xcite .
since we started with dirac s quantization condition ( [ eq - singv ] ) it is not surprising that we are consistent with quantum mechanics with one exception .
we construct a continuous wave function which models the displacement of each line pair in ( [ eq - psix ] ) in the average rest frame ( @xmath282 ) of the electron : @xmath783 this function is a signed scalar .
it interpolates between the displacements of the various line pairs to form a continuous function , as in standard quantum mechanics .
but contrary to the expectations of standard quantum mechanics , this wave function does not spread out with increasing time .
the electron at rest is at equilibrium with the ether but subject to zitterbewegung which will manifest itself in slow motion after sufficiently long times .
the whole wave function moves , but does not spread .
this could be a measureable difference between our model and the standard model of quantum mechanics .
thus we see that our electron s wave function is intrinsically localized in position space , but has a finite spread , and therefore must have some spread in momentum space .
hence true plane - wave states do not exist .
therefore we need to modify our discussion of the quantization of the photon ( equation ( [ eq - pepg ] ) and figures [ fig - wfelectron]-[fig - photonfixedz ] ) since we used a plane wave electron for simplicity .
but there is really no essential difference from our previous discussion .
the non - plane - wave electron state is a superposition of plane wave electrons with various values of @xmath784 and we have @xmath785 for each @xmath784 in the superposition .
this leads to the quantization of the photon ( [ eq - pepg ] ) as before .
we can also obtain the quantization of matter waves by assuming that all matter waves are localized ( to any required accuracy consistent with the uncertainty principle ) in both position and momentum spaces .
if we simply look on a larger scale at figure [ fig - photonfixedt ] and replace the photon by a localized matter wave , then just as for the photon we will have @xmath786 where we do not know in advance the precise value of @xmath784 , but presume it will correspond to a value of @xmath787 that is probable according to the momentum space wave function of our particle . in figure [ fig - photonfixedt ] ( concentrating on the matter wave instead of the photon ) we pick a path of integration along the @xmath106-axis , defined to be parallel to @xmath787 , and integrate over one complete wavelength . from the fact that @xmath158 ( table [ table - lpdatsurface ] ) is very small , we can do this to any required accuracy , and then turn to integrating over a @xmath368-leg , which is perpendicular to the momentum .
we then use a @xmath106-leg which is far away , well outside the particle s wave function , where the contribution to the total integral is negligible .
we then complete the integral with a @xmath368-leg as before .
hence we obtain as in section [ ssec - quantphoton ] @xmath788 then using the standard formalism@xcite we find the momentum space wave function of our electron @xmath789 in its rest frame . from our previous discussion ( section [ ssec - masselectron ] ) both @xmath790 and @xmath791
are spherically symmetric , if we sum over spin states .
hence the general wave function of the free electron can be written in the standard form , where @xmath792 where @xmath793 is no longer restricted to the rest frame . the full wave function , including time dependence is given by @xmath794 where @xmath784 is the total 4-momentum of the electron . in position space we have @xmath795 just as in the normal formalism .
this leads to schr " odinger s equation and the standard quantum mechanics of angular momentum . but
until we have a model of the magnetic moment of the electron , we can not fully discuss spin .
( we could , of course , just take the value of the magnetic moment of the electron from experiment or the dirac equation , as standard quantum mechanics does . )
for instance for the hydrogen atom , an electron in the potential energy of a heavy proton , we obtain @xmath796 \
\psi_{qm } \ = \
i\hbar\ { { \partial \psi_{qm } } \over { \partial t}}\ \ \ \label{sch}\ ] ] the schr " odinger equation just like in standard quantum mechanics . the energy spectra for the two models are identical . in both models we have to add the effects of spin to get the measured spectra
if we have succeeded in calculating @xmath19 , the entire concept of the planck mass is specious because @xmath19 is not fundamental .
hence general relativity and the hot big bang model , in particular , need modification .
but we can at least heuristically consider the topic of inflation in our model by using weinberg s result@xcite for evolution of the universe @xmath797 where @xmath51 is the energy density of the universe , @xmath798 is the pressure , and @xmath176 is the cosmic scale factor ( sometimes called the radius of the universe ) .
we suggest that the twoline potential energy ( [ eq - twolinepot ] ) was the energy source for inflation . in the early universe , after the line pairs started to condense
, we presume that they were radiation dominated , with many excitations traveling on them , so that both @xmath51 and @xmath798 were positive .
the line pair density should have been extremely high so that the fractional concentration of superconductor approached one ( @xmath770 ) and the dielectric constant of the ether was very large ( @xmath799 ) .
thus the effective charge of the electron approached zero and from ( [ eq - ftotal ] ) the gravitational force was initially _
repulsive_. therefore , heuristically , we can think of the value of @xmath19 in ( [ eq - evolutionr ] ) as negative and exponential expansion resulted .
when the universe expanded to the point where the fractional concentration of superconductor became less than 1 , @xmath186 decreased , the force of gravity became attractive and the exponential expansion stopped , as long as @xmath800 this should have been true through the eras of radiation dominance and matter dominance until fairly recent times .
using ( [ eq - fulldensity ] ) and ( [ eq - dndanum ] ) , we estimate the expansion factor during inflation to be @xmath22 , with a large uncertainty .
dark energy consists of line pairs which stretch across the universe .
the dark energy density is the energy density of the ether @xmath801 note from ( [ eq - tensiontau0 ] ) that the energy / length due to the quantization currents has been included in @xmath632 .
dark energy is completely dark .
the only way it can be detected directly is by its gravitational effects , such as changes in a mass moving through it .
the dark energy density decreases with time as the universe expands and the tension of the all the line pairs in the universe decreases slowly .
the pressure due to the longitudinal tension is given by @xmath802 so for the longitudinal tension alone we would have the ratio of pressure to energy density given by @xmath803 if there were no transverse tension , we would take the average pressure over three directions and find @xmath804 but there is tension in one of the transverse directions .
we can easily calculate the transverse tension which the quantization currents need to balance in order to keep @xmath632 constant , for a length @xmath210 of a line pair .
if @xmath805 locates the center of line 2 from line 1 ( section [ sec - two ] ) , this tension is @xmath806 the quantization currents could provide this transverse tension simply by flowing in opposite directions on the outer surfaces of the line pair or they could communicate with other line pairs as shown in figure [ fig - currentsforapart ] . in the latter case
the transverse tension would be transmitted to other line pairs and @xmath807 implying that @xmath808 for the ether and from ( [ eq - evolutionr ] ) the ether can contribute to the current acceleration of the universe .
each line pair in this case is under tension in two directions .
we use ( [ eq - me ] ) , ( [ eq - g ] ) , and ( [ eq - rholambda ] ) to estimate the line - pair density of the background ether in terms of the dark energy density , the mass of the electron , and newton s constant @xmath19 @xmath809 we have taken @xmath810 from the wmap collaboration@xcite @xmath811 in order to obtain an estimate for the properties of the background ether , we guess that @xmath812 so that @xmath813 we take @xmath433 to be the mean free path in air@xcite @xmath814 we obtain @xmath815 consistent with our minumum value from table [ table - lpdatearth ] .
we note that the line pair density of the background ether is of order 20 times the line - pair density of the sun at its surface . from ( [ eq - dldv ] )
we obtain @xmath816 . using ( [ eq - rholambda ] ) and
( [ eq - fulldensity ] ) we find the line - pair tension of the universe at the current time : @xmath817 then again using @xmath818 we find from ( [ eq - rasmall ] ) @xmath819 we note that our guess for @xmath682 ( [ eq - yminguess ] ) and ( [ eq - ymintyp ] ) looks reasonable since @xmath682 must be @xmath820 , or it would affect the measured pointlike structure of the electron . to estimate the fractional superconducting concentration of the background ether , we consider the fractional concentration in one volume @xmath821 and find it to be negligible .
so the dielectric constant of the ether is 1 near the earth and in all circumstances except the most extreme .
since this model predicts the weakening of the gravitational force near strong gravitational sources because of the increasing dielectric constant of the ether , in principle it could account for dark matter . but the very low fractional concentration of superconductor in the background ether from ( [ eq - fb ] ) ( which should be similar in contribution to stars in our galaxy ) appears to rule this out .
we need to discuss what happens to the line pairs , which begin in a star , when it rotates .
if a line pair points along the axis of rotation , nothing happens .
if it is off axis pointing parallel to the axis , it will twist about the axis of rotation .
but this is no problem since line pairs can move through each other .
if a line pair is off axis and pointing perpendicular to the axis , it will stretch and contract as the star rotates .
this is no problem since the first line pair is matched with another line pair pointing in the opposite direction which stretches when the first contracts and vice versa .
hence no net work is done by the star in rotating however , as a star rotates , each bit of mass @xmath91 does exert a force on the line pairs in its wave function given by ( [ eq - fonpairs ] ) @xmath822 ( where @xmath823 is the cylindrical radius ) .
thus in extreme cases , line pairs will be concentrated on the axis of a spinning mass .
hence we achieve a natural qualitative understanding of agn ( active galactic nuclei ) jets in this model .
the m87 galaxy ( in table [ table - lpdatsurface ] ) has an agn / black - hole at it s center of a mass @xmath824 , which sends out about @xmath825 line pairs .
many of these line pairs become concentrated on the axis of rotation , apparently sufficiently to reverse the force of gravity , producing jets . the velocities in these jets
are high@xcite , because the particles involved have very low mass , according to ( [ eq - me ] ) because @xmath826 becomes small .
( actually if the force of gravity reverses , the dominant contributions to the mass ( [ eq - metau0 ] ) from ( [ eq - ftotal ] ) are repulsive , but we can still expect the mass to be small . )
we have tied a string onto every electron in the universe , therfore electrons are _
distinguishable_. this removes the theoretical underpinnings of the exclusion principle , which apparently pauli never believed@xcite anyway . but something must replace these underpinnings .
we suggest that the pole force of the electron does part of the job .
the pole has the strength of a dirac monopole ( if @xmath502 near the earth ) .
screening currents tend to amplify the strength of the pole if you are in _ contact _ with the electron , i.e. inside the screening currents .
so there should be _ very _ strong forces between electrons ( i.e , any objects ) in contact .
we suggest the pole field of the electron is responsible for the impenetrability of matter . if you push down on a table , we suggest that the force which pushes back on you is the contact force of the pole fields of the electrons in the table .
( hence our lines sense pressure only at the ends . )
we postulate that the rest of the job is done by the line forces between electrons .
the line force has infinite range , as does the spin glass spin - spin correlation length@xcite .
we also suggest that the line force is responsible for the _ exchange _ force of ferromagnetism .
elliot lieb has said@xcite that the reason the pauli exclusion principle is not a force is that you can not write it in hamiltonian form .
so calling it a force is perhaps appropriate for this classical model .
the line force between two electrons provides a strong attractive force between separated electrons of opposite spin .
apparently any attractive force between electrons can lead to cooper pairs@xcite , so we expect the line forces between two electrons to contribute to the formation of cooper pairs , and hence that form of superconductivity .
we note that line pairing should show an isotope effect since nuclei are involved .
the isotope effect played a strong role in establishing the original cooper pair mechanism .
the quantum of flux in a cooper pair system is @xmath827 , because the charge is due to two electrons .
we have to explain why flux quanta of @xmath827 can exist in type ii superconductors , when only flux quanta of @xmath828 are allowed by the ether .
we note that @xmath827 is a point of unstable equilibrium for line pairs in the ether .
if the curl of the phase field is filamentary , the ether currents do nt know which way to change .
i.e. whether to bring the flux to @xmath64 or @xmath828 we postulate that this position of unstable equilibrium with the ether currents is enough to let the cooper pair currents keep the flux steady at @xmath827 .
it is interesting to note that if line forces between electrons contribute strongly to the formation of cooper pairs , then line pairs are pretty much the same thing as cooper pairs in a superconductor and hence appropriate to the generation of mass@xcite according to some authors .
the integer quantum hall effect can be taken@xcite as direct evidence for the quantization of magnetic flux in units of @xmath828 . in the fractional quantum hall effect
, it is suggested@xcite that composite fermions are formed when one electron binds an even number of flux quanta ( @xmath828 ) .
the fractional quantum hall effect is then explained when these composite fermions undergo the integer quantum hall effect , as if they were electrons .
the currents needed to contain these flux quanta are not obviously available , so we suggest that the quantum hall effects present direct evidence for our unpaired lines of flux ( section [ ssec - static ] ) .
we suggest that the flux is contained by the superconducting fields which we postulate in section [ ssec - superconfield ] .
the force that binds the electron to the lines of flux is then given by ( [ eq - ftotal ] ) .
we have seen that this same force on line pairs is the force of gravity .
we obtain schr " odinger s equation for the hydrogen atom , and by implication for heavier atoms .
we obtain newton s second law of motion for the electron , and by assumption for more massive objects .
our model of gravity agrees with newtonian gravity within our solar system
. therefore differences between the predictions of this model and the standard model should be hard to find .
we focus here on experiments which could show a difference .
we show in table [ table - relevantexperiments ] relevant experimental results and indicate if the standard model and this model are consistent with these experiments . 1 .
search for flux quantization outside cooper pair superconductors : + the most relevant experiment to this model is a search for flux quantization performed by tonomura _ et .
they conclude that magnetic flux is not quantized in a torroidal ferromagnet
. if this result is correct , it rules out this model , which is based on flux quantization .
this experiment is one in a series of beautiful electron holography experiments that has studied the aharanov - bohm effect in fine detail .
+ however , we suggest that this group of experiments can not rule out this model .
the reason is that this model changes the nature of the experimenter s probes .
we attach a line of flux onto each of their electrons .
examination of their methods@xcite shows that they did not pay attention to the spins of their probe electrons .
consequently the lines of flux attached to their probe electrons may have spoiled their results .
their geometry in dealing with normal samples was likely different from that which they used for measurements with superconductors .
+ the above argument applies to any search for flux quantization which uses electrons .
+ we consider the argument for flux quantization in section [ ssec - fq ] to be very strong .
we also note that flux quantization has apparently been observed outside of superconductors in the quantum hall effects
g-2 of the electron : + we list in table [ table - relevantexperiments ] the measurement of @xmath829 of the electron , for which the standard model has achieved very high precision agreement .
so far , there is no prediction for this model , because we do not have a model of the magnetic moment of the electron .
dark energy : + measurements of dark energy are pretty clearly not consistent with the standard model , but are consistent with this model of the ether ( section [ sec - darkenergy ] ) .
4 . inflation : + inflation in the standard model requires a scalar field , which has never been observed , plus reheating@xcite after the scalar field does its job .
this model , at least heuristically ( section [ sec - inflation ] ) , proceeds from the string era naturally , with no new requirements , since the strings are still here .
agn jets : + this model , at least qualitatively ( section [ sec - rotations ] ) , explains why the rotation axis of a black hole is special , why agn jets form , and why they have very high velocities .
explanations using the standard model are doubtful .
spin - spin correlation length + the spin - spin correlation length in spin glasses is infinite@xcite .
every spin couples equally with every other spin regardless of how far they are apart .
this is widely considered to be ridiculous , but it is not taken as a conflict with the standard model , because spin glasses are very complicated systems . in our model
an infinite correlation length is natural .
we suggest that these complicated systems which disagree with the standard model show a pattern .
fractional quantum hall effect : + the fractional quantum hall effect is another case where the standard model is in apparent disagreement with experiment , but the disagreement is not taken very seriously because solids are complicated systems . our model has a natural interpretation of this effect ( section [ sec - quantumhall ] ) .
xperiment standard model this model + + flux quantization search@xcite yes no but see section [ ssec - existing ] + g-2 of the electron yes no prediction + dark energy no yes + inflation requires scalar field , reheating yes + agn jets doubtful yes + spin - spin correlation length no , but complicated yes + fractional qh effect no , but complicated yes + [ table - relevantexperiments ] in addition we suggest the following experiments to measure a difference between this model and the standard model : 1 .
find the line : + this model predicts that charged particles come with an attached line . in accelerators
presumably the pair to the beam particle ( with which the beam particle forms a line pair ) , because of energy considerations , is among the image charge in the beam pipe .
the stability of the beam should be strongly affected by the image charge in the beam pipe .
apparently this is true . if a short section of ceramic beam pipe were to be used , the beam would probably be lost.@xcite it may be possible to directly observe the line in accelerators
measure g accurately : + look for a material dependence ( of order a part in @xmath830 ) of the measured values of @xmath19 . see appendix [ sec - rigorousgravity ] .
3 . study the pomeron : + if this model is correct charged particles are not isolated and unitarity will be violated . in light of this possibility , the rising total cross section is interesting .
this rise may be associated with the pomeron .
4 . study the spreading of the electron wave function : + this model predicts that the wave function of an isolated electron will not spread .
look for the pole of the electron : + from ( [ eq - ymintyp ] ) we see that measurements down to a scale of @xmath831 may prove sufficient .
experiments able to make measurements on this scale may be able to measure the screening function @xmath832 in ( [ eq - ftotal ] ) .
we collect here the remaining problems that we anticipate to be the most serious for this model : 1 .
the disagreement with the flux quantization experiments of tonomura _ et al . _
is discussesd in section [ ssec - existing ] .
the violation of the theoretical underpinnings of the pauli exclusion principle is discussed in section [ sec - pauli ] .
3 . with the size of the outer radius of a line pair estimated by equation ( [ eq - ra ] )
the tension in an unpaired line ( [ eq - t ] ) is ridiculously high .
this has observable consequences .
other numbers appear to be fairly reasonable . if this model is correct , a way must be found to fix this problem .
this problem seems similar in severity to a well known problem in the standard model
the energy in the electric field of a point charge is infinite .
the bare electron is modeled as the end of a superconducting line of magnetic flux .
electric charge must be added by hand to this bare electron , but with the correct charge the electron also has the correct spin , @xmath18 .
so the bare electron is almost a dyon , but with a spin - dependent magnetic charge .
if the dielectric constant of the ether is 1 near the earth , the effective magnetic charge of the electron has the strength of a dirac monopole , but is strongly screened by the ether superconducting line pairs from all the other electrons in the universe .
the bare electron interacts with the ether and undergoes zitterbewegung , very rapid random walk motion , traveling near the speed of light in its average rest frame .
the wave function of the electron is the response of the ether to the average electromagnetic field of the bare electron , explaining wave - particle duality in a natural way .
quantum mechanics results from this wave nature .
the mass of the electron is the potential energy of its wave function ( in its rest frame ) .
the charge and mass of the electron are predicted to vary with the dielectric constant of the ether .
this may affect the interpretation of the cosmological red shift .
the photon is modeled as a helical wave , one wavelength long , traveling on a line pair .
the spin of the helical photon is calculated to be @xmath17 .
the ether is a mesh of superconducting line pairs , of a new type which do not conduct electrons or cooper pairs .
these line pairs are under tension , so could represent dark energy . in general , quantization conditions arise because the ether is a multiply connected superconductor . using a toy version of our model with measured values of newton s constant @xmath19 , the mass of the electron and the dark energy density of the universe
, we estimate the line pair density near the earth to be @xmath833 .
this is a factor of @xmath22 larger than would be expected without inflation .
the gravitational force exerted by a star on a positron is due to the electromagnetic field of the positron interacting with the superconducting line pairs coming from the star .
this force is independent of the sign of the positron s charge because it is _
induced_. the electric force is attractive but the magnetic force is repulsive .
hence the gravitational force is predicted to weaken , then reverse , as the dielectric constant of the ether becomes large .
the ether exerts a force equal to @xmath779 on an accelerating mass which leads to newton s second law of motion .
we have proven that both the principles of gauge invariance and microscopic causality are violated by the superconducting fields of this model . because this model is acausal
, it is also nonlocal and even self interactions are finite .
we suggest that the pauli exclusion principle is actually due to a very strong spin - dependent magnetic force : the pole force and line force between electrons .
we suggest that the pole force , a contact force , is responsible for the impenetrability of matter .
i thank my colleagues at stony brook for helpful discussions .
this includes chris quigg who introduced me to the dirac string long ago .
any mistakes , of course , are my own .
this work was supported by the national science foundation .
it is useful to define the angle @xmath834 so that @xmath835 gives the instantaneous slope of the string at any point for solution ( [ eq - yzline ] ) moving in the @xmath314 plane .
@xmath836 here @xmath40 is the direction of the magnetic field ( [ eq - bfieldline ] ) at @xmath41 , the point in question .
also @xmath40 is the direction of the centerline at the relevant point ( @xmath837 ) .
( for simplicity , we assume here that this point is unique as will be true in the case of sufficiently large radius of curvature . )
we follow the discussion of section [ ssec - supercurrent ] in seeking the derivatives of @xmath838 with respect to position and time at the field point @xmath41 .
this unit vector gives the direction of the magnetic field at @xmath41 and is assumed to be parallel transported from point @xmath38 on the centerline to @xmath41 for this purpose .
note that the derivatives of @xmath317 will be _ different _ at point @xmath41 than on the centerline ( point * y * ) because the radius of curvature will in general be different for a curving line .
if @xmath40 changes , the direction of its change must be perpendicular to itself : @xmath839\ = \
d\theta[{\hat y } \ell_z-{\hat z } \ell_y ] \ = \ d\theta\ \hat{\theta}\ ] ] we define the instantaneous radius of curvature @xmath176 by the equation @xmath840 where we define the radius to be _ negative _ if the change in @xmath834 is positive , assuming positive @xmath841 .
we make this choice of sign because a line with decreasing @xmath834 in @xmath314 motion matches the configuration of helical motion and it is convenient to consider this radius of curvature positive .
since @xmath210 changes only in the @xmath40 direction , we have @xmath842 hence we can evaluate the partial derivatives of the components of @xmath40 , for solution ( [ eq - yzline ] ) motion in the @xmath314 plane : @xmath843 @xmath844 @xmath845 @xmath846 since @xmath847 does not change in the rest frame of the string , we use section [ ssec - supercurrent ] to find the partial derivatives of the components of @xmath317 with respect to time . @xmath848 @xmath849 at a given point in time the instantaneous radius of curvature changes with the change in the position of the field point @xmath41 according to the equation @xmath850 because at a given time the center of curvature is fixed , and therefore for solution ( [ eq - yzline ] ) the partial derivatives with respect to @xmath368 and @xmath106 are given by @xmath851 @xmath852 and the partial derivative with respect to @xmath200 is @xmath853 from ( [ eq - lhat ] ) we can find general expressions for the direction of the change with decreasing @xmath834 ( hence @xmath854 , defining positive @xmath176 ) and the expression for @xmath350 , the radius of curvature of the centerline . taking the differential of ( [ eq - lhat ] ) with respect to @xmath106 we find @xmath855 @xmath856 @xmath857 and so noting @xmath858 we find @xmath859 for a displacement in the @xmath106-direction along the centerline .
we note that the @xmath368-component of @xmath860 is always negative , but that the radius of curvature in @xmath314 motion can change dynamically between positive and negative values .
if we consider the position of the centerline as a function of @xmath106 ( as an independent variable_not _ a field coordinate ) we can relate the radius of curvature on the centerline to the shape ( second derivative ) of the trajectory : @xmath861 @xmath862 @xmath863 we use this equation to show how the radius of curvature is transformed in going from the lab to the line s rest frame : @xmath864 we use a polar coordinate system with origin at the center of curvature to find @xmath865 . from the definition of the curl @xmath866 where @xmath867 is the normal ( given by the right hand rule ) to the circulation of @xmath317 around the closed path @xmath868 , divided by the area enclosed
( see figure [ fig - curl ] . )
we find @xmath869 } \over { dr\ ( r+dr/2)\ d\phi } } \ = \ { { { \hat n}\ dr\ d\phi } \over { dr\ r\ d\phi } } \ ] ] @xmath870 where @xmath871 in the case of figure [ fig - curl ] . to 2.5 in .,title="fig : " ] in common mode geometry appendix [ ssec - instradius ] refers to @xmath872 which points generally in the @xmath873-direction .
we need to discuss how @xmath874 differs , since @xmath875 . from ( [ eq - lhat ] )
we find @xmath876 @xmath877 @xmath878 @xmath879 @xmath880 @xmath881 and all the spatial first derivatives are reversed ( when referenced to a fixed coordinate system , as must be true ) compared to the case of @xmath872 and expressed in terms of @xmath872 or @xmath874 as appropriate .
but the radius of curvature on the centerline @xmath350 ( and by extension @xmath176 ) is the same for both lines . for helical geometry from ( [ eq - lhat1 ] ) and ( [ eq - lhat2 ] ) we find , in the rest frame of _ both _ lines @xmath882 @xmath883
so for line 1 @xmath884 the radius of curvature is a constant on the centerline . in the rotating rest frame
we use the time of an observer at the center of rotation who is at rest in the lab .
the derivatives of @xmath885 can be calculated from @xmath886 but if we only need the curl , we can use ( [ eq - curl ] ) . for line 2 @xmath887 @xmath888 and
we see that the radius of curvature on the centerline is also a constant for line 2 , with the same constant value of @xmath889 as for line 1 .
the derivatives of @xmath890 can be calculated from @xmath891 in order to calculate the electromagnetic field from ( [ eq - supercurrentyz ] ) , we need to make use of the fact that @xmath667 and @xmath892 are related by the fact that @xmath823 is a cylindrical radius . selecting events which are simultaneous in the lab to define @xmath317 and @xmath823 , and requiring @xmath893 in the rest frame of the line , we find @xmath894 using this relationship and ( [ eq - supercurrentyz ] ) in ( [ eq - generalsol ] ) we find for the electromagnetic field : @xmath895 @xmath896 @xmath897 @xmath898 - \hat{z}\delta y\gamma^2[1 - 2\ell_y^2\beta^2 ] \bigr\}\ ] ] @xmath899 \biggr\}\ ] ] so the terms in @xmath334 and @xmath335 are not consistent with ( [ eq - efieldline ] ) and ( [ eq - bfieldline ] ) . in helical geometry
if we attempt to use ( [ eq - generalsol ] ) to try to calculate the electromagnetic field in the rest frame of line 1 , we obtain @xmath900 @xmath901 which is consistent with ( [ eq - eline * ] ) but differs from ( [ eq - bline * ] ) by a term which is second order in @xmath902 and consequently very small .
we consider the dependence of the energy of a curved but static line on the radius of curvature .
we use a curvilinear coordinate system which bends with the line and has @xmath834 , @xmath51 and @xmath639 as independent variables with the elements of length @xmath903 , @xmath904 , and @xmath905 . the cylindrical radius of curvature @xmath176 at the point @xmath41
is given by @xmath906 where @xmath350 is the radius of curvature of the centerline .
we calculate the magnetic energy in a length @xmath907 of the line where @xmath908 we find @xmath909 @xmath910 so we obtain @xmath911 as in ( [ eq - t0 ] ) .
the energy of the line is independent of its curvature .
one side of the centerline stretches , but the other side contracts , with no net change in energy .
for common mode motion with the line separation perpendicular to the motion , we have shown that the interaction terms in ( [ eq - intcommode ] ) sum to zero and therefore solutions of type ( [ eq - yzline ] ) can propagate at the speed of light on the two lines of an isolated pair .
now we would like to consider a line pair moving in the @xmath314 plane with the line separation also in the @xmath314 plane . in common mode
the separation of the centerlines is fixed at @xmath912 perpendicular to both centerlines .
this geometry has the complication that the corresponding points on the two lines have _ different _ values of @xmath913 though each is a solution of type ( [ eq - yzline ] ) .
so @xmath914 @xmath915 where @xmath372 is a constant and @xmath916 .
using arguments similar to appendix [ ssec - instradius ] and remembering that here @xmath106 is an independent variable not a field coordinate , we find @xmath917 @xmath918 where @xmath350 is the radius of curvature of the _ common mode centerline _ of both lines .
so @xmath919 @xmath920 we find that ( [ eq - lm1 ] ) and ( [ eq - lm2 ] ) become @xmath921 @xmath922 so the total lagrangian becomes @xmath923 identical to @xmath251 in ( [ eq - lm ] ) . as in section [ sssec - commonmodesol ]
the common mode interaction terms add to zero when both lines are referenced to the common mode centerline .
the two lines do not bend so that they have precisely the same shape .
they have the same center of curvature when referenced to the common mode centerline , so that their individual centerlines have _ different _ radii of curvature but the radii of curvature of the fields are identical at any point hence we get the same equations of common mode motion with the separation perpendicular to the plane of the motion as we get if the separation is in the plane of the motion .
therefore we presume that the orientation of @xmath402 makes no difference to the common mode motion , and common mode solutions ( gravitational waves ) of type ( [ eq - circularline ] ) can also propagate at the speed of light on an isolated line pair .
following the reasoning of section [ ssec - motionbpp ] we have shown how the equations of motion of the the bare positron - pole ( [ eq - bpptmotion]-[eq - bpp3motion ] ) are solved in the case that @xmath924 ( where @xmath162 and @xmath56 refer to external fields ) . if @xmath317 is at an angle @xmath925 from the magnetic induction field @xmath56 and @xmath926 as shown in figure [ fig - orientation ] , then from ( [ eq - bpplmotion ] ) and ( [ eq - bpp3motion ] ) we find @xmath927 @xmath928 + b_0 \sqrt{b_0 ^ 2 + e^2 cos^2(\theta+\alpha ) - b^2 cos^2(\theta ) } } \over { b_0 ^ 2 + e^2cos^2(\theta+\alpha ) } } \label{eq - transcendental}\ ] ] which is the transcendental equation which must be solved for @xmath834 given @xmath248 , @xmath929 , and @xmath584 .
if @xmath584 is small , we presume that @xmath834 will be small and @xmath579 is fixed , since all the cosine terms in ( [ eq - transcendental ] ) are equal to @xmath114
. then we have @xmath930 @xmath931 showing that the orientation of @xmath838 is easily determined if @xmath584 is small . to 3.0 in the electric charge at the end of each line
is easily incorporated into the principle of least action using the electric charge current in ( [ eq - actionnoh ] ) .
the only unnatural requirement for this model is putting the charge and pole at ( nearly ) the same position at the end of each line , so that the end of a line becomes a almost a dyon , with a spin dependent magnetic charge . with both electric and magnetic charge ,
equation ( [ eq - motionepparts ] ) , the boundary condition to be satisfied at the end of each line , becomes : @xmath932 we evaluate the boundary condition just outside the end of each line so that the electric charge is included .
the pole self field is continuous so we can evaluate it just inside the line as we did in section [ ssec - selfintbareelectronpole ] .
the self field of the electric charge will be finite if we average over the cross section of the end of the line .
then , since we switch to the point charge approximation to evaluate the motion of the bare positron , we can argue that the electric self energy is zero from symmetry .
it is easy to see that the cross terms in the self energy are zero . evaluating them in the rest frame of the positron where @xmath933
, we see that the pole only interacts with a magnetic field , whereas the charge only interacts with an electric field .
hence the self - energy terms are unchanged from the case of the positron - pole and the equation of motion becomes @xmath934 \label{eq - eqofmotionwcharge}\ ] ] where @xmath935 is a small number if @xmath936 near the earth
implying that the motion of the bare positron is dominated by the pole , which is much stronger than the charge .
we can show this directly by combining terms in ( [ eq - eqofmotionwcharge ] ) to find @xmath937 that the motion reduces to the case of the bare positron - pole .
we simply make the substitutions @xmath938 @xmath939 and the equation of motion of the bare positron reduces to the equation of motion of the bare positron - pole , in terms of these transformed fields .
this model of the positron ( using positive charge for convenience ) has an electric charge of @xmath193 , a screened magnetic pole field of effective magnetic charge @xmath194 , and a line field that ensures that the total magnetic flux leaving the positron is zero .
hence our positron ( assuming @xmath940 near the earth ) has the expected electric charge but no true magnetic charge .
it has a magnetic dipole moment , which we are currently unable to model
. therefore our model of the positron should behave like a physical positron if we assume that the positron s measured magnetic dipole moment is consistent with our model .
the short answer to this question is no .
the string tension is quantized and therefore can not transmit an increase in force to the positron .
the string is perfectly elastic .
if you pull on it in the direction of the string , you simply increase the string s length and do not affect the positron s motion .
however , if you pull on the string perpendicularly to the string , you can affect the orientation of the string , and thereby change the positron s motion .
so this fact brings out the importance of spin and developing a model of the magnetic moment that will be needed to make this model complete .
following the line of reasoning which led to ( [ eq - dndastar ] ) and hence to ( [ eq - g ] ) , the line pair density coming from a star of arbitrary material is given by @xmath941 where @xmath942 is one atomic mass , @xmath943 is the mass number of the material ( number of nucleons ) and @xmath944 is the relative atomic mass of the material .
the brackets indicate the mass weighted mean and the subscript @xmath945 indicates that the ratio of the line pair density from a star to its mass depends on the material of the active mass ( source of the field ) .
then the value of @xmath946 becomes @xmath947 the size of the material dependence factor is typically @xmath948 .
this factor approaches @xmath689 for hydrogen , but it is difficult to measure @xmath19 with a gas .
there are wild fluctuations@xcite in existing measurements of @xmath19 and this model may be able to help@xcite bring understanding to these measurements . this model can not assess whether or not the gravitational force is proportional to passive mass ( on which the field acts ) because we do not model the nucleus .
the discussion of section [ ssec - calcg ] indicates that when a line pair passes a charged particle , it will exert an attractive force on this particle , and this force will switch directions as the line pair passes .
hence the force of a line pair on a particle _ will _ depend on the velocity of the line pair .
if line pairs in our surroundings dominantly come from stars we can see , the fixed " stars , we expect that the force of the ether on a charged particle _ will _ depend on the velocity of the particle , instead of only on the acceleration of the particle .
however , we have seen that the average line pair ( section [ ssec - estimatelpd ] ) is about a factor of @xmath22 larger than the size of our visible universe , probably because of inflation .
this indicates that the total volume of our universe is about a factor of @xmath949 greater than the volume of our visible universe !
matter which went over our horizon early in the inflationary process will now be moving at velocities far beyond the velocity of light relative to us , but may have stars which are the source of line pairs near us .
thus we postulate that the line pairs near us are moving isotropically with uniformly distributed velocities at least up to the velocity of light .
this causes lorentz invariance .
apparently these line pairs are not accelerating with respect to our inertial system@xcite . c. bennet _ et al .
_ , apjs * 148 * ( 2003 ) , p. 24
. m. k. gaillard , p. d. grannis , and f. j. sciulli , reviews of modern physics * 71 * , s96 ( march 1999 ) .
b. abbott _
et al . _ ,
0 collaboration , phys .
letters * 82 * , 2457 ( 1999 ) .
see , for example , donald h. perkins , _ introduction to high energy physics _ , addison - wesley ( 1987 ) , p. 21 and b. hatfield , _ quantum field theory of point particles and strings _ , addison - wesley ( 1992 ) , p. 479 . n. bohr , phys . rev . * 48 * , 696 ( 1935 ) .
n. bohr , discussion with einstein on epistomological problems in atomic physics " , in _ albert einstein : philosopher - scientist _ , p. a. schilp , ed . ,
library of living philosophers , evanston , ill .
( 1949 ) , p. 232 .
a. einstein , b. podolsky , and n. rosen , phys . rev . * 47 * , 777 , ( 1935 ) .
albert einstein,reply to criticism " , in _ albert einstein : philosopher - scientist _ , p. a. schilp , ed . , library of living philosophers , evanston , ill .
( 1949 ) , p. 672 .
see , for example , n. d. mermin , reviews of modern physics * 65 * , 803 ( 1993 ) .
a. m. dirac , phys . rev . * 74 * , 817 ( 1948 ) .
j. d. jackson , _ classical electrodynamics _ , john wiley & sons , inc . ( 1975 ) .
( we follow the conventions of the second edition rather than the third because the third edition does not use a consistent set of units . )
a. tonomura , _ the quantum world unveiled by electron waves _
, world scientific ( 1998 ) , p. 133 .
d. j. fixsen _
_ , astrophys
. j. * 420 * , 445 ( 1994 ) .
michael r. douglas , jhep 05(2003)46 , p. 55 .
roberto valandro , arxiv:0801.05842v2 , ph .
d. thesis , p. 10
this is a readable introduction to the landscape .
ernst mach , _ the science of mechanics _ , written in may , 1883 , translated by thomas j. mccormack , open court ( 1960 ) , p. 284 .
p. a. m. dirac , proc .
london * a133 * , 60 ( 1931 ) .
the fact that a spinor changes into its negative under a @xmath950 rotation is irrelevant here .
we are not rotating the wave function .
f. london , _ superfluids _ , vol .
i , john wiley & sons , inc .
( 1950 ) , p. 152 .
some parts of our point of view have been expressed by m. h. saffouri , nuovo cimento * 96a * , 1 ( 1986 ) .
we generally follow the line " terminology of faraday and maxwell ( but not line of force " ) .
see j. c. maxwell , _ a treatise on electricity and magnetism _
, preface to the first edition , p. ix , feb . 1 , 1873 ,
dover ( 1954 ) .
j. d. jackson , _ classical electrodynamics _ , john wiley & sons ,
( 1975 ) , pp .
h. lamb , _ hydrodynamics _ , dover ( 1945 ) .
h. helmholtz ( translated by p. g. tait ) , phil . mag . and j. of science * 33 * , 485 ( 1867 ) .
w. thomson , phil . mag .
* 10 * , 155 ( 1880 ) .
w. thomson , phil . mag . and j. of science * 34 * , 15 ( 1867 ) .
a. a. abrikosov , rev .
* 76 * , 975 ( 2004 ) .
d. saint - james , g. sarma , and e. j. thomas , _ type ii superconductivity _ , pergammon ( 1969 ) , p. 48 . t. van duzer and c. w. turner , _ principles of superconductive devices and circuits _ , elsevier ( 1981 ) , p. 310 .
see , for instance , g. arfken , _ mathematical methods for physicists _ , academic press ( 1970 ) , p. 510 .
f. london , _ superfluids _ , vol .
i , john wiley & sons , inc.(1950 ) , p. 29 .
j. h. van vleck , _ the theory of electric and magnetic susceptibilities _ , oxford university press ( 1932 ) ,
g. dorfman , _ diamagnetism and the chemical bond _ , elsevier ( 1965 ) , p. 1 .
a. tonomura , _ the quantum world unveiled by electron waves _
, world scientific ( 1998 ) , p. 123 .
c. kittel , _ introduction to solid state physics _ , @xmath951 edition , pp .
373 - 4 ( 1976 ) .
the fact that @xmath952 as @xmath953 implies that @xmath954 is _ not _ a special point in taking the curl of @xmath54 even though the curl of @xmath955 is proportional to a delta function . equation ( [ eq - bline ] ) correctly represents the singularity at @xmath78 .
the phase field is a particular gauge field .
see c. n. yang , phys .
letters * 33 * , 445 ( 1974 ) and t. t. wu and c. n. yang , phys . rev . * d12 * , 3843 ( 1975 ) .
we presume that the phase fields of two lines add linearly . hence the only regions which are important in determining whether or not the wave functions of the lines are single valued , are the regions with non - zero curl .
f. london , _ superfluids _ , vol .
i , john wiley & sons , inc .
( 1950 ) , p. 64 .
we actually quote the relativistic generalization of the standard london equations .
london gives this relativistic generalization here .
j. d. jackson , _ classical electrodynamics _ , john wiley & sons , inc .
( 1975 ) , p. 118 .
a. tonomura , _ the quantum world unveiled by electron waves _ , world scientific ( 1998 ) ,
s. weinberg , _ gravitation and cosmology _ , john wiley & sons , inc . ( 1972 ) , p. 469 .
philips , _ the physics of stars _ ,
john wiley & sons , inc .
( 1999 ) , p. 62 .
e. w. kolb , _ dynamics of the inflationary era _ , fermilab - conf/303-a ( 1999 ) . albert einstein,autobiographical notes " , in _ albert einstein : philosopher - scientist _ , p. a. schilp , ed .
, library of living philosophers , evanston , ill .
( 1949 ) , p. 25 .
j. c. maxwell , _ a treatise on electricity and magnetism _ , volume ii ( written in 1873 ) , dover ( 1954 ) p.822 .
a. d. linde , _ inflation and quantum cosmology _ , academic press ( 1990 ) , p. 15 .
m. b. hindmarsh and t. w. b. kibble , rep .
. phys . * 58 * , 487 ( 1995 ) . p. j. e. peebles , _ principles of physical cosmology _ , princeton university press ( 1993 ) ,
r. hofstadter , h. r. fechter , and j. a. mcintyre , phys . rev . * 92 * , 978 ( 1953 ) .
s. chakrabarty , d. bandyopadhay , s. pal , phys .
* 78 * , 2898 ( 1997 ) .
a. vilenkin , physics reports * 121 * , 283 ( 1985 ) .
j. d. jackson , _ classical electrodynamics _ , john wiley & sons , inc .
( 1975 ) , p. 228 .
j. d. jackson , _ classical electrodynamics _ , john wiley & sons ,
( 1975 ) , p. 226 .
a tonomura _ et al .
lett . * 51 * , 331 ( 1983 ) .
a. tonomura , _ the quantum world unveiled by electron waves _
, world scientific ( 1998 ) , p. 118 .
r. becker , _ electromagnetic fields and interactions _
, blaisdell ( 1964 ) , p. 94 .
p. goddard , j. goldstone , c. rebbi and c.b .
thorn , nuclear physics * b56 * ( 1973 ) pp .
111 - 114 .
a. o. barut , _ electrodynamics and classical theory of fields and particles _ , dover ( 1964,1980 ) , p. 60 .
j. d. bjorken and s. d. drell , _ relativistic quantum mechanics _
, mcgraw - hill ( 1964 ) , p. 52 .
a. o. barut and g. l. bornzin , nuclear physics * b81 * , 480 ( 1974 ) .
there is a sign mistake in equation ( 12 ) of this reference .
j. d. jackson , _ classical electrodynamics _ , john wiley & sons , inc .
( 1975 ) , p. 222 .
the nonrelativistic approximation certainly fails at the ends of the line .
so ( 4.24 ) should be thought of as valid away from the ends of the line , if the transverse motion is nonrelativistic . c. w. misner , k. s. thorne , and j. a. wheeler , _ gravitation _ , freeman ( 1973 ) , p. 165 .
j. d. bjorken and s. d. drell , _ relativistic quantum fields _ , mcgraw - hill ( 1965 ) , p. 70 .
a. tonomura , _ the quantum world unveiled by electron waves _
, world scientific ( 1998 ) , p. 143 . c. w. misner , k. s. thorne , and j. a. wheeler , _ gravitation _ , freeman ( 1973 ) , p. 53 .
j. e. peebles , _ principles of physical cosmology _ , princeton university press ( 1993 ) ,
m. b. hindmarsh and t. w. b. kibble , rep .
. phys . * 58 * , 486 ( 1995 ) .
j. d. jackson , _ classical electrodynamics _ , john wiley & sons , inc .
( 1975 ) , p. 654 .
j. schwinger , science * 165 * , 757 ( 1969 ) .
j. d. jackson , _ classical electrodynamics _ , john wiley & sons , inc .
( 1975 ) , p. 252 .
j. d. jackson , _ classical electrodynamics _ , john wiley & sons , inc . ( 1975 ) , p. 256 . a. s. goldhaber , phys
. rev . * 140b * , 1407 ( 1965 ) .
we presume , but do not prove , that the response can be derived from ( [ eq - actionnoh ] ) using the superconducting current which is induced on the outer surface of each line pair ( and extending a few penetration depths into the pair ) .
r. becker , _ electromagnetic fields and interactions _ , blaisdell ( 1964 ) , p. 111 . if we approximate a line pair by a series of superconducting spheres we can see that the polarizability per unit length of a line pair should be approximately given by @xmath956 where @xmath957 is the cylindrical outer radius of a line pair .
see figure [ fig - lpstars12 ] .
j. d. jackson , _ classical electrodynamics _ , john wiley & sons , inc .
( 1975 ) , p. 185 .
c. kittel , _ introduction to solid state physics _ , @xmath951 edition , p. 218
fulvio melia , _ the black hole at the center of our galaxy _ , princeton university press ( 2003 ) , p. 49 .
fulvio melia , _ the black hole at the center of our galaxy _ , princeton university press ( 2003 ) ,
this variation can not be measured via the gravitational force because of the weak principle of equivalence .
use of another type of force is required .
m. zeilik , s. gregory , and e. smith , _ introductory astronomy and astrophysics _ ,
saunders ( 1992 ) , p. 279 .
d. elmegreen , _ galaxies and galactic structure _ , prentice hall ( 1998 ) , p. 150 .
f. c. witteborn and w. m. fairbank , phys .
* 19 * , 1049 ( 1967 ) .
b. barish and r. weiss , _ ligo and the detection of gravitational waves _ , physics today * 52 * , 42 ( 1999 ) .
s. weinberg , _ gravitation and cosmology _ , john wiley & sons , inc . ( 1972 ) , p. 68 .
j. d. jackson , _ classical electrodynamics _ , john wiley & sons , inc .
( 1975 ) , p. 503 .
b. f. shutz , _
first course in general relativity _ , cambridge university press ( 1998 ) , p. 214
this idea was first suggested to me by m. c. mccarthy .
a. messiah , _ quantum mechanics _ ,
john wiley & sons , inc .
( 1958 ) , p. 118 .
r. resnick and d. halliday , _
physics for students of science and engineering _ , john wiley & sons , inc .
( 1961 ) , p. 515 .
s. weinberg , _ gravitation and cosmology _ , john wiley & sons , inc . ( 1972 ) , p. 472 .
fulvio melia , _ the black hole at the center of our galaxy _ , princeton university press ( 2003 ) ,
c. n. yang , colloquium , may 4 , 1999 , state university of new york at stony brook . j. mydosh , _ spin glasses : an experimental introduction _ , taylor and francis ( 1993 ) , p. 144 .
elliot lieb , colloquium , march 2 , 1999 , state university of new york at stony brook .
leon cooper , private communication .
i. aitchison and a. hey , _
gauge theories in particle physics _ , adam hilger ltd .
( 1982 ) , p. 205 .
klaus von klitzing , rev .
phys . * 58 * , 522 ( 1985 ) .
j. k. jain and r. k. kamila , _ composite fermions : particles of the lowest landau level _ ,
in _ composite fermions _ , ed . by o heinonen ,
world scientific ( 1998 ) .
a. tonomura _
et al . _ ,
lett . * 51 * , 331 ( 1983 ) .
a. tonomura , _ electron holography _ , springer ( 1998 ) .
matthew sands , private communication .
quinn , nature(london ) * 408 * 919 ( 2000 ) .
r. mccarthy , in preparation for physical review . | this is an attempt to construct a _ classical _ microscopic model of the electron which underlies quantum mechanics .
an electron is modeled , not as a point particle , but as the end of an electromagnetic string , a line of flux .
these lines stretch across cosmic distances , but are almost unobservable because they condense into pairs which form the ether .
photons are modeled to propagate on these line pairs , which act effectively as wave guides .
these line pairs are also responsible for the force of gravity which is electromagnetic in character . using our classical interpretation of an electron as a quantized electromagnetic field
, we find that we achieve a natural interpretation of not only quantum mechanics , but also mass , spin , mach s principle , gravity , self interactions , newton s second law of motion , the quantization of the photon , wave - particle duality , inflation , dark energy , agn jets , the pauli exclusion principle , superconductivity , and the quantum hall effects .
mass generation and inflation are discussed without the use of scalar fields .
we attempt to calculate newton s gravitational constant g , the mass of the electron , the spin of the electron , and the spin of the photon all in terms of planck s constant , the charge of the electron , the outer radius of a line , and the distribution of matter in the universe .
the predictions of this model and the standard model are both compared to experiment , emphasizing areas where the predictions differ .
this model s remaining problems are discussed .
this model is a classical field theory of interacting strings which explains a large number of diverse phenomena , even though it does not discuss the strong or weak interactions .
since the ether of this model is quite well determined , it might be useful in restricting the landscape of vacua in more complete string theories . |
this is an addition to a series of papers @xcite , where we develop quaternionic analysis from the point of view of representation theory of the conformal lie group and its lie algebra . since the paper @xcite serves as a starting point for this article , we highlight the most relevant results .
let @xmath3 denote the algebra of quaternions , @xmath4 and @xmath5 then let @xmath6\ ] ] be the space of polynomial functions on @xmath7 .
we can think of @xmath8 as a quaternionic analogue of laurent polynomials @xmath9 $ ] . there is a natural action of the lie algebra @xmath1 on @xmath8 denoted by @xmath10 ( see lemma [ rho - algebra - action ] ) . according to theorem 7 in @xcite ( restated here as theorem [ thm7 ] )
, @xmath11 decomposes into three irreducible components : @xmath12 this decomposition is followed by a derivation of projectors of @xmath8 onto its irreducible components @xmath13 , @xmath14 and @xmath15 .
these projectors are regarded as quaternionic analogues of the cauchy s formula for the second order pole @xmath16 the cases of @xmath13 and @xmath15 are fairly straightforward . on the other hand , the case of @xmath14
is more subtle and requires a certain regularization of infinities which is well known in four - dimensional quantum field theory as `` scalar vacuum polarization '' .
this vacuum polarization is usually encoded by a feynman diagram shown in figure [ vacuum ] and plays a key role in renormalization theory ( see , for example , @xcite ) .
corollary 6 in @xcite ( restated here as proposition [ cor6 ] ) asserts that @xmath17 where @xmath18 s are the matrix coefficients of @xmath19 . in this article
we replace the matrix coefficients of @xmath19 with the matrix coefficients of the discrete series representations of @xmath20 denoted by @xmath21 and consider a space @xmath22 ( we denote by @xmath23 and @xmath24 the subspaces of ( [ split - zh ] ) spanned by @xmath25 , where @xmath21 s are the matrix coefficients of holomorphic and antiholomorphic discrete series respectively . )
we regard the space ( [ split - zh ] ) as a split quaternionic analogue of @xmath8 .
our first result is theorem [ decomposition - thm ] , which shows that this space is invariant under the @xmath10-action of @xmath2 and decomposes into six irreducible components that we denote by @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 and @xmath31 .
then in proposition [ kernel - prop ] we find the reproducing kernels for each of these irreducible components .
finally , in theorem [ projector - thm ] we obtain the @xmath2-equivariant projectors onto the irreducible components of the space ( [ split - zh ] ) .
this result can be regarded as a split analogue of corollary 14 and theorem 15 in @xcite . perhaps surprisingly , unlike the case of @xmath14
, there are absolutely no convergence issues for the reproducing kernels or the integral operators realizing the projectors onto the irreducible components .
let @xmath32 denote the space of complexified quaternions : @xmath33 , it can be identified with the algebra of @xmath34 complex matrices : @xmath35 for @xmath36 , we write @xmath37 and think of it as the norm of @xmath38 .
let @xmath39 be the group of invertible complexified quaternions , @xmath40 .
let @xmath41 denote the space of @xmath42-valued functions on @xmath32 ( possibly with singularities ) which are holomorphic with respect to the complex variables @xmath43 .
we recall the action of @xmath44 on @xmath41 given by equation ( 49 ) in @xcite : @xmath45 differentiating this action , we obtain an action of the lie algebra @xmath46 , which we still denote by @xmath10 .
[ rho - algebra - action ] let @xmath47 , where @xmath48 . the lie algebra action @xmath10 of @xmath2 on @xmath41 is given by @xmath49 this lemma implies that @xmath50 preserves the spaces
@xmath51 \qquad \qquad \text{and } \\ { \textit{{{\fontencoding{ot2}\fontfamily{wncyr}\selectfont zh}}}}&= \{\text{polynomial functions on $ { { \mathbb}h_{{\mathbb}c}}^{\times}$}\ } = { \mathbb}c[z_{11},z_{12},z_{21},z_{22},n(z)^{-1}].\end{aligned}\ ] ] recall that @xmath18 s denote the matrix coefficients of @xmath19 described by equation ( 27 ) of @xcite ( cf
. @xcite ) .
these are polynomial functions on @xmath32 .
the parameters range as follows : @xmath52 we recall some results from @xcite .
[ cor6 ] the functions @xmath53 form a vector space basis of @xmath54 $ ] .
[ thm7 ] the representation @xmath11 of @xmath2 has the following decomposition into irreducible components : @xmath55 where @xmath56 ( see figure [ decomposition - fig ] ) .
( 110,70 ) ( 10,10)(10,0)10 ( 10,20)(10,0)10 ( 10,30)(10,0)10 ( 10,40)(10,0)10 ( 10,50)(10,0)10 ( 10,60)(10,0)10 ( 60,0)(0,1)70 ( 0,10)(1,0)110 ( 58,10)(0,1)55 ( 60,8)(1,0)45 ( 52,10)(0,1)55 ( 5,8)(1,0)35 ( 41.4,11.4)(-1,1)36.4 ( 48.6,8.6)(-1,1)43.6 ( 58,10)(58,8)(60,8 ) ( 40,8)(43.8,8)(41.4,11.4 ) ( 52,10)(52,5.2)(48.6,8.6 ) ( 62,67)@xmath57 ( 107,12)@xmath58 ( 3,23)@xmath13 ( 23,63)@xmath14 ( 102,43)@xmath15
in this section we essentially replace the classical quaternions @xmath3 with split quaternions @xmath59 by replacing the matrix coefficients @xmath60 of @xmath19 with the matrix coefficients of @xmath20 , which we denote by @xmath61 in order to avoid conflict of notations .
we are primarily interested in a real form of @xmath32 called split quaternions : @xmath62 this algebra is studied in detail in @xcite .
the `` unit sphere '' in @xmath59 is a hyperboloid @xmath63 which is a lie group isomorphic to @xmath64 .
another important group sitting inside @xmath32 is @xmath65 we orient @xmath66 so that @xmath67 is a positive basis of the tangent space at @xmath68 ( same orientation as in @xcite ) , where @xmath69 we regard groups @xmath70 and @xmath71 as subgroups of @xmath44 : @xmath72 @xmath73 the matrix coefficients of the discrete series representations of @xmath20 are rational functions on @xmath32 described by equation ( 13 ) of @xcite ( cf .
@xcite ) : @xmath74 where the integral is taken over the unit circle @xmath75 traversed once in the counterclockwise direction .
the parameters @xmath76 , @xmath77 and @xmath78 range over @xmath79 when @xmath80 , we get the matrix coefficients of the holomorphic discrete series representations . and
when @xmath81 , we get the matrix coefficients of the antiholomorphic discrete series representations . while the integral expressions for @xmath18 s and @xmath21 s are exactly the same , the indices @xmath76 , @xmath77 and @xmath78 have different ranges , and we insist on using different notations , since the former are matrix coefficients of @xmath19 and the latter are matrix coefficients of @xmath20 .
we define split analogues of the space @xmath8 : @xmath82 ( @xmath83 stands for `` holomorphic '' and @xmath84 stands for `` antiholomorphic '' ) . as in subsection 3.6 in @xcite
, we define a @xmath2-invariant symmetric bilinear pairing for functions on @xmath0 @xmath85 where @xmath86 is a holomorphic 4-form on @xmath32 @xmath87 we have the following orthogonality relations : @xmath88 where the indices @xmath89 are @xmath90 , @xmath91 , @xmath92 or @xmath93 , @xmath94 and similarly for @xmath95 .
in particular , the functions @xmath96 s are linearly independent . since @xmath97 is proportional to @xmath98 , the functions @xmath99 indeed lie in @xmath0 .
hence the spaces @xmath23 and @xmath24 are dual to each other under the pairing ( [ pairing ] ) .
in this section we discuss a split analogue of theorem [ thm7 ] : [ decomposition - thm ] the spaces @xmath23 and @xmath24 are invariant under the @xmath10-action of @xmath2 .
we have the following decompositions into irreducible components : @xmath100 where @xmath101 ( see figure [ decomposition - fig2 ] ) .
( 110,70 ) ( 10,10)(10,0)10 ( 10,20)(10,0)10 ( 10,30)(10,0)10 ( 10,40)(10,0)10 ( 50,0)(0,1)70 ( 0,60)(1,0)110 ( 5,42)(1,0)35 ( 42,5)(0,1)35 ( 40,42)(42,42)(42,40 ) ( 48,5)(0,1)35 ( 51.4,41,4)(1,-1)36.4 ( 48,40)(48,44.8)(51.4,41.4 ) ( 60,42)(1,0)45 ( 58.6,38.6)(1,-1)33.6 ( 60,42)(55.2,42)(58.6,38.6 ) ( 52,67)@xmath57 ( 106,62)@xmath58 ( 7,45)@xmath26 and @xmath29 ( 70,45)@xmath28 and @xmath31 ( 53,1)@xmath27 and @xmath30 since the proof is essentially the same as that of theorem 7 in @xcite , we only give an outline . note that the basis elements of @xmath23 and @xmath24 consist of functions of the kind @xmath102 where the functions @xmath103 range over a basis of harmonic functions which are homogeneous of degree @xmath57 .
recall that we consider @xmath70 and @xmath71 as subgroups of @xmath44 via ( [ u(1,1)xu(1,1 ) ] ) and ( [ su(1,1)xsu(1,1 ) ] ) . for
@xmath58 and @xmath76 fixed , these functions span an irreducible representation of the lie algebra @xmath104 , which remains irreducible when restricted to @xmath105 , belongs to the holomorphic discrete series when @xmath92 and the antiholomorphic discrete series when @xmath93 . 0.2 0.2 0.2 0.2 by lemma [ rho - algebra - action ] , matrices of the kind @xmath106 with @xmath107 act by @xmath108 and , by equation ( 27 ) in @xcite , @xmath109 with @xmath110 and @xmath111 being harmonic and having degrees @xmath112 and @xmath113 respectively . using lemma [ rho - algebra - action ] again , matrices of the kind @xmath114 with @xmath115 act by @xmath116 and , by equation ( 28 ) in @xcite , @xmath117 with @xmath118 and @xmath119 being harmonic and having degrees @xmath113 and @xmath112 respectively .
since @xmath120 it follows that @xmath121 where @xmath122 hence @xmath118 and its conjugate @xmath111 become zero if and only if @xmath123 .
the actions of @xmath124 , @xmath125 , @xmath126 and @xmath127 are illustrated in figure [ actions ] . in the diagram
describing @xmath128 the vertical arrow disappears when @xmath129 and the diagonal arrow disappears when @xmath130 or @xmath123 .
similarly , in the diagram describing @xmath131 the vertical arrow disappears when @xmath123 or @xmath132 and the diagonal arrow disappears when @xmath133 .
this proves that the spaces @xmath23 , @xmath26 , @xmath27 , @xmath28 , @xmath24 , @xmath29 , @xmath30 and @xmath31 are @xmath2-invariant .
this also proves that @xmath26 , @xmath27 and @xmath28 are irreducible @xmath2-invariant subspaces of @xmath23 and that @xmath29 , @xmath30 and @xmath31 are irreducible @xmath2-invariant subspaces of @xmath24 .
theorem 53 in @xcite gives explicit isomorphisms of @xmath2-modules @xmath134 one immediately notices the similarity of figures [ decomposition - fig ] and [ decomposition - fig2 ] illustrating the decompositions of theorems [ thm7 ] and [ decomposition - thm ] respectively .
the figures look similar because the matrix coefficients @xmath18 s of @xmath19 and @xmath21 s of @xmath20 span irreducible representations of the diagonal subalgebras for @xmath76 fixed , are homogeneous of degree @xmath57 , except @xmath135 in the case of @xmath19-coefficients and @xmath136 in the case of @xmath20-coefficients .
the `` border conditions '' separating irreducible components @xmath137 or @xmath138 and @xmath139 or @xmath140 are formally the same in the cases of @xmath8 , @xmath23 and @xmath24 .
in this section we obtain the reproducing kernels for the irreducible components of @xmath23 and @xmath24 . recall that @xmath141 and @xmath142 are certain open olshanskii semigroups lying inside @xmath143 which were described in subsection 3.3 in @xcite . for our purposes
they can be defined as @xmath144 @xmath145 we have the following matrix coefficient expansions for the reproducing kernels of @xmath27 and @xmath30 respectively : @xmath146 which converges pointwise absolutely whenever @xmath147 and @xmath148 which converges pointwise absolutely whenever @xmath149 .
the sums are taken first over all @xmath77 , @xmath78 , then over @xmath150 , and then over @xmath151 .
[ kernel - prop ] suppose that @xmath152 is @xmath20-conjugate to a diagonal matrix @xmath153 .
we have the following matrix coefficient expansions for the reproducing kernels of @xmath26 , @xmath28 , @xmath29 and @xmath31 respectively : 1 .
[ 1 ] when @xmath147 and @xmath154 we have a pointwise absolutely convergent series @xmath155 2 .
when @xmath147 and @xmath156 we have a pointwise absolutely convergent series @xmath157 3 .
when @xmath149 and @xmath154 we have a pointwise absolutely convergent series @xmath158 4 .
when @xmath149 and @xmath156 we have a pointwise absolutely convergent series @xmath159 we will prove part _ [ 1 ] _ only , the other parts can be proved in the same way .
the proof is similar to that of proposition 54 in @xcite . since @xmath147 , @xmath152 is @xmath20-conjugate to a diagonal matrix @xmath160 for some @xmath161 with @xmath162 . using equations ( 33 ) and ( 34 ) from @xcite and letting the indices @xmath58
, @xmath76 , @xmath77 , @xmath78 run over @xmath163 we obtain : @xmath164 from these expansions of reproducing kernels and orthogonality relations ( [ orthogonality ] ) we immediately obtain the following extension of theorem 57 from @xcite , which can be regarded as a split analogue of corollary 14 and theorem 15 in @xcite . 1 .
if @xmath167 and @xmath168 , this operator provides the projector onto @xmath27 ; 2 .
if @xmath167 and @xmath169 , this operator provides the projector onto @xmath30 ; 3 .
if @xmath170 and @xmath168 , @xmath171 , this operator provides the projector onto @xmath26 ; 4 .
if @xmath172 and @xmath168 , @xmath173 , this operator provides the projector onto @xmath28 ; 5 .
if @xmath174 and @xmath169 , @xmath171 , this operator provides the projector onto @xmath29 ; 6 .
if @xmath175 and @xmath169 , @xmath173 , this operator provides the projector onto @xmath31 ; [ fl1 ] i. frenkel , m. libine , _ quaternionic analysis , representation theory and physics _
, advances in math * 218 * ( 2008 ) , 1806 - 1877 ; also arxiv:0711.2699 .
i. frenkel , m. libine , _ split quaternionic analysis and the separation of the series for @xmath177 and @xmath178 _ , advances in math * 228 * ( 2011 ) , 678 - 763 ; also arxiv:1009.2532 .
i. frenkel , m. libine , _ quaternionic analysis and the schrdinger model for the minimal representation of @xmath179 _ , imrn , 2012 ( 2012 ) , 4904 - 4923 ; also arxiv:1110.2106 . i. frenkel , m. libine , _ anti de sitter deformation of quaternionic analysis and the second - order pole _ , to appear in imrn , doi : 10.1093/imrn / rnu083 ; also arxiv:1404.7098 . v. a. smirnov , _ evaluating feynman integrals _ , springer tracts in modern physics , 211 , springer - verlag , berlin , 2004 . n. ja .
vilenkin , _ special functions and the theory of group representations _ , translated from the russian by v. n. singh , translations of mathematical monographs , vol . 22 american mathematical society , providence , ri 1968 . | this is an addition to a series of papers @xcite , where we develop quaternionic analysis from the point of view of representation theory of the conformal lie group and its lie algebra . in this paper
we develop split quaternionic analogues of certain results from @xcite .
thus we introduce a space of functions @xmath0 with a natural action of the lie algebra @xmath1 , decompose @xmath0 into irreducible components and find the @xmath2-equivariant projectors onto each of these irreducible components . |
complex network studies have traditionally focused on single networks in which nodes represent agents and links represent the connections between agents .
recent efforts have focused on complex systems that are comprised of interconnected networks , a configuration that more accurately represents real - world networks @xcite .
real - world power grids , for example , are almost always coupled with communication networks .
power stations need communication nodes for control and communication nodes need power stations for electricity .
the influence of coupled networks on cascading failures has been widely studied @xcite .
when a node at one end of an interdependent link fails , the node at the other end of the link usually fails . a non - consensus opinion model of two interconnected networks that allows the opinion interaction rules within each individual network to differ from those between the networks
was recently studied @xcite .
this model shows that opinion interactions between networks can transform non - consensus opinion behavior into consensus opinion behavior . in this paper
we investigate the susceptable - infected - susceptable ( sis ) behavior of a spreading virus , a dynamic process in interconnected networks that has received significant recent attention @xcite .
an interconnected networks scenario is essential when modeling epidemics because diseases spread across multiple networks , e.g. , across multiple species or communities , through both contact network links within each species or community and interconnected network links between them .
dickison et al .
@xcite study the behavior of susceptible - infected - recovered ( sir ) epidemics in interconnected networks . depending on the infection rate in weakly and strongly coupled network systems , where each individual network follows the configuration model and interconnections
are randomly placed , epidemics will infect none , one , or both networks of a two - network system .
mendiola et al .
@xcite show that in sis model an endemic state may appear in the coupled networks even when an epidemic is unable to propagate in each network separately . in this work
we will explore how both the structural properties of each individual network and the behavior of the interconnections between them determine the epidemic threshold of two generic interconnected networks . in order to represent two generic interconnected networks ,
we represent a network @xmath16 with @xmath1 nodes using an @xmath17 adjacency matrix @xmath18 that consists of elements @xmath19 , which are either one or zero depending on whether there is a link between nodes @xmath13 and @xmath14 . for the interconnected networks , we consider two individual networks @xmath9 and @xmath10 of the same size @xmath1 .
when nodes in @xmath9 are labeled from @xmath20 to @xmath1 and in @xmath10 labeled from @xmath21 to @xmath22 , the two isolated networks @xmath9 and @xmath10 can be presented by a @xmath0 matrix @xmath23{cc}a_{1 } & \mathbf{0}\\ \mathbf{0 } & a_{2}\end{array } \right ] $ ] composed of their corresponding adjacency matrix @xmath18 and @xmath24 respectively .
similarly , a @xmath0 matrix @xmath25{cc}\mathbf{0 } & b_{12}\\ b_{12}^{t } & \mathbf{0}\end{array } \right ] $ ] represents the symmetric interconnections between @xmath9 and @xmath10 .
the interconnected networks are composed of three network components : network @xmath18 , network @xmath24 , and interconnecting network @xmath26 . in the sis model ,
the state of each agent at time @xmath27 is a bernoulli random variable , where @xmath28 if node @xmath13 is susceptible and @xmath29 if it is infected .
the recovery ( curing ) process of each infected node is an independent poisson process with a recovery rate @xmath30 .
each infected agent infects each of its susceptible neighbors with a rate @xmath3 , which is also an independent poisson process .
the ratio @xmath31 is the effective infection rate .
a phase transition has been observed around a critical point @xmath32 in a single network . when @xmath33 , a non - zero fraction of agents will be infected in the steady state , whereas if @xmath34 , infection rapidly disappears @xcite .
the epidemic threshold via the n - intertwined mean - field approximation ( nimfa ) is @xmath35 , where @xmath36 is the largest eigenvalue of the adjacency matrix , also called the spectral radius @xcite . for interconnected networks ,
we assume that the curing rate @xmath30 is the same for all the nodes , that the infection rate along each link of @xmath9 and @xmath10 is @xmath3 , and that the infection rate along each interconnecting link between @xmath9 and @xmath10 is @xmath4 , where @xmath8 is a real constant ranging within @xmath37 without losing generality .
we first show that the epidemic threshold for @xmath38 in interconnected networks via nimfa is @xmath39 , where @xmath7 is the largest eigenvalue of the matrix @xmath40 .
we further express @xmath7 as a function of each network component @xmath18 , @xmath24 , and @xmath26 and their eigenvalues / eigenvectors to reveal the contribution of each component network .
this is a significant mathematical challenge , except for special cases , e.g. , when @xmath41 and @xmath26 commute , i.e. , @xmath42 ( see sec . [ special cases ] ) .
our main contribution is that we analytically derive for the epidemic characterizer @xmath43 ( a ) its perturbation approximation for small @xmath8 , ( b ) its perturbation approximation for large @xmath8 , and ( c ) its lower and upper bound for any @xmath8 as a function of component network @xmath18 , @xmath24 , and @xmath26 and their the largest eigenvalues / eigenvectors . numerical simulations in sec .
[ simulation ] verify that these approximations and bounds well approximate @xmath7 , and thus reveal the effect of component network features on the epidemic threshold of the whole system of interconnected networks , which provides essential insights into designing interconnected networks that are robust against the spread of epidemics ( see sec . [ conclusion ] ) .
sahneh et al .
@xcite recently studied sis epidemics on generic interconnected networks in which the infection rate can differ between @xmath9 and @xmath10 , and derived the epidemic threshold for the infection rate in one network while assuming that the infection does not survive in the other .
their epidemic threshold was expressed as the largest eigenvalue of a function of matrices .
our work explains how the epidemic threshold of generic interconnected networks is related to the properties ( eigenvalue / eigenvector ) of each network component @xmath18 , @xmath24 , and @xmath26 without any approximation on the network topology .
graph spectra theory @xcite and modern network theory , integrated with dynamic systems theory , can be used to understand how network topology can predict these dynamic processes .
youssef and scoglio @xcite have shown that a sir epidemic threshold via nimfa also equals @xmath44 .
the kuramoto synchronization process of coupled oscillators @xcite and percolation @xcite also features a phase transition that specifies the onset of a remaining fraction of locked oscillators and the appearance of a giant component , respectively . note that a mean - field approximation predicts both phase transitions at a critical point that is proportional to @xmath44 .
thus we expect our results to apply to a wider range of dynamic processes in interconnected networks .
in the sis epidemic spreading process , the probability of infection @xmath45 $ ] for a node @xmath13 in interconnected networks @xmath16 is described by @xmath46 via nimfa , where @xmath19 and @xmath47 is an element of matrix @xmath41 and @xmath26 respectively .
its matrix form becomes @xmath48 the governing equation of the sis spreading process on a single network @xmath18 is @xmath49 whose epidemic threshold has been proven @xcite to be @xmath50 which is a lower bound of the epidemic threshold @xcite .
hence , the epidemic threshold of interconnected networks by nimfa is @xmath51 which depends on the largest eigenvalue of the matrix @xmath6 .
the matrix @xmath6 is a weighted matrix , where @xmath52 .
note that the nimfa model is an improvement over earlier epidemic models @xcite in that it applies no approximations to network topologies , and thus it allows us to identify the specific role of a general network structure on the spreading process .
the spectral radius @xmath7 as shown in the last section is able to characterize epidemic spreading in interconnected networks . in this section
we explore how @xmath7 is influenced by the structural properties of interconnected networks and by the relative infection rate @xmath8 along the interconnection links .
specifically , we express @xmath7 as a function of the component network @xmath18 , @xmath24 , and @xmath26 and their eigenvalues / eigenvectors .
( for proofs of theorems or lemma , see the appendix . )
we start with some basic properties related to @xmath7 and examine several special cases in which the relation between @xmath43 and the structural properties of network components @xmath18 , @xmath24 and @xmath26 are analytically tractable .
the spectral radius of a sub - network is always smaller or equal to that of the whole network .
hence , @xmath53 [ lowerboundasymptotic]@xmath54 the interconnection network @xmath26 forms a bipartite graph . [ bipartite ] the largest eigenvalue of a bipartite graph @xmath25{cc}\mathbf{0 } & b_{12}\\ b_{12}^{t } & \mathbf{0}\end{array } \right ] $ ] follows @xmath55 where @xmath56 is possibly asymmetric @xcite .
[ lemma_regular ] when @xmath9 and @xmath10 are both regular graphs with the same average degree @xmath57 $ ] and when any two nodes from @xmath9 and @xmath10 respectively are randomly interconnected with probability @xmath58 , the average spectral radius of the interconnected networks follows @xmath59=e[d]+\alpha np_{i}\ ] ] if the interdependent connections are not sparse . a dense erds - rnyi ( er ) random network approaches a regular network when @xmath1 is large .
lemma [ lemma_regular ] , thus , can be applied as well to cases where both @xmath9 and @xmath10 are dense er random networks . for any two commuting matrices @xmath41 and @xmath26 , thus @xmath42 , @xmath60 @xcite .
this property of commuting matrices makes the following two special cases analytically tractable .
[ aaii ] when @xmath61{cc}a_{1 } & \mathbf{0}\\ \mathbf{0 } & a_{1}\end{array } \right ] + \alpha\left [ \begin{array } [ c]{cc}\mathbf{0 } & i\\ i & \mathbf{0}\end{array } \right ] $ ] , i.e. , the interconnected networks are composed of two identical networks , where one network is indexed from @xmath20 to @xmath1 and the other from @xmath21 to @xmath22 , with an interconnecting link between each so - called image node pair @xmath62 from the two individual networks respectively , its largest eigenvalue @xmath63 . when @xmath61{cc}a_{1 } & \mathbf{0}\\ \mathbf{0 } & a_{1}\end{array } \right ] + \alpha\left [ \begin{array } [ c]{cc}\mathbf{0 } & i\\ i & \mathbf{0}\end{array } \right ] , $ ] matrix @xmath41 and @xmath64 are commuting @xmath65{cc}\mathbf{0 } & a_{1}\\ a_{1 } & \mathbf{0}\end{array } \right ] = \alpha ba\ ] ] therefore , @xmath66 .
the network @xmath26 is actually a set of isolated links .
hence , @xmath67 .
[ aaaa]when @xmath61{cc}a_{1 } & \mathbf{0}\\ \mathbf{0 } & a_{1}\end{array } \right ] + \alpha\left [ \begin{array } [ c]{cc}\mathbf{0 } & a_{1}\\ a_{1 } & \mathbf{0}\end{array } \right ] $ ] , its largest eigenvalue @xmath68 . when @xmath61{cc}a_{1 } & \mathbf{0}\\ \mathbf{0 } & a_{1}\end{array } \right ] + \alpha\left [ \begin{array } [ c]{cc}\mathbf{0 } & a_{1}\\ a_{1 } & \mathbf{0}\end{array } \right ] , $ ] matrix @xmath41 and @xmath64 are commuting @xmath65{cc}\mathbf{0 } & a_{1}^{2}\\ a_{1}^{2 } & \mathbf{0}\end{array } \right ] = \alpha ba\ ] ] therefore @xmath69 .
when @xmath41 and @xmath26 are not commuting , little can be known about the eigenvalues of @xmath7 , given the spectrum of @xmath41 and of @xmath26 .
for example , even when the eigenvalue of @xmath41 and @xmath26 are known and bounded , the largest eigenvalue of @xmath7 can be unbounded @xcite .
we now denote matrix @xmath6 to be @xmath71 . applying the rayleigh inequality @xcite to the symmetric matrix @xmath72 yields @xmath73 where equality holds only if @xmath74 is the principal eigenvector of @xmath71 .
[ theorem_lowerbound1 ] the best possible lower bound @xmath75 of interdependent networks @xmath71 by choosing @xmath74 as the linear combination of @xmath11 and @xmath12 , the largest eigenvector of @xmath18 and @xmath24 respectively , is @xmath76 where @xmath77 . when @xmath78 , the lower bound becomes the exact solution @xmath79 .
when the two individual networks have the same largest eigenvalue @xmath80 , we have @xmath81 [ theorem_lowerbound2 ] the best possible lower bound @xmath82 by choosing @xmath74 as the linear combination of @xmath11 and @xmath12 , the largest eigenvector of @xmath18 and @xmath24 respectively , is @xmath83 where @xmath84 . in general , @xmath85 the largest eigenvalue is lower bounded by @xmath86 [ theorem_higherorder ] given a vector @xmath74 , @xmath87when @xmath88 is an even integer and @xmath89 .
furthermore , @xmath90 hence , given a vector @xmath74 , we could further improve the lower bound @xmath91 by taking a higher even power @xmath88 .
note that theorem [ theorem_lowerbound1 ] and [ theorem_lowerbound2 ] express the lower bound as a function of component network @xmath18 , @xmath24 and @xmath26 and their eigenvalues / eigenvectors , which illustrates the effect of component network features on the epidemic characterizer @xmath92 .
[ theorem_upperbound ] the largest eigenvalue of interdependent networks @xmath93 is upper bounded by @xmath94 this upper bound is reached when the principal eigenvector of @xmath95 coincides with the principal eigenvector of @xmath96 if @xmath97 and when the principal eigenvector of @xmath98 coincides with the principal eigenvector of @xmath99 if @xmath100 in this subsection , we derive the perturbation approximation of @xmath101 for small and large @xmath8 , respectively , as a function of component networks and their eigenvalues / eigenvectors .
we start with small @xmath8 cases .
the problem is to find the largest eigenvalue @xmath102 of @xmath71 , with the condition that @xmath103{c}(w-\lambda i)z=0\\ z^{t}z=1 \end{array } \right.\ ] ] when the solution is analytical in @xmath8 , we express @xmath104 and @xmath74 by taylor expansion as @xmath105 substituting the expansion in the eigenvalue equation gives @xmath106 where all the coefficients of @xmath107 on the left must equal those on the right . performing the products and reordering the series we obtain @xmath108 this leads to a hierarchy of equations @xmath109 the same expansion must meet the normalization condition @xmath110 or equivalently , @xmath111 where @xmath112 represents the scalar product .
the normalization condition leads to a set of equations @xmath113 for any @xmath114 and @xmath115 .
let @xmath116 and @xmath117 denote the largest eigenvalue and the corresponding eigenvector of @xmath118 respectively .
we examine two possible cases : ( a ) the non - degenerate case when @xmath119 and ( b ) the degenerate case when @xmath120 and the case @xmath121 is equivalent to the first .
[ theorem_perturbation_nondegenerate ] for small @xmath8 , in the non - degenerate case , thus when @xmath119 , @xmath122 where @xmath123{cc}x^{t } & \mathbf{0}^{t}\end{array } \right ) $ ] .
note that in ( [ secondorder_a ] ) @xmath26 is symmetric and @xmath124 is positive definite and so is @xmath125 .
hence , this second order correction @xmath126 is always positive .
[ theorem_perturbation_degenerate ] for small @xmath8 , when the two component networks have the same largest eigenvalue @xmath127 , @xmath128 in the degenerate case , the first order correction is positive and the slope depends on @xmath56 , @xmath12 , and @xmath11 . when @xmath18 and @xmath24 are identical , the largest eigenvalue of the interdependent networks becomes @xmath129 when @xmath23{cc}a_{1 } & \mathbf{0}\\ \mathbf{0 } & a_{1}\end{array } \right ] $ ] and @xmath25{cc}\mathbf{0 } & i\\ i & \mathbf{0}\end{array } \right ] $ ] , our result ( [ degenerate ] ) in the degenerate case up to the first order leads to @xmath63 , which is an alternate proof of lemma [ aaii ] . when @xmath23{cc}a_{1 } & \mathbf{0}\\ \mathbf{0 } & a_{1}\end{array } \right ] $ ] and @xmath25{cc}\mathbf{0 } & a_{1}\\ a_{1 } & \mathbf{0}\end{array } \right ] $ ] , ( [ degenerate ] ) again explains lemma [ aaaa ] that @xmath130 .
[ alphainfinity ] for large @xmath8 , the spectral radius of interconnected networks is @xmath131 where @xmath132 is the eigenvector belonging to @xmath133 and @xmath134 the lemma [ alphainfinity ] follows by applying perturbation theory @xcite to the matrix @xmath135 and the rayleigh principle @xcite , which states that @xmath136 , for any normalized vector @xmath132 such that @xmath137 , with equality only if @xmath132 is the eigenvector belonging to the eigenvalue @xmath36 .
in this section , we employ numerical calculations to quantify to what extent the perturbation approximation ( [ perturbation_non ] ) and ( [ degenerate ] ) for small @xmath8 , the perturbation approximation ( [ perturb_alarge ] ) for large @xmath8 , the upper ( [ upperbound ] ) and lower bound ( [ lowerbound2 ] ) are close to the exact value @xmath138 .
we investigate the condition under which the approximations provide better estimates .
the analytical results derived earlier are valid for arbitrary interconnected network structures . for simulations ,
we consider two classic network models as possible topologies of @xmath9 and @xmath10 : ( i ) the erds - rnyi ( er ) random network @xcite and ( ii ) the barabsi - albert ( ba ) scale - free network @xcite .
er networks are characterized by a binomial degree distribution @xmath139=\binom{n-1}{k}p^{k}(1-p)^{n-1-k}$ ] , where @xmath1 is the size of the network and @xmath140 is the probability that each node pair is randomly connected . in scale - free networks ,
the degree distribution is given by a power law @xmath139=ck^{-\lambda}$ ] such that @xmath141 and @xmath142 in ba scale - free networks .
in numerical simulations , we consider @xmath143 .
specifically , in the ba scale - free networks @xmath144 and the corresponding link density is @xmath145 .
we consider er networks with the same link density @xmath146 .
a coupled network @xmath16 is the union of @xmath9 and @xmath10 , which are chosen from the above mentioned models , together with random interconnection links with density @xmath58 , the probability that any two nodes from @xmath9 and @xmath10 respectively are interconnected .
given the network models of @xmath9 and @xmath10 and the interacting link density @xmath58 , we generate 100 interconnected network realizations . for each realization , we compute the spectral radius @xmath93 , its perturbation approximation ( [ perturbation_non ] ) and ( [ degenerate ] ) for small @xmath8 , the perturbation approximation ( [ perturb_alarge ] ) for large @xmath8 , upper bound ( [ upperbound ] ) and lower bound ( [ lowerbound2 ] ) for any @xmath8 .
we compare their averages over the 100 coupled network realizations .
we investigate the degenerate case @xmath147 where the largest eigenvalue of @xmath9 and @xmath10 are the same and the non - degenerate case where @xmath148 respectively .
baerasymp we consider the non - degenerate case in which @xmath9 is a ba scale - free network with @xmath149 , @xmath10 is an er random network with the same size and link density @xmath150 , and the two networks are randomly interconnected with link density @xmath58 .
we compute the largest average eigenvalue @xmath151 $ ] and the average of the perturbation approximations and bounds mentioned above over 100 interconnected network realizations for each interconnection link density @xmath152 $ ] such that the average number of interdependent links ranges from @xmath153 to @xmath154 and for each value @xmath8 that ranges from @xmath155 to @xmath156 with step size @xmath157 . for a single ba scale - free network , where the power exponent @xmath158 , the largest eigenvalue is @xmath159 where @xmath160 is the maximum degree in the network @xcite . the spectral radius of a single er random graph is close to the average degree @xmath161 when the network is not sparse . when @xmath162 , @xmath163 .
the perturbation approximation is expected to be close to the exact @xmath93 only for @xmath164 and @xmath165 . however , as shown in fig .
[ baerapprox](a ) , the perturbation approximation for small @xmath8 approximates @xmath93 well for a relative large range of @xmath8 , especially for sparser interconnections , i.e. , for a smaller interconnection density @xmath58 .
figure [ baerapprox](b ) shows that the exact spectral radius @xmath93 is already close to the large @xmath8 perturbation approximation , at least for @xmath166 .
baerbounds as depicted in fig .
[ baerbounds ] , the lower bound ( [ lowerbound2 ] ) and upper bound ( [ upperbound ] ) are sharp , i.e. , close to @xmath93 for small @xmath8 .
the lower and upper bounds are the same as @xmath93 when @xmath167 . for large @xmath8 ,
the lower bound better approximates @xmath93 when the interconnections are sparser .
another lower bound @xmath168 , i.e. , lemma [ lowerboundasymptotic ] , is sharp for large @xmath8 , as shown in fig .
[ lowerasymp ] , especially for sparse interconnections .
we do not illustrate the lower bound ( [ lowerbound1 ] ) because the lower bound ( [ lowerbound2 ] ) is always sharper or equally good .
the lower bound @xmath169 considers only the largest eigenvalue of the interconnection network @xmath26 and ignores the two individual networks @xmath9 and @xmath10 . the difference @xmath170 @xmath171 according to the large @xmath8 perturbation approximation , is shown in fig .
[ lowerasymp ] to be larger for denser interconnections .
it suggests that @xmath9 and @xmath10 contribute more to the spectral radius of the interconnected networks when the interconnections are denser in this non - degenerate case . for large @xmath8 ,
the upper bound is sharper when the interconnections are denser or when @xmath58 is larger , as depicted in figure [ baerbounds](b ) .
this is because @xmath172 .
when the interconnections are sparse , @xmath93 is close to the lower bound @xmath169 and hence far from the upper bound .
lowerasympbaer most interdependent or coupled networks studied so far assume that both individual networks have the same number of nodes @xmath1 and that the two networks are interconnected randomly by @xmath1 one - to - one interconnections , or by a fraction @xmath173 of the @xmath174one - to - one interconnections where @xmath175 @xcite .
these coupled networks correspond to our sparse interconnection cases where @xmath176 , when @xmath133 is well approximated by the perturbation approximation for both small and large @xmath8 .
the spectral radius @xmath93 increases quadratically with @xmath8 for small @xmath8 , as described by the small @xmath8 perturbation approximation .
the increase accelerates as @xmath8 increases and converges to a linear increase with @xmath8 , with slope @xmath133 .
here we show the cases in which @xmath9 , @xmath10 , and the interconnections are sparse , as in most real - world networks .
however , all the analytical results can be applied to arbitrary interconnected network structures .
babaasymp we assume the spectrum @xcite to be an unique fingerprint of a large network .
two large networks of the same size seldom have the same largest eigenvalue . hence , most interconnected networks belong to the non - degenerate case .
degenerate cases mostly occur when @xmath9 and @xmath10 are identical , or when they are both regular networks with the same degree .
we consider two degenerate cases where both network @xmath9 and @xmath10 are er random networks or ba scale - free networks . both er and ba networks lead to the same observations .
hence as an example we show the case in which both @xmath9 and @xmath10 are ba scale - free networks of size @xmath177 and both are randomly interconnected with density @xmath178 $ ] , as in the non - degenerate case .
figure [ babaasym](a ) shows that the perturbation analysis well approximates @xmath93 for small @xmath8 , especially when the interconnection density is small .
moreover , the small @xmath8 perturbation approximation performs better in the non - degenerate case , i.e. , is closer to @xmath93 than in degenerate cases [ see fig .
[ baerapprox](a ) ] .
bababounds similarly , fig .
[ bababounds ] shows that both the lower and upper bound are sharper for small @xmath8 .
the lower bound better approximates @xmath101 for sparser interconnections whereas the upper bound better approximates @xmath93 for denser interconnections .
thus far we have examined the cases where @xmath9 , @xmath10 , and the interconnections are sparse , as is the case in most real - world networks . however , if both @xmath9 and @xmath10 are dense er random networks and if the random interconnections are also dense , the upper bound is equal to @xmath93 , i.e. , @xmath179 ( see lemma [ lemma_regular ] ) .
equivalently , the difference @xmath180 is a constant @xmath147 independent of the interconnection density @xmath58 . in both the non - degenerate and
degenerate case , @xmath93 is well approximated by a perturbation analysis for a large range of small @xmath8 , especially when the interconnections are sparse , and also for a large range of large @xmath8 .
the lower bound ( [ lowerbound2 ] ) and upper bound ( [ upperbound ] ) are sharper for small @xmath8 .
most real - world interconnected networks are sparse and non - degenerate , where our perturbation approximations are precise for a large range of @xmath8 , and thus reveal well the effect of component network structures on the epidemic characterizer @xmath93 .
we study interconnected networks that are composed of two individual networks @xmath9 and @xmath10 , and interconnecting links represented by adjacency matricies @xmath18 , @xmath24 , and @xmath26 respectively .
we consider sis epidemic spreading in these generic coupled networks , where the infection rate within @xmath9 and @xmath10 is @xmath3 , the infection rate between the two networks is @xmath4 , and the recovery rate is @xmath30 for all agents . using a nimfa we show that the epidemic threshold with respect to @xmath38 is @xmath39 , where @xmath23{cc}a_{1 } & \mathbf{0}\\ \mathbf{0 } & a_{2}\end{array } \right ] $ ] is the adjacency matrix of the two isolated networks @xmath9 and @xmath10 .
the largest eigenvalue @xmath7 can thus be used to characterize epidemic spreading .
this eigenvalue @xmath7 of a function of matrices seldom gives the contribution of each component network .
we analytically express the perturbation approximation for small and large @xmath8 , lower and upper bounds for any @xmath8 , of @xmath43 as a function of component networks @xmath18 , @xmath24 , and @xmath26 and their largest eigenvalues / eigenvectors . using numerical simulations
, we verify that these approximations or bounds approximate well the exact @xmath7 , especially when the interconnections are sparse and when the largest eigenvalues of the two networks @xmath9 and @xmath10 are different ( the non - degenerate case ) , as is the case in most real - world interconnected networks . hence , these approximations and bounds reveal how component network properties affect the epidemic characterizer @xmath43 .
note that the term @xmath181 contributes positively to the perturbation approximation ( [ degenerate ] ) and the lower bound ( [ lowerbound2 ] ) of @xmath7 where @xmath11 and @xmath12 are the principal eigenvector of network @xmath9 and @xmath10 .
this suggests that , given two isolated networks @xmath9 and @xmath10 , the interconnected networks have a larger @xmath7 or a smaller epidemic threshold if the two nodes @xmath13 and @xmath14 with a larger eigenvector component product @xmath15 from the two networks , respectively , are interconnected .
this observation provides essential insights useful when designing interconnected networks to be robust against epidemics .
the largest eigenvalue also characterizes the phase transition of coupled oscillators and percolation .
our results apply to arbitrary interconnected network structures and are expected to apply to a wider range of dynamic processes .
we wish to thank onr ( grant n00014 - 09 - 1 - 0380 , grant n00014 - 12 - 1 - 0548 ) , dtra ( grant hdtra-1 - 10 - 1 - 0014 , grant hdtra-1 - 09 - 1 - 0035 ) , nsf ( grant cmmi 1125290 ) , the european epiwork , multiplex , congas ( grant fp7-ict-2011 - 8 - 317672 ) , motia ( grant jls-2009-cips - ag - c1 - 016 ) and linc projects , the deutsche forschungsgemeinschaft ( dfg ) , the next generation infrastructure ( bsik ) and the israel science foundation for financial support .
in any regular graph , the minimal and maximal node strength are both equal to the average node strength .
since the largest eigenvalue is lower bounded by the average node strength and upper bounded by the maximal node strength as proved below in lemma [ boundlamda ] , a regular graph has the minimal possible spectral radius , which equals the average node strength .
when the interdependent links are randomly connected with link density @xmath58 , the coupled network is asymptotically a regular graph with average node strength @xmath57+\alpha np_{i}$ ] , if @xmath58 is a constant .
[ boundlamda]for any @xmath17 weighted symmetric matrix @xmath71 , @xmath182\leq\lambda_{1}(w)\leq\max s_{r}\ ] ] where @xmath183 is defined as the node strength of node @xmath184 and @xmath185 $ ] is the average node strength over all the nodes in graph @xmath16 .
the largest eigenvalue @xmath186 follows @xmath187 when matrix @xmath71 is symmetric and the maximum is attained if and only if @xmath11 is the eigenvector of @xmath71 belonging to @xmath93 . for any other vector @xmath188 it holds that @xmath189 . by choosing the vector @xmath190 we have @xmath191\ ] ] where @xmath192 is the element in matrix @xmath71 and @xmath185 $ ] is the average node strength of the graph @xmath16 .
the upper bound is proved by gerschgorin circle theorem .
suppose component @xmath193 of eigenvector @xmath11 has the largest modulus
. the eigenvector can be always normalized such that @xmath194 where @xmath195 for all @xmath14 .
equating component @xmath193 on both sides of the eigenvalue equation @xmath196 gives @xmath197 when none of the elements of matrix @xmath71 are negative . since any component of @xmath11 may have the largest modulus , @xmath198
we consider the @xmath199 vector @xmath74 as @xmath200{cc}c_{1}x^{t } & c_{2}y^{t}\end{array } \right ] $ ] the linear combination of the principal eigenvector @xmath11 and @xmath12 of the two individual networks respectively , where @xmath201 , @xmath202 , @xmath203 such that @xmath204 and compute @xmath205{cc}c_{1}x^{t } & c_{2}y^{t}\end{array } \right ] \left [ \begin{array } [ c]{cc}a_{1 } & \alpha b_{12}\\ \alpha b_{12}^{t } & a_{2}\end{array } \right ] \left [ \begin{array } [ c]{c}c_{1}x\\ c_{2}y \end{array } \right ] \\ & = c_{1}^{2}x^{t}a_{1}x+c_{2}^{2}y^{t}a_{2}y+2\alpha2c_{1}c_{2}x^{t}b_{12}y\\ & = c_{1}^{2}\lambda_{1}\left ( a_{1}\right ) + c_{2}^{2}\lambda_{1}\left ( a_{2}\right ) + 2c_{1}c_{2}\xi\end{aligned}\ ] ] where @xmath77 . by rayleigh s principle @xmath206 .
we could improve this lower bound by selecting @xmath74 as the best linear combination ( @xmath207 and @xmath208 ) of @xmath11 and @xmath12 .
let @xmath209 be the best possible lower bound @xmath210 via the optimal linear combination of @xmath11 and @xmath12 .
thus , @xmath211 we use the lagrange multipliers method and define the lagrange function as @xmath212 where @xmath213 is the lagrange multiplier .
the maximum is achieved at the solutions of @xmath214 note that @xmath215 which leads to @xmath216 hence , the maximum @xmath209 is achieved at the solution of @xmath217 that is @xmath218{cc}\lambda_{1}\left ( a_{1}\right ) -\lambda_{l } & \xi\\ \xi & \lambda_{1}\left ( a_{2}\right ) -\lambda_{l}\end{array } \right ) = 0\ ] ] this leads to @xmath219 the maximum is obtained when @xmath220{cc}\sqrt{\frac{\lambda_{1}\left ( a_{2}\right ) -\lambda_{l}}{\lambda_{1}\left ( a_{1}\right ) + \lambda_{1}\left ( a_{2}\right ) -2\lambda_{l}}}x^{t } & \sqrt{\frac{\lambda_{1}\left ( a_{1}\right ) -\lambda_{l}}{\lambda_{1}\left ( a_{1}\right ) + \lambda_{1}\left ( a_{2}\right ) -2\lambda_{l}}}y^{t}\end{array } \right]\ ] ] by rayleigh s principle @xmath221 .
we consider @xmath74 as linear combination @xmath200{cc}c_{1}x^{t } & c_{2}y^{t}\end{array } \right ] $ ] of @xmath11 and @xmath12 .
the lower bound @xmath222{cc}c_{1}x^{t } & c_{2}y^{t}\end{array } \right ] \left [ \begin{array } [ c]{cc}a_{1}^{2}+\alpha^{2}b_{12}b_{12}^{t } & \alpha\left ( a_{1}b_{12}+b_{12 } a_{2}\right ) \\ \alpha\left ( a_{1}b_{12}+b_{12}a_{2}\right ) ^{t } & a_{2}^{2}+\alpha ^{2}b_{12}^{t}b_{12}\end{array } \right ] \left [ \begin{array } [ c]{c}c_{1}x\\ c_{2}y \end{array } \right ] \\ & = c_{1}^{2}x^{t}a_{1}^{2}x+c_{2}^{2}y^{t}a_{2}^{2}y+\alpha^{2}\left ( c_{1}^{2}x^{t}b_{12}b_{12}^{t}x+c_{2}^{2}y^{t}b_{12}^{t}b_{12}y\right ) + 2\alpha c_{1}c_{2}x^{t}\left ( a_{1}b_{12}+b_{12}a_{2}\right ) y\\ & = c_{1}^{2}\lambda_{1}^{2}\left ( a_{1}\right ) + c_{2}^{2}\lambda_{1 } ^{2}\left ( a_{2}\right ) + 2c_{1}c_{2}\theta+\alpha^{2}\left ( c_{1 } ^{2}\left\vert b_{12}^{t}x\right\vert _ { 2}^{2}+c_{2}^{2}\left\vert b_{12}y\right\vert _ { 2}^{2}\right)\end{aligned}\ ] ] where @xmath84 .
let @xmath209 be the best possible lower bound @xmath223 via the optimal linear combination ( @xmath207 and @xmath208 ) of @xmath11 and @xmath12 .
thus , @xmath224 we use the lagrange multipliers method and define the lagrange function as @xmath225 where @xmath213 is the lagrange multiplier .
the maximum is achieved at the solutions of @xmath226 which lead to @xmath227 .
hence , the maximum @xmath209 is achieved at the solution of @xmath228 that is @xmath218{cc}\lambda_{1}^{2}\left ( a_{1}\right ) + \alpha^{2}\left\vert b_{12 } ^{t}x\right\vert _ { 2}^{2}-\lambda_{l } & \theta\\ \theta & \lambda_{1}^{2}\left ( a_{2}\right ) + \alpha^{2}\left\vert b_{12}y\right\vert _
{ 2}^{2}-\lambda_{l}\end{array } \right ) = 0\ ] ] this leads to @xmath229 hence , @xmath230 which is obtained when @xmath231 any vector @xmath74 of size @xmath22 with @xmath232 can expressed as a linear combination of the eigenvectors @xmath233 of matrix
@xmath71 @xmath234 where @xmath235 .
hence , @xmath236 hence , @xmath90 according to lyapunov s inequality , @xmath237 \right ) ^{1/s } \leq\left ( e\left [ \left\vert x\right\vert ^{t}\right ] \right ) ^{1/t}\ ] ] when @xmath238 .
taking @xmath239=c_{i}^{2}$ ] , we have @xmath240 \right ) ^{1/s } \leq\left ( e\left [ \left\vert x\right\vert ^{k}\right ] \right ) ^{1/k } = \sum_{i=1}^{2n}c_{i}^{2}\frac{\lambda_{i}^{k}}{\lambda_{1}^{k}}\ ] ] since @xmath88 is even and @xmath241 .
@xmath242{cc}x^{t } & y^{t}\end{array } \right ] \left ( a+\alpha b\right ) \left [ \begin{array } [ c]{c}x\\ y \end{array } \right ] \\ & = \max_{x^{t}x+y^{t}y=1}\left ( \left [ \begin{array } [ c]{cc}x^{t } & y^{t}\end{array } \right ] a\left [ \begin{array } [ c]{c}x\\ y \end{array } \right ] + \alpha\left [ \begin{array } [ c]{cc}x^{t } & y^{t}\end{array } \right ] b\left [ \begin{array } [ c]{c}x\\ y \end{array } \right ] \right ) \\ & \leq\max_{x^{t}x+y^{t}y=1}\left ( x^{t}a_{1}x+y^{t}a_{2}y\right ) + \alpha\max_{x^{t}x+y^{t}y=1}\left [ \begin{array } [ c]{cc}x^{t } & y^{t}\end{array } \right ] b\left [ \begin{array } [ c]{c}x\\ y \end{array } \right ] \\ & = \max\left ( { \lambda}_{1}{(a_{1}),\lambda}_{1}{(a_{2})}\right ) + \alpha\lambda_{1}(b)\end{aligned}\ ] ] the inequality is due to the fact that the two terms are maximized independently .
the second term@xmath243 is equivalent to the system of equations@xmath103{c}b_{12}y=\lambda_{1}(b)x\\ b_{12}y=\lambda_{1}(b)x\\ x^{t}x+y^{t}y=1 \end{array } \right.\ ] ] or @xmath103{c}b_{12}^{t}b_{12}y=\lambda_{1}(b)^{2}y\\ b_{12}b_{12}^{t}x=\lambda_{1}(b)^{2}x\\ x^{t}x+y^{t}y=1 \end{array } \right.\ ] ] which is to find the maximum eigenvalue ( or more precisely the positive square root ) of the symmetric positive matrix @xmath95 @xmath244 this actually proves lemma [ bipartite ] , the property @xmath245 of a bipartite graph @xmath246 in the non - degenerate case , @xmath250 hence , @xmath251\end{aligned}\ ] ] where the first @xmath1 elements of @xmath252 are @xmath11 and the rest @xmath1 elements are all zeros .
let us look at the first order correction .
imposing the identity for the first order expansion in ( [ eigenequa ] ) gives @xmath253 furthermore , we impose the normalization condition to @xmath74 ( see ( [ aaa ] ) ) , which leads to @xmath254 the first order correction to the principal eigenvector is orthogonal to the zero order . plugging this result in ( [ 1ordercorrection ] ) @xmath255 that is @xmath256 since @xmath123{cc}x^{t } & \mathbf{0}^{t}\end{array } \right ) $ ] and @xmath25{cc}\mathbf{0 } & b_{12}\\ b_{12}^{t } & \mathbf{0}\end{array } \right ] $ ]
, the first order correction is this non - degenerate case is null @xmath257 equation ( [ 1ordercorrection ] ) allows us to calculate also the first order correction to the eigenvector @xmath258 @xmath259 is invertible out of its kernel @xmath260 ( that is the linear space generated by @xmath252 ) and since @xmath261 we have @xmath262 let us look for the second order correction . imposing the identification of the second order term of ( [ eigenequa ] ) we obtain @xmath263
projecting this vectorial equation on @xmath252 provides the second order correction to @xmath104 @xmath264 substituting ( [ xfirstordercorrection ] ) gives @xmath265 which can be further expressed as a function of the largest eigenvalue / eigenvector of individual network @xmath266 or their interconnections @xmath56 .
since @xmath267{ll}0 & b_{12}\\ b_{12}^{t } & 0 \end{array } \right ) \left ( \begin{array } [ c]{l}x\\ 0 \end{array } \right ) = \left ( \begin{array } [ c]{l}0\\ b_{12}^{t}x \end{array } \right)\ ] ] we have @xmath268{ll}0 & b_{12}x^{(0)}\end{array } \right ) \left ( \begin{array } [ c]{ll}(\lambda^{(0)}i - a_{1 } ) & 0\\ 0 & ( \lambda^{(0)}i - a_{2 } ) \end{array } \right ) ^{-1}\left ( \begin{array } [ c]{l}0\\ b_{12}^{t}x^{(0)}\end{array } \right ) \\ & = \left ( \begin{array } [ c]{ll}0 & ( x^{(0)})^{t}b_{12}\end{array } \right ) \left ( \begin{array } [ c]{ll}(\lambda^{(0)}i - a_{1 } ) & 0\\ 0 & ( \lambda^{(0)}i - a_{2 } ) \end{array } \right ) ^{-1}\left ( \begin{array } [ c]{l}0\\ b_{12}^{t}x^{(0)}\end{array } \right ) \\ & = \left ( \begin{array } [ c]{ll}0 & ( x^{(0)})^{t}b_{12}\end{array } \right ) \left ( \begin{array } [ c]{ll}(\lambda^{(0)}i - a_{1})^{-1 } & 0\\ 0 & ( \lambda^{(0)}i - a_{2})^{-1}\end{array } \right ) \left ( \begin{array } [ c]{l}0\\ b_{12}^{t}x^{(0)}\end{array } \right ) \\ & = ( x^{(0)})^{t}b_{12}\left ( \lambda^{(0)}i - a_{2}\right ) ^{-1}b_{12 } ^{t}x^{(0)}\ ] ] which finishes the proof . in this case , the solution @xmath252 of the zero order expansion equation @xmath248 can be any combination of the largest eigenvector of the two individual networks @xmath11 and @xmath12 : @xmath269 and @xmath270 .
the first order correction of the largest eigenvalue correction in the non - degenerate case ( [ lamdafirstordernon ] ) holds as well for the generate case @xmath271 which is however non - zero in the degenerate case due to the structure of @xmath252 and is maximized by the right choice of @xmath272 and @xmath273 .
thus , @xmath274 where @xmath275 is maximum when @xmath276 one may also evaluate the second order correction to the largest eigenvalues of the degenerate case .
the following results we derived in the non - degenerate case hold as well for the degenerate case @xmath103{c}\lambda^{(2)}=\left ( z^{(0)},bz^{(1)}\right ) \\
az^{(1)}+bz^{(0)}=\lambda^{(0)}z^{(1)}+\lambda^{(1)}z^{(0)}\end{array } \right.\ ] ] the latter equation allows to calculate the first order correction to the dominate eigenvector @xmath277 : @xmath278 any linear equation admits solutions when the constant term @xmath279 is orthogonal to the kernel of the adjoint matrix of @xmath280 .
@xmath281 apart from pathological cases , each the two interactive nets have non degenerate maximum eigenvalues ( @xmath282 ) corresponding to the dominant eigenvectors @xmath283 and @xmath284 : @xmath103{lll}a_{1}x^{(0 ) } & = & \lambda^{(0)}x^{(0)}\\ a_{2}y^{(0 ) } & = & \lambda^{(0)}y^{(0 ) } ; \end{array } \right.\ ] ] in this case , the kernel of the matrix @xmath280 is just the linear space generated by the maximum eigenvalue of the single nets : @xmath285{l}ax^{(0)}\\ by^{(0)}\end{array } \right ) .\ ] ] as stated , the constant term is orthogonal to the entire kernel : @xmath286{ll}ax^{(0 ) } & by^{(0)}\end{array } \right ) \cdot\left ( b-\lambda^{(1)}\right ) z^{(0)}=a\left ( \left ( b_{12}^{t}x , y\right ) -\lambda^{(1)}\right ) + b\left ( \left ( x , b_{12 } y\right ) -\lambda^{(1)}\right ) = 0.\ ] ] therefore the solution of the previous equation exists and all solutions differ by a vector in @xmath287 .
it is worth stressing that the value of @xmath288 does not depend on this extra terms , that is @xmath288 is invariant under the transformation : @xmath103{l}x^{(1)}\rightarrow x^{(1)}+ax^{(0)}\\ y^{(1)}\rightarrow y^{(1)}+by^{(0 ) } ; \end{array } \right . ; \ ] ] providing the normalization condition is respected : @xmath289 that is @xmath290 that leads to @xmath291 .
let us apply the transformation to @xmath288 : @xmath292{l}b_{12 } y^{(0)}\\ b_{12}^{t } x^{(0)}\end{array } \right ) ^{t } \left ( \begin{array } [ c]{l}x^{(1)}\\ y^{(1)}\end{array } \right ) \to\hat{\lambda}^{(2)}= \frac{1}{\sqrt{2}}\left ( \begin{array } [ c]{l}b_{12 } y^{(0)}\\ b_{12}^{t } x^{(0)}\end{array }
\right ) ^{t } \left ( \begin{array } [ c]{l}x^{(1)}+ax^{(0)}\\ y^{(1)}-ay^{(0)}\end{array } \right)\ ] ] @xmath293 = \lambda^{(2)}\ ] ] therefore we are allowed to select a definite solution as we where fixing a gauge .
we will impose the orthogonality of @xmath294 with @xmath283 and @xmath295 with @xmath284 .
the linear operator @xmath296 and @xmath297 operate identically over all vectors orthogonal to @xmath283 ( as all constant terms are in our case ) , while it behaves as an identity in the linear space generated by @xmath283 . therefore
to fix the gauge one may substitute @xmath297 with @xmath298 .
the same argument holds for @xmath24 .
this allows us to provide @xmath288 an explicit expression and to calculate it algebraically .
equation [ th12egen1 ] in components reads : @xmath103{lll}(\lambda^{(0)}i - a_{1 } ) x^{(1 ) } & = & b_{12}y^{(0)}-\lambda^{(1)}x^{(0)}\\ ( \lambda^{(0)}i - a_{2 } ) y^{(1 ) } & = & b^{t}_{12}x^{(0)}-\lambda^{(1)}y^{(0 ) } ; \end{array } \right.\ ] ] that , after fixing the gauge , becomes : @xmath103{lll}(\lambda^{(0)}i - a_{1}+ x^{(0)}(x^{(0)})^{t})x^{(1 ) } & = & b_{12 } y^{(0)}-\lambda^{(1)}x^{(0)}\\ ( \lambda^{(0)}i - a_{2}+ y^{(0)}(y^{(0)})^{t})y^{(1 ) } & = & b^{t}_{12 } x^{(0)}-\lambda^{(1)}y^{(0 ) } ; \end{array } \right.\ ] ] and hence the first - order correction to the dominant eigenvector can be @xmath103{lll}x^{(1 ) } & = & ( \lambda^{(0)}i - a_{1}+ x^{(0)}(x^{(0)})^{t})^{-1 } ( b_{12}y^{(0)}-\lambda^{(1)}x^{(0 ) } ) \\ y^{(1 ) } & = & ( \lambda^{(0)}i - a_{2}+ y^{(0)}(y^{(0)})^{t})^{-1 } ( b^{t } _ { 12}x^{(0)}-\lambda^{(1)}y^{(0 ) } ) ; \end{array } \right.\ ] ] the second order correction can be finally be calculate algebraically , resorting to the spectral properties of the isolated networks , @xmath299{l}b_{12}y^{(0)}\\ b_{12}^{t}x^{(0)}\end{array } \right ) ^{t}\left ( \begin{array } [ c]{ll}(\lambda^{(0)}i - a_{1}+x^{(0)}(x^{(0)})^{t})^{-1 } & 0\\ 0 & ( \lambda^{(0)}i - a_{2}+y^{(0)}(y^{(0)})^{t})^{-1}\end{array } \right ) \left ( \begin{array } [ c]{l}b_{12}y^{(0)}-\lambda^{(1)}x^{(0)}\\ b_{12}^{t}x^{(0)}-\lambda^{(1)}y^{(0)}\end{array } \right)\ ] ] that is : @xmath300 s. c. ferreira , c. castellano , and r. pastor - satorras , `` epidemic thresholds of the susceptible - infected - susceptible model on networks : a comparison of numerical and theoretical results , '' arxiv:1206.6728v1 ( 2012 ) . | most real - world networks are not isolated . in order to function fully , they are interconnected with other networks , and this interconnection influences their dynamic processes . for example , when the spread of a disease involves two species , the dynamics of the spread within each species ( the contact network ) differs from that of the spread between the two species ( the interconnected network ) .
we model two generic interconnected networks using two adjacency matrices , a and b , in which a is a @xmath0 matrix that depicts the connectivity within each of two networks of size @xmath1 , and b a @xmath0 matrix that depicts the interconnections between the two . using an n - intertwined mean - field approximation , we determine that a critical susceptable - infected - susceptable ( sis ) epidemic threshold in two interconnected networks is @xmath2 , where the infection rate is @xmath3 within each of the two individual networks and @xmath4 in the interconnected links between the two networks and @xmath5 is the largest eigenvalue of the matrix @xmath6 . in order to determine how the epidemic threshold is dependent upon the structure of interconnected networks , we analytically derive @xmath7 using perturbation approximation for small and large @xmath8 , the lower and upper bound for any @xmath8 as a function of the adjacency matrix of the two individual networks , and the interconnections between the two and their largest eigenvalues / eigenvectors .
we verify these approximation and boundary values for @xmath7 using numerical simulations , and determine how component network features affect @xmath7 .
we note that , given two isolated networks @xmath9 and @xmath10 with principle eigenvectors @xmath11 and @xmath12 respectively , @xmath7 tends to be higher when nodes @xmath13 and @xmath14 with a higher eigenvector component product @xmath15 are interconnected .
this finding suggests essential insights into ways of designing interconnected networks to be robust against epidemics . |
exascale computing @xcite is expected to revolutionize computational science and engineering by providing 1000x the capabilities of currently available computing systems , while having a similar power footprint .
the total performance of the 500 systems in the 44th top500 list ( 18 nov 2014 , http://top500.org/ ) is about 0.3 exaflops .
the hpc community @xcite is now working towards the development of the first exaflop computer , expected around 2020 , after reaching the petaflop milestone in 2008 .
however , only a few hpc applications are so far able to fully exploit the capabilities of petaflop systems @xcite .
examples of typical scalability for commonly used hpc applications in our organizations are provided in table [ tab : hpcapps ] . as the existing hpc applications are the major hpc asset , it is important and challenging to increase their scalability and lifetime by making them exascale - ready before 2020 .
[ cols="<,<,<,>",options="header " , ] [ tab : hpcapps ] the major challenge for preparing hpc applications for exascale is that there is no exascale system available yet .
currently all we have are assumptions about exascale systems .
therefore the commonly used measurement - based approaches for reasoning about performance issues are not applicable .
pre - exascale systems ( known as _
summit _ and _ sierra _ ) that ibm @xcite is developing for the u.s .
department of energy will exceed 100 petaflops and may provide hints about the extreme - scale architectures of the future .
this paper argues that efforts for preparing hpc applications for exascale should start before such systems become available .
we identify challenges that need to be addressed and recommend solutions in areas that are relevant for porting hpc application to future exascale computing systems , including formal modeling , static analysis and optimization , runtime analysis and optimization , and autonomic computing .
we suggest that porting of hpc applications should be made by successive , stepwise improvements based on the currently available assumptions and data about exascale systems .
this approach should support application improvement each time new information about future exascale systems becomes available , including the time when the application is actually deployed and runs on a concrete exascale system . a high - level application representation that captures key functional and non - functional properties in conjunction with the abstract machine model will enable programmers and tools to reason about and perform application improvements , and will serve as input to runtime systems to handle performance and energy optimizations and self - aware fault management .
a tunable abstract machine model encapsulates current assumptions for future exascale systems and enables a priori application improvement before the concrete execution platform is known . at runtime ,
the model is a posteriori tuned to support activities such as feedback - oriented code improvement or dynamic optimization .
major contributions of this paper include , * identification of challenges and recommendation of solutions in formal modeling ( section [ subsec : modeling ] ) , static analysis and optimization ( section [ subsec : static ] ) , runtime analysis and optimization ( section [ subsec : runtime ] ) , autonomic computing ( section [ subsec : autonomic ] ) ; * a conceptual framework for preparing hpc applications for exascale that supports _ a priori _ application improvements before the concrete execution platform is known as well as _
a posteriori _ optimization at runtime ( section [ sec : approach ] ) ; * a discussion of the related work ( section [ sec : related - work ] ) .
in this section we identify challenges and recommend solutions in formal modeling , static analysis and optimization , runtime analysis and optimization , and autonomic computing .
our goal is to adapt hpc application code to exascale execution platforms to achieve good utilization of resources . for this , we need to address questions such as : 1 . what would happen if we change application or hardware layout ?
2 . what would happen if we change some parameters of the execution platform ? 3 . what would happen if we use a different execution platform ? unfortunately , answering these questions can not be done experimentally at the concrete level because such platforms do not yet exist .
an alternative is to address these questions at an abstract level , focusing only on _ relevant information _ without actually executing the program .
we believe that relevant information in this context is not what the code aims to achieve ( the result of the computation ) but its corresponding _ resource footprints _ , that is , how computational tasks communicate and synchronize , the amount of resources ( such as memory and computing time ) these tasks require , and how they access and move data . in order to adapt the hpc code to a particular architecture
we need to capture such resource footprints of software modules at different levels of granularity ( e.g. , program statements , blocks in procedure bodies and whole procedures ) , and be able to compare different task compositions .
consequently , the modeling language must feature massively parallel operators over such task - level resource footprints @xcite .
a similar notion of resource footprints and composition can be used to express the properties of the architecture in a machine model to capture the resources that the architecture can make available to the code .
working with resource footprints can be supported by an abstract behavioral specification language @xcite , in which models describe both tasks and deployments .
these models can be used to predict the non - functional behavior of code before it is deployed , and to compare deployments using formal methods .
this requires a formal semantics for the specification language that can be used to devise static analysis techniques .
when developing code from scratch using a model - based approach , the resource footprints can be specified in tandem with the standard model in a model - driven development @xcite . however , when building such models from existing hpc code , monitoring profiles of low- and medium - scale systems can be used to extract resource footprints that approximate the resource consumption in terms of probabilistic distributions .
the application of formal methods to parallel programs for analyzing functional properties , such as safety and liveness , has a long tradition . for _ non - functional properties _ , such as execution time and energy consumption , most performance analysis approaches use _ monitoring _ and present statistical information to the user .
these approaches are helpful to improve hpc application code , but they also have some shortcomings : 1 .
due to non - determinism , different program executions might lead to different observations . as a consequence , these methods are not able to provide reliable probabilistic information about average or worst - case execution times .
they are based on execution on a real platform , thus they can not be used to predict performance on exascale computers , which are not available yet .
3 . these methods can be used to identify execution bottlenecks , but they can not explain the reasons for these bottlenecks , and thus they do not offer any concrete support for code improvement . we expect that _ formal methods _ can address these limitations to provide performance analysis tools that considerably go beyond the state - of - the - art
. a major step in this direction will be the usage of resource footprints which describe both hpc applications and execution platforms as abstract probabilistic models .
formal analyses can be applied to these models to predict their probabilistic behavior .
while a range of techniques are available for non - probabilistic programs , the analysis of parallel probabilistic programs still need development effort . to achieve a reasonable balance between scalability and precision for challenging hpc applications
, it seems fundamental to use _ hybrid approaches _ @xcite that combine techniques such as static analysis , dynamic analysis , simulation , ( parametric ) model checking @xcite , counterexample - guided abstraction refinement @xcite , deductive approaches , etc . to deliver the envisaged performance analysis tools , we face the following challenges : ( 1 ) determining the computation of cost properties that are given by means of probabilistic distributions ; ( 2 ) the inference of average cost in addition to the traditional worst - case cost ; ( 3 ) take into account the underlying platform through a set of probabilistic parameters ; ( 4 ) deal with massive and heterogeneous parallelism @xcite which is challenging for program analysis in general ; and ( 5 ) develop multi - objective resource usage analyses and optimizations .
formal modeling and static analysis should be enhanced with analysis of measurements at runtime .
plenty of tools ( for instance , http://www.vi-hps.org ) have been developed for performance measurement and analysis of hpc applications at runtime .
however , these tools will experience several issues when applied to exascale .
the collection rates and the overall volume of monitoring data in an exascale computing environment will exceed the scalability of present performance tools .
therefore , throttling the data volume will have to be applied online in order to store as less data as possible and as much as necessary for later _ post mortem _ analysis
. however , simple profiling will not be sufficient due to loss of temporal information , thus a hybrid approach will have to be applied that performs on - the - fly trace analysis in order to discard irrelevant data , while retaining the same amount of information .
the metric classification should be based on the formal model ( see section [ subsec : modeling ] ) .
such an approach will provide a generic insight into the performance of an hpc application that can be used for detecting performance bottlenecks .
the instrumentation and hardware counter monitoring should follow a similar procedure where source code probing should be applied automatically by using tools such as opari @xcite . while many tools for collecting metrics of computing performance have been developed , very few analysis tools exist for energy consumption metrics in adequate accuracy and time resolution necessary for the runtime performance analysis @xcite .
currently , common approaches ( see for example prace best practices @xcite ) for optimizing hpc applications require per - case inspection of runtime performance measurement data , such as profiling and tracing data .
after the critical region has been determined , diverse heuristic approaches , such as `` _ _ trial and error _ _ '' , `` _ _ educated guess _ _ '' or `` _ _ rule of thumb _ _ '' , are applied to make changes in the affected source code sections .
the most significant limitation of these heuristics and knowledge - based approaches is that , 1 .
all changes are made directly and manually in the source code , and 2 .
the effect of the changes does not always lead to an improved performance which makes necessary the repeating of all steps several times .
moreover , exascale computers pose a multi - objective optimization problem , weighing out the effects of several sometimes incongruent requirements .
therefore , a systematic and automatic approach for the optimization problem is essential to find the optimal solution .
another problem is that the critical section in an application typically changes with the optimization iterations and/or with upscaling , due to the law of diminishing returns , which makes the manual analysis and source code changes even more laborious and inefficient , even if done by an experienced hpc developer .
thus instrumentation , collection / measurement and analysis steps should be automated , for example based on high - level scalable tools @xcite , and integrated into a feedback loop ( see section [ subsec : autonomic ] ) . during the execution of an application
, failures may occur or the application performance may be below the expectation .
these issues are addressed typically by programmers in a `` _ _ trial and error _ _ '' manner , i.e. by manually changing and adapting their code to handle the failures and improve the performance .
our proposed framework ( section [ sec : approach ] ) provides means for model - based failure handling or performance improvement based on _
autonomic computing_. autonomic computing addresses self - managing characteristics of distributed computing resources with the facilities to adapting to unpredictable changes while hiding management complexity to operators and users @xcite . among the explored categories , _ advanced - control based methods _ and more specifically _ distributed controllers _ are the first candidates to realize autonomic computing in exascale systems .
we propose to devise methodologies to efficiently collect runtime information balancing the amount and cost for storage of monitoring data with the quality of monitored data necessary to make deductions about the application behavior ( e.g. trace analysis ) .
the goal is thereby to define methodologies to scale current monitoring tools to exascale , balancing between quality and volume of monitoring data . combining the information from both static code analysis and runtime analysis , as outlined above
, we will iteratively apply objective - oriented transformations to legacy application code at a formal level based on the exascale dsl model ( see section [ subsec : modeling ] ) . to this end
, we will automate the analysis , optimization and transformation processes by implementing a generic feedback loop independent of the concrete programming language , algorithms used and target hardware architecture .
a feedback loop driver enables to link the static analysis tool , the runtime analysis tool , the knowledge database and multi - parameter multi - objective optimization . as output , a set of rules ( policies )
is generated which is then applied to transform the formal application model and to adapt the runtime environment parameters ( cf .
figure [ fig : diagram ] ) .
after the transformations a new application executable is built and started in the adapted runtime environment .
this described loop is iterated until convergence of the optimization .
in this section we propose a conceptual framework for porting hpc applications to exascale computing systems .
furthermore , we highlight benefits of our conceptual framework in the context of exascale computing . our proposed approach for preparing hpc applications for exascale is depicted in figure [ fig : diagram ] .
the usage of a _ domain - specific exascale language ( dsel ) _ facilitates the programmer to express non - functional aspects ( like required time to solution , resilience or energy - efficiency ) of the execution of scalable parallel hpc codes .
dsel has a formal operational semantics that enables the formal analysis of the code .
the aim of the _ scalable model - based analyzer ( sma ) _ is to address non - functional properties of hpc codes , with a particular focus on scalability while complying with the crucial dimensions of resource consumption for exascale computing : time , energy , and resilience .
the sma is responsible for analyzing resource consumption in terms of time , energy , and resilience , based on developed dsels .
the _ exascale runtime data collector ( erdc ) _ is responsible for scalable monitoring to _ extract important monitoring data _ through the utilization of various techniques like _ filtering _ , _ streaming _ , or _ data mining_. the runtime information is used to verify or to tune the model of the code via the _ autonomous feedback loop ( afl)_. to endow the system with self - adaption , control - theoretical concepts are incorporated in autonomic computing paradigm .
based on the autonomic technology for application optimization , programmers will be less dependent on the currently used `` _ _ trial - and - error _ _ '' approach .
our approach considers optimization opportunities during the application life cycle comprising improvements based on static code analysis , deployment - time optimization , and run - time optimization .
the developed models are used to identify the potential for improvement of the scalability for hpc applications under study and suggest application modifications that may lead to better scalability .
exascale computing is not simply the continuation of a computational capability trend that has been proven true for the last five decades .
first , while clock rate scaling is limited , complex multicore architectures and parallel computing still follow moore s law .
second , exascale computing capability will finally allow complex real - life simulations and data analytics .
the latter will greatly expand the horizons of scientific discovery and enable the new data - driven economy to become a reality .
however , the exascale promise faces a series of obstacles , with the most difficult being energy , scalability , reliability and programmability .
our proposal is to develop a holistic , unifying and mathematically founded framework to systematically attack the roots of these problems .
that is , instead of attacking these problems separately , we propose a holistic approach to study them as a multi - parametric problem which will allow us to deeply understand their interplay and thus make the right decisions to navigate in this complex landscape .
the benefits are targeting the full spectrum of actors and beneficiaries .
system developers will have a much better path to design , while end users and application developers will benefit from increased scalability , performance , reliability and programmability .
hpc centers will see a great increase in overall system usability and an energy budget that is affordable .
this in turn has the potential to greatly limit and contain the overall impact of high end hpc to the environment .
in a prospective analysis of issues with extreme scale systems @xcite , the importance of concurrency , energy efficiency and resilience of software , as well as software hardware co - design has been elucidated .
focusing on energy - aware hpc numerical applications , the exa2green project ( http://exa2green-project.eu ) has developed energy - aware performance metrics @xcite , as well as energy - aware basic algorithm motifs such as linear solvers @xcite .
further work will strongly benefit from these results .
different power measurement interfaces available on current architecture generations have been evaluated and the role of the sampling rate has been discussed @xcite . the autotune approach @xcite employs the periscope tuning framework @xcite to automate performance analysis and tuning of hpc applications with the goal to improve performance and energy efficiency .
therein , both performance analysis and tuning are performed automatically during a single run of the application . the deep project @xcite has developed a novel exascale - enabling supercomputing architecture with a matching software stack and a set of optimized grand - challenge simulation applications .
the goal of the deep architecture is to enable unprecedented scalability and with an extrapolation to millions of cores to take the deep concept to an exascale level .
the follow - up deeper project ( http://www.deep-er.eu ) is mainly focusing on i / o and resiliency aspects . the cresta project ( http://www.cresta-project.eu ) has adopted a co - design strategy for exascale , including all aspects of hardware architectures , system and application software . a major asset from the cresta project
is the score - p measurement system @xcite on which an integration and automation of performance analysis tools ( cf .
section [ subsec : runtime ] ) can be based .
in addition , efforts have been made on developing a domain - specific language for expressing parallel auto - tuning specifications and an adaptive runtime support framework .
exascale computing will revolutionize high - performance computing , but the first exascale systems are not expected to appear before 2020 . in this paper we have argued that the effort for preparing hpc applications for exascale should start now .
we have proposed that porting of hpc applications should be made by successive , stepwise improvements based on the currently available assumptions and data about exascale systems .
this approach should support application improvement each time new information about future exascale systems becomes available , including the time when the application is actually deployed and runs on a concrete exascale system .
we have identified challenges that need to be addressed and recommended solutions in key areas of interest for our approach including : formal modeling , static analysis and optimization , runtime analysis and optimization , and autonomic computing .
our future research will address the development of a framework that supports the conceptual framework presented in this paper .
e. dhollander , j. dongarra , i. foster , l. grandinetti , and g. joubert , eds . , _ transition of hpc towards exascale computing _ , ser .
advances in parallel computing.1em plus 0.5em minus 0.4emios press , 2013 , vol .
24 .
department of energy selects ibm `` data centric '' systems to advance research and tackle big data challenges , '' 2014 .
[ online ] .
available : http://www-03.ibm.com/press/us/en/pressrelease/45387.wss s. pllana , i. brandic , and s. benkner , `` a survey of the state of the art in performance modeling and prediction of parallel and distributed computing systems , '' _ international journal of computational intelligence research ( ijcir ) _ , vol . 4 , no . 1 ,
pp . 1726 , january 2008 .
e. b. johnsen , r. hhnle , j. schfer , r. schlatte , and m. steffen , `` abs : a core language for abstract behavioral specification , '' in _
9th international symposium on formal methods for components and objects ( fmco10 ) _ , ser .
lecture notes in computer science , vol .
6957.1em plus 0.5em minus 0.4emspringer , 2011 , pp
. 142164 .
e. arkin , b. tekinerdogan , and k. m. imre , `` model - driven approach for supporting the mapping of parallel algorithms to parallel computing platforms , '' in _ model - driven engineering languages and systems _ ,
lecture notes in computer science.1em plus 0.5em minus 0.4em springer , 2013 , vol . 8107 , pp .
757773 .
s. pllana and t. fahringer , `` on customizing the uml for modeling performance - oriented applications , '' in _ uml _ , ser .
lecture notes in computer science , vol .
2460.1em plus 0.5em minus 0.4emspringer , 2002 , pp .
259274 .
s. pllana , s. benkner , e. mehofer , l. natvig , and f. xhafa , `` towards an intelligent environment for programming multi - core computing systems , '' in _ euro - par 2008 workshops - parallel processing _ ,
lecture notes in computer science , vol .
5415.1em plus 0.5em minus 0.4emspringer , 2008 , pp
. 141151 .
t. fahringer , s. pllana , and j. testori , `` teuta : tool support for performance modeling of distributed and parallel applications , '' in _ computational science - iccs 2004 _ , ser .
lecture notes in computer science.1em plus 0.5em minus 0.4emspringer berlin heidelberg , 2004 , vol . 3038 , pp .
456463 .
s. pllana , s. benkner , f. xhafa , and l. barolli , `` hybrid performance modeling and prediction of large - scale parallel systems , '' in _
parallel programming , models and applications in grid and p2p systems _ , ser .
advances in parallel computing.1em plus 0.5em minus 0.4emios press , 2009 , vol .
17 , pp . 5482 .
n. jansen , f. corzilius , m. volk , r. wimmer , e. brahm , j .-
katoen , and b. becker , `` accelerating parametric probabilistic verification , '' in _ quantitative evaluation of systems ( qest14 ) _ , ser .
lecture notes in computer science , vol .
8657.1em plus 0.5em minus 0.4emspringer , 2014 , pp
. 404420 . c. dehnert , n. jansen , r. wimmer , e. brahm , and j .-
katoen , `` fast debugging of prism models , '' in _ international symposium on automated technology for verification and analysis ( atva14 ) _ , ser .
lecture notes in computer science , vol .
8837.1em plus 0.5em minus 0.4emspringer , 2014 , pp . 146162 .
s. benkner , s. pllana , j. traff , p. tsigas , u. dolinsky , c. augonnet , b. bachmayer , c. kessler , d. moloney , and v. osipov , `` peppher : efficient and productive usage of hybrid computing systems , '' _ micro , ieee _ , vol .
. 5 , pp . 2841 , sept 2011 .
c. kessler , u. dastgeer , s. thibault , r. namyst , a. richards , u. dolinsky , s. benkner , j. traff , and s. pllana , `` programmability and performance portability aspects of heterogeneous multi-/manycore systems , '' in _ design , automation test in europe conference exhibition ( date ) , 2012 _ , march 2012 , pp .
14031408 .
m. sandrieser , s. benkner , and s. pllana , `` using explicit platform descriptions to support programming of heterogeneous many - core systems , '' _ parallel computing _ , vol .
38 , no . 1 - 2 , pp .
5256 , january 2012 .
j. dokulil , e. bajrovic , s. benkner , s. pllana , m. sandrieser , and b. bachmayer , `` efficient hybrid execution of c++ applications using intel(r ) xeon phi(tm ) coprocessor , '' _ corr _ , vol .
abs/1211.5530 , 2012 .
[ online ] .
available : http://arxiv.org/abs/1211.5530 p. alonso , r. badia , j. labarta , m. barreda , m. dolz , r. mayo , e. quintana - orti , and r. reyes , `` tools for power - energy modelling and analysis of parallel scientific applications , '' in _
2012 41st international conference on parallel processing ( icpp12)_.1em plus 0.5em minus 0.4emieee , 2012 , pp .
420 429 .
a. bohra and v. chaudhary , `` vmeter : power modelling for virtualized clouds , '' in _ 2010 ieee international symposium on parallel distributed processing , workshops and phd forum ( ipdpsw10)_.1em plus 0.5em minus 0.4emieee , 2010 , pp . 1 8 .
m. geimer , f. wolf , b. wylie , and b. mohr , `` scalable parallel trace - based performance analysis , '' in _ recent advances in parallel virtual machine and message passing interface _ ,
lecture notes in computer science .
1em plus 0.5em minus 0.4emspringer , 2006 , vol . 4192 , pp .
303312 .
b. mohr and f. wolf , `` kojak : a tool set for automatic performance analysis of parallel programs , '' in _ euro - par 2003 parallel processing _ , ser .
lecture notes in computer science.1em plus 0.5em minus 0.4em springer , 2003 , vol .
2790 , pp .
13011304 .
f. wolf and b. mohr , `` automatic performance analysis of hybrid mpi / openmp applications , '' in _ eleventh euromicro conference on parallel , distributed and network - based processing_.1em plus 0.5em minus 0.4emieee , 2003 , pp .
1322 . c. bekas and a. curioni , `` a new energy aware performance metric , '' _ computer science - research and development _ ,
vol . 25 , no . 3 - 4 , pp . 187195 , 2010 .
[ online ] .
available : http://dx.doi.org/10.1007/s00450-010-0119-z p. klavk , a. c. i. malossi , c. bekas , and a. curioni , `` changing computing paradigms towards power efficiency , '' _ philosophical transactions of the royal society a _ , vol . 372 , no .
2018 , 2014 .
[ online ] .
available : http://dx.doi.org/10.1098/rsta.2013.0278 m. e. m. diouri , m. f. dolz , o. glck , l. lefvre , p. alonso , s. cataln , r. mayo , and e. s. quintana - ort , `` assessing power monitoring approaches for energy and power analysis of computers , '' _ sustainable computing : informatics and systems _ , vol . 4 , no . 2 , pp . 68 82 , 2014
r. miceli , g. civario , a. sikora , e. csar , m. gerndt , h. haitof , c. navarrete , s. benkner , m. sandrieser , l. morin , and f. bodin , `` autotune : a plugin - driven approach to the automatic tuning of parallel applications , '' in _ applied parallel and scientific computing _ , ser .
lecture notes in computer science.1em plus 0.5em minus 0.4emspringer , 2013 , vol . 7782 , pp .
328342 .
s. benedict , v. petkov , and m. gerndt , `` periscope : an online - based distributed performance analysis tool , '' in _ tools for high performance computing 2009_.1em plus 0.5em minus 0.4emspringer , 2010 , pp .
n. eicker , t. lippert , t. moschny , and e. suarez , `` the deep project - pursuing cluster - computing in the many - core era , '' in _ 2013 42nd international conference on parallel processing ( icpp13 ) _ , 2013 , pp . 885892 .
[ online ] .
available : http://dx.doi.org/10.1109/icpp.2013.105 x. aguilar , j. doleschal , a. gray , a. hart , d. henty , t. hilbrich , d. lecomber , s. markidis , h. richardson , and m. schliephake , `` the exascale development environment : state of the art and gap analysis ( cresta white paper ) , '' 2013 , accessed 2015 - 02 - 04 .
[ online ] .
available : http://cresta-project.eu/ | while the hpc community is working towards the development of the first exaflop computer ( expected around 2020 ) , after reaching the petaflop milestone in 2008 still only few hpc applications are able to fully exploit the capabilities of petaflop systems . in this paper
we argue that efforts for preparing hpc applications for exascale should start before such systems become available .
we identify challenges that need to be addressed and recommend solutions in key areas of interest , including formal modeling , static analysis and optimization , runtime analysis and optimization , and autonomic computing .
furthermore , we outline a conceptual framework for porting hpc applications to future exascale computing systems and propose steps for its implementation . |
cold fronts were discovered as sharp features in the x - ray surface brightness distribution in chandra observations of the clusters a2142 and a3667 ( markevitch et al . , 2000 , vikhlinin , markevitch & murray , 2001a ) , see also markevitch et al .
similar features have now been found in several other clusters ( e.g. sun et al . , 2002 ,
kempner , sarazin & ricker , 2002 ) .
unlike shocks , these features have lower gas temperature on the x - ray brighter side of the discontinuity .
for that reason they are called `` cold fronts '' .
it is believed that some cold fronts are formed when a subcluster merges with another cluster and the ram pressure of gas flowing outside the subcluster gives the contact discontinuity the characteristic curved shape . indeed , features resembling cold fronts
are found in numerical simulations of cluster formation ( bialek , evrard & mohr , 2002 , nagai & kravtsov , 2003 ) .
ablation of the gaseous cloud by the hot gas causes characteristic differential motion of the gas inside the subcluster , which transports the low entropy gas from the subcluster core towards the contact discontinuity , thus enhancing the jump in temperature and surface brightness across the discontinuity ( heinz et al .
2003 ) . here
we address the question of the front stability . as was pointed out by vikhlinin et al .
( 2001a , b ) , vikhlinin and markevitch ( 2002 ) the observed fronts are narrow and could be unstable to the kelvin - helmholtz ( kh ) instability .
for a3667 however the front appears to be narrow ( less than @xmath05 kpc in width ) up to @xmath030from the stagnation point .
magnetic fields can act as a stabilizing agent thus allowing indirect estimates of the field strength near the front ( vikhlinin , markevitch & murray,2001b ) . on the other hand numerical simulations without magnetic field ( and in particular relatively high resolution simulations by heinz et al .
2003 ) do not show instability of the front within 20 - 30 .
while this discrepancy could be due to numerical effects , we revisit the question of the front stability below and show that the characteristic convex geometry and finite ( small ) intrinsic width of the interface may help to stabilize the discontinuity with respect to kh instability .
the structure of the paper is as follows . in section 2
we derive a simple expression for the growth of the kh instability along the curved interface . in section 3
we argue that diffusion processes are likely to set an approximately constant width for the interface . in section 4
we discuss astrophysical applications .
the last section summarizes our findings .
for an infinitely thin plane parallel interface separating semi - infinite layers of two incompressible fluids with densities @xmath1 and @xmath2 the dispersion relation for kelvin - helmholtz instability is ( landau & lifshitz , 1987 ) : @xmath3 where @xmath4 is a wave number , @xmath5 is the wavelength of the perturbation , @xmath6 is the velocity of fluid with density @xmath1 , while fluid with density @xmath2 is at rest and @xmath7 .
the frequency @xmath8 is measured in the frame of the fluid at rest ( i.e. with density @xmath2 ) . for cold fronts in clusters
we adopt the geometry of the flow following vikhlinin et al .
2001a as shown in fig.[fig : geom ] .
the front has a curvature radius @xmath9 and the velocity @xmath6 along the interface is assumed to follow the law @xmath10 , where @xmath11 is the angle along the interface . for the potential flow of incompressible fluid @xmath12 , where @xmath13 is the velocity of the cold gas through the hot gas . for a particular case of the cold front in a3667 , vikhlinin et el .
( 2001b ) argue that due to compressibility @xmath14 is more close to @xmath13 . for simplicity
we assume that the densities of both fluids remain constant along the interface .
consider the perturbation of the interface separating cold and hot gas . from eq.([eq : dr ] ) it follows that perturbations are advected along the interface with the increasing velocity @xmath15 , where @xmath16 .
this implies that the wave number @xmath17 of the perturbation decreases as : @xmath18 where @xmath19 is the wave number at some initial position @xmath20
. this result can be formally obtained from the lowest order wkb ( wentzel - kramer - brillouin ) approximation ( see appendix ) , applicable if @xmath21 .
apart from the exponentially growing factor , the next order wkb approximation allows one to estimate slower changes of the perturbation amplitude @xmath22 caused by the changes of the velocity along the interface and stretching of fluid elements in the azimuthal direction ( see appendix ) . for estimates
we assume that @xmath23 , and @xmath24 . i.e. @xmath25 , where @xmath26 is the initial amplitude
. thus one can evaluate the growth of the perturbation propagating along the interface as : @xmath27 where @xmath28 is the increment of kh instability and one has to evaluate it using @xmath17 from eq.([eq : stra ] ) .
an analogous expression was used by inogamov & chekhlov ( 1991 ) for the growth of perturbations for rayleigh - taylor instability . from eq .
( [ eq : dr ] ) and ( [ eq : stra ] ) it is clear that the increment @xmath29 does not increase as the perturbation propagates , since the increase of the velocity amplitude is compensated by the decrease of the wave number .
thus @xmath30 , where @xmath31 .
replacing @xmath32 with @xmath33 one can write an explicit expression for @xmath34 : @xmath35 .\end{aligned}\ ] ] thus the final expression for the growth factor for a perturbation starting at @xmath20 with the wave number @xmath36 measured at position @xmath37 is : @xmath38 |_{\theta_1}^{\theta_2 } \right \}. \label{eq : gft}\end{aligned}\ ] ] one can compare this expression with eq .
( 10 ) from vikhlinin & markevitch ( 2002 ) - i ) an additional factor appears in front of the exponential and more importantly ii ) the argument of the exponential is changed due to the variation of the velocity along the interface , thus * reducing the growth factor * for a fixed @xmath36 and @xmath20 .
we note here that from the point of view of the degree of nonlinearity of a perturbation one can use the ratio of the amplitude to the wavelength : @xmath39 . therefore the expression for the degree of nonlinearity of a perturbation will contain an additional factor proportional to @xmath40 in front of the exponential . alternatively , since @xmath41
, we may regard the growth factor as a function of @xmath42 specified at the angle @xmath37 .
the formal condition for the growth factor ( 5 ) to diverge for small @xmath20 is @xmath43 .
thus in spite of the * suppression * of the growth factor for a fixed * initial * wave number , the growth factor can be infinitely large for a fixed * final * wave number .
there are however several factors which can limit the growth of instability .
we argue below that a small , but finite , thickness of the interface is the most important factor in the conditions relevant for cold fronts in clusters .
any real interface can not be infinitely thin . if the thickness of the interface is @xmath44 , then only modes with @xmath45 are unstable .
thus for a fixed @xmath37 and @xmath46 the minimum angle at which kh instability starts is such that @xmath47 . in section 3 we argue that diffusive processes across the interface may set a finite thickness of the interface , which is almost independent of the position ( angle ) .
one can approximately account for the finite width of the interface by using @xmath48 as a lower bound in eq.([eq : gft ] ) .
but in fact the largest contribution to the growth factor comes from small angles and this part of the integral has to be evaluated more accurately than simply introducing a cutoff in @xmath11 .
one can use for instance the dependence of increment on the wave number and thickness in the form given by rayleigh ( see `` the theory of sound '' , e.g. 1945 edition ) for two fluids with similar densities : @xmath49 where @xmath50 is the condition for instability .
thus the growth factor is : @xmath51 where @xmath52 is such that @xmath53 . in this case
the growth factor has to be calculated numerically . in fig.[fig : growth ] we show the total growth factor calculated for several values of @xmath37 and @xmath44 . for @xmath54
30the instability grows very strongly if the thickness of the interface is less than @xmath01.5 - 2% of the curvature radius .
we note that all the above expressions for the growth factor do not depend explicitly on the absolute value of @xmath14 , indicating that the appearance of the front may be similar for clouds moving with different velocities , if the relation between the width and the curvature of the interface is the same .
an interesting question is if the nonlinear evolution of kh instability itself can be responsible for establishing the width of the interface ( along the lines of reasoning given by nulsen , 1982 ) of order of few percent of @xmath9 , smoothing an interface and suppressing growth of longer modes .
we leave this question for subsequent studies and in the next section consider the width of the interface set by diffusion .
for flow past a flat plate ( with high reynolds number ) the width of the boundary layer is known to vary as the square root of the distance @xmath55 from the front edge of the plate ( blasius law ) : @xmath56 where @xmath57 is the viscosity . for a rounded body ,
the flow velocity increases linearly with distance from the stagnation point , @xmath58 , making the width of the boundary layer constant ( landau & lifshitz , 1987 ) .
thus @xmath59 , where @xmath9 is the curvature radius . substituting further @xmath60 , where @xmath61 and @xmath62 are the rms velocity and mean free path for ions and assuming that @xmath63 one gets an order of magnitude estimate of the expected width @xmath64 . in the absence of magnetic field the viscosity of the hot gas is larger than for the cold gas and most of the velocity drop may take place on the hotter side in the form of a broad layer .
we note here that observationally it is almost impossible to `` see '' such a layer .
analogously one can consider the case of a scalar diffusion ( e.g. thermal conduction ) with a constant diffusion coefficient @xmath65 .
assuming that conditions inside the cloud ( where the fluid is not moving ) are frozen and seeking a stationary solution of a diffusion equation in outer gas it is easy to show that @xmath66 .
both expressions , when applied to the case of a3667 and compared with the upper limit on the front width ( vikhlinin et al . , 2001a ) , indicate significant suppression of transport processes across the front .
more accurate calculations and a detailed comparison with observations are beyond the scope of this paper and will be reported elsewhere . in what follows
we assume that the interface has a constant width independent of @xmath11 and the expression ( [ gfray ] ) for the growth factor is applicable .
we now apply the above results to the particular case of the cold front in a3667 .
we set the following parameters based on the results of vikhlinin et al .
, 2001a , b , vikhlinin & markevitch 2002 : @xmath67 kpc , @xmath68 , @xmath69 .
they give the width of the front ( upper limit ) in terms of the standard deviation of a gaussian smoothing , as @xmath70 kpc . in our notation , @xmath71 . in fig.[fig : a3667 ] we plot the growth factor calculated for several values of @xmath72 and angle @xmath37 .
this is largely the same figure as fig.[fig : growth ] but now in natural units . for comparison
we show ( upper dotted curve ) the growth factor calculated for an infinitely thin interface , assuming fixed wave number @xmath17 and @xmath7330 .
one can see that accounting for varying wave number and the finite thickness of the interface ( within the observational limits , i.e. @xmath70 kpc ) drastically reduces the growth factor . for these reduced values
it is not obvious that small initial perturbations can grow strongly nonlinearly and disturb the sharp appearance of the front around @xmath5430 . for larger angles or for smaller widths of the interface
the growth factors are large . while we assumed above that the gas inside the cloud is still , internal motions are likely to appear during the formation of the front and especially during later periods of the cloud evolution when continuous stripping at the sides of the cloud is compensated by the cold gas flow along the interface to replenish the losses .
e.g. in simulations by heinz et al .
( 2003 ) the velocity of the gas inside the cloud @xmath00.2 of the velocity of the surrounding gas at the moment when the morphology of the front and the temperature distribution resemble those observed in a3667 as follows from the bernoulli equation .
due to gravity the velocity of the motions in cold gas should be lower . ] .
the internal motions have a two - fold effect : i ) they reduce the shear rate and ii ) they increase the group speed of the perturbation in the interface , both of which reduce the total growth of a perturbation arriving at the angle @xmath37 .
accounting for internal motions can easily reduce the argument of the exponential in equation ( [ gfray ] ) by 30 - 50% thus strongly reducing the final growth factor compared to fig.[fig : a3667 ] . in the above discussion
we have completely neglected the role of gravity , although we assumed that it plays an important role in preserving the integrity of the cold cloud when it passes through the hot gas . for an infinitely thin interface
, gravity modifies the dispersion relation ( [ eq : dr ] ) to : @xmath74 where @xmath75 is gravitational acceleration . for a fixed @xmath42 and @xmath37 and
given that @xmath76 and @xmath77 the role of gravity is more important for small angles @xmath11 , effectively introducing a lower limit on @xmath20 .
the growth factor calculated with the above dispersion relation and using the values of @xmath75 and @xmath14 from vikhlinin & markevitch ( 2002 ) is very large , implying that gravity does not strongly suppress kh instability , in agreement with their conclusion . for an interface of finite thickness
, gravity makes the region near the stagnation point stable for perturbations of any @xmath17 .
one can estimate the richardson number of the flow assuming that density and velocity both vary over a layer of thickness @xmath44 : @xmath78 the sufficient condition for stability is @xmath79 everywhere in the flow . again plugging in the numbers from vikhlinin & markevitch ( 2002 ) we find that for @xmath80 kpc the front is stable up to angles of @xmath06,11,14respectively .
simple estimates show that account for this stable region would further reduce the growth factor compared to fig.[fig : a3667 ] , although not dramatically .
this is expected given that gravity alone ( for the thin interface case - see above ) does not strongly suppress kh instability .
accurate calculations of the growth factor taking account of gravity and the finite thickness of the interface require knowledge of the density and velocity profiles of the interface and are beyond the scope of this paper .
we note only that stretching of the flow in the stable region near the stagnation point insures that initial perturbations in the unstable regions further downstream are small .
since we are considering motion near the stagnation point the velocities are small compared to the gas sound velocity ( even for colder gas ) and compressibility does not play a significant role .
further downstream the compressibility starts to be more important and it reduces the growth factor of the perturbation ( see dispersion relation in miles , 1958 , gerwin , 1968 or nulsen 1982 ) .
e.g. for the density ratio @xmath81 the quantity @xmath82}{\rm re[\omega]}$ ] is equal to @xmath83 for small shear velocity and it drops to @xmath84 when the shear velocity approaches the sound speed of the colder gas ) is proportional to this factor ] .
we have shown that accounting for wave number changes along the curved interface and finite ( but small ) width of the interface significantly reduces the kh instability growth factor .
we also argue that if diffusion sets the intrinsic width of the interface then this width does not vary much along the interface ( at least for small angles ) . for a set of parameters relevant for the front in a3667
the growth factor is not large enough to guarantee nonlinear growth at angles @xmath030 , provided that the intrinsic width of the front is of the order of a few percents of the curvature radius ( i.e. not much less than the existing observational limit ) .
therefore dynamically important magnetic fields may not be necessary to stabilize this particular cold front .
the growth factor ( and therefore possible limits on the magnetic fields ) * depends critically on the width of the interface * and much more weakly on the angular extent over which the front remains sharp .
a factor of 2 - 3 improvement in the observational constraints on the interface width is therefore extremely important . for a given width of the interface the value of the growth factor does not depend strongly on the absolute value of the shear velocity .
this probably explains the ubiquity of fronts both in observations and hydrodynamical simulations .
we are grateful to sebastian heinz , rashid sunyaev and alexey vikhlinin for numerous useful discussions .
we would like to thank the referee , paul nulsen , for a number of important and constructive comments and suggestions .
j. j. , evrard a. e. , mohr j. j. , 2002 , , 578 , l9 gerwin r. a. , 1968 , rvmp , 40 , 652 heinz s. , churazov e. , forman w. , jones c. , briel u. g. , 2003 , mnras , 346 , 1 inogamov n. a. , chekhlov a. v. , 1991 , pzhtf , 17 , 58 j. c. , sarazin c. l. , ricker p. m. , 2002 , , 579 , 236 landau l.d . , lifshitz , e.m .
, course of theoretical physics , vol . 6 : fluid mechanics , pergamon , new york , 1987 markevitch m. , ponman t. j. , nulsen p. e. j. , bautz m. w. , burke d. j. , david l. p. , davis d. , donnelly r. h. , forman w. r. , jones c. , kaastra j. , kellogg e. , kim d .- w .
, kolodziejczak j. , mazzotta p. , pagliaro a. , patel s. , van speybroeck l. , vikhlinin a. , vrtilek j. , wise m. , zhao p. , 2000
, , 541 , 542 m. , vikhlinin a. , forman w. , 2003 , in matter and energy in clusters of galaxies , stuart bowyer and chrong - yuan hwang , eds . , asp conference proceedings , v.301 , 37 .
miles , j.w .
1958 , j. fluid mech . , 4 , 538 d. , kravtsov a. v. , 2003 , , 587 , 514 nulsen p. e. j. , 1982 , mnras , 198 , 1007 rayleigh , j. w. s. the theory of sound , vol .
2 , new york : dover , 1945 . m. , murray s. s. , markevitch m. , vikhlinin a. , 2002 , , 565 , 867 a. , markevitch m. , murray s. s. , 2001a , , 551 , 160 a. , markevitch m. , murray s. s. , 2001b , , 549 , l47 vikhlinin a. a. , markevitch m. l. , 2002 , astl , 28 , 495
consider a plane boundary between two incompressible , inviscid , irrotational fluids with densities @xmath1 and @xmath2 moving in the @xmath55 direction with velocities @xmath85 and @xmath86 .
the fluid 1 is accelerating along the interface : @xmath87 . to match pressures at the boundary we have to assume that the velocity of fluid 2 is @xmath88 and is also accelerating along the interface .
we now consider small perturbations of the boundary @xmath89 and velocity potentials of two fluids @xmath90 .
the governing equations are : @xmath91 where @xmath92 and @xmath93 are the velocity components in @xmath55 and @xmath94 directions for two flows .
the second equation is the kinematic boundary condition and the third equation is the bernoulli equation .
these equations are evaluated at the boundary @xmath95 .
we are seeking a solution of these equations in the form @xmath96 and @xmath97 , which decays for @xmath98 , where @xmath99,@xmath100 , and all @xmath101 are slowly varying functions and @xmath102 is a small parameter . using these expressions and keeping only the terms of order @xmath103 we recover a dispersion relation , which we write in a `` spatial '' form , denoting @xmath104 , @xmath105 : @xmath106 now collecting higher order terms and using the relations @xmath107 from the laplace equation and @xmath108 from the bernoulli equation we get a transport equation for spatial variations of the amplitude @xmath100 : @xmath109 thus @xmath110 incompressibility implies that @xmath111 .
the last equality is due to assumed form of acceleration .
therefore @xmath112 i.e. the amplitude decreases with the distance as @xmath113 .
if additional uniform stretching in the 3rd dimension is present then one has to set @xmath114 . in the case of interest for us ( potential flow past a sphere ) one can simply set @xmath115 . substituting this into eq .
( [ tr ] ) we get @xmath116 | cold fronts sharp discontinuities recently discovered by chandra in many clusters of galaxies are believed to be due to a hot gas flow over a colder gravitationally bound gas cloud .
we analyze the stability of the fronts with respect to kelvin - helmholtz instability and show that an intrinsic width of the interface of the order of a few per cent of the curvature radius strongly limits the growth of perturbation . for the best studied case of a front in the cluster abell 3667 we conclude that current observational data on the width and extent of the front can be explained even in the absence of dynamically important magnetic fields .
epsf # 1= # 1#2=.99 = .99 # 1#2#3=.31 = .31 = .31 [ firstpage ] galaxies : clusters : individual : a3667 , x - rays : galaxies : clusters |
9 g. nakamoto _ et al . _ , j. phys .
soc . jpn . * 64 * , 4834 ( 1995 ) . | previously measured thermopower data of cenisn exhibit a significant sample dependence and non - monotonous behavior in magnetic fields . in this paper
we demonstrate that the measured thermopower @xmath0 may contain a contribution from the huge nernst coefficient of the compound , even in moderate fields of 2 t. a correction for this effect allows to determine the intrinsic field dependence of @xmath0 .
the observed thermopower behavior can be understood from zeeman splitting of a v - shaped pseudogap in magnetic fields . *
introduction : * the orthorhombic system cenisn has been classified as a kondo semimetal , in which an anisotropic pseudogap opens in the density of states ( dos ) below approximately 10 k @xcite .
various experimental probes confirm the presence of a finite quasiparticle dos at the fermi level , such as the metal - like resistivities of samples with high purity @xcite and the linear - in-@xmath1 dependence of the thermal conductivity below 0.3 k @xcite .
magnetic fields of the order of 10 t along the easy magnetic @xmath2 axis suppress the gap formation significantly , while fields along @xmath3 and @xmath4 are less effective @xcite .
the thermopower @xmath5 of cenisn is highly anisotropic and exhibits a significant sample dependence @xcite .
below 10 k , the absolute values of @xmath5 are enhanced , which has been attributed to the gap formation in this temperature range @xcite .
consequently , application of a magnetic field of 8 t along the easy @xmath2 axis has been found to induce a significant lowering of the thermopower along @xmath2 , @xmath6 , at low @xmath1 @xcite .
however , the experimental data for @xmath7 and @xmath8 with @xmath9 are inconsistent .
first measurements at 4.2 k and 1.3 k showed a lowering of @xmath8 in magnetic fields @xcite , while investigations on samples of higher purity revealed an increasing @xmath8 upon increasing field @xcite .
likewise , for @xmath7 at 1.3 k either a monotonous decrease up to 10 t @xcite or an increase in 4 t with a subsequent decrease in 8 t @xcite was found . for correlated semimetals as cenisn
a large nernst coefficient has been predicted @xcite .
the corresponding transverse thermal voltage can become comparable in magnitude to the longitudinal one even in moderate magnetic fields .
in such a case , small deviations from the ideal contact geometry may give rise to a non - negligible nernst contribution to the measured thermopower , an effect which has not been considered previously @xcite .
it is expected to be most relevant for @xmath7 , which exhibits significantly lower absolute values at low temperatures than @xmath6 and @xmath8 . in this work
we discuss the intrinsic behavior of @xmath10 for @xmath9 and @xmath4 determined from measurements in positive and negative fields up to 7 t. + + * experiments : * the investigated samples originate from the single crystal # 5 , which was grown by the czochralski technique and subsequently purified by the solid - state electrotransport ( sse ) technique as described in ref .
@xcite .
thermopower and nernst coefficient have been measured with a steady - state method .
the temperature gradient along the samples was determined using a chromel - aufe@xmath11 thermocouple . in order to correct for spurious contributions due to a non - ideal contact geometry
all measurements were performed in @xmath12 .
two configurations are presented : the heat current @xmath13 was applied along @xmath3 with @xmath9 ( sample 1 ) and @xmath14 ( sample 2 ) . + + * results : * and @xmath9 .
exemplary error bars indicate the uncertainty of the data at low @xmath1 and high @xmath15 . above 4 k
it is typically less than 1 @xmath16v / k .
the inset compares the zero - field data obtained on two samples of cenisn .
( b ) the in - field antisymmetric contribution of the measured thermopower @xmath17 in comparison to the nernst signal @xmath18 in 2 t. @xmath19 can be scaled to @xmath18 by a factor of 8 .
[ cns - svst],scaledwidth=96.0% ] the zero - field thermopower of cenisn measured along @xmath3 , @xmath10 , is shown in the inset of fig .
[ cns - svst ] .
the data sets obtained on two different samples agree relatively well in the whole investigated temperature range . above 10 k , the thermopower exhibits a similar @xmath1 dependence as reported for a high - purity single crystal without sse treatment @xcite : @xmath10 is positive with maxima at 20 k and 100 k. the two maxima are attributed to kondo scattering from the ground - state doublet and thermally populated cef levels . a contribution from paramagnon drag
was also suggested as a possible origin for the maximum at 20 k @xcite .
toward lower @xmath1 , @xmath20 changes sign at 8 k and goes through a large negative minimum of @xmath21 @xmath16v / k at around 3.5 k. this value represents the largest negative @xmath5 ever observed for a cenisn sample .
below 1.7 k the thermopower assumes again positive values .
a similar temperature dependence has been observed in samples grown by a czochralski technique without sse treatment @xcite .
these samples exhibit a negative @xmath7 between 2.5 and 7 k , however , with significantly smaller absolute values of at most -6 @xmath16v / k .
by contrast , investigations on single crystals of similar quality as those presented here yielded a thermopower @xmath10 with opposite sign and a maximum value of 8 @xmath16v / k at 3.5 k @xcite . ,
@xmath22 ( a ) , and the thermopower value at @xmath22 , @xmath23 ( b ) , for different orientations of @xmath15 .
the dashed lines are meant guides to the eye .
the solid line in the left plot is a linear fit to the data for @xmath24 .
the error bars given for @xmath23 represent the scattering of the data around @xmath22 .
[ cns - min_vsb],scaledwidth=96.0% ] the effect of a magnetic field along @xmath2 on the low-@xmath1 thermopower @xmath7 is shown in the main plot of fig .
[ cns - svst ] . with increasing @xmath15 the minimum at @xmath25 k
shifts to lower @xmath1 and the absolute values at the minimum @xmath26 are enhanced for @xmath27 t. in 7 t a weak lowering of @xmath26 is observed , which , however , is of the order of the uncertainty in the data .
application of a magnetic field along @xmath4 gives rise to a similar evolution of @xmath0 ( not shown ) .
compared to the configuration @xmath9 , the shift of the minimum is less pronounced and no lowering of @xmath26 is observed around 7 t , ( cf .
fig .
[ cns - min_vsb ] ) . fig .
[ cns - min_vsb]a shows the field dependence of @xmath22 for @xmath9 and @xmath14 .
it clearly reveals that the shift of the minimum is stronger for @xmath9 . a linear extrapolation of @xmath28 to zero temperature yields a critical field of 14 t.
this value is comparable to the energy - gap quenching field along @xmath2 of 18 t determined from resistivity measurements @xcite .
therefore , it is supposed that the shift of the minimum is related to the closing of the gap in field .
fig .
[ cns - min_vsb]b shows the evolution of the thermopower value at the minimum , @xmath29 for @xmath9 and @xmath30 .
while @xmath26 first increases with increasing field , a saturation and subsequent reduction is observed at higher @xmath15 .
again , the effect of a field along @xmath2 is more pronounced than that of @xmath31 .
it is suspected that @xmath26 for both orientations is further reduced in higher @xmath15 as the minimum shifts to lower @xmath1 .
+ + * discussion : * the strong sample dependence of @xmath0 of cenisn reported in literature has been related to the differing purity of the investigated crystals @xcite . however , two other effects have to be taken into account .
firstly , in view of the sensitive direction dependence of @xmath0 it can not be excluded that tiny misorientations are responsible for at least part of the reported variations .
the large negative @xmath7 observed around 3 k in the present investigation could be easily diminished by a small contribution from the huge positive @xmath6 and @xmath8 of up to 70 @xmath16v / k expected in the same temperature range @xcite . in this context , the small discrepancy of about 20 % between @xmath32 of sample 1 and 2 might be attributed to this effect . secondly , the large nernst signal expected for cenisn can contribute to the thermopower voltage for a non - ideal contact geometry . in the present investigation ,
this effect was already appreciable in 2 t as demonstrated in fig .
[ cns - svst]b . below 8 k
the measured thermopower curves @xmath33 for positive and negative fields differ significantly .
the in - field antisymmetric contribution @xmath34 $ ] exhibits a temperature dependence similar to that of the nernst signal @xmath35 in 2 t and can be scaled to it by a factor of 8 .
this corresponds to a misorientation of the contacts by only @xmath36 . for the presented data
this effect has been corrected for by averaging between @xmath37 and @xmath38 .
however , it has generally not been accounted for in previous investigations @xcite .
in particular , the drastic and non - monotonous change in @xmath7 around 2 k reported for crystals of similar quality as those investigated here @xcite might be to some extent influenced by the huge nernst signal in this @xmath1 range .
therefore , the current study represents the first investigation with the intrinsic temperature dependence of @xmath7 in fields up to 7 t. for different @xmath15 .
( c ) field dependence of the logarithmic derivative of the dos at @xmath39 .
[ cns - fig4x1],scaledwidth=96.0% ] application of magnetic fields @xmath40 is found to induce a systematic shift of the thermopower minimum at 3.5 k to lower temperatures ( cf .
fig .
[ cns - min_vsb]a ) .
the estimated critical field along @xmath2 of 14 t as well as the stronger effect for @xmath9 compared to @xmath41 confirms that the minimum is related to the gap formation in cenisn . in this context , the enhancement of @xmath42 in parallel with the closing of the gap in field ( fig .
[ cns - min_vsb]b ) is surprising .
apparently , application of a magnetic field does not only suppress the gap but also influences the residual dos inside the gap .
a similar effect has been found from investigations of the specific - heat @xmath43 of cenisn @xcite , which revealed an enhancement of @xmath43 at low @xmath1 in applied magnetic fields @xmath9 .
this observation had been interpreted within a simple model assuming zeeman splitting of a modified v - shaped dos @xcite .
an illustration , how the same mechanism can give rise to an enhancement of @xmath44 at low @xmath15 is depicted in fig .
[ cns - fig4x1 ] : application of a magnetic field induces a shift of the sub - bands for the spin - up and spin - down states to different energetic directions , as shown for an arbitrary field in fig . [ cns - fig4x1]a .
thus , the resulting total dos around the gap structure depends sensitively on the field magnitude ( fig .
[ cns - fig4x1]b ) .
if the fermi level is situated slightly off the symmetry line of the gap structure , the slope of the dos at @xmath45 becomes field dependent ( fig .
[ cns - fig4x1]c ) . within this simplified picture , the absolute values of the thermopower @xmath46
@xcite increase in small magnetic fields and decreases for higher @xmath15 .
it is admitted , that the sketched behavior is much too simple to explain in detail the behavior of @xmath47 observed for cenisn .
nevertheless , the presented picture demonstrates that the increase in @xmath44 is not fundamentally contradictory to the suppression of the gap in magnetic fields .
zeeman splitting appears a possible and simple mechanism to understand the observed behavior .
it seems likely that a realistic band structure and the allowance for thermal broadening enables a more sophisticated description of the data . in conclusion ,
the minimum in @xmath10 of cenisn at 3 k is attributed to the gap formation and the observed field dependence is related to the suppression of the gap in magnetic fields . |
we propose a feature allocation model [ broderick , jordan and pitman ( @xcite ) ] to describe tumor heterogeneity using next - generation sequencing ( ngs ) data .
we use a variation of the indian buffet process ( ibp ) [ @xcite ] .
the feature allocation in our model is latent .
that is , the features are not directly observed .
we record point mutations as single nucleotide variants ( snvs ) , each of which is defined as a dna locus that possesses a variant sequence from that on the reference human genome .
we use the feature allocation model to describe unobserved haplotypes , defined as a collection of single nucleotide variants ( snvs ) scaffolded on a homologous genome . in a tumor sample
, having more than two haplotypes is evidence of heterogeneous cell subpopulations with a distinct genome .
this is the case because humans are diploid and we would therefore only observe up to two haplotypes if all cells in a tumor sample were genetically homogeneous . in the proposed application of feature allocation models to inference for tumor heterogeneity ,
the haplotypes are the features and the snvs are the experimental units that select the features .
the number of features is unknown .
each tumor sample is composed of an unknown proportion of each of these haplotypes .
the top level sampling model for the observed snv counts is then defined as binomial sampling with a proportion for each snv that is implied by this composition . in summary , we solve a deconvolution problem to explain the observed snv frequencies for each sample by compositions of latent haplotypes .
heterogeneity in cancer tissue has been hypothesized over the past few decades [ @xcite ] and has been demonstrated elegantly using ngs technology over the past few years [ @xcite ] .
genetic variation in a tumor occurs due to evolutionary processes that drive tumor progression .
specifically , tumors include distinct clonal subpopulations of cells that arise stochastically by a sequence of randomly acquired mutations .
substantial genetic heterogeneity between tumors ( inter - tumor ) or within a tumor ( intra - tumor ) can be explained by differences in clonal subpopulations and varying proportions of those subpopulations [ @xcite ] .
for example , @xcite reported clonal genomic heterogeneity in breast cancers .
data derived from ngs experiments include snvs , small indels and copy number variations [ @xcite ] .
many researchers use snv data to investigate genes and genomic regions related to cancer phenotypes [ @xcite ] . in this paper
, we utilize whole - genome sequencing data measuring variant allele fractions ( vafs ) at snvs to understand tumor heterogeneity by proposing inference on how haplotypes may be distributed within a tumor . in an ngs experiment ,
millions of short dna reads are generated and are then aligned to the reference genome . at certain positions of the genome ,
some or all of the mapped reads will show a sequence different from what is in the reference genome . at each genomic locus , the proportion of short reads bearing a variant sequence is called the vaf . if the vaf at a locus is nonzero , an snv may be `` called '' at that locus , based on statistical inference [ @xcite ] .
the raw experimental data comprises the total number of reads ( @xmath0 ) that are mapped to the locus and the number of those reads ( @xmath1 ) indicating a variation from the reference sequence .
then vaf @xmath2 . if a tumor sample is homogeneous , that is , having a single clone , the vaf values of all the snvs should be close to 0 , 0.5 or 1 , reflecting the three possible homozygous and heterozygous alleles ( i.e. , aa , ab , bb ) at any snv . different vaf values imply heterogeneity of the cellular genome in the tumor sample ( see figure [ fig : data_1 ] for an example ) . we propose to study inference to deconvolute the vafs from multiple snvs and back out the latent haplotypes .
@c@ + + + we propose a bayesian feature allocation model to characterize such cellular heterogeneity in a way that explains the observed ngs data .
we construct a matrix of binary features ( equivalently , haplotypes ) as shown in figure [ fig : demo_z ] . in the figure
, columns correspond to haplotypes and rows correspond to snvs .
we define haplotype @xmath3 by a binary vector @xmath4 of indicators of whether ( @xmath5 ) or not ( @xmath6 ) a variant sequence is observed at the snv @xmath7 . here
we view snv as a genetic locus on which either a variant or reference dna sequence could be observed .
figure [ fig : demo_z ] illustrates the definition of five haplotypes ( @xmath8 , columns ) with @xmath9 snvs ( rows ) . in the figure , black ( white ) indicates @xmath5 ( @xmath6 ) .
for example , snv 1 in figure [ fig : demo_z ] possesses a variant sequence in the two haplotypes @xmath10 and @xmath11 .
on the other hand , snv 9 possesses variant sequences in four haplotypes : @xmath12 and @xmath13 .
a prior probability model on such a binary matrix @xmath14 $ ] is known as a feature allocation model . here , we assume that @xmath15 is unknown and place a prior on @xmath15 . assuming that samples are composed of different proportions of @xmath15 haplotypes , we aim to fit the observed vafs of the snvs to infer these proportions .
for example , we may observe that one sample is dominated by haplotypes @xmath16 and @xmath13 , while another is dominated by haplotypes @xmath17 and @xmath18 . if the samples are from the same tumor , the differences in haplotypic compositions are evidence of _ intra - tumor _ heterogeneity ; on the other hand , differences in samples from different tumors imply _
inter - tumor _ heterogeneity .
therefore , the proposed inference provides a unified framework to address inference for both biological concepts .
importantly , the characterization of haplotypes is based on selected snvs only .
otherwise inference for heterogeneity between tumors in different patients would not be biologically meaningful , as cellular genomes and haplotypes are not expected to be shared across patients .
however , for tumors in the same class of disease , snvs in local disease - related genomic regions may be common to all or some of the tumors , thereby allowing for the proposed inference .
there are currently few approaches that address the problem of tumor heterogeneity . @xcite and @xcite recognized that a tumor sample is a mixture of normal cells and tumor cells , and developed a method to estimate tumor purity levels for paired tumor - normal tissue samples using dna sequencing data
. none of the two methods considers more than two samples or unpaired samples .
purbayes [ @xcite ] accounts for intra - tumor heterogeneity , but it does not provide inference on the subpopulation configurations as inference on the latent matrix @xmath19 under the proposed model .
pyclone [ @xcite ] , a recently published method , proposes inference to cluster snvs with different vafs .
an underlying assumption of pyclone is that snvs can be arranged in clusters that inform about subclones .
a key component of pyclone is the use of clustering models such as the dirichlet process for inference on these clusters .
while such clusters are informative about heterogeneity , inference that is provided by pyclone is not meant to identify subclones or haplotypes .
the primary aim of pyclone is inference on mutation clusters , defined as a group of snvs with similar variant allele fractions .
in contrast , our proposed feature allocation model explicitly models the haplotypic genomes of subclones , allowing overlapping snvs shared between different subclones .
we do not use nonoverlapping snv clusters as the building block for subclones .
that is , instead of first estimating the snv clusters and then constructing subclones based on clusters , we directly infer the subclonal structure based on haplotypes .
we show in later examples the distinction between pyclone and our proposed method .
the remainder of the paper is organized as follows : section [ sec : model_sel ] describes the proposed bayesian feature allocation model and a model selection criterion to select the number of subclones .
section [ sec : simulation ] describes simulation studies .
sections [ sec : example ] and [ sec : lung_cancer ] report data analyses of in - house data sets to illustrate inter - tumor heterogeneity and intra - tumor heterogeneity , respectively .
the last section concludes with a final discussion . and @xmath20 .
let @xmath21 denote an @xmath22 matrix of count data from an ngs genome sequencing experiment , with @xmath23 denoting the number of reads that bear a variant sequence at the location of snv @xmath7 for tissue sample @xmath24 , @xmath25 and @xmath26 .
we assume a binomial sampling model .
let @xmath27 denote the total number of reads in sample @xmath24 that are mapped to the genomic location of snv @xmath7 .
we assume @xmath28 in figure [ fig : demo_bin ] , @xmath20 and @xmath29 .
we do not model @xmath27 , that is , we treat @xmath27 as fixed , and only consider a sampling model for @xmath23 conditional on @xmath27 ( modeling @xmath27 would not contribute any information on tumor heterogeneity based on snvs ) .
conditional on @xmath27 , the observed counts @xmath23 are independent across @xmath7 and @xmath24 . the model in ( [ eq : like ] )
is illustrated in figure [ fig : demo_bin ] .
we build a prior probability model for @xmath30 in two steps , using the notion that each sample is composed of a mixture of different haplotypes . and
each haplotype , in turn , is characterized by the haplotypes consisting of the snvs .
let @xmath31 denote the proportion of haplotype @xmath3 in sample @xmath24 and let @xmath32 denote a latent indicator of the event that snv @xmath7 bears a variant sequence for haplotype @xmath3 .
note that @xmath33 corresponds to a black block in figure [ fig : demo_z ] . then @xmath30 is written as a sum over @xmath15 latent haplotypes @xmath34 the construction of the haplotypes , including the number of haplotypes , @xmath15 , and the indicators @xmath35
are latent . the key term , @xmath36 , indirectly infers haplotypes by explaining @xmath30 as arising from sample @xmath24 being composed of a mix of hypothetical haplotypes which do ( @xmath5 ) or do not ( @xmath6 ) include a mutation for snv @xmath7 .
the indicators @xmath35 are collected in a @xmath37 binary matrix @xmath38 .
the number of latent haplotypes , @xmath15 , is unknown .
conditional on @xmath15 , the binary matrix @xmath38 describes @xmath15 latent tumor haplotypes present in the observed samples .
joint inference on @xmath15 , @xmath39 and @xmath40 explains tumor heterogeneity .
in addition , we introduce a background haplotype , labeled as haplotype @xmath10 , which includes all snvs .
the background haplotype accounts for experimental noise and haplotypes that appear with negligible abundance .
specifically , @xmath41 in ( [ eq : model_p ] ) relates to this background haplotype , with @xmath42 being the relative frequency of observing a mutation at an snv due to noise and artifact ( we assume equal frequency for all snvs ) and @xmath43 being the proportion in sample @xmath24 .
the prior on @xmath43 is defined later . for @xmath42
, we assume @xmath44 with @xmath45 to inform a small @xmath42 value a priori .
we start the prior construction with a prior for the number of haplotypes , @xmath15 .
we consider a geometric distribution , @xmath46 where @xmath47 .
conditional on @xmath15 , we use a feature allocation model for a binary matrix @xmath19 .
we first define the model for any given @xmath15 and start with feature - specific selection probabilities , @xmath48 let @xmath49 .
the selection probabilities are used to define @xmath50 as @xmath51 where @xmath52 is the number of snvs bearing variant sequences for haplotype @xmath3 .
a limit of the model , as @xmath53 , becomes a constructive definition of the indian buffet process ( ibp ) [ @xcite ] .
the model is symmetric with respect to arbitrary indexing of the snvs , simply because of the symmetry in ( [ eq : cond_z ] ) and ( [ eq : p_mu ] ) .
note that @xmath54 is possible with positive prior probability .
next , we consider a prior distribution for abundances associated with the haplotypes defined by @xmath19 . the haplotypes are common for all tumor samples , but the relative weights in the composition ( [ eq : model_p ] ) are different across tissue samples .
we assume dirichlet priors for the relative weights @xmath31 , defined as follows .
let @xmath55 denote an ( unscaled ) abundance level of haplotype @xmath3 in tissue sample @xmath24 .
we assume @xmath56 for @xmath57 and @xmath58 .
we then define @xmath59 as the relative weight of haplotype @xmath3 in sample @xmath24 .
this is equivalent to @xmath60 for @xmath61 , where @xmath62 .
recall the binomial sampling likelihood ( [ eq : like ] ) with success probability , @xmath30 . given @xmath15 , @xmath19 and @xmath63 ,
we define @xmath30 in ( [ eq : model_p ] ) . in words , @xmath30 is determined by @xmath15 , @xmath19 and @xmath40 with the earlier describing the latent haplotypes and the latter specifying the relative abundance of each haplotype in sample @xmath24 .
let @xmath64 , where @xmath65 .
markov chain monte carlo ( mcmc ) posterior simulation proceeds by sequentially drawing random numbers for the parameters in @xmath66 .
given @xmath15 , such mcmc simulation is straightforward .
specifically , gibbs sampling transition probabilities are used to update @xmath35 , and metropolis hastings transition probabilities are used to update @xmath67 and @xmath42 .
it is possible to improve the mixing of the markov chain by updating all columns in row @xmath7 of the matrix @xmath19 jointly by means of a metropolis
hastings transition probability that proposes changes in the entire row vector @xmath68 .
the construction of transition probabilities that involve a change of @xmath15 is more challenging , since the dimension of @xmath19 and @xmath69 changes as @xmath15 varies .
we use a reversible jump ( rj ) mcmc algorithm for posterior simulation [ @xcite ] .
we first define a proposal distribution @xmath70 for @xmath15 , and then introduce a proposal distribution @xmath71 for @xmath66 conditional on the proposed @xmath72 .
the latter potentially involves a change in dimension of the parameter vector .
we found that high posterior correlation of @xmath19 and @xmath63 conditional on @xmath15 greatly complicated the construction of a practicable rj scheme . to overcome this
, we use an approach similar to @xcite .
we split the data into a minimal training set @xmath73 with @xmath74 , @xmath75 , and a test data set , @xmath76 with @xmath77 etc . in the implementation we use
@xmath78 generated from @xmath79 .
let @xmath80 denote the posterior distribution under @xmath15 using the training sample .
we use @xmath81 in two instances .
first , we replace the original prior @xmath82 by @xmath83 and , second , we also use it as proposal distribution @xmath84 .
the test data is then used to evaluate the acceptance probability .
the strategy can be characterized as model comparison by fractional bayes factors [ @xcite ] and is related to a similar approach proposed in @xcite for model comparison with intrinsic bayes factors .
both are originally proposed for model comparison with noninformative priors , but can be modified to facilitate mcmc across models as we need it here . [ cols="^,^ " , ]
tumors are heterogeneous tissues .
the traditional way to identify this heterogeneity has been to sequence multiple samples from the tumor .
using such data to study the coexistence of genetically different subpopulations across tumors and within a tumor can shed light on cancer development .
identifying subpopulations within a tumor can lead to clinically important insights .
for example , @xcite found that a chemotherapy affects subclonal heterogeneity in chronic lymphocytic leukemia .
they also observed that the presence of a certain subpopulation may adversely affect clinical outcome .
we have proposed a model - based approach based on a feature allocation model .
the feature allocation model allows us to impute inference about different components of tumor tissues based on ngs data .
the identified components are not necessarily unique because there might be other possible solutions which can lead to the same hypothetical mutation frequencies . instead of reporting a single solution
, the proposed approach provides a full probabilistic description of all possible solutions as a coherent posterior probability model over @xmath15 , @xmath19 and @xmath63 .
a number of extensions are possible for the present model .
first , the number of snvs examined in this paper was relatively limited ( about 100 ) , although the total number of snvs that were found in the whole exome of a tissue is on the order of about 50,000 .
other than computational complexity , there is no bar in principle on expanding the model to analyze the whole snv complement of the exome
. it could also be instructive to quantify the cellular diversity of the tumor based on findings from various regions of the exome .
another important extension of the model is in the basic representation of subclones and haplotypes .
the current model uses a binary matrix to record whether a variant sequence for an snv is present or absent in a haplotype .
a variation of the model could instead record for each subclone whether an snv is absent ( @xmath6 ) , heterozygous ( @xmath5 ) or homozygous ( @xmath85 ) .
that is , @xmath19 would become a trinary matrix .
other extensions of the model are to consider each snv position to have four possible bases , @xmath86 , to introduce dependence among mutations or to formally model the noise in variant calling .
each of these extensions is currently in development .
for example , incorporating explicit error probabilities in variant calls is possible .
similar to our previous work [ @xcite ] , we could replace the binomial likelihood ( [ eq : like ] ) in the proposed model with a bernoulli likelihood , for each read , where the probability associated with a read depends on quality scores of base calling and read mapping .
we will consider this extension as future work .
tumor genome sequencing projects have typically looked for specific genes to be mutated or not .
the inherent assumption here , so far unproven , is that the overall effect of carcinogenesis could be explained by a handful of changes in a small number of genes .
our model takes the opposite approach and allows us to examine the whole genome ( or exome ) and , by considering vaf patterns , to construct reasonable models for the tissue .
we believe this holistic approach to the analysis might provide more robust conclusions and biomarkers than the gene - by - gene approach .
church , d. m. , schneider , v. a. , graves , t. , auger , k. , cunningham , f. , bouk , n. , chen , h .- c . ,
agarwala , r. , mclaren , w. m. , ritchie , g. r. , _ et al . _
( 2011 ) . modernizing reference genome assemblies .
_ plos biology _ * 9 * , 7 , e1001091
. gerlinger , m. , rowan , a. , horswell , s. , larkin , j. , endesfelder , d. , gronroos , e. , martinez , p. , matthews , n. , stewart , a. , tarpey , p. , varela , i. , phillimore , b. , begum , s. , mcdonald , n. , butler , a. , jones , d. , raine , k. , latimer , c. , santos , c. , nohadani , m. , eklund , a. , spencer - dene , b. , clark , g. , pickering , l. , stamp , g. , gore , m. , szallasi , z. , downward , j. , futreal , p. , and swanton , c. ( 2012 ) .
intratumor heterogeneity and branched evolution revealed by multiregion sequencing . _
n engl j med .
_ * 366 * , 10 , 883892 .
ji , y. , xu , y. , zhang , q. , tsui , k .- w . , yuan , y. , norris jr , c. , liang , s. , and liang , h. ( 2011 ) .
bm - map : bayesian mapping of multireads for next - generation sequencing data .
_ biometrics _ * 67 * , 4 , 12151224 .
kanehisa , m. , goto , s. , furumichi , m. , tanabe , m. , and hirakawa , m. ( 2010 ) . for representation and analysis of molecular networks involving diseases and drugs .
_ nucleic acids research _ * 38 * , suppl 1 , d355d360 .
landau , d. , carter , s. , stojanov , p. , mckenna , a. , stevenson , k. , lawrence , m. , sougnez , c. , stewart , c. , sivachenko , a. , wang , l. , wan , y. , zhang , w. , shukla , s. , vartanov , a. , fernandes , s. , saksena , g. , cibulskis , k. , tesar , b. , gabriel , s. , hacohen , n. , meyerson , m. , lander , e. , neuberg , d. , brown , j. , getz , g. , and wu , c. ( 2013 ) . evolution and impact of subclonal mutations in chronic lymphocytic leukemia
. _ cell _ * 152 * , 4 , 714726 .
li , h. , handsaker , b. , wysoker , a. , fennell , t. , ruan , j. , homer , n. , marth , g. , abecasis , g. , durbin , r. , _ et al . _
the sequence alignment / map format and samtools .
_ bioinformatics _ * 25 * , 16 , 20782079 .
mckenna , a. , hanna , m. , banks , e. , sivachenko , a. , cibulskis , k. , kernytsky , a. , garimella , k. , altshuler , d. , gabriel , s. , daly , m. , _
et al . _
the genome analysis toolkit : a mapreduce framework for analyzing next - generation dna sequencing data .
_ genome research _ * 20 * , 9 , 12971303 .
navin , n. , krasnitz , a. , rodgers , l. , cook , k. , meth , j. , kendall , j. , riggs , m. , eberling , y. , troge , j. , grubor , v. , _
et al . _
( 2010 ) . inferring tumor progression from genomic heterogeneity .
_ genome research _ * 20 * , 1 , 6880 .
roth , a. , khattra , j. , yap , d. , wan , a. , laks , e. , biele , j. , ha , g. , aparicio , s. , bouchard - ct , a. , and shah , s. p. ( 2014 ) .
pyclone : statistical inference of clonal population structure in cancer .
_ nature methods _ .
russnes , h. g. , navin , n. , hicks , j. , and borresen - dale , a .-
insight into the heterogeneity of breast cancer through next - generation sequencing . _ the journal of clinical investigation _ * 121 * , 10 , 38103818 .
su , x. , zhang , l. , zhang , j. , meric - bernstam , f. , and weinstein , j. n. ( 2012 ) .
purityest : estimating purity of human tumor samples using next - generation sequencing data .
_ bioinformatics _ * 28 * , 17 , 22652266 .
teh , y. w. , grr , d. , and ghahramani , z. ( 2007 ) . stick - breaking construction for the indian buffet process . in _ proceedings of the international conference on artificial intelligence and statistics
11 .
wheeler , d. a. , srinivasan , m. , egholm , m. , shen , y. , chen , l. , mcguire , a. , he , w. , chen , y .- j . ,
makhijani , v. , roth , g. t. , _
et al . _
. the complete genome of an individual by massively parallel dna sequencing .
_ nature _ * 452 * , 7189 , 872876 . | we develop a feature allocation model for inference on genetic tumor variation using next - generation sequencing data . specifically , we record single nucleotide variants ( snvs ) based on short reads mapped to human reference genome and characterize tumor heterogeneity by latent haplotypes defined as a scaffold of snvs on the same homologous genome . for multiple samples from a single tumor , assuming that each sample is composed of some sample - specific proportions of these haplotypes , we then fit the observed variant allele fractions of snvs for each sample and estimate the proportions of haplotypes . varying proportions of haplotypes across samples is evidence of tumor heterogeneity since it implies varying composition of cell subpopulations .
taking a bayesian perspective , we proceed with a prior probability model for all relevant unknown quantities , including , in particular , a prior probability model on the binary indicators that characterize the latent haplotypes .
such prior models are known as feature allocation models .
specifically , we define a simplified version of the indian buffet process , one of the most traditional feature allocation models .
the proposed model allows overlapping clustering of snvs in defining latent haplotypes , which reflects the evolutionary process of subclonal expansion in tumor samples .
./style / arxiv - general.cfg , , |
although pbtase@xmath0 was discovered several decades ago , @xcite its electronic structure and physical properties have only been studied in detail over the past few years .
@xcite structurally , pbtase@xmath0 can be thought of as alternating stacking of hexagonal tase@xmath0 and pb layers with the @xmath9 space group .
the crystal structure of pbtase@xmath0 is non - centrosymmetric . @xcite initially , only the values of resistivity and hall coefficient at 300 k of a pressed powder pellet of pbtase@xmath0 and resistive onset of superconductivity at @xmath10 k were reported .
@xcite recently pbtase@xmath0 was identified as a topological , nodal semimetal with strong spin - orbit coupling .
@xcite its superconducting transition temperature in polycrystalline samples @xcite and in single crystals @xcite based on thermodynamic and transport measurements was established to be @xmath11 k. thermal conductivity and london penetration depth measurements @xcite suggested a nodeless superconducting gap structure for this material .
low temperature magnetoresistance was found to be relatively high , anisotropic , and sublinear in magnetic field ; the anisotropy of the upper critical field , @xmath12 , from the resistivity measurements was reported to be temperature dependent , with the values of @xmath13 .
@xcite the complex , non - trivial band structure of pbtase@xmath0 @xcite suggests possible sensitivity of its physical properties to applied pressure .
indeed , a non - monotonic , v - shaped , pressure dependence of @xmath8 was observed in a polycrystalline sample of pbtase@xmath0 that had a relatively broad resistive superconducting transition .
@xcite this behavior of @xmath14 was suggested to result from a lifshitz transition under pressure .
non - monotonic behavior of @xmath8 as a function of pressure has been observed in a number of materials , including elements .
@xcite a lifshitz transition ( a change of the fermi surface topology ) @xcite has been invoked to explain such evolution of @xmath8 with pressure .
@xcite recently , for complex superconductors , other possible causes for non - monotonic behavior of @xmath8 under pressure were discussed and studied , e.g. crossing long range magnetic order phase lines in @xmath15 phase diagrams @xcite or pressure - induced changes in superconducting pairing symmetry or gap structure .
@xcite to clarify the intrinsic @xmath14 behavior of pbtase@xmath0 under pressure and to better address the physics associated with this behavior , in this work we perform measurements of in - plane resistivity of high quality pbtase@xmath0 single crystals under pressures up to @xmath16 gpa in zero and applied magnetic field .
in addition to @xmath14 , and pressure dependence of the normal state resistivity , these measurements allow us to follow the evolution of the upper critical field , @xmath17 near the @xmath18 ( that was instrumental in e.g. studies of kfe@xmath0as@xmath0 and fese @xcite ) and of the low temperature , normal state magnetoresistance .
additionally , motivated by the low temperature results discussed below , ambient pressure x - ray diffraction and transmission electron microscopy measurements at elevated temperatures were performed .
experimental studies were complimented by the first - principles calculations of the stability of different crystallographic phases under pressure .
pbtase@xmath0 single crystals were grown by chemical vapor transport , using pbcl@xmath0 as a transport agent .
more details about the synthesis are presented in ref . .
the quality of the samples was attested by rather high residual resistivity ratios , @xmath19 and sharp superconducting transitions in zero field .
the electrical contacts for standard 4-probe ac resistivity measurements were made using pt wires and combination of du pont 4929n silver paste and epo - tek h20e silver epoxy .
the current was flowing in the @xmath20 plane and magnetic field was applied along the @xmath21-axis .
the resistivity measurements at ambient and high pressure were performed in a quantum design physical property measurement system ( ppms-9 ) . for resistivity measurements under pressure a be - cu / ni - cr - al hybrid piston cylinder pressure cell similar to that used in ref . was used .
a 40 : 60 mixture of light mineral oil and n - pentane was used as a pressure medium .
this medium solidifies at room temperature at @xmath22 gpa @xcite , well above the pressure range used in this work .
the pressure at room temperature was evaluated using manganin resistive gauge , whereas at low temperatures the superconducting transition of pure pb @xcite was used to determine pressure .
additionally , low - field dc magnetization under pressure down to 1.8 k was measured in a quantum design magnetic property measurement system ( mpms-5 ) squid magnetometer using a commercial , hmd , be - cu piston - cylinder pressure cell @xcite . in these measurements
daphne oil 7373 was used as a pressure medium and superconducting pb as a low - temperature pressure gauge @xcite . for magnetization measurements
a stack of single crystals was oriented with @xmath23 .
an additional set of resistivity measurements was performed under he - gas pressure .
four - point ac electrical resistivity measurements were carried out simultaneously on two pbtase@xmath0 crystals with approximate dimensions [email protected]@xmath240.05 mm@xmath25 to hydrostatic pressures as high as 0.37 gpa in a he - gas high - pressure system .
an excitation current of 1 ma ( rms ) at 17 hz was applied using a keithley 6221 constant ac / dc current source and the small voltage detected by a stanford research sr830 digital lock - in amplifier .
experiments were carried out both at constant temperature with varying pressures or at constant pressure with varying temperatures .
a janis supervaritemp cryostat was used to reach temperatures below the ambient . to generate hydrostatic pressure
the samples were placed in the 7 mm diameter bore of a be - cu high - pressure cell ( unipress , warsaw ) connected to a three - stage harwood compressor system .
he gas from the compressor is fed into the pressure cell via a 3 mm od / 0.3 mm i d be - cu capillary tube .
a calibrated digital manganin gauge at ambient temperature ( harwood model dj-320/42 ) accurately determines the pressure .
a sizable dead volume " at ambient temperature reduces the decrease of he - gas pressure on cooling from ambient to low temperatures .
high temperature x - ray diffraction ( xrd ) was performed using a panalytical xpert pro xrd system with an anton paar htk-1200n furnace under flowing helium after evacuating and backfilling the system with he .
larger single crystals were ground to a few tens of @xmath26 m but retained their micaceous morphology hence the xrd only exhibited ( 00l ) reflections . copper and cobalt radiation was used .
heating rates were either 3 k / min or 5 k / min .
since the xrd was not able to resolve changes of the _ in - plane _ lattice with heating , additional studies using transmission electron microscopy ( tem ) at different temperatures using a fei tecnai g2 f200 instrument operating at 200 kv were performed . for these studies
a single crystal was thinned via ar ion milling .
the thin region was co - planar with the basal planes , providing an orthogonal view of the lattice expansion on heating compared to the xrd results , i.e. , the ( hk0 ) reflections .
the in situ heating / cooling was performed on a gatan heating stage up to @xmath27 k. continuous recording of the selected area diffraction was obtained on cooling . the first - principles density functional theory ( dft )
@xcite calculations were performed using the vienna ab - initio simulation package ( vasp ) @xcite with projector - augmented wave ( paw ) pseudopotential method @xcite and plane wave basis .
the generalized - gradient approximation parameterized for solids ( pbesol ) @xcite was used for the exchange - correlation energy functional .
spin - orbit coupling was included in the calculations .
the energy cutoff is 320 ev and the monkhorst - pack s scheme @xcite was used for brillouin zone sampling with a @xmath28-point mesh of @xmath29 for the ground state @xmath9 structure and equivalent @xmath28-point meshes for other structures .
all crystal structures were fully relaxed until the forces on each atom were smaller than 0.01 ev / and external pressure was smaller than 0.1 gpa .
ambient pressure resistivity data ( fig . [ f1]a ) are grossly consistent with those reported in refs . .
although there is a region of @xmath30 behavior in resistivity , it does not persist to the temperatures close to the superconducting transition , where the behavior with temperature to higher power ( close to 4 ) was observed ( fig .
[ f1]b ) . the upper critical field , @xmath31 , was determined from electrical transport and magnetization measurements ( see appendix a for details ) . here , as well as in the literature , @xcite
there is an apparent discrepancy between the results of thermodynamic and transport measurements . turning to superconducting properties under pressure ,
the evolution of the superconducting transition under pressure , as measured by resistivity and low field magnetization , is presented in fig .
[ f3 ] . in the resistivity
measurements the transitions at all pressures , other than 0.24 gpa , are sharp . from both measurements
it appears that the @xmath14 has a step - like behavior . indeed , as it is seen in fig .
[ f4 ] , both measurements result in consistent data , and a clear , sharp step in @xmath14 is observed at about 0.25 gpa .
the broad , two - step - like resistive transition at 0.24 gpa corresponds to this apparent phase boundary .
normal state resistivity at 5 k , just like @xmath8 , has a step - like behavior ( fig .
the initial slope of transport @xmath31 ( see appendix b ) has a step - like change at @xmath32 gpa as well .
furthermore , the normal state properties show discontinuities at @xmath33 gpa . the field dependent magnetoresistance data ( fig .
[ f7 ] ) clearly fall on two different manifolds , for @xmath34 gpa and @xmath35 gpa , that have different functional dependencies of @xmath36 .
a careful look at zero field , temperature - dependent resistivity data taken in the piston - cylinder cell at small , from 0.05 to 0.39 gpa at room temperature , pressures ( fig .
[ f9 ] ) reveals a clear , hysteretic in temperature , sharp transition .
although qualitatively these data point to an additional phase line , probably of the structural transition , that has a very steep pressure dependence , a quantitative analysis of these data is hindered by an experimental issue associated with a use of piston - cylinder pressure cells over an extended temperature range .
due to differential thermal contraction of the materials of the cell and the medium , the pressure inside the cell decreases on cooling @xcite .
this pressure drop depends on multiple factors , including the cell materials and design , medium and the pressure range . at low pressures ,
even when using manganin as a room temperature pressure gauge and pb as a low temperature pressure gauge , the evaluation of pressure at intermediate temperatures has substantial error bars . to address the pressure dependence of the apparent structural transition in a quantitative manner , a set of measurements in a he gas system
was performed ( fig .
[ f10 ] ) .
the hysteretic nature of the transition is seen both in temperature sweeps and pressure sweeps .
these signatures are sharp and well defined . from the pressure sweeps ( fig . [ f10](b ) ) it is seen that ( a ) at lower temperatures the transition shifts to higher pressures ; ( b ) as has been seen in figure [ f9 ] , the size of the resistance jump becomes smaller at lower temperatures and it appears not to be detected anymore in the pressure sweep at @xmath37 k. fig .
[ f11 ] shows that both absolute and relative resistance jump decrease at lower temperatures , and it is not clear if at 50 k the jump disappears , or just becomes on the level of the noise in the data .
the pressure dependence of the apparent structural transition measured in gas pressure system is presented in fig .
the second order polynomial fits to the @xmath38 data extrapolate to 0.25 gpa for the decreasing temperature manifold to intercept the @xmath39 k axis . in a similar manner ,
both manifolds extrapolate to @xmath2 k at ambient pressure .
this is a rather rough extrapolation , additional , preferably structural , data at high temperatures are required to verify the nature of the transition . in situ xrd
was performed at different elevated temperatures . as mentioned above , the pbtase@xmath0 powder retained the micaceous morphology , so only ( 00l ) reflections could be detected and followed as a function of temperature .
the experimental results are shown in fig .
the ( 00l ) reflections show a clear , step - like , contraction in the @xmath21-axis lattice on heating which is reversible upon cooling .
the temperature of this transition is @xmath40 k. to resolve changes of the in - plane lattice with heating , tem measurements were performed ( fig .
[ f16 ] ) .
the thin region studied with tem was co - planar with the basal planes , providing an orthogonal view of the lattice expansion on heating compared to the xrd results , i.e. , the ( hk0 ) reflections .
continuous recording of the selected area diffraction was obtained on cooling .
the sample was slowly cooled over the temperature range of @xmath41 k to @xmath42 k at @xmath43 k / min .
an abrupt contraction of the basal plane lattice of @xmath44 occurred @xmath2 k. no other changes in the diffraction pattern were observed , indicating that like in the xrd there is a discrete change in the cell parameters occurring at @xmath2 to 430 k but no obvious change in the space group ( fig .
[ f16 ] , lower panels ) .
tem results suggest that the basal plane undergoes a normal expansion with heating and contraction on cooling in contrast to the xrd results which shows a large contraction in the @xmath21-axis at @xmath2 k. aberration corrected scanning transmission electron microscopy using a fei titan themis 300 cubed 300 stem / tem shows that the room temperature atomic decoration is fully consistent with the space group # 187 , @xmath9 , ( fig .
[ f17 ] ) .
the image suggest minimal chemical disorder .
there is a rather large gap between the pb layers relative to the ta / se inner layers which form an open network of edge sharing prisms .
this large gap between these layers maybe responsible for the lattice contraction in the @xmath21-axis with heating . altogether on heating through the structural transition at @xmath45 k the @xmath46-axis increases by @xmath47 , whereas the @xmath21-axis decreases by @xmath48 , leading to a decrease of the unit cell volume by @xmath49 .
based on the data discussed above , the pressure - temperature the phase diagram for pbtase@xmath0 ( fig .
[ f13 ] ) appears to have two superconducting phases with the boundary between them defined by a structural phase transition that has extremely steep pressure dependence . since the normal state resistivity , magnetoresistance and the initial slopes of @xmath50 are different in these two phases ,
clearly , the electronic structure is affected .
a bloch - grneisen fit of resistivity ( over the 50 k - 300 k temperature range ) ( fig .
[ f14 ] ) .
suggests that as a result of the structural phase transition the debye temperature increases , so that the lattice becomes stiffer .
this is largely consistent with an overall decrease of the unit cell volume at ambient pressure on heating through the transition . despite this
, @xmath8 is lower in the new structural phase , suggesting that either the change in the electronic subsystem is the dominant contribution to the @xmath8 decrease , or that the in - plane phonons are more important for superconductivity than the out - of - plane phonons .
the latter hypothesis is in agreement with the observed large gaps between the layers ( fig .
[ f17 ] ) . since available experimental techniques
do not allow for an unambiguous identification of the high pressure ( high temperature ) phase , we have performed first - principles calculations that address relative stability of several related hexagonal phases under pressure .
in addition to the known @xmath9 structure of pbtase@xmath0 , we consider 3 other low - energy structures ( within 50.0 mev / atom with respect to that of the @xmath9 structure ) coming from our crystal structure optimization scheme with one and two formula units and hexagonal symmetry constraints ( fig . [ f18 ] ) .
these structures can also be obtained through modification of the @xmath9 structure .
the @xmath51 structure can be obtained from the @xmath9 structure by moving the pb atom from the @xmath3-wyckoff position to the @xmath52-wyckoff position .
similarly the @xmath53 structure can be obtained by shifting the pb from the @xmath3-wyckoff position to the @xmath4-wyckoff position . by doubling unit cell of the @xmath9 structure along the @xmath21-axis lattice vector and
then moving the upper half of unit cell by 1/3 along the long diagonal of the basal plane @xmath54 , the @xmath55 structure can be obtained .
the @xmath51 and @xmath53 structures are in @xmath9 space group symmetry whereas the @xmath55 structure is in @xmath56 space group symmetry .
the lattice parameters and wyckoff positions of these structures are given in table [ t1 ] .
the lateral lattice constants of these 3 structures are similar to that of the @xmath9 structure , but their lattice constants along the @xmath21 direction ( @xmath57 for the @xmath55 structure ) are smaller than that of the @xmath9 structure by 0.534 , 0.487 , and 0.246 respectively , where the @xmath51 structure has the shortest lattice parameter @xmath21 . in order to compare with experimental xrd results , we simulated the xrd ( 0 0 l)-peaks of all four structures considered .
it is interesting to note that although the @xmath55 structure has a doubled unit cell along the @xmath21-direction in comparison with other structures , it shows only ( 0 0 l)-peaks with even - l . as shown in fig .
[ f19 ] , the xrd ( 0 0 l)-peaks versus @xmath58-spacing patterns from all four structures are very similar . in figure [ f19 ]
the peak indexes of the @xmath55 structure are ( 0 0 2l ) of the labeled ones and the diffraction intensities of all structures are normalized by the intensity of their ( 0 0 2)-peak .
figure [ f20 ] shows the relative formation enthalpies of different pbtase@xmath0 structures as a function of pressure with respect to that of the @xmath9 structure . at zero and low pressure ,
the stable structure is the @xmath9 structure .
as the pressure is increased , the @xmath53 structure becomes more stable with a structural transition from the @xmath9 to the @xmath53 at 3 gpa . at ambient conditions ,
the formation energy of the @xmath53 structure is 25.9 mev / atom higher than that of the @xmath9 ground state structure .
this energy difference is about the room temperature thermal energy .
we note that the @xmath55 structure is lower in energy than the @xmath53 structure at ambient conditions .
however , the @xmath53 structure becomes more stable than the @xmath55 structure when the external pressure is greater than 2.7 gpa , which is just below the transition pressure from the @xmath9 structure to the @xmath53 structure .
thus these calculations suggest that the experimentally observed high pressure ( high temperature ) structure is the @xmath53 structure .
we would like to note that structural transition pressure from the dft calculations is higher than that observed in experiment .
this discrepancy is probably due to the systematic error in dft calculation of pressure .
for example , in the literature the predicted structural transition pressure of si from cubic diamond ( si - i ) to - sn ( si - ii ) phase can be several gpa off the experimentally observed value , depending on exchange - correlation functional used.@xcite but the sequence of phase transition , i.e. si - i to si - ii to si - v to si - vi to si - vii , from dft calculation is consistent with experiment .
@xcite + _ to summarize _ , pbtase@xmath0 exhibits a structural , sharp , first order , phase transition at a very moderate pressure of @xmath1 gpa at low temperatures .
pbtase@xmath0 has a @xmath59 k / gpa making it one of the more pressure sensitive transitions found in inorganic compounds .
the structural phase transition line extends to @xmath2 k at ambient pressure as evidenced by transmission electron microscopy and x - ray diffraction at elevated temperatures . upon transition to the new phase ( on increase of temperature at ambient pressure )
the @xmath21-axis decreases and the @xmath46-axis increases , resulting in a slight , @xmath49 decrease of the unit cell volume .
the new , high temperature / high pressure phase has similar crystal structure . as suggested by the first - principles calculations , in this structure pb shifts from @xmath3 to @xmath4 wyckoff position with ta and se positions remaining the same .
the superconductivity appears to be robust , it persists through the structural phase transition into the other phase with slight , step - like decrease of @xmath8 .
ambient pressure , low temperature resistivity , @xmath60 , measured in constant fields and @xmath61 , measured at constant temperatures , are shown in figs . [ f2]a ,
b . two criteria , onset and offset ( see fig .
[ f2]b ) were used to evaluate the @xmath12 .
additionally , zero - field - cooled magnetization was measured in different applied fields ( fig . [ f2]c ) and @xmath31 was determined from these measurements as well .
the summary of these results , the ambient pressure @xmath31 for @xmath23 inferred from electrical transport and magnetization measurements are shown in fig .
[ f2]d and are compared to the literature data @xcite .
the resistively determined @xmath31 is comparable with the literature data ( with possible contributions to the difference from the sample shape and slight mis - orientation ) .
similarly , results from magnetization are very close to the published upper critical field determined from specific heat .
however , both in the literature @xcite and in this work , even though the low field / zero field @xmath8 values are quite similar , @xmath31 from electrical transport [ @xmath62 and thermodynamic [ @xmath63 measurements are noticeably different .
the origin of this difference is not clear at this point .
the temperature - dependent upper critical field for @xmath23 was measured resistively as a function of pressure ( fig .
[ f5 ] ) . whereas there is an apparent difference between thermodynamic and transport @xmath12 values ( see above ) , as well as an upward curvature of @xmath31 and the limited range of the data for the higher pressures
, we can still compare the change of the initial , close to @xmath64 , slope of transport @xmath12 as a function of pressure ( fig .
[ f6 ] ) . in agreement with the @xmath14 behavior ,
both bare , @xmath65 , and normalized , @xmath66 , initial slopes of the upper critical field have step - like change at @xmath32 gpa .
we thank the ms . junying shen for assistance in the he - gas measurements .
work at the ames laboratory was supported by the u.s .
department of energy , office of science , basic energy sciences , materials sciences and engineering division .
the ames laboratory is operated for the u.s .
department of energy by iowa state university under contract no .
de - ac02 - 07ch11358 , including a grant of computer time at the national energy research scientific computing centre ( nersc ) in berkeley , ca .
the work at washington university was supported by the national science foundation ( nsf ) through grants no .
dmr-1104742 and 1505345 .
guang bian , tay - rong chang , raman sankar , su - yang xu , hao zheng , titus neupert , ching - kai chiu , shin - ming huang , guoqing chang , ilya belopolski , daniel s. sanchez , madhab neupane , nasser alidoust , chang liu , baokai wang , chi - cheng lee , horng - tay jeng , chenglong zhang , zhujun yuan , shuang jia , arun bansil , fangcheng chou , hsin lin , and m. zahid hasan , nat .
* 7 * , 10556 ( 2016 ) .
e. hassinger , g. gredat , f. valade , s. ren de cotret , a. juneau - fecteau , j .- ph .
reid , h. kim , m. a. tanatar , r. prozorov , b. shen , h .- h .
wen , n. doiron - leyraud , and louis taillefer , phys .
b * 86 * , 140502 ( 2012 ) .
valentin taufour , neda foroozani , makariy a. tanatar , jinhyuk lim , udhara kaluarachchi , stella k. kim , yong liu , thomas a. lograsso , vladimir g. kogan , ruslan prozorov , sergey l. budko , james s. schilling , and paul c. canfield , phys . rev .
b * 89 * , 220509 ( 2014 ) . .
inset : low temperature part of the data showing the superconducting transition .
( b ) low temperature resistivity of pbtase@xmath0 at ambient pressure plotted as @xmath67 vs @xmath68 .
the dashed line is a guide for the eye .
inset : _ log - log _ plot of the low temperature resistivity after subtraction of the residual resistivity @xmath69 .
lines show @xmath68 and @xmath70 behavior .
, width=453 ] correspond to the pressure runs that have slightly different room temperature pressure values but result in the same @xmath71 low temperature value as measured by pb manometer . for magnetization measurements a deviation from linear , normal state magnetization ( indicated by circle )
was taken as a @xmath8 criterion , see inset to the panel ( b ) .
, title="fig:",width=377 ] correspond to the pressure runs that have slightly different room temperature pressure values but result in the same @xmath71 low temperature value as measured by pb manometer . for magnetization measurements a deviation from linear , normal state magnetization ( indicated by circle ) was taken as a @xmath8 criterion , see inset to the panel ( b ) .
, title="fig:",width=377 ] as determined from resistivity and magnetization measurements ( left y - axis ) , and normal state resistivity at 5 k ( right y - axis ) as a function of pressure .
dashed vertical line shows @xmath72 gpa location .
other dashed lines are guides for the eye.,width=453 ] k , field dependent magnetoresistance at different pressures ( listed in the upper panel in units of gpa ) , plotted as @xmath61 and @xmath73 .
note : for both panels the four p = 0 data sets , as well as the @xmath74 gpa data set form an upper manifold , and the @xmath35 gpa data sets form a lower manifold.,width=453 ] data taken in applied magnetic fields from @xmath77 to @xmath78 koe with 0.1 koe steps .
( b ) field dependent resistivity in fixed magnetic fields .
onset and offset criteria are shown for 3 k dataset .
( c ) low temperature zero - field - cooled magnetization data taken in different magnetic fields .
arrows mark the superconducting transition temperatures .
( d ) zero pressure @xmath31 data for @xmath23 obtained from resistivity ( onset and offset criteria ) - red circles @xmath61 , black circles @xmath60 , and magnetic susceptibility data - stars , plotted together with the literature data @xcite from resistivity ( triangles ) and heat capacity ( rhombi ) measurements in magnetic field.,title="fig:",width=302 ] data taken in applied magnetic fields from @xmath77 to @xmath78 koe with 0.1 koe steps .
( b ) field dependent resistivity in fixed magnetic fields .
onset and offset criteria are shown for 3 k dataset .
( c ) low temperature zero - field - cooled magnetization data taken in different magnetic fields . arrows mark the superconducting transition temperatures .
( d ) zero pressure @xmath31 data for @xmath23 obtained from resistivity ( onset and offset criteria ) - red circles @xmath61 , black circles @xmath60 , and magnetic susceptibility data - stars , plotted together with the literature data @xcite from resistivity ( triangles ) and heat capacity ( rhombi ) measurements in magnetic field.,title="fig:",width=302 ] data taken in applied magnetic fields from @xmath77 to @xmath78 koe with 0.1 koe steps .
( b ) field dependent resistivity in fixed magnetic fields .
onset and offset criteria are shown for 3 k dataset .
( c ) low temperature zero - field - cooled magnetization data taken in different magnetic fields . arrows mark the superconducting transition temperatures .
( d ) zero pressure @xmath31 data for @xmath23 obtained from resistivity ( onset and offset criteria ) - red circles @xmath61 , black circles @xmath60 , and magnetic susceptibility data - stars , plotted together with the literature data @xcite from resistivity ( triangles ) and heat capacity ( rhombi ) measurements in magnetic field.,title="fig:",width=302 ] data taken in applied magnetic fields from @xmath77 to @xmath78 koe with 0.1 koe steps .
( b ) field dependent resistivity in fixed magnetic fields .
onset and offset criteria are shown for 3 k dataset .
( c ) low temperature zero - field - cooled magnetization data taken in different magnetic fields .
arrows mark the superconducting transition temperatures .
( d ) zero pressure @xmath31 data for @xmath23 obtained from resistivity ( onset and offset criteria ) - red circles @xmath61 , black circles @xmath60 , and magnetic susceptibility data - stars , plotted together with the literature data @xcite from resistivity ( triangles ) and heat capacity ( rhombi ) measurements in magnetic field.,title="fig:",width=302 ] | transport and magnetic studies of pbtase@xmath0 under pressure suggest existence of two superconducting phases with the low temperature phase boundary at @xmath1 gpa that is defined by a very sharp , first order , phase transition . the first order phase transition line can be followed via pressure dependent resistivity measurements , and is found to be near 0.12 gpa near room temperature .
transmission electron microscopy and x - ray diffraction at elevated temperatures confirm that this first order phase transition is structural and occurs at ambient pressure near @xmath2 k. the new , high temperature / high pressure phase has a similar crystal structure and slightly lower unit cell volume relative to the ambient pressure , room temperature structure .
based on first - principles calculations this structure is suggested to be obtained by shifting the pb atoms from the @xmath3 to @xmath4 wyckoff position without changing the positions of ta and se atoms .
pbtase@xmath0 has an exceptionally pressure sensitive , structural phase transition with @xmath5 k / gpa near 4 k , this first order transition causes an @xmath6 k ( @xmath7 ) step - like decrease in @xmath8 as pressure is increased through 0.25 gpa . |
the properties of quark matter at extreme conditions have been an active research area both theoretically and experimentally . at high temperature , the quark - gluon plasma(qgp )
has long been searched by colliding two nuclei at sufficiently high energy . on the other side
, we expect that quark matter becomes color superconducting through a phase transition at high baryon density but low temperature@xcite , which is the typical condition inside compact stars . in a typical metallic superconductor , the electrons pair with equal chemical potential near the fermi surface .
the situation with a quark matter , however , is much more complicated . while the quark - quark interaction favors pairing between quarks of different flavors , the mass difference among @xmath1 , @xmath2 and @xmath3 together with the charge neutrality requirement induces a substantial mismatch among their fermi momenta at the baryon density inside a compact star and thereby reduces the available phase space for cooper pairing .
a number of exotic color superconductivity phases in the presence of mismatch have been proposed in the literature@xcite , but a consensus point of view of the true ground state has not been reached .
the single flavor pairing@xcite , which is free from the fermi momentum mismatch , is an interesting alternative in this circumstance and will be considered here .
since the quark - quark interaction is attractive in the color anti - triplet channel , the color wave function of the pair is anti - symmetric . for the equal helicity pairing to be considered in this article
, the parity of the orbital wave function has to be odd as required by the pauli principle . except for the color - spin - lock phase examined in @xcite ,
the energy gap will not be spherical .
the odd parity prevents the diquark wave function from realizing the full pairing potential .
the energy scale of the color superconductivity is therefore reduced
. at ultra - high baryon densities , asymptotic freedom of qcd ensures the validity of the weak coupling expansion , which has been carried out for csc by a number of authors@xcite .
the dominant pairing interaction is mediated by one - gluon exchange and can be decomposed into partial waves , as is shown in eq.([33 ] ) below .
quantitative results of the transition temperature in the equal helicity channel with an arbitrary angular momentum has been obtained from the first principle @xcite , and read @xmath4 \label{1}\ ] ] where @xmath5 is the number of flavor , @xmath6 is the number of colors , @xmath7 is the chemical potential , @xmath8 is the running coupling constant of qcd and @xmath9 is the euler constant .
the @xmath10-dependent constant @xmath11 the equal helicity pairing of odd parity picks up @xmath12 as the transition temperature .
we have @xmath13 .
natural analogy is drawn with the superfluidity of @xmath14 .
but important difference between the pairing potential in quark matter and that in @xmath14 has to be considered before ascertaining the angular dependence of the energy gap .
the forward singularity of the one - gluon exchange renders the pairing strength _
equal _ for _ all _ partial waves ( the same leading order term inside the bracket of ( [ 1 ] ) ) .
the paring potential in @xmath14 , however , is entirely in the channel of @xmath15 .
the transition temperature was determined from the pairing instability of the diquark scattering amplitude in the normal phase . in a perturbative treatment ,
the dyson - schwinger equation for the scattering amplitude is linear and the partial wave decomposition in eq.([1 ] ) is legitimate .
this is not the case with the gap equation below @xmath16 , which is nonlinear . a non - spherical solution of the gap equation , in general , will be a mixture of different partial waves unless the the pairing is dominated in one angular momentum channel .
the adjective `` non - spherical '' refers to the magnitude of the energy gap of the quasi - particle .
so the @xmath3-wave and the csl gaps are spherical and are therefore free from the angular momentum mixing .
the gap equations in these cases are linear with respect to the angular dependence .
a nonspherical gap function , @xmath17 , for the equal helicity pairing at @xmath18 contains all spherical harmonics of odd parity with the same azimuthal quantum number @xmath19 .
we have @xmath20 where @xmath21 is the gap function of 2sc in the absence of the mismatch , @xmath22 the direction of the relative momentum of the two quarks in a cooper pair and the angular dependent factor @xmath23 with @xmath10 the _ total _ angular momentum of the cooper pair .
carrying the formulation of @xmath14 over to qcd amounts to drop all higher multipoles except that of @xmath15 , which will not satisfy the gap equation of qcd .
it was argued in the literature that @xmath24 but @xmath25 for @xmath26 , this , as will be shown below , is not the case .
instead , we find that the function @xmath27 satisfies a nontrivial integral equation and thus @xmath28 for all odd @xmath10 s
. therefore the angular momentum mixing does occurs in the subleading order of the gap function .
the angular momentum mixing will modify all non - spherical `` spin-1 '' csc examined in the literature , we shall focus our attention in this paper to the equal helicity pairing with zero azimuthal quantum number , i.e. the analog of the polar phase of @xmath14 .
the subscript @xmath19 of @xmath17 and @xmath29 will be suppressed below . even though this phase is unstable , it is the simplest one to illustrate the mixing mechanism .
the current work is organized as follows . in the next section
, we shall clarify the concept of the angular momentum mixing with a toy model of non - relativistic and spinless fermions . in the sect .
iii the gap equation for the single flavor csc will be derived by minimizing the cjt free energy of qcd .
this gap equation will be reduced to an nonlinear integral equation for the function @xmath30 in the subsequent section and the numerical solution will be presented in the sect , v. we conclude the paper in the sect .
some technical details are deferred to the appendices .
our units are @xmath31 and 4-vectors are denoted by capital letters , @xmath32 with @xmath33 the matsubara energy , which becomes continuous at @xmath34 . throughout the article , we shall follow the definition of the leading order and the subleading order in @xcite . upon taking the logarithm of the transition temperature or the magnitude of the gap function
, the order @xmath35 will be referred to as the leading one and the @xmath36 term to the subleading one .
to clarify the concept of the angular momentum mixing , we consider a toy model of nonrelativistic and spinless fermions . in terms of the creation and annihilation operators ,
the model hamiltonian reads @xmath37 where @xmath38 with @xmath39 the mass and @xmath7 the chemical potential , @xmath40 is a coupling constant , @xmath41 is the normalization volume and the summation @xmath42 extends to states with @xmath43 and @xmath44 with @xmath45 a uv cutoff ( debye frequency for electronic superconductors ) .
the angular dependent form factor @xmath46 can be expanded in series of the legendre polynomials , @xmath47 introducing the order parameter @xmath48 with @xmath49 the ground state and expanding the interaction term of ( [ toy ] ) to the linear order of the fluctuation @xmath50 , we end up with the mean - field hamiltonian @xmath51,\ ] ] where we have introduced the gap function via @xmath52 we have @xmath53 and @xmath54 following from their definitions . upon a bogoliubov transformation , we find that @xmath55 with @xmath56 and the ground state energy @xmath57 with @xmath58 the condensation energy density given by @xmath59 .
\label{cond}\ ] ] substituting ( [ bogoliubov ] ) into ( [ gapdef ] ) , we obtain the gap equation @xmath60 in the weak coupling limit , @xmath61 and @xmath62 with @xmath63 the density of states on the fermi surface , but the magnitude of @xmath46 remains of order one .
we have @xmath64 also , the support of the gap function extends only to a narrow band around the fermi surface .
we may ignore the dependence of @xmath65 on the magnitude @xmath66 and switch the argument of @xmath65 from @xmath67 to @xmath22 . following ( [ weak ] ) , the integrations over @xmath0 in ( [ cond ] ) and ( [ gapeq ] ) can be carried out readily and we end up with @xmath68 and @xmath69 the gap equation ( [ gapeqwk ] ) is nonlinear because of the logarithm . in what follows , we consider two extremes of @xmath46 , each of which gives rise to an exact solution to the gap equation ( [ gapeqwk ] ) .
we present only the solution that is invariant under time reversal , i.e. the one with zero azimuthal quantum number .
_ case i _ : @xmath70 it corresponds to the partial wave expansion ( [ partial ] ) with @xmath71 and @xmath72 for @xmath26 .
the angular dependence of the pairing force in @xmath73 is of this type .
the gap equation reads @xmath74 and its solution of zero azimuthal quantum number is given by @xmath75 with @xmath76 the angle with respect to a prefixed direction in space and @xmath77 the condensation energy density @xmath78 this solution corresponds to the polar phase of @xmath73 @xcite .
since the gap function contains only the partial wave of @xmath15 , there is no angular momentum mixing .
the additional term in the exponent of ( [ polarhe ] ) , @xmath79 , comes from the logarithm of ( [ gapeqlog ] ) .
_ case ii _ : @xmath80 this corresponds to a singularity of the two body scattering amplitudes in the forward direction .
we have @xmath71 for all @xmath10 in ( [ partial ] ) .
the last step of ( [ toy2 ] ) follows from the addition theorem and the completeness of the spherical harmonics .
the gap equation ( [ gapeqwk ] ) becomes @xmath81 which implies a constant @xmath82 and yields a solution of odd parity and zero azimuthal quantum number .
@xmath83 the condensation energy density in this case reads @xmath84 we refer to this case as the case with the angular momentum mixing because the gap function ( [ toymixing ] ) contains all partial waves .
carrying the solution of the case i to the case ii amounts to drop all partial waves other than that of @xmath15 and would lead to a lower magnitude of the condensation energy ( [ condi ] ) .
the case with qcd is similar to the case ii above since the forward singularity of the diquark scattering renders the pairing strength of all partial waves equal to the leading order .
the running coupling constant @xmath8 of qcd corresponds to @xmath85 here and the angular momentum mixing shows up in the @xmath36 term of @xmath86 .
therefore we expect angular momentum mixing to the subleading order of the angular dependence of the gap function .
besides being an ultra relativistic system , the csc of qcd differs from the toy model considered above in two aspects .
the forward singularity of qcd also brings about the energy dependence of the gap , so the gap equation ( [ gapeq ] ) will be replaced by the eliashberg equation derived by minimizing the cjt effective action of qcd .
secondly , the pairing strength of each partial wave does fall off with an increasing @xmath10 in the sub - leading order of the pairing potential .
it is this falling off that makes the amount of the angular momentum mixing numerically small for the solution considered in this article .
the qcd lagrangian for one flavor of massless quark is given by @xmath87 where , @xmath88 is the quark spinor in dirac and color space and @xmath89 .
the covariant derivative acting on the fermion field is @xmath90 , where @xmath8 is the running coupling constant , @xmath91 is the gauge potential , @xmath92 is the @xmath93-th @xmath94 generator with @xmath95 the @xmath93-th gell - mann matrix .
@xmath96 is the field strength tensor . introducing the nambu - gorkov spinor @xmath97 where @xmath98 is the charge - conjugate spinor and @xmath99 , the cjt effective action reads@xcite @xmath100=-\frac{1}{2 } \{{\rm trln}d^{-1}+{\rm tr}(d_{0}^{-1}d-1)-{\rm trln}s^{-1}-{\rm tr}(s_{0}^{-1}s-1)-2\gamma_2[d , s]\ } \label{2}\ ] ] where @xmath101 and @xmath102 are the full gluon and quark propagators , @xmath103 and @xmath104 are the inverse tree - level propagators for gluons and quarks , respectively .
@xmath105 is the sum of all two - particle irreducible(2pi)vacuum diagrams built with @xmath101 , @xmath102 and the tree - level quark - gluon vertex @xmath106 .
we have @xmath107 where the first term corresponds to the sunset diagram of fig.[fig : sset ] and the contribution from ... is beyond the subleading order of the gap function@xcite . the sunset diagram of eq.([sun ] ) .
straight line and wavy line denotes quark and gluon propagators respectively . ]
the stationary points of the cjt effective action are determined by @xmath108 which gives rise to the dyson - schwinger equation for gluons and quarks , @xmath109 where @xmath110 and @xmath111 are the gluon and quark self - energy defined via @xmath112 and @xmath113 . instead of solving the two equations of ( [ 5 ] )
simultaneously , we shall reduce the cjt effective free energy with the aid of the first equation , leaving the gap function arbitrary . the gap equation ( which is the nambu - gorkov off diagonal part of the second equation of ( [ 5 ] ) ) will be derived after the cjt free energy is fully simplified under the weak coupling approximation . substituting the first equation of ( [ 5 ] ) into ( [ 2 ] ) , the second term in eq .
( [ 2 ] ) cancels the last term .
we have @xmath100=-\frac{1}{2}[{\rm trln}d^{-1}-{\rm trln}s^{-1}-{\rm tr}(s_{0}^{-1}s-1 ) ] \label{6}\ ] ] in nambu - gorkov space , the inverse free quark propagator is @xmath114^ { - 1 } } & 0 \\ 0 & { [ g_0^ - ] ^ { - 1 } } \end{array } \right ) \label{7}\ ] ] where @xmath115^{-1}=(p_0\pm\mu)\gamma_0-\vec{\gamma}\cdot\vec{p}. \label{8}\ ] ] on writing the quark self - energy @xmath116 the full quark propagator , @xmath117 can be obtained explicitly by inverting the matrix @xmath118 . for the single flavor pairing , the simplest choice of the off - diagonal block of eq .
( [ 9 ] ) reads @xmath119 and @xmath120(see theorem 2 in@xcite ) , where @xmath121 is the 2nd gell - mann matrix and @xmath65 is a function of the energy and the momentum , i.e. @xmath122 .
@xmath65 is even in @xmath123 and odd in @xmath67 . by using the energy projectors of massless fermions @xmath124 and ignoring the contribution from the wave - function renormalization , the ng blocks of the propagator ( [ 10 ] ) take the form @xmath125 @xmath126 because of the @xmath121 of ( [ 13 ] ) ,
the excitation in the third color direction is ungapped .
now , we proceed to simplify the cjt free energy under the weak coupling approximation .
denote by @xmath127 the free energy density of the normal phase , we have @xmath128 where the condensate energy density @xmath129\ ] ] is the part of @xmath130 responsible to the gap equation . following the procedure of @xcite
, we approximate @xmath131\simeq{\rm tr}[d_n\delta\pi ] \label{16}\ ] ] where @xmath132 with @xmath133 the hard - dense - loop ( hdl ) resummed gluon self - energy in normal phase and @xmath134 the corresponding hdl gluon propagator . in the coulomb gauge ,
the hdl gluon propagator is @xmath135 where @xmath136 are the longitudinal and transverse propagators respectively and are diagonal in adjoint color space , i.e. @xmath137 .
consequently , we only need the 00-component , @xmath138 , and the transverse projection of the ij - components , @xmath139 the gluon self - energy in super phase reads @xmath140 \label{20}\ ] ] where @xmath141 and @xmath142 with @xmath143 and @xmath144 . substituting eq .
( [ 10 ] ) into eq.([20 ] ) , we find that nambu - gorkov space , @xmath145 + { \rm tr}[\bar\gamma_{a}^{\mu}g^-(p)\bar\gamma_{b}^{\nu}g^-(p^{\prime})]\\ & & + { \rm tr}[\gamma_{a}^{\mu}\xi^-(p)\bar\gamma_{b}^{\nu}\xi^+(p^{\prime } ) ] + { \rm tr}[\bar\gamma_{a}^{\mu}\xi^+(p)\gamma_{b}^{\nu}\xi^-(p^{\prime})]\big\ } \label{22}\end{aligned}\ ] ] since the hdl gluon propagators are diagonal in color space , we only need the diagonal terms of eq.([22 ] ) to deal with eq.([16 ] ) .
the explicit form of each diagonal term of ( [ 22 ] ) reads @xmath146=\frac{g^2}{2}t\sum\limits_{p , p^{\prime } } { \rm tr}\big[\gamma^\mu\lambda_{p}^{+}\gamma_0\gamma^{\nu}\lambda_{p^{\prime}}^{+}\gamma_0 \big]w_a^+(p , p^{\prime}),\label{23a } \\ & & { \rm tr}[\bar\gamma_{a}^{\mu}g^-(p)\bar\gamma_{a}^{\nu}g^-(p^{\prime})]=\frac{g^2}{2}t\sum\limits_{p , p^{\prime } } { \rm tr}\big[\gamma^\mu\lambda_{p}^{-}\gamma_0\gamma^{\nu}\lambda_{p^{\prime}}^{-}\gamma_0 \big]w_a^-(p , p^{\prime}),\label{23b } \\ & & { \rm tr}[\gamma_{a}^{\mu}\xi^-(p)\bar\gamma_{a}^{\nu}\xi^+(p^{\prime})]=-\frac{g^2}{2}t\sum\limits_{p , p^{\prime } } { \rm tr}\big[\gamma^\mu\gamma_5\lambda_{p}^{+}\gamma^{\nu}\gamma_5\lambda_{p^{\prime}}^{- } \big]w_a(p , p^{\prime}),\label{23c } \\ & & { \rm tr}[\bar\gamma_{a}^{\mu}\xi^+(p)\gamma_{a}^{\nu}\xi^-(p^{\prime})]=-\frac{g^2}{2}t\sum\limits_{p , p^{\prime } } { \rm tr}\big[\gamma^\mu\gamma_5\lambda_{p}^{-}\gamma^{\nu}\gamma_5\lambda_{p^{\prime}}^{+ } \big]w_a(p , p^{\prime}).\label{23d}\end{aligned}\ ] ] where the repeated color indexes on lhs are _ not _ to be summed .
the quantities @xmath147 and @xmath148 on rhs of eqs.([23a]-[23d ] ) are given by @xmath149 , a=4,\cdot\cdot\cdot,7 \end{gathered}\\ \hfill\\ \begin{gathered } \frac{1}{6}\frac{p_0\pm\epsilon_p}{p_{0}^{2}-\varepsilon_{p}^{2}}\frac{p_0^{\prime}\pm\epsilon_{p^{\prime}}}{p_{0}^{\prime2}-\varepsilon_{p^{\prime}}^{2 } } + \frac{1}{3}\frac{p_0\pm\epsilon_p}{p_{0}^{2}-\epsilon_{p}^{2}}\frac{p_0^{\prime}\pm\epsilon_{p^{\prime}}}{p_{0}^{\prime2}-\epsilon_{p^{\prime}}^{2 } } \hspace{0.1cm},a=8 \end{gathered } \end{array } \right .
\label{24a}\ ] ] and @xmath150 where @xmath151 and @xmath152 .
since the dominant contributions in the weak coupling arise from the quasiparticles , we have ignored the contributions from the quasi - antiparticles in the calculations above .
the trace over dirac space is straightforward @xmath153 = -{\rm tr}\big[\gamma^0\gamma_5\lambda_{p}^{\pm}\gamma^{0}\gamma_5\lambda_{p^{\prime}}^{\mp}\big]=1+\hat{p}\cdot\hat{p^{\prime } } , \label{25a}\\ \hfill\nonumber\\ & & \sum\limits_{i}{\rm tr}\big[\gamma^{i}\lambda_{p}^{\pm}\gamma_0\gamma^{i}\lambda_{p^{\prime}}^{\pm}\gamma_0\big ] = \sum\limits_{i}{\rm tr}\big[\gamma^{i}\gamma_5\lambda_{p}^{\pm}\gamma^i\gamma_5\lambda_{p^{\prime}}^{\mp}\big]=3-\hat{p}\cdot\hat{p^{\prime } } , \label{25b}\\ & & { \rm tr}\big[\vec{\gamma}\cdot\vec{k}\lambda_{p}^{\pm}\gamma_0\vec{\gamma}\cdot\vec{k}\lambda_{p^{\prime}}^{\pm}\gamma_0\big ] = { \rm tr}\big[\vec{\gamma}\cdot\vec{k}\gamma_5\lambda_{p}^{\pm}\vec{\gamma}\cdot\vec{k}\gamma_5\lambda_{p^{\prime}}^{\mp}\big ] = ( 1+\hat{p}\cdot\hat{p^{\prime}})\frac{(p - p^{\prime})^2}{k^2 } \label{25c}\end{aligned}\ ] ] it can be shown that the contribution from eq.([24a ] ) to @xmath58 is suppressed by an order @xmath8 relative to that from eq.([24b ] ) and will be ignored here .
we neglect also the dependence of the gap function on the magnitude of the momentum , but keep the dependence on the energy and the momentum orientation .
then the integrals over @xmath0 and @xmath154 can be carried out easily .
since we are only interested in the zero temperature , the matsubara sum becomes an integral over the euclidean energy .
we find @xmath155=-\frac{6\bar{g}^2\mu^4}{32\pi^4}\int d\nu \int d\nu^{\prime}\int d^2\hat{p}\int d^2\hat{p^{\prime}}\frac{\phi(\nu,\hat{p})\phi(\nu^{\prime},\hat{p^{\prime}})}{\sqrt{(\nu^2+\phi^2(\nu,\hat{p}))(\nu^{\prime2}+\phi^2(\nu^{\prime},\hat{p^{\prime } } ) } ) } \times\big[d_l(\nu-\nu^{\prime},\theta)+d_t(\nu-\nu^{\prime},\theta)\big ] \label{26}\ ] ] where @xmath156 .
making use of the nambu - gorkov formalism in eq.([7]-[15 ] ) , the rest terms of the condensate energy density eq.([6 ] ) can be evaluated readily .
@xmath157 \label{27}\ ] ] @xmath158 the final expression of the condensation energy density reads @xmath159 where @xmath160 contains the contribution from both magnetic and electric gluons , i.e. @xmath161 the gap equation can be derived by minimizing @xmath58 with respect to the gap function @xmath162 , @xmath163 and we end up with @xmath164 a consistent derivation of the gap equation up to the subleading order requires both the contribution from @xmath165 and that from the diagonal block of ( [ 9 ] ) to be kept .
the net result is to replace the first term inside the square root on rhs of ( [ qcdgapeq ] ) by @xmath166 with @xmath167 the wave function renormalization of the normal phase . but
it will not interfere with the angular dependence of the gap function to the subleading order as will be shown in the next section .
although the pairing strength are equal to the leading order of the qcd running coupling constant , similar to the case ii of the toy model , the subleading terms fall off with an increasing @xmath10 .
this makes the solution to the gap equation ( [ qcdgapeq ] ) highly nontrivial . in what follows
, we shall isolate the energy dependence and the angle dependence of the pairing potential @xmath160 . a differential equation with respect to the matsubara energy
will be derived from ( [ qcdgapeq ] ) that fixes the gap function up to an arbitrary function of the angle .
this function will be determined then by ( [ qcdgapeq ] ) with @xmath65 a known function of the matsubara energy .
proceeding with the partial wave analysis , we expand @xmath168 in series of legendre polynomials@xcite : @xmath169 where @xmath170 and @xmath171 is given by eq.([subleading ] ) .
using the completeness relation @xmath172 and the identity ( proved in the appendix a ) @xmath173 with @xmath174 an arbitrary function of @xmath175 , the gap equation ( [ qcdgapeq ] ) becomes @xmath176\big\ } \label{37}\end{aligned}\ ] ] where @xmath177 , @xmath178 with @xmath179 a fixed spatial direction and a uv cutoff , @xmath180 is introduced . in deriving ( [ 37 ] ) , we have assumed that the gap depends on @xmath181 only , so the integration over the azimuthal angle of @xmath182 can be carried out explicitly . the gap equation ( [ 37 ] )
can be further simplified by using the approximation of son @xcite @xmath183 with @xmath184 .
it is convenient to introduce @xmath185 on writing @xmath186 and @xmath187 the gap equation([37 ] ) becomes @xmath188 taking the derivative of both sides with respect to @xmath189 , we find @xmath190 which implies the boundary condition @xmath191 as @xmath192 for all @xmath181 .
another derivative of ( [ 40 ] ) yield the ordinary differential equation @xmath193 which is universal for all @xmath181 .
it follows from eq.([39 ] ) that the gap equation is equivalent to @xmath194 the solution to ( [ 42 ] ) subject to the condition ( [ 41 ] ) contains an arbitrary function of @xmath181 to be determined by ( [ 43 ] ) .
no further approximation has been made up to now .
the solution to the differential equation ( [ 42 ] ) proceeds in the same way as that for a spherical gap . to the leading order
, the equation can be approximated by a linear one , @xmath195 where @xmath196 is to be determined by the condition @xmath197 .
its solution that satisfies the boundary condition([41 ] ) and the continuity up to the first order derivative reads @xmath198,\hspace{0.3 cm } { \rm for}\hspace{0.2 cm } \xi < b(x),\\ \hfill \\ \phi_0(x),\hspace{2.5 cm } { \rm for}\hspace{0.2 cm } \xi\ge b(x ) . \end{array } \right .
\label{45}\ ] ] where @xmath199 it follows from eq.([40 ] ) then that @xmath200,\hspace{0.3 cm } { \rm for}\hspace{0.2 cm } \xi < b(x),\\ \hfill \\ 0,\hspace{3.3 cm } { \rm for}\hspace{0.2 cm } \xi\ge b(x ) .
\end{array } \right . \label{47}\ ] ] the angle dependent factor @xmath201 introduced in eq.([gap2 ] ) is defined by @xmath202 where @xmath203 is the @xmath3-wave gap given by @xmath204 where the contribution from the wave - function renormalization is ignored . up to the subleading order , the differential equation([42 ] )
reads @xmath205\phi^{(0)}. \label{48}\ ] ] we find that @xmath206 where @xmath207 and @xmath208 are the two linearly independent solutions to the eq.([44 ] ) , @xmath209,\hspace{0.3 cm } { \rm for}\hspace{0.2 cm } \xi < b(x),\\ \hfill \\ 1,\hspace{2.3 cm } { \rm for}\hspace{0.2 cm } \xi\ge b(x ) . \end{array } \right.\ ] ] @xmath210,\hspace{0.3 cm } { \rm for}\hspace{0.2 cm } \xi < b(x),\\ \hfill \\ \bar g\xi,\hspace{2.4 cm } { \rm for}\hspace{0.2 cm } \xi\ge b(x ) .
\end{array}\right.\ ] ] and @xmath211v(\xi^{\prime},x ) \phi^{(0)}(\xi^\prime , x ) , \label{50a}\ ] ] @xmath212u(\xi^{\prime},x ) \phi^{(0)}(\xi^\prime , x ) .
\label{50b}\ ] ] at the point @xmath213 , we have @xmath214 therefore @xmath215 \label{52}\ ] ] to the subleading order . since @xmath216 according to eq.([55 ] )
, we have @xmath217 and @xmath218+o(g ) \label{53}\ ] ] substituting eqs.([53_1 ] ) and ( [ 53 ] ) into eq.([43 ] ) , we obtain the gap equation to the subleading order @xmath219\phi_0(x)+\int\limits_{-1}^{1}dx^{\prime}\frac{\phi_0(x^{\prime})-\phi_0(x)}{|x - x^{\prime}|}=0 \label{54}\ ] ] then the integral equation for @xmath220 , @xmath221 follows from ( [ 55 ] ) .
few comments are in order : 1 ) the spherical gap , @xmath222 is a trivial solution to eq.([56 ] ) and there is no angular momentum mixing .
2 ) the `` spin-1 '' gap , carried over from the polar phase of @xmath223 , @xmath224 , fails to satisfy this equation .
3 ) eq.([56 ] ) conserves the parity . in another word , its solution can be either an even or an odd function of @xmath181 .
4 ) if the wavefunction renormalization is restored , there will be an additional subleading term on rhs of ( [ 48 ] ) and an additional subleading term on rhs of @xmath225 of eq.([51 ] ) .
this term , when substitute into eq.([43 ] ) , will cancel the corresponding contribution to @xmath203 @xmath226 leaving the integral equation ( [ 56 ] ) intact .
the solution to the integral equation eq.([56 ] ) can be obtained from a variational principle . upon substitution of eq.([45 ] ) with @xmath227 into eq.([29 ] ) , the condensate energy density becomes a functional of @xmath228(details in appendix b ) , i. e. @xmath229,\ ] ] where @xmath230&=&\int\limits_{-1}^{1}dxf^2(x)\big[{\rm ln}|f(x)|-\frac{1}{2}\big]+\frac{3}{2}\int\limits_{-1}^{1}dx\int\limits_{-1}^{1}dx^{\prime}\frac{\big[f(x)-f(x^{\prime})\big]^2 } { |x - x^{\prime}|}\\ & = & 2\int\limits_{0}^{1}dxf^2(x)\big[{\rm ln}|f(x)|-\frac{1}{2}\big]+3\int\limits_{0}^{1}dx\int\limits_{0}^{1}dx^{\prime}\big\ { \frac{\big[f(x)-f(x^{\prime})\big]^2 } { |x - x^{\prime}|}+\frac{\big[f(x)+f(x^{\prime})\big]^2 } { x+x^{\prime}}\big\ } \label{a11 } \label{functional}\end{aligned}\ ] ] with the last equality following from the odd parity of @xmath220 , i. e. @xmath231 .
readers may easily verify that the variational minimum of eq.([functional ] ) does solve eq.([56 ] ) . before the numerical solution
, we consider a trial function @xmath232 and substitute it into the target functional ( [ functional ] ) .
the minimization yields @xmath233 at which @xmath234\simeq -3.989\times 10^{-6}. \label{initial}\ ] ] the trial function ( [ trial ] ) is what people carried over from the polar phase of @xmath223 .
the `` -6 '' of the exponent of ( [ prefactor ] ) comes from the pairing strength of the @xmath0-wave and the `` @xmath79 '' stems from the logarithm of ( [ 56 ] ) .
the latter contribution was reported in @xcite .
the trial function ( [ trial ] ) with @xmath0-wave alone is not optimal . the free energy will be lowered further by including higher partial waves of odd @xmath10 as we shall see . to find the variational minimum , we discretize the integral of eq.([functional ] ) by dividing the domain @xmath235 into @xmath236 intervals with @xmath237 where @xmath238 .
we have then @xmath239 with @xmath240 \label{a13}\ ] ] where we have dropped the limit @xmath241 of the first term inside the curly bracket of eq .
( [ functional ] ) .
@xmath242 is a function of @xmath6 variables .
the stationary condition @xmath243 yields @xmath244f_j-3\delta x\sum\limits_{i\neq j}\big(\frac{1}{|x_i - x_j|}+\frac{1}{x_i+x_j}\big)f_i=0 \label{a15}\ ] ] which is a discrete version of eq .
( [ 56 ] ) .
regarding @xmath245 s as given , the equation for @xmath246 is of the form @xmath247 with @xmath93 and @xmath248 positive .
it has one and only one solution for @xmath249 .
we start with the trial function ( [ trial ] ) , @xmath250 as an initial configuration and update each @xmath246 by solving eq.([a16 ] ) .
this way we lower the value of the target functional @xmath251 in each step and approach the solution to ( [ 56 ] ) eventually .
the process converges rapidly and our numerical solution to ( [ 56 ] ) is shown as solid line in fig.[fig : epsart ] , which depart from the trial function ( dashed line ) slightly .
we find the minimum value of the target functional @xmath234\simeq -4.130\times 10^{-6}.\ ] ] which drops from ( [ initial ] ) by @xmath252 percent .
it is instructive to examine the angular momentum contents of our solution in the partial wave expansion @xmath253 the coefficients of the first three partial waves , @xmath254 , calculated by substituting the numerical solution into the formula @xmath255 are displayed in table i. while the gap function contains all partial waves of odd @xmath10 , the component of @xmath15 is the biggest .
this is anticipated because the pairing strength of the all partial waves are equal in leading order but fall off with an increasing @xmath10 in the subleading order as is shown in the partial wave expansion ( [ 33 ] ) .
c|ccc + & @xmath256 & @xmath257 & @xmath258 + + + @xmath259 & @xmath260&@xmath261 & @xmath262 + + the angular dependence of the gap function with angular momentum mixing . the dashed line and the solid line are the initial configuration and the final numerical results respectively . ]
in summary we have explored the angular dependence of the gap function for a non - spherical pairing of csc .
because of the equal strength of the pairing potential mediated by one - gluon exchange for all partial waves to the leading order of qcd running coupling constant and the nonlinearity of the gap equation , a non - spherical gap function can not be restricted to one angular momentum channel only .
other multipoles are bound to show up , which renders the angular dependence of the gap nontrivial . on the other hand , the pairing strength to
the subleading order decreases with increasing angular momentum @xmath10 .
the mixing effect will not be as big as that in the soluble toy model we introduced for the purpose of clarification . for the single flavor csc , we worked out the angular momentum mixing effect explicitly for the gap function with zero azimuthal quantum number at zero temperature .
an nonlinear integral equation for the nontrivial angular dependence was derived and its solution was obtained numerically .
the gap function in this case reads @xmath263 where @xmath203 is given by eq.([delta0 ] ) and @xmath264 is plotted in fig.[fig : epsart ] .
the drop of the free energy of the modified polar phase by the mixing , however , is numerically small . the magnitude of its condensation energy is smaller than that of the csl phase by a factor of 1.48 instead of the factor 1.54 reported in @xcite .
the csl phase remains stable . in this sense
our results at the moment is of theoretical values only .
there are many other candidate pairing states between quarks of the same flavor @xcite . among them
are the states with a nonzero azimuthal quantum number and the pairing between quarks of opposite helicities .
the former is analogous to the @xmath265 phase of @xmath223 and may be present in a compact star with a strong magnetic field .
the pairing force in the unequal - helicity channel is stronger @xcite .
the angular momentum mixing effect is generic in all nonspherical pairing states and the integral equation ( [ 56 ] ) can be readily generalized to these cases .
there may be phenomenological implications of the angular momentum mixing . a systematic survey of the angular momentum mixing effect in all `` spin-1 '' csc states covered in @xcite will be reported in another paper .
another place where the angular momentum mixing shows up is the csc - loff state in the presence of fermi momentum mismatch .
it has been speculated @xcite that the forward singularity will increase the upper limit of the mismatch value that supports a loff pairing .
the new threshold was found in @xcite , motivated by the nearly equal pairing strength of all partial wave channels .
the same mechanism works for the gap equation of loff pairing .
its free energy will be lowered by the angular momentum mixing and the lower edge of the loff window is expected to be shifted to a lower value of the mismatch parameter .
we would like to extend our gratitude to d. rischke , t. sch@xmath266fer and a. schmitt for stimulating discussions and valuable comments .
we are also benefitted from conversations with j.r .
li and q. wang .
the work of d. f. h. and h. c. r. is supported in part by nsfc under grant no .
10575043 and by us department of energy under grants de - fg02 - 91er40651-taskb .
the work of d. f. h. is also supported in part by educational committee of china under grant ncet-05 - 0675 and project no .
the integral formula of @xmath171 is@xcite @xmath267 it is convenient to introduce @xmath268 where @xmath269 is an infinitesimal quantity .
we have @xmath270 . for the first term on rhs ,
we expand @xmath271 according to @xmath272 therefore eq . ( [ b2 ] ) reads @xmath273 evaluating the summation in eq.([36 ] ) is straightforward now @xmath274 then for an arbitrary function @xmath201 , @xmath275 the eq .
( [ 36 ] ) is obtained by taking the limit @xmath276 .
in this appendix , we shall derive the expression eq.([29 ] ) ) of the condensation energy density with the angular momentum mixing . substituting eq.([33 ] ) into the first term of eq .
( [ 29 ] ) , we find @xmath277\frac{\phi(\nu,\hat{p})\phi(\nu^{\prime},{\hat p^{\prime } } ) } { \sqrt{(\nu^2+\phi^2(\nu,\hat{p}))(\nu^{\prime2}+\phi^2(\nu^{\prime},{\hat p^{\prime}})})}\\ \nonumber & = & -\frac{\bar g^2\mu^2}{16\pi^3}\big\{\int d\nu^{\prime}\int d\nu\int d^2{\hat p}{\rm ln}\frac{\omega_c}{|\nu-\nu^{\prime}|}\frac{\phi(\nu,\hat{p})\phi(\nu^{\prime},\hat{p } ) } { \sqrt{(\nu^2+\phi^2(\nu,\hat{p}))(\nu^{\prime2}+\phi^2(\nu^{\prime},\hat{p})})}\\ & & -\frac{3}{4\pi}\int d\nu^{\prime}\int d\nu\int d^2{\hat p^{\prime}}\int d^2{\hat p}\frac{1}{1-{\hat p^{\prime}}\cdot{\hat p}}\big[\frac{\phi(\nu , { \hat p})}{\sqrt{\nu^2+\phi^2(\nu , { \hat p})}}-\frac{\phi(\nu , { \hat p^{\prime}})}{\sqrt{\nu^2+\phi^2(\nu , { \hat p^{\prime}})}}\big]\big[(\nu\leftrightarrow\nu^{\prime})\big]\big\ } \label{a1}\end{aligned}\ ] ] because of the eveness of @xmath278 in @xmath279 , we have @xmath280[\nu^{\prime2}+\phi^2(\nu^{\prime},\hat p)]}}\\ & & -\frac{3}{4\pi}\int d^2{\hat p^{\prime}}\int d^2{\hat p}\int\limits_{0}^{\omega_0}d\nu^{\prime}\int\limits_{0}^{\omega_0}d\nu\frac{1}{1-{\hat p}\cdot{\hat p^{\prime}}}\big[\frac{\phi(\nu , { \hat p})}{\sqrt{\nu^2+\phi^2(\nu , { \hat p})}}-\frac{\phi(\nu , { \hat p^{\prime}})}{\sqrt{\nu^2+\phi^2(\nu , { \hat p^{\prime}})}}\big ] \big[(\nu\leftrightarrow\nu^{\prime})\big]\big\ } \label{a3}\end{aligned}\ ] ] where the approximation ( [ son ] ) has been applied to the forward logarithm . for the gap function of zero azimuthal quantum number ,
@xmath281 depends only on @xmath282 .
we find that @xmath283 ^ 2}{|x - x^{\prime}|}\big\ } \label{a4}\ ] ] where , @xmath284 has been defined in eq.([38 ] ) and @xmath189 and @xmath93 have been defined in ( [ newvar ] ) .
the integral over @xmath285 followed by the integral by part over @xmath189 leads to @xmath286 ^ 2}{|x - x^{\prime}|}\big\ } \label{a5}\ ] ] making use of eq .
( [ 38 ] ) and ( [ 53 ] ) , we have @xmath287+{\bar g^2}\int\limits_{a}^{\infty}d\xi\frac{\phi^2(\xi , x)}{\sqrt{1+\frac{\phi^2(\xi , x)}{\omega_c}e^{2\xi } } } \label{a6}\ ] ] and thus @xmath288 substituting @xmath227 into ( [ a5 ] ) , we obtain that @xmath289 ^ 2 } { |x - x^{\prime}|}\big\}-\frac{\mu^2}{2\pi^2}\int\limits_{0}^{\omega_0}d\nu\frac{\phi^2(\nu , x)}{\sqrt{\nu^2+\phi^2(\nu , x ) } } \label{a8}\ ] ] then the condensate energy density with the angular momentum mixing reads @xmath290\\ & = & \frac{\mu^2\delta_0 ^ 2}{2\pi^2}\big\{\int\limits_{-1}^{1}dxf^2(x)\big[{\rm ln}|f(x)|-\frac{1}{2}\big]+\frac{3}{2}\int\limits_{-1}^{1}dx\int\limits_{-1}^{1}dx^{\prime}\frac{\big[f(x)-f(x^{\prime})\big]^2 } { |x - x^{\prime}|}\big\ } \label{a9}\end{aligned}\ ] ] the minimization of this free energy give rise to eq.([56 ] ) of the text . w. e. brown , j. t. liu , and h - c .
ren , phys . rev . *
d62 * , 054016 ( 2000 ) .
g. d. mahan , many particle physics , 1990 , plenum press , new york .
w. e. brown , j. t. liu , and h - c .
ren , phys . rev . *
d62 * , 054013 ( 2000 ) . | we study the angular momentum mixing effects in the color superconductor with non - spherical pairing .
we first clarify the concept of the angular momentum mixing with a toy model for non - relativistic and spinless fermions .
then we derive the gap equation for the polar phase of dense qcd by minimizing the cjt free energy .
the solution of the gap equation consists of all angular momentum partial waves of odd parity .
the corresponding free energy is found to be lower than that reported in the literature with @xmath0-wave only . |
quantum chromodynamics ( qcd ) is the fundamental theory of the strong interaction .
qcd is an asymptotically free theory @xcite , i.e. , interactions between quarks and gluons become weaker as the mutual distance decreases or as the exchanged momentum increases . consequently , at very large temperatures and/or densities , the interactions which confine quarks and gluons inside hadrons should become sufficiently weak to release them @xcite .
the phase where quarks and gluons are deconfined is termed the _ quark - gluon plasma _ ( qgp ) .
lattice qcd calculations have established the existence of such a phase of strongly interacting matter at temperatures larger than @xmath0 mev and zero net - baryon density .
depending on the number of quark flavors and the masses of the quarks , the transition between ordinary hadronic matter and the qgp could be a thermodynamic phase transition of first order , of second order , or simply a crossover transition .
the qgp was certainly present in the evolution of the early universe .
the early universe was very hot , but close to net - baryon free . in the opposite limit of small temperature and large baryon density
, the qgp may exist even nowadays in the interior of compact stellar objects such as neutron stars .
the main effort in present - day nuclear physics is to create the qgp under controlled conditions in the laboratory via collisions of heavy nuclei at ultrarelativistic energies @xcite .
the temperatures and net - baryon densities reached in nuclear collisions depend on the bombarding energy .
they interpolate between the extreme conditions of the early universe on one side and compact stellar objects on the other .
if at all , the qgp is only transiently created in a nuclear collision ; free quarks and gluons will not be released .
therefore , deciding whether a qgp was formed or not is not easy .
detectors in the laboratory can only measure hadrons , leptons , or photons .
the bulk of the particles emerging from a nuclear collision are hadrons with transverse momenta of order @xmath1 gev .
they carry information about the final stage of the collision after hadronization of the qgp .
the formation of the latter can only indirectly influence this final stage , for instance by modifying the collective dynamics of the system through a softening of the equation of state in the hadronization transition @xcite .
very few hadrons are emitted with transverse momenta of the order of several gev .
they arise from the fragmentation of jets and may carry information also about the earlier stages of the collision . of particular interest
is the situation where the jet has to traverse hot and dense matter possibly formed in the collision and is quenched by multiple rescattering in the medium @xcite . from this `` jet - quenching '' process one
may indirectly learn about the properties of the hot and dense environment .
finally , leptons and photons only interact electromagnetically .
once formed in the early stage of the collision , they leave the system unimpededly and carry information about this stage to the detector @xcite .
the difficulty is to disentangle the thermal radiation from a hot , equilibrated qgp @xcite from the initial production of leptons and photons in the very first , highly energetic partonic collisions and from the thermal radiation of hot hadronic matter @xcite . in order to decide whether a qgp was formed , one has to have detailed knowledge about its properties .
otherwise it is impossible to find an unambiguous signature for qgp formation in nuclear collisions . in this review
, i present an overview of the _ thermodynamical properties _ of the qgp .
section [ secphasediag ] contains a detailed discussion of the phase diagram of strongly interacting matter .
the present status of knowledge is shown schematically in fig .
[ phasediagramnm ] . depending on the temperature , @xmath2 , and the quark chemical potential , @xmath3
, strongly interacting matter may occur in three distinct phases : the hadronic phase , the qgp , and color - superconducting quark matter .
the ground state of ( infinite ) nuclear matter is at @xmath4 mev .
there is a line of first - order phase transitions emerging from this point and terminating in a critical endpoint at a temperature of order @xmath5 mev . at this point ,
the transition is of second order .
this phase transition is the nuclear liquid - gas transition @xcite . to the left of the line nuclear matter
is in the gaseous phase , and to the right in the liquid phase . above the critical endpoint , there is no distinction between these two phases . for temperatures below @xmath6 mev and quark chemical potentials below @xmath7 mev ( corresponding to net - baryon densities which are a few times the ground state density of nuclear matter ) , strongly interacting matter is in the hadronic phase . quite similar to the liquid - gas transition , there is a line of first - order phase transitions which separates the hadronic phase from the qgp and terminates in a critical endpoint where the transition is of second order .
this endpoint is approximately at @xmath8 mev , cf .
[ nonzeromu ] . for smaller quark chemical potentials ( smaller net - baryon densities ) ,
the transition becomes crossover , and there is no real distinction between hadronic matter and the qgp . as will be discussed in detail in sec .
[ secphasediag ] , the position of the critical endpoint depends on the value of the quark masses .
finally , at large quark chemical potential ( large baryon density ) and small temperature , quark matter becomes a color superconductor .
there can be multitude of color - superconducting phases , depending on the symmetries of the order parameter for condensation of quark cooper pairs .
the discussion in sec .
[ secphasediag ] is qualitative and is meant to give an overview of the phase structure of strongly interacting matter at nonzero temperature and quark chemical potential .
the discussion in the following sections is both more quantitative as well as technical and focusses on the properties of the qgp and color - superconducting quark matter .
the early universe evolved close to the temperature axis in the phase diagram of strongly interacting matter .
matter in the core of compact stellar objects , like neutron stars , is close to the quark chemical potential axis , at values of @xmath3 around 400 500 mev .
nuclear collisions at bombarding energies around @xmath9 agev explore a region of temperatures and quark chemical potentials around @xmath10 mev .
collisions at current rhic energies of @xmath11 agev are expected to excite matter in a region around and above @xmath12 mev .
collision energies in between these two extremes cover the intermediate region and , in particular , may probe the critical endpoint .
section [ lattqcd ] presents a review of lattice qcd . after a brief introduction to the basic principles , results on the order of the qcd phase transition , the equation of state of strongly interacting matter ,
the heavy quark free energy , and mesonic spectral functions are collected . for technical reasons ,
most lattice qcd calculations have been done at zero quark chemical potential .
an extension to nonzero values of @xmath3 is difficult and has been started only fairly recently .
first results will also be discussed .
lattice qcd is in principle an exact method to solve qcd .
if one had sufficiently large computer power , one could not only decrease the lattice spacing and increase the size of the system to come sufficiently close to the continuum and thermodynamic limit , one could also sample over a sufficiently large number of configurations to make the statistical errors arbitrarily small .
however , one still has to interpret the results in physical terms . in this respect ,
analytic approaches to solve qcd have a certain advantage over lattice qcd . in an analytic approach ,
one has complete control over the physical assumptions entering the calculation .
section [ seciii ] gives an overview of what is known about the qgp from analytic calculations .
the most simple approach from a conceptual ( albeit not technical ) point of view is to determine the thermodynamical properties of the qgp by a perturbative computation of the qcd partition function in terms of a power series in the strong coupling constant , @xmath13 .
this can be done up to terms of order @xmath14 . at order @xmath15 ,
the perturbative series breaks down @xcite , and the remaining nonperturbative contribution has to be determined , for instance , from a lattice qcd calculation .
those terms of the perturbative series , which are analytically computable , are rapidly decreasing in magnitude at high temperatures where the strong coupling constant is small , @xmath16 .
this gives rise to the hope that only the first few terms of the perturbative series are actually necessary to obtain physically reasonable values for the qcd partition function . for temperatures of order @xmath0 mev ,
however , @xmath17 and the perturbative series is obviously not convergent .
therefore , one has tried other ways to compute the partition function , either by expanding around a nonperturbative ground state or by resumming certain classes of diagrams to improve the convergence properties of the perturbative series . in both approaches ,
quarks and gluons are considered as quasiparticles with a dispersion relation which is modified as compared to the one in the vacuum .
still another approach is to construct an effective theory for qcd which can be solved exactly or at least within a many - body approximation scheme .
all these approaches will be reviewed in sec .
[ seciii ] .
section [ csc ] contains an introduction to color superconductivity at large quark chemical potentials . in this case , analytic calculations are well under control , because corrections can be systematically computed in terms of powers of @xmath13 . after a derivation of the gap equation for the color - superconducting gap function ,
the excitation spectrum in a color superconductor is presented .
the solution of the gap equation is discussed and the magnitude of the gap parameter is determined . as in ordinary superconductors ,
quark cooper pairs break up , if the thermal motion becomes too large .
the critical temperature for the transition between the normal- and the superconducting phase is computed .
finally , the properties of gluons and photons in color superconductors are discussed .
section [ concloutl ] concludes this review with a brief summary of the material and an outlook towards future directions of research in this area .
a lot of the material in this review can also be found in other places . the standard review for properties of hot and dense , strongly interacting matter is the seminal work of gross , pisarski , and yaffe @xcite .
the contents as well as more recent developments have found their way into textbooks @xcite . for early reviews focussing on the properties of the qgp ,
see ref .
an introduction to lattice qcd and recent results can be found in ref .
@xcite . the present status of lattice qcd is also reviewed in ref .
resummation techniques which attempt to compute the qcd partition function analytically are discussed in great detail in ref .
@xcite .
the present review tries to give a balanced overview of all subfields concerned with the equilibrium properties of the qgp .
therefore , the presentation is not as detailed as in a more specialized review .
on the other hand , i tried to explain the basics in somewhat greater detail than usually found in the literature .
my hope is that in this way , this review will become useful for early - stage researchers working in both theory as well as experiment , and for all researchers who would like to get an overview of the theoretical activity related to equilibrium properties of the qgp .
the only somewhat more specialized and thus more technical part is the section on color superconductivity .
this field has seen a lot of activity only fairly recently , but there are already a couple of excellent reviews @xcite . these , however , focus mainly on the basics of the phenomenon of color superconductivity and its phenomenological implications .
in contrast , sec .
[ csc ] contains a very detailed discussion of how to compute properties of the quasiparticle excitations in a color superconductor in the weak - coupling limit . by clarifying some of the technical details ,
i hope to remove the obstacles that may prevent other researchers to enter this rapidly evolving and rather exciting new field of strongly interacting matter physics . due to space restrictions this review
is forced to omit many things that could ( and possibly , should ) also have been said about strongly interacting matter at high temperature and/or density .
fortunately , most of these have already been covered in review articles .
these are , for instance , non - equilibrium properties of the qgp @xcite or the physics of instantons in non - abelian gauge theories @xcite .
another important topic which is not mentioned in this work , but for which excellent reviews exist @xcite , are the experimental signatures for the qgp .
recent developments in the field of color superconductivity are mainly focussed on deriving effective theories for quarks around the fermi surface .
these greatly simplify calculations and allow to systematically study effects of nonzero quark masses , for details , see ref .
@xcite . finally , the list of references is , necessarily , far from complete .
i would like to apologize to all authors whose work should have been ( but was not ) mentioned .
the units are @xmath18 .
i work in euclidean space - time at nonzero temperature @xmath2 , i.e. , space - time integrals are @xmath19 , where @xmath20 is the 3-volume of the system .
energy - momentum integrals are @xmath21 ; @xmath22 labels the matsubara frequencies @xmath23 for bosons and @xmath24 for fermions , @xmath25 .
i denote 4-vectors with capital letters , but unless mentioned otherwise , retain a notation familiar from minkowski space : @xmath26 , where @xmath27 , and @xmath28 , where @xmath29 , with the metric tensor @xmath30 .
3-vectors @xmath31 have length @xmath32 and direction @xmath33 .
in order to understand the phase structure of strongly interacting matter , one has to know its equation of state . in the grand canonical ensemble ,
the equation of state is determined by the grand partition function @xcite @xmath34\,\ , , \ ] ] where @xmath3 is the quark chemical potential associated with ( net ) quark number conservation .
the qcd lagrangian is given by @xmath35 for @xmath36 colors and @xmath37 flavors , @xmath38 is the @xmath39-dimensional spinor of quark fields , @xmath40 is the dirac conjugate spinor , @xmath41 are the dirac matrices and @xmath42 is the quark mass matrix . the covariant derivative is defined as @xmath43 , with the strong coupling constant @xmath13 , the gluon fields @xmath44 , and the generators @xmath45 of the local @xmath46 $ ] symmetry .
( throughout this paper , i indicate local , i.e. gauged , symmetries by square brackets . )
the latter are usually taken as @xmath47 , where @xmath48 are the gell - mann matrices .
the gluonic field strength tensor is defined as @xmath49 where @xmath50 are the structure constants of @xmath46 $ ] .
the term @xmath51 in eq .
( [ lqcd ] ) will not be specified further .
it comprises gauge fixing terms and the contribution from faddev - popov ghosts . the number density operator associated with the conserved ( net ) quark number is @xmath52 . for any finite volume @xmath20 and nonzero temperature @xmath2 ,
the partition function is defined for a compact euclidean space - time volume @xmath53 . for the sake of simplicity ( but without loss of generality ) , assume that the spatial volume @xmath20 is a box , @xmath54 , with @xmath55 being the length of the box in one spatial dimension .
all fields are then usually taken to be periodic in spatial directions , @xmath56 , where @xmath57 stands generically for @xmath38 , @xmath58 , and @xmath44 .
bosonic fields , such as gluons , are periodic also in temporal direction , @xmath59 , while fermionic fields , such as quarks , are antiperiodic in temporal direction , @xmath60 . from the grand partition function
, one can derive other thermodynamic quantities , for instance the pressure , @xmath61 in the thermodynamic limit , @xmath62 is an extensive quantity ( @xmath63 ) and the dependence of the pressure on @xmath20 drops out .
phase transitions are determined by studying the derivatives of the pressure with respect to @xmath2 and @xmath3 for a given point @xmath64 in the phase diagram of the independent thermodynamic variables temperature and chemical potential . for a phase transition of first order , the first derivatives @xmath65 are discontinuous while the pressure @xmath66 is continuous at the point @xmath64 . here , @xmath67 is the entropy density and @xmath22 the ( net ) quark number density . for a phase transition of second order ,
the second derivatives are discontinuous , while @xmath66 and its first derivatives are continuous . in this way
, phase transitions of arbitrarily high order can be defined .
one speaks of a crossover transition , if thermodynamic properties change rapidly within a narrow range of values @xmath2 and @xmath3 , but the pressure and all its derivatives remain continuous . usually , the points @xmath64 where a phase transition occurs are continuously connected and thus determine a _ line _ of phase transitions in the phase diagram .
these lines usually start on either the @xmath2 or the @xmath3 axis .
they may terminate for nonzero values of @xmath2 and @xmath3 .
two examples for this behavior , the liquid - gas transition in nuclear matter and the quark - hadron transition , have already been seen in the phase diagram of strongly interacting matter , fig .
[ phasediagramnm ] , and will be discussed in more detail in the following .
the liquid - gas transition in nuclear matter is a consequence of the fact that nuclear matter assumes its ground state at a nonvanishing baryon density @xmath68 at zero temperature @xmath69 .
the underlying microscopic mechanism for this phenomenon is a competition between attractive and repulsive forces among nucleons , with the attraction winning at this particular value of the baryon density .
( this is good , because otherwise there would be no stable atomic nuclei , precluding the existence of our universe as we know it . ) in infinite , isospin - symmetric nuclear matter , nucleons in the ground state are bound by @xmath70 mev ( if one neglects the coulomb repulsion ) , i.e. , the energy per baryon is @xmath71 mev , where @xmath72 is the energy density , @xmath73 the baryon density , and @xmath74 mev is the rest mass of the nucleon .
nuclear matter is mechanically stable in the ground state , such that the pressure vanishes , @xmath75 .
from the fundamental relation of thermodynamics , @xmath76 , one then concludes that the baryon chemical potential in the ground state is identical to the energy per baryon , @xmath77 mev . since a baryon consists of three quarks , @xmath78 and @xmath79 .
hence , the ground state of nuclear matter is the point @xmath80 mev in the nuclear matter phase diagram . obviously , it costs energy to compress nuclear matter to baryon densities @xmath81 .
such an increase in energy leads to an increase in pressure . at zero temperature , this can be immediately seen from the identity @xmath82 . since the pressure is a monotonous function of the thermodynamic variables , and since it vanishes in the ground state , there are only two possibilities for the behavior of @xmath66 for densities @xmath83 : either the pressure remains zero as the density decreases , @xmath84 , or the pressure further decreases such that @xmath85 .
the latter possibility implies that the system becomes mechanically unstable .
this can be prevented by fragmenting nuclear matter into droplets .
these droplets are mechanically stable , i.e. , the density inside each droplet is equal to @xmath86 and the pressure vanishes .
the total baryon density in the system can be further reduced by simply decreasing the droplet density .
the pressure in such a system remains zero down to arbitrarily small densities , because compression just results in a decrease of the space between droplets .
thus , @xmath84 from @xmath87 to @xmath88 , and then @xmath89 for @xmath90 . at small , but nonzero temperatures ,
this picture remains valid , with the additional possibility to thermally evaporate single nucleons from the surface of the droplets . at small temperatures and densities below the ground state density ,
one thus has a mixed phase of nucleons and droplets of nuclear matter .
this is reminiscent of water which , at room temperature and normal pressure , consists of a mixed phase of water molecules and water droplets . changing the density one can alter the relative fraction of molecules and droplets . beyond the density where droplets fill the entire volume one enters the liquid phase , while below the density where the last droplet fragments into molecules one enters the gas phase .
this behavior is typical for a first - order phase transition . in this case , this is the so - called liquid - gas transition in water .
nuclear matter shows a similar behavior , featuring a `` gaseous '' phase of nucleons at small chemical potentials ( densities ) and a `` liquid '' phase of nuclear matter at large chemical potentials ( densities ) , cf .
[ phasediagramnm ] . at small temperatures
the transition between the two phases is of first order .
thus , in the @xmath64 phase diagram there is a line of first - order phase transitions extending from the nuclear ground state @xmath91 mev up towards larger values of @xmath2 and smaller values of @xmath3 .
as for water , this line terminates at a critical point where the transition becomes of second order .
the critical temperature is of the order of 10 mev . as for water
, one can not distinguish between the gaseous and the liquid phase for temperatures above this critical temperature .
the existence of the liquid - gas phase transition has been confirmed in heavy - ion collision experiments at bevalac and gsi energies ( @xmath9 agev ) , although the precise value for the critical temperature and the critical exponents remain a matter of debate @xcite .
the liquid - gas transition is also a feature of phenomenological models for the nuclear interaction , for instance , the walecka model @xcite . in the following section another phase transition in strongly interacting matter
is discussed , which very much resembles the liquid - gas transition in that it ( most likely ) is of first order for small temperatures and terminates in a critical point where the transition becomes of second order .
this transition is the so - called quark - hadron transition . for a non - interacting , translationally invariant system a convenient basis of states
are the single - particle momentum eigenstates . due to the pauli principle
, the density in a fermionic system can only be increased by successively filling states with higher momentum .
the highest filled state defines the fermi surface of the system , and the corresponding momentum is the fermi momentum , @xmath92 . for non - interacting systems at zero temperature , the single - particle density @xmath22 is given in terms of the fermi momentum as @xmath93 where @xmath94 counts the internal degrees of freedom of the fermion ( like spin , color , flavor , etc . ) .
thus , at large densities the fermi momentum becomes large . in a cold
, dense fermionic system particles can only scatter elastically if their momenta lie on the fermi surface , as states below the fermi surface are not accessible due to the pauli principle ( the so called `` pauli - blocking '' phenomenon ) , and states above the fermi surface are not accessible due to energy conservation .
if the fermi momentum exceeds the qcd scale parameter @xmath95 mev , scattering events between nucleons start to probe distances of the order 1 fm or less , i.e. , the nucleonic substructure of quarks and gluons becomes visible .
the fermi momentum in the ground state of nuclear matter can be inferred from eq .
( [ density ] ) to be @xmath96 mev .
this is already of the order of @xmath97 .
nevertheless , a description of nuclear matter in terms of nucleonic degrees of freedom is certainly feasible around the ground state . at which densities
does a description in terms of quark and gluon degrees of freedom become more appropriate ?
the `` volume '' occupied by a single nucleon can be estimated from its charge radius to be @xmath98 .
on the other hand , the specific volume of the system in the ground state is @xmath99 . in this sense ,
nuclear matter in the ground state is dilute .
however , increasing the density to about @xmath100 , the system becomes densely packed with nucleons . at larger densities
, they will even start to overlap .
therefore , around densities of a few times nuclear matter ground state density , one expects that a description of the system in terms of quarks and gluons is more appropriate
. similar arguments also apply to a system at nonzero temperature , even when the net - baryon number density is small . at nonzero temperature ,
nuclear matter consists not only of nucleons but also of other , thermally excited hadrons . for a non - interacting system in thermodynamical equilibrium and neglecting quantum statistics ,
the hadron number densities are proportional to @xmath101 , where @xmath102 labels the hadron species , @xmath103 is their mass , @xmath104 is their chemical potential , and @xmath105 is a modified bessel function of the second kind . for nonzero temperature and small net - baryon number density ,
the lightest hadrons , the pions , are most abundant . at nonzero temperature and small baryon chemical potential ,
the typical momentum scale for scattering events between hadrons is set by the temperature @xmath2 .
if the temperature is on the order of or larger than @xmath97 , scattering between hadrons starts to probe their quark - gluon substructure .
moreover , since the particle density increases with the temperature , the hadronic wave functions will start to overlap for large temperatures .
consequently , above a certain temperature one expects a description of nuclear matter in terms of quark and gluon degrees of freedom to be more appropriate .
the picture which emerges from these considerations is the following : for quark chemical potentials @xmath3 which are on the order of 350 mev or smaller , and for temperatures @xmath106 mev , nuclear matter is a gas of hadrons .
( at very small temperatures @xmath107 mev , there is a gaseous and a liquid nucleonic phase , cf .
[ lgtrans ] . ) on the other hand , for @xmath108 , nuclear matter consists of quarks and gluons .
the natural question which emerges is , whether the `` hadron phase '' and the `` quark - gluon phase '' ( the qgp ) are separated by a phase transition in the thermodynamic sense .
the rigorous way to decide this question is by identifying an order parameter which is nonzero in one phase and zero in the other . this will be discussed in more detail in the following .
let us first study the pure @xmath109 $ ] gauge theory , i.e. , qcd without dynamical quarks ( sometimes also termed the @xmath110 case ) .
in this theory , there is a phase transition between a low - temperature and a high - temperature phase , cf .
[ translatt ] .
the order parameter for this transition is the expectation value @xmath111 of the polyakov loop ( or wilson line ) @xmath112 \right\}\,\,,\ ] ] where @xmath113 stands for path - ordering .
the expectation value of an operator @xmath114 in the grand canonical ensemble is defined as @xmath115\,\ , .\ ] ] the expectation value @xmath111 vanishes in the low - temperature phase . if one interpretes this expectation value as @xmath116 , where @xmath117 is the free energy of an infinitely heavy quark @xcite , then @xmath118 implies that the free energy is infinite , corresponding to confinement of color charges . in the high - temperature phase , @xmath119 , which implies that the free energy of an infinitely heavy quark is finite
this indicates the liberation of colored degrees of freedom , i.e. , deconfinement .
the expectation value of the polyakov loop is therefore the order parameter for the deconfinement transition . for an @xmath109 $ ] gauge theory
the action has a global @xmath120 symmetry : the action does not change when multiplying all time - like links at a given spatial position @xmath121 by an element @xmath122 of the center @xmath120 of the gauge group @xmath109 $ ] . in the high - temperature phase ,
the nonzero expectation value of the polyakov loop breaks this symmetry spontaneously .
in the low - temperature phase , @xmath118 , and this symmetry is restored . for two colors , @xmath123 , the effective theory in the critical region around the phase transition is given by a @xmath124 spin model , i.e. , it is in the same universality class as the ising model @xcite .
this model has a second - order phase transition . for @xmath125 ,
the effective theory is that of a @xmath126 spin model @xcite , i.e. , it is in the universality class of the 3-state potts model which has a first - order phase transition @xcite .
the transition temperature was computed to be @xmath127 mev @xcite , see also sec .
[ translatt ] . in the presence of dynamical quarks ,
@xmath128 , the picture becomes somewhat more complicated .
the fermionic term in the qcd lagrangian ( [ lqcd ] ) breaks the @xmath120 symmetry explicitly , and thus there is strictly speaking no order parameter for deconfinement .
nevertheless , the qcd transition in the presence of massless dynamical quarks has an order parameter , which is related to the chiral symmetry of qcd . while the qcd lagrangian ( [ lqcd ] ) is chirally symmetric when @xmath129 , the ground state of qcd is not , i.e. , chiral symmetry is spontaneously broken .
it is instructive to review these arguments in more detail . in the chiral limit , where the quark mass matrix is zero , @xmath129 ,
the qcd lagrangian ( [ lqcd ] ) is invariant under global chiral @xmath130 rotations of the quark fields . to see this ,
decompose the quark spinors into right- and left - handed spinors , @xmath131 where @xmath132 are chirality projectors . then perform a @xmath133 transformation on the right / left - handed quark spinors , @xmath134 where @xmath135 are the parameters and @xmath45 the generators of @xmath133 .
the lagrangian ( [ lqcd ] ) remains invariant under this transformation , @xmath136 .
the chiral group @xmath130 is isomorphic to the group @xmath137 of unitary vector and axial transformations , where @xmath138 , @xmath139 , i.e. , @xmath140 , @xmath141 .
any unitary group is the direct product of a special unitary group and a complex phase , @xmath142 .
thus , @xmath143 .
the vector subgroup @xmath144 of this symmetry group corresponds to quark number conservation . as physical states
trivially conserve quark number , this subgroup does not affect the chiral dynamics and can be omitted from the further symmetry consideration .
this leaves an @xmath145 symmetry .
the axial @xmath146 symmetry is broken explicitly by instantons ( the so - called @xmath146 anomaly of qcd ) @xcite , leaving an @xmath147 symmetry which determines the chiral dynamics .
since instantons are screened in a hot and/or dense medium @xcite , the @xmath146 symmetry may become effectively restored in matter .
then , the chiral symmetry is again enlarged to @xmath145 .
nonzero quark masses break the chiral symmetry of the qcd lagrangian explicitly . the quark mass term in eq .
( [ lqcd ] ) is @xmath148 where flavor indices @xmath149 are explicitly written , a sum over repeated indices is implied , and the properties @xmath150 of the chirality projectors were used .
now suppose all quark masses were equal , @xmath151 .
performing chiral @xmath145 rotations of the quark fields , one observes that the mass term ( [ massterm ] ) preserves an @xmath152 symmetry .
all axial symmetries are explicitly broken .
if less than @xmath37 quark masses are equal , say @xmath153 , the preserved vector symmetry is @xmath154 . in nature , where @xmath155 , one only has the well - known ( approximate ) @xmath156 isospin symmetry .
consequently , exotic hadrons with strange , charm , bottom , or top degrees of freedom are not degenerate in mass with their non - strange counterparts .
the mass term @xmath157 in the qcd lagrangian is of the same form as the term @xmath158 in spin models , which couples the spin @xmath159 to an external magnetic field @xmath160 .
obviously , the operator @xmath161 corresponds to the spin @xmath159 , while the quark mass matrix @xmath162 assumes the role of the external magnetic field @xmath160 .
thus , the expectation value @xmath163 is the analogue of the expectation value of the spin , the magnetization @xmath164 . while the mass term explicitly breaks the chiral symmetry @xmath145 of the qcd lagrangian to an ( approximate ) @xmath156 symmetry , the external magnetic field in spin models explicitly breaks the rotational symmetry @xmath165 to @xmath166 .
the analogy between qcd and spin models , however , extends further than this . even in the absence of external magnetic fields , in spin models with ferromagnetic interactions rotational symmetry
is spontaneously broken due to a nonvanishing magnetization @xmath167 in the ferromagnetic phase .
analogously , in the qcd vacuum , chiral symmetry is spontaneously broken by a nonvanishing expectation value @xmath168 .
let us introduce the so - called chiral condensate @xmath169 and its complex conjugate , @xmath170 , via @xmath171 a nonvanishing expectation value @xmath172 is then equivalent to @xmath173 .
just like the mass term in the qcd lagrangian , a nonvanishing chiral condensate breaks the chiral symmetry . in the chiral limit , @xmath174 ,
nothing distinguishes one quark flavor from another and , consequently , @xmath175 .
( in principle , there is another possibility how a chiral condensate could break the chiral symmetry , for a more detailed discussion see below and ref .
this chiral condensate breaks the chiral @xmath130 symmetry spontaneously to @xmath176 . to see this , note that the chiral condensate is still invariant under vector transformations , @xmath177 , if @xmath178 , but not under axial transformations , @xmath179 , if @xmath180 . according to goldstone s theorem
, the breaking of the global chiral symmetry leads to the occurrence of massless modes , the so - called goldstone bosons .
the number of goldstone bosons is identical to the number of broken generators . in the qcd vacuum , where the @xmath146 anomaly is present , the breaking pattern is @xmath181 , i.e. , in this case there are @xmath182 broken generators , corresponding to the generators of the broken axial symmetry @xmath183 . for @xmath184 ,
there is no global chiral symmetry that could be broken , and thus no goldstone boson . for @xmath185 ,
the goldstone bosons are the three pions , the lightest hadronic species . in nature , the pions are not completely massless , because the chiral symmetry is explicitly broken by the small , but nonzero quark mass term in the qcd lagrangian .
this turns the goldstone bosons into so - called pseudo - goldstone bosons . for @xmath186 ,
the pseudo - goldstone bosons correspond to the pseudoscalar meson octet , comprising pions , kaons , and the eta meson .
since chiral symmetry is more strongly broken by the larger strange quark mass , the pseudo - goldstone bosons carrying strangeness are heavier than the pions . for @xmath187 ,
the explicit symmetry breaking by the heavy exotic quark flavors is so strong that the would - be goldstone bosons are actually heavier than the ordinary ( i.e. non - goldstone ) non - strange bosons . in spin models ,
rotational symmetry is restored above some critical temperature and the magnetization vanishes .
the magnetization is the order parameter for this so - called ferromagnet - diamagnet phase transition . by analogy
, one expects a similar mechanism to occur in qcd , i.e. , @xmath169 to vanish above some critical temperature .
the symmetry of the ground state is then restored to the original chiral symmetry , i.e. , @xmath147 , if the @xmath146 anomaly is still present , or @xmath188 , if instantons are sufficiently screened at the transition in order to effectively restore the @xmath146 symmetry .
lattice qcd calculations show that this expectation is indeed fulfilled : there is a phase transition between the low - temperature phase where chiral symmetry is broken and the high - temperature phase where it is restored , for details see sec .
[ lattqcd ] . just like the magnetization in spin models ,
the chiral condensate @xmath169 is the order parameter for this so - called chiral phase transition . in the case of massless quarks ,
@xmath174 , based on universality arguments one can analyse the order of the chiral transition in the framework of a linear sigma model for the order parameter field @xmath169 @xcite .
this linear sigma model is an effective theory , i.e , all terms allowed by the original chiral symmetry must in principle appear , @xmath189 where the effective potential @xmath190 ^ 2 + \lambda_2 \ , { \rm tr } \left ( \phi^\dagger \phi\right)^2 - c \ , \left ( { \rm det } \phi + { \rm det } \phi^\dagger \right ) + \ldots \,\,\ ] ] determines the ground state of the theory . in eq .
( [ leff ] ) it was assumed that the first term is canonically normalized . however , since lorentz symmetry is explicitly broken in a medium at nonzero temperature , the coefficient @xmath191 in eq .
( [ leff ] ) may in general be different from one . in eq .
( [ veff ] ) , @xmath192 denote higher - dimensional operators which are irrelevant for the discussion of the order of the phase transition . for @xmath193 , the chiral symmetry of @xmath194
is @xmath147 , while for @xmath195 , it is @xmath145 .
thus , the @xmath146 anomaly is present for @xmath196 , and absent for @xmath195 .
while these chiral symmetries are manifest in the lagrangian ( [ leff ] ) , the ground state of the theory respects them only for @xmath197 and @xmath198 . for @xmath195 and @xmath199 the chiral symmetry
is spontaneously broken by a nonvanishing vacuum expectation value for the order parameter .
consequently , if the linear sigma model is to describe the chiral transition in qcd , one has to ensure that @xmath199 for @xmath195 .
there are still two possibilities how the order parameter can break the symmetry . as shown in ref .
@xcite , if @xmath200 the ground state is given by @xmath201 , while for @xmath202 the ground state is given by @xmath203 .
( the choice of the 1-direction in right- and left - handed flavor space is arbitrary . )
nature realizes the first possibility . for the case
@xmath193 , no general arguments can be made ; whether the ground state of the theory breaks chiral symmetry spontaneously depends on the particular values for the coupling constants @xmath204 and the number of flavors @xmath37 . for @xmath205 , there is no difference between the two quartic invariants in eq .
( [ veff ] ) , and one may set @xmath206 . in the presence of the @xmath146 anomaly ,
@xmath193 , there is also no chiral symmetry , and the transition is crossover , due to the linear term @xmath207 , which tilts the effective potential such that the ( thermal ) ground state @xmath208 .
if the @xmath146 anomaly is absent , @xmath195 , the effective theory for the order parameter falls in the same universality class as that of the @xmath166 heisenberg magnet , and thus the transition is of second order . for @xmath185 and in the presence of the @xmath146 anomaly ,
the chiral symmetry is @xmath209 , which is isomorphic to @xmath210 .
the effective theory for the order parameter is in the universality class of the @xmath210 heisenberg magnet .
consequently , the transition is of second order @xcite .
if the @xmath146 symmetry is effectively restored at the phase transition temperature , the symmetry group is larger , @xmath211 , which is isomorphic to @xmath212 , and the transition is of first order .
lattice qcd calculations determine the transition temperature to be @xmath213 mev @xcite . for @xmath186 , the chiral
transition is of first order , both when the @xmath146 symmetry is explicitly broken by instantons or when it is effectively restored at the transition . in the first case ,
the effective theory features a cubic invariant in the order parameter field ( the term @xmath214 ) , which drives the chiral transition first order @xcite . in the second case , the transition is fluctuation - induced of first order @xcite .
this also holds for @xmath187 , irrespective of whether the @xmath146 symmetry is explicitly broken or not . for @xmath215 ,
lattice qcd calculations find the transition temperature to be @xmath216 mev @xcite , cf .
[ translatt ] . note that nonvanishing quark masses can also be accounted for by adding a term @xmath217\ ] ] to the right - hand side of eq .
( [ leff ] ) . as discussed above ,
this term is the analogue of the term @xmath158 in spin models .
consequently , the external `` magnetic field '' @xmath218 is proportional to the quark mass matrix @xmath162 . nonvanishing quark masses lead to the term ( [ additional ] ) in the effective theory for the order parameter field
this term is linear in @xmath219 , such that the effective potential is tilted .
this may render a first or second order phase transition a crossover transition ( similar to the case @xmath205 with @xmath146 anomaly discussed in sec .
[ dynquarks ] , where a tilt in the potential is induced by the linear term @xmath207 ) .
for instance , the second - order transition for qcd with @xmath185 massless flavors is rendered crossover by a nonvanishing quark mass .
the first - order phase transition for qcd with @xmath186 massless flavors can also become crossover , if the quark masses are sufficiently large . in the real world , the up and down quark are approximately of the same mass , while the strange quark is much heavier .
it is customary to put @xmath220 and identify first - order regions , second - order lines , and crossover regions in an @xmath221 diagram , see fig .
[ quark - mass ] . to simplify the following discussion , only the case where the @xmath146 anomaly is present will be considered . the origin in fig .
[ quark - mass ] corresponds to the massless 3-flavor case , and the transition is of first order .
the upper left corner corresponds to the massless 2-flavor case , since the strange quark is infinitely heavy . here
, the transition is of second order .
the lower right corner is the case of one massless flavor .
the transition is crossover .
the upper right corner , where all quark flavors are infinitely heavy , corresponds to the pure gauge theory . at this point
the transition is of first order .
the first - order regions around the origin and the upper right corner extend somewhat into the @xmath222 plane and are bounded by critical lines where the transition is of second order . along these critical lines ,
the second - order phase transitions are in the universality class of the ising model , @xmath124 . in between the critical lines ,
the transition is crossover .
there is also a second - order phase transition line ( with a phase transition in the @xmath210 universality class ) extending downwards from the upper left corner along the @xmath223 axis .
there is a tricritical point where this line meets the second - order phase transition line bordering the first - order region around the origin @xcite .
it is an interesting question whether the real world , where @xmath224 , is still inside the first - order phase transition region or already in the crossover region .
there are ongoing lattice qcd studies to decide this question , which at present favor the latter possibility @xcite , see also sec .
[ translatt ] for more details .
so far , the quark - hadron phase transition was studied at @xmath225 .
let us finally discuss the case of nonzero quark chemical potential .
for many years , lattice qcd studies at nonzero chemical potential were hampered by numerical problems related to the so - called sign problem of the fermion determinant .
only recently there have been attempts to compute the order of the phase transition , as well as thermodynamical properties , at nonzero quark chemical potential ; for details , see sec .
[ nonzeromu ] .
so far , these calculations have been done on fairly small lattices with rather heavy quarks .
consequently , they show a crossover transition at @xmath226 . this crossover transition extends to the point @xmath227 mev , see also fig .
[ endpoint ] .
this point is a critical point where the transition is of second order .
it is in the unversality class of the ising model , i.e. , @xmath124 . for smaller temperatures and larger chemical potentials ,
the transition becomes of first order .
the critical point will move towards the @xmath2 axis when the quark masses are decreased . from the discussion in sec .
[ quarkmassdiagram ] one can not exclude the possibility that , for realistic quark masses , the first - order phase transition line emerges directly from the @xmath2 axis . finally , the question arises whether the line of first - order phase transitions extends all the way down to @xmath69 , and if so , at which point it hits the @xmath3 axis .
renormalization group arguments @xcite suggest that the behavior at zero temperature is very similar to the one at nonzero temperature : the transition is of first order for @xmath228 as well as for @xmath229 in the absence of the @xmath146 anomaly , while it could be of second order for @xmath185 in the presence of the @xmath146 anomaly . in the latter case , however , it would remain of second order along the whole phase transition line , and only the universality class would change from @xmath210 critical behavior at nonzero temperature to gaussian critical behavior at zero temperature .
since lattice qcd calculations @xcite indicate that the transition becomes of first order for temperatures below the critical point , and since it is hard to imagine that the transition switches back to second order as the temperature decreases further , this possibility can most probably be ruled out .
note , however , that if quark - gluon matter is in a color - superconducting phase to the right of the qcd phase transition line , other possibilities emerge , for details , see sec .
[ cscphases ] . in any case , model calculations @xcite within a nambu jona - lasinio ( njl ) model @xcite support the picture that the transition remains of first order below the critical point all the way down to the @xmath3 axis .
the value of @xmath3 , where the first - order phase transition line meets the @xmath3 axis , depends sensitively on the parameters used in these model calculations . its actual value should not be taken too seriously , because the njl model with quark degrees of freedom does not have a phase where matter consists of nucleons .
instead , the transition to quark matter happens at a quark chemical potential which is of the order of the ground state quark chemical potential , @xmath230 mev . since we know
that nucleonic matter exists , this behavior is clearly unphysical .
the critical point has recently received a fair amount of attention @xcite . for a second - order phase transition in the @xmath124 universality class
, there must be one massless degree of freedom . the fact that this degree of freedom is massless causes critical fluctuations .
these fluctuations were suggested to be an experimental signature for the critical point in heavy - ion collisions . which physical particle does the massless degree of freedom correspond to ?
for realistic quark masses , the pions are not massless in the vacuum , and it is unlikely that they become massless at the critical point ( the pions usually get more massive when the temperature is increased ) . moreover , since isospin is still a good symmetry at the critical point , all pions would simultaneously become massless .
then , one would have three massless modes instead of just one .
consequently , the pions can not assume the role of the massless mode .
since the critical point exists even when considering only @xmath185 quark flavors and since , for @xmath185 , there is only one other degree of freedom in the effective theory besides the pions , it must be this degree of freedom that becomes massless : the scalar @xmath231 meson @xcite . in the vacuum , this meson has a mass of about @xmath232 mev @xcite . if it becomes massless , it can be copiously produced .
when these @xmath231 mesons decouple from the collision region , they assume their vacuum masses and rapidly decay into pions . besides critical fluctuations
, another signature for the critical point would thus be the late emission of a large amount of pions in a heavy - ion collision .
there are other phases in the phase diagram of nuclear matter , which have recently received much attention in the literature , the so - called color - superconducting phases in sufficiently cold and dense quark matter @xcite .
color superconductivity occurs , because there is an attractive interaction between two quarks at the fermi surface @xcite .
then , by cooper s theorem , these quarks form cooper pairs which condense in the new ground state of the system . at least at
asymptotically large quark chemical potentials , the existence of an attractive interaction between quarks at the fermi surface , and thus the existence of color superconductivity , can be rigorously proven .
due to asymptotic freedom @xcite , when @xmath233 , the strong coupling constant of qcd , evaluated at the scale @xmath3 , becomes small , @xmath234 , such that the dominant interaction between quarks is given by single - gluon exchange .
the scattering amplitude for single - gluon exchange in an @xmath46 $ ] gauge theory is proportional to @xmath235 where @xmath236 are the fundamental colors of the two quarks in the incoming channel , and @xmath237 their respective colors in the outgoing channel . under the exchange of the color indices of either the incoming or the outgoing quarks the first term is antisymmetric , while the second term is symmetric . in group
theoretical language , for @xmath238 $ ] eq . ( [ tata ] ) represents the coupling of two fundamental color triplets to an ( antisymmetric ) color antitriplet and a ( symmetric ) color sextet , @xmath239^c \times [ { \bf 3 } ] ^c = [ { \bf \bar{3 } } ] _
a^c + [ { \bf 6 } ] _
s^c\,\ , .\ ] ] the minus sign in front of the antitriplet contribution in eq .
( [ tata ] ) signifies the fact that this channel is attractive , while the sextet channel is repulsive .
therefore , one expects that quark cooper pairs condense in the color - antitriplet channel .
this argument holds rigorously at asymptotically large densities .
the highest densities of nuclear matter that can be achieved in the laboratory through heavy - ion collisions , or that occur in nature in the interior of neutron stars , are of the order of ten times nuclear matter ground state density . at these densities ,
the quark chemical potential is still fairly small , @xmath240 gev . for phenomenology
it is therefore important to answer the question whether color superconductivity also exists at the ( comparatively moderate ) densities occurring in nature .
there is no rigorous way to answer this question , as an extrapolation of the above asymptotic argument becomes unreliable when @xmath241 .
nevertheless , calculations in the framework of the njl model @xcite show that color superconductivity does seem to occur also at moderate densities . in this case , the attractive interaction could be mediated by instanton ( instead of single - gluon ) exchange .
color superconductivity is a much more complicated phenomenon than ordinary superconductivity . from a very qualitative point of view , in comparison to electrons , quarks carry additional quantum numbers such as color and flavor .
the wave function of a cooper pair has to be antisymmetric under the exchange of the two fermions forming the pair .
consequently , the possible color and flavor representations of the two - fermion state have to be chosen in a way which respects this antisymmetry .
this requirement helps to classify all possible color - superconducting condensates @xcite .
this classification will be presented in the following .
the baryon density in heavy - ion collisions or in neutron stars is at most of the order of ten times the nuclear ground state density .
therefore , the quark chemical potential is unlikely to assume values beyond @xmath242 gev . at zero temperature
, one can only have fermions if the fermi energy @xmath3 exceeds their mass .
thus , only the three lightest quark flavors , up , down , and strange , play any role ; charm , bottom , and top quarks are too heavy . for small temperatures , these heavy flavors can be thermally excited , but their abundance is exponentially suppressed @xmath243 , cf . sec .
[ qualarg ] .
therefore , they will be excluded from the following consideration . for @xmath242
gev , up and down quarks can be treated as truly ultrarelativistic particles , as @xmath244 . to first approximation , also the strange quark will be considered to be massless .
corrections due to the strange quark mass can be treated perturbatively , the correction factor being of order @xmath245 @xcite . for ultrarelativistic particles , spin @xmath246 and angular momentum @xmath55 are not separately good quantum numbers , only the total spin @xmath247 is .
therefore , possible cooper pair wave functions should be classified according to their total spin @xmath248 .
let us first focus on the spin - zero channel , @xmath249 .
the @xmath249 representation of the @xmath250 spin group is totally antisymmetric .
therefore , the remaining color and flavor part of the cooper pair wave function has to be symmetric under the simultaneous exchange of color and flavor indices in order to fulfill the requirement of overall antisymmetry .
if one assumes that quarks pair in the antisymmetric color - antitriplet channel , one has no choice but to also chose an antisymmetric flavor representation . for @xmath251 flavor ,
this is impossible , as there is no flavor symmetry group with an antisymmetric representation .
one therefore has to consider at least @xmath185 quark flavors ( for instance , up and down ) , where the most simple representation is the antisymmetric flavor singlet @xmath252_a^f$ ] representation of the @xmath156 flavor group .
therefore , the most simple @xmath249 quark cooper pair condensate has the form @xmath253 here , @xmath254 are the color indices of the quarks forming the cooper pair , while @xmath255 are the corresponding flavor indices .
the two totally antisymmetric tensors on the right - hand side ensure that the condensate belongs to the @xmath256^c_a$ ] in color , as well as the @xmath252^f_a$ ] representation in flavor space .
the color - superconducting phase represented by the condensate ( [ 2sc ] ) is commonly called the `` 2sc '' phase ( for `` @xmath257-flavor color * s*uper*c*onductor '' ) .
condensation of quark cooper pairs occurs if the quantity @xmath258 on the right - hand side of eq .
( [ 2sc ] ) is nonzero , @xmath259 .
thus , the quark cooper pair condensate carries a fundamental color index @xmath260 .
this indicates that the local @xmath238 $ ] color symmetry is spontaneously broken by the quark cooper pair condensate , similarly to the spontaneous breaking of the global chiral symmetry @xmath147 by the chiral condensate ( [ chiralcond ] ) in the qcd vacuum discussed in sec .
[ dynquarks ] . in this sense
, @xmath258 is the order parameter for color superconductivity .
it is zero in the phase of unbroken @xmath238 $ ] symmetry , and nonzero in the broken phase where condensation of quark cooper pairs occurs .
of course , a local symmetry can never be truly broken spontaneously @xcite . however , after fixing the gauge , spontaneous breaking does occur , just like in ordinary superconductivity or in the standard model of electroweak interactions . in ordinary superconductivity ,
the condensation of electron cooper pairs breaks the @xmath261 $ ] gauge symmetry of electromagnetism , while in the standard model , the higgs field breaks @xmath262 \times [ u(1)_y]$ ] to @xmath261 $ ] .
analogously , the quark cooper pair condensate ( [ 2sc ] ) breaks the @xmath238 $ ] color gauge symmetry .
since quarks carry baryon and electric charge , the cooper pair condensate ( [ 2sc ] ) in principle also breaks the global @xmath144 of baryon number conservation and the local @xmath261 $ ] of electromagnetism . in the discussion of chiral symmetry breaking ,
these symmetries were never broken because the chiral condensate consists of a quark and an antiquark which carry opposite baryon and electric charge .
the chiral condensate is thus a singlet under @xmath144 and @xmath261 $ ] and consequently preserves these symmetries .
this is different for a color - superconducting condensate which consists of two quarks .
it turns out , however , that there exists a `` rotated '' baryon number @xmath263 symmetry and a `` rotated '' electromagnetic @xmath264 $ ] symmetry , which are formed from the original baryon number and electromagnetic symmetries and the eighth generator of @xmath238 $ ] @xcite .
this is similar to electroweak symmetry breaking , where the @xmath261 $ ] symmetry of electromagnetism is a combination of the @xmath265 $ ] hypercharge symmetry and the third generator of @xmath262 $ ] .
thus , the @xmath144 and @xmath261 $ ] symmetries are not really broken in the 2sc phase , but `` rotated '' .
the rotation angle is the analogue of the weinberg angle in the standard model of electroweak interactions ; for more details , see sec .
[ gluephoton ] .
spontaneous symmetry breaking in gauge theories does not lead to goldstone bosons .
rather , what would have been a goldstone mode will be `` eaten '' by a gauge boson which in turn becomes massive and thus acquires an additional longitudinal degree of freedom .
there are as many massive gauge bosons as there would have been goldstone modes due to spontaneous symmetry breaking . in ordinary superconductors ,
the electromagnetic @xmath261 $ ] symmetry is broken , which has only one generator .
consequently , there is a single goldstone mode which is `` eaten '' by the single gauge boson present in this case , the photon .
the photon acquires a so - called meissner mass .
what happens physically is that magnetic fields are damped on length scales of the order of the inverse meissner mass , which in turn leads to the meissner effect , the expulsion of magnetic flux from the superconductor . in the standard model of electroweak interactions , @xmath262 \times [ u(1)_y]$ ]
is broken to @xmath261 $ ] , i.e. , there are three goldstone modes which in turn lead to the massive gauge bosons of the weak interaction , @xmath266 and @xmath267 .
the photon is massless .
this is required , since it is the gauge boson of the residual @xmath261 $ ] symmetry of electromagnetism .
analogously , in a color superconductor one expects some of the gluons to become massive .
exactly how many gluons acquire a mass depends on the pattern of symmetry breaking . for the condensate ( [ 2sc ] )
, one can clarify this via the following argument . by a global color rotation
, one can always orient the order parameter @xmath258 to point in the 3-direction in color space ( more precisely , the _ anti_-3-direction , as condensation occurs in the color _ anti_-triplet channel ) , @xmath268 physically , this means that if we call color 1 red , color 2 green , and color 3 blue , red up ( or down ) quarks and green down ( or up ) quarks condense to form an anti - blue cooper pair condensate .
blue up and down quarks do not participate in condensation .
the condensate ( [ 2scorient ] ) does not break the color @xmath238 $ ] gauge symmetry completely .
the residual symmetry is a local @xmath269 $ ] symmetry in the space of the first two colors ( in our conventions , red and green ) . including electromagnetism , the symmetry breaking pattern for the condensate ( [ 2scorient ] ) is therefore @xmath238 \times [ u(1)_{\rm em } ] \rightarrow [ su(2)_c ] \times [ \tilde{u}(1)]$ ] . consequently , there are @xmath270 broken generators , which lead to five massive gluons .
the remaining three gluons must remain massless as they correspond to the gauge bosons of the residual local @xmath269 $ ] symmetry .
this is also borne out by an explicit calculation of the gluon meissner masses in the 2sc phase @xcite .
the gauge boson of the local @xmath264 $ ] symmetry ( the `` rotated '' photon ) is also massless . for more details , see sec .
[ gluephoton ] . for @xmath215 flavors ,
condensation of quark cooper pairs becomes considerably more interesting .
first , to preserve the antisymmetry of the cooper pair wave function the two quarks have to be in the @xmath256_a^f$ ] representation of the global @xmath271 flavor symmetry .
consequently , the quark cooper pair condensate has the form @xmath272 the difference to eq .
( [ 2sc ] ) is that , to ensure antisymmetry in flavor space , one is required to use the totally antisymmetric tensor of rank 3 , @xmath273 , rather than its rank-2 counterpart .
consequently , an additional flavor index @xmath274 appears in the order parameter , @xmath275 .
a nonvanishing order parameter automatically implies that not only the local @xmath238 $ ] color , but also the global @xmath271 flavor symmetry is broken . the situation is not unlike the one encountered in superfluid helium-3 @xcite .
superfluid helium-3 forms cooper pairs with spin @xmath276 and angular momentum @xmath277 .
( both spin and angular momentum are good quantum numbers , as helium-3 is a non - relativistic system . )
consequently , the order parameter breaks the global @xmath278 of spin as well as the global @xmath279 of angular momentum .
this breaking can occur in many possible ways , giving rise to a plethora of phases in superfluid helium-3 .
similarly , one would expect many different phases to occur in a 3-flavor color superconductor .
however , in fact there are only two possibilities , one of which is likely to be realized in nature @xcite . to see this ,
note the formal similarity between the order parameter @xmath280 and the one encountered in chiral symmetry breaking , @xmath169 , in sec .
[ dynquarks ] . while @xmath169 transforms under @xmath147 ( in the presence of the @xmath146 anomaly ) , @xmath280 transforms under @xmath238 \times su(3)_v$ ] .
consequently , the effective lagrangian for @xmath280 is of the same form ( [ leff ] ) as for @xmath169 .
the two possible patterns of symmetry breaking occurring in such an effective theory were already mentioned above . if the coupling constant @xmath200 , the order parameter assumes the form @xmath281 in contrast to the 2sc case , where blue quarks remained unpaired , now all quark colors and flavors participate in the pairing process .
the order parameter ( [ cfl ] ) is similar to the one for chiral symmetry breaking , where @xmath282 .
( in the ground state , @xmath283 . )
similar to the chiral symmetry breaking pattern @xmath181 , @xmath284 , the condensate ( [ cfl ] ) breaks @xmath238 \times su(3)_v$ ] to the vectorial subgroup @xmath285 .
the condensate is still invariant under vector transformations in color and flavor space , or in other words , a transformation in color requires a simultaneous transformation in flavor to preserve the invariance of the condensate .
therefore , the discoverers @xcite of this 3-flavor color - superconducting state termed it the `` * c*olor-*f*lavor-*l*ocked '' , or short , `` cfl '' state .
the residual @xmath285 symmetry is no longer a local symmetry , so that there are no massless gluons . the complete pattern of symmetry breaking , including the @xmath144 symmetry of baryon number and the @xmath261 $ ] symmetry of electromagnetism is @xmath238 \times su(3)_v \times u(1)_v \times [ u(1)_{\rm em } ] \rightarrow su(3)_{c+v } \times [ \tilde{u}(1)]$ ] .
( this notation is slightly ambiguous : the @xmath261 $ ] symmetry is generated by the quark charge operator @xmath286 which is traceless .
thus , the _ global _ part of the @xmath261 $ ] symmetry is actually a subgroup of the _ global _ @xmath271 flavor symmetry .
however , the _ local _ part of @xmath261 $ ] is not .
therefore , here and in the following i choose to explicitly denote the @xmath261 $ ] symmetry group . ) in the cfl case , unlike the 2sc case , baryon number is broken , but a rotated electromagnetic @xmath264 $ ] is again preserved .
the symmetry breaking pattern leads to nine goldstone bosons , eight of which are `` eaten '' by the gluons , i.e. , all gluons acquire a meissner mass .
there is one goldstone boson from the breaking of the @xmath144 symmetry . in the chiral limit ,
the flavor symmetry of qcd is actually not just @xmath271 but @xmath287 . assuming that also the @xmath146 symmetry of qcd is effectively restored at large quark densities , quark cooper pair condensation of the form ( [ cfl ] ) breaks @xmath238 \times su(3)_r \times su(3)_\ell \times u(1)_v \times u(1)_a \times [ u(1)_{\rm em}]$ ] to @xmath288 $ ] , i.e. , not only the local color , but also the global chiral symmetry is broken .
in addition to the eight massive gluons , there are also ten real goldstone bosons , eight from the breaking of the @xmath289 chiral symmetry , and one each from the breaking of @xmath144 and @xmath146 .
closer inspection @xcite shows that the excitation spectrum in the cfl state bears a striking resemblance to the one in the hadronic phase .
let us first focus on the fermionic sector . in the cfl phase , there are nine gapped fermionic quasiparticles ( cf .
also sec .
[ secexspec ] ) , eight of which are degenerate in mass .
these correspond to the baryon octet in the qcd vacuum .
( for this argument we have to assume that there is no explicit @xmath271 flavor - symmetry breaking in the qcd vacuum . )
the ninth quasiparticle is twice as heavy and does not have a counterpart in hadronic matter , but this is not a reason to worry , as such a particle would have a large decay width into lighter particles . in the bosonic sector , there are nine goldstone bosons from the breaking of the axial @xmath290 symmetry , which correspond to the pseudoscalar nonet in the hadronic phase .
only the tenth goldstone boson from the breaking of @xmath144 does not have a counterpart in the qcd vacuum .
such a boson exists , however , in dense nuclear matter , where a superfluid @xmath291 condensate may form , which also breaks the @xmath144 baryon number symmetry .
the preceding arguments have led to the conjecture of `` continuity '' between hadron and quark matter @xcite .
this conjecture states that , since there is no difference in symmetry between quark matter in the cfl state and ( @xmath271 flavor - symmetric ) hadronic matter , there need not be any phase boundary between these two phases at all .
of course , this requires that there is no other color - superconducting phase , for instance the 2sc state , which separates cfl matter from hadronic matter . for @xmath202 ,
the order parameter is given by @xmath292 where the 3-direction is arbitrary .
this condensate breaks @xmath238 \times su(3)_v$ ] to @xmath269 \times su(2)_v$ ] . in this case ,
@xmath269 $ ] is still a local symmetry .
consequently , like in the 2sc case only five gluons become massive , and blue up , down , and strange quarks do not participate in the formation of cooper pairs . additionally , there are also five goldstone bosons from the breaking of the flavor symmetry .
there is a more technical and a more physical argument , why the cfl state is most likely realized in nature . from the more technical point of view
, one can show @xcite that , to one - loop order , in qcd @xmath293 and @xmath200 . from the more physical point of view
, the cfl state is energetically favored because _ all _ quark colors and flavors ( instead of just a few ) acquire a gap at the fermi surface .
the gain in condensation energy is thus expected to be larger than for a state with a condensate of the form ( [ 3sc ] ) .
although a single quark flavor can not form cooper pairs with total spin @xmath249 , it can pair in the @xmath294 channel .
( an exhaustive discussion of possible pairing channels for a single quark flavor is given in ref .
this channel corresponds to the symmetric @xmath295_s^j$ ] representation of the @xmath250 spin group .
if one still assumes pairing to occur in the color @xmath256_a^c$ ] channel , the cooper pair wave function is , as required , overall antisymmetric . the condensate is a 3-vector in space which points in the direction of the spin of the cooper pair .
it has the form @xcite @xmath296 where @xmath297 denotes the spatial component of the spin vector .
condensation breaks the local color @xmath238 $ ] symmetry and the global @xmath298 spin symmetry .
this is similar to superfluid helium-3 where condensation breaks @xmath299 ( see discussion above ) .
while many different phases arise , let us just mention two which are quite similar to the ones discussed in the context of three and two quark flavors , the so - called `` * c*olor-*s*pin-*l*ocked '' or csl phase , where the order parameter assumes the form @xmath300 and the so - called polar phase , where @xmath301 in the csl phase , the order parameter ( [ csl ] ) is strikingly similar to the one in the cfl phase , cf .
( [ cfl ] ) .
all quark colors participate in the formation of cooper pairs .
also the symmetry breaking pattern is similar , @xmath238 \times so(3)_j \times u(1)_v \times [ u(1)_{\rm em } ] \rightarrow so(3)_{c+j}$ ] .
the main difference is that now there is no rotated electromagnetism @xmath264 $ ] , cf .
[ gluephoton ] .
consequently , all eight gluons _ and _ the photon become massive @xcite . in the polar phase , the order parameter resembles that of eq .
( [ 3sc ] ) . neglecting electromagnetic interactions for the moment ,
the symmetry breaking pattern is @xmath238 \times so(3)_j \times u(1)_v \rightarrow [ su(2)_c ] \times so(2)_j \times \tilde{u}(1)$ ] . like in the 2sc phase ,
the residual @xmath269 $ ] color symmetry is a local symmetry , and there are three massless and five massive gluons .
the breaking of the rotational @xmath298 symmetry to @xmath302 also leads to two real goldstone bosons .
baryon number is not broken , but merely rotated . including the @xmath261 $ ] of electromagnetism there is a small subtlety which is explained in more detail in sec .
[ gluephoton ] : if there is only a single flavor present , or if all flavors carry the same electric charge , a rotated electromagnetic @xmath264 $ ] symmetry exists . if there are at least two flavors which differ in charge , the @xmath261 $ ] symmetry is broken .
table [ table1 ] summarizes the results of this section for the 2sc , the cfl , the csl , and the polar phase of color - superconducting quark matter .
.color - superconducting phases in dense quark matter . for the polar phase
, there is an additional @xmath264 $ ] symmetry , if all flavors in the system carry the same electric charge . [ cols="<,^,^,^ " , ] the final step is to diagonalize the mass matrices @xmath303 , @xmath304 for electric and magnetic gluons . since
only the eighth gluon mixes with the photon , this diagonalization is achieved by a simple orthogonal rotation in the @xmath305 block corresponding to the indices @xmath306 .
the resulting diagonal ( squared ) debye masses @xmath307 and ( squared ) meissner masses @xmath308 , as well as the ( square of the ) cosine of the rotation angles @xmath309 , @xmath310 are shown in table [ tableunmix ] . in the case of an unbroken @xmath264 $ ] symmetry , cf .
table [ table1 ] , the `` rotated '' ( magnetic ) photon is massless .
the `` rotated '' gluon remains massive , but its degeneracy with the other massive gluons is lifted .
the case of the polar phase is special . if there is only one quark flavor , or all quark flavors have the same electric charge , the results shown in tables [ debyemass ] [ tableunmix ] hold . in this case , the rotated photon is massless , and there is indeed an unbroken @xmath264 $ ] symmetry .
however , in the case of two or more quark flavors with different electric charges , the results change @xcite .
let us assume that the chemical potentials of all quark flavors are identical .
then the ( squared ) gluon masses are the same , but in the mixed masses , the factor @xmath311 has to be replaced by @xmath312 , while in the photon masses , the factor @xmath313 is replaced by @xmath314 . in this case , it is not hard to realize that a diagonalization of the meissner mass matrix does not lead to a massless rotated photon . there is therefore no @xmath264 $ ] symmetry .
consequently , spin - one color - superconducting quark matter exhibits a meissner effect , while color superconductors with spin - zero cooper pairs do not .
due to the smallness of the ratio @xmath315 for spin - one color superconductors , these are very likely of type i , i.e. , the magnetic field is completely expelled .
this is in contrast to the standard model of a neutron star , where the core is assumed to be a type - ii superconductor and thus threaded by magnetic flux tubes .
it was recently argued in ref .
@xcite that the short precession period of some pulsars contradicts this assumption and requires the core of the pulsar to be a type - i superconductor .
the question then is whether the core could be made of spin - one color - superconducting quark matter @xcite .
in this review , i have presented the current knowledge of the equilibrium properties of strongly interacting matter at large temperatures and/or densities . in particular , i have qualitatively discussed the phase diagram .
i have presented calculations of thermodynamic properties of strongly interacting matter , both via lattice qcd , as well as within analytic approaches .
finally , i have given an overview of color superconductivity in weak coupling .
our knowledge of the qcd phase transition and the qgp at zero quark chemical potential has tremendously increased over the last few years .
lattice qcd calculations are well under control for the pure @xmath238 $ ] gauge theory , and the quality of the data is such that an extrapolation to the continuum limit as well as to the thermodynamic limit has become possible .
lattice calculations with dynamical fermions are more challenging and , consequently , the data are not of the same quality as for the pure gauge theory .
the main problem is that , with present methods of putting fermions on the lattice , the pion comes out too heavy .
since pions dominate the equation of state in the hadronic phase , calculations of the pressure below the chiral restoration temperature do not yet reflect the correct physics .
the challenge for the future is to improve the methods such that the pion mass on the lattice is close to the value in nature . besides a reliable computation of the equation of state
, this will also allow to decide the question about the order of the qcd phase transition in nature . another important development in lattice
qcd is to extend the investigation of thermodynamic properties to nonzero quark chemical potentials . for many years
, the fermion sign problem has impeded progress in this direction .
recent attempts , like multiparameter reweighting , taylor expansion around @xmath225 , or analytic continuation from imaginary values of @xmath3 , have made an attempt to work around this problem .
much work remains to be done to improve these methods in order to correctly determine the location of the critical point in the @xmath64 plane .
this is of great phenomenological importance : in order to find a signal for the first - order phase transition to the qgp in nuclear collisions , one has to tune the bombarding energy such that one probes the region of the phase diagram , which is to the right of the critical point .
nuclear collisions at very high energies most likely probe the region to the left , i.e. , the crossover region of the quark - hadron transition . by definition
, there is no qualitative difference between hadronic and qgp phase in this region , and a clear signal for the qgp will be hard to identify .
although the question about the location of the critical point is important and can be investigated with the above mentioned lattice methods , these methods circumvent the fermion sign problem rather than solving it . moreover , they are only applicable for quark chemical potentials
@xmath3 from zero up to values of order @xmath2 . therefore , one will ultimately have to find a true solution which also works at small temperatures and large chemical potential , so that the color - superconducting quark matter phase can be explored .
analytic approaches to compute the equation of state of strongly interacting matter at high temperature have advanced rapidly in recent years .
the equation of state is now known to all orders which are perturbatively computable .
work is in progress to determine the remaining nonperturbative contribution of order @xmath15 .
resummation techniques have been applied to compute the thermodynamic properties of strongly interacting matter . at large temperatures
@xmath316 , they suggest that the qgp is a gas of weakly interacting quasiparticles . however , when approaching the critical temperature from above , the approaches based on resummation techniques fail to describe lattice qcd data . at the moment , one seems to be forced to either abandon field - theoretical rigor in favor of simple quasiparticle models with sufficiently many fit parameters to reproduce the data , or turn to an alternative description , such as the polyakov loop model , which is physically less intuitive .
it remains to be shown how this model is related to the quasiparticle picture at large temperatures .
color superconductivity is a rapidly evolving field .
it is fairly likely that color - superconducting quark matter can be found in the core of compact stellar objects .
it remains to explore how this phase influences the properties of the star .
much work has still to be done , for instance to compute the transport properties of color - superconducting matter and the phase diagram under the conditions of electric and color neutrality .
although njl - type models may give a qualitative picture of possible scenarios , they are unreliable when one wants to draw quantitative conclusions .
the task is to improve existing weak - coupling calculations or apply nonperturbative techniques to obtain further knowledge about this interesting , exotic phase of strongly interacting matter .
i would like to thank a. dumitru , z. fodor , s. hands , o. kaczmarek , f. karsch , e. laermann , p. levai , r. pisarski , a. rebhan , t. schfer , c. schmidt , a. schmitt , y. schrder , and a. steidl for valuable discussions and for help in preparing the figures .
i greatly appreciate the hospitality of the int seattle , where part of this work was done . 99 -2pt d.j .
gross and f. wilczek , ; h.d .
politzer , j.c .
collins and m.j .
perry , see , for instance : proc . of the 15th int . conf . on ultra - relativistic nucleus - nucleus collisions ,
_ quark matter 2001 _ ( eds .
hallman , d.e .
kharzeev , j.t .
mitchell , and t.s .
ullrich ) , l. van hove , ; s. pratt , ; m. kataja , v. ruuskanen , l.d .
mclerran , and h. von gersdorff , ; c.m . hung and e.v .
shuryak , ; d.h .
rischke and m. gyulassy , , for a review , see for instance : a. kovner and u.a .
wiedemann , hep - ph/0304151 , to appear in : _ quark - gluon plasma 3 _ , eds .
r.c hwa and x.n .
wang ( world scientific , singapore , 2003 ) l.d .
mclerran ad t. toimela , p. arnold , g.d .
moore and l.g .
yaffe , , , j. kapusta , p. lichard , and d. seibert , , erratum see , for instance : eos collaboration , , , ; indra collaboration ( j.d .
frankland et al . ) , ; s. das gupta , a.z .
mekjian , and m.b .
tsang , nucl - th/0009033 ( submitted to _ adv .
kapusta , _ finite temperature field theory _
( cambridge university press , cambridge , 1983 ) a.d .
linde , d.j .
gross , r.d .
pisarski , and l.g .
yaffe , m. le bellac , _ thermal field theory _
( cambridge university press , cambridge , 1996 ) b. mller , _ the physics of the quark - gluon plasma _ , ( lecture notes in physics vol .
225 , springer , berlin , 1985 ) ; l.d .
mclerran , ; h. stcker and w. greiner , f. karsch , e. laermann and o. philipsen , hep - ph/0303042 j .-
blaizot , e. iancu , and a. rebhan , hep - ph/0303185 , to appear in : _ quark - gluon plasma 3 _ , eds .
r.c hwa and x.n .
wang ( world scientific , singapore , 2003 ) k. rajagopal and f. wilczek , hep - ph/0011333 ; m. alford , d. litim and c. manuel , t. schfer and e.v .
shuryak , j.w .
harris and b. mller , ; s.a .
bass , m. gyulassy , h. stcker , and w. greiner , t. schfer , hep - ph/0304281 b.d .
serot and j.d .
walecka , l.d .
mclerran and b. svetitsky , , l.g . yaffe and b. svetitsky ,
see , for instance : s.j .
knak jensen and o.g .
mouritsen , g. t hooft , a.j .
paterson , ; r.d .
pisarski and d.l .
stein , r.d .
pisarski and f. wilczek , s. gavin , a. gocksch , r.d .
pisarski , s.d.h .
hsu and m. schwetz , z. fodor and s.d .
katz , , o. scavenius , a. mocsy , i.n .
mishustin , and d.h .
rischke , y. nambu and g. jona - lasinio , m.a .
stephanov , k. rajagopal , and e.v .
shuryak , , review of particle physics by particle data group , b.c .
barrois , d. bailin and a. love , m.g .
alford , k. rajagopal , and f. wilczek , ; r. rapp , t. schfer , e. shuryak , and m. velkovsky , r.d .
pisarski and d.h .
rischke , r.d .
pisarski and d.h .
rischke , s. elitzur , t. schfer and f. wilczek , g.w .
carter and d. diakonov , d.h .
rischke , d. vollhardt and p. wlfle , _ the superfluid phases of helium 3 _ ( taylor & francis , london , 1990 ) m.g .
alford , k. rajagopal , and f. wilczek , t. schfer and f. wilczek , m.g .
alford , j.a .
bowers , j.m .
cheyne , and g.a .
cowan , t. schfer , a. schmitt , q. wang , and d.h .
rischke , a. schmitt , q. wang , and d.h .
rischke , nucl - th/0301090 m.g .
alford , j. berges , and k. rajagopal , p. bedaque , i. shovkovy and m. huang , ; e. goubankova , w.v .
liu , and f. wilczek , m.g .
alford , j.a .
bowers , and k. rajagopal , h. mther and a. sedrakian , w.v .
liu and f. wilczek , w.e .
brown , j.t .
liu , and h .- c .
ren , , see , for instance : m. creutz , _ quarks , gluons and lattices _ ( cambridge university press , cambridge , 1983 ) k. wilson , k. symanzik , ; b. beinlich , f. karsch , and e. laermann , p. hasenfratz and f. niedermayer , j. kogut and l. susskind , r. narayanan and h. neuberger , , d.b .
kaplan , p. hasenfratz , s. hauswirth , k. holland , t. jorg , f. niedermayer , and u. wenger , c. gattringer , f. karsch and p. hasenfratz , r.d .
pisarski , ; a. dumitru and r.d .
pisarski , , , f. karsch , e. laermann , and a. peikert , c. schmidt , c.r .
allton , s. ejiri , s.j .
hands , o. kaczmarek , f. karsch , and e. laermann , hep - lat/0209009 f.r .
brown , f.p .
butler , h. chen , n.h .
christ , z. dong , w. schaffer , l.i .
unger , and a. vaccarino , jlqcd collaboration ( s. aoki et al . ) , f. karsch , e. laermann , and c. schmidt , f. karsch and c. schmidt , private communication f. karsch , e. laermann and a. peikert , g. boyd , j. engels , f. karsch , e. laermann , c. legeland , m. ltgemeier , and b. petersson , , o. kaczmarek , f. karsch , e. laermann , and m. ltgemeier , o. kaczmarek , f. karsch , and p. petreczky , and f. zantow , , hep - lat/0301015 f. karsch , p. petreczky , s. stickan , and i. wetzorke , ceres collaboration , , m. asakawa , t. hatsuda , and y. nakahara , , g. aarts and j.m .
martinez resco , , c.r .
allton , s. ejiri , s.j .
hands , o. kaczmarek , f. karsch , e. laermann , c. schmidt , and l. scorzato , p. de forcrand and o. philipsen , ; m. delia , m .-
lombardo , i.m .
barbour , s.e .
morrison , e.g. klepfish , j.b .
kogut , and m .-
lombardo , c.n .
yang and t.d .
lee , z. fodor , s.d .
katz , and k.k .
szabo , hep - lat/0208078 s. chandrasekharan and u.j .
wiese , ; m.g .
alford , s. chandrasekharan , j. cox , and u.j .
wiese , s.j .
hands and d.n .
walters , b. freedman and l. mclerran , e. shuryak , j. kapusta , t. toimela , p. arnold and c. zhai , , c. zhai and b. kastening , e. braaten and a. nieto , t. toimela , , erratum k. kajantie , m. laine , k. rummukainen , and y. schrder , , , , b. kastening , ; t. hatsuda , r.r .
parwani , , j. letessier and j. rafelski , p. ginsparg , ; t. appelquist and r.d .
pisarski , ; s. nadkarni , , e. braaten and a. nieto , , see , for instance : p. arnold and l.g .
yaffe , , ; d. bdeker , , ; p. arnold , g.d .
moore , and l.g .
yaffe , , a. vuorinen , hep - ph/0305183 d.h .
rischke , m.i .
gorenstein , a. schfer , h. stcker , and w. greiner , d.h .
rischke , j. schaffner , m.i .
gorenstein , a. schfer , h. stcker , and w. greiner , a. chodos , r.l .
jaffe , k. johnson , c.b .
thorn , and v.f .
weisskopf , a.l .
fetter and j.d .
walecka , _ quantum theory of many particle systems _ ( mcgraw - hill , new york , 1971 ) v. goloviznin and h. satz , w. greiner and d.h .
rischke , a. peshier , b. kmpfer , o.p .
pavlenko , and g. soff , p. levai and u.w .
heinz , r.a .
schneider and w. weise , ; t. renk , r.a .
schneider and w. weise , a. peshier , b. kmpfer , and g. soff , , ; k.k . szabo and a.i .
toth , ; a. rebhan and p. romatschke , j. fingberg , u.m .
heller , and f. karsch , e. braaten and r.d .
pisarski , , ; j. frenkel and j.c .
taylor , , t. altherr and u. kraemmer , ; j .-
blaizot and j .- y .
ollitrault , ; h. vija and m.h .
thoma , ; c. manuel , .
d.h . rischke and i.a .
shovkovy , u.w .
heinz , k. kajantie , and t. toimela , h.a .
weldon , r.d .
pisarski , h.a .
weldon , , ; g. baym , j .-
blaizot , and b. svetitsky , e. braaten and r.d .
pisarski , j.m .
cornwall , r. jackiw , and e. tomboulis , g. baym , b. vanderheyden and g. baym , a. arrizabalaga and j. smit , r.e .
norton and j.m .
cornwall , h. van hees and j. knoll , , j .-
blaizot , e. iancu , and a. rebhan , , , a. peshier , j.o .
andersen , e. braaten , and m. strickland , , ; j.o .
andersen , e. braaten , e. petitgirard , and m. strickland , cp - pacs collaboration ( m. okamoto et al . ) , p.m. stevenson , ; a. okopiska , ; f. karsch , a. patks , and p. petreczky , ; s. chiku and hatsuda , j.p .
blaizot , e. iancu , and a. rebhan , r. baier and k. redlich , b. mller , n. petropoulos , j.t .
lenaghan and d.h .
rischke , j.t .
lenaghan , j. schaffner - bielich , and d.h .
rischke , d. rder , j. ruppert , and d.h .
rischke , b. alles , m. delia , and a. di giacomo , j.t .
lenaghan , j. baacke and s. michalski , j. ruppert and d.h .
rischke , ( work in progress ) c. manuel , t. fugleberg , r.d .
pisarski and d.h .
rischke , c. manuel , q. wang and d.h .
rischke , d. boyanovsky and h.j .
de vega , d.h .
rischke , d.t .
son , ; t. schfer and f. wilczek , ; d.k .
hong , v.a .
miransky , i.a .
shovkovy , and l.c.r .
wijewardhana , m. prakash , j.m .
lattimer , j.a .
pons , a.w .
steiner , and s. reddy , d.h .
rischke , a. schmitt , q. wang , and d.h .
rischke , ( in preparation ) d.h .
rischke , d.h .
rischke , d.t .
son , and m.a .
stephanov , b. link , astro - ph/0302441 | our current knowledge of the quark - gluon plasma in thermodynamical equilibrium is reviewed .
the phase diagram of strongly interacting matter is discussed , with emphasis on the quark - hadron phase transition and the color - superconducting phases of quark matter .
lattice qcd results on the order of the phase transition , the thermodynamical functions , the heavy quark free energy , mesonic spectral functions , and recent results for nonzero quark chemical potential are presented . analytic attempts to compute the thermodynamical properties of strongly interacting matter , such as perturbation theory , quasiparticle models ,
`` hard - thermal - loop '' ( htl)-resummed perturbation theory , the polyakov - loop model , as well as linear sigma models are discussed .
finally , color - superconducting quark matter is considered in the limit of weak coupling . the gap equation and the excitation spectrum
are derived .
the solution of the gap equation , gap parameters in various color - superconducting phases , and critical temperatures for the transition to normal - conducting quark matter are presented .
a summary of gluon and photon properties in color superconductors is given . |
understanding the way in which galaxies form is a key challenge facing cosmologists today .
this can be approached in different ways , of which the most important can be the study of the clustering of galaxies and the bound systems defined by their spatial distribution , and the statistical measures of the properties of galaxies and how these vary with their environment .
there have been several attempts to characterise galaxy properties as a function of local density @xcite . in particular ,
studies of the variation of the luminosity function include the luminosity function of wall and void galaxies in the 2dfgrs @xcite , in the sdss @xcite , and theoretical variations in the luminosity function as a function of local galaxy density @xcite .
in this paper we concentrate on the study of the variations of the luminosity function with local density , which we characterise by the distance to the nearest group of galaxies , normalised by the group radius .
we approach this problem by first studying numerical simulations with semi - analytic galaxies , to later corroborate with mock 2dfgrs catalogues the possibility of applying the method to the real 2dfgrs galaxies , thus providing a new test of the semi - analytic model of galaxy formation .
it should be noted that according to the present understanding of the galaxy formation process , galaxies form exclusively in dark - matter haloes @xcite .
therefore , our analysis does not refer to groups as any dark matter halo , but instead to associations of at least two galaxies with @xmath10 .
this , according to recent measurements of the halo occupation number statistics ( * ? ? ?
* in sdss ) , can only be found in haloes with @xmath11m@xmath12 .
other studies @xcite shows a dependence of the halo mass function on the environment . since the properties of the semi - analytic galaxies depend only on their host halo mass ( )
, we can also expect a dependence of the luminosity function on the local environment .
this paper is organised as follows .
we start by describing the numerical simulation and semi - analytic galaxies used in this work in section [ sec : sim ] , where we also describe the selection of haloes used to characterise the local galaxy density in the simulation box . in section
[ sec : method ] we describe the method and algorithms used to determine the luminosity function corresponding to the different semi - analytic galaxy environments .
we present the results from the numerical simulation galaxies in section [ sec : simres ] , and study the possibility of detecting such results in the 2dfgrs by applying the method to mock 2dfgrs catalogues in section [ sec : mocks ] . in section [ sec : results ] we show the results from the real 2dfgrs catalogue , and finally in section [ sec : conc ] we conclude our work with a summary of our main conclusions .
the numerical simulation used in this work , kindly provided by the cosmology group at durham group , comprises data on dark matter particles and on galform semi - analytic galaxies .
the numerical simulation box is @xmath13h@xmath4mpc a side , and contains @xmath14 particles .
the process by which the distribution of dark matter particles was populated by galaxies can be found in cole et al .
( 2000 ) , and can be summarised as follows .
after haloes of at least @xmath15 dark matter particles are identified in the simulation using a friends - of - friends algorithm , the galform algorithm produces a monte - carlo merger history for each halo which defines the galaxies that populate it .
we emphasize that the properties of these galaxies depend only on the halo mass , and not on the environment in which each halo resides .
these galaxies are characterised by magnitudes in several bands ( of which we only make use of the @xmath16 band ) , and stellar formation rate parameters , plus spatial positions and peculiar velocities .
the total number of galaxies in the simulation box is approximately @xmath17 million . due to the dynamical limitations of the numerical simulation ,
the galaxy sample is limited by an absolute magnitude limit of @xmath10 .
in this paper we characterise the density at which a galaxy resides by locating the dark - matter halo that is closest to the galaxy position in units of halo / group radius , denoted by @xmath1 , calculated as half the mean projected distance between pairs of galaxies in a group .
we use a normalised distance ( @xmath0 ) in order to take into account the different halo sizes .
this is motivated by the clear indications from numerical simulations that the galaxy density around dark - matter haloes depends on the normalised distance and on the halo concentration parameter ( * ? ? ?
figure [ fig : dens ] shows the density profile around haloes for the dark - matter ( dotted line ) using nfw profile , and for galaxies in halos with different mass ranges in the simulation ( solid lines ) .
the galaxy density profile shows no important variation with the halo mass . as can be seen ,
even though the halo concentrations vary between the different samples of haloes shown in the figure , there is little difference between the dark - matter and galaxy density profiles , even for @xmath18 , the dark - matter density is only lower than the galaxy density by about a factor of @xmath19 .
this fundamental result supports our hypothesis that a sample of galaxies lying within a given range of normalised distances to halo centres are characterised by the same local galaxy or dark - matter density .
= 11.truecm the procedure to define our semi - analytic galaxy subsamples can be described as follows : 1 .
calculate the distance from each galaxy to the nearest haloes .
2 . assign this galaxy to the halo that presents the shortest normalised distance @xmath0 to the galaxy .
3 . divide the galaxy sample according to the normalised distance to the closest halo .
this process produces galaxy subsamples that are characterised by a complicated spatial distribution which is depicted in figure [ fig : geometry ] from different viewpoints ( top left : front view , top right : side - view , lower panels , angle views ) ; in this example we have three shells of equal density for three groups of galaxies ( each sphere corresponds to a different group ) . as we are interested in assessing the possibility of applying this method to real data , the different observational biases present in real data samples have to be taken into account as well .
one of the effects that will prove particularly important for our work is that of peculiar velocities in redshift surveys , which causes the well known finger - of - god effect around groups and clusters of galaxies .
this effect makes a galaxy that is sitting in the middle of the cluster potential well appear spuriously far away from the cluster centre .
this will lead us to induce that this galaxy belongs to a subsample characterised by a low density environment due to its spuriously large separation from the halo centre and will introduce a systematic error in our measurements . in order to avoid this problem , we simply remove a cylinder centered on the halo centre and aligned with the direction of the line of sight . in the case of the simulation box ,
we take this direction to be the z - axis .
note that in figure [ fig : geometry ] we have removed these cylinders from the spheres surrounding each group .
the galaxy subsample selection criteria and the removal of a cylinder shaped volume to avoid the effects of peculiar velocities makes it difficult to estimate the normalization of the luminosity function . therefore , in cases when an estimate of the volume is needed we also produce a catalogue of random positions which populate the same regions occupied by the galaxy subsample of interest . such random catalogues
are produced using the same halo centres in the numerical simulation and following the steps used for defining the galaxy subsamples , but this time using the random particles instead of semi - analytic galaxies .
= 11.truecm the luminosity functions ( lf hereafter ) are calculated using the standard sty ( sandage , tammann & yahill , 1979 ) , step wise maximum likelihood ( swml , efstathiou , ellis , & peterson , 1988 ) and @xmath20 ( vmax , schmidt , 1968 , huchra & sargent , 1973 ) methods .
as willmer ( 1997 ) provides a detailed review on the three different methods , we only present their most relevant characteristics .
the vmax method is the most popular of the three , and weights each galaxy by the volume corresponding to the maximum distance out to which it would remain below the magnitude limit of the survey .
this method also provides the normalization of the lf , but it does not take into account the fact that galaxies present a clustered distribution .
another setback of the vmax method is that it requires the counts of galaxies to be binned in magnitude , which makes data of individual galaxies to be lost .
the sty method is a maximum likelihood estimator , which has the advantage of calculating a likelihood for each individual galaxy , as well as being an unbiased estimator of the lf for a clustered spatial distribution of galaxies .
one drawback is that this method does not provide the normalization of the lf and assumes a fixed functional form for the shape of the lf , usually a schechter function ( schechter , 1976 ) .
the swml method is similar to the sty method , in that it is also a maximum likelihood method , but it does not assume a specific shape of the lf since it uses galaxies binned in consecutive absolute magnitude intervals . in the case of the sty and swml methods , the normalization is calculated using the estimators given by @xcite .
we will quantify the luminosity function by quoting the values of the parameters @xmath6 and @xmath7 in the schechter formula : @xmath21 where @xmath22 , and the parameter @xmath6 refers to the characteristic luminosity of the sample of galaxies ; @xmath7 is the faint - end slope and indicates the relative importance of a population of low luminosity galaxies .
for instance , a large , negative value of @xmath23 indicates a large population of dwarf galaxies ; @xmath24 is the normalization of the lf which we choose not to calculate since this parameter traces the density profile around groups of galaxies or dark - matter haloes .
the mass function of haloes in the numerical simulation is well described by the jenkins et al .
( 2001 ) formula , which is characterised by a large population of low mass haloes , and a few , rare , high mass objects with @xmath25m@xmath12 .
observational catalogues face little difficulties in detecting the high mass objects , but can not observe galaxies fainter than the apparent magnitude of the survey , which translates into an approximate lower mass cut - off in the catalogue .
for example , in the 2dfgrs percolation inferred galaxy group catalogue ( 2pigg ) , this is @xmath26m@xmath12 ( eke et al .
2004 ) . in order to illustrate what would be found in the dependence of the luminosity function with normalised distance to the halo centres , we first select from the simulation , haloes with a distribution of masses resembling that of an observational sample of groups .
in particular , we use the distribution of masses in the full 2pigg presented by eke et al .
( 2004 ) . in figure
[ fig : selgroup ] we show the distribution of masses in the simulation and in a 2dfgrs mock catalogue .
we then select haloes from the simulation so that they show the same distribution of masses as the 2pigg . as can be seen , this is very similar to placing a lower limit on the halo mass of @xmath11m@xmath12 .
= 8.truecm it should be noted that as the semi - analytic model for galaxy formation only takes into account the mass of the dark matter halo when populating it with galaxies , the dependence of the luminosity function with normalised distance responds to variations in the mass function of low mass haloes around larger dark - matter haloes ( in order to improve clarity , we will refer to the latter as simulated clusters ) .
figure [ fig : massf ] shows the variations in halo mass function as the normalised distance ( @xmath0 ) to a nearby simulated cluster increases .
the differences in amplitude respond mainly to the variation in density around simulated clusters ; for instance , in the cluster central regions , the density is about @xmath27 orders of magnitude larger than the mean in the simulation , thus explaining the @xmath27 order of magnitude difference between the dotted and solid lines . in this figure
there is a striking lack of high mass haloes in the mass function of haloes in the simulated cluster outer regions compared to the inner haloes at @xmath28 ; this is caused by the selection of clusters in the simulation , which contains all the high mass haloes with masses @xmath29h@xmath4m@xmath12 .
the resulting mass functions show larger abundances of high mass haloes in the cluster centres , a slight decrease in high mass haloes for the intermediate distances , and again an increase for the larger distances from the simulated cluster centres . on the low mass end of the mass function
, it can be seen that the inner haloes present the largest population , whereas this diminishes as we go towards haloes in regions further away from the cluster centres .
variations in the mass function of low mass haloes around simulated clusters may be biased in the sense that simulations show that the radial distribution of sub - haloes within their parent haloes is substantially less centrally concentrated than that of the dark matter @xcite .
this is not a setback to our procedures since we only argue that the normalised distance between haloes and clusters provides a reasonable proxy for local density .
furthermore , we have shown that semi - analytic galaxies trace the same density profile regardless of the mass of the central cluster .
= 8.truecm given that the properties of semi - analytic galaxies depend only on its host halo mass , the variations in the mass function of haloes on normalised distance to simulated clusters will imprint variations in the lf .
for instance , the haloes in the outermost regions contain larger abundance at @xmath30m@xmath12 which contributes satellites galaxies which may increase the value of @xmath7 .
the higher upper limit in halo mass will also induce a change in @xmath6 since larger haloes tend to host the brightest galaxies . in order to study the variations of the lf in the simulation ,
we proceed to rank the semi - analytic galaxies according to their normalised distance to the closest halo centre , and produce several subsamples at @xmath31 , @xmath32 , and so forth . in a first instance , we use the real - space galaxy positions .
later on , we apply redshift - space distortions by displacing the galaxy z - coordinate positions by an amount equal to the z - component of their peculiar velocities ( in units of h@xmath4mpc ) .
( 230,450 ) ( 0,0 ) ( 0,215 ) the results in real - space are displayed in solid lines in figure [ fig : cube ] , where the top panel shows the dependence of @xmath6 on @xmath0 , and the bottom panel , the dependence of the parameter @xmath7 .
error bars correspond to @xmath33-@xmath34 errors calculated using the jacknife method .
all errors in this work are calculated using this method , a choice that is supported by numerous works reporting that jacknife errors are comparable to the dispersion of results from large numbers of realistic mock catalogues ( see for instance , padilla et al . , 2005 ) . as can be seen
, the dependence of @xmath6 indicates a population of very bright galaxies near the simulated cluster centres , which dims significantly as we approach a normalised distance of @xmath35 .
interestingly , at @xmath36 , the characteristic luminosity increases to reach a constant value of @xmath37 , which is consistent with the value of @xmath6 characterising the full sample of semi - analytic galaxies in the simulation .
this in perfect agreement with what was found for the mass function of haloes in the same normalised distance ranges .
qualitatively , a very similar behaviour can be seen for the parameter @xmath7 , which is very negative at the centre of the groups , indicating a large population of low luminous galaxies or semi - analytic satellites ( low and high mass haloes respectively ) . at @xmath8
, @xmath7 reaches its highest value ( or flattest slope ) @xmath38 , and settles at @xmath39 at larger normalised separations .
the full sample of galaxies in the simulation shows @xmath40 ; therefore the galaxies at large separations from groups show a comparable @xmath6 to the full sample of galaxies , but a comparatively low population of dwarf galaxies .
these findings encode a wealth of information regarding the galaxy formation process and possible further evolution . even though the semi - analytic model for galaxy formation is a relatively simple one , in the sense that it only requires the mass of a host halo in order to populate it with galaxies
, it manages to produce answers that can be reconciled with several physical processes that affect the evolution of galaxies . for instance , the dependence of the parameters @xmath6 and @xmath7 can be understood as follows .
the high luminosity galaxies in the centres of groups along with a large population of dwarfs is a reasonably well documented fact observed in redshift surveys @xcite and individual clusters @xcite , and is thought to be a consequence of subsequent mergers of haloes .
this preferentially takes place in the centres of groups and clusters of galaxies ( massive haloes ) , where the galaxy density is high . the large population of dwarf galaxies
is thought to be produced by the close interaction of haloes , tidal stripping @xcite or other processes considered in the semi - analytic model .
note that since this simulation only takes into account dark - matter interactions , no gas processes are taken into account .
a new interesting result that can be appreciated in figure [ fig : cube ] is how extended is the influence of simulated clusters over their surrounding medium .
as can be seen , the population of low luminosity galaxies ( i.e. low mass haloes ) is slightly reduced at distances @xmath35 ( faint @xmath6 and flat @xmath7 ) with respect to the field population , marking the radius of influence of clusters .
the mechanisms that produce such effects take place in the cluster centres , so this loss of galaxies may be the consequence of galaxy / halo passages through the near centres of clusters .
there is also a clear depletion of low luminosity galaxies ( low absolute values of @xmath7 ) , which can also be understood in terms of passages of low mass haloes ( dwarfs ) through the group centres of which a fraction is cannibalised by larger haloes .
finally , at separations @xmath41 , the behaviour of both @xmath6 and @xmath7 tends to constant luminosity function parameters for larger separations .
notice that the values of the field lf parameters in the simulation depend on input values that have not been tuned to reproduce exact lf parameters of 2dfgrs data .
we also check whether the variations in luminosity function depend on the mass of the nearby simulated clusters .
figure [ fig : cube ] shows in solid squares and triangles , the resulting variations in @xmath6 and @xmath7 with @xmath0 when the mass of the centre haloes varies ( solid squares for low mass haloes , solid triangles for high mass haloes ) . as can be seen , the characteristic luminosity does not change significantly in the central or more external regions , being the values of @xmath6 consistent when taking as centres either low , high or all the haloes selected .
the major difference can be seen at intermediate distances , @xmath8 , where low mass haloes are surrounded by lower luminosity galaxies than higher mass haloes ( by almost half a magnitude ) . regarding the distance at which @xmath6 tends to a roughly constant value , it can be argued that lower mass haloes influence a larger volume in terms of their radius , @xmath1 , since @xmath6 remains at low luminosity values out to @xmath42 .
the variation in the luminosity function slope @xmath7 shows that at intermediate distances , low mass haloes contain less low luminosity galaxy neighbours than high mass haloes . on the other hand , at low @xmath43 ,
the population of faint galaxies decreases dramatically in low mass haloes . in this case
the low mass halo sample contains objects with @xmath44h@xmath4m@xmath12 , which for the magnitude limit of the semi - analytic galaxies in this simulation , corresponds to haloes that preferentially contain @xmath27 or less galaxies , whereas large mass haloes contain large amounts of satellite galaxies which account for this difference .
as a first check , we repeat the study performed in the previous section to the full simulation box data , but taking the galaxy positions in redshift - space . in this case , as can be seen in figure [ fig : parcube ] shown by the crosses , the effect of peculiar velocities smears the relation seen in the cube in real - space , to a degree that makes it difficult to arrive at any conclusions regarding a dependence of the luminosity function parameters on normalised distance .
this is produced by the high velocity dispersions that characterise the centres of galaxy groups and clusters , which produce the finger - of - god effect in redshift surveys , and that spuriously displaces galaxies corresponding to the centres of groups to the group outer regions , thus contaminating the relation between @xmath6 and @xmath7 with @xmath0 .
this is a very important effect that needs to be taken into account when analysing redshift space data .
( 230,450 ) ( 0,0 ) ( 0,215 ) in order to correct for this effect , we proceed to remove from our analysis all those galaxies that lie within a cylindrical volume centred on the group centres and aligned with the direction over which the galaxy positions were displaced by their peculiar velocities ( the z - axis in this case ) .
we experiment with different cylinder radii and show different symbols in figure [ fig : parcube ] ( see the key ) . as can be seen
, this technique is able to partially recover the qualitative shape of the dependence of both @xmath6 and @xmath7 with @xmath0 found using the full simulation box in real - space .
notice that as we increase the cylinder radii to reduce the peculiar velocity effect , we lose data of central galaxies in groups , so we test with different cylinder radii .
we select an optimal value for this parameter such that it minimises the effects from redshift - space distortions and low - number statistics , @xmath45 .
this is the value of @xmath46 that we will adopt from now in the analysis of both mock and real catalogues , unless otherwise stated . before going any further
, we describe the characteristics of the observational data .
the 2dfgrs survey ( * ? ? ?
* & 2003 ) comprises a total of @xmath47 low resolution spectra of galaxies distributed in two regions near the north and south galactic poles ( ngp and sgp regions respectively ) covering a total of about @xmath48 @xmath49 on the sky .
the apparent magnitude limit of the survey is of the order of @xmath50 with plate - to - plate variations on the sky of about @xmath51 magnitudes .
the parent imaging 2dfgrs catalogue is the automated plate machine survey , apm @xcite .
the survey presents a complicated angular completeness mask which has also been made available by the 2dfgrs team and which we use in all of our following analyses .
when measuring luminosity functions using 2dfgrs data , we pay special atention to remove angular incompleteness effects by weighting each galaxy by the inverse of the incompleteness factor at its position ( this information is part of the public 2dfgrs data ) .
this incomplentness factor includes effects from fiber collisions , and therefore our results are expected to be free from such effects .
the galaxy groups to be used in our analysis are taken from the 2pigg catalogue @xcite , which contains groups of galaxies identified in the 2dfgrs with at least 2 members .
there is a total of @xmath52 groups in the 2pigg catalogue .
the advantage in using this particular sample of groups relies in the careful calibration of the friends - of - friends ( fof ) method in eke et al . (
2004 ) using detailed mock 2dfgrs catalogues .
the distribution of group @xmath1 radii show only a few hundred groups with @xmath53 arcseconds , with a typical @xmath54 arcminutes .
this indicates that there is little effect on our results from this problem , which we also correct for by applying a completeness weight . at any rate , our conclusions from the 2dfgrs are only relevant at distances @xmath55 from the group centres ; a study of the group central regions would need a more careful treatment of this problem
. we will now test whether we can use the 2dfgrs groups and galaxies to detect the variations in the lf of section 4 using mock catalogues .
we use the semi - analytic galaxies in the numerical simulation to produce mock 2dfgrs and 2pigg catalogues .
we place an observer at a random position in the simulation box , then simulate the observing biases including variable magnitude limits , angular incompleteness masks , and @xmath56 corrections from norberg et al .
the final mock 2dfgrs catalogue contains comparable numbers of members in the ngp and sgp areas .
we then feed the mock 2dfgrs catalogue into the fof algorithm kindly provided by vincent eke , and produce a mock 2pigg catalogue .
one of the advantages of the mock 2pigg catalogue over the real 2pigg is that we also know the true mass of each galaxy group .
the true mass of a group is that of the halo in the numerical simulation box corresponding to the galaxy group in the mock catalogue .
therefore , these mock catalogues match as closely as possible the properties of the observational dataset , and will make it possible for us to assess whether the properties of the luminosity function of galaxies at different normalised distances from galaxy groups can be measured with the real 2dfgrs dataset .
= 11.truecm the first check we perform on the mock catalogues is that we actually recover the schechter function parameters of the underlying luminosity function in the numerical simulation .
figure [ fig : levelngp ] shows the likelihood contours on the @xmath57 plane that correspond to the @xmath58 ( solid lines ) , and @xmath27-@xmath34 ( dotted lines ) levels , using different methods for estimating the luminosity function .
the large empty circle represents the simulation underlying luminosity function , and as can be seen , the swml method provides the most accurate answer from the mock ngp region . after checking that we can recover the galaxy luminosity function within at least @xmath27-@xmath34 of the underlying values , we proceed to rank galaxies according to their minimum normalised distance to galaxy groups , just as we did with the semi - analytic galaxies in the simulation box ( in the real and mock catalogues we actually use the projected positions of group member galaxies to estimate the group radius , @xmath1 ) .
the resulting luminosity function parameters can be seen in figure [ fig : parmock ] , for for a cylindrical volume of radius @xmath45 , that is removed from the analysis in order to avoid the finger - of - god effects . for comparison
, this figure also shows the results from the full numerical simulation box in real - space .
as expected , it is difficult to recover exactly the behaviour for @xmath7 and @xmath6 seen in the simulation box .
moreover , there is no apparent trend in the schechter parameters with @xmath0 , except for a marginal signature of a minimum in the @xmath7 parameter at 3 @xmath0 , which is in agreement with the simulation box results .
still , it is extremely important that the results are not seriously affected by spurious variations of @xmath6 and @xmath7 , incompatible with the simulation box results .
therefore , any features in @xmath6 and @xmath7 seen in the observational data will likely correspond to actual underlying changes in these parameters .
( 450,230 ) ( 0,0 ) ( 235,0 )
we start this section by corroborating that our algorithm provides reliable measurements of the lf by first calculating the 2dfgrs lf using the full sample of galaxies and comparing it to previous measurements . applying the swml algorithm we find the schechter parameter combination @xmath59 , @xmath60 , and @xmath61 .
these results are in agreement ( @xmath62-@xmath34 ) with measurements of the @xmath16 galaxy luminosity function from the eso slice project by zucca et al .
( 1997 ) , who find @xmath63 , @xmath64 and @xmath65 ; and a measurement of the 2dfgrs galaxy luminosity function by norberg et al .
( 2002 ) , who obtain @xmath66 , @xmath67 and @xmath68 .
it should be noted that we do not intend to provide a new measurement of the galaxy luminosity function in the 2dfgrs , this is only to demonstrate that the code used in this work provides reasonable answers .
the reader should bear in mind that the semi - analytic model is an approach that aims to reproduce the outcome of many real processes involved in the formation of galaxies , and that this result is tied to a particular cosmological model .
therefore , on a direct comparison between mocks and real data , several characteristics of the numerical and semi - analytical models should be taken into account ( for instance , differences in the strength of the peculiar velocity field between mock and real data can produce different responses to the method when using real data ) .
therefore , the actual variations of the schechter parameters as a function of @xmath0 may be completely different to what we found in the semi - analytic simulation .
we bear these possible differences in mind and proceed to apply the method to the real 2dfgrs and 2pigg catalogues following the same steps from the previous section . after ranking the 2dfgrs galaxies by @xmath0 we measure the luminosity functions and present the schechter parameters @xmath7 and @xmath6 in figure [ fig : par2df ] . in this figure
, the left panels show the dependence with normalised distance , and the right panels the dependence on @xmath69 , which assuming a nfw profile around the 2pigg galaxy groups , corresponds to the dark matter density in terms of the mean density .
results from the ngp region are similar to those from the mock catalogues in that there is no clear trend in the schechter parameters with @xmath0 given the error bars .
however , the sgp region shows a dip in both , @xmath6 and @xmath7 at the expected @xmath9 . at @xmath70
the schechter parameters converge to values marginally consistent with the field luminosity function parameters @xcite .
this result is very encouraging and may be interpreted as a possible , new success of the semi - analytic galaxy formation model .
the implications for this are dramatic since only one parameter , namely the mass of a dark - matter halo can reproduce an intuitive life cycle for a galaxy / halo in the nearby regions of clusters , including passages , cannibalism , and stripping of dark - matter from tidal encounters .
still , the reader should be reminded that this is only a marginal detection on the sgp region of the 2dfgrs .
( 430,450 ) ( 0,0 ) ( 0,215 ) ( 235,0 ) ( 235,215 ) the right panels of figure [ fig : par2df ] show the dependence of the schechter parameters on density contrast @xmath69 .
when analysing these results it should be kept in mind that the density contrast quoted in these panels corresponds to that of a nfw profile centred on the galaxy groups .
were this profile a fair representation of the dark - matter density profile of real galaxy groups , then our values of @xmath69 would correspond to the actual dark - matter density in which our samples of galaxies are embedded .
these results are compatible and complementary to those from mo et al .
( 2004 ) , who study the changes in @xmath6 and @xmath7 in the range @xmath71 .
there is a clear overlap between our results and mo et al . at @xmath72 , and we extend this measurement to @xmath73 .
( 450,230 ) ( 0,0 ) ( 235,0 ) we perform a final analysis of the luminosity function around groups .
this time we take the full 2dfgrs catalogue and divide our sub - samples of galaxies at different @xmath0 by the luminosity of the nearest galaxy group at @xmath74 . figure [ fig:2dflum ] shows the results , and indicates that the luminosity function of galaxies around the brighter groups contain an excess of dwarf galaxies when compared to fainter groups , at least for @xmath75 , as indicated by the more negative values of the parameter @xmath7 .
also , as expected , the typical luminosity of objects in brighter groups is higher than in fainter groups , as indicated by the more negative values of @xmath6 in bright groups . in this case
, this effect is seen at all values of @xmath0 .
if we compare the lf for different halo masses in figure [ fig : cube ] with the luminosity dependence in figure [ fig:2dflum ] we can see that brighter groups behave similarly than massive haloes except for the inner regions @xmath76 where the results in both figures do not match ; however , at @xmath77 we can see the dip of @xmath7 and @xmath6 at larger normalised distance for low mass haloes or less luminous haloes .
it has been shown that the total luminosity of a group is an excellent tracer of group mass ( eke et al .
2004 , padilla et al . 2004 ) , as our results also show .
this figure also shows a marginal detection of a more extended dip in @xmath6 for the low luminous clusters , although it would be necessary to study a larger sample of groups and galaxies than the 2dfgrs dataset to be able to measure the upturn in @xmath6 at larger separations .
we presented a study on the variations of the luminosity function of galaxies around clusters of galaxies in numerical simulations with semi - analytic galaxies .
the aim of this work was to quantify the extent of the influence of clusters on their environments and to assess whether such effects can be reproduced by a semi - analytic model where the main parameter determining the population of galaxies are the halo masses . in order to do this ,
we select a population of haloes from the numerical simulation which are set to be the clusters of our simulation .
these are characterised by the same distribution of mass found for the groups in a mock 2pigg catalogue . therefore
, the variations we find in the semi - analytic model can be considered as predictions which we also test in this work by applying the method to the 2dfgrs galaxies and 2pigg groups . as mentioned above ,
the main parameter taken into account by the semi - analytic model when producing galaxies is the mass of the halo .
therefore , the variations in the lf parameters @xmath6 and @xmath7 correspond to variations in the mass function of haloes at different ranges in @xmath0 , and when we measure the variations in the luminosity function we are actually measuring changes in the mass function with @xmath0 .
a good agreement between the variations in the simulated and observed lfs indicates that the mass of a halo correlates very well with the different processes a galaxy may have been subject to during its life .
the semi - analytic galaxy luminosity function shows important variations with the distance to the centres of clusters in the simulation box .
we find that galaxies are brighter in the cluster inner regions , and that they show a minimum in characteristic luminosity at intermediate distances , @xmath8 . at larger separations , the characteristic luminosity increases and
tends to a constant value , @xmath78 .
the minimum in @xmath6 is interpreted as a measure of the zone of influence of clusters in terms of their radius .
we also find a similar behaviour in @xmath7 , which indicates a massive abundance of low luminosity galaxies ( low mass haloes and/or semi - analytic satellites ) in the cluster central regions , which decreases to a minimum at intermediate distances , and increases again at @xmath36 to reach a constant value , somewhat lower in absolute value than for the full semi - analytic galaxies .
we interpret these results as clusters being capable of cannibalising small haloes from nearby regions out to @xmath79 virial radii from their centre , and also of stripping the mass of normal haloes .
these two effects would explain the high luminosity central galaxies surrounded by numerous low luminosity satellites .
the former can be the result of successive mergers of haloes , whereas the latter can be thought of as captured low mass haloes from nearby regions , and relatively high mass haloes from nearby regions which have been tidally stripped off part of their mass .
this would also explain the lack of high and low luminosity semi - analytic galaxies from the region around @xmath8 . by properly taking into account observational biases
, we checked whether the variations found in the luminosity function of galaxies around groups in the simulation box can be measured using available data from redshift surveys .
first we study the effect that the velocity dispersion of galaxies in groups induces in our results when the galaxy positions are measured in redshift space . when no corrections for this effect are attempted , we find that the dependence of the schechter parameters is almost completely erased
. one way to avoid this is the removal of cylinders centred on each group , aligned with the line of sight .
the most convenient value for the projected radius of the cylinder is about @xmath80 .
once we demonstrate that it is possible to find variations in the luminosity function from redshift data ( at least qualitatively ) , we test whether this can still be detected using the 2dfgrs galaxies and 2pigg groups , which are available to the astronomical community
. the 2dfgrs is affected by a series of observational biases such as a complicated incompleteness mask and radial selection function , which can make it even more difficult to detect the variations in the luminosity function .
therefore , we produce mock catalogues which reproduce as closely as possible all the observational biases present in the real data .
when we do this , we find that the dependence of the schechter parameter on the normalised distance to group centres is almost completely lost .
this indicates that the removal of a cylindrical volume around groups is not enough to recover the underlying variations in the luminosity function , but that the measurements do not return spurious answers either . when applying the method to real data , we find that the luminosity function of galaxies in the 2dfgrs ngp strip shows no variations in @xmath6 or @xmath7 with the normalised distance to the group centres .
however , we do find a minimum value in @xmath6 and @xmath7 in the luminosity function of the sgp region , in agreement with our results from the semi - analytic model .
further agreement between the data and the semi - analytic model is found in the correlation between the luminosity of galaxies and that of their neighbour galaxy group , which is seen in the simulation as a correlation between halo mass and the luminosity of the surrounding galaxies .
also , the data is marginally consistent with a dip in @xmath6 for the low luminosity 2pigg groups , again in agreement with the results from the full simulation box .
these results strongly indicate that the interpretation given to the simulation results could actually have taken place in the formation history of 2dfgrs galaxies .
we have shown in this work that the variation of the galaxy luminosity function around groups and clusters of galaxies is a powerful tool for understanding the processes that galaxies are subject to in the vicinities of clusters .
it is left to forthcoming efforts to assess whether the variations in the lf are also accompanied by changes , for instance , in the surface brightness of galaxies , as this galaxy property is also believed to be influenced by the presence of a nearby cluster or group of galaxies ( or by its absence , as in a low surface brightness galaxy ) .
this as well as other galaxy properties can be studied to complement the results obtained in this work .
a future application of the method presented here that would benefit from improved statistics is the study of the regions surrounding groups in the sloan digital sky survey ( abazajian et al .
2004 ) , which in the very near future will double the number of galaxies available in the 2dfgrs .
this method can also be applied to deep , wide area surveys ( such as the galex medium spectroscopic survey , morrissey et al .
2005 ) which will make it possible to detect variations of galaxy properties around groups with redshift providing a study of the evolution of the influence of clusters on their surrounding environments throughout the history of the universe .
this work was supported in part by the daa - eso grant at puc , ndp was supported by a proyecto postdoctoral fondecyt no . 3040038 , gg is supported by proyecto regular fondecyt no .
rg was partially supported by proyecto regular fondecyt no .
we acknowledge support from the fondap center for astrophysics . we have benefited from helpful discussions with d. g. lambas .
we thank the referee for helpful comments and suggestions . | we present a study on the variations of the luminosity function of galaxies around clusters in a numerical simulation with semi - analytic galaxies , attempting to detect these variations in the 2df galaxy redshift survey .
we subdivide the simulation box in equal - density regions around clusters , which we assume can be achieved by selecting objects at a given normalised distance ( @xmath0 , where @xmath1 is an estimate of the halo radius ) from the group centre .
the semi - analytic model predicts important variations in the luminosity function out to @xmath2 . in brief , variations in the mass function of haloes around clusters ( large dark - matter haloes with @xmath3h@xmath4m@xmath5 ) lead to cluster central regions that present a high abundance of bright galaxies ( high @xmath6 values ) as well as low luminosity galaxies ( high @xmath7 ) ; at @xmath8 there is a lack of bright galaxies , which shows the depletion of galaxies in the regions surrounding clusters ( minimum in @xmath6 and @xmath7 ) , and a tendency to constant luminosity function parameters at larger cluster - centric distances .
we take into account the observational biases present in the real data by reproducing the peculiar velocity effect on the redshifts of galaxies in the simulation box , and also by producing mock catalogues .
we find that excluding from the analysis galaxies wich in projection are close to the centres of the groups provides results that are qualitatively consistent with the full simulation box results .
when we apply this method to mock catalogues of the 2df galaxy redshift survey ( 2dfgrs ) and the 2pigg catalogue of groups , we find that the variations in the luminosity function are almost completely erased by the finger - of - god effect ; only a lack of bright galaxies at @xmath8 can be marginally detected in the mock catalogues .
the results from the real 2dfgrs data shows a more clear detection of a dip in @xmath6 and @xmath7 for @xmath9 , consistent with the semi - analytic predictions .
large - scale structure of the universe , methods : n - body simulations , galaxies : kinematics and dynamics cosmology : theory |
small - x qcd evolution , historically based on dglap @xcite and bfkl @xcite dynamics , has been widely investigated in the past years @xcite , leading to a better understanding of the two approaches just mentioned and to robust resummed predictions for the gluon density and splitting function @xcite .
there is now a remarkable consensus @xcite among various resummation approaches on the resulting gluon evolution kernel , and a satisfactory comparison of some of them to experimental data @xcite .
the basic idea underlying the progress of the resummation approaches just mentioned , lies in the observation @xcite that the bfkl kernel embodies an infinite number of subleading contributions which are collinear singular .
these terms are parametrically large and need to be taken into account in order to achieve consistency with the renormalization group ( rg ) .
the techniques for incorporating such terms differ in detail according to the various authors , but lead eventually to similar results .
however , all approaches developed so far limit themselves to a consistent resummation scheme only for the evolution of the gluon density .
the quark - sea contribution is obtained on the basis of @xmath1-factorization of the @xmath16 dipole @xcite in the dis factorization scheme and/or by adding the @xmath17 contribution to the next - to - leading-@xmath18 ( nl@xmath3 ) bfkl kernel .
the main purpose of the present paper is to devise a resummed small-@xmath3 evolution scheme in a coupled matrix form , so as to treat gluons and quarks on the same footing and in a collinear factorization scheme which is as close as possible to a predetermined one , e.g. the @xmath2 scheme .
our matrix approach , in the collinear limit , has the advantage that it complies automatically with the matrix factorization of the integrated partonic densities in the singlet evolution , and thus is able to incorporate the known low - order anomalous dimensions for any value of @xmath19 , the moment index . on the other hand , in the high - energy limit , the ( gauge - invariant ) unintegrated partonic densities
are well defined by @xmath1-factorization around different values of @xmath20 . to be precise ,
the gluon unintegrated density is defined around @xmath21 and the quark around @xmath22 .
therefore , in the leading high energy region that is around @xmath21 we are able to include the known ll@xmath3+nl@xmath3 bfkl kernel in the gluon channel _ only _ , thus leaving the quark entries somewhat unconstrained from the @xmath1-factorization standpoint .
note however that assuming a bfkl framework in matrix form is a demanding requirement , because @xmath1-factorization implies resummation formulae for the anomalous dimension matrix up to nl@xmath3 level .
therefore , incorporating both exact low - order anomalous dimensions ( say , in the @xmath2 scheme ) and nl@xmath3 expressions ( from the exact bfkl kernel ) imposes on our matrix kernel some nontrivial consistency relations expressing the requirement that collinear and high - energy schemes do not conflict with each other .
they are discussed in detail in the following , and we find that they are satisfied in the @xmath2 scheme up to nlo level , while they are marginally violated at nnlo , by small terms in the @xmath23 entry of relative order @xmath10 .
for this reason we restrict ourselves , in this paper , to the nlo - nl@xmath3 level .
this means that , starting at nnlo , our splitting functions will be in some matrix scheme which is not the @xmath2 scheme .
higher order effects on the scheme change have been studied in @xcite and have been found to be comparable to renormalization scale uncertainties .
besides the nlo - nl@xmath3 information mentioned above , we impose the general requirement of consistency with the renormalization group in both ordered and antiordered configurations of exchanged partonic momenta @xmath1 s .
this is best expressed in the @xmath24 symmetry of the kernel , where @xmath25 is conjugated to @xmath26 .
we enforce this symmetry by the so - called consistency constraint @xcite which introduces an @xmath20-dependence in the ( leading ) kernel , so as to resum those parts @xcite of the higher order bfkl kernels which are required by the rg .
this procedure follows previous papers @xcite and is used in particular for the @xmath9 matrix element of the kernel . despite all such requirements
, there is a considerable ambiguity in our approach which is tied up to the matrix structure , because the @xmath25 and @xmath20 dependences of the various matrix elements of the kernel are constrained only to a limited extent by the collinear and high - energy limits .
we therefore introduce in sec .
[ sec : bmf ] some further requirements , mostly related to the pole structure of the kernel in the @xmath25 and @xmath20 variables , by requiring it to have at most simple poles .
this assumption is quite natural in the case of @xmath27 because of the bfkl limit , and follows by the @xmath20-expansion method @xcite in the @xmath28 case .
the leading twist pole structure of the kernel at lowest order in @xmath29 is basically @xmath30 $ ] , where @xmath31 denotes the lo dglap anomalous dimension matrix . the full kernel to second order in @xmath29
is constructed in secs .
[ sec : bmf ] and [ sec : knla ] according to the requirements stated above .
it also contains running coupling effects at the scales suggested by the nl@xmath3 bkl kernel and by the rg . in the frozen @xmath29 case we calculate in sec .
[ s : fcad ] the resulting anomalous dimension matrix , and its eigenvalues @xmath32 .
the leading eigenvalue at high energies , @xmath33 , contains important resummation effects in the @xmath34 variable .
we also obtain resummation formulae for @xmath35 and @xmath36 , the latter being specific to our matrix approach and not directly obtained on the basis of the nl@xmath3 bfkl kernel only . in sec .
[ s : cfrgf ] we present results for the effective eigenvalue ( or characteristic ) functions @xmath37 as inverse functions of the anomalous dimensions , and for the hard pomeron exponent @xmath38 . in the case with running coupling described in sec .
[ s : nrrc ] , we provide the resummed dglap splitting function matrix in @xmath3 space , obtained by the numerical deconvolution method proposed in @xcite and generalized to the matrix case .
details of the matrix kernel and of the anomalous dimension expressions are left to appendices a - c .
the general purpose of this paper is to provide integro - differential matrix equations for unintegrated parton distributions , which interpolate between dglap evolution equations @xcite in the hard scale variable @xmath39 and the high - energy bfkl evolution equation @xcite in the rapidity - like variable @xmath40 . despite the high - energy and collinear factorization constraints ,
the above interpolation is subject to considerable ambiguities , due to the following facts : * off - shell , unintegrated densities are defined in a gauge - invariant way by @xmath1-factorization of gluon and quark exchanges around different values of the moment index @xmath41 , namely @xmath42 for the gluon and @xmath43 for the quark . of the quark exchange , which leads to an energy dependence of the cross - section of type @xmath44 at high energies . for alternative approaches to the definition of unintegrated densities
see @xcite . ]
therefore , in the high - energy region that is around the leading value @xmath42 only some effective gluon equation ( which incorporates the high - energy quark contributions ) is constrained by the bfkl limit , whose kernel has been calculated perturbatively @xcite .
this makes the interpolation of the kernel to generic @xmath20 values more ambiguous for the quark entries . *
the collinear limit constrains the matrix kernel for both quarks and gluons , but in a factorization - scheme dependent way , and only in the strongly ordered region of transverse momenta @xmath45 and in the anti - ordered one .
this limit only restricts the singularities of the kernel in the variable @xmath25 ( conjugated to @xmath39 ) so as to reproduce the low order anomalous dimension matrix .
in addition , the collinear @xmath46 anti - collinear relationship implies the existence of a @xmath24 symmetry , whose form is however quite general , depending again on the factorization - scheme . in order to tame the ambiguities of the off - shell continuation mentioned above
, we shall use a few basic criteria which we shall argue can be consistently imposed and correspond to a factorization - scheme choice for both high - energy and collinear limits .
let us refer to a matrix kernel @xmath47 , acting on @xmath1-space , such that the parton green s function is given by @xmath48^{-1}_{ab}({\boldsymbol{k}},{\boldsymbol{k}}_0 ) \ ; , \quad ( a , b = q , g ) \;.\ ] ] in the frozen @xmath29 limit , the kernel matrix elements are diagonalised in @xmath25-space and given by the eigenvalue function @xmath49 .
our first basic assumption is that in the collinear limit @xmath50 and @xmath20 fixed , the matrix kernel @xmath51 shows simple poles only , in the form of a @xmath25-expansion @xmath52 here , the @xmath53 singularity is natural because of the dglap limit , and would be the only term present in a pure evolution equation in @xmath26 .
in fact , it implies ( sec .
[ ss : fklla ] ) that the one - loop anomalous dimension matrix @xmath54 is given by @xmath55 , the coefficient of the @xmath53 pole of lowest order in @xmath29 ( while the higher order terms @xmath56 involve @xmath57 as well ) .
note however that higher powers of @xmath58 could have been present also , anomalous dimension relation @xmath59 could be replaced in the subleading @xmath20-dependence , thus producing higher powers of @xmath58 . ] and do actually occur in the normal formulation of the nl@xmath3 bfkl kernel @xcite . by eq .
( [ eq : gammastruct ] ) we explicitly exclude such possibility in our matrix kernel , while the bfkl kernel will be recovered by proper algebraic manipulations ( sec .
[ sec : knla ] ) .
our second assumption is analogous to ( [ eq : gammastruct ] ) with @xmath20 and @xmath25 interchanged . in the high - energy limit of @xmath60 with @xmath25
kept fixed we require simple pole singularities in the @xmath20-expansion @xmath61 where , in addition , @xmath62 the @xmath63 singularity is natural because of the bfkl limit and would be the only term present in a pure evolution equation in @xmath40 .
it implies that the eigenvalue function of the ll@xmath3 bfkl kernel is given by @xmath64 , the coefficient of the @xmath63 singularity of lowest order in @xmath29 ( while the nl@xmath3 bfkl kernel , discussed in sec .
[ sec : knla ] , involves @xmath65 also )
. therefore , higher order singularities in @xmath20 which are present in the anomalous dimension at higher order will be obtained ( sec .
[ s : fcad ] ) by using the rough anomalous dimension relation @xmath66 in the subleading @xmath25-dependence .
the fact that only @xmath67 and @xmath68 possess the @xmath63 singularity is related to the fact that in usual factorization schemes @xcite , only @xmath36 and @xmath69 show a ll@xmath3 dependence on the @xmath34 variable .
there is a third important assumption , which deals with the relationship between collinear and anti - collinear orderings of exchanged transverse momenta .
both orderings are to be incorporated in our off - shell formulation and simple kinematical considerations show that the variable conjugated to @xmath39 in the reverse ordering is @xmath70 .
therefore , values of @xmath25 and @xmath70 must be related by some symmetry , and we shall assume , out of simplicity , @xmath71 it is perhaps useful to recall the formal basis for the symmetry ( [ eq : symmetry ] )
. let us write the @xmath1-factorization formula for the @xmath72 differential cross - section in the form @xcite @xmath73 where we have lumped in the @xmath74 superscripts the dependence on the hard scales @xmath75 of the process .
then the change of energy - scale from @xmath76 to , say , @xmath77 can be incorporated by the change of kernel @xmath78}({\boldsymbol{k}},{\boldsymbol{k}}';{\omega } ) = \left(\frac{k}{k'}\right)^{\omega}{\mathsf{k}}^{[k k_0]}({\boldsymbol{k}},{\boldsymbol{k}}';{\omega } ) \ ; , \ ] ] or , at frozen @xmath29 , by the @xmath20-dependent shift @xcite of the corresponding eigenvalue functions to denote eigenvalue functions of kernels considered in or related to previous works on small-@xmath3 resummations in the gluon - channel .
we keep the symbol @xmath51 to denote both kernels and eigenvalue functions specifically designed for this matrix formulation .
note also that the @xmath79 s are perturbative coefficients of expansions in @xmath80 , therefore differing in normalization by a factor @xmath81 from their @xmath51 counterparts in eq .
( [ eq : kexpansion ] ) , because @xmath82 . ]
@xmath83}({\gamma},{\omega } ) = \chi^{[k k_0]}({\gamma}-{{\textstyle\frac{\omega}{2}}},{\omega } ) \ , .\ ] ] on the other hand , in the one - channel case namely when one considers only gluon dynamics the @xmath84 symmetry of @xmath85 implies the @xmath86 symmetry of the kernel @xmath87 and the @xmath88 symmetry of the eigenvalue functions @xmath89}({\gamma},{\omega})$ ] . therefore , at energy - scale @xmath77 , the @xmath24 symmetry of eq .
( [ eq : symmetry ] ) holds for @xmath90}({\gamma},{\omega})$ ] , whose superscript will be dropped from now on . in the matrix case , the thorough discussion of sec .
[ ss : gfdrs ] shows that the collinear @xmath46 anti - collinear symmetry of the matrix kernel is expected to have the more general form @xmath91 therefore , eq . ( [ eq : symmetry ] )
is obtained by choosing the similarity transformation
@xmath92 so as to have @xmath93 and represents yet another restriction of our off - shell scheme . )
must be supplemented by a corresponding choice of impact factors in order to satisfy the symmetry for observable cross - sections . in the realistic nlo - nl@xmath3 case the similarity transformation @xmath92 in eq .
( [ eq : similarity ] ) is expected to be an operator in @xmath1-space also . ] in the following we shall show in more detail how to construct the matrix kernel so as to satisfy the known collinear / high - energy limits with lo - ll@xmath3 accuracy ( secs .
[ ss : fklla ] , [ ss : gfdrs ] ) and nlo - nl@xmath3 accuracy ( sec .
[ sec : knla ] ) , within the scheme restrictions provided by assumptions ( [ eq : gammastruct ] , [ eq : omegastruct ] , [ eq : symmetry ] ) .
we note from the start that eqs .
( [ eq : gammastruct ] ) and ( [ eq : omegastruct ] ) impose consistency relations on the anomalous dimensions , which show up in a novel nl@xmath3 resummation formula for @xmath36 , to be discussed in detail in secs .
[ ss : rf ] and [ ss : cr ] . in order to discuss the above features in more detail in the frozen @xmath29 limit
, we introduce the triple expansion @xmath95 where @xmath96 are @xmath97 matrices in the @xmath98 indices , and we note that @xmath99 , consistently with the simple pole assumption of eqs .
( [ eq : gammastruct ] , [ eq : omegastruct ] ) .
we also use the notation @xmath100 , @xmath101 , @xmath102 and @xmath103 to mean partially resummed coefficients , as already done before . in this paper
we limit ourselves to two terms ( @xmath104 ) in the frozen @xmath29-expansion , which will be able to accommodate the ll@xmath3 and nl@xmath3 bfkl kernels .
however , running coupling effects will be introduced by various scale choices for the various terms ( sec .
[ ss : runningc ] ) and this implies in general an infinite series when expanding around a fixed scale .
let us first show how to construct @xmath105 so as to be consistent with the collinear and high - energy limit at lo - ll@xmath3 accuracy .
we denote by @xmath106 the anomalous dimension matrix , with the expansion @xmath107 where we recall the small-@xmath20 behaviour ( @xmath108 ) @xmath109 and , in the @xmath2 scheme , @xmath110 with the two eigenvalues @xmath111 we then write the generalised bfkl equation for the unintegrated parton densities @xmath112 in the form @xmath113 where the source @xmath114 is local in @xmath1-space .
it is then straightforward ( cf .
[ ss : adm ] and app .
[ a : read ] ) to derive , for frozen @xmath29 , dglap type equations for the integrated densities @xmath115 of type @xmath116 where [ eq : gamma_from_ck ] @xmath117 and so on .
this identification relies on the expansion ( [ eq : kexpansion ] ) , which in turn relies on the assumed single @xmath25-pole structure of ( [ eq : gammastruct ] ) , see app .
[ a : read ] .
we note that this procedure implies that the part of the anomalous dimensions , proportional to @xmath118 with @xmath119 , at a given order of perturbation theory can be generated from the lower orders .
( [ eq : gamma_from_ck ] ) can be used to constrain recursively the @xmath50 singularities of @xmath120 for a given set of low order anomalous dimensions , for instance in the @xmath2 scheme . in particular
, it determines @xmath121 , as noticed after eq .
( [ eq : gammastruct ] ) , but does not fix @xmath122 , which is therefore a scheme - changing parameter . at lo - ll@xmath3 level , we choose the parameterization @xmath123\chi_c^{\omega}({\gamma } ) \end{pmatrix } \ ; , \ ] ] where superscript to denote @xmath20-shifted @xcite eigenvalue functions @xmath124 , @xmath125 being the `` left projection '' of the eigenvalue function @xmath126 with singularities in the half - plane @xmath127 only . ]
@xmath128 reduces to the leading bfkl eigenvalue function in the @xmath42 limit ( @xmath129 is the digamma function ) , and @xmath130 is a simple interpolation of @xmath131 and @xmath132 poles .
higher twist terms will be needed at the nl level in sec . [
sec : knla ] .
let us note that in eq .
( [ eq : k0param ] ) we have already incorporated assumptions ( [ eq : omegastruct ] ) and ( [ eq : symmetry ] ) at this level . in fact , the form ( [ eq : k0param ] ) is consistent with eq .
( [ eq : omegastruct ] ) because of the bfkl limit of @xmath133 in eq .
( [ eq : chi_0 ] ) and of the absence of @xmath63 poles in @xmath134 and @xmath135 . furthermore , the symmetry ( [ eq : symmetry ] ) is present by construction . on the other hand , the simple form of @xmath136 in eq .
( [ eq : chi_c ] ) and the fact that it appears unchanged in the @xmath137 , @xmath138 and @xmath23 entries and in the part of the @xmath9 entry which has no @xmath63 singularities are all `` off - shell '' features which we choose for simplicity reasons .
the fact that @xmath134 and @xmath139 are proportional to @xmath136 helps to satisfy the momentum conservation sum rule .
indeed , since @xmath140 we obtain @xmath141 where all quantities are evaluated at @xmath142 and @xmath143 is some higher - twist kernel , having singularities at @xmath144 .
it is then easy to show that the sum - rule violation is at most @xmath145 , instead of @xmath146 . )
( [ eq : qplusgdot ] ) would imply that the sum rule violation is higher - twist only . in the matrix case
, a higher twist component is expected on top of the perturbative component discussed here . ]
this feature will be improved in sec .
[ sec : knla ] by modifying the parameterization ( [ eq : k0param ] ) at nlo - nl@xmath3 level so as to reduce the violation to @xmath147 .
as @xmath149 , which may be collinear , anti - collinear or some mixture.,scaledwidth=30.0% ] the role of the symmetry ( [ eq : symmetry ] ) in implementing rg properties deserves a special discussion for the matrix kernel .
in fact , its typical effect is to produce two pole terms in @xmath25 and @xmath70 ( see eq .
( [ eq : chi_c ] ) ) , which are supposed to describe the correct product of @xmath106-matrices for both direct and reverse orderings .
the situation is pretty clear for collinearly ordered particles .
in fact , referring to fig .
[ fig : abcd ] , with @xmath3 decreasing from right to left we have a sequence of splitting functions such as @xmath150 describing @xmath151 going to @xmath152 , @xmath152 to @xmath153 and so on , as predicted from @xmath51 also .
on the other hand , in the anti - collinear limit the dglap splitting functions need to account for the opposite splittings , @xmath152 to @xmath151 , etc . @xmath154
then , one would naively expect that the anti - collinear pole in the kernel be associated with @xmath155 .
this seems to work fine as long as we consider a complete chain of anti - collinear splittings .
problems arise however when trying to join collinear and anti - collinear chains .
firstly there is an issue of colour factors : the anomalous dimension @xmath156 implicitly includes a factor @xmath157 for the number of varieties of parton @xmath158 that can be produced ( @xmath159 if @xmath158 is a gluon , @xmath160 for an ( anti)quark ) . for each exchanged particle in fig .
[ fig : abcd ] that factor should be included exactly once . in the collinear limit
it is included in the branching to the right of a given exchange ( e.g. for @xmath152 it is included in @xmath161 ) , while in the anti - collinear limit , as written in ( [ eq : naive_anticollinear ] ) , it is included to the left ( in @xmath162 ) . if we are to consider a single evolution from right to left containing both collinear and anti - collinear splittings we should ensure that the @xmath157 factors are consistently included to one side , for example in the branching to the right of the exchange .
one then needs to correct the splitting function @xmath163 for an anti - collinear @xmath164 splitting by a factor @xmath165 . the second issue that arises relates to high - energy factorization .
for each exchanged gluon we have a factor @xmath63 . in a sequence of collinear branchings that @xmath63 factor
is associated with the splitting function to the right of the gluon exchange ( e.g. if @xmath152 is a gluon then it is included in @xmath161 ) , while for anti - collinear branchings it comes from the splitting function to the left ( i.e.from @xmath162 ) .
this causes problems if we have an anti - collinear splitting to the left of a gluon exchange and a collinear one to the right , since _ both _ will include a @xmath63 factor for the intermediate exchanged gluon .
however it is necessary for the gluonic part of our green function to be consistent with high - energy factorization , which systematically assigns an exchanged gluon s @xmath63 divergence to the larger-@xmath3 part of the diagram , i.e. to the right of the exchanged gluon in the collinear limit .
therefore in the case of an anti - collinear branching we should multiply @xmath163 by a factor @xmath166 where @xmath167 , so as to ensure that @xmath168 is never associated with a @xmath63 factor , while @xmath158 has it when @xmath158 is a gluon .
note that @xmath169 is arbitrary ( other than that it should be a non - zero constant for @xmath60 ) since high - energy factorization is not defined for quarks around @xmath42 we shall discuss its choice below .
the outcome of this discussion is that colour factor and the high - energy factorization corrections can be combined by introducing a similarity transformation matrix @xmath170 and defining a ` refactored ' splitting function matrix @xmath171 for anti - collinear splittings in an evolution that will combine both collinear and anti - collinear splittings : @xmath172 a matrix kernel will therefore have collinear and anti - collinear structure of the form @xmath173 and will satisfy the collinear @xmath46 anti - collinear symmetry in the general form ( [ eq : gensymmetry ] ) . the fact that the diagonal entries ( in particular the @xmath9 element ) of @xmath106 and @xmath174 are identical is consistent with our expectation that the single - channel ( @xmath175 ) limit should coincide with bfkl , which is symmetric in @xmath176 .
the structure of colour factors and @xmath63 ensures that chains containing collinear and anti - collinear splittings will have the expected sets of colour factors and overall @xmath63 factors .
note finally that we can further specify @xmath92 so as to satisfy the symmetry in the form ( [ eq : symmetry ] ) .
since @xmath177 we can simplify eq .
( [ eq : basic_matrix_kernel ] ) by exploiting the arbitrariness of @xmath178 and setting it to @xmath179 thus providing , at leading level , a fully symmetric collinear structure @xmath180 as assumed in eq .
( [ eq : symmetry ] ) .
let us recall that , while the lo anomalous dimension matrix and the ll@xmath3 expression of @xmath69 are factorization - scheme independent , the nlo , nl@xmath3 expressions do depend on the scheme ( except possibly for the eigenvalue @xmath33 in the frozen @xmath29 limit ) .
this opens up the possibility of constructing the kernel so as to reproduce the nlo , nl@xmath3 anomalous dimensions in a given scheme , say @xmath2 scheme
. however , we have to comply with the restrictions ( [ eq : gammastruct ] , [ eq : omegastruct ] , [ eq : symmetry ] ) , in particular the requirements of a simple @xmath25-pole structure at fixed @xmath20 , a simple @xmath20-pole at fixed @xmath25 , and absence of @xmath63 singularity in @xmath134 and @xmath135 .
this means , at nlo , that the @xmath182 terms of @xmath183 and @xmath35 can not be reproduced by an explicit @xmath63 term in @xmath184 , but should result from the @xmath25-dependence of @xmath185 and @xmath186 , where @xmath187 are free scheme choice parameters .
in other words , we have to adjust the subleading @xmath25-dependence of @xmath105 so as to reproduce the known @xmath2 anomalous dimensions at nl@xmath3 level in the form
@xmath188 the above discussion shows that we have to change the parameterization ( [ eq : k0param ] ) at next - to - leading level so as to allow a more general subleading @xmath25-dependence .
we choose the following one @xmath189 \end{pmatrix } \;,\ ] ] where @xmath190 is a higher - twist kernel possessing the symmetry ( [ eq : symmetry ] ) , e.g. @xmath191 and @xmath192 is an @xmath20-dependent coefficient which we require to be regular for @xmath193 and vanishingly small as @xmath194 , e.g .- space function @xmath195 rapidly vanishes for @xmath196 ( cf .
( [ dqgz ] ) ) , thus not disturbing the large-@xmath3 behaviour of the model . ]
@xmath197 the form ( [ eq : chi_mat_0_all_prerequisites ] ) allows one to choose @xmath198 so as to reproduce the known nl@xmath3 expressions ( [ eq : gamma0smallx ] ) and ( [ eq : gamma1smallx ] ) of @xmath183 and @xmath35 in the @xmath2 scheme , up to order @xmath182 .
note that the logic is here reversed with respect to the dis scheme , in which @xmath199 is directly calculated by @xmath1-factorization ( see @xcite ) and nl@xmath3 resummation formulae for @xmath183 and @xmath35 are derived . by the @xmath50 behaviour @xmath200 we derive , from eq .
( [ eq : gamma_from_ck ] ) for @xmath201 , that @xmath202 . in more detail , the nl@xmath3 coefficient @xmath203 , according to eq .
( [ gamma_1 ] ) is given by @xmath204}_{qg } = \left [ { { } _ { 0}{\mathsf{k}}_1^{(0 ) } } + { { } _ { 0}{\mathsf{k}}_0^{(1 ) } } \ ; { { } _ { 1}{\mathsf{k}}_0^{(0 ) } } + { { } _ { 1}{\mathsf{k}}_0^{(1 ) } } \ ; { { } _ { 0}{\mathsf{k}}_0^{(0 ) } } \right]_{qg } \;.\ ] ] by expanding the matrix products in terms of their matrix elements and by taking into account the conditions ( [ eq : quarkrowcond ] ) , only the last term in the r.h.s .
( [ eq : gamma_qg1nlx ] ) does not vanish , yielding @xmath205 2{c_a}= 2{c_a}\left [ \frac43 { t_f}+ { \delta_{qg}}\right ] \;,\ ] ] where we used the explicit expressions ( [ eq : def_chi_ht ] ) , ( [ eq : constcoef ] ) and ( [ c_c ] ) .
the @xmath2-scheme value @xmath206 of eq .
( [ eq : gamma1smallx ] ) is then recovered provided @xmath207 note that the corresponding expression for @xmath208 involves @xmath209 , thus respecting the colour charge relation @xmath210 which is apparent in eq .
( [ eq : gamma1smallx ] ) .
one could repeat the above procedure in other factorisation schemes as well . in the @xmath211 scheme ,
e.g. , the value @xmath212 is recovered by setting @xmath213 having fixed @xmath183 and @xmath35 at nlo - nl@xmath3 level , the remaining constraints ( exact nlo anomalous dimension matrix and exact bfkl kernel at nl@xmath3 level ) are fixed by a proper choice of @xmath184 . because of the expansion in eq .
( [ eq : omegastruct ] ) @xmath214 where @xmath215 has only @xmath9 and @xmath23 entries , the resolvent @xmath216^{-1}\ ] ] has the @xmath9 matrix element proportional to a pure @xmath40 evolution form , with the kernel @xmath217_{gg } \nonumber \\ & = \left [ { { } _ { 0}{\mathsf{k } } } + { { } _ { 0}{\mathsf{k}}}\;{{}_{1}{\mathsf{k } } } \right]_{gg } + { { \cal o}\left({\bar{\alpha}_\mathrm{s}}^3\right)}\ ; , \label{eq : equivnlkern}\end{aligned}\ ] ] where @xmath218 .
we thus arrive at the identification @xmath219 which parallels eq .
( [ eq : gamma_from_ck ] ) for the perturbative expansion , with the difference that it concerns the @xmath9 entry only , as is appropriate to the @xmath1-factorization of gluon exchange .
( [ eq : nlx_ident ] ) is used as in the single - channel case @xcite to derive @xmath220 from the known expression of the nl@xmath3 bfkl kernel explicitly considered in sec .
[ ss : runningc ] are to be subtracted out . ]
@xmath221 and of the @xmath105 kernel ( [ eq : chi_mat_0_all_prerequisites ] ) .
explicitly @xmath222 where @xmath223 is the regular part of the @xmath9 anomalous dimension .
this result reduces to the corresponding one - channel @xcite @xmath175 subtraction @xmath224}\chi_c)$ ] after setting @xmath225 .
we encounter at this point a consistency relation on the factorization scheme for @xmath201 , due to the fact that we want to incorporate both @xmath201 and @xmath221 in a kernel satisfying the simple - pole assumptions of sec .
[ ss : bckc ] , as better discussed in sec .
[ ss : cr ] .
note in fact that , given @xmath105 , eq ( [ eq : gamma_from_ck ] ) determines @xmath226 in terms of @xmath201 and eq .
( [ eq : nlx_ident ] ) determines @xmath220 in terms of @xmath221 .
therefore , @xmath227_{gg}$ ] the @xmath25-pole part with the @xmath63 singularity is determined in two independent ways , which should provide the _ same _ result , in the form @xmath228 } = { { } _ { 0}\!\left ( { \gamma}_{+,1}\right ) } - ( { { } _ { 0}{\mathsf{k}}}_0\;{{}_{1}{\mathsf{k}}}_0)_{gg}^{(0 ) } \ ; , \label{eq : ggconscons}\ ] ] where we have used the fact that the simple - pole part of @xmath221 predicts the nl@xmath3 part of @xmath229 , that is @xmath230 we show in sec .
( [ ss : cr ] ) that the consistency equation ( [ eq : ggconscons ] ) is identically satisfied , provided @xmath231_{qg}$ ] , as assumed in eq .
( [ eq : gamma1qg ] ) . therefore
, both dis and @xmath2 schemes satisfy eq .
( [ eq : ggconscons ] ) and can be accommodated by a proper matrix kernel at nl level .
more precisely , we start from the `` ansatz '' @xmath232 where + _
( i ) _ the function @xmath233 at @xmath42 is equal to @xmath234_{gg}}{(2{c_a})^2 } \ ] ] [ see eq .
( [ chi1tilde ] ) for an explicit expression ] and is extrapolated to generic @xmath20 values by the @xmath20-shift @xcite procedure of left and right projections as follows @xmath235_l({\gamma } ) + [ \tilde{\chi}_1]_l(1+{\omega}-{\gamma } ) \;,\ ] ] so as to satisfy the symmetry ( [ eq : symmetry ] ) ; + _ ( ii ) _ in order to minimize momentum sum rule violations , we have added to the high - energy pole @xmath63 in front of @xmath236 a low - energy term @xmath237 : their sum vanishes at @xmath238 ; + _ ( iii ) _ the @xmath239 matrix is fixed by matching @xmath240 to the known nlo @xmath2 splitting functions ( cf .
( [ gamma_1 ] ) ) : @xmath241 @xmath242 being the coefficient of the simple pole at @xmath243 of @xmath233 .
we note that @xmath136 in eq .
( [ eq : our_k1_struct ] ) is again chosen for simplicity reasons according to the symmetry ( [ eq : symmetry ] ) .
the final form of the next - to - leading matrix kernel is then @xmath244 we shall choose the running coupling scales of our kernel as in @xcite .
we thus associate @xmath245 ( @xmath246 ) to the ll@xmath3 bfkl kernel @xmath247 , and @xmath248 ( @xmath249 ) to all other ones .
the choice of the intermediate gluon momentum transfer is suggested by the nl@xmath3 bfkl kernel itself , which contains the beta - function dependent term @xmath250 ( quoted for a renormalisation scale choice @xmath251 ) , corresponding to the @xmath1-space kernel @xmath252_{\mathrm{reg}}\ ] ] where the regularization procedure is explained in @xcite . since the term in eq .
( [ krun ] ) is accounted for by expanding @xmath245 up to nl order , the expression in eq .
( [ chirun ] ) should be subtracted out from the nl kernel considered before .
more precisely , the kernel @xmath221 in eq .
( [ eq : nlx_ident ] ) is meant to have the eigenvalue @xmath253 + \frac{b}{2 } \left ( \chi_0 ' + \chi_0 ^ 2 \right ) \;,\ ] ] where the expression in square brackets is the eigenvalue function at energy - scale @xmath77 , obtained by the @xmath20-expansion of eq .
( [ eigenvchange ] ) , and @xmath254 is the customary nl eigenvalue @xcite at energy - scale @xmath255 , given by the expression @xmath256
-\frac{1}{4 } \chi_0''({\gamma } ) -\frac{1}{4 } \left(\frac{\pi}{\sin \pi { \gamma } } \right)^2 \frac{\cos \pi { \gamma}}{1 - 2{\gamma } } \left[3+\left(1+\frac{2{t_f}}{{c_a}^3}\right)\frac{2 + 3{\gamma}(1-{\gamma})}{(1 + 2{\gamma})(3 - 2{\gamma})}\right ] \\ &
\quad + \left(\frac{67}{36}-\frac{\pi^2}{12 } -\frac{5{t_f}}{9{c_a } } \right ) \chi_0({\gamma } ) + \frac{3}{2 } \zeta(3 ) + \frac{\pi^3}{4\sin \pi{\gamma } } \nonumber\\ & \quad - \sum_{n=0}^{\infty } ( -1)^n \left [ \frac{\psi(n+1+{\gamma})-\psi(1)}{(n+{\gamma})^2 } + \frac{\psi(n+2-{\gamma})-\psi(1)}{(n+1-{\gamma})^2 } \right ] \ ; . \label{eq : nllorg}\end{aligned}\ ] ] it follows that the overall kernel has the structure undergo the transformation @xmath257 and @xmath258 , with @xmath259 , while quadratic powers of @xmath29 are modified as @xmath260 . ] @xmath261 0 & k_0^{\omega}\end{pmatrix } + { \hat{\alpha}}(k_>^2 ) \begin{pmatrix } { \gamma}_{qq,0 } k_c^{\omega } & { \gamma}_{qg,0 } k_c^{\omega}+ { \delta_{qg}}{k_{\mathrm{ht}}}^{\omega}\\[1ex ] { \gamma}_{gq,0 } k_c^{\omega } & \left({\gamma}_{gg,0}-\frac{2{c_a}}{{\omega}}\right ) k_c^{\omega}\end{pmatrix } \\[1ex ] \nonumber & \quad + { \hat{\alpha}}^2(k_>^2 ) \left [ \left({\gamma}_1-{\mathsf{k}}_0^{(1)}{\mathsf{k}}_0^{(0)}\right ) k_c^{\omega}+ ( 2{c_a})^2\left(\frac1{{\omega}}-\frac2{1+{\omega}}\right ) \begin{pmatrix } 0 & 0 \\[1ex ] 0 & \tilde{k}_1^{\omega}- \tilde{\chi}_1^{(0 ) } k_c^{\omega}\end{pmatrix } \right ] \;.\end{aligned}\ ] ] @xmath262 , @xmath263 and @xmath264 are the @xmath1-dependent kernels corresponding to the characteristic functions @xmath133 , @xmath136 and @xmath265 respectively , and their explicit expressions in @xmath266-space are provided in app . [ a : kcf ] ( cf .
( [ k0action ] ) , ( [ k_c_bis ] ) and ( [ k_ht ] ) ) . the eigenvalue of the @xmath9 entry of @xmath184 at @xmath42 , thanks to eq .
( [ chi1bfkl ] ) and to the subtraction procedure in eq .
( [ eq : nlsubtr ] ) , is provided by the expression @xmath267 + \frac12 \chi_0({\gamma } ) \frac{\pi^2}{\sin^2(\pi{\gamma } ) } \nonumber \\ & - \chi_0({\gamma } ) \frac{a_{gg}(0)/2{c_a}}{{\gamma}(1-{\gamma } ) } -\frac{{c_f}/{c_a}}{{\gamma}(1-{\gamma } ) } \left [ \frac{{\gamma}_{qg,0}(0)/2{c_a}}{{\gamma}(1-{\gamma } ) } + \frac{{\delta_{qg}}/{c_a}}{(1+{\gamma})(2-{\gamma } ) } \right ] \ ; , \label{chi1tilde}\end{aligned}\ ] ] which then acquires the @xmath20-dependent shift , as explained previously .
note that @xmath268 is the same as that in @xcite in the @xmath269 limit , and differs from it by the @xmath4-dependent terms in @xmath270 , @xmath271 and @xmath272 .
note also that cubic and quadratic poles cancel out in @xmath268 , because of the @xmath20-shift of the collinear poles and of their factorization , which are embodied in our formalism ( cf .
sec [ ss : cr ] ) .
the remaining single poles of @xmath273 are provided by @xmath242 , which is obtained as follows .
let us introduce the constant ( in @xmath25 ) coefficients @xmath274 of the characteristic functions : @xmath275 then , according to eqs .
( [ eq : chi_0],[eq : chi_c ] ) , we have [ constants ] @xmath276 note that @xmath277 vanishes by virtue of the ll@xmath3 expansion @xmath278 , in other words it is a scheme - independent coefficient . on the other hand , all other quantities in eq .
( [ constants ] ) are scheme - dependent , i.e. , they depend on the particular choice we adopted for shifting the poles of @xmath279 and on the definition of @xmath280 as in eqs .
( [ eq : chi_0 ] , [ eq : chi_0 ] ) . by using the above notation
we derive the expansions of the bfkl eigenvalue function @xmath281 the expansion of the subtraction term in eq .
( [ eq : nlsubtr ] ) @xmath282 + { { \cal o}\left({\gamma}^0\right)}\ ] ] and the expression of the pole term in @xmath184 @xmath283 where all quantities in the three formulas above are evaluated at @xmath42 .
note again that @xmath242 satisfies the consistency relation in eq .
( [ eq : ggconscons ] ) , proved in sec .
[ ss : cr ] .
the remaining expressions used in the previous subsections are easily obtained by the following detailed formulas [ k01k00ij ] @xmath284_{qq } & = c_c \left({\gamma}_{qg } { \gamma}_{gq } + { \gamma}_{qq}^2 \right ) + { \delta_{qg}}{\gamma}_{gq } { \chi_{\mathrm{ht } } } & & \sim { \gamma}_{qq,1 } \sim \frac{{c_f}}{{c_a}}{\gamma}_{qg,1 } \\ [ { \mathsf{k}}_0^{(1 ) } { \mathsf{k}}_0^{(0)}]_{qg } & = c_c { \gamma}_{qg } \left({\gamma}_{qq}+{\gamma}_{gg}\right ) + { \delta_{qg}}{\gamma}_{gg } { \chi_{\mathrm{ht } } } & & \sim { \gamma}_{qg,1 } \\ [ { \mathsf{k}}_0^{(1 ) } { \mathsf{k}}_0^{(0)}]_{gq } & = { \gamma}_{gq } \left [ c_0 \ , \frac{2{c_a}}{{\omega } } + c_c \left ( { \gamma}_{qq } + a_{gg } \right ) \right ] \\ [ { \mathsf{k}}_0^{(1 ) } { \mathsf{k}}_0^{(0)}]_{gg } & = c_0 { \gamma}_{gg } \ , \frac{2{c_a}}{{\omega } } + c_c \left [ a_{gg } { \gamma}_{gg } + { \gamma}_{gq } { \gamma}_{qg } \right ] & & \sim { \gamma}_{gg,1 } - ( 2{c_a})^2 \frac{\tilde{\chi}_1^{(0)}}{{\omega } } \ ; , \end{aligned}\ ] ] where @xmath285 stands for @xmath286 and @xmath287 means asymptotic in the @xmath60 limit . from the above formulas we can compute the high - energy limit of @xmath184 @xmath288 the introduction of running coupling may change our expectations on how momentum conservation is satisfied by our kernel . since we incorporate nlo anomalous dimensions , as given by eqs .
( [ eq : k1ansatz ] ) , we expect violations at next order , that is at relative order @xmath289 .
this is also the order at which running coupling effects start to matter in the actual derivation of anomalous dimensions , being related to a commutator of two values of @xmath29 , evaluated at different scales ( cf .
[ a : read ] ) .
therefore , no problems arise at nlo .
since we do not explicitly consider incorporating nnlo results in this paper , we do not make an effort to improve energy - momentum conservation at frozen @xmath29 .
we should add that the consistency relations on nl@xmath3 terms which restrict our scheme start imposing a collinear scheme restriction at order @xmath290 which is derived in sec .
[ ss : rf ] and is violated , even if marginally , in the @xmath2-scheme .
in this section we want to discuss a number of issues concerning the anomalous dimension matrix in the case of frozen coupling , in which the whole matrix can be analytically calculated in terms of the kernel matrix elements in @xmath291-space .
this allows us to compute the two eigenvalues @xmath292 and their inverses , the effective eigenvalue functions @xmath37 , as well as their eigenvectors .
we obtain in this way the hard pomeron exponent @xmath38 and the resummation formulae for the matrix elements of the anomalous dimension matrix .
the latter , at a given level of the ll@xmath3 hierarchy , must be consistent with the exact low order anomalous dimensions we have used in constructing the kernel , thus providing consistency relations for the collinear and @xmath1-factorization schemes .
here we find , at nl@xmath3 level , that such relations are identically satisfied by our construction at nlo , while they put a nontrivial constraint on the @xmath290 term of @xmath36 at nnlo .
if @xmath29 is frozen , the matrix kernel @xmath87 is scale invarant and its resolvent admits the @xmath25-representation @xmath293 where @xmath294 is the characteristic function matrix of @xmath87 .
we introduce the eigenvalues @xmath295 and eigenvectors @xmath296 of @xmath51 in the usual way @xmath297 where the @xmath29- and @xmath20-dependences of all the above quantities are understood .
one can then write the spectral decomposition @xmath298 where @xmath299 are the orthogonal projectors on the eigenspaces of @xmath51 , and are given by @xmath300 @xmath301 being the left - eigenvectors of @xmath51 satisfying @xmath302 when @xmath303 s and @xmath304 s are evaluated at the same value of @xmath25 .
the behaviour in @xmath305-space of @xmath306 is determined by the @xmath25-poles in eq .
( [ eq : resolvent ] ) .
these poles are found at @xmath307 such that @xmath308 and are interpreted as anomalous dimension eigenvalues of @xmath306 . by applying a driving term @xmath309 to the green s function @xmath306 , at leading - twist level
i.e. , taking into account only the two rightmost poles in the half - plane @xmath127 obeying @xmath310 one obtains the vector of ( integrated ) quark and gluon densities @xmath311 we want to show that @xmath312 satisfies the dglap - type evolution equation @xmath313 in terms of a well - defined resummed anomalous dimension matrix @xmath106 .
in fact , by inserting the expression ( [ eq : gf0 ] ) into both sides of eq .
( [ eq : dglap - type ] ) , the equality is satisfied provided @xmath314 \pi_l({\gamma}_l ) f_0({\gamma}_l ) = 0 \ ; , \qquad(l = + , - ) \;.\ ] ] it might seem that @xmath106 is dependent on the initial condition @xmath309 . this is not the case , because whatever the choice of @xmath309 , the projector @xmath315 projects @xmath309 into a vector proportional to @xmath316 , and eq .
( [ eq : gammacondition ] ) reduces to @xmath317 i.e. , @xmath106 is the ( unique ) matrix whose eigenvectors are @xmath318 relative to the eigenvalues @xmath319 .
@xmath320 in more detail , the eigenvalues @xmath321 are provided by @xmath322 = 0\ ] ] and the eigenvectors are @xmath323 therefore , the _ full _ expression of the anomalous dimension matrix is @xmath324 ( { \gamma}_+ - { \gamma}_- ) r & { \gamma}_+ + r\rho{\gamma}_- \end{pmatrix } \;.\end{aligned}\ ] ] we obtain the relationships @xmath325 all the above formulas are exact in the frozen coupling case , and do not depend on our particular assumptions on @xmath51 .
now , by taking into account the structure of @xmath51 described in the previous sections , we compute the anomalous dimension matrix elements at nl@xmath3 level . to this purpose
, we note that the eigenvalues of @xmath51 are defined by @xmath326 and that both @xmath327 and @xmath328 are of order @xmath63 , so that up to nl@xmath3 level we have @xmath329 we note that the equation @xmath330 reduces to the usual bfkl determination of @xmath33 because @xmath331 can be replaced by @xmath332 in the nl@xmath3 term .
furthermore , the equation @xmath333 is dominated by its @xmath25-pole part , yielding @xmath334 this provides the lowest order determination of @xmath335 so that , up to nl@xmath3 level only the one - loop term of @xmath35 contributes and no small-@xmath3 enhancements are present .
the coefficient @xmath336 is now calculable from eq .
( [ uminus ] ) and , up to nl@xmath3 level , we obtain @xmath337 \ ; , \quad r_1 = c_c { \gamma}_{-,0 } + \frac{{\mathsf{k}}_{gq,1}^{(0)}}{2{c_f } } - \frac{{\mathsf{k}}_{gg,1}^{(0)}}{2{c_a } } \ ; , \quad \tilde{r}_1 = \frac{a_{gq}}{2{c_f } } - \frac{a_{gg}+\frac{{c_f}}{{c_a } } { \gamma}_{qg,0}}{2{c_a } } \;.\ ] ] note that , since eq.([uminus ] ) is evaluated at @xmath338 , @xmath336 does not contain @xmath63 enhancements and is generally calculable from fixed order perturbation theory .
the coefficient @xmath339 , on the other hand , is calculated at @xmath33 ( cf .
( [ uplus ] ) ) , and so contains resummation of nl@xmath3 terms to all orders in @xmath29 . given that @xmath338 and @xmath339 are nl@xmath3 quantities , it follows that also @xmath183 and @xmath35 are nl@xmath3 . from the previous equations we obtain the resummation formulae @xmath340 which are well known @xcite .
in addition , the matrix kernel predicts @xmath341 \label{gagqnl } \\
\nonumber & = \frac{{c_f}}{{c_a } } \left [ { \gamma}_{gg}^{{{\mathrm{nl}x } } } + \frac{{\alpha_\mathrm{s}}}{2\pi}\frac{{c_f}}{{c_a } } { \gamma}_{qg,0}(0 ) + \frac{{\alpha_\mathrm{s}}}{2\pi } \frac{{\bar{\alpha}_\mathrm{s}}}{{\omega } } r_1 + { \bar{\alpha}_\mathrm{s}}\tilde{r}_1 \right ] + { { \cal o}\left({\alpha_\mathrm{s}}^4\right ) } .\end{aligned}\ ] ] note that nl@xmath3 running coupling contributions are shown in appendix a to start at order @xmath342 . the above resummation formula for @xmath36 is easily checked to be identically valid at @xmath146 and @xmath145 .
at @xmath147 it yields the relation @xmath343 which characterises the class of schemes described by our matrix formulation , and appears to be not satisfied in the @xmath2 scheme @xcite , , it turns out that , in the @xmath2-scheme , @xmath344 $ ] . ] even though the violation , of relative order @xmath10 , is numerically less than 0.5% for @xmath345 . strictly speaking
, this implies that the @xmath2 scheme at nnlo can not be incorporated in the present matrix approach .
however , one could think of adding the small violation just mentioned by a matching procedure .
they arise in general because of the joined requirements of simple @xmath20-poles and @xmath25-poles inposed on our kernel .
for instance , by the @xmath25-pole hypothesis we determine the @xmath25-pole parts of @xmath184 and @xmath346 by the equations @xmath347 these expressions should be consistent with the @xmath20-pole hypothesis so that higher order poles in @xmath20 , possibly occurring in the r.h.s . of eq .
( [ gammapolepart ] ) , should cancel out .
furthermore , by the @xmath20-pole hypotesis , we determine the @xmath20-pole part of @xmath273 by a subtraction of the bfkl kernel , as follows : @xmath348 once again , this should be consistent with the @xmath25-pole hypothesis , so that quadratic ( and possibly cubic ) @xmath25-poles in @xmath221 should cancel out on the r.h.s . , and furthermore the simple pole should be consistent with eq .
( [ gammapolepart ] ) , that is @xmath349 let us start proving the consistency relation for eq .
( [ gammapolepart ] ) .
generally speaking , they are equivalent to recursive relations on the @xmath20-singuarities of @xmath56 or , in other words , to the resummation formulas proved in sec .
[ ss : rf ] .
for instance , the assumed absence of @xmath20-poles in @xmath350 implies the nl@xmath3 resummation formulas for @xmath351 : @xmath352 as predicted by eqs .
( [ gaqgnl ] , [ gaqqnl ] ) .
the @xmath23 , @xmath9 entries are slightly more complicated . at order @xmath182 ,
@xmath353 is determined by eq .
( [ gammapolepart ] ) , so that no consistency condition arises in the @xmath23 entry .
however , @xmath354 is already determined by eq .
( [ omegapolepart ] ) , so that the consistency condition ( [ simplepolebfkl ] ) arises .
the latter is verified because the simple - pole part of @xmath221 is simply @xmath229 @xcite , so that eq .
( [ simplepolebfkl ] ) reduces to the identity ( [ eq : ggconscons ] ) , which implies , at order @xmath182 , @xmath355_{gg } = \frac{2{c_f}}{{\omega } } \big({{}_{1}{\mathsf{k}}}_0^{(1)}\big)_{qg } \ ; , \label{idqg1}\end{aligned}\ ] ] where the r.h.s .
reduces , by eq .
( [ gamma1qa ] ) , to @xmath356 , as it should .
furthermore , at order @xmath290 we have consistency conditions for @xmath357 and @xmath358 .
the former is identically satisfied , by some algebra similar to eq .
( [ idqg1 ] ) , because @xmath359 as given in eq .
( [ gammaplus ] ) .
the latter is instead non trivial and , after a similar algebra , reduces to @xmath360 as already proved in eq .
( [ eq : nnloconstraint ] ) , with the same consequences .
we finally note that cubic and quadratic @xmath25-poles are absent in ( [ omegapolepart ] ) because of the identity , valid up to order @xmath361 , @xmath362 here the cubic poles at @xmath243 already cancel out in the l.h.s .
, because of the @xmath363 subtraction needed to switch energy - scale @xmath364 , due to the @xmath20-shift ( [ eigenvchange ] ) .
the remaining quadratic poles are given by the r.h.s .
, because of normal collinear factorization , and of absence of @xmath53 singularities in @xmath365 at energy - scale @xmath77 ( eq . ( [ chi0dot ] ) ) .
in this section we present numerical results of some phenomenologically relevant quantities which can be obtained by using the matrix kernel @xmath366 developed in the previous sections .
we recall that the final form of @xmath105 and @xmath184 can be found in eqs .
( [ eq : chi_mat_0_all_prerequisites ] ) and ( [ eq : k1ansatz ] ) respectively , and that the detailed implementation of the running coupling is found in eq .
( [ runningkernel ] ) .
we state once more that our matrix kernel incorporates exactly the dglap and bfkl properties at nlo and nl@xmath3 accuracy .
however , in order to see the impact of the nlo contributions and to compare with previous resummation approaches , we will consider also results obtained from the kernel with only lo anomalous dimensions ( but still in nl@xmath3 approximation ) .
the corresponding kernel which we refer to as nl@xmath3-lo model is built with the same @xmath105 given in eq .
( [ eq : chi_mat_0_all_prerequisites ] ) but with @xmath184 including only the @xmath367 term in the @xmath9 entry , as can be read from eq .
( [ eq : our_k1_struct ] ) by setting @xmath368 .
we shall first investigate the high - energy @xmath369 behaviour of the @xmath72 differential cross section given in eq .
( [ hefact ] ) at fixed and equal value of the two hard scales @xmath370 , by determining the growth exponent ( hard pomeron ) @xmath371 in the limit of frozen coupling . in this limit
, we can use the representation ( [ eq : resolvent ] ) for the green s function @xmath372 and , by using the spectral decomposition introduced in sec .
[ ss : adm ] , we obtain ( the @xmath29-dependence is understood ) @xmath373 the @xmath20-integral gets contributions from the singularities ( labelled by the index @xmath374 ) of the integrand at @xmath375 due to the vanishing of the denominator @xmath376 thus providing @xmath377 in the limit @xmath378 the @xmath25-integral is dominated by the saddle - point @xmath379 such that @xmath380 for the particular values of @xmath381 and @xmath374 such that @xmath382 is maximum .
it turns out that those values correspond to the leading - twist component of the ( @xmath383)-branch of the eigenvalue function @xmath384 , namely the solutions of @xmath385 with @xmath386 for @xmath387 . as a result ,
in the high - energy limit the cross section has the power - like behaviour @xmath388 where the process dependent coefficient @xmath389 is constant or at most logarithmic in @xmath390 , and the growth exponent @xmath371 is determined by the conditions ( [ vanishden],[gammasaddlepoint ] )
. we can recast eq .
( [ gammasaddlepoint ] ) into an equivalent relation for the function @xmath391 .
in fact , by taking the total @xmath25-derivative of eq .
( [ vanishden ] ) we can express @xmath392 thus obtaining the following conditions for the hard pomeron exponent : [ omscondition ] @xmath393 the above conditions in turn can be translated into analogous conditions for the determinant of the operator @xmath394 . in fact , from the relations @xmath395\end{aligned}\ ] ] eqs .
( [ omscondition ] ) are equivalent to [ detcondition ] @xmath396 & = 0 \\ \partial_{\gamma}\det[1-{\mathsf{k}}({\gamma_{\mathrm{s}}},\bar{{\omega}}_i ) ] & = 0 \;.\end{aligned}\ ] ] we have numerically solved the implicit equations ( [ detcondition ] ) in our matrix formulation in a range of @xmath29 up to 0.4 , both at nl@xmath3-nlo and nl@xmath3-lo accuracy , for two values of @xmath397 and 4 . the results for @xmath371 versus @xmath29 are shown in fig .
[ f : oms ] , where we compare with results obtained from our previous one - channel approach .
the nl@xmath3-lo curve at @xmath397 almost overlaps to the old one - channel result , thus showing the stability of the matrix formulation and the continuity with the one - channel approach .
in fact , since at @xmath397 the kernel is diagonal , only the @xmath9 entry determines @xmath371 .
the small discrepancy is due to : _
( i ) _ the momentum - conserving factor in front of @xmath268 in eq .
( [ eq : our_k1_struct ] ) ( it was a plain @xmath63 in the one - channel case ) ; _ ( ii ) _ a non vanishing two - loop anomalous dimension ( only for the low - energy part , actually ) ( we enforced vanishing anomalous dimension in the one - channel case ) . by including the nlo contributions we obtain a moderate increase of the pomeron intercept , which slightly diminishes when quarks are also taken into account . as already noted in the previous section , the contributions to the integral representation of the cross section stem from those values of @xmath20 and @xmath25 such that @xmath398 = 0 \;.\ ] ] which provides a relation between the moment index @xmath20 and the anomalous dimension variable @xmath25
. solving eq .
( [ impleq ] ) for either @xmath20 or @xmath25 defines the _ effective characteristic function _ and its dual _ effective anomalous dimension _ @xmath399 while in the one - channel case we have only one perturbative branch of those functions , corresponding to the bfkl eigenvalue function @xmath400 and to the larger eigenvalue @xmath401 of the anomalous dimension matrix , in the matrix formulation we expect two branches
. the second branch corresponds to the smaller eigenvalue @xmath402 which is dual to a second effective characteristic function @xmath403 . in fig .
[ f : chieff ] we show the two branches of the effective characteristic functions obtained in the nl@xmath3-nlo and nl@xmath3-lo cases .
we have considered here the asymmetric @xmath20-shift corresponding to the energy - scale @xmath77 , with @xmath404 .
the @xmath405 s are characterised by the typical minimum around @xmath406 whose value is nothing but @xmath38 . on the other hand ,
the @xmath407 s appear as steeply decreasing functions located around @xmath408 in the region shown in our plots . the continuity of the resummation procedure when going from the one - channel to the two - channel formulation at @xmath175 is illustrated in fig . [ f : chieff]a by the overlapping of the ( @xmath409)-branch of the nl@xmath3-lo curve to the circles corresponding to the one - channel scheme - b effective eigenvalue function .
the nlo terms provide a slight increase of the ( @xmath409)-branch in the region @xmath410 , and a small decrease of the ( @xmath411)-branch at @xmath412 . at @xmath397
there is a crossing point of the two branches at negative @xmath25 in either approximations .
the inclusion of quarks removes the crossing ( with a mechanism similar to the degenerate level splitting in quantum mechanics ) causing @xmath407 to be always on the left of @xmath405 , as can be seen in fig .
[ f : chieff]b .
quantitatively , the quark contribution lowers both @xmath405 in the region around the minimum ( compare the two inserts in fig .
[ f : chieff ] ) and @xmath407 .
note the two fixed points at @xmath413 = @xmath414 and @xmath415 of the ( @xmath409)-branches . in the one - channel case these fixed points corresponds to momentum conservation in the collinear and anti - collinear limits respectively .
in the two - channel formulation they imply that the anomalous dimension eigenvalue @xmath416 ; however , momentum sum rule is satisfied provided the corresponding left - eigenvector @xmath417 of the anomalous dimension matrix ( cf .
( [ vplus ] ) ) be @xmath418 at @xmath238 , i.e. , provided @xmath419 .
actually , our matrix model presents a small violation of the momentum sum rule .
in fact , by exploiting the fact that @xmath416 and by using eqs .
( [ eq : dglapeq ] ) and ( [ gammaspectral ] ) we have ( at @xmath238 ) @xmath420 the computation of the prefactor @xmath421 versus @xmath29 shown in tab .
[ t : mnc ] gives us an estimate of the relative amount of momentum non - conservation ; the violation is of order @xmath422 for the nl@xmath3-lo scheme , and of order @xmath289 for the scheme with nlo terms included .
._estimate of momentum sum rule violation .
the quantity @xmath423 at @xmath142 has been computed for various values of @xmath29 ( column 1 ) in the nl@xmath3-lo ( column 2 ) and nl@xmath3-nlo ( column 3 ) schemes .
column 4 ( resp .
5 ) shows that the nl@xmath3-lo ( nl@xmath3-nlo ) violation is of order @xmath422 ( @xmath289 ) .
[ t : mnc ] _ [ cols="^,^,^,^,^",options="header " , ] we close this section by showing in table [ tab : mom - sum - viol - running ] the degree of momentum sum - rule ( msr ) violation in the splitting functions for three values of @xmath29 . from
just a small number of values it is difficult to establish the exact scaling law , , because as @xmath29 decreases one needs ever higher relative accuracy to accurately determine the rapidly vanishing msr - violating component . ] and in particular it is difficult to determine the relative admixture of higher - twist and perturbative components in the msr violations .
nevertheless , one sees that the msr violation vanishes very rapidly as @xmath29 decrease , suggesting that a significant component of it is non - perturbative in origin .
this conclusion is borne out by studies which show that the amount of msr violation depends somewhat also on @xmath424 , the infrared cutoff scale for the coupling .
we have proposed here a matrix evolution equation for the flavour singlet , unintegrated quark and gluon densities , which generalizes the dglap and bfkl equations in the relevant limits . the matrix approach ( secs .
[ sec : bmf ] and [ sec : knla ] ) is supposed to unify collinear and high - energy factorizations in both partonic channels , and is not necessarily guaranteed to actually work , because of the various crossed consequences that the above factorizations have : consider , for instance , the anomalous dimension resummation formulae arising from @xmath1-factorization @xcite and the @xmath24 symmetry of the bfkl kernel @xcite arising from collinear factorization .
it is therefore a nontrivial result of this paper that our resummed splitting functions do satisfy collinear factorization in matrix form , as shown in secs .
[ s : fcad ] and [ s : nrrc ] . in this respect ,
our approach defines , by the matrix evolution , some unintegrated densities that are appropriate both in the collinear and in the small-@xmath3 limits . it would be interesting to explore the relationship of such explicit construction with alternative studies @xcite .
furthermore , we want to incorporate exact low - order anomalous dimensions in our matrix kernel , say in the @xmath2 scheme .
we find , in this context , a new kind of consistency relations on the kernels , due to a possible clash of exact low - order expressions with a novel nl@xmath3 resummation formula for @xmath36 , arising in the matrix evolution ( sec .
[ s : fcad ] ) .
we prove such relations to be satisfied by our construction in the @xmath2 scheme at nlo , but marginally violated by @xmath10-suppressed terms at nnlo .
we are thus able to complete our construction with exact nlo anomalous dimensions and nl@xmath3 kernel , and we postpone the analysis of the nnlo accuracy , which is however nearly incorporated ( in the nl@xmath3 approximation ) in our nlo@xmath425 version .
the frozen-@xmath29 features of our matrix model are characterized by the previously mentioned resummation formulae of sec .
[ s : fcad ] , and by the hard pomeron exponent and effective eigenvalue functions of sec .
[ s : cfrgf ] .
one should notice the basic continuity of our matrix approach with the single - channel case in the @xmath269 limit , and the corresponding agreement of the leading effective eigenvalue function with the abf approach .
additionally , we provide here the subleading effective eigenvalue at @xmath426 , corresponding to the @xmath338 eigenvalue of the anomalous dimension .
we are finally able to provide the whole matrix of resummed splitting functions in sec .
[ s : nrrc ] . roughly speaking
, the outcome shows that resummation effects are small in the @xmath5 entries up to @xmath3-values as small as @xmath427 , while the shallow dip is the main qualitative feature of both @xmath7 entries , with resummation effects starting below @xmath428 .
we notice that the above results are in the @xmath2 scheme up to nlo ( including approximately nnlo in their nlo@xmath425 version ) , but generally differ from it at higher orders .
however , resummation effects in the scheme change have been studied in @xcite and turn out to be of the order of the scale uncertainty . for this reason
we believe that the above results can be safely used in the study of structure functions and other cross - sections , by supplementing them with the corresponding coefficient functions or impact factors .
we wish to thank guido altarelli , richard ball , stefano catani , stefano forte and al mueller for various conversations on topics related to this paper .
we are grateful to the galileo galilei institute for theoretical physics in arcetri for hospitality during the workshop on `` high density qcd '' , while part of this work was being done .
this paper is supported in part by a prin grant from miur ( italy ) and by the french agence nationale de la recherche , grant anr-05-jcjc-0046 - 01 .
a.m.s has been supported by the polish committee for scientific research grant no .
kbn 1 p03b 028 28 .
in this appendix we show how eqs .
( [ eq : gamma_from_ck ] ) are obtained .
let us first rewrite eqs .
( [ eq : integrateddensities],[eq : dglapeq ] ) in @xmath25-space , by noting that @xmath429 : @xmath430 where `` h.t . '' stands for higher - twist contributions characterised by being regular at @xmath131 .
it follows that , in matrix notation , @xmath431 and , by induction , @xmath432 secondly , we consider eq .
( [ eq : genbfkleq ] ) and expand the matrix kernel @xmath51 in powers of gamma ( according to the notations following eq .
( [ eq : kexpansion ] ) ) , obtaining @xmath433 by comparing eqs .
( [ dglapgammaspace ] ) and ( [ eqbfklgammaexpnsn ] ) we derive the implicit equation @xmath434 which allows us to determine the effective anomalous dimension matrix @xmath106 in terms of the matrix kernel @xmath51 .
it is now straightforward to compute the perturbative coefficients @xmath56 defined in eq .
( [ eq : gammaexpansion ] ) . by expanding eq .
( [ gammakrel ] ) to first order in @xmath435 yields ( @xmath436 ) @xmath437[orderalh1]from which eq .
( [ gamma_0 ] ) follows . by expanding eq .
( [ gammakrel ] ) to second order in @xmath435 yields ( @xmath438 ) @xmath439 at frozen coupling , the operators @xmath440 and @xmath435 commute . by then collecting the @xmath441 terms and remembering that @xmath442 we get @xmath443 from which eq .
( [ gamma_1 ] ) follows .
a similar iteration procedure produces eq .
( [ gamma_2 ] ) and higher orders . in the running coupling case
, we have an additional commutator term starting at second order , namely @xmath444 { \mathsf{k}}_0^{(0 ) } \;.\ ] ] by using the expansion @xmath445 the commutator reads @xmath446 = - \beta_0 { \hat{\alpha}}^2 [ { \mathsf{k}}_0^{(1 ) } , \log\textstyle{\frac{k^2}{\mu^2 } } ] + { { \cal o}\left({\hat{\alpha}}^3\right ) } \;,\ ] ] thus producing , by eq .
( [ commuterm ] ) , a contribution of order @xmath447 to the anomalous dimension matrix .
note however that the nl@xmath3 term of order @xmath290 vanishes , because the @xmath9 entry of @xmath122 has no leading @xmath63 term . extending the above procedure to higher orders
, we see that at each order ( say n@xmath448lo ) the anomalous dimension gets a new term @xmath449 but also a series of other terms which are combinations of the anomalous dimensions at the lower orders @xmath450 .
singlet anomalous dimensions and splitting functions at lowest order appear in many of our formulas .
the former are given by ( @xmath451 ) @xmath452 \label{pqq } \\
p_{qg,0}(z ) & = 2 { t_f}[z^2 + ( 1-z)^2 ] \label{pqg } \\
p_{gq,0}(z ) & = c_f \frac{1+(1-z)^2}{z } \label{pgq } \\
p_{gg,0}(z ) & = 2{c_a}\left [ \frac1{z } + \frac1{(1-z)_+ } -2+z - z^2 \right ] + \frac{11{c_a}- 4{t_f}}{6 } \delta(1-z ) \;. \label{pgg}\end{aligned}\ ] ] the anomalous dimensions , i.e. , the mellin transforms of the splitting functions , are given by @xmath453 \label{gamma0_qq } \\
{ \gamma}_{qg,0}({\omega } ) & = 2 { t_f}\left ( \frac1{{\omega}+1 } - \frac2{{\omega}+2 } + \frac2{{\omega}+3 } \right ) \label{gamma0_qg } \\ { \gamma}_{gq,0}({\omega } ) & = { c_f}\left [ \frac2{{\omega } } - \frac{2}{{\omega}+1 } + \frac1{{\omega}+2 } \right ] \label{gamma0_gq } \\ { \gamma}_{gg,0}({\omega } ) & = 2{c_a}\left [ \frac1{{\omega } } + \psi(1 ) - \psi({\omega}+1 ) - \frac2{{\omega}+1 } + \frac1{{\omega}+2 } - \frac1{{\omega}+3 } + \frac{11}{12 } \right ] - \frac{2{t_f}}{3 } \;. \label{gamma0_gg}\end{aligned}\ ] ] the gluon anomalous dimensions @xmath454 and @xmath455 are singular at @xmath42 .
the regular parts are defined by subtraction of the @xmath20-pole , namely @xmath456 their values at @xmath42 are @xmath457 a_{gq}(0 ) & a_{gg}(0 ) \end{pmatrix } = \begin{pmatrix } 0 & \frac{4 { t_f}}{3 } \\[2ex ] -\frac{3 { c_f}}{2 } & -\frac{11{c_a}+ 4{t_f}}{6 } \end{pmatrix } \;.\ ] ]
our method of resumming energy - scale dependent terms relies on the introduction of improved kernels whose characteristic functions @xmath458 are @xmath20-dependent . in general such characteristic functions
are symmetric in the @xmath459 transformation and have the following structure : @xmath460 \;,\ ] ] where the left - projection @xmath125 contains collinear ( and possibly higher - twist ) singularities only in the half - plane @xmath461 .
, since @xmath462 is the correct bjorken scaling variable in the collinear limit @xmath463 we are interested to .
this causes the @xmath20-shift to apply only on the @xmath464 argument of @xmath125 and asymmetric kinematical constraints in the @xmath465 variable as shown below . ]
the @xmath20-dependence in the argument of the second @xmath125 term imposes kinematical constraints on the longitudinal momentum fraction variable @xmath465 conjugated to @xmath20 .
in fact , by denoting by @xmath466 and @xmath467 the inverse mellin transforms of @xmath468 and @xmath125 respectively , we have @xmath469 where @xmath470 , @xmath471 and @xmath472 . since @xmath473 , the kinematical constraint @xmath474 follows .
the lowest - order matrix kernel @xmath105 in @xmath475-space can be derived from eq .
( [ eq : chi_mat_0_all_prerequisites ] ) and reads @xmath476(z;k , k ' ) & = { \hat{\alpha}}(q^2 ) \begin{pmatrix } 0 & 0 \\[2ex ] 0 & 2{c_a}k_0(z;k , k ' ) \end{pmatrix } \\
\nonumber & + { \hat{\alpha}}(k_>^2 ) \left [ z ' \begin{pmatrix } p_{qq,0}(z ' ) & p_{qg,0}(z ' ) \\[2ex ] p_{gq,0}(z ' ) & p_{gg,0}(z ' ) -\frac{2{c_a}}{z ' } \end{pmatrix } k_c(k , k ' ) + z ' \begin{pmatrix } 0 & { \delta_{qg}}(z ' ) { k_{\mathrm{ht}}}(k , k ' ) \\[2ex ] 0 & 0 \end{pmatrix } \right]\end{aligned}\ ] ] having defined @xmath477 the terms in square brackets in eq .
( [ k0zk ] ) correspond to the operator @xmath478 introduced in eq .
( [ integraleq ] ) .
the action of the first term @xmath479 on a test function @xmath480 is @xmath481 \;,\end{aligned}\ ] ] while the action of the terms @xmath482 , e.g. the higher - twist one , is @xmath483 in order to obtain @xmath484 according to eq .
( [ eq : k1ansatz ] ) , we start by computing the first term proportional to @xmath485 .
the inverse mellin transform of @xmath486 is just the matrix of the two - loop singlet splitting functions in the @xmath2-scheme @xcite .
the inverse mellin tranform of the subtraction @xmath487 can be either computed by inverting the expressions listed in eq .
( [ k01k00ij ] ) , or by convolution in @xmath465-space of the corresponding factors . here
we choose the second method , by computing first the analytic expressions of all factors in eq .
( [ k01k00ij ] ) and then the numerical convolution of the ensuing functions .
we already obtained the mellin transform of @xmath192 in eq .
( [ dqgz ] ) ; the transforms of the @xmath156 factors are just the one - loop splitting functions reported in eqs .
( [ gamma0_qq]-[gamma0_gg ] ) ; the remaining functions are listed below : @xmath488\ , .
\label{inv_chiht}\end{aligned}\ ] ] the numerical coefficients @xmath492 can be found in eq .
( [ eq : azero ] ) , @xmath493 in eq .
( [ eq : delta_qg_msbar ] ) and @xmath242 in eq .
( [ chi1t0 ] ) .
furthermore , the computation of the kernel @xmath494 requires the subtraction from the nl@xmath3 bfkl kernel @xcite of the running coupling terms and of additional kernels corresponding to the characteristic functions on the r.h.s .
( [ eq : chireg ] ) .
they are given by @xmath495\ , , \label{k_0dot0 } \\
\chi_0\chi_c & \to \frac1{k_>^2 } \log \big(\frac{k_>^2}{k_<^2}-1\big ) - \frac1{k_<^2 } \log \big(1-\frac{k_<^2}{k_>^2}\big)\ , , \label{k_0c } \\
\chi_c^2 & \to \frac1{k_>^2}\left ( \log\frac{k_>^2}{k_<^2 } + 2 \right)\ , , \label{k_c_sq } \\ \chi_c { \chi_{\mathrm{ht}}}&\to \frac1{k_>^2}\left ( 1 - \frac1{3}\frac{k_<^2}{k_>^2 } \right)\ , .
\label{k_c_ht}\end{aligned}\ ] ] the resulting expression for @xmath494 is @xmath496 + \right . \nonumber \\
& - \frac{1}{32 } \left ( 1 + \frac{2{t_r}{n_f}}{{c_a}^3 } \right ) \left[\frac{2}{{k'}^2 } + \frac{2}{k^2 } + \left(\frac{1}{{k'}^2 } - \frac{1}{k^2 } \right ) \log\left(\frac{k^2}{{k'}^2}\right)\right ] + \nonumber \\ & \left . - \left[3 + \left(\frac{3}{4 } - \frac{({k'}^2+k^2)^2}{32{k'}^2 k^2}\right ) \right ] \left ( 1 + \frac{2{t_r}{n_f}}{{c_a}^3 } \right ) \int_0^{\infty }
\frac{{\mathrm{d}}y}{k^2 + y^2 { k'}^2 } \ ; \log\left|\frac{1+y}{1-y}\right| \right\ } + \nonumber \\ & + \frac{3}{2 }
\zeta(3 ) \delta(k^2-k'^2 ) + \frac{4 { { \mathrm{li}_2}}(1-k_{<}^2/k_{>}^2)}{|{k'}^2 - k^2| } + \nonumber \\ & -4 \frac{a_{gg}(0)}{2{c_a}}{\rm sgn}({k}^2-{k'}^2 ) \left ( \frac{1}{k^2 } \log\frac{|{k'}^2-k^2|}{{k'}^2 } - \frac{1}{{k'}^2 } \log\frac{|{k'}^2-k^2|}{{k}^2 } \right ) + \nonumber \\ & -\frac{{c_f}}{{c_a } } \frac1{k_>^2 } \left [ \frac{{\gamma}_{qg,0}(0)}{2{c_a } } \left ( \log\frac{k_>^2}{k_<^2 } + 2 \right ) + \frac{{\delta_{qg}}}{2{c_a}}\left ( 1 - \frac1{3}\frac{k_<^2}{k_>^2 } \right ) \right ] - \tilde{\chi}_1^{(0 ) } \frac1{k_>^2 } \;. \label{k1reg}\end{aligned}\ ] ] where @xmath497 denotes the azimuthal average of the ll@xmath3 bfkl kernel whose action on a test function @xmath498 is given by @xmath499(k ) = \int{\mathrm{d}}k'{}^2 \ ; \frac{1}{|{k'}^2-k^2| } \left [ f(k ' ) - \frac{2 k_{<}^2}{k'{}^2 + k^2 } f(k ) \right ] \;.\ ] ] finally , we provide the eigenvalue function of @xmath494 with kinematical constraints , which is given by @xmath500 , and occurs directly in eq .
( [ runningkernel ] ) .
the left projection @xmath501 of the eigenvalue function ( [ eq : chireg ] ) can be computed starting from the expression of @xmath502 in eq .
( a.13 ) of ref .
@xcite and noticing that : + _ ( i ) _ here we have more terms to subtract , namely those proportional to @xmath503 and @xmath242 ; + _ ( ii ) _ in ref .
@xcite we subtracted the @xmath504-part of the double pole by letting @xmath505 , while here we keep only @xmath506 in front of @xmath507 since the @xmath504-dependent double pole is subtracted by the @xmath508 term .
therefore , in order to complete the calculation we need the left projections of the following kernels : @xmath509_l({\gamma } ) & = [ \chi_{cl}({\gamma})]^2 + 2 \chi_{cl}({\gamma } ) = \frac1{{\gamma}^2 } + \frac{2}{{\gamma}}\ , , \\ [ \chi_c { \chi_{\mathrm{ht}}}]_l({\gamma } ) & = \chi_{cl}({\gamma } ) - \frac1{2}{\chi_{\mathrm{ht}}}{}_l({\gamma } ) = \frac1{{\gamma } } - \frac1{3(1+{\gamma } ) } \;.\end{aligned}\ ] ] the final result is @xmath510 \left[\psi'({\gamma } ) - \frac{a_{gg}(0)}{2{c_a}}\chi_c({\gamma } ) + \frac{67}{36}-\frac{\pi^2}{12}-\frac{5}{18}\frac{{n_f}}{{c_a } } \right ] \nonumber \\ & \quad + \frac12\psi''({\gamma } ) + \pi({\gamma } ) - \phi_{l}({\gamma } ) + \frac{\pi^2}{8 } \left[\psi\big(\frac{1+{\gamma}}{2}\big ) -\psi\big(\frac{{\gamma}}{2}\big ) \right ] + \frac{3}{4}\zeta(3 ) \nonumber \\ & \quad + \frac1{32 } \left\ { -3m({\gamma } ) + \left(1+\frac{{n_f}}{{c_a}^3}\right ) \left [ \frac14\left(\frac1{{\gamma}^2}-\frac1{(1-{\gamma})^2}\right ) -\frac12\left(\frac1{{\gamma}}-\frac1{1-{\gamma}}\right ) \right .
\nonumber \\ & \qquad\qquad\qquad\qquad\qquad\qquad\qquad\ ; \left.\left .
+ \frac1{32}\big(m({\gamma}+1)+m({\gamma}-1)\big ) -\frac{11}{16}m({\gamma } ) \right ] \right\ } \nonumber \\ & \quad -\frac{{c_f}}{2{c_a}^2}\left [ { \gamma}_{qg,0}(0 ) \left ( \frac1{{\gamma}^2 } + \frac{2}{{\gamma } } \right ) + { \delta_{qg}}\left ( \frac1{{\gamma } } - \frac1{3(1+{\gamma } ) } \right ) \right ] - \frac{\tilde{\chi}_1^{(0)}}{{\gamma } } \ ; , \label{chi1regl}\end{aligned}\ ] ] where @xmath511 \;.\end{aligned}\ ] ] the explicit form of @xmath494 in ( @xmath512 ) space is obtained by use of eq .
( [ eq : k_z_kt ] ) , or by introducing the kinematical constraints on eq .
( [ k1reg ] ) directly .
the final form of @xmath513 follows from eq .
( [ runningkernel ] ) .
gribov and l.n .
lipatov , sov .
. phys . * 15 * ( 1972 ) 438 ; + g. altarelli and g. parisi , nucl .
b * 126 * ( 1977 ) 298 ; + yu.l .
dokshitzer , sov .
jetp * 46 * ( 1977 ) 641 .
lipatov , sov .
* 23 * ( 1976 ) 338 ; + e.a .
kuraev , l.n .
lipatov and v.s .
fadin , sov .
jetp * 45 * ( 1977 ) 199 ; + i.i . balitsky and l.n .
lipatov , sov .
* 28 * ( 1978 ) 822 ; + l.n .
lipatov , sov .
jetp * 63 * ( 1986 ) 904 .
v. s. fadin and l. n. lipatov , jetp lett .
* 49 * ( 1989 ) 352 [ yad . fiz .
* 50 * ( 1989 sjnca,50,712.1989 ) 1141 ] ; nucl .
b * 406 * ( 1993 ) 259 ; nucl .
b * 477 * ( 1996 ) 767 .
+ v. s. fadin , r. fiore and a. quartarolo , phys .
d * 50 * ( 1994 ) 2265 ; phys .
d * 50 * ( 1994 ) 5893 .
+ v. s. fadin , r. fiore and m. i. kotsky , phys .
b * 359 * ( 1995 ) 181 ; phys .
b * 387 * ( 1996 ) 593 ; phys .
b * 389 * ( 1996 ) 737 .
+ v. s. fadin , m. i. kotsky and l. n. lipatov , budker - inp-1996 - 92 , arxiv : hep - ph/9704267 . + v. del duca , phys . rev .
d * 54 * ( 1996 ) 989 ; phys .
d * 54 * ( 1996 ) 4474 .
fadin , r. fiore , a. flachi and m.i .
kotsky , phys .
b * 422 * ( 1998 ) 287 .
s. catani , m. ciafaloni and f. hautmann , phys .
b * 242 * ( 1990 ) 97 ; nucl .
b * 366 * ( 1991 ) 135 .
+ g. camici and m. ciafaloni , phys .
b * 386 * ( 1996 ) 341 ; nucl .
b * 496 * ( 1997 ) 305 [ erratum - ibid .
b * 607 * ( 2001 ) 431 ] .
v. s. fadin and l. n. lipatov , phys .
b * 429 * ( 1998 ) 127 .
g. camici and m. ciafaloni , phys .
b * 412 * ( 1997 ) 396 [ erratum - ibid .
b * 417 * ( 1998 ) 390 ] ; phys .
b * 430 * ( 1998 ) 349 .
s. catani and f. hautmann , nucl .
b * 427 * ( 1994 ) 475 .
g. p. salam , jhep * 9807 * ( 1998 ) 019 .
m. ciafaloni and d. colferai , phys .
b * 452 * ( 1999 ) 372 .
m. ciafaloni , d. colferai and g. p. salam , phys .
d * 60 * ( 1999 ) 114036 .
m. ciafaloni , d. colferai , g. p. salam and a. m. stato , phys .
d * 68 * ( 2003 ) 114003
. m. ciafaloni , d. colferai , g. p. salam and a. m. stato , phys .
b * 576 * , ( 2003 ) 143 .
m. ciafaloni , d. colferai , g. p. salam and a. m. stato , phys .
b * 587 * , ( 2004 ) 87 .
m. ciafaloni and d. colferai , jhep * 0509 * ( 2005 ) 069 .
m. ciafaloni , d. colferai , g. p. salam and a. m. stato , phys .
b * 635 * ( 2006 ) 320 .
g. altarelli , r. d. ball and s. forte , nucl .
b * 575 * , ( 2000 ) 313 ; g. altarelli , r. d. ball and s. forte , nucl .
b * 599 * , ( 2001 ) 383 .
g. altarelli , r. d. ball and s. forte , nucl .
b * 621 * , ( 2002 ) 359 ; + g. altarelli , r. d. ball and s. forte , nucl .
b * 674 * , ( 2003 ) 459 ; + g. altarelli , r. d. ball and s. forte , nucl .
b * 742 * , ( 2006 ) 1 .
thorne , _ phys .
rev . _ * d 64 * ( 2001 ) 074005 ; _ phys .
* b 474 * ( 2000 ) 372 .
s. alekhin _ et al .
_ , `` hera and the lhc - a workshop on the implications of hera for lhc physics : proceedings part a , '' arxiv : hep - ph/0601012 .
c. d. white and r. s. thorne , phys .
d * 75 * , ( 2007 ) 034005 .
b. andersson , g. gustafson , h. kharraziha and j. samuelsson , _
z. phys .
_ * c 71 * ( 1996 ) 613 .
m. ciafaloni , _ phys .
* b 429 * ( 1998 ) 363 ; m. ciafaloni and d. colferai , _ nucl .
* b 538 * ( 1999 ) 187 .
a. vogt , s. moch and j. a. m. vermaseren , nucl .
b * 691 * ( 2004 ) 129 .
g. curci , w. furmanski and r. petronzio , nucl .
b * 175 * , ( 1980 ) 27 ; w. furmanski and r. petronzio , phys .
b * 97 * , ( 1980 ) 437 . | we propose a matrix evolution equation in @xmath0-space for flavour singlet , unintegrated quark and gluon densities , which generalizes dglap and bfkl equations in the relevant limits .
the matrix evolution kernel is constructed so as to satisfy renormalization group constraints in both the ordered and antiordered regions of exchanged momenta @xmath1 , and incorporates the known nlo anomalous dimensions in the @xmath2 scheme as well as the nl@xmath3 bfkl kernel .
we provide a hard pomeron exponent and effective eigenvalue functions that include the @xmath4-dependence , and give also the matrix of resummed dglap splitting functions .
the results connect smoothly with those of the single - channel approach .
the novel @xmath5 splitting functions show resummation effects delayed down to @xmath6 , while both @xmath7 entries show a shallow dip around @xmath8 , similarly to the @xmath9 single - channel results .
we remark that the matrix formulation poses further constraints on the consistency of a bfkl framework with the @xmath2 scheme , which are satisfied at nlo , but marginally violated by small @xmath10-suppressed terms at nnlo .
cern - ph - th/2007 - 104 + dff 436/07/07 + july 2007 * a matrix formulation for small-@xmath11 singlet evolution * m. ciafaloni@xmath12 , d. colferai@xmath13 , g.p .
salam@xmath14 and a.m. stato@xmath15 + @xmath12 _ cern , department ph - th , ch-1211 geneva 23 , switzerland
_ + @xmath13 _ dipartimento di fisica , universit di firenze , 50019 sesto fiorentino ( fi ) , italy _ ; + _ infn sezione di firenze , 50019 sesto fiorentino ( fi ) , italy _
+ @xmath14 _ lpthe , universit p. et m. curie
paris 6 , universit d. diderot
paris 7 , cnrs umr 7589 , paris , france _
+ @xmath15 _ department of physics , pennsylvania state university , university park , 16802 pa , usa _ ; + + |
@xcite found families of discrete velocities in normal galaxies and fitted them to a model in which they were all harmonically related to @xmath1 the speed of light . in his model
the discrete components detected are octave - spaced sub - harmonics of @xmath1 , and related harmonics , and the redshift is assumed to arise from a time - dependent decay from an origin at the planck scale .
his observed velocity periods can be accurately described by the relation p = @xmath2 , where d is the number of doublings and t defines the different families .
there can be many different families but the most common , or basic , of these make up what is referred to by @xcite as the t = 0 family of discrete values .
these were found in common spiral galaxies .
they showed discrete velocity components near 4.5 , 9 , 18 , 36 , 72 , 145 , 290 , etc .
, km s@xmath3 ( hereafter tifft s t=0 period ) .
although the other period groups were less common , they showed this same octave - related , or doubling nature .
we have recently examined the radial velocities and distances of 55 spiral galaxies and 36 type ia supernovae ( sneia ) galaxies with accurate distances @xcite .
both groups showed evidence for discrete , intrinsic velocity components related to those reported by tifft .
however , the galaxies we studied extended to much larger distances and showed that the entire radial velocity was not quantized , as previously thought by tifft . instead , the discrete components _ were superimposed on top of the hubble flow . _
the observed discrete spectral shifts were always positive ( redshifts ) .
the origin of these discrete velocities is unknown .
when the intrinsic components were removed from the measured radial velocities , a hubble constant of h@xmath4 = 58 km s@xmath3 mpc@xmath3 was obtained for the local universe , and the velocity - distance plot contained less scatter .
this h@xmath5-value is 20% lower than that obtained in the hubble key project @xcite .
the difference between these two hubble constants is due to the fact that @xcite interpreted the scatter on the hubble plot as random peculiar velocities containing both positive and negative doppler components .
because what is meant by `` peculiar velocity '' is of fundamental importance to this analysis we define the following two terms to avoid future confusion .
the term `` peculiar velocities '' is used here in the conventional sense to refer to the velocity scatter about the hubble line and will hereafter be referred to as pvs . because we assume here that these pvs may contain discrete intrinsic components that can cause them to fall along lines that run parallel to the hubble line , our analysis involves the determination of the rms deviation relative to the nearest of these intrinsic lines instead of one hubble line .
we define the scatter about these discrete lines as the `` true peculiar velocities '' and they are hereafter referred to as tpvs _ _ here we examine the radial velocities of 83 sci galaxies .
we use the same analysis procedure as used previously so the results can be easily compared .
we first describe the procedure in detail in order to convince the reader that we are really detecting discrete velocity components .
b- and i - band tully - fisher ( tf ) distances , and v@xmath0 velocities for the 83 sci galaxies were obtained from table vi of @xcite .
the tf distances for these galaxies were calculated utilizing the type dependent tf relations of @xcite , who found that galaxies of morphological types sbci , sbc i - ii , sci , sci - ii and seyferts have systematically underestimated tf distances utilizing a standard single tf equation .
the 83 sci galaxies were drawn from the samples of @xcite and @xcite and are all classified as sbc or sc of luminosity class i or i - ii in the leda database .
tf scatter was significantly reduced with the use of type dependent tf relations for the calibrator samples @xcite and for cluster samples @xcite .
the tf scatter of the calibrators in the b - band was only @xmath6 mag for the sci group and @xmath7 mag for the sb / sciii group when using the type dependent tf relations compared with @xmath8 mag with a traditional single calibration @xcite .
the scatter quoted here represents the empirical scatter of the tf relation relative to calibrator .
from our previous work and tifft s , we can make the following initial assumptions to limit the number of free parameters .
first , the hubble slope obtained must be h@xmath4 = 58 km s@xmath3 mpc@xmath3 .
second , since we are examining spiral galaxies , the discrete components present will be predominantly those predicted by tifft s most common ( t = 0 ) group , and third , the discrete components will be superimposed on top of the hubble flow .
these assumptions severely restrict what the data must show if our previous results are to be confirmed .
the first of these assumptions is an important one because it allows us to differentiate between real and random features . as in previous analyses @xcite , in a given source sample , v@xmath0 , the radial velocity of each source relative to the cosmic microwave background reference frame ,
was first plotted vs distance .
fig 1 shows such a plot using peculiar velocities ( pvs ) that were randomly generated .
the hubble slope used in the generation of this sample ( h@xmath4 = 85 ) was taken from the actual source distribution and is the slope obtained for the data when the lines in fig 1 are ignored and a linear relation is fitted to the data points . in this figure the solid lines represent tifft s relevant t = 0 discrete components superimposed on top of an h@xmath4 = 58 hubble slope .
since the superimposed velocities were randomly generated there should be no evidence in fig 1 that the source velocities fall along the discrete velocity grid lines , and clearly there is not .
as in all previous analyses we then calculated the rms deviation of the sources from their nearest discrete velocity grid line .
we then repeated this calculation for a range of hubble constants .
when the velocity distribution does not contain tifft s discrete components , then , as the hubble slope is varied , _
one _ rms minimum will be obtained .
_ this minimum will occur for that hubble slope that positions the densest portion of the discrete velocity grid lines on top of the densest portion of the source distribution . _ the resulting rms deviation plot is shown in fig 2 . as predicted , one rms minimum is obtained and it occurs near a hubble slope of h@xmath4 = 70 .
this then , is how the rms deviation plot is expected to appear when tifft s discrete velocity components are @xmath9 present in the radial velocities . _ _ _ _ we now demonstrate how the rms deviation plot in fig 2 will be affected if tifft s
t = 0 discrete velocities @xmath10 present in the source velocities . to do this
we generated a second data set containing tifft s t = 0 components superimposed on top of the hubble slope .
the generated velocities contained no distance errors , no true peculiar velocities ( tpvs ) , and were superimposed on top of a hubble slope of h@xmath4 = 50 .
a slope of 50 was used to show that there is nothing magical about the value 58 .
fig 3 shows these velocities plotted versus distance , using the same distance values as in fig 1 . in fig 3 the velocities
are clearly aligned along tifft s t = 0 discrete velocity lines for h@xmath4 = 50 .
the rms deviation of the source velocities from their nearest discrete velocity lines in fig 3 was then calculated for a range of hubble slopes as was done in fig 2 .
as the hubble slope is varied , a low rms value will be obtained as before for that h@xmath4-value that positions the densest portion of the lines near the center of the source velocity distribution .
this is again expected to occur near h@xmath4 = 70 . however , with tifft s discrete components present in the data , a second dip in the rms deviation plot will occur at the hubble slope where the source velocities are positioned along the discrete velocity grid lines .
the resulting rms deviation plot is shown in fig 4 , where a second dip is now visible corresponding to a hubble slope of 50 , indicated by the arrow .
this second dip in the rms deviation curve at h@xmath4 = 50 in fig 4 is thus the defining signature of discrete velocities in the data .
the scatter in the source distances about the discrete velocity lines will determine the depth of the second rms dip .
this rms dip therefore does not have to be deeper than the first dip near h@xmath4 = 70 to indicate that discrete velocities are present . for real data ,
however , it does have to occur at , or very near , h@xmath4 = 58 , since this is the value found in all previous analyses of real galaxy data .
although it is theoretically possible for an rms dip to be produced by random fluctuations in the source distances and peculiar velocities , if these fluctuations are truly random , _ so will be the value of the hubble constant at which the rms dip occurs .
_ also , the amplitude of random dips is expected to decrease , as the number of sources in the sample increases . _ _
since , as noted above , the depth of the discrete component signature ( i.e. the second rms dip ) is directly proportional to the size of the distance uncertainties , as these increase the second dip will become shallower , until it eventually disappears , as in fig 2 .
there are good reasons to suspect that some of the sources will have larger distance errors than others .
as long as these sources are present in the sample they are expected to reduce the visibility of the second rms dip .
we therefore first looked at a test sample drawn from the 83 sci galaxies that we have a good reason to believe will have some of the most accurate distances . by taking the difference between the b- and i - band distances for each source it is found that these differences range from 0 to 57.6 mpc for the 83 sci galaxies .
clearly , from this result , some distances have to have large errors . by making the simple assumption that the distances are likely to be more accurate when both b and i - band distances agree
, we obtained a sample that we felt should have accurate distances by choosing the 23 sources with the smallest b - i differences .
the mean b - i difference for this sample was 0.8 mpc . in fig 5 the v@xmath0 velocities of these 23 sources
are plotted vs their b - band distances .
t = 0 discrete velocity lines are included on the plot .
a solid line joins sources assumed to contain the same discrete velocity component .
it is apparent from this plot that the sources are aligned along the lines representing the discrete velocity components for a hubble slope near 58 .
if most of the scatter in the plot is then due to the discrete intrinsic components , the real dispersion ( due to tpvs and distance errors ) will be given by the scatter about the discrete lines .
it must be much smaller and will not be clearly visible until the discrete components are removed ( see fig 7 ) .
the measurement error for velocity is likely to be less than 35 km s@xmath3 @xcite .
in fact , the mean velocity error in fig 5 is 17.6 km s@xmath3 from the velocity errors listed in the leda database for these sources .
although magnitude measurement errors are also known , they are expected to bear little resemblance to the true distance errors that are obtained after the various corrections are applied .
the rms deviation plot for the sources in fig 5 is given by the solid curve in fig 6 and it also clearly shows the signature of tifft s discrete velocities by the presence of a second rms dip near h@xmath4 = 58 . also plotted in fig 6 are two other curves .
the dashed curve is the rms deviation plot obtained by replacing the pvs of the 23 sources in fig 5 with randomly generated ones .
the dotted curve is an rms deviation plot obtained by retaining the original pvs and generating random distance errors .
both random data sets were generated so that the resulting scatter was similar to that for the real data ( assuming one hubble line with no discrete components ) .
both plots obtained using randomly generated data show only one minimum .
no second minimum is expected since tifft s t = 0 components are not present in the randomly generated data . in fig 7
the discrete components ( as defined by the lines in fig 5 ) have been removed from the radial velocities leaving only the component v@xmath11 , due to the hubble flow , with its uncertainty in distance and any small tpvs remaining .
the dashed lines in fig 7 are assumed to represent the peak - to - peak scatter in the distances , as a function of distance , for the 23 galaxies with smallest b - i differences .
the peak - to - peak scatter in fig 7 corresponds to a 1 @xmath12 uncertainty in the distance modulus of @xmath13 mag , if h@xmath4 = 58 , assuming no tpvs .
after conversion to distance , these uncertainties ( 1 sigma ) have been included in fig 7 .
this is consistent with the cepheid calibrators for which the 12 sci group calibrators had a b - band scatter of @xmath6 mag ( 1 sigma ) .
the 23 galaxies in fig 7 represent those galaxies in the sample that are likely to have some of the smallest distance uncertainties .
the scatter obtained for the entire sample of 83 galaxies is expected to be larger .
typical scatter , when type dependence is not accounted for is at least @xmath14 magnitude @xcite .
when type dependence is taken into account , as here , our value of @xmath13 magnitude is in good agreement with what is expected .
these results , after removal of the discrete velocity components , indicate that any tpv components present can not be large .
the solid line in fig 7 represents a linear fit to the data which gave a hubble constant of h@xmath15 .
previous analyses of galaxies gave values of h@xmath4 = 57.9 for sneia galaxies , 57.5 for sb galaxies and 60.0 for sc galaxies @xcite .
with equal weighting these results give a mean value of h@xmath4 = [email protected] km s@xmath17 mpc@xmath17 .
fig 8 shows the plots obtained for the rms deviation in v@xmath0 velocities for all 83 sci galaxies for both b- and i - band tf distances .
there is no evidence in the i - band curve for a second rms deviation dip at h@xmath4 = 58 , and the minimum rms deviation found is near 580 km s@xmath3 . in the b - band curve there is a clear second dip near h@xmath4 = 58 , indicated by the arrow , and its minimum rms deviation is now close to 420 km s@xmath3 . although this is much higher than the rms - value of 190 obtained for the 23 sources selected because their b and i - band distances were similar , this is not surprising since the full sample contains sources with a wide range of inclination angles and rotation velocities .
it is well known that distance errors can vary significantly with inclination @xcite .
one of the strongest arguments favoring our interpretation that the detection of a second dip in the rms deviation in v@xmath0 as h@xmath4 is varied is due to discrete velocities present in the data , is the fact that in the present analysis , as well as in all previous analyses where galaxy tf distances were examined , this second dip has always occurred at the same hubble slope of h@xmath4 = 58 .
although the occurrence of rms dips due to random clumping in the source distribution is possible , there are few arguments that can be made to explain the consistent appearance of randomly produced dips at the same hubble slope
. there would appear to be only two reasons why dips would always appear at the same hubble slope in the real data .
one is if the discrete components are real , and the other is if the analysis procedure is somehow affecting the result .
the latter case was examined extensively in test 2 of a previous paper @xcite , where it was clearly demonstrated , by averaging rms deviation curves for 10 randomly generated samples , that our reduction procedure was not introducing a second rms dip at a particular hubble slope . in test 1 of that paper
we also demonstrated that the randomly generated samples gave only one rms dip .
we note again that the depth of the second rms dip , the signature of the presence of discrete components , is unrelated to the overall scatter in the data ( i.e. to the size of the pvs generated ) .
it is entirely dependent upon how closely aligned the data points are to tifft s discrete velocity lines ( i.e. how small the tpvs and distance errors are ) .
if a second rms dip appears in the real data it means that the pvs can not be purely random , but must contain discrete components that align the sources along the `` discrete velocity '' lines .
we have further demonstrated again in fig 6 of this paper that the second rms dip is not seen when the discrete components are not present in the data , and it is therefore not produced by the analysis procedure . we have also demonstrated in figs 3 and 4 that the location of the second dip is determined by the data .
we conclude that the second rms dip at h@xmath4 = 58 that we have now detected in four completely separate galaxy groups is due to the presence of discrete velocity components in the radial velocities of these galaxies .
the fact that the discrete velocities found are identical to those reported previously by tifft strongly reinforces this claim .
we have examined the velocities and distances of a source sample containing 83 sci galaxies .
we find evidence that the velocity components of tifft s most common t=0 velocity period are present in the radial velocities of these galaxies .
we find for a particular sub - sample of 23 sci galaxies , chosen because their distances are likely to be accurate , that after the discrete velocities are removed , the distance scatter remaining in the hubble plot corresponds to a mean magnitude error in their distance modulus of @xmath18 magnitude .
this is in good agreement with the value expected when morphological type is taken into consideration .
we also find , for the 83 sources , that b - band distances are considerably more accurate than i - band . as a result of this
we suggest that our technique may be a valuable one to use to study the tully - fisher distance errors as a function of galaxy inclination and rotation velocity .
bell , m.b . & comeau , s.p .
2003a , astro - ph/0305112 bell , m.b . & comeau , s.p .
2003b , astro - ph/0305060 freedman et al .
2001 , , 553 , 47 giovanelli r. et al .
1997 , aj 113 , 22 jorgensen , i. , franx , m. and kjaergaard , p. 1995
, mnras , 276 , 1341 mathewson , d. & ford , v. 1996 , , 107 , 97 russell , d.g .
2003 , apss ( submitted ) , astro - ph/0310284 russell , d.g .
2004 , apj , 607 , 241 sakai , s. et al .
2000 , , 529 , 698 tifft , w.g .
1976 , , 206 , 38 tifft , w.g .
1977a , , 211 , 31 tifft , w.g .
1977b , , 211 , 377 tifft , w.g .
1996 , , 468 , 491 tifft , w.g .
1997 , , 485 , 465 tully , r.b . & pierce , m.j .
2000 , , 533 , 744 | we examine the v@xmath0 velocities and tully - fisher distances of 83 sci galaxies and find that their velocities contain the same discrete velocity components identified previously by tifft . after removing the discrete velocity components ,
the scatter remaining on the hubble plot is in good agreement with the tully - fisher distance uncertainties .
although there is , as yet , no physical explanation for these discrete components , the fact that they appear to be identical to those found previously by tifft suggests that their reality needs to be taken seriously . |
within the framework of the standard model of the electroweak and strong interactions the neutrinos are massless to all orders in perturbation theory . however , it is widely believed , based on the interpretation of current observations , that the neutrinos are massive and light , i.e. , @xmath31 - 100 ev ) @xcite .
if confirmed by future experiments ( e.g. , the next generation of underground observatories and the long - baseline oscillation experiments ) the massive neutrinos would provide an unambiguous signal of physics beyond the standard model ( sm ) . for example , the small neutrino masses are often attributed to some sort of a see - saw mechanism involving an intermediate scale @xmath4 gev ) right - handed neutrino ( or some other structure at that scale ) , i.e. , @xmath5 @xcite .
the neutrino mass in that case is a signal of the physics at intermediate scales .
a different and most interesting possibility is that the neutrino mass is a signal of weak - scale physics . indeed
, this could be the case within a supersymmetric framework @xcite as it could incorporate a new and novel ( but often overlooked ) mechanism for the generation of neutrino masses which was first proposed by hall and suzuki @xcite .
supersymmetric theories naturally accommodate weak - scale scalars and are a strongly motivated candidate for an extension of the sm at a scale @xmath6 gev@xmath7 tev ) @xcite .
thus , an investigation of neutrino mass generation within this framework is well motivated and would provide an alternative and unique interpretation of a neutrino mass signal . in turn
, the latter could provide valuable information on the supersymmetric structure both at weak and planckian scales . in this work
we will propose a new realization of that mechanism , argue for its naturalness within the context of supersymmetric model building , and examine its phenomenological and cosmological implications .
( we will assume the supersymmetric extension of the sm throughout this work . )
the neutrino mass generation follows in a straightforward manner if one allows a low - energy ( bilinear ) mass superpotential , @xmath8 , which is the most general one in the fields of the minimal supersymmetric extension ( mssm ) and the sm @xmath9 gauge group .
the ( renormalizable ) superpotential reads in that case @xmath10 where quark and lepton doublet chiral superfields by @xmath11 and @xmath12 , the up- and down - quark singlets by @xmath13 and @xmath14 , and the lepton singlets by @xmath15 , respectively . @xmath16 and @xmath17 denote the @xmath18 and @xmath19 higgs doublets , respectively .
a chiral superfield and its scalar component will be denoted below by the same symbol .
] @xmath20 is [ listing @xmath9 quantum numbers ] the ( 1,2,-1/2 ) four - vector in field space , the four - vector @xmath21 is of the order of magnitude of the weak scale and @xmath22 , and @xmath23 is the ( trilinear ) yukawa part of the superpotential .
one neutrino species ( see below ) is now massive at tree level as a result of the superpotential mass term ( in a similar manner to the fermion partners of the higgs bosons the higgsinos ) .
the details of the yukawa superpotential @xmath23 would only affect the loop - level corrections to the masses .
it is quite striking that the sm massless fermions
the neutrinos are the only sm fermions that are allowed by the symmetries to have a supersymmetric mass . in general , supersymmetrtic mass parameters are naturally of the order of the grand - scale or are zero , e.g. , in string theory and grand - unified models .
hence , low - energy mass parameters ( which are of the order of the weak - scale ) in ( [ sp1 ] ) represent a perturbation to the above expectation and they parameterize the high - energy physics in a similar way to the dimensionful supersymmetry breaking soft parametrs which set the scale @xmath22 ( see below for a discussion of the @xmath2-problem ) .
once mass parameters are introduced , the term @xmath24 ( rather than only @xmath25 , which is the only superpotential mass term included in the mssm case ) is the most natural choice .
the neutrino mass then arises from weak - scale parameters , but in fact , parameterizes the high - energy physics . naively , one might have expected that ( unless @xmath26 ) @xmath27 . however , a simple condition @xcite , which , as we will show , may be realized dynamically in universal models for the soft supersymmetry breaking parameters , guarantees that this is not the case and that neutrinos as light as @xmath28100 ev@xmath29 are obtained .
other proposals for weak - scale neutrino mass generation in a supersymmetric framework @xcite include mass generation due to the soft supersymmetry breaking sector , spontaneous lepton number breaking , and yukawa interactions .
the supersymmetric neutrino mass mechanism @xcite , that we will examine here , embodies various aspects of all other proposals .
the neutrino mass generation involves , in our case , several delicate issues which require some elaboration .
while doing so , we will establish the outline of our proposed models .
_ r - parity violation : _ the superpotential @xmath30 given in ( [ sp1 ] ) does not preserve lepton number , and thus , the discrete @xmath31 @xmath1-parity @xcite @xmath32 ( where @xmath33 , @xmath34 , and @xmath12 are the particle spin , baryon and lepton numbers , respectively ) that is typically imposed in the mssm ( and forbids weak - scale neutrino mass generation ) is also broken . unlike in generic broken @xmath1-parity models ,
the breaking here is restricted and does not lead to unacceptable proton - decay rates .
specifically , the ( tree - level ) neutrino mass generation is insensitive to the details of @xmath23 , and the yukawa superpotential could still have ( approximately ) its sm form and preserve this or some other symmetry so that the proton is long lived .
( in fact , for that purpose it is sufficient to preserve at the renormalizable level only baryon number , e.g. , this is the case in the @xmath35 baryon parity @xcite and similar @xcite models . ) for our purpose it is enough to consider @xmath36 where ( suppressing family indices ) @xmath37 below , we will show that this is indeed a natural choice in the models that we consider .
_ rotations and lepton - number redefinitions : _ lepton number violation can be rotated from @xmath8 onto @xmath23 , introducing the lepton number violating terms the yukawa couplings are sometimes denoted by @xmath38 and @xmath39 .
our notation aims at stressing the relation in our case between @xmath40 and @xmath41 ( see below ) . ]
@xmath42e , \label{wlnv}\ ] ] where commutation with respect to generation indices is implied in the second yukawa term ( which vanishes in the case of one lepton generation ) .
@xmath1-parity is now replaced , e.g. , by baryon parity .
such rotations are useful when discussing the implications of the model at low energies ( where all relevant symmetries are broken )
. the new yukawa couplings are not arbitrary in that case and lead to the same phenomenology .
however , rotations are impossible if @xmath30 is invariant under a symmetry which does not commute with the @xmath43 symmetry of @xmath44 rotations in the field space , i.e. , a symmetry which distinguishes the higgs and lepton superfields .
we will assume that this is the case at high energies , e.g. , at planckian scales .
we will show later that such a `` non - commuting '' symmetry is , in fact , a desired feature and that it can be related to the ad hoc absence of an arbitrary @xmath40 ( and also to the absence of lepton number violation in the soft terms ) .
in particular , one may not need to impose an additional symmetry on @xmath23 .
note also that the rotation leading to ( [ wlnv ] ) is not scale invariant and that @xmath45 will regenerate lepton number violation in @xmath8 via renormalization group scaling .
it suggests that one can define the sm lepton number consistently only at low - energies .
for that reason , and as a result of our symmetry assumption , we distinguish between the high- ( mssm ) and low - energy ( sm ) definitions of the leptons .
the former are defined by the superpotential ( [ wy ] ) while the latter will be chosen ( after all symmetries are broken and rotations are possible ) in most cases as the perpendicular directions to the relevant weak - scale expectation value .
this distinction will prove as a useful model - building tool . _
the @xmath2-problem and the choice of a symmetry : _ an arbitrary ( high - energy ) vector with a magnitude @xmath46 could lead to highly suppressed neutrino masses and there is no need to significantly suppress @xmath47 etc .
( this will be demonstrated in the following sections ) . in particular , the neutrino mass is related in our framework to the solution of the celebrated @xmath2-problem @xcite to which we alluded above , i.e. , explaining @xmath48 and @xmath49 , which is the manifestation of the correlation between the supersymmetry preserving [ with a natural scale of @xmath50 and softly breaking [ with an ad hoc scale @xmath51 ) ] sectors .
the correlation is imposed by requiring the correct electroweak symmetry breaking pattern @xcite .
( usually the problem is phrased in terms of @xmath52 . )
we will adopt a somewhat ambitious approach and require that the solution to the @xmath2-problem is determined by the same symmetry that was proposed above [ and which does not commute with the @xmath43 ] .
in particular , we will consider a continuous @xmath54 r - symmetry ( the @xmath1-charge is defined as above ) that is known to be relevant for the solution of the @xmath2-problem @xcite , and under which @xmath55 and @xmath56 effectively carry different @xmath1-charges . in principle , we could use other @xmath0 symmetries to break the @xmath43 , e.g. , a peccei - quinn symmetry @xcite or the stringy inspired @xmath0 s of ref .
@xcite . in the latter case @xmath21 could be forbidden by the new @xmath0 gauge symmetry but induced by weak - scale expectation values in a similar manner to the authors original proposal for the generation of @xmath55 .
we will concentrate in this work on the @xmath54 case .
generalization of our proposal to models in which other symmetries play a similar role in solving the @xmath2-problem is straightforward .
_ operator classification : _ in the symmetry framework @xmath55 and @xmath57 are realized as nonrenormalizable operators ( nro s ) present in the effective low - energy superpotential @xcite .
the @xmath1-symmetry allows us to classify three categories of mass terms in @xmath30 [ which is normalized transformation law @xmath58 , @xmath59 , and @xmath60 , for the scalar , fermion and auxiliary components of a chiral superfield @xmath61 with r - charge @xmath1 , respectively . ] to carry @xmath62charge @xmath63 : [ mus ] @xmath64 @xmath65 @xmath66 where we have assumed the standard @xmath1-charge assignments as above ( substituting @xmath67 ) , i.e. , @xmath68 and @xmath69
. thus , the @xmath1-symmetry provides us with a convenient book - keeping tool .
note that the proposal @xmath70 @xcite is trivially realized in this framework . generically
( but see exceptions below ) , one expects that the higher @xmath1-charge @xmath2 carries the more suppressed it is . furthermore , since lepton number violation is in the form of nonrenormalizable operators , and based on dimensional arguments , one also expects that , e.g. , @xmath71 .
[ note that @xmath72 . ]
thus , @xmath23 is effectively @xmath41 with an accidental discrete @xmath73 symmetry .
similar arguments hold for the dimensionless couplings in the kahler potential , and typically lepton number violation in the soft terms is also suppressed . as a result of the symmetry selection rules , @xmath1-number violation , and in particular , lepton number violation , is contained in @xmath8 as advocated .
the singlet @xmath2-parameters may not be suppressed , and the right handed neutrino @xmath74 ( or other singlets ) and vector - like exotic @xmath61 and @xmath75 supermultiplets with standard @xmath1-charge assignments are expected to be heavy .
( the latter points are , in fact , relevant for the dynamical suppression of the neutrino masses discussed below . )
we are now in a position to outline our proposal : the @xmath54 symmetry selection rules require that there is a @xmath76 hierarchy between the symmetry violation in the supersymmetric mass terms and in the yukawa terms and the kahler potential . as a result
, it leads to the generation of @xmath2 terms of the right magnitude , accidental @xmath73 symmetry in the yukawa terms , and to a consistent definition of the low - energy lepton number .
in addition , the neutrino masses are suppressed dynamically .
we will examine the above issues in greater detail and expand our discussion in the following sections .
in section ii we briefly discuss possible realizations ( and their problems ) of the @xmath54 symmetry , and present some simple examples that realize ( or are exceptions to ) the main features discussed above , i.e. , the suppression of non - minimal yukawa interactions and the hierarchy among the different @xmath2 terms .
we will show that the nature of the hierarchy depends on whether or not the relevant operators involve @xmath77-terms . in section iii
we discuss the neutral fermion mass matrix and the sufficient conditions for small neutrino masses , which include the alignment ( in field space ) of @xmath78 with @xmath21 .
we then proceed and show in section iv that the latter condition and @xmath79 are achieved in a straightforward manner in universal models with radiative ( electroweak ) symmetry breaking ( rsb ) @xcite via the `` dynamical alignment '' mechanism .
( for simplicity , we discuss a model with only the third generation of leptons and quarks . )
it is also pointed out that grand - unification relations , as well as an intermediate - scale right - handed neutrino , generically destabilize this result . in section
v we show that lepton number ( and individual lepton number ) violations in the model are typically proportional to either the neutrino mass or to small mssm yukawa couplings , and thus , they are suppressed .
we briefly address the issue of possible discovery channels , and stress that there could be an enhancement of dijet and multijet cross - sections .
cosmology and astrophysics implications are discussed in section vi .
it is shown that weak - scale neutrino mass generation implies that the supersymmetric partners ( superpartners ) of ordinary matter and gauge fields are not stable on cosmological time scales .
we will alternate in our discussion between the high - energy ( sections ii and iv ) and low - energy ( sections iii , v and vi ) definitions of the leptons and lepton number , as appropriate . in section vii we comment on possible family dependencies in more complete models and on their implications .
we conclude in section viii , where we suggest that the proposal for supersymmetric neutrino mass generation is most elegantly understood in terms of symmetry principles that are also relevant to the ( generalized ) @xmath2-problem , can be naturally incorporated in simple supersymmetric models , removes the need to ad hoc generate an intermediate scale for the right - handed neutrinos , e.g. , in string models , and requires one to consider unorthodox scenarios for supersymmetric particle astrophysics , cosmology , and for superpartner collider signatures .
we have chosen a spontaneously broken @xmath54 symmetry as our primary tool in deriving selection rules for the low - energy effective theory .
we also have shown above that it allows one to distinguish three classes of mass ( i.e. , @xmath2- ) parameters in the effective low - energy superpotential ; it distinguishes the ( high - scale ) lepton and higgs doublets ; and that , on the basis of dimensional arguments , non - singlet ( i.e. , lepton - number violating ) yukawa couplings are typically suppressed by an inverse power of the planck mass in comparison to non - singlet @xmath2-parameters .
the presence of such a symmetry is intimately connected to ( dynamical ) supersymmetry breaking @xcite and the @xmath1-symmetry could play a natural rule in the solution of the ( either mssm or our generalized ) @xmath2-problem if the @xmath54 symmetry is unbroken and the @xmath2-problem is trivially solved @xcite . however , such models are strongly constrained and face many phenomenological ( e.g. , generation of a gluino mass ) and model - building ( e.g. , realization of rsb ) difficulties ( see , for example , ref .
@xcite ) . ] @xcite .
hence , our choice of `` book - keeping '' tool is well motivated .
however , similarly to the peccei - quinn case @xcite , a spontaneously broken continuous @xmath1-symmetry implies an unwanted @xcite ( pseudo ) goldstone boson the @xmath1-axion .
the troublesome presence of the @xmath1-axion can be resolved @xmath80 if it is an invisible axion , i.e. , the @xmath1-symmetry is broken at a scale @xmath81 gev ( from stellar evolution ) and below the planck scale @xmath82 gev ( from its contribution to the energy density ) @xcite ( a higher bound @xmath83 may exist in some cases @xcite ) ; @xmath84 if the axion receives its mass from nro s that explicitly break the symmetry - charge and which is perhaps needed to cancel the vacuum energy . in that case
the axion mass is related to the size of the soft supersymmetry breaking parameters .
in fact , explicit breaking by a large constant in the gravitational sector is a common solution to the @xmath54 problem in supergravity models , e.g. , in the polonyi model @xcite . ]
@xcite ; @xmath85 if the symmetry is gauged ( and anomaly free ) at planckian scales @xcite ; or @xmath86 if the axion is rendered heavy by different means than the higgs mechanism which operates in @xmath85 . the first option could be the natural choice if the low - energy global symmetry results from an anomalous high - energy @xmath0 ( e.g. , in a string theory ) , and thus , could be stable with respect to gravitational corrections . in that case @xmath87 where the symmetry - breaking scale @xmath88 is the scale parameter that enters nro s with positive powers , and thus , controls the size of couplings in the low - energy effective theory .
if the axion scale is related to a typical supergravity breaking scale then it naturally falls in this range @xcite .
the second option is motivated by the observation that global symmetries are not likely to be exact in the presence of gravity or if the @xmath54 symmetry is accidental and due to a higher symmetry and renormalizability @xcite , but it implies that a priory we do not have an handle on the choice of nro s .
thus , it may undermine our motivation .
nevertheless , one could still distinguish @xmath89mev@xmath29 operators , that are sufficient to generate an acceptable axion mass , from @xmath90 operators , which are the relevant ones in our case .
[ recall that the @xmath91 s need not be suppressed necessarily more than the ordinary @xmath92 parameter and could be @xmath93 . ] in this scenario there are no constraints on @xmath88 .
( note that if the axion is hidden , i.e. , with only gravitational interactions with ordinary matter , then effects of any operator that may be needed to render it massive are suppressed in the observable sector by an additional inverse power of @xmath94 . )
the third option is quite attractive .
the symmetry is anomaly free and the goldstone boson is not an axion .
however , the anomaly cancelation equations involve the gauge fermions and the gravitino , in addition to the ordinary , exotic and hidden matter fermions ( and similarly the @xmath14-terms ) .
it was recently shown that the symmetry must be broken at a planckian scale so that one can tune to a flat @xmath14-term direction @xmath95 [ and avoid @xmath96 masses for ordinary fields ] @xcite .
thus , @xmath97 the anomaly equations are difficult to solve and require ( many ) new sm singlet fields and/or exotic fields with non - trivial sm charge assignments ( and , in some cases , non - standard and family - dependent @xmath1-charge assignments for the ordinary sm fields ) @xcite . while the singlet fields could be hidden in the hidden sector , @xmath96 mass terms for the ( observable - sector ) exotics are forbidden by the symmetry , and in the existing examples with exotics there are @xmath89tev@xmath29 colored fields .
we will not consider explicitly models with such fields and will constrain our investigation to the model with minimal matter content .
however , we note that unless such fields interact with @xmath44 , they are irrelevant for our purposes and do not alter our discussion .
lastly , the fourth option could be realized , e.g. , if the axion is a hidden - sector field and the @xmath54 is anomalous with respect to a hidden sector qcd group .
the axion would then acquire a large mass ( of order of the hidden sector qcd scale @xmath98 gev , assuming dynamical supersymmetry breaking in the hidden sector ) @xcite .
in this case one also expects @xmath99 below , we will not promote any option in particular , but comment on their different implications where relevant .
we will assume standard @xmath1-charge assignments , unless otherwise is specified , that the scale @xmath88 is in the range suggested by ( [ theta1 ] ) ( [ theta3 ] ) , and that the non - vanishing fields are hidden fields ( with only gravitational interactions with ordinary matter ) .
the latter is motivated by the assumption of gauge confinement and dynamical supersymmetry breaking in the hidden sector , and as discussed above , could ease some of the problems .
having discussed the possible frameworks in which the symmetry can be realized , we now turn to a discussion of some examples .
nro s can be induced in our case by scalar vacuum expectation values ( vev s ) , non - vanishing @xmath77-terms , and fermion condensates @xcite , leading to many possible scenarios which would relate differently to supersymmetry breaking .
since we assume that only sm hidden singlets participate , then the operators ( which are nonrenormalizable in that case ) are suppressed by powers of the planck mass . for simplicity
, we will assume that @xmath100 gev ) is the only large scale suppressing the operators .
note also that while the form of the operators ( i.e. , the selection rules ) is dictated by the symmetries [ and in particular , @xmath54 ] , their non - vanishing values ( i.e. , the vev s ) may be a result of @xmath54 breaking or of the breaking of a different symmetry at a lower scale . in the latter case @xmath88 could be lower then the @xmath54 scale .
[ this observation is most relevant if the gauged @xmath54 option is realized . ]
the most simple example is the case in which @xmath102 is given by the vacuum expectation value of the scalar component @xmath103 of a hidden sector chiral superfield @xmath104 with @xmath1-charge @xmath105 .
the @xmath2-parameters , @xmath106 for @xmath107 , respectively , are given in that case by ( omitting hereafter dimensionless couplings and coefficients in the nonrenormalizable operators ) @xmath108}}. \label{e1}\ ] ] we first discuss the most simple case in which each of the @xmath109 s depends only on one field , i.e. , @xmath110 for @xmath111 ( @xmath112 ) , respectively .
@xmath113 is also suppressed by negative powers of @xmath114 because @xmath115 is a hidden sector field ( otherwise , all symmetries allow a renormalizable term @xmath116 in the supepotential ) .
however , it need not be suppressed by more than one power of @xmath114 , i.e. , @xmath117 .
it could be a weak - scale or a planck - scale parameter , depending if @xmath118 gev or @xmath119 . if there are no @xmath120 fields then @xmath121 . @xmath92 and @xmath91 are weak - scale parameters , leading to the constraint @xmath122/[\ln(\langle z_{1_{1},\,1_{2}}\rangle / m)]$ ] , i.e. , @xmath123 for @xmath124 gev and @xmath125 for @xmath126 gev . if @xmath92 and @xmath91 are both given by the same field @xmath127 , then @xmath128 must be even and @xmath129 . in that case the electroweak symmetry breaking is induced by the scalar neutrinos ( sneutrinos ) rather than @xmath16 which is now decoupled from @xmath17 .
( we discuss electroweak symmetry breaking in section iv . ) more generally , the operators could involve several fields such that @xmath130 , for @xmath107 , respectively .
let us first assume , however , that all ( non - vanishing ) singlets have positive @xmath1 charges . then ( because of the holomorphicity of the superpotential ) the situation is similar to the one field case , i.e. , @xmath131 ( and @xmath121 unless there is a non - vanishing @xmath120 field ) .
more importantly , @xmath132 in this case .
lastly , there could be fields with negative @xmath1-charges [ in particular , if the @xmath54 is gauged and anomaly free ] .
this most general case could allow for , e.g. , @xmath133^{l } < \mu_{h}$ ] and @xmath134 [ and also @xmath135 with @xmath136 .
however , one could also have now @xmath137 ( @xmath138 is a baryon number violating coupling ) .
both , @xmath91 and @xmath139 ( as well as @xmath138 ) vanish in the mssm limit @xmath140 . otherwise , one has to impose an additional symmetry on @xmath23 .
for example , @xmath92 , @xmath91 and @xmath141 $ ] have ( standard ) peccei - quinn charges for @xmath16 and @xmath17 and @xmath142 for all other ordinary matter .
the anomalous @xmath0 would be given in that case by some linear combination of the ( anomalous ) @xmath143 and @xmath54 .
] of @xmath144 and @xmath145 , respectively .
thus , a peccei - quinn symmetry would forbid in this case the dangerous yukawa couplings .
the chiral superfields discussed in the previous example could also have non - vanishing @xmath77-terms which are of the order of magnitude of the supersymmetry breaking scale , @xmath147 ( i.e. , they contribute to supersymmetry breaking ) . in that case
, @xmath115 could also provide a source for the soft supersymmetry breaking @xmath34-terms , @xmath148 , and @xmath149 ( for example , see @xcite ) .
an interesting scenario arises in the case that @xmath150 ( we discuss below cases in which @xmath21 is a mixture of non - vanishing @xmath151 and @xmath152 components ) with @xmath153 , where @xmath154 , @xmath155 and @xmath156 or @xmath157 .
the parameter @xmath158 is dominated by the @xmath159 contribution and could be of a different order of magnitude than @xmath21 .
an interesting example is the case of @xmath160 gev ) and @xmath161 , i.e. , @xmath162 .
also , a small @xmath91 does not necessarily imply in this case a small lepton number violation in the scalar potential .
[ we comment on models with @xmath163 gev ) ( e.g. , @xcite ) below . ]
note that @xmath158 are the only lepton - number violating soft supersymmetry breaking parameters in our models .
all three classes of @xmath2 parameters could be obtained from a single @xmath165 chiral superfield , @xmath115 , with non - vanishing scalar and @xmath77 ( auxiliary ) components , and whose fermion component , @xmath166 , condenses .
it may be difficult to incorporate such a scenario in a realistic model .
nevertheless , it is worth noting .
the different @xmath2 parameters are now given by [ e3 ] @xmath167 @xmath168 @xmath169 and from ( [ e3a ] ) one has @xmath170 gev )
. for @xmath171 gev)@xmath172 one has @xmath173 , and there is no clear hierarchy between the two parameters .
lastly , if the condensate @xmath174 , where @xmath175 gev is the supersymmetry breaking scale , then contrary to the generic situation ( see example a ) , @xmath176 .
as we discuss in section iv , the couplings of the right - handed neutrino need to be suppressed in this case .
the relation ( [ e3c ] ) would also allow for weak - scale vector - like exotics .
it is interesting to note that , regardless of its origin , @xmath177 gev in all operators in this case .
regarding the yukawa coupling @xmath139 , the holomorphicity of the superpotential forbids @xmath178 .
thus , one has @xmath179 .
if @xmath115 was not an hidden - sector field , then the renormalizable superpotential term @xmath180 would have been allowed ( leading to @xmath181 gev ) . in such a case
one needs to impose an additional symmetry , e.g. , the usual @xmath31 @xmath1-parity with @xmath182 , which allows for ( [ e3a ] ) ( [ e3c ] ) but forbids the @xmath183 renormalizable term . in the case of a @xmath186 superfield , @xmath187 implies that @xmath188 gev ) .
thus , @xmath189 and like in example a , the sneutrinos could contribute significantly to electroweak symmetry breaking .
having an additional field @xmath31 with @xmath190 , and if the @xmath191 gev)@xmath172 is non - vanishing , one can obtain instead [ e4 ] @xmath192 @xmath193 @xmath194 @xmath195 are @xmath90 parameters in that case if @xmath196 ( and there are two scale parameters @xmath197 gev and @xmath198 gev , as is often the case in supergravity models ) , which implies that @xmath199 [ and again , there is no clear hierarchy between ( [ e4a ] ) and ( [ e4b ] ) ] . however , one has to forbid in that case the @xmath200 combinations @xmath201 , etc .
, which would lead to unacceptably large @xmath202 and @xmath203 . for example , a continuous hidden sector @xmath0 symmetry with identical charge assignments to both fields would allow ( [ e4a ] ) and ( [ e4b ] ) but forbid the dangerous couplings ( as well as @xmath113 ) .
our main lesson is that the typical situation @xmath204 and @xmath205 is easily found in simple examples of @xmath54 selection rules ( but exceptions exist ) . if @xmath206 then typically @xmath207 .
however , if there are also non - singlet @xmath77-terms which do not vanish , then no clear hierarchy exists . in some cases additional symmetries
may be required in order to suppress @xmath139 , but typically @xmath54 is sufficient for that purpose . ( the additional symmetries ,
if required , could correspond to typical symmetries that are often found in models . )
we presented a few scenarios in which some , or all , of the @xmath112 are generated .
many more scenarios exist .
the relevant scenario would be determined by the realization of the @xmath54 symmetry , the hidden singlet @xmath1-charges , and by the relations between the fields that spontaneously break @xmath54 and supersymmetry breaking .
if the sm fields do not have the standard @xmath1-charge assignments @xcite that we assume , then some of the examples given above may need to be revised , depending on the charge assignments chosen .
if it is a family - dependent assignment then one could distinguish ( unlike in our case ) between @xmath208 , @xmath209 , and @xmath210 .
( we discuss family dependences in section vii . )
lastly , if there is a hidden @xmath120 chiral superfield @xmath115 with @xmath211 gev ) then @xmath92 ( but not is generated in this manner then it is proportional to @xmath30 and has @xmath1-charge @xmath212 . ]
@xmath91 ) could be partially generated by a kahler potential source @xcite , smearing any correlation between @xmath92 and @xmath91 .
the ratio @xmath213 , which is a ( high - energy ) order parameter of the models , need not be suppressed in order to suppress the tree - level neutrino mass .
this is a trivial result of the observation that there are no tree - level masses for the neutrinos if @xmath21 and @xmath214 are aligned in field space @xcite .
( this observation is also the basis for the recent works of refs .
we elaborate on that observation in this section , and we show in the next section that nearly parallel @xmath21 and @xmath78 four - vectors arise dynamically in simple models for the soft supersymmetry breaking parameters ( dynamical alignment ) .
discussion of laboratory and cosmological implications is postponed to sections v and vi , respectively . an important distinction ( which allows for the neutrino mass generation ) between the models that are discussed here and
the mssm is that all seven neutral fermions mix in our case .
that is , the supersymmetric partners of the @xmath34 and @xmath215 gauge bosons and neutral higgs
the bino , wino ( the gauginos ) and the two higgsinos , respectively ( the neutralinos ) , mix with each other and with the three neutrinos once elelectroweak symmetry is broken .
hence , one has a seven - dimensional majorana mass matrix for the neutral fermions , which involves ( at tree - level ) * the bino and wino soft supersymmetry breaking masses @xmath216 and @xmath217 , respectively , * the higgsino and neutrino masses @xmath21 , * and the gauge - matter mixing parameters @xmath218 and @xmath219 .
( we define @xmath220 , @xmath221 , @xmath222 , @xmath223 , and @xmath224 . )
the determinant of the neutral fermion mass matrix vanishes : the @xmath43 symmetry is broken down to @xmath225 by two fundamental vectors , @xmath21 and @xmath226 , each of which can render only one physical state massive , and there are two massless states .
one could define , for example , the @xmath227 and @xmath2 neutrinos as those states , @xmath228 .
let us then eliminate the ( tree - level ) massless states and discuss the residual five - dimensional non - degenerate mass matrix .
it is convenient to define for that purpose , following banks et al .
@xcite , @xmath229 .
the determinant ( of the residual mass matrix ) is proportional to @xmath230 , which is the second ( or the low - energy ) order parameter of the model .
that is , if @xmath21 and @xmath226 are aligned [ and @xmath43 is broken down to only @xmath231 then only one state is rendered massive and there is an additional massless state @xmath232 .
otherwise , the lightest neutral fermion , which we define to be the @xmath233 neutrino @xmath234 , has a mass @xmath235 ( assuming , for simplicity , @xmath236 ) .
it is sufficiently suppressed and phenomenologically acceptable if @xmath237 ( where @xmath238 mev [ @xmath239 ev ] is the laboratory @xcite [ cosmological ev bound applies to stable neutrinos and assumes the critical energy density @xmath240 . if the neutrinos constitute only a part of that energy density , then the upper bound scales down accordingly , i.e. , @xmath241 ev . ] ( energy density ) @xcite ] upper bound on the neutrino mass ) .
the alignment condition and all other sufficient conditions for the suppression of the neutrino mass are more easily seen in the rotated basis in which @xmath242 ( assuming one lepton generation ) .
the tree - level neutrino mass is given by the ratio @xmath243 . ] ( for the residual mass matrix ) in that basis @xmath244 where @xmath245 are the hypercharge and @xmath225 couplings , respectively , and the rotated @xmath246 \approx \mu_{l_{\tau}}[(\nu_{l_{\tau}}/\mu_{l_{\tau } } ) - ( \nu_{h_{1}}/\mu_{h_{1}})]$ ] ( assuming @xmath247 ) .
thus , there are light states if @xcite * @xmath248 @xmath249 , i.e. , @xmath21 and @xmath250 are parallel ( and @xmath251 ) or @xmath252 ( and @xmath253 ) , * @xmath254 @xmath255 , * and @xmath256 the gauginos are heavy and their mixing with the matter fermions is negligible , i.e. , @xmath257 , in which case @xmath258 ( see above footnote ) .
it is , however , difficult to realize the latter case because @xmath2 and the gaugino masses both contribute to the quadratic terms in the scalar potential , and thus , are related by the minimization conditions described below and are typically of the same order of magnitude .
the case @xmath254 is quite interesting and was discussed in refs . @xcite .
( see also ref .
@xcite for a recent discussion and references of the equivalent situation in the mssm .
however , the constraints on the parameter space in our case could be different than in the mssm . )
at least two of the neutral fermions , a higgsino and a neutrino , are now light .
the light higgsino contribution to the @xmath115 width constrains the model to the region of @xmath259 . from rsb one
has that the scalar mixing parameter @xmath158 ( see section iv ) is not proportional to @xmath21 and can not vanish in that case .
thus , the higgs boson mass would partially come from mixing ( @xmath260 ) with the scalar neutrinos . to consistently realize such a scenario would require a departure from the simple models for the soft terms that we consider below .
we will comment on that scenario again when discussing the scalar potential , but we do not consider it in detail .
perturbativity of yukawa couplings and dynamical alignment ( which are considered in section iv ) constrain @xmath261 from above ( see below ) .
thus , the remaining case @xmath262 is typically realized for nearly aligned vectors , i.e. , @xmath249 .
( this is crucial for the discussion of the phenomenological implications of the model in sections v and vi . ) nevertheless , there are additional ( secondary ) suppressions of @xmath263 if @xmath264 or if the gaugino mass parameters are large .
for example , in fig . 1 we show @xmath265 as a function of @xmath266 for different values of the gaugino masses and of @xmath261 ( but no rsb is yet required ) . note , in particular , the high precision required in the alignment for moderate values of @xmath261 .
the alignment condition , however , can be slightly relaxed for large values of @xmath261 .
[ fig : fig1 ] we consider in this work only the tree - level mass matrix .
the one - loop matrix has new non - diagonal entries [ that explicitly break the residual @xmath225 symmetry ] and all seven states are massive with generically one very light neutrino and no tree - level or @xmath14-vertex one - loop masses as in the general case . ] and one neutrino with a mass @xmath267 ( i.e. , it could be of the order of the tree - level mass ) @xcite .
thus , even though two neutrinos are massless here , all neutrinos are massive in a more complete calculation ( which is not necessary for our purposes ) .
we would also like to point out that in four - family models , which are constrained by the @xmath115-width to have @xmath268 gev , @xmath266 is not constrained by the tree - level neutrino mass ( which could partially account for @xmath269 ) but by the size of the corresponding loop masses for the ordinary neutrinos ( as well as by other phenomenological considerations ) .
in a general lepton number violating extension of the mssm the scalar neutrinos ( sneutrinos ) will typically acquire vev s so that lepton number is broken both explicitly and spontaneously .
containing all lepton number violation in @xmath8 ( as in our case ) is , however , sufficient to spontaneously break lepton number .
( in this section lepton number refers to the high - energy lepton number . ) below , we will organize the higgs - lepton scalar potential and its minimization equations in a convenient way , and show that indeed the sneutrinos aqcuire vev s .
we would then list the conditions for achieving the required dynamical alignment between @xmath21 and @xmath226 , and show that they are satisfied in a large class of models , but not in generic grand - unified theories and in models with an intermediate - scale right - handed neutrino .
we will also comment on perturbativity constraints on the yukawa couplings ( which are different than in the mssm ) . for simplicity
, we will discuss models with only the third generation of quarks and leptons .
the generalization , however , is straightforward .
( renormalization - group scaling of general broken @xmath73 extensions of the mssm were discussed recently in ref .
@xcite . )
it is convenient to define [ defs ] @xmath270 @xmath271 @xmath272 and @xmath273 as before .
( note that @xmath158 has a squared mass dimension . )
the selection rules of section ii typically imply lepton number conservation by dimensionless couplings of the low - energy kahler potential .
thus , the scalar masses @xmath274 and @xmath275 are the only source of explicit lepton number violation in the scalar potential .
( the yukawa @xmath77 terms must involve , in our case , charged degrees of freedom , and thus , vanish at the minimum . ) hence , the higgs - lepton scalar potential can be written as a straightforward generalization of the mssm higgs scalar potential , i.e. , @xmath276 where @xmath277 is the one - loop correction ( that is included in our numerical procedures and @xmath278 yukawa couplings , which in our case , are given by the corresponding corrections in the mssm . ] ) and [ defs2 ] @xmath279
@xmath280 @xmath281 where on the right - hand side @xmath282 is the soft supersymmetry breaking squared mass of scalar @xmath283 .
note that the ( tree - level ) scalar potential ( [ potential ] ) has , in the absence of explicit lepton number breaking ( @xmath284 ) , the unbounded direction @xmath285 @xcite , as well as a flat direction @xmath286 @xcite . the mssm @xmath14-flat direction @xmath287 ( i.e. , @xmath288 ) is also relevant here and could also lead to a flat direction ( one could eliminate @xmath289 in that case by redefinitions of @xmath290 and @xmath291 ) . lepton number could be spontaneously broken in the case of a flat direction , even if it is not broken explicitly .
this possibility was discussed in the context of supersymmetric majoron models @xcite .
we will exclude these directions , which are difficult to realize once rsb is included , from our analysis .
it is convenient to minimize @xmath292 with respect to @xmath293 , @xmath294 , and the angle @xmath295 .
two of the minimization equations reduce to those of the mssm , i.e. , @xmath296 @xmath297 .
\label{min2}\ ] ] in particular , at the minimum @xmath298 , independent of @xmath299 .
the third equation reads @xmath300 \cos\beta + 2b\sin ( \alpha - \delta)\sin \beta = 0 . \label{min3}\ ] ] for the boundary conditions [ bc ] @xmath301 @xmath302 there is only one @xmath43 direction in field space which is determined by the angle @xmath303 .
hence , @xmath226 must align along that direction and the solution to ( [ min3 ] ) ( for @xmath304 , @xmath305 , and away from the flat direction ) is given by @xmath306 .
the alignment is achieved dynamically in that case .
note that the @xmath21 is a parameter vector in the low - energy theory and the @xmath43 symmetry is broken explicitly and spontaneously down to ( the low - energy lepton symmetry ) @xmath307 , and ( the rotated low - energy ) @xmath308 contains the seven pseudo goldstone bosons with mass @xmath309 . unlike models with no explicit breaking ( but only spontaneous breaking along the flat direction )
there are no light majorons .
the boundary conditions ( [ bc ] ) are often found in models for the soft supersymmetry breaking terms .
the second condition is trivially realized if @xmath310 ( i.e. , @xmath34-proportionality ) , as is also often the case .
the crucial point ( which was also noted independently by hempfling @xcite ) is that if @xmath311 and @xmath310 at some planckian scale , then the deviations from these relations at the weak scale due to renormalization group scaling are proportional to the square of the @xmath278 yukawa coupling @xmath312 , i.e. , @xmath313 , \label{rge}\ ] ] where @xmath314 is the trilinear soft mass parameter @xmath315 .
thus , deviations from universality are strongly suppressed ( unless @xmath252 , i.e. , @xmath316 , or if the soft parameters @xmath317 ) , but are still sufficient to generate the small neutrino mass .
this is the case in the ( extended ) mssm with the so - called universal boundary conditions at the grand scale ( i.e. , @xmath318 , @xmath319 , and @xmath320 ) .
we solved the system ( [ min1 ] ) ( [ min3 ] ) iteratively and @xmath303 , we use eq .
( [ min1 ] ) to solve for @xmath2 , eq .
( [ min2 ] ) to solve for the proportionality parameter @xmath278 , and eq .
( [ min3 ] ) to solve ( numerically ) for @xmath295 . ] in our model for various universal boundary conditions and confirmed the above assertions and claims .
we also included in the numerical procedures the one - loop radiative corrections ( see above ) and the corrected @xmath278 and @xmath233 yukawa couplings ( see below ) .
unlike in the mssm , @xmath321 , but rather @xmath322 , leading to the weak - scale perturbativity constraint ( requiring @xmath323 , i.e. , that it is below its quasi fixed - point ) @xmath324 it can also be written as @xmath325 .
( [ rge ] ) and ( [ hb ] ) imply that the models are realized more naturally for small and moderate values of @xmath261 ( while typical grand - unified models with right - handed neutrinos are realized for large @xmath261 @xcite ) .
phenomenological implications , which are discussed in the following sections , constrain @xmath326 from above , i.e. , @xmath327 , which is stronger than any constraint that one could derive from ( [ hb ] ) . also ,
unlike in the mssm , the leptonic yukawa couplings @xmath328 but are found by requiring the three light eigenstates of the charged ( color singlet ) fermion mass matrix to have the correct masses , i.e. , mixing effects can slightly modify their values .
[ since the leptonic yukawa couplings have nearly flat renormalization curves ( for not too large @xmath261 ) , this effect could be important when considering relations between yukawa couplings at planckian scales . ]
we account for these effects in our numerical procedures .
[ fig : fig2 ] in fig .
2 we show the neutrino mass as a function of @xmath261 for @xmath329 . the corresponding alignment , i.e. , comparison of ( the output ) @xmath330 and ( the input ) @xmath331 is shown in fig . 3 for @xmath332 ( bullets ) and 0.9 ( squares ) .
4 we further explore the parameter space for @xmath332 . by observation
, we can draw the following conclusions , some of which are somewhat surprising : * the typical suppression factor is of @xmath333 , i.e. , @xmath334 mev )
. however , for @xmath335 one finds suppression as strong as @xmath336 [ or even @xmath337 , i.e. , @xmath338 ev ) .
the functional dependence on the soft parameters is complicated . in general
, we find that @xmath339 and @xmath340 are slightly prefered .
the former leads to heavy scalars which are less sensitive to the renormalization - group corrections , while the latter suppresses the @xmath341 contributions in ( [ rge ] ) [ which is a more important consideration than the large @xmath342 effect in ( [ det ] ) ] . * the smaller values of @xmath331 do not imply necessarily smaller neutrino masses and vice versa .
* the alignment condition is indeed satisfied poorly for @xmath252 , in contradiction to its excellent satisfaction for small and intermediate values of @xmath261 .
this leads to a relatively wide ( narrow ) range for @xmath263 in the former ( latter ) case . *
the poor alignment for @xmath252 and the suppression of @xmath343 are competing effects which allow to realize small neutrino masses for large @xmath261 in some cases .
[ fig : fig3 ] [ fig : fig4 ]
let us stress that our results hold in a larger class of models , i.e. , in models with universality of the @xmath44 soft masses and the proportionality @xmath344 .
the lepton - higgs universality is guaranteed if the kahler potential is ( approximately ) invariant under the @xmath43 , e.g. , in string models it would require that all @xmath44 components have a unique set of modular weights . no universal boundary conditions for any other fields are required . in grand - unified models , however ,
the soft parameters are scaled at planckian scales according to renormalization group equations which are invariant under the grand - unified symmetry and which are determined by its superpotential couplings . in generic models
the leptonic and higgs fields are embedded in different representations ( often with different dimensions ) and have different yukawa interactions with sometimes large yukawa couplings .
for example , @xmath16 would typically couple to the large higgs representation that break the unified symmetry and/or render its su(5 ) partner , the color triplet , heavy .
thus , the higgs and lepton soft masses are subject to different scaling laws at large scales .
they could also receive different contributions from @xmath0 @xmath14-terms is gauged the respective ( residual ) @xmath14-terms could lead to higgs - lepton non universality .
similar problems would also arise in the case of low - energy stringy @xmath0 s , unless the @xmath14-terms vanish or are negligible .
thus , global symmetries are better suited for our purposes .
] if the grand unified symmetry is broken to some group @xmath345 [ which is not the hypercharge @xmath0 ] between the planck and unification scales .
the @xmath43 symmetry would be broken in @xmath292 in more that one direction and @xmath226 and @xmath21 would not be aligned .
for example , we find that in the @xmath347 models of ref .
@xcite the scaling violation of the @xmath43 due to the @xmath348 term destroys , in most cases , the dynamical alignment .
similarly , a right - handed neutrino , @xmath74 , will introduce the yukawa terms @xmath349 , and unless @xmath350 for all three generations ( but in unified models typically @xmath351 ) or @xmath352 , it would again introduce new @xmath43 breaking .
thus , it is difficult to reconcile supersymmetric neutrino masses ( with dynamical alignment ) and right - handed neutrino see - saw models .
( similar arguments hold for an arbitrary high - energy @xmath40 . ) in general models for the soft terms one could always tune the parameters to satisfy ( [ min3 ] ) .
for example , if @xmath160 gev ) then @xmath353 may not be proportional to @xmath2 ( see example b of section ii ) .
it could lead to large mixing in the scalar sector , but would require some tuning of parameters ( or to invoke extended symmetries ) in order to still guarantee @xmath354 .
we do not find this possibility attractive and do not pursue it here .
having established the smallness of the neutrino mass and its consistency with the laboratory bound @xmath355 mev @xcite ( and in many cases with the cosmological bound @xmath356 ev @xcite ) in the previous sections , we will not discuss other phenomenological aspects of the models in detail but only survey possible constraints .
we will show that in our framework the constrained couplings and amplitudes are proportional to some power of @xmath357 or to small ( mssm ) yukawa couplings , and thus , are evaded in large sections of the parameter space .
in particular , due to the large number of parameters , it is difficult to efficiently constrain the models ( until supersymmetric particles and their decays are observed and characterized ) : see , for example , ref .
@xcite for detailed discussions of constraints etc . ,
apply , in our case , only to the rotated vev s after the sm lepton number is defined ( see below ) . ] from rare decays , weak current universality , etc .
, on these and on more general models of lepton number violation .
the discussion is simplified once again if one performs a low energy @xmath43 rotation . for our purposes
here it is convenient to define the low - energy higgs field @xmath308 such that @xmath358 , and the three leptonic fields @xmath359 correspond to the three perpendicular directions in field space . in the degenerate limit @xmath251 one then has @xmath360 .
all fermion and scalar mass matrices are block diagonal in the gauge-(low - energy ) higgs fields and in the ( low - energy ) leptonic fields .
the rotated neutral and charged states constitute the `` physical '' lepton and higgs doublets consistently , and one can define the ( low - energy ) sm lepton number in this basis .
( if @xmath361 then the discrepancy between low and high energy definition of leptons is maximized , leading to possible constraints on their ratio . ) the rotation generates @xmath40 [ see ( [ wlnv ] ) ] and thus , @xmath362 and @xmath30 has an accidental @xmath35 ( baryon parity @xcite ) symmetry ( but only at the renormalizable level ) .
since @xmath266 is small but not exactly zero ( the alignment , and in particular , if achieved dynamically , is not scale invariant and is not expected to be exact ) there are perturbations to the naive limit , and some mixing between the ( low - energy ) leptonic and gauge - higgs sectors arises , e.g. , the @xmath233 could contain a small admixture of the wino and the z could have off - diagonal couplings to a neutrino and a gaugino .
however , the effects are of the order of magnitude of the neutrino mass suppression factor , however , @xmath253 allows for larger values of @xmath266 . in that case , experimental constraints could be effective in constraining the models . ]
@xmath363 to some power .
note that since one of the loop - induced neutrino masses could be of the order of the tree - level mass , experimental constraints should be applied to the mixing parameters extracted from the one - loop mass matrices .
other constraints result from the new yukawa interaction contributions to flavor changing neutral ( fcnc ) and charged currents .
the superpotentials @xmath40 and @xmath41 are related by a rotation and @xmath364 . as a result , @xmath365 and @xmath366
are diagonalized simultaneously ( and @xmath367 has two generation indices ) . in particular ,
the smallness of relevant entries in @xmath368 is directly related to their smallness in the @xmath369 matrices .
the only new contributions to hadronic fcnc arise from loops , and in many cases ( e.g. , @xmath34 meson mixing ) are suppressed as in the mssm @xcite .
( there are some differences due to the different masses of the higgs and the leptonic fields that propagate in the loops . ) in addition , the amplitudes for leptonic fcnc processes , e.g. , @xmath370 , are typically proportional in our case to the mixing angle @xmath266 and/or to the electron mass @xcite ( and , in general , depend sensitively on @xmath22 ) .
thus , they also lead to only mild constraints .
charged current processes and constraints , e.g. , weak current universality , neutrinoless double-@xmath371 decay , etc . , are induced at tree and loop levels .
they typically constrain @xmath372 , where @xmath373 is the relevant sm fermion mass ( or lead to even weaker constraints ) .
given the origin of @xmath374 in our case , these constrains are typically satisfied and could lead at most to an upper bound on the ratio @xmath375 .
( note that not all @xmath376 combinations are present in the basis that @xmath365 is diagonal and that we do not specify any textures for the mssm yukawa matrices . ) also , all constraints scale with the inverse mass of the relevant ( virtual ) superpartner that mediates the process , and are further weakened as the mass parameters reach the few hundred gev mark .
it is reasonable , however , to require that the ratio @xmath377 so that the couplings in @xmath40 are of the same order of magnitude ( or smaller ) as the couplings in @xmath41 .
we conclude that a dedicated analysis and searches could constrain the order parameters of the model : @xmath266 ( i.e. , deviations from the alignment ) and the ratio @xmath378 , which determine the gauge - lepton mixing and the size of the couplings in @xmath40 , respectively .
however , it will lead to only mild restrictions on the model parameter space , and it is not called for at present ( given our poor knowledge of the supersymmetric spectrum parameters ) .
the former would then constrain the size of neutrino masses and thus , serve in the future as a consistency test of the model relating the neutrino and supersymmetric spectra .
it could also constrain @xmath261 ( see above footnote ) . the constraints on the ratio @xmath213
are , on the other hand , a function of @xmath261 ( recall that @xmath379 , where @xmath380 is the relevant fermion mass ) .
the ratio @xmath213 and the size of @xmath381 have also important implications for collider signatures from exotic superpartner decays @xcite , in particular , from the decays of lightest superpartner ( the lsp ) .
for example , there could be a single production of a scalar lepton in hadron colliders which would then decay to two jets or two jets and either like - sign dileptons or ( degraded ) missing energy ( for example , see ref .
such decays serve , in general , as a test of all supersymmetric lepton number violating models ( assuming that supersymmetry is established and characterized by experiment ) in both hadron and electron colliders . of particular interest
is the case in which the scalar neutrino @xmath382 ( that can be singly produced ) is the lsp .
the gauginos and higgsinos are then heavy and the sneutrino would decay exclusively to two jets ( assuming @xmath383 ) .
its charged @xmath225 partner , which has a similar mass , will also decay preferably to jets .
superpartner decays are characterized in that case by a @xmath384 signal . in the case of the tevatron , for example
, the lsp dijet signal[multiblock footnote omitted ] could be only one order of magnitude below the corresponding qcd signal @xcite ( but would be further suppressed if the couplings are small ) .
however , the enhancement of the total dijet cross - section is more significant ( and is smeared over a larger energy range ) if there are more than one relevant threshold [ e.g. , its charged @xmath225 partner and possibly scalars of all three lepton doublets ] .
these observations may require further attention which is only weakly constrained ( in the absence of lepton number violation ) , but is forbidden here .
] in view of recent indications of enhancements in the dijet inclusive cross section @xcite and mass spectrum @xcite measurements .
in particular , the situation here is quite different than in @xmath1-parity conserving supersymmetric theories , where any new contributions to the inclusive dijet cross section are only at the loop level and were shown recently to be small or negative@xcite .
we concluded above that no significant constraints arise on the model from laboratory and collider data , but stressed the unconventional decays and signatures of the supersymmetric particles .
we also pointed out that the lsp is not expected to be stable . here , we will survey some astrophysical and cosmological aspects of the model .
we also elaborate on the lsp and neutrino decays .
in particular , we show that the lsp ( and thus , all other mssm superpartners ) is cosmologically irrelevant in any supersymmetric model with explicit ( weak - scale ) lepton number violation which is the source of the neutrino mass .
radiative neutrino decays are shown to lead to only moderate constraints on the order parameter @xmath213 ( unless @xmath261 is large ) .
we do not consider here some issues which are more model dependent , e.g. , that of the primordial baryon asymmetry in the universe ( but see , for example , @xcite ) .
_ neutrino oscillations : _ current observations imply not only massive neutrinos but also neutrino mixing , so that the neutrinos could oscillate @xcite .
indeed , we find model dependent mixing between the three neutrinos .
it was recently pointed out @xcite that the mass and mixing patterns in the models with supersymmetric neutrino masses are relevant for the solution of the solar and atmospheric neutrino problems @xcite .
most importantly , we would like to stress that the patterns that arise could be quite different from those in typical see - saw and similar models .
they are not proportional to any yukawa couplings and the hierarchy among the neutrino masses is not related to either the hierarchy among the quarks or the charged leptons .
_ lsp decays : _ since @xmath73 is effectively broken to the baryon parity in our models , the lsp , i.e. , the lightest neutralino or scalar , is not stable but decays @xcite through small mixing ( if it is a neutralino ) , and more probably , via @xmath366 and @xmath367 interactions . in order to render the lsp stable on a cosmological time scale ( @xmath386 sec )
one would have to fine tune @xmath387 for a neutralino lsp ( and even more so for a scalar lsp ) , and we neglected the decay via mixing which is further suppressed .
( see ref .
@xcite for life time formulae ) .
it would imply , in our case , an unacceptable fine tuning of @xmath213 .
but moreover , even if the neutrino masses are generated by only yukawa loops @xcite , they would be essentially massless given the constrain ( [ lsp ] ) , which is thus unacceptable .
we therefore conclude that the lsp decays on short time scales and is cosmologically irrelevant .
( we find that a neutralino lsp could , however , still be stable in our model on collider scales . )
note that this argument can not be cured by any symmetry that allows for @xmath388 since it is sufficient to have a single lepton number violating yukawa coupling that does not vanish .
_ neutrino mass and decays : _ the neutrino energy density and over - closure of the universe considerations constrain @xmath356 ev @xcite ( unless it decays on cosmological time scales , i.e. , @xmath390 sec , in which case one could also have @xmath391 mev ) @xcite ) .
the neutrino decays , similarly to the lsp , via mixing to neutrinos ( i.e. , @xmath392 ) and via its small one - loop magnetic moment operator @xmath393 ( i.e. , @xmath394 ) .
the former may be beneficial in the case of a @xmath28mev ) neutrino but is strongly suppressed .
the latter is the dominant mode with ( adopting the calculation of ref .
@xcite to our case ) @xmath395 where we assumed @xmath396 .
[ more generally , @xmath397 is replaced by @xmath398 . ] a comprehensive analysis of such decays was presented in ref .
@xcite , leading to the constraint @xmath399(sec ) @xmath400(mev ) ( for @xmath401 ev ) . from ( [ neutrino ] )
one has the corresponding constraint ( for @xmath401 ev ) , @xmath402 again , no significant constraints arise on the order parameter @xmath213 .
( note that for a given neutrino mass only the combination @xmath403 is constrained . for large @xmath261 , i.e. , @xmath404
, the constraints could be significant . )
_ dark matter sources : _ one might have hoped that the model would provide a closed framework in which the primary candidates for cold and hot dark matter , the lsp and the neutrino , respectively , are related by mixing and/or decay chains and thus , have correlated abundances .
such a scenario could make , for example , critical density models for the universe more plausible .
however , we have concluded above that the lsp is cosmologically irrelevant and the only ( hot ) dark matter candidates in the model are the two heavier neutrinos ( which in some sense are the actual lsp ) .
this is somewhat a disappointing aspect of the model .
nevertheless , there could still be other sources for cold dark matter , e.g. , light components of the axion superfield ( as was proposed in the case of a peccei - quinn axion @xcite ) .
lastly , it is interesting to note that our models could reverse the generic situation in the @xmath160 gev ) lepton number conserving models which typically suffer from a slow lsp annihilation rate @xcite .
throughout this work we have assumed universal @xmath1-charge assignment to all matter fields . as a result ,
the vector @xmath405 was @xmath307 symmetric .
there are many ways in which the @xmath307 could be broken , most obviously so , if the @xmath1-charge assignment is family dependent .
it was suggested that a family dependent assignment may be needed in order to cancel the anomalies of a gauged @xmath54 theory and that it is related to the fermion mass problem ( but , as we noted above , it could lead to dangerous @xmath14-terms ) . in ref .
@xcite yukawa operators involving the first two families were forbidden by the assignment @xmath406 but the possibility of dynamical couplings was not discussed .
a more ambitious program was pursued in ref .
@xcite where it was suggested that all yukawa couplings ( but not @xmath21 ) are dynamical variables that depend on the @xmath54 breaking parameter @xmath88 ( in a similar manner to the horizontal symmetry approach @xcite ) .
however , being an @xmath1-symmetry complicates the anomaly cancelation equations and ref.@xcite chose not to consider ( and satisfy ) the complete set of equations , in particular , those involving hidden fields .
given the ambiguous status of the gauged @xmath54 , we chose the simplest possible @xmath54 charge assignments and did not consider the mssm yukawa couplings as dynamical variables .
nevertheless , it is very likely that the solution of the fermion mass problem , and in particular , if it relies on some symmetry principles and/or if the yukawa couplings are dynamical variables , would break explicitly the @xmath307 symmetry in @xmath8 as well , e.g. , by forbidding certain field combinations .
it could also break the higgs - lepton soft mass universality from a non - minimal kahler functions or @xmath14-terms , in which case the dynamical alignment would fail .
however , the latter breaking is constrained ( from fcnc ) to @xmath407 type breaking of the @xmath307 , and one can impose the additional condition that breaking the above @xmath307 by the leptonic soft - terms is negligible . for example , this would be the case if the @xmath307 is broken mainly in the right - handed lepton singlet sector ( but then @xmath8 is still @xmath307 symmetric ) , or if it is a global or discrete symmetry ( and the kahler function is , e.g. , minimal ) .
therefore , we do not consider the @xmath307 symmetry as a necessary result of our model . the above discussion also affects the phenomenology of the model .
in particular , if @xmath408 and only @xmath233 number is ( significantly ) broken , then the constraints on @xmath213 are quite weak .
( models in that spirit were discussed , e.g. , in ref .
in conclusion , we have shown that the neutrino mass could arise from a generalized supersymmetric mass term @xmath24 in the weak - scale superpotential , on the condition that higgs - lepton universality in the scalar potential is broken only weakly .
this is indeed the situation , for example , in universal models for the soft supersymmetry breaking parameters .
the restricted form of @xmath1-parity breaking in the model was realized , as an example , in the framework of a spontaneously broken @xmath54 symmetry , which is often present in models of dynamical supersymmetry breaking .
the symmetry framework provides also a solution to the generalized @xmath2-problem while suppressing lepton number violation in the yukawa interactions and in the kahler potential . in hidden sector models
the @xmath54 scale is an intermediate scale .
the neutrino mass is related to the intermediate scale physics in that case in a very different fashion than in see - saw models . in particular
, it could be a result of supersymmetry breaking .
it offers a new mechanism for the neutrino mass generation in supergravity and superstring theories , which does not require to introduce any intermediate - scale see - saw structure .
the neutrino mass suppression is achieved dynamically ( and is sensitive to @xmath261 ) .
the resulting mass range for the neutrino agrees with not only the weaker collider limits but also with cosmological considerations .
the dynamical suppression is typically ( but not always ) destroyed in grand - unified models , models with an intermediate right - handed neutrino , and in models with arbitrary lepton - number violating yukawa couplings , once renormalization effects are taken into account .
we were able to define the sm lepton number at the weak - scale ( after all symmetries were broken ) .
the resulting yukawa superpotential has an accidental @xmath35 baryon - parity symmetry , but the yukawa couplings are not arbitrary and , in general , satisfy all experimental constraints .
one can still constrain the models two order parameters . however , given the above and the large number of parameters in supersymmetric models , the constraints are mild ( unless @xmath252 ) .
1 . the patterns of neutrino masses that arise could be quite different from those that arise in typical see - saw models .
the lsp is not stable on cosmological time - scales , an observation which holds more generally in supersymmetric models with weak - scale origin of the neutrino masses .
3 . if the lsp is a scalar neutrino it most probably decays in the detector , enhancing jet production and leading to a typical @xmath384 signal . it is pleasure to thank herbi dreiner , ralf hempfling , and yossi nir for discussions of their works , and to paul langacker and masahiro yamaguchi for their comments on the manuscript .
this work was partially supported by a fellowship ( n. p. ) and grant no .
sfb-375 - 95 ( research in astroparticle physics ) of the deutsche forschungsgemeinschaft , and by european union grant no .
sc1-ct92 - 0789 ( h .- p .
65 for reviews and references , see p. langacker , _ physics beyond the standard model iv _ , proceedings of the conference , lake tahoe , calif . , 1994 ,
f. gunion , t. han , and j. ohnemus ( world scientific , singapore , 1995 ) p. 487 , hep - ph/9503327 ; j. w. f. valle , invited talk given at the fourth international workshop on theoretical and phenomenological aspects of underground physics ( taup 95 ) , toledo , spain , 1995 , ftuv-96 - 13 , hep - ph/9602369 . m. gell - mann , p. ramond , and r. slansky , in _ supergravity _ , proceedings of the workshop , stony brook , new york , 1979 , eds .
p. van nieuwenhuizen and d. freedman ( north holland , amsterdam , 1979 ) p. 315 ; t. yanagida , in _ proceedings of the workshop on unified theories and baryon number in the universe _ , tsukuba , japan , 1979 , eds .
a. sawada nad a. sugamoto ( kek reprot no .
79 - 18 , tsukuba , 1979 ) . for review and references ,
nilles , phys . rep .
110 , 1 ( 1984 ) ; _ testing the standard model _ , eds .
m. cvetic and p. langacker ( world scientific , singapore , 1991 ) p. 633 ; _ the building blocks of creation _ , eds .
s. raby and t. walker ( world scientific , singapore , 1994 ) p. 291 .
h. e. haber and g. l. kane , phys .
117 ( 1985 ) 75 .
p. nath et al . , _ applied n = 1 supergravity _
( world scientific , singapore , 1984 ) ; g. g. ross , _ grand unified theories _
( benjamin , new york , 1984 ) ; r. n. mohapatra , _ unifcation and supersymmetry _
( springer , new york , 1986 , 1992 ) .
l. j. hall and m. suzuki , nucl .
b231 ( 1984 ) 419 .
t. banks , y. grossman , e. nardi , and y. nir , phys .
d 52 ( 1995 ) 5319 . after completion of this work we also received their paper f. m. borzumati , y. grossman , e. nardi , and y. nir , wis-96 - 21-ph , hep - ph/9606251 .
i. lee , nucl .
b246 ( 1984 ) 120 . c. s. aulakh and r. n. mohapatra , phys .
lett . 119b ( 1982 ) 136 .
j. ellis , g. gelmini , c. jarlskog , g. g. ross , and j. w. f. valle , phys .
lett . 150b ( 1985 ) 142 .
g. g. ross and j. w. f. valle , phys .
lett . 151b ( 1985 ) 375 .
k. s. babu and r. n. mohapatra , phys .
64 ( 1990 ) 1705 ; phys .
d 42 ( 1990 ) 3778 ; _ ibid .
_ 43 ( 1991 ) 2278 .
r. barbieri , m. m. guzzo , a. masiero , and d. tommasini , phys .
b 252 ( 1990 ) 251 ; e. roulet and d. tommasini , _ ibid . _ 256 ( 1991 ) 218 .
s. dawson , nucl .
b261 ( 1985 ) 297 .
g. farrar and p. fayet , phys .
lett . 76b ( 1978 ) 575 .
l. e. ibanez and g. g. ross , nucl .
b368 ( 1992 ) 3 .
d. kapetanakis , p. mayr , and h .-
nilles , phys .
b 282 ( 1992 ) 95 .
j. e. kim and h .-
nilles , phys . lett . 138b ( 1984 )
150 . for a recent discussion , see , n. polonsky and a. pomarol , phys .
d 51 ( 1995 ) 6532 . g. f. giudice and a. masiero , phys
b 206 ( 1988 ) 480 .
e. j. chun , j. e. kim , and h .-
nilles , nucl .
b370 ( 1992 ) 105 ; j. e. kim and h .-
nilles , mod .
lett . a 9 ( 1994 ) 3575 .
w. buchmuller and d. wyler , phys .
121b ( 1983 ) 321 ; l. j. hall and l. randall , nucl . phys .
b352 ( 1991 ) 289 ; m. dine and d. a. mcintire , phys .
d 46 ( 1992 ) 2594 .
r. peccei and h. quinn , phys .
38 ( 1977 ) 1440 .
m. cvetic and p. langacker , iassns - hep-95 - 90 , hep - ph/9511378 ; iassns - hep-95 - 113 , hep - ph/9602424 .
j. a. casas and c. munoz , phys .
b 306 ( 1993 ) 288 .
i. affleck , m. dine , and n. seiberg , nucl .
b256 ( 1985 ) 557 .
a. e. nelson and n. seiberg , nucl .
b416 ( 1994 ) 46 .
j. l. feng , n. polonsky , and s. thomas , phys .
b 370 ( 1996 ) 95 .
s. weinberg , phys .
40 ( 1978 ) 223 ; f. wilczek , phys .
40 ( 1978 ) 279 . for review and references ,
see j. e. kim , phys .
149 ( 1987 ) 1 .
p. j. steinhardt and m. s. turner , phys .
lett . 129b ( 1983 )
51 ; k. babu , s. m. barr , and d. seckel , phys .
b 336 ( 1994 ) 213 ; g. dvali , ifup - th-21 - 95 , hep - ph/9505253 ; m. kawasaki , t. moroi , and t. yanagida , ut-730 , hep - ph/9510461 .
j. bagger , e. poppitz , and l. randall , nucl .
b426 ( 1994 ) 3 .
j. polonyi , budapest report no .
kfki-93 ( 1977 ) .
a. h. chamseddine and h. dreiner , nucl .
b458 ( 1996 ) 65 ; d. j. castano , d. z. freedman , and c. manuel , nucl .
b461 ( 1996 ) 50 .
s. hawking , comm .
43 ( 1975 ) 199 ; phys .
lett . 195b ( 1987 ) 337 ; g. v. lavrelashvili , v. rubakov , and p. tinyakov , jetp lett .
46 ( 1987 ) 167 ; s. giddings and a. strominger , nucl .
b307 ( 1988 ) 854 ; s. coleman , _ ibid . _ b310 ( 1988 ) 643 .
t. banks and l. j. dixon , nucl . phys .
b307 ( 1988 ) 93
. g. d. coughlan , w. fischler , e. w. kolb , s. raby , and g. g. ross , phys .
lett . 113b ( 1983 ) 59 .
t. banks , d. b. kaplan , and a. e. nelson , phys . rev .
d 49 ( 1994 ) 779 . for general formulae for the soft supersymmetry breaking parameters and for references , see , for example , v. s. kaplunovsky and j. louis , phys .
b 306 ( 1993 ) 269 .
d. e. brahm , l. j. hall , and s. d. h. hsu , phys .
d 42 ( 1990 ) 1860 .
r. hempfling , mpi - pht/95 - 59 , hep - ph/9511288 .
the aleph collabortion , d. buskulic et al .
b 349 ( 1995 ) 585 .
s. s. gerstein and ya .
b. zeldovich , zh .
pisma 4 ( 1972 ) 174 ; r. cowsik and j. mcclelland , phys .
( 1972 ) 669 ; p. hut , phys .
lett . 96b ( 1977 ) 85 ; k. sato and h. kobayashi , prog .
58 ( 1977 ) 1775 ; b. w. lee and s. weinberg , phys .
39 ( 1977 ) 165 ; m. i. vysotsky , a. d. doglov , and ya . b. zeldovich , jetp lett . 26
( 1977 ) 188 . v. barger , m. s. berger , r. j. n. phillips , and t. wohrmann , madph-95 - 910 , hep - ph/9511473 ; hempfling @xcite ; h. dreiner and h. pois , eth - th/95 - 30 , hep - ph/9511444 ; b. de carlos and p. l. white , susx - th/96 - 003 , hep - ph/9602381 .
note that the renormalization effects in the soft parameters discussed by de carlos and white do not appear in our case .
b. gato , j. leon , j. perez - mercader , and m. quiros , nucl .
b260 ( 1985 ) 203 .
a. bouquet and p. salati , nucl .
b284 ( 1987 ) 557 , and refernces therein .
d. comelli , a. masiero , m. pietroni , and a. riotto , phys .
b 324 ( 1994 ) 397 .
a. yu smirnov and f. vissani , phys .
b 341 ( 1994 ) 173 ; a. brignole , h. murayama , and r. rattazzi , _ ibid . _ 335 ( 1994 ) 345 .
hall and suzuki @xcite ; dawson @xcite ; r. n. mohapatra , phys .
d 34 ( 1986 ) 3457 ; v. barger , g. f. giudice , and t. han , phys .
d 40 ( 1989 ) 2987 ; r. barbieri , d. e. brahm , l. j. hall , and s. d. h. hsu , phys .
b 238 ( 1990 ) 86 ; brahm et al .
@xcite ; i. hinchliffe and t. kaeding , phys .
d 47 ( 1993 ) 279 ; l. roszkowski , in _ physics and experiments with linear @xmath409 colliders _ , eds . f. a. harris , s. l. olsen , s. pakvasa , and x. tata ( world scientific , singapore , 1993 ) p. 854 , hep - ph/9309208 ; g. bhattachrayya , j. ellis and k. sridhar , mod .
lett . a 10 ( 1995 ) 1583 ; g. bhattachrayya and d. choudhury , _ ibid .
_ 1699 ; c. e. carlson , p. roy , and m. sher , phys
b 357 ( 1995 ) 99 ; m. hirsch , h. v. klapdor - kleingrothaus , and s. g. kovalenko , phys .
lett . 75 ( 1995 ) 17 ; k. s. babu and r. n. mohapatra , _ ibid .
_ 2276 ; f. de campos , m. a. garcia - jareno , a. s. joshipura , j. rosiek , and j. w. f. valle , nucl .
b451 ( 1995 ) 3 ; m. nowakowski and a. pilaftsis , _ ibid .
_ b461 ( 1996 ) 19 ; banks et al .
, @xcite ; k. agashe and m. graesser , lbl-37823 , hep - ph/9510439 ; a. yu . smirnov and f. vissani , ic-96 - 16 , hep - ph/9601387 ; f. vissani , to be published in _ the proceedings of v hellenic school and workshops on elementary particle physics _ , corfu , greece , sept .
1995 , ic-96 - 32 , hep - ph/9602395 ; de carlos and white @xcite ; d. choudhury and p. roy , mpi - pth/96 - 20 , hep - ph/9603363 ; m. chaichian and k. huitu , hu - seft - r-1996 - 09 , hep - ph/9603412 .
see , for example , s. bertolini , f. m. borzumati , a. masiero , and g. ridolfi , nucl .
b353 ( 1991 ) 591 .
s. dimopoulos , r. esmailzadeh , l. j. hall , j .-
merlo , and g. d. starkman , phys .
d 41 ( 1990 )
2099 . h. dreiner and g. g. ross , nucl .
b365 ( 1991 ) 597 .
h. dreiner and r. j. n. phillips , nucl .
b367 ( 1991 ) 591 ; d. p. roy , phys .
b 283 ( 1992 ) 270 ; j. butterworth and h. dreiner , _ ibid .
_ b397 ( 1993 ) 3 ; h. dreiner , m. guchait , and d. p. roy , phys . rev .
d 49 ( 1994 ) 3270 ; v. barger , m. s. berger , p. ohmann , and r. j. n. phillips , phys .
d 50 ( 1994 ) 4299 ; g. bhattachrayya , d. choudhury , and k. sridhar , phys .
b 355 ( 1995 ) 193 ; m. guchait and d. p. roy , tifr / th/96 - 10 , hep - ph/9603219 . s. dimopoulos and l. j. hall , phys . lett .
b 207 ( 1988 ) 210 ; barger et al . @xcite ; r. godbole , p. roy , and x. tata , nucl .
b401 ( 1993 ) 67 ; v. barger , w .- y .
keung , and r. j. n. phillips , phys .
b 364 ( 1995 ) 27 ; n. k. krasnikov , hep - ph/9511464 ; d. choudhury cern - th/95 - 250 , hep - ph/9511466 .
a. a. bhatti , fermilab - conf-95/192-e ; the cdf collaboration , f. abe et al .
, fermilab - pub-96/020-e , hep - ex/9601008 .
r. m. harris , fermilab - conf-95/152-e , hep - ex/9506008 ; the cdf collaboration , f. abe et al .
( 1995 ) 3538 .
v. barger , m. s. berger , and r. j. n. phillips , madph-95 - 920 , hep - ph/9512325 ; p. kraus and f. wilczek , caltech-68 - 2032 , hep - ph/9601279 ; j. ellis and d. a. ross , cern - th/96 - 108 , hep - ph/9604432 .
( the light gluino case was discussed by i. terekhov and l. clavelli , uahep964 , hep - ph/9603390 . )
b. a. campbell , s. davidson , j. ellis , and k. a. olive , phys .
b 256 ( 1991 ) 457 ; h. dreiner and g. g. ross , nucl . phys .
b410 ( 1993 ) 188 .
a. s. joshipura and m. nowakowski , phys .
d 51 ( 1995 ) 2421 ; _ ibid . _ 5271
. g. f. giudice , in _ proceedings of the xii warsaw symposium on elementary particle physics _ , kazimierz , poland , 1989 , eds .
z. ajduk , s. pokorski , and a. k. wroblewski ( world scientific , singapore , 1990 ) p. 608 . for review and references ,
see i. z. rothstein , talk given at the joint meeting of the international symposium on particles , strings and cosmology and the ixx johns hopkins workshop on current problems in particle theory , baltimore , md , 1995 , hep - ph/9506443 .
k. enqvist , a. masiero , and a. riotto , nucl .
b373 ( 1992 ) 95 .
k. rajagopal , m. s. turner , and f. wilczek , nucl .
b358 ( 1991 ) 447 ; j. e. kim , phys .
67 ( 1991 ) 3465 ; e. j. chun , j. e. kim , and h .-
nilles , phys .
b 287 ( 1992 ) 123 ; d. h. lyth , phys . rev .
d 48 ( 1993 ) 4523 ; e. j. chun and a. lukas , phys .
b 357 ( 1995 ) 43 ; s. chang and h. b. kim , ufift - hep-96 - 8 , hep - ph/9604222 .
e. j. chun , phys .
b 367 ( 1996 ) 226 .
l. ibanez and g. g. ross , phys .
b 332 ( 1993 ) 100 ; p. binetruy and p. ramond , _ ibid .
_ 350 ( 1995 ) 49 ; v. jain and r. shrock , _ ibid .
_ 352 ( 1995 ) 83 ; itp - sb-95 - 22 , hep - ph/9507238 ; y. nir , phys .
b 354 ( 1995 ) 107 . e. ma and p. roy , phys .
d 41 ( 1990 ) 988 ; e. ma , in _ proceedings of the tallinn symposium on neutrino physics _ , lohusalu , estonia , 1993 , eds . i. ots and l. palgi ( estonian acad .
phys . , 1994 )
, hep - ph/9310280 . | in supersymmetric models a tree - level neutrino mass could originate from the ( weak - scale ) superpotential .
we propose and examine a realization of that idea , which arises naturally in the framework of a spontaneously broken @xmath0 @xmath1-symmetry .
the solution to the neutrino mass problem could shed light in this framework on the possible resolution of the @xmath2 problem .
furthermore , the suppression of the neutrino mass in comparison to the weak scale arises dynamically and need not be encoded in the superpotential .
the latter mechanism operates , for example , in universal models for the soft supersymmetry breaking terms .
phenomenological and cosmological implications of the model are also discussed , some of which are shown to hold more generally .
we also note that future signatures could include observable enhancement of dijet and multijet production rates and a correlation between the supersymmetric and neutrino spectra . |
v sge is a blue star with a mean brightness around 11 mag which has been shown to vary between 9.614.7 mag since its discovery in 1902 .
it shows wide eclipses at a period of 051419 , a small secondary eclipse , and complex emission line behaviour ( herbig 1965 ) .
extinction estimates vary between @xmath3 ( herbig 1965 ) and @xmath4 ( verbunt 1987 ) implying a distance of 0.72.7 kpc .
supersoft x - ray binaries ( ssb ; see greiner 1996 and references therein ; van teeseling 1998 ) were established as a new class of accreting binaries during the early 90ies with rosat ( trmper 1991 ; greiner 1991 ) and are thought to contain white dwarfs accreting mass at rates sufficiently high to allow stable nuclear surface burning of the accreted matter ( van den heuvel 1992 ) .
ssb have luminosities of @xmath5 10@xmath610@xmath7 ergs s@xmath2 , but their characteristic temperatures of 2040 ev imply strong attenuation by the interstellar medium .
thus , most of the known ssb are located in external galaxies ( e.g. greiner 1996 ) making detailed optical observations difficult .
it is therefore of great interest to identify galactic ssb .
it has recently been suggested ( steiner & diaz 1998 ; patterson 1998 ) that v sge has spectroscopic and photometric properties which are very similar to those of ssb . this suggestion is based on characteristics which are typical for ssb , but are rare or even absent among canonical cataclysmic variables : ( 1 ) the presence of both o and n emission lines , ( 2 ) a he@xmath8/h@xmath9 emission line ratio @xmath10 , ( 3 ) rather high absolute magnitudes and very blue colours , and ( 4 ) orbital lightcurves which are characterized by a wide and deep eclipse .
the suggestion of the similarity of v sge to ssb is almost entirely based on optical and ultraviolet data . in this paper
, we investigate the archival rosat data of v sge and discuss them in the context of the long - term optical behaviour of v sge . hoard ( 1996 ) reported the detection of v sge as a soft x - ray source in the nov .
1992 rosat observation , but did not perform a spectral fit .
verbunt et al .
( 1997 ) already reported the non - detection of v sge during the rosat all - sky survey .
clrccccccr date & obs - id@xmath11 & t@xmath12 & offaxis & cr@xmath13 & hr1@xmath14 & hr2@xmath14 & x - ray & optical & d@xmath15 + & & ( sec ) & angle & ( cts / s ) & & & state & state & + oct . 1931 , 1990 &
& 50 & 055 & @xmath160.054 & & & & bright & + nov . 23/24 , 1991 & 400155p & @xmath1710235 & 039 & [email protected] & [email protected] & [email protected] & hard & bright & @xmath1916 + nov .
1012 , 1992 & 300182p & @xmath1727745 & 309 & [email protected] & [email protected] & [email protected] & soft & intermediate & 28 + apr .
1924 , 1994 & 300311h & @xmath1724610 & 018 & @xmath160.00044 & & & & bright & +
may 1213 , 1994 & 300311h & 4700 & 018 & [email protected] & & & very soft & faint & 1 + oct . 18/19 , 1994 & 300311h-1 & @xmath1718440 & 017 & [email protected] & & & hard & bright & 5 + may 12/13 , 1997 & 300582h & @xmath1715700 & 017 & [email protected] & & & soft & intermediate & 1 + @xmath11 the letter after the observation i d number gives the rosat detector : p = pspc , h = hri .
+ @xmath13 count rates in the corresponding detector in the 0.12.4 kev range ( pspc : channels 11240 ) .
upper limits are 3@xmath20 confidence level .
note the different pspc to hri count rate conversion factors of 2.7:1 and 7.8:1 for hard and soft spectrum sources .
+ @xmath14 hardness ratios with hr1 = @xmath21 and hr2 = @xmath22 , where @xmath23 kev ) , @xmath24 kev ) , @xmath25 kev ) , and @xmath26 kev ) are the counts in the given energy range .
+ @xmath15 distance between best - fit x - ray and optical position . for the optical position @xmath27(2000.0)=2020147 ,
@xmath28(2000.0)=+210610 has been used as determined from the second generation dss .
this position differs from the simbad position by @xmath29=12 and @xmath30=4 .
[ log ] v sge has been the target of three dedicated pointed pspc and hri observations ( one of these splits into 3 separate observation intervals ) , and in addition is in the field of view of another pspc observation ( table [ log ] ) .
the results of these observations are quite diverse : v sge has not been detected during the rosat all - sky survey in 1990 and a long rosat hri pointing in april 1994 , but has been detected during all other observations , even in a much shorter hri observation .
thus , v sge shows strong x - ray variability with an amplitude of a factor of 140 .
in addition , the x - ray spectral characteristics during two rosat pspc pointings obtained 1 yr apart show a remarkable difference : at one occasion v sge has a ` supersoft ' x - ray spectrum , at another occasion the spectrum is very hard .
the diversity of x - ray measurements looks more ordered when it is compared with the optical brightness of v sge .
this binary system is included in the roboscope program of automatic long - term monitoring the results of which led to the classification of three distinct optical states : bright state ( v@xmath1611 mag ) , intermediate state ( v@xmath311112 mag ) and faint state ( v@xmath3212 mag ) ( robertson 1997 ) .
we have combined the optical lightcurve obtained by these observations with data collected in the vsolj database ( www.kusastro.kyoto-u.ac.jp/vsnet/ ) and plotted these in fig .
[ lc ] together with the times of the rosat observations .
this suggests that during optical bright state v sge is a hard , but rather faint x - ray source , while during optical faint state v sge is a more luminous and very soft x - ray source . during the intermediate optical state also the x - ray spectrum is intermediate with respect to the very soft and hard spectrum .
to obtain an idea about the x - ray spectral parameters during the soft x - ray state , we fit the nov .
1992 pspc spectrum with a solar - abundance lte @xmath33 white dwarf atmosphere model ( van teeseling 1994 ) .
the @xmath34 contours are shown in fig .
[ logg9 ] .
the @xmath35 contour suggests a temperature @xmath36k and a bolometric luminosity @xmath37ergs@xmath2 ( with @xmath38kpc ) , but lower temperatures and higher luminosities are still acceptable within the 90% confidence contour .
if we require that the soft x - ray absorbing column is at least @xmath39@xmath40 as derived from the @xmath41absorption dip ( @xmath42 ; verbunt 1987 ) , we find @xmath43k and @xmath44ergs@xmath2 . with @xmath45@xmath40 , only for @xmath46200000k a luminosity of @xmath47ergs@xmath2 is reached .
it is possible , however , that because of the very high orbital inclination the soft x - ray absorption is much larger than the ultraviolet absorption .
a similar discrepancy is known for cal87 ( cf .
hutchings 1995 ; parmar 1997 ) .
if we relax the absorption constraint and assume that v sge is a ssb with @xmath47ergs@xmath2 , we find @xmath48k and a radius @xmath49 cm consistent with a white dwarf . however , because of a very high orbital inclination , the white dwarf may be completely obscured from view by the accretion disk rim , in which case the observable luminosity ( from x - rays scattered into the line of sight ) may be much less than @xmath50ergs@xmath2 .
we note that there is no significant modulation of the soft x - rays on the orbital period .
a factor of 45 increase in hri count rate occurred within less than three weeks in april / may 1994 during which the optical brightness decreased and v sge eventually became a very soft x - ray source .
though the rosat observations during this optical state transition have been performed with the hri , the grossly different spectral shapes are easy to recognize ( fig .
[ logg9 ] , bottom ) .
the anti - correlation of soft x - ray emission with optical brightness is reminiscent of the behaviour of the ssb 513(reinsch 1996 ; southwell 1996 ) .
513 turns on as a supersoft x - ray source only during @xmath51mag optical dips , which occur every 100200 days and last about @xmath52 days .
this behaviour has been explained by assuming that the shell - burning white dwarf in 513 has normally expanded to a few @xmath53 cm and radiates its luminosity in the extreme - ultraviolet . during the optical faint states , the white dwarf contracts with almost constant bolometric luminosity to @xmath54 cm , and radiates in the soft x - ray band .
the model that has been suggested for 513cannot explain the observational data of v sge .
first , the optical brightness changes of v sge are very rapid : both the faint-/bright - state transitions as well as the succession of different faint states may occur on timescales of @xmath55day ( compared to the smooth decline of several days in 513 ) .
such very rapid changes are only possible if the white dwarf envelope expands and contracts on the kelvin - helmholtz timescale and the mass of the expanding envelope is rather small ( @xmath56 ) .
such a small envelope mass is difficult to accept for a white dwarf with stable shell burning ( e.g. prialnik & kovetz 1995 ) .
second , the expected optical eclipse would become deeper when the system becomes brighter , opposite to what has been observed ( patterson 1998 ) . before we speculate on a possible explanation for the observed x - ray properties of v sge , we note the similarity of the v sge behaviour to that of vy scl stars ( as has been noted with respect to the optical behaviour also by robertson 1997 ) .
in a recent survey of the available rosat data of vy scl stars ( greiner 1998 ) a relatively hard x - ray spectrum was found for vy scl stars during optical bright state ( see also van teeseling 1996 ) .
moreover , observations of the vy scl star v751 cyg during its 1997 optical faint state have revealed luminous and very soft x - ray emission ( greiner 1998 ) , similar to that of v sge in its faint state .
inspection of the change in eclipse depth from faint to bright state ( e.g. fig . 5 in patterson 1998 ) shows that it is possible to reproduce this change by an increase of uneclipsed flux , while the eclipsed light ( presumably from the irradiated accretion disk ) remains almost constant . if the flux from the irradiated disk ( and therefore also from the irradiated secondary )
remains unchanged , this would suggest that a brightening of v sge is caused by an increasing amount of extended luminous ( outflowing ? ) matter
. this would also be consistent with the emission lines , which only show partial eclipse effects , and which increase in strength when v sge brightens ( e.g. herbig 1965 ) , indicating either an increasing amount of line emitting matter or an increasing amount of ionizing flux .
( mauche ( 1997 ) note that the heii@xmath57/civ@xmath58 and nv@xmath59/civ@xmath58 line ratios decrease when v sge brightens . ) additional extended gas could increase the amount of soft x - ray absorption and make the soft x - ray component fainter and harder and make it even completely undetectable . because it is impossible to produce the hardness ratios _ and _ count rate observed in nov .
1991 by simply adding more absorption to an absorbed hot - white - dwarf spectrum , we conclude that in nov .
1991 an additional harder x - ray component was present .
the hardness ratios and count rate of the nov .
1991 pspc observation can be explained perfectly with a thermal bremsstrahlung spectrum , absorbed with @xmath60@xmath40 , a temperature of a few kev , and a 0.12.4kev luminosity of @xmath1ergs@xmath2 ( at 1kpc ) .
the same bremsstrahlung component may also have been present during the soft x - ray state in nov .
1992 without significantly affecting the confidence intervals in fig .
[ logg9 ] of the spectral parameters of the soft component .
we note that such a bremsstrahlung component is not inconsistent with the x - ray flux of an evolved secondary in a 12hr binary ( cf .
dempsey 1993 ) . a simple wind model for the recently observed radio flux density of v sge implies a mass - loss rate of the order of 10@xmath61/yr ( lockley 1997 ) . with their ( assumed ) terminal velocity of 1500 km / s
this wind zone is completely opaque for x - rays up to 0.7kev , even if the wind is assumed to be circumbinary instead of arising from one component .
since the radio measurement has been obtained during optical high state , it supports the above described scenario .
we note that the colliding wind scenario as discussed in lockley ( 1997 ) would predict a positive correlation between optical and x - ray emission , contrary to our finding .
vitello & shlosman ( 1993 ) have modeled the uv line shapes of v sge assuming a biconical accretion disk wind , and need high mass - loss rates to explain the observations , while they did not consider the possibility of a luminous shell - burning white dwarf .
we also note that irradiation - induced winds with a rate of 10@xmath6210@xmath61/yr are expected in ssb ( van teeseling & king 1998 ) .
we conclude that it is possible to explain the optical and x - ray behaviour of v sge by assuming a variable amount of extended uneclipsed matter which contributes significantly in the optical ( by reprocessing of soft x - rays ? ) and may completely absorb the ` supersoft ' x - ray component during the optical bright / hard x - ray state .
the x - ray properties of v sge , in any case , support the presence of a hot luminous white dwarf , possibly with hydrogen shell - burning , and do not hinder the addition of v sge to the class of supersoft x - ray binaries .
more detailed modelling of the changing optical light curves and ultraviolet and optical spectra are necessary to test this scenario .
we thank kent honeycutt for providing his roboscope data as published in robertson ( 1997 ) in electronic format .
jg and avt are supported by the german bundesministerium fr bildung , wissenschaft , forschung und technologie ( bmbf / dlr ) under contract no .
fkz 50 qq 9602 3 and 50or96098 , respectively . the project is supported by bmbf / dlr and the max - planck - society .
this research has made use of the simbad database , operated at cds , strasbourg , france and the digitized sky survey ( dss ) produced at the space telescope science institute under us government grant nag w-2166 . | we investigate the rosat x - ray properties of v sge , which has been proposed to be related to supersoft x - ray binaries . during optical bright states , v sge is a faint hard x - ray source , while during optical faint states ( @xmath0 mag ) , v sge is a ` supersoft ' x - ray source .
spectral fitting confirms that v sge s x - ray properties during its soft x - ray state may be similar to those of supersoft x - ray binaries , although a much lower luminosity can not be excluded .
it is possible to explain the different optical / x - ray states by a variable amount of extended uneclipsed matter , which during the optical bright states contributes significantly to the optical flux and completely absorbes the soft x - ray component .
an additional , perhaps permanent , hard x - ray component , such as a bremsstrahlung component with a 0.12.4kev luminosity of @xmath1ergs@xmath2 , must be present to explain the x - ray properties during the optical bright / hard x - ray state . |
* fusion * - fusion cross sections can be calculated from the equation @xcite @xmath0 where @xmath1 is the center of mass energy , @xmath2 is the reduced wavelength and @xmath3 .
the cross section is proportional to @xmath4 , the area of the quantum wave .
each part of the wave corresponds to different impact parameters having different probabilities for fusion . as the impact parameter increases , so does the angular momentum , hence the reason for the @xmath5 term .
@xmath6 is the probability that fusion occurs at a given impact parameter , or angular momentum .
sometimes , for a better visualization , or for extrapolation to low energies , one uses the concept of _ astrophysical s - factor _ , redefining the cross section as @xmath7 , \label{se}\ ] ] where @xmath8 , with @xmath9 being the relative velocity .
the exponential function is an approximation to @xmath10 for a square - well nuclear potential plus coulomb potential , whereas the factor @xmath11 is proportional to the area appearing in eq .
[ eqnf ] .
* many reaction channels * - eq .
[ eqnf ] does not work in most situations . only by including coupling to other channels
, the fusion cross sections can be reproduced . in coupled channels
schemes one expands the total wavefunction for the system as @xmath12 where @xmath13 form the channel basis , @xmath14 is a dynamical variable ( e.g. , the distance between the nuclei ) , and @xmath15 are intrinsic coordinates . inserting this expansion in the schrdinger equation yields a set of cc equations in the form @xcite @xmath16 where @xmath17 is whatever
potential couples the channels @xmath18 and @xmath19 and @xmath20 is some sort of transition energy , or transition momentum . in the presence of continuum states , continuum - continuum coupling ( relevant for breakup channels ) can be included by discretizing the continuum .
this goes by the name of _ continuum discretized coupled - channels _ ( cdcc ) calculations .
there are several variations of cc equations , e.g. , a set of differential equations for the wavefunctions , instead of using basis amplitudes .
coupled channels calculations with a large number of channels in continuum couplings are somewhat challenging : the phases of matrix elements can add destructively or constructively , depending on the system and on the nuclear model .
such suppressions or enhancements are difficult to interpret .
* radiative capture * - for reactions involving light nuclei , only a few channels are of relevance . in this case ,
a real potential is enough for the treatment of fusion .
for example , radiative capture cross sections of the type @xmath21 and @xmath22 ( @xmath23electric ( magnetic ) l - pole ) transitions can be calculated from @xcite @xmath24 where @xmath25 is an em operator , and @xmath26 is a multipole matrix element involving bound ( @xmath27 ) and continuum ( @xmath28 ) wavefunctons . for electric multipole transitions ( @xmath29 ) ,
@xmath30 where @xmath31 are radial wavefunctions .
the total direct capture cross section is obtained by adding all multipolarities and final spins of the bound state ( @xmath32 ) , @xmath33 where @xmath34 are spectroscopic factors . * asymptotic normalization coefficients * - in a microscopic approach , instead of single - particle wavefunctions one often makes use of overlap integrals , @xmath35 , and a many - body wavefunction for the relative motion , @xmath36 .
both @xmath35 and @xmath37 might be very complicated to calculate , depending on how elaborated the microscopic model is .
the variable @xmath38 is the relative coordinate between the nucleon and the nucleus @xmath39 , with all the intrinsic coordinates of the nucleons in @xmath39 being integrated out .
the direct capture cross sections are obtained from the calculation of @xmath40 ) , @xmath41 , \label{whitt}%\end{aligned}\ ] ] where ( ) are for protons and [ ] for neutrons .
the binding energy of the @xmath42 system is related to @xmath43 by means of @xmath44 , @xmath45 is the whittaker function and @xmath46 is the modified bessel function . in eq .
[ whitt ] , @xmath47 is the _ asymptotic normalization coefficient _
( anc ) . in the calculation of @xmath48
above , one often meets the situation in which only the asymptotic part of @xmath35 and @xmath49 contributes significantly to the integral over @xmath38 . in these situations , @xmath50 is also well described by a simple two - body scattering wave ( e.g. coulomb waves ) .
therefore the radial integration in @xmath48 can be done accurately and the only remaining information from the many - body physics at short - distances is contained in the asymptotic normalization coefficient @xmath47 , i.e. @xmath51 .
we thus run into an _ effective theory for radiative capture cross sections _
, in which the constants @xmath47 carry all the information about the short - distance physics , where the many - body aspects are relevant @xcite .
it is worthwhile to mention that these arguments are reasonable for proton capture at very low energies , because of the coulomb barrier .
* resonating group method * - one immediate goal can be achieved in the coming years by using the _ resonating group method _ ( rgm ) or the generator coordinate method ( gcm ) .
these are a set of coupled integro - differential equations of the form @xcite @xmath52 g_{\alpha'}({\bf r'})=0,\label{rgm}\ ] ] where @xmath53 . in these equations @xmath54 is the hamiltonian for the system of two nuclei ( a and b ) with the energy @xmath1 , @xmath55 is the wavefunction of nucleus a ( and b ) , and @xmath56 is a function to be found by numerical solution of eq .
[ rgm ] , which describes the relative motion of a and b in channel @xmath14 .
full antisymmetrization between nucleons of a and b are implicit .
modern nuclear shell - model calculations are able to provide the wavefunctions @xmath55 for light nuclei @xcite .
the astrophysical s - factor for the reaction @xmath57be@xmath58p@xmath59b was calculated @xcite and excellent agreement was found with the experimental data in both direct and indirect measurements @xcite .
the low- and high - energy slopes of the s - factor obtained with a many - body microscopic calculation @xcite is well described by the fit @xmath60 where e is the relative energy ( in mev ) of p@xmath61be in their center - of - mass .
this equation corresponds to a pad approximant of the s - factor .
a subthreshold pole due to the binding energy of @xmath62b is responsible for the denominator @xcite .
figure [ s17ber ] show the world data on @xmath63be(p,@xmath64)@xmath62b compared to a few of the theoretical calculations .
the recent compilation published in ref .
@xcite recommends @xmath65 ev b.
* transfer reactions * - transfer reactions @xmath66 are effective when a momentum matching exists between the transferred particle and the internal particles in the nucleus .
thus , beam energies should be in the range of a few 10 mev per nucleon .
low energy reactions of astrophysical interest can be extracted directly from breakup reactions @xmath67 by means of the _ trojan horse method _ ( thm ) @xcite . if the fermi momentum of the particle @xmath39 inside @xmath68 compensates for the initial projectile velocity @xmath69 , the low energy reaction @xmath70 is induced at very low ( even vanishing ) relative energy between @xmath71 and @xmath39 .
basically , this technique extends the method of transfer reactions to continuum states .
very successful results using this technique have been reported @xcite .
another transfer method , coined as anc technique @xcite relies on fact that the amplitude for the radiative capture cross section @xmath72 is given by @xmath73 where @xmath74 is the integration over the internal coordinates @xmath75 , and @xmath76 , of @xmath27 and @xmath39 , respectively . for low energies , the overlap integral @xmath77 is dominated by contributions from large @xmath78
thus , what matters for the calculation of the matrix element @xmath79 is the asymptotic value of @xmath80 , where @xmath81 is the anc and @xmath82 is the whittaker function .
this coefficient is the product of the _ spectroscopic factor _ and a normalization constant which depends on the details of the wave function in the interior part of the potential .
thus , @xmath81 is the only unknown factor needed to calculate the direct capture cross section .
these normalization coefficients can be found from : 1 ) analysis of classical nuclear reactions such as elastic scattering [ by extrapolation of the experimental scattering phase shifts to the bound state pole in the energy plane ] , or 2 ) peripheral transfer reactions whose amplitudes contain the same overlap function as the amplitude of the corresponding astrophysical radiative capture cross section .
one of the many advantages of using transfer reaction techniques over direct measurements is to avoid the treatment of the _ electron screening problem _ @xcite . *
intermediate energy coulomb excitation * - the coulomb excitation cross section is given by @xmath83 where @xmath84 is the reduced transition probability of the projectile nucleus , @xmath85 @xmath86 is the multipolarity of the excitation , and @xmath87 .
the relativistic corrections to the rutherford formula for @xmath88 ( relevant for collisions at @xmath89 mev / nucleon and above ) has been investigated in ref .
it was shown that the scattering angle increases by up to 6% when relativistic corrections are included in nuclear collisions at 100 mev / nucleon .
the effect on the elastic scattering cross section is even more drastic : up to @xmath90 for center - of - mass scattering angles around 0 - 4 degrees . the orbital integrals @xmath91 ) contain the information about relativistic corrections .
inclusion of absorption effects in @xmath91 ) due to the imaginary part of an optical nucleus - nucleus potential where worked out in ref .
these orbital integrals depend on the lorentz factor @xmath92 , with @xmath28 being the speed of light , on the multipolarity @xmath93 , and on the _ adiabaticity parameter _
@xmath94 , where @xmath95 is the excitation energy ( in units of @xmath96 ) and @xmath27 is the impact parameter . ref .
@xcite has shown that at 10 mev / nucleon the relativistic corrections are important only at the level of 1% . at 500 mev / nucleon ,
the correct treatment of the recoil corrections is relevant on the level of 1% .
thus the non - relativistic treatment of coulomb excitation @xcite can be safely used for energies below about 10 mev / nucleon and the relativistic treatment with a straight - line trajectory @xcite is adequate above about 500 mev / nucleon .
however at energies around 50 to 100 mev / nucleon , accelerator energies common to most radioactive beam facilities , it is very important to use a correct treatment of _ recoil and relativistic effects _ , both kinematically and dynamically . at these energies
, the corrections can add up to 50% .
these effects were also shown in ref .
@xcite for the case of excitation of giant resonances in collisions at intermediate energies .
a reliable extraction of useful nuclear properties , like the electromagnetic response ( b(e2)-values , @xmath64-ray angular distribution , etc . ) from coulomb excitation experiments at intermediate energies requires a proper treatment of special relativity @xcite .
the dynamical relativistic effects have often been neglected in the analysis of experiments elsewhere ( see , e.g. @xcite ) .
the effect is highly non - linear , i.e. a 10% increase in the velocity might lead to a 50% increase ( or decrease ) of certain physical observables .
a general review of the importance of the relativistic dynamical effects in intermediate energy collisions has been presented in ref .
@xcite . * the coulomb dissociation method * - the coulomb dissociation method is quite simple .
the ( differential , or angle integrated ) coulomb breakup cross section for @xmath97 follows from eq .
[ cross_2 ] .
it can be rewritten as @xmath98 where @xmath99 is the energy transferred from the relative motion to the breakup , and @xmath100 is the photo nuclear cross section for the multipolarity @xmath101 and photon energy @xmath99 .
the function @xmath102 , sometimes called _ virtual photon numbers _ , depends on @xmath99 , the relative motion energy , nuclear charges and radii , and the scattering angle @xmath103 .
@xmath102 can be reliably calculated @xcite for each multipolarity @xmath101 .
time reversal allows one to deduce the radiative capture cross section @xmath104 from @xmath105 .
this method was proposed in ref .
@xcite and has been tested successfully in a number of reactions of interest for astrophysics .
the most celebrated case is the reaction @xmath57be@xmath58p@xmath59b @xcite , followed by numerous experiments in the last decade ( see e.g. ref .
@xcite ) . eq .
[ cdmeth ] is based on first - order perturbation theory .
it also assumes that the nuclear contribution to the breakup is small , or that it can be separated under certain experimental conditions .
the contribution of the nuclear breakup has been examined by several authors ( see , e.g. @xcite ) .
@xmath62b has a small proton separation energy ( @xmath106 kev ) . for such loosely - bound systems it had been shown that multiple - step , or higher - order effects , are important @xcite .
these effects occur by means of continuum - continuum transitions .
detailed studies of dynamic contributions to the breakup were explored in refs .
@xcite and in several other publications which followed .
the role of higher multipolarities ( e.g. , e2 contributions @xcite in the reaction @xmath57be@xmath58p@xmath59b ) and the coupling to high - lying states has also to be investigated carefully .
it has also been shown that the influence of giant resonance states is small @xcite .
* knock - out reactions * - the early interest in knockout reactions came from studies of nuclear halo states , for which the narrow momentum distributions of the core fragments in a qualitative way revealed the large spatial extension of the halo wave function .
it was shown @xcite that the longitudinal component of the momentum ( taken along the beam or @xmath107 direction ) gave the most accurate information on the intrinsic properties of the halo and that it was insensitive to details of the collision and the size of the target .
in contrast to this , the transverse distributions of the core are significantly broadened by diffractive effects and by coulomb scattering . for experiments that observe the nucleon produced in elastic breakup , the transverse momentum is entirely dominated by diffractive effects , as illustrated @xcite by the angular distribution of the neutrons from the reaction @xmath108be(@xmath109be,@xmath110be+n)x . in this case , the width of the transverse _ momentum distribution _ reflects essentially the size of the target @xcite .
total knockout cross sections for removing the @xmath111 halo neutron of @xmath112c , bound by 1.218 mev , in the reaction @xmath113be(@xmath112c,@xmath114c@xmath115 ) .
the solid curve is obtained with the use of free nucleon - nucleon cross sections .
the dashed curve includes the geometrical effects of pauli blocking .
the dashed - dotted curve is the result using the brueckner theory , and the dotted curve is a phenomenological parametrization . ] to test the influence of the medium effects in nucleon knockout reactions , we consider the removal of the @xmath111 halo neutron of @xmath112c , bound by 1.218 mev .
the reaction studied is @xmath113be(@xmath112c,@xmath114c@xmath115 ) .
the total cross sections as a function of the bombarding energy are shown in figure [ sko15c ] .
the solid curve is obtained with the use of free nucleon - nucleon cross sections .
the dashed curve includes the geometrical effects of pauli blocking .
the dashed - dotted curve is the result using the _ brueckner theory _ , and the dotted curve is the phenomenological parametrization of the free cross section .
longitudinal momentum distribution for the residue in the @xmath113be(@xmath109be,@xmath110be ) , reaction at 250 mev / nucleon .
the dashed curve is the cross section calculated using the nn cross section from the brueckner theory and the solid curve is obtained the free cross section . ] in figure [ dsdpz ] we plot the longitudinal momentum distributions for the reaction @xmath113be(@xmath109be,@xmath110be ) , at 250 mev / nucleon @xcite .
the dashed curve is the cross section calculated using the nn cross section from the brueckner theory and the solid curve is obtained the free cross section .
one sees that the momentum distributions are reduced by 10% , about the same as the total cross sections , but the shape remains basically unaltered . if one rescales the dashed curve to match the solid one , the differences in the width are not visible @xcite .
there were many questions not addressed in this review , such as the role of central nucleus - nucleus collisions in determining phase transition , equation of state , and a quark - gluon plasma , all topics or relevance in astrophysics .
the review was more focused on the role of short - lived , exotic nuclei .
there are many important scientific questions to be addressed both experimentally and theoretically in nuclear physics of exotic nuclei with relevance for astrophysics .
these questions provide extreme challenges for experiments and theory . on the experimental side ,
producing the beams of radioactive nuclei needed to address the scientific questions has been an enormous challenge .
pioneering experiments have established the techniques and present - generation facilities have produced first exciting science results , but the field is still at the beginning of an era of discovery and exploration that will be fully underway once the range of next- generation facilities becomes operational .
the theoretical challenges relate to wide variations in nuclear composition and rearrangements of the bound and continuum structure , sometimes involving near - degeneracy of the bound and continuum states .
the extraction of reliable information from experiments requires a solid understanding of the reaction process , in addition to the structure of the nucleus . in astrophysics ,
new observations , for example the expected onset of data on stellar abundances , will require rare - isotope science for their interpretation .
h. esbensen , g.f .
bertsch , and k. snover , phys .
rev . lett . 94,042502 ( 2005 ) ; moshe gai , phys .
96 , 159201 ( 2006 ) ; h. esbensen , g.f .
bertsch , and k. snover , phys .
94 , 042502 ( 2005 ) . | the reactions of relevance for stellar evolution are difficult to measure directly in the laboratory at the small astrophysical energies . in recent years
indirect reaction methods have been developed and applied to extract low - energy astrophysical s - factors .
these methods require a combination of new experimental techniques and theoretical efforts , which are the subject of this review . |
in the complementary metal oxide semiconductor ( cmos ) devices , sio@xmath3 has been the first and only choice of gate oxide material for a long time . however , with the continued decrease in the feature size of the cmos devices , sio@xmath3 is no longer reliable as a gate oxide due to the high tunneling leakage current through it at small thickness .
major research efforts are continuously being carried out in search for a suitable replacement of the sio@xmath3 as a gate material.@xcite several promising candidates are metal oxides such as hfo@xmath3,@xcite zro@xmath3@xcite and al@xmath3o@xmath5,@xcite all of which have high value of dielectric constant @xmath0 .
high-@xmath0 metal oxides in their amorphous ( a- ) form are more preferable as a gate oxide over their crystalline form due to several important advantages they provide : ( i ) isotropic physical properties ; ( ii ) no crystalline domain boundary which leads to less defects at the interface with the si substrate ; and ( iii ) good compatibility with the conventional cmos fabrication process . some alloy structures of these high-@xmath0 metal oxides are also being studied extensively@xcite and , in fact , a hf based alloy material has already been in its third generation of production as a gate oxide in the semiconductor industry and further improvements of thermal stability , dielectric constant and material preparations are underway.@xcite in clear contrary to the abundance of experimental results in the literature , theoretical studies of the structure and dielectric properties of amorphous metal oxides and their alloys are quite limited .
one of the reasons is the difficulties in generating reasonable and reliable amorphous structures with the available theoretical methods .
experimentally , the high-@xmath0 materials are deposited on silicon substrate by vapor deposition followed by annealing processes at around @xmath6 k.@xcite simulating the deposition process and the resulting amorphous high-@xmath0 material structures by molecular dynamics ( md ) is extremely time consuming because .
the as - deposited a - hfo@xmath3 and a - al@xmath3o@xmath5 films have been generated by kinetic monte carlo ( kmc ) methods@xcite with certain assumptions such as the atoms are kept fixed on pre - determined crystal sites and less important dynamics processes of atomic motion are omitted .
both md and kmc are useful for the simulation of the initial atomic layer deposition processes of the amorphous film growth . to generate bulk amorphous structures , _ ab inito _
md simulations can be used in a melt and quench scheme , a - hfo@xmath3,@xcite a - zro@xmath3@xcite and their silicates@xcite have been successfully simulated this way . the activation - relation technique is another method to generate structures of continuous disordered systems@xcite and has been applied to generate a - zro@xmath3 structures.@xcite in this method , one optimizes the structures many times to determine a local energy minima ( in configurational space ) using a force field which can be calculated from quantum mechanical first principles or described by an empirical potential function . for
the former the calculation cost is high while for the latter a reliable potential function is required .
recently , several extended - tersoff potentials@xcite were proposed for certain metal oxides and applied to generate their amorphous structures from classical md.@xcite amorphous structures are long range disordered systems which typically requires much larger than @xmath7 atoms to simulate .
another issue , perhaps more difficult to deal with , is the simulation time scale that is needed for the atoms to evolve into the amorphous structure .
while _ ab initio _ methods have simulated up to picoseconds time scale , one certainly wishes to investigate the structures and electronic properties of metal oxides at much larger time scales and much larger sizes in order to reveal the long time limit of the structural evolution and the physical properties of the resulting amorphous material .
it is the purpose of this work to report such an investigation . generated by classical md starting from the cubic crystal form of hfo@xmath3 with different lattice constants , corresponding to different densities .
if the density is high , the system recrystallizes to its cubic form ( a ) or any other crystal form ( b ) ; if the density is low , the system tends to form disordered structure with big holes ( d ) , whereas with moderate density of @xmath8 g/@xmath9 which corresponds to lattice constant of @xmath10 for crystalline hfo@xmath3 , a reasonable amorphous structure is generated ( c ) .
blue spheres represent hf atoms , red spheres are for o atoms.,title="fig : " ] + in particular , we calculate the lattice dielectric constants of bulk amorphous metal oxides by a combined approach of classical md - for structure evolution , and quantum mechanical first principles dft - for electronic structure analysis .
we overcome the size and time scale difficulty by classical md simulations in a melt and quench scheme and generate amorphous structures of hfo@xmath3 , zro@xmath3 and their alloys in the form of hf@xmath1zr@xmath2o@xmath3 with different values of the concentration x ( @xmath11 ) .
the classical md is based on the interatomic born - mayer - buckingham potential function.@xcite this potential has been developed over many years and extensively applied to investigate the growth processes of y@xmath3o@xmath5 , zro@xmath3 and mg - al - o thin films.@xcite detailed analysis of the amorphous structures are carried out including the coordination number , the radial distribution function and the potential energy per atom of the structures .
using the atomic structures generated this way , we carry out first principles density functional theory ( dft ) calculations of the lattice dielectric constant and the results are averaged over ten independent samples for each material .
the sample averaged dielectric constants show very good agreement with the corresponding experimental data .
the combined approach of classical md and dft is an excellent way to deal with the problem of large sizes and large time scales required by the materials physics of metal oxides .
several advantages can be summarized : ( i ) classical md is inexpensive for simulating systems containing large number of atoms which is important for a proper characterization of the amorphous structures having long range disorders , here we have gone up to 768 atoms although several thousand atoms can be easily handled ; ( ii ) in the melt and quench scheme a much longer relaxation time scale , here to several nano seconds , can be evolved after the material is heated above the melting point .
such a long time scale is important for erasing the memory of the initial structure thereby the final outcome becomes independent of the initial conditions .
here we show that the same kind of amorphous structure is obtained by evolving from different initial structures including from a completely random structure .
( iii ) since the calculated physical properties of amorphous material should be averaged over an ensemble of samples due to the randomness of the structure , the combined classical - quantum approach allows such configurational average to be done . here
we show that the calculated dielectric constants of the metal oxides become spatially isotropic after configurational average , but it is typically anisotropic if only one sample is calculated .
the paper is organized as follows . in sec .
ii we present the classical md method as employed in this work . in sec .
iii we discuss the structural properties of our simulated amorphous structures , and also present our results for the lattice dielectric constants of these structures .
section iv concludes the paper .
structure obtained by the classical md simulations .
green , red and blue lines in ( b ) represent hf - o , o - o and hf - hf rdf , respectively .
the results are obtained from structures having 768 atoms.,title="fig : " ] +
in this section we discuss the potential function used in our classical md simulations along with a short description of the melt and quench scheme . in our md method
the interactions between the atoms are described by the born - mayer - buckingham potential function@xcite as shown below : @xmath12 the cation - cation interactions are assumed to be purely coulombic , the parameters for @xmath13 interactions and cation - anion interactions are presented in table .
all the parameters are obtained from ref .
we employed wolf _ et .
s method @xcite for the summation of the coulomb interaction term .
at the initial stage of the simulation , all the atoms are assigned velocities taken from a maxwell s distribution at the room temperature @xmath14 .
the system is then heated up to a temperature @xmath15 k , followed by an evolution of @xmath16 ns at the same temperature ; and then quenched back to the room temperature @xmath14 .
the temperature of the system is adjusted by resetting the velocity of a randomly selected atom at the interval of @xmath17 according to the following expression:@xcite @xmath18 where @xmath19 is a parameter between @xmath20 and @xmath21 , @xmath22 is a velocity selected randomly from the maxwell s distribution at temperature @xmath23 or @xmath14 via a random number @xmath24 .
after the system is quenched to @xmath14 , it is allowed to evolve for @xmath25 ps to calculate the radial distribution function ( rdf ) . finally , the system is cooled down to @xmath20 k by a damping method,@xcite and thus concludes the procedure for generating one amorphous structure of a particular metal oxides . in the simulations ,
the initial structure is modeled with a @xmath26 super cell of hf@xmath1zr@xmath2o@xmath3 containing 96 atoms ; or @xmath27 containing 768 atoms .
a periodic boundary condition is imposed on the supercell .
the initial atomic configuration is the cubic crystalline form or in the form of a random structure by _ randomly _ placing the constituents atoms ( with the correct ratio ) inside the simulation box .
these drastically different initial conditions verify that the md simulation time scales given above are large enough since the same kind of final amorphous structures are obtained .
ccccc elements & charge & a(ev ) & @xmath28( ) & c + hf & + 4 & 1454.6 & 0.350 & 0 + zr & + 4 & 1453.8 & 0.350 & 0 + o & -2 & 22764.3 & 0.149 & 20.4 + [ tab1 ]
to generate reliable amorphous structures it is important to find the optimal density to use for the initial structure .
hence , we first perform a series of independent simulations on the 96-atom super cell , starting from initial structures of different lattice constants ( _ i.e. _ different densities ) of cubic hfo@xmath3 and zro@xmath3 in the range from @xmath29 to @xmath30 with an interval of @xmath31 . the density of @xmath8 g/@xmath9 for hfo@xmath3 and @xmath32 g/@xmath9 for zro@xmath3 .
if we start with a lattice constant less than @xmath33 , the systems recrystallize to a cubic form ( fig.1a ) or some other crystal form ( fig.1b ) ; while with a lattice constant greater than @xmath34 , the systems tend to form disordered structures with big holes inside ( fig.1d ) .
our optimized density values are close to those obtained previously from _ ab initio _ calculations which were @xmath35 g/@xmath9 for hfo@xmath3 @xcite and @xmath36 g/@xmath9 for zro@xmath3.@xcite hence , the lattice constant of 5.6 is used as initial condition ( cubic form ) in subsequent melt and quench evolution to the amorphous structures of hf@xmath1zr@xmath2o@xmath3 . to verify the generated amorphous structures we calculate the coordination numbers and the radial distribution functions .
results are presented in fig .
[ fig2 ] for a - hfo@xmath3 obtained from the 768-atom supercell . from fig .
[ fig2]a we note that the three - coordinated o atoms and six - coordinated hf atoms dominate the a - hfo@xmath3 structures which agree with the known _ ab initio _
results.@xcite from fig .
[ fig2]b we find that the first peaks in the hf - o , o - o and hf - hf rdfs are at 2.1 , 2.8 , and 3.6 , respectively , again compare well with _
_ studies.@xcite these comparisons indicates that the classical md can reliably generate amorphous structures of the metal oxides .
importantly , the efficiency of classical md allows us to generate ensembles of amorphous samples of hf@xmath1zr@xmath2o@xmath3 for each given concentration @xmath4 , which is essential for statistical analysis of the dielectric properties ( see below ) .
zr@xmath2o@xmath3 averaged over @xmath25 samples for each value of @xmath4 ( solid red line ) .
the potential energies per atom for each individual sample are also shown in the figure ( circles ) .
the potential energies of these samples reside in the range of @xmath37 ev to @xmath38 ev , which are about 0.3 ev higher than that of the cubic hfo@xmath3 .
the very small differences in the potential energies among the samples with the same value of @xmath4 confirms the reliability of the structures generated.,title="fig : " ] + for further analysis , we calculate the potential energy per atom of the a - hf@xmath1zr@xmath2o@xmath3 structures for different values of @xmath4 ranging from @xmath20 to @xmath21 with an interval of @xmath39 . for each value of @xmath4 , we generated @xmath25 different amorphous samples by varying the initial conditions , and then calculated the potential energy per atom for each sample which are presented in fig .
the results show that the potential energies of a - hf@xmath1zr@xmath2o@xmath3 structures as generated are located in the range from @xmath37 ev to @xmath38 ev which is a reasonable and small spread .
we also generate a structure of a - hfo@xmath3 starting from an initial system of @xmath40 hf atoms and @xmath41 o atoms placed _ randomly _ in a cubic simulation box .
the calculated potential energy of this particular a - hfo@xmath3 structure comes to @xmath42 ev which is in energy range of the a - hfo@xmath3 structures that were generated from the cubic hfo@xmath3 initial conditions .
it is very satisfactory that the generated amorphous structures are not dependent on the initial condition , indicating that presented above are adequate . ) for the amorphous structures of hf@xmath1zr@xmath2o@xmath3 .
all the black symbols are for the values in each direction of @xmath43 , @xmath44 and @xmath45 averaged over @xmath25 samples , while the square symbols connected with a solid red line represent the values which are averaged over all the three directions for each value of @xmath4 . , title="fig : " ] + a most important property of the high-@xmath0 metal oxides is their dielectric constant . in this section
we present our calculations of the lattice dielectric constant ( static ) of hf@xmath1zr@xmath2o@xmath3 amorphous structures generated in the last section .
we found that the electronic contribution to the dielectric constant is substantially smaller than the lattice contribution and does not change very much for different structures , hence it will not be discussed until when we calculate the total dielectric constant to compare with the measured data ( see below ) .
ccccc oxide & amorphous & cubic & monoclinic & tetragonal + [ 0.5ex ] hfo@xmath3 & 24.52 & 31.08 & 14.49 & 125.95 + zro@xmath3 & 35.76 & 33.80 & 16.14 & 26.72 + [ tab2 ] for each amorphous structure the dielectric constant is calculated for an ensemble of @xmath25 different samples ( the 96-atom super cell ) .
the sample averaged values are presented in table [ tab2 ] for a - hfo@xmath3 and a - zro@xmath3 along with the dielectric constants for all @xmath46 of their crystal forms .
we notice that the lattice dielectric constants of a - hfo@xmath3 and a - zro@xmath3 are close to that of their cubic form , suggesting that some cubic - like short range order content may dominate the amorphous structures .
indeed , recently experimental measurements@xcite on sputtering samples of a - hfo@xmath3 discovered a cubic - like short range coordination in the amorphous microstructure which led to high measured dielectric constant @xmath47 .
our calculated lattice @xmath0 ( see table [ tab2 ] ) is 24.52 , adding the electronic contribution which we found to be around 4.4 , the total @xmath48 which agrees quite well with the measured data . to confirm that the @xmath0 of a - hfo@xmath3 being close to that of the cubic hfo@xmath3 , is not due to the cubic initial condition of the structure simulation
, we have evolved a single a - hfo@xmath3 structure from a completely random initial condition ( see previous section ) : its lattice dielectric constant is found to be 21.785 , 25.43 and 25.163 in the three spatial directions respectively ; its electronic contribution is found to be 4.45 .
these values are totally consistent with that obtained from the cubic initial conditions . in comparison to previous theoretical first principles
investigations,@xcite our @xmath0 values are higher .
this difference may be explained by noting that in ref .
[ ] the amorphous structures were obtained from a relaxed monoclinic structure with cubic lattice vectors which tends to have lower @xmath0 values .
we should mention that direct comparison with experimental measurement is often difficult due to scarcity of data for dielectric constants of amorphous bulk structures .
the ensemble averaged lattice dielectric constants of all the amorphous structures generated for different values of the concentration @xmath4 are presented in fig .
we observe a gradual increase ( solid red line ) in the values of dielectric constant from a - hfo@xmath3 , through hf@xmath1zr@xmath2o@xmath3 , to a - zro@xmath3 .
we also note that the spatial resolved dielectric constants in @xmath43 , @xmath44 and @xmath45 directions ( represented by the black symbols in fig .
[ fig4 ] ) are quite _ isotropic _ which is expected for amorphous structures .
it is very important to note that an isotropic dielectric constant is obtained only when the values are averaged over the ensemble of samples of the same structure ( same @xmath4 ) . for a single sample
, we always observe large spreads in the values of dielectric constant in different directions .
this finding underpins the major advantage of employing a rather inexpensive classical md to efficiently generate amorphous structures for the statistical average of the physical properties .
using classical md simulations within a melt and quench scheme , the amorphous structures of hfo@xmath3 and zro@xmath3 , along with their alloy structures of the form of hf@xmath1zr@xmath2o@xmath3 , have been generated .
a suitable potential function , namely the born - mayer - buckingham potential , has been adopted for these simulations .
the calculated coordination numbers and radial distribution functions of the generated structures agree well with the previous studies .
the values of the ensemble averaged lattice dielectric constants of these amorphous structures are nearly isotropic , close to that of their cubic form , and compare well with available experimental data.@xcite on the other hand , a single sample always produces anisotropic @xmath0 values .
our results indicate that the combined approach of classical md - for structure evolution , and quantum mechanical first principles dft - for electronic structure analysis , is a reliable technique for investigating high-@xmath0 metal oxides .
we would like to thank h.z .
shao of fudan university , c. yin of dec and j.n .
zhuang of the university of hong kong for useful discussions .
this work is supported by the university grant council ( contract no .
aoe / p-04/08 ) of the government of hksar , the small project funding of hku ( yw ) , nserc of canada ( hg ) , and cifar ( hg ) .
x. zhao and d. vanderbilt , phys .
b * 65 * , 233106 ( 2002 ) . d. ceresoli and d. vanderbilt , phys . rev .
b * 74 * , 125108 ( 2006 ) .
x. zhao and d. vanderbilt , phys .
b * 65 * , 075105 ( 2002 ) .
x. zhao , d. ceresoli , and d. vanderbilt , phys .
b * 71 * , 085107 ( 2005 ) .
g. mazaleyrat , a. estve , l. jeloaica , m.d .
rouhani , comput .
, * 33 * , 74 ( 2005 ) .
scopel , a.j.r .
da silva , and a. fazzio , phys .
b * 77 * , 172101 ( 2008 )
zhu , t. tamagawa , m. gibson , t. furukawa , and t.p .
ma , ieee electron device lett . *
23 * , 649 ( 2002 ) .
m.y . ho , h. gong , g.d .
wilk , b.w .
busch , m.l .
green , w.h .
lin , a. see , s.k .
lahiri , m.e .
loomans , p.i .
risne , and t. gustafsson , appl . phys
. lett . * 81 * , 4218 ( 2002 ) .
y. xiong , h. tu , j. du , m. ji , x. zhang , l. wang , appl . phys .
* 97 * , 012901 ( 2010 ) .
h. momida , t. hamada , y. takagi , t. yamamoto , t. uda , and t. ohno , phys .
b * 75 * , 195105 ( 2007 ) ; h. momida , t. hamada , and t. ohno , jpn . j. appl
46 * , 3255 ( 2007 ) .
p. broqvist and a. pasquarello , microelectron
. eng . * 84 * , 2416 ( 2007 ) ; p. broqvist and a. pasquarello , appl .
. lett . * 90 * , 082907 ( 2007 ) .
perevalov , v.a .
gritsenko , s.b .
erenburg , a.m. badalyan , h. wong , and c.w .
kim , j. appl .
* 101 * , 053704 ( 2007 ) .
bayer , s. hofmann , j.d .
dutson , s.j .
wakeham , m.j .
thwaites , w.i .
milne , and a.j .
flewitt , appl .
lett . * 98 * , 252903 ( 2011 ) .
a. dkhissi , a. estve .
c. mastail , s. olivier , g. mazaleyrat , l. jeloaica , and m.d .
rouhani , j. chem .
comput . , * 4 * , 1915 ( 2008 ) .
chen , z.f .
hou , and x.g .
gong , appl .
. lett . * 95 * , 102905 ( 2009 ) .
d. vanderbilt , x. zhao , d. ceresoli , thin solid films * 486 * , 125 ( 2005 ) .
barkema , n. mousseau , phys .
lett . * 77 * , 4358 ( 1996 ) ; n. mousseau , g.t .
barkema , phys .
b * 61 * , 1898 ( 2000 )
. t.p .
perumal , v. sridhar , k.p.n .
murthy , k.s .
easwarakumar , s. ramasamy , phys .
a , * 309 * , 35 ( 2002 ) . t. arima , k. fukuyo , k. idemitsu , y. inagaki , j. mol . liq . * 113 * , 67 ( 2004 ) . m. kilo , m.a .
taylor , c. argirusis , g. borchardt , r.a .
jackson , o. schulz , m. martin , m. weller , solid state ionics * 175 * , 823 ( 2004 ) .
v. georgieva , i.t .
todorov , a. bogaerts , chem .
phys . lett . *
485 * , 315 ( 2010 ) .
v. georgieva , m. saraiva , n. jehanathan , o.i .
levelev , d. depla , and a. bogaerts , j. phys .
phys . , * 42 * , 065107 ( 2009 ) . | high-@xmath0 metal oxides are a class of materials playing an increasingly important role in modern device physics and technology . here
we report theoretical investigations of the properties of structural and lattice dielectric constants of bulk amorphous metal oxides by a combined approach of classical molecular dynamics ( md ) - for structure evolution , and quantum mechanical first principles density function theory ( dft ) - for electronic structure analysis . using classical md based on the born - mayer - buckingham potential function within a melt and
quench scheme , amorphous structures of high-@xmath0 metal oxides hf@xmath1zr@xmath2o@xmath3 with different values of the concentration @xmath4 , are generated .
the coordination numbers and the radial distribution functions of the structures are in good agreement with the corresponding experimental data .
we then calculate the lattice dielectric constants of the materials from quantum mechanical first principles , and the values averaged over an ensemble of samples agree well with the available experimental data , and are very close to the dielectric constants of their cubic form . |
schur s q - functions are very interesting analogues of the ( standard ) schur functions @xmath3 in several combinatorial and representation - theoretic contexts
. examples of their analogy include the shifted rsk correspondence , the shifted littlewood - richardson rule , and the character theory of representations of queer lie superalgebras . in this note
we study when certain shifted littlewood - richardson coefficients ( the @xmath4 in the language of @xcite ) are zero or one . studying questions of the same nature
has been of interest to a number of authors . in particular , one should mention stembridge s recent classification of multiplicity - free products of schur functions @xcite which was generalized to p - functions in @xcite .
the related questions as to when two ribbon schur functions are equal and when a schur function is equal to a skew schur function were answered in @xcite and @xcite . here
we show that the latter problem has a simple answer for schur s q - functions as well .
+ this note is organized as follows . in the next section
we give all the required definitions . in the third section
we prove our main result .
+ * acknowledgement . *
the author thanks the referee for reading the paper very carefully .
our notation is compatible with stembridge s paper @xcite . a strictly decreasing sequence @xmath5 of positive integers
is called a _ distinct partition _ of @xmath6 if the sum of the @xmath7 s is equal to @xmath6 .
the @xmath7 s are called the _ parts _ of @xmath8 . a partition is represented by a shifted young diagram as follows : there are @xmath7 boxes in the @xmath9-th row , after @xmath10 empty positions .
we denote this diagram by @xmath11 .
[ one ] let @xmath12 .
then the diagram @xmath11 which represents @xmath8 is given below .
consider an ordered alphabet @xmath14 the letters @xmath15 will be referred to as _ marked _ , whereas the letters @xmath16 will be referred to as _
unmarked_. [ gsyt ] by a generalized shifted young tableau ( gsyt ) of shape @xmath11 we mean a filling of a given young diagram @xmath11 with letters from @xmath17 such that the following properties hold : 1 . the rows and columns are weakly increasing .
2 . each row contains each marked letter at most once .
each column contains each unmarked letter at most once .
let @xmath18 be a given gsyt .
we define @xmath19 to be the monomial @xmath20 where @xmath21 is the total number of occurrences of @xmath22 or @xmath23 in @xmath18 .
schur s q - function @xmath1 is equal to @xmath24 where the sum is over all gsyt of shape @xmath11 .
the sequence @xmath25 is called the _ content _ of @xmath18 . +
now suppose @xmath8 and @xmath26 are two distinct partitions .
assume that for each @xmath9 , @xmath27 , so that @xmath28 lies inside @xmath11 .
the _ shifted skew diagram _ associated to @xmath29 is the set - theoretic difference of @xmath11 and @xmath28 .
it is represented by @xmath30 .
[ two]let @xmath8 be as in example [ one ] and let @xmath31 .
then @xmath30 is represented by the following skew young diagram : @xmath32 we may assume that @xmath33 is placed on the cartesian plane in the usual way such that the centers of boxes lie on the lattice of points with integer coordinates .
let @xmath34 denote the box whose center is at the point with coordinates @xmath35 .
the following easy lemma includes some basic properties of shifted skew diagrams .
[ propd ] let @xmath33 be an arbitrary shifted skew diagram on the cartesian plane . for any integer @xmath36 define @xmath37 1 .
if @xmath38 and @xmath39 then there exists an integer @xmath40 such that @xmath41 .
2 . for a fixed @xmath42 such that @xmath43 , let @xmath44 be the smallest number for which @xmath45 .
assume @xmath46 . then for any @xmath47 such that @xmath48 , the following statement is true : * @xmath49 and for any @xmath22 if @xmath50 then @xmath51 .
schur s skew q - function @xmath52 is equal to a summation similar to ( [ qschur ] ) , where the summation is now on all shifted skew tableaux @xmath18 , with underlying diagram @xmath33 filled by the alphabet @xmath17 , such that they satisfy the properties of definition [ gsyt ] .
the function @xmath52 can be expressed as a linear combination of the functions @xmath53 for various @xmath54 as follows .
we have @xmath55 where the summation is over all distinct partitions @xmath54 . here
@xmath56 is the number of _ amenable _ tableaux of shape @xmath30 and content @xmath54 .
we define the amenable tableaux in definition [ amenable ] below . however , before doing so , we need some notation . for a given ( possibly skew )
gsyt such as @xmath18 , the row word of @xmath18 is the word obtained by reading the rows of @xmath18 consecutively from left to right starting with the bottom row .
we denote the row word of @xmath18 by @xmath57 .
now , let @xmath58 be an arbitrary word of length @xmath6 such that for any @xmath22 we have @xmath59 .
define a function @xmath60 as follows .
@xmath61 by convention , we assume @xmath62 for any @xmath63 .
[ amenable ] let @xmath64 be an integer .
a word @xmath58 is called @xmath65-amenable iff it satisfies the following properties : 1 . for any @xmath66 if @xmath67 then @xmath68 .
2 . for any @xmath69 ,
if @xmath67 then @xmath70 .
3 . if @xmath71 is the smallest number such that @xmath72 , then @xmath73 .
if @xmath71 is the smallest number such that @xmath74 , then @xmath75 .
a word @xmath76 is called amenable if it is @xmath65-amenable for any @xmath64
. * remark . *
suppose @xmath58 is @xmath65-amenable for some @xmath64 .
then it follows from definition [ amenable ] that if @xmath77 then @xmath78 .
( to prove this , first we show that definition [ amenable ] implies @xmath79 for any @xmath80 .
then we pick the largest @xmath71 such that @xmath81 , and we show that definition [ amenable ] implies that there must exist a @xmath82 such that @xmath83 .
the details of the argument are left to the reader . ) consequently , if @xmath58 is amenable , then @xmath84 a given gsyt is called amenable iff its row word is amenable .
in this section we prove the main statement of this note . * @xmath87 is arbitrary and @xmath88 . *
@xmath89 and @xmath90 * @xmath91 and + @xmath92 where @xmath93 are integers such that @xmath94 , @xmath95 . * @xmath96 and @xmath97 where @xmath93 are integers such that @xmath98 and @xmath99 . * remark .
* the reader should note that there are overlaps among the cases for special values of @xmath93 . moreover , whether or not @xmath33 is strange only depends on @xmath29 and not on @xmath8 and @xmath26 individually .
theorem [ main ] identifies @xmath33 by identifying all possible differences @xmath29 , but not all possibilities of @xmath8 and @xmath26 .
the latter problem is not hard once we have theorem [ main ] .
+ the proof of theorem [ main ] will be given throughout this section . from equation ( [ skewqr ] ) it follows that @xmath33 is strange if and only if there exists a unique amenable filling of @xmath30 .
our approach is to rule out various possibilities for the shape of @xmath30 by demonstrating the existence of at least two different amenable fillings in each case .
+ we give a simple algorithm to construct an amenable tableau of any given shape @xmath30 .
the output of the algorithm is an amenable tableau of content @xmath54 for some distinct partition @xmath54 .
note that @xmath54 is generated by the algorithm and is not an input . + * notation .
* let @xmath34 be a box in a diagram .
the operation of filling ( or replacing the current entry of ) @xmath34 by the letter @xmath100 is represented by @xmath101 .
the operation of filling ( or replacing the entry of ) @xmath34 by the current entry of @xmath102 is represented by @xmath103 .
+ the algorithm is given below .
set @xmath104 and @xmath105 .
2 . set @xmath106 .
3 . put a @xmath65 or a @xmath107 in any of the boxes in @xmath108 according to the following rule : @xmath109 4 .
remove all boxes of @xmath108 from @xmath110 .
let @xmath111 be the diagram obtained after removing boxes .
if @xmath111 has no boxes , then stop .
increase @xmath65 by one .
go back to step 2 .
+ let @xmath65 be a positive integer .
then each connected component of @xmath108 forms a `` path '' of boxes in a connected component of the diagram @xmath110 .
( a connected component of @xmath110 ( or @xmath108 ) is a maximal subset of boxes of @xmath110 ( or @xmath108 ) which can be ordered in a sequence so that each box has a common edge with at least one of the boxes preceding it in the sequence . )
each of these paths can be directed as follows .
let @xmath112 be a connected component of @xmath108 .
we know that @xmath112 is a path of boxes .
the first box of this path is the box @xmath113 such that @xmath114 and @xmath115 .
the last box of @xmath112 is the box @xmath113 such that @xmath116 and @xmath117 .
[ checktwo ] let @xmath18 be a gsyt of shape @xmath33 and let @xmath128 be the row word of @xmath18 . for any integer @xmath64 , form the word @xmath129 from @xmath76 by dropping all the letters in @xmath76 which are not in @xmath130 .
( for example , if @xmath131 then @xmath132 . )
then @xmath76 is amenable if and only if @xmath129 is either @xmath65-amenable or empty for any @xmath64 . properties of definition [ gsyt ] are satisfied trivially .
it remains to show that this filling is amenable . by lemma [ checktwo ]
it suffices to check @xmath65-amenability of @xmath133 for any @xmath65 .
it is easily seen that if @xmath33 is disconnected but each connected component of @xmath18 is amenable , then @xmath18 is amenable too .
therefore we can assume @xmath33 is connected .
next we show that @xmath134 is amenable .
the argument in the general case is similar .
let the length of @xmath134 be @xmath6 .
amenability of @xmath134 follows from the following facts : 1 . for any box filled by a @xmath135 ,
there is a box above it in the same column which is filled by a @xmath136 .
this implies that @xmath137 for any @xmath138 .
2 . since the last entry of the path @xmath139
is filled by a @xmath136 and there is no @xmath135 below it , we have @xmath140 .
3 . there is exactly one @xmath141 in every row of @xmath33 in which @xmath139 has boxes , except for the lowest row among them .
the same statement holds for @xmath142 and @xmath143 .
4 . set @xmath144 .
then @xmath145 for integers @xmath146 .
moreover @xmath147 .
a slight modification of the algorithm of section [ algorithm ] can be applied to show that @xmath33 is not a strange diagram whenever there exist disconnected @xmath108 s .
more accurately , we have the following lemma .
[ discon ] let @xmath18 be a gsyt obtained by applying the algorithm of section [ algorithm ] to @xmath33 .
suppose that for some @xmath65 , @xmath108 is disconnected .
then @xmath33 has an amenable filling different from the one given by the algorithm .
let @xmath148 , @xmath149 , where @xmath150 s are mutually disjoint paths , ordered such that for any @xmath9 , @xmath150 lies to the northeast of @xmath151 .
note that the @xmath150 s belong to mutually disjoint connected components of the diagram @xmath110 .
since the filling of @xmath108 is obtained by the algorithm , the last box of @xmath152 is filled by a @xmath65 .
we claim that if we change it to a @xmath107 , the new filling is still amenable . to this end
we use lemma [ checktwo ] .
one can see that changing a @xmath65 to a @xmath107 does not affect @xmath65-amenability of @xmath133 unless we are changing the last box of @xmath153 .
moreover @xmath154 remains @xmath155-amenable since @xmath154 can be written as @xmath156 where @xmath157 is the row word of the intersection of @xmath158 with the connected component of @xmath110 which contains @xmath159 . a bookkeeping argument using the facts that the words @xmath160 are @xmath155-amenable and @xmath161 is `` not far from ''
being @xmath155-amenable completes the proof . from lemma [ discon ]
we conclude that if @xmath33 is a strange diagram then for any @xmath65 the set @xmath108 which is obtained by the algorithm of section [ algorithm ] is in fact a ( connected ) path .
this fact simplifies the case - by - case analysis of possible shapes of @xmath33 .
probably a few remarks are necessary before starting the next section : throughout the next section , at several points we give procedures to modify an existing filling and claim that the new filling will be amenable too .
the proofs of amenability of these new fillings are very similar in nature . however , most of these amenability proofs are not given because they are tedious but very simple .
the main ideas are using lemma [ checktwo ] and bookkeeping arguments . throughout the next section
there are several illustrations which help the reader understand the effect of procedures on the fillings .
the figures usually demonstrate a union of @xmath162 for some @xmath163 . only the boxes whose entries change by the procedures are shown .
the dots show that there may or may not be boxes in the direction of the dots .
see the following example .
let @xmath164 denote the largest integer such that @xmath165 .
suppose @xmath166 and the union @xmath167 is illustrated by the diagram @xmath168 this means that there may or may not be a vertical path of boxes which belong to @xmath169 in the location of the dots .
for instance @xmath167 may actually be @xmath170 now again suppose @xmath166 and the union @xmath167 is represented by @xmath171 this means that if the path @xmath169 extends from the top right corner of @xmath167 , then its direction will be vertical in the beginning .
( however , its direction does not have to remain vertical all the time . ) for instance , @xmath167 may actually be @xmath172 again let @xmath166 and suppose @xmath167 is illustrated by the diagram @xmath173 this means that @xmath169 may extend either horizontally or vertically ( but clearly not both , since @xmath169 is a path ) .
let @xmath18 be the filling of @xmath33 obtained by the algorithm of section [ algorithm ] .
by lemma [ discon ] , @xmath108 is a connected path for any @xmath65 .
let @xmath164 be the largest integer for which @xmath174 is nonempty .
then the following possibilities exist for @xmath174 : * @xmath174 consists of a single box .
* @xmath174 lies within a single row and has at least two boxes .
* @xmath174 lies within a single column and has at least two boxes .
* @xmath174 is none of the above .
the following figures show how to obtain a new amenable filling of @xmath33 when @xmath174 has a `` turning point '' .
if the path @xmath174 changes its direction at some point clockwise , i.e. if for integers @xmath175 we have @xmath176 then all we need to do is @xmath177 if it changes its direction counterclockwise , i.e. if @xmath178 , then we should find the smallest @xmath179 such that @xmath180 and then do @xmath181 and @xmath182 .
we will assume @xmath33 is a strange diagram and @xmath186 .
let @xmath34 be the box in @xmath174 .
then by lemma [ propt ] and lemma [ paths ] , @xmath187 contains @xmath188 and @xmath189 .
now we have the following two possibilities : 1 .
then it follows that @xmath191 should be the last box of @xmath192 , as @xmath192 can not proceed to @xmath193 by lemma [ propd ] .
this in turn yields three new possibilities : * @xmath194 . in this case
@xmath192 has two boxes only .
* @xmath195 . * @xmath196 .
2 . @xmath197 .
then @xmath191 should be filled by @xmath198 and @xmath199 .
one of the following six cases may take place .
* @xmath200 .
* @xmath201 .
* @xmath202 but @xmath203 . * @xmath204 but @xmath196 . * @xmath204 but @xmath195 . * @xmath205 . + assuming
@xmath166 we can illustrate @xmath206 in cases a to i by the following figures .
@xmath207 note that in the cases d , e , f , g and h given above we can modify the filling to get another amenable one as shown in the following figures .
therefore if @xmath206 is one of those cases then @xmath33 can not be strange . @xmath208 the only remaining possibility is case i. if @xmath209 then @xmath33 is equal to the underlying diagram of figure i , and has a unique amenable filling .
assume @xmath210 . next we prove proposition [ casei ] .
[ casei ] suppose @xmath33 is a strange diagram and @xmath216 is as in case i. then for any @xmath71 , @xmath217 is a @xmath218-hook for some @xmath219 . moreover , if for some @xmath220 we have @xmath221 , then @xmath222 .
we use backwards induction . assume the statement holds for @xmath223 where @xmath224 .
we prove it for @xmath225 .
suppose that @xmath226 is an @xmath227-hook for some @xmath228 .
as before , assume @xmath229 .
lemma [ discon ] , lemma [ paths ] and lemma [ propd ] imply that * @xmath235
. then @xmath236 should be the last box of @xmath225 .
we will show that @xmath237 .
suppose , on the contrary , that @xmath238 . then either @xmath239 or @xmath240 .
however , @xmath33 can not be strange in either of the cases , as shown below . * * @xmath240 .
let @xmath241 .
then the following procedure provides a new filling . * * * set @xmath242 and @xmath243 . * * * if @xmath244 then stop . * * * find smallest @xmath245 such that @xmath246 .
if there is no such @xmath179 , then stop . * * * if @xmath247 then do @xmath248 and @xmath249 . *
* * increase @xmath250 by one .
go back to step 2 .
+ the idea behind the procedure can be more concretely described as follows .
we eliminate the last box of @xmath225 , add this box to @xmath251 , and find new paths @xmath252 according to the algorithm of section [ algorithm ] .
+ the following figure demonstrates the above procedure when @xmath253 and @xmath254 .
@xmath255 * * @xmath239 .
then the following procedure gives the new amenable filling .
* * * set @xmath256 .
* * * if @xmath257 then replace the entry of @xmath258 by that of @xmath259 . if @xmath242 then replace the entry of @xmath258 by @xmath260 .
if @xmath261 then stop .
* * * decrease @xmath250 by one .
go back to step 2 .
+ + the amenability of the diagram obtained by the procedure follows from statements similar to those that appear in the proof of lemma [ amenamen ] .
+ for @xmath166 and @xmath262 the following figure demonstrates the effect of the procedure . @xmath263 * @xmath264 . in this case
if @xmath265 then we can argue as in the case @xmath235 to show that @xmath237 .
the following figure illustrates what happens when @xmath166 , @xmath262 and @xmath240 .
@xmath266 + therefore we may assume @xmath267 .
next we show that @xmath268 suppose on the contrary that @xmath269 .
then one of the following cases occurs . in all of them
we show that @xmath33 is not strange . * * @xmath270 .
we can get a new amenable filling by the operations @xmath271 and @xmath272 . * * @xmath273 . in this case
we can obtain a new amenable filling as follows .
find the smallest @xmath179 such that @xmath274 . then do @xmath275 , @xmath276 and @xmath277 .
* * @xmath278 , @xmath279 but @xmath240 .
the new amenable filling is obtained by @xmath280 the following figure illustrates this procedure when @xmath281 . @xmath282 * * @xmath278 , @xmath279 but @xmath239 . in this case
exactly the same procedure that was given in the analysis of case a2 can provide a new filling .
we illustrate it below with @xmath166 and @xmath283 .
@xmath284 * remark .
* the two cases of corollary [ caseicor ] may overlap . moreover , not all of them are such that @xmath206 is as in case i. however , we have written it in the given form above so that we can refer to it at other points of this manuscript .
for example see corollary [ iicor ] .
it suffices to show that there does not exist any @xmath287 such that @xmath288 .
assume the contrary , and let @xmath287 be the smallest integer such that @xmath288 .
let @xmath163 be such that @xmath289 . from lemma [ propd ] , lemma [ propt ] and lemma [ discon ]
it follows that @xmath290 .
moreover , by lemma [ paths ] it follows that for any @xmath22 such that @xmath291 , @xmath292 taking @xmath293 implies that @xmath294 and the path @xmath295 makes a clockwise turn at some point .
let @xmath296 be such that @xmath297 and @xmath298 .
for example we can take @xmath299 .
it follows from lemma [ paths ] that for any @xmath300 such that @xmath301 , @xmath302 now the following procedure provides a new amenable filling .
* set @xmath303 . * if @xmath304 then do @xmath305 and @xmath306 . if @xmath307 and the entry of @xmath308 is either @xmath76 or @xmath309 where @xmath310 , then do @xmath311 . if @xmath312 then stop .
* increase @xmath250 by one .
go back to step 2 .
the idea of the procedure is more concretely explained as follows .
we eliminate one of the boxes of @xmath295 , add it to @xmath313 , and then find new paths @xmath314 according to the algorithm of section [ algorithm ] .
the following figure demonstrates how the procedure can be applied to a case with @xmath166 and @xmath283 .
in fact it can be seen that there are at least two ways to get new amenable fillings .
@xmath315 let , as before , @xmath229 .
suppose @xmath33 is not given by partitions of proposition [ caseb ] . according to the proof of proposition [ caseb ] ,
this means that there exists a @xmath250 such that @xmath317 .
we can assume @xmath250 is the smallest integer with this property .
let @xmath163 be such that @xmath318 .
let @xmath319 be the largest integer for which @xmath320 does not lie within a single row .
note that if there is no such @xmath319 then it follows that @xmath321 .
set @xmath322 .
an argument similar to that of proposition [ caseb ] shows that @xmath323 should make a clockwise turn ; i.e. there exist @xmath324 such that @xmath325 .
now one of the following possibilities happens .
* @xmath326 .
then it follows that @xmath327 . in this case
we use the following procedure .
1 . set @xmath303 .
2 . if @xmath328 then do @xmath329 . otherwise , do @xmath330 and stop .
3 . decrease @xmath250 by one .
go back to step 2 .
+ + the idea of the procedure is similar to that of the procedure given in case b4 of proposition [ casei ] .
* @xmath331 .
then it follows that @xmath332 . in this case
we use the following procedure . * * set @xmath242 and do @xmath333 . * * if @xmath334 , then do @xmath335 and @xmath336 . * * if @xmath337 and @xmath338 then do @xmath335 and stop . * * if @xmath339 then do @xmath340 and stop .
* * increase @xmath250 by one .
go back to step 2 .
+ the idea behind the procedure is to eliminate one of the boxes of @xmath323 , modify the filling of @xmath323 slightly , attach the box removed from @xmath323 to @xmath341 and find new paths @xmath342 according to the algorithm of section [ algorithm ] .
[ [ case - ii - p_m - lies - within - a - single - row - and - has - at - least - two - boxes ] ] case ii : @xmath174 lies within a single row and has at least two boxes ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ @xmath347 . in this case the boxes @xmath348 and @xmath349 form a path in @xmath192 .
an argument similar to the one given in case b of proposition [ casei ] proves that there are no other boxes in @xmath192 .
next we state the following proposition which is similar in statement and proof to proposition [ casei ] .
[ case1row ] suppose that @xmath33 is a strange diagram such that @xmath206 is as in case d1 .
then for any @xmath350 , @xmath217 is a @xmath351-hook for some @xmath219 .
moreover , if there exists a @xmath220 such that @xmath221 , then @xmath222 .
it suffices to show that for any @xmath71 , @xmath217 lies within a single row .
the proof is by contradiction .
assume the contrary , and let @xmath71 be the largest integer such that @xmath217 does not lie within one row .
from lemma [ propt ] it follows that if @xmath353 is the largest integer for which there exists a box @xmath354 then there exists a @xmath355 such that @xmath356 , where @xmath357 .
now set @xmath358 and apply the procedure given in case c1 of the proof of proposition [ casea ] to get a new amenable filling .
[ [ case - iii - p_m - lies - within - a - single - column - and - has - at - least - two - boxes ] ] case iii : @xmath174 lies within a single column and has at least two boxes ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ the analysis in this case is pretty similar to the previous cases .
suppose @xmath174 consists of the boxes @xmath359 where @xmath344 .
then for any @xmath360 , @xmath361 .
moreover , @xmath362 .
next we show that none of the boxes @xmath363 can belong to @xmath192 .
in fact in each of the following cases we show that @xmath33 can not be a strange diagram .
. then we get a new filling as follows .
find the smallest @xmath179 such that @xmath364 and do @xmath365 * @xmath200 .
then we obtain a new filling by @xmath366 * @xmath367 but @xmath368 .
then a new filling can be obtained as in case a1 or case b3 of proposition [ casei ] .
we leave the details to the reader .
it suffices to show that there does not exist any @xmath179 such that @xmath372 .
suppose there exists such a @xmath179 , and without loss of generality assume that @xmath179 is the smallest integer with this property .
suppose @xmath373 . then by lemmas [ paths ] and
[ propt ] it follows that @xmath374 now the following procedure provides a new amenable filling . in the previous sections we showed that the only diagrams which could possibly be strange are those listed in theorem [ main ] . in this section
we prove that all of those diagrams are indeed strange .
first we need a simple property of any arbitrary gsyt . from definition [ amenable ]
it follows that for any @xmath64 , @xmath385 .
suppose @xmath386 .
then it follows from definition [ amenable ] that if @xmath387 ( respectively @xmath388 ) is the smallest integer such that @xmath389 ( respectively @xmath390 ) then @xmath391 .
the fourth part of definition [ amenable ] implies that @xmath392 for any @xmath393 . since @xmath79 for any @xmath394
, it follows that @xmath395 .
but then @xmath390 , which contradicts the second part of definition [ amenable ] .
it is obvious that if @xmath396 then @xmath33 is strange .
next we prove that if @xmath33 is given by partitions of corollary [ caseicor ] then @xmath33 is a strange diagram .
we will present the argument only for the case @xmath397 and @xmath398 where @xmath93 are integers such that @xmath94 , @xmath95 .
the other case can be treated in a very similar fashion .
we give the proof of proposition [ semirec0 ] through lemma [ semirec1 ] , lemma [ semirec2 ] and lemma [ semirec3 ] .
let us assume that @xmath34 is the top right box of the diagram @xmath33 .
let @xmath6 denote the number of boxes of @xmath33 .
we consider an arbitrary amenable filling of @xmath33 and show that it has to be the one obtained by the algorithm of section [ algorithm ] .
we prove this lemma by induction on @xmath65 . for @xmath403 ,
the above statement follows from the fact that we should have @xmath404 and moreover no two @xmath141 s can lie within the same row .
next assume @xmath405 .
let @xmath406 be chosen as above .
then by induction hypothesis , the boxes @xmath407 are filled by @xmath65 . moreover
, none of the rows which lie above the box @xmath408 can contain an element of @xmath409 which is strictly larger than @xmath65 , because their rightmost boxes are filled by elements less than or equal to @xmath65 .
consequently , @xmath408 should be filled by an element of @xmath410 .
( note that @xmath411 contains a @xmath65 . )
this in turn limits the entries of @xmath412 and @xmath413 to @xmath410 .
however , if @xmath412 is filled by a @xmath414 then @xmath413 should also be filled by a @xmath414 , which is a contradiction by definition [ gsyt ] .
therefore @xmath412 has to be filled by a @xmath155 .
lemma [ semirec1 ] implies that for any @xmath405 , if there is at least one box in @xmath33 filled by a @xmath65 , then there are at least @xmath415 boxes in @xmath33 which are filled by a @xmath65 .
note that since @xmath33 has @xmath416 columns , there can be at most @xmath416 boxes which are filled by a @xmath136 .
lemma [ strict ] implies that there can be at most @xmath417 boxes which are filled by a @xmath65 .
we will see below that actually the latter possibility takes place .
more accurately , we show that if there is at least one @xmath65 in the filling of @xmath33 , then there are exactly @xmath417 of them . if @xmath420 then by lemma [ semirec1 ] the box @xmath421 is filled by a @xmath422 . but
@xmath421 is the rightmost box of the lowest row of @xmath33 , and since the rows and columns of a filling are weakly increasing , all of the boxes have to be filled by elements of @xmath409 which are less than or equal to @xmath422 .
now lemma [ semirec1 ] implies that @xmath423 contains a @xmath422 and @xmath424 contains a @xmath425 . because of the shape of @xmath33 we can see that @xmath426 . since @xmath424 contains a @xmath425 , the entry of @xmath427 should belong to @xmath428 . but the entry of @xmath427 can not be @xmath429 because this implies that the leftmost occurence of an element of @xmath428 in the row word of the filling is marked .
consequently , @xmath427 is filled by a @xmath422 which implies that @xmath430 .
now we have @xmath431 which implies that @xmath432 for any @xmath433 .
if @xmath434 , then lemma [ semirec1 ] implies that @xmath435 has to be filled by @xmath436 . now @xmath437 and therefore it should be filled by an element of @xmath409 which is larger than @xmath436 .
consequently , @xmath438 .
now by lemma [ strict ] , we have @xmath439 which implies that @xmath440 . in particular , @xmath441 if and only if @xmath442 .
let @xmath443 .
by lemma [ semirec2 ] , for any @xmath444 there are exactly @xmath417 boxes which contain a @xmath65 .
lemma [ semirec1 ] determines the position of @xmath415 of these boxes uniquely .
therefore for any @xmath445 , the location of precisely one box containing a @xmath65 is left to be determined .
in fact it turns out that after the location of this box is determined , the entire filling is also determined uniquely .
see lemma [ semirec3 ] below .
let @xmath454 be the entry of the box @xmath102 of the statement of the lemma .
since the rows and columns are weakly increasing we have @xmath455 where the inequalities are interpreted in the ordering of @xmath409 .
from definition [ amenable ] it follows that @xmath456 should be unmarked for any @xmath457 .
in fact assume on the contrary that @xmath458 for some @xmath459 .
let @xmath460 and @xmath179 be such that @xmath456 be the entry of the box @xmath102
. by definition [ amenable ] there should exist a box @xmath461 such that @xmath462 and @xmath461 is filled by an @xmath459 . from the shape of @xmath33 it follows that @xmath463 .
but then @xmath464 and it is impossible to fill the latter box such that definition [ gsyt ] holds .
the rest of this section is devoted to the proof of proposition [ llll ] .
suppose @xmath33 has @xmath212 columns and @xmath6 boxes .
this means that @xmath469 and @xmath470 .
let @xmath128 be the row word of the filling of @xmath33 .
@xmath76 can be expressed as @xmath58 where each @xmath471 . since each column of @xmath33 contains at most one box filled by @xmath136 , @xmath472 .
lemma [ strict ] implies that @xmath473 let @xmath34 denote the box in the lowest row of @xmath33 .
( note that there is only one box in the lowest row . ) for any @xmath474 let @xmath475 denote the entry of the box @xmath476 .
since rows and columns are weakly increasing , the @xmath475 s form a decreasing sequence in @xmath409 ; i.e. @xmath477 first we show that all the @xmath475 s are unmarked .
suppose on the contrary that for integers @xmath9 and @xmath459 , @xmath479 .
then by definition [ amenable ] , there should be a box @xmath480 such that @xmath481 , @xmath482 and the entry of @xmath480 is @xmath459 .
but then it will be impossible to choose an entry for @xmath483 so that definition [ gsyt ] holds .
therefore @xmath475 can not be marked for any @xmath9 . from equation ( [ aster ] ) it is clear that @xmath484 .
we use induction on @xmath9 .
namely , we assume that the entries of the boxes in @xmath485 are given according to the rules stated in the lemma , and then we use induction on @xmath9 to prove that the entry of @xmath510 is given according to the rules in the statement of the lemma as well .
obviously @xmath513 , since it lies above the box @xmath514 whose entry belongs to @xmath515 .
this amounts for the basis of induction , i.e. @xmath516 .
next suppose @xmath63 .
then @xmath510 lies to the right of @xmath517 and above @xmath518 .
however , by our assumption about @xmath485 , the entries of the latter two boxes , which belong to @xmath485 , are known .
a simple argument based on monotonicity of rows and columns implies that if @xmath519 is the entry of @xmath510 then @xmath520 .
it suffices to show that @xmath521 is impossible .
the proof is by contradiction .
suppose on the contrary that @xmath521 .
by induction hypothesis we know that @xmath522 is filled by @xmath523 , which implies that it is impossible to choose an entry for @xmath524 such that definition [ gsyt ] holds .
consequently , @xmath525 . * case 2 . *
first note that @xmath510 lies to the right of @xmath518 and above @xmath526 .
but the entries of the latter two boxes are known , and a simple argument based on monotonicity of rows and columns implies that if @xmath519 is the entry of @xmath510 then @xmath527 .
next we show that @xmath519 can not be equal to @xmath528 .
suppose on the contrary that @xmath529 .
suppose @xmath519 corresponds to the letter @xmath530 of the row word of the filling .
then by a bookkeeping argument we have @xmath531 by lemma [ chalghuz ] and lemma [ malghuz ] there exists a @xmath287 such that the box @xmath532 is filled by an @xmath512 . obviously @xmath533 .
let the entry of @xmath532 correspond to @xmath534 .
then there should exist an integer @xmath535 such that @xmath536 and @xmath537 , because otherwise we will have @xmath538 and @xmath539 which contradicts the second property of definition [ amenable ] .
however , existence of @xmath535 leads to a contradiction too .
in fact @xmath540 should correspond to a box @xmath461 such that either @xmath541 or @xmath542 and @xmath543 .
however , if @xmath543 and @xmath544 then @xmath461 lies above @xmath545 for some @xmath546 , and therefore the entry of @xmath461 should be strictly less than @xmath511 .
( note that if @xmath63 then by induction hypothesis @xmath547 and if @xmath516 then @xmath548 . )
if @xmath549 and @xmath541 then @xmath461 lies below @xmath532 and therefore its entry should be strictly larger than @xmath512 . finally , if @xmath550 and @xmath541 , then @xmath551 but it is impossible to choose its entry such that definition [ gsyt ] holds .
therefore all of these cases lead to a contradiction .
billera , louis and thomas , hugh and van willigenburg , stephanie , _ decomposable compositions , symmetric quasisymmetric functions and equality of ribbon schur functions _ , preprint ( 2005 ) , accepted in adv . math . | we find a simple criterion for the equality @xmath0 where @xmath1 and @xmath2 are schur s q - functions on infinitely many variables . |
resolving the emission regions of distant quasars is beyond the current capabilities of present - day telescopes , as direct imaging of accretion disks requires angular resolutions on the order of tens of nanoarcseconds at @xmath7 @xmath1 1 . until the spatial resolution of telescopes reaches this limit
, we will have to rely on indirect methods of mapping the emission regions of quasars .
such methods include light - travel time arguments , reverberation mapping of the broad line region ( blandford & mckee 1982 ; peterson 1993 , netzer & peterson 1997 ) , reverberation mapping of the fe k@xmath8 emission region ( young & reynolds 2000 ) , occultation measurements of the central x - ray source by orbiting compton thick clouds ( risaliti et al , 2007 ) , and microlensing of the continuum and line emission regions ( e.g. , grieger et al . 1988 and 1991 ; schneider , ehlers & falco 1992 ; gould & gaudi 1997 ; agol & krolik 1999 ; yonehara et al .
1999 ; mineshige & yonehara 1999 ; chartas et al . 2002a ; popovic et al . 2003 ; blackburne et al . 2006 ; pooley et al .
2006 , 2007 ; kochanek et al . 2007 ;
jovanovic et al . 2008 ; morgan et al .
2008a , 2008b ) .
we observe microlensing as uncorrelated variability in the brightnesses of the images of a lensed quasar , where the amplitude of the variability is determined by the size of the emission regions ( e.g. , lewis et al .
1998 ; popovic & chartas 2005 ) .
the largest components , the radio , dust and optical line emission regions , should show little or no microlensing variability .
the thermal continuum emission from the disk should show greater variability at shorter wavelengths corresponding to smaller disk radii and higher temperatures , and this has been observed by poindexter et al .
( 2008 ) and anguita et al .
the non - thermal x - ray emission is thought to be dominated by inverse compton scattering of uv photons from the disk ( e.g. , reynolds & nowak 2003 ) , some of which is reprocessed into fe k@xmath8 line emission ( e.g. , george & fabian 1991 ; mushotzky et al .
1993 ) but the geometry and scale of these emission regions is not well understood . the effects of microlensing on the x - ray fluxes of lensed images have been reported for many lenses at this point , including q0957 + 561 ( e.g. , chartas et al .
1995 ) , rx j0911 + 0511 ( e.g. , morgan et al .
2001 ; chartas et al . 2001 ) , h1413 + 117 ( e.g. , chartas et al
. 2004 , 2007 ) , pg 1115 + 080 ( pooley et al . 2006 ; morgan et al .
2008b ) , and rx j1131@xmath01231 ( e.g. , blackburne et al . 2006 ; kochanek et al . 2007 ; dai et al .
2008 ) . the differences between the x - ray flux ratios and optical flux ratios can be used to estimate the source size after accounting for absorption by the interstellar medium of the lens ( e.g. , dai & kochanek 2008 ) and the effects of substructures ( satellites ) in the lens halo ( mao & schneider 1998 ) .
pooley et al .
( 2007 ) used observed x - ray and optical flux ratios for ten four - image lensed quasars to show that the x - ray source is much more strongly microlensed than the optical source , indicating that the x - ray emission region is more compact than the optical .
they also found that their estimates of the optical source sizes were significantly larger than would be expected based on the observed , magnification - corrected optical fluxes .
morgan et al .
( 2008a ) measured the sizes of the optical emission regions of 11 quasars using microlensing to find that the sizes scaled with black hole mass as expected from thin disk theory ( shakura & sunyaev 1973 ) and had approximately the sizes expected from thin disk theory , but confirmed the pooley et al .
( 2007 ) result that these sizes , both microlensing and theoretical , are larger than expected given the observed optical flux and the same disk model .
accurate microlensing estimates of the sizes of the x - ray emission regions requires x - ray light curves rather than isolated epochs .
first , the absolute magnifications of lensed images are not well understood because substructure in the lens galaxy can modify the fluxes from the predictions of lens models ( e.g. , mao & schneider 1998 ; dalal & kochanek 2002 ) . while comparing optical and x - ray flux ratios
provides estimates for the difference in size between the emission regions , the lack of a secure estimate of the intrinsic flux ratios makes it difficult to set an absolute scale .
second , quasars have intrinsic variability that appears in the images with relative time delays , which means that instantaneous flux ratios are contaminated by the effects of time variability modulated by the delays .
third , while the source size ultimately determines the amplitude of microlensing variability , the instantaneous values are a combination of the source size and the location of the source in the complex microlensing magnification patterns . monitoring the variability minimizes these problems and should lead to far more accurate estimates of the sizes .
we have been monitoring roughly 25 lenses in the optical to measure time delays ( e.g. , kochanek et al .
2006 ) and to study quasar structure ( e.g. , morgan et al . 2008a
; poindexter et al . 2008 ) .
from this sample we selected a small subset with reasonable x - ray fluxes for monitoring with the _ chandra x - ray observatory_. given a well - sampled optical light curve that can be used to determine the size of the optical / uv emission region and act as a microlensing variability `` reference , '' we can afford to sample the x - ray microlensing variability relatively sparsely . in this paper
we present the x - ray observations of the lensed quasars rx j1131@xmath01231 ( @xmath9 ) and he 1104@xmath01805 ( @xmath10 .
the optical data for rx j1131@xmath01231 and he 11041805 are presented in morgan et al .
( 2006 ) and poindexter et al .
( 2007 ) respectively , and the optical microlensing results are presented in morgan et al .
( 2008a ) and poindexter et al .
( 2008 ) . here
we carry out a microlensing analysis of he 1104@xmath01805 , while a companion paper , dai et al .
( 2009 ) , does so for rx j1131@xmath01231 .
we discuss the observations and data analysis in 2 , study the microlensing of he 11041805 in 3 and summarize our results in 4 . throughout this paper
we adopt a @xmath11-dominated cosmology with @xmath12 = 70 km s@xmath13 mpc@xmath13 , @xmath14 = 0.7 , and @xmath15 = 0.3 .
we monitored rx j1131@xmath01231 and he 1104@xmath01805 with the advanced ccd imaging spectrometer ( acis ; garmire et al .
2003 ) on board the _ chandra x - ray observatory _ ( hereafter _ chandra _ ) using short ( @xmath1 5 ksec ) snapshot observations .
we obtained 5 epochs for each lens in 2006 placing the sources on the back - illuminated s3 chip of acis .
we combined these with archival observations of rx j1131@xmath01231 in april 2004 and he 1104@xmath01805 in february 2000 .
a log of the observations that includes observation dates , observed count rates , total exposure times , and observation identification numbers is presented in table 1 .
we analyzed the data using the standard software ciao 4.0 provided by the cxc .
we used standard cxc threads to screen the data for status , grade , and time intervals of acceptable aspect solution and background levels .
we removed the @xmath16 025 spatial randomization applied to the event positions by the standard pipeline and instead used the sub - pixel resolution techniques developed by tsunemi et al .
( 2001 ) and mori et al .
( 2001 ) in order to improve the image resolution .
the _ chandra _ spectra of rx j1131@xmath01231 and he 1104@xmath01805 were fit with a variety of models employing ` xspec ` version 12 ( arnaud 1996 ) . for all spectral models of rx j1131@xmath01231 and he 1104@xmath01805 we included galactic absorption due to neutral gas ( dicky & lockman 1990 ) with column densities of @xmath17= 3.6 @xmath4 10@xmath18 @xmath19 and 4.6 @xmath4 10@xmath18 @xmath19 , respectively .
we use 90% confidence level uncertainties unless otherwise stated . for estimating the x - ray counts of images b , c , and d of rx j1131@xmath01231 we extracted events from circular regions with radii of 1.5 arcsec slightly off - center from the images to reduce contamination from nearby images ( see panel ( c ) of fig . 1 )
. to estimate the x - ray counts of image a of rx j1131@xmath01231 , which lies between images b and c , we extracted events within a 0.75 arcsec circular region centered on a. aperture corrections were applied to all images to account for counts not included in the extraction regions . for estimating the x - ray counts of images a and b of he 1104@xmath01805 we extracted events from circular regions centered on the images with radii of 1.5 arcsec . the backgrounds for rx j1131@xmath01231 and
he 1104@xmath01805 were determined by extracting events within an annulus centered on the mean location of the images with inner and outer radii of 7.5 arcsec and 50 arcsec , respectively .
contamination between images was minimized and made insignificant by the application of the sub - pixel resolution technique and the use of appropriately sized and off - center extraction apertures .
we also corrected for pile - up , " an instrumental effect that occurs when two or more x - ray photons strike individual or neighboring ccd pixels within one frame time .
neglecting these corrections can cause spectral distortion , grade migration and distortion of the image psf .
we used the forward spectral - fitting tool lynx ( chartas et al .
2000 ) to estimate the fraction of events lost due to the pile - up effect .
table 1 summarizes the observed 0.210 kev band source counts .
the pile - up corrections needed to correctly estimate these counts can be significant .
for example the counts for images b and c in the 2004 april 12 observation of rx j1131@xmath01231 have pile - up corrections of 37% and 16% respectively . in their analysis of the 2004 observation of rx j1131@xmath01231 , blackburne et al .
( 2006 ) did not include pile - up corrections , so our present results should be more reliable .
in particular , pile - up usually leads to energy spectra that appear harder , and this likely explains the differences in the x - ray spectral hardness ratios reported by blackburne et al .
( 2006 ) and our present results . in figure 1
we show the lucy - richardson deconvolved images in the 0.28 kev bandpass of the _ chandra _ observations of rx j1131@xmath01231 and he 1104@xmath01805 . to estimate the relative x - ray image positions of rx j1131@xmath01231 and he 1104@xmath01805 we modeled the _
chandra _
images using point - spread functions ( psfs ) generated by the simulation tool ` marx ` ( wise et al 1997 ) .
the x - ray event locations were binned with a bin - size of 00246 to sample the psf sufficiently ( an acis pixel subtends 0491 ) .
the simulated psfs were fitted to the _ chandra _ data by minimizing the @xmath20-statistic formed between the observed and simulated images . in table 2
we compare the x - ray image separations to the observed nicmos values reported by morgan et al .
( 2006 ) for rx j1131@xmath01231and by lhar et al .
( 2000 ) for he 1104@xmath01805 .
we conclude that the x - ray and optical image positions are consistent given the estimated uncertainties . the deconvolved images of rx j1131@xmath01231 and he 1104@xmath01805 do not show any additional lensed images . to obtain quantitative limits on possible additional images located at the centers of rx j1131@xmath01231 and he 1104@xmath01805we extracted the 0.210 kev counts within 0.4 arcsec circles centered on the images at the mean lens position of each quasar .
the backgrounds in the central source extraction regions are dominated by the contamination from the images the instrumental and cosmic backgrounds in the 0.4 arcsec extraction circles are significantly lower than the contamination . to estimate the fractional contamination per image we used the 0.210 kev counts in 0.4 arcsec circular apertures placed 1 and 2 arcsec north of image b in rx j1131@xmath01231 and 1 and 2 arcsec northwest of image a in he 1104@xmath01805 . the detected 0.210 kev counts in the central 0.4 arcsec apertures are consistent with the estimated background and contamination from the bright images . by combining all the data for each lens we set 3@xmath6 upper limits on the 0.210 kev flux of any central , odd image of @xmath21 and @xmath22 ergs s@xmath13 @xmath19 for rx j1131@xmath01231 and he 1104@xmath01805 , respectively .
these limits corresponded to a 3@xmath6 limit on the flux ratio of a factor of 50 relative to image c in rx j1131@xmath01231 , and a factor of 30 relative to image a in he 1104@xmath01805 . while relatively tight , these limits are not strong enough to constrain the central surface density of the lens ( keeton 2003 ) .
for example , in the models we use for he 1104@xmath01805 below , the expected flux ratio between image a and any central image is @xmath23 .
we performed fits to the individual spectra of the images of rx j1131@xmath01231 and he 1104@xmath01805 using events in the 0.48 kev energy range with a model that consisted of a simple power law modified by galactic absorption .
due to the moderate s /
n of the spectra , we performed these fits using the cash statistic which does not require binning of the data , although if we instead use @xmath24 statistics we find similar results .
the best - fit parameters of these fits and the unabsorbed 0.22 kev and 2 - 10 kev fluxes are presented in tables 2 and 3 .
we found no significant variability ( within the 90% errors ) of the photon indices ( @xmath25 ) of the spectra of the images of rx j1131@xmath01231 and he 1104@xmath01805 with the exception of the 2004 observation of rx j1131@xmath01231 that showed significant differences of @xmath25 between images a / b / c with @xmath26 and image d with @xmath27 .
we also considered models that included neutral absorption at the redshifts of the sources or lenses of rx j1131@xmath01231 and he 1104@xmath01805 .
we do not detect any additional neutral absorption at these redshifts .
this is consistent with the marginal detection of absorption in he 1104@xmath01805 in our earlier analysis ( dai et al .
2006 ) . in figure
2 we show the _ chandra _ image of the 2004 observation of the lensed system rx j1131@xmath01231 and the surrounding field . to reduce background contamination and to enhance possible soft extended x - ray emission we filtered the image to include only photons with energies ranging between 0.4 and 3.0 kev .
the image was binned with a bin size of 0.5 arcsec and adaptively smoothed with the tool csmooth developed by ebeling et al .
csmooth smooths a two - dimensional image with a circular gaussian kernal of varying radius .
we confirm the extended soft x - ray emission reported by morgan et al .
( 2006 ) centered at 11 32 1.6 , @xmath012 31 6.5 ( j2000 ) and about 158 arcsec northeast of image d of rx j1131@xmath01231 .
this emission comes from a foreground cluster at @xmath28 .
morgan et al .
( 2006 ) also reported a possible detection at a low significance level ( 3 @xmath6 ) of extended emission centered 33 arcsec southwest of rx j1131@xmath01231 at 11 31 50.1 , @xmath012 32 23 ( j2000 ) .
we combined all six observations of rx j1131@xmath01231 listed in table 1 , for a total exposure time of 33.8 ksec , to better limit this possible source .
all the observations from 2006 were performed in sub - array mode to mitigate pile - up and with the reduced field of view they do not include the foreground cluster to the northeast of the lens .
figure 3 shows the adaptively smoothed image including only photons with energies from 0.4 to 3.0 kev .
we find no significant sources of extended emission near the lens beyond the foreground cluster .
the 0.22 .
kev @xmath29 upper flux and luminosity limits are @xmath30 ergs @xmath19 s@xmath13 and @xmath31 ergs s@xmath13 , respectively , assuming a cluster with a temperature of @xmath32 kev at the lens redshift with an abundance of @xmath33 solar and an extraction radius of 30 arcsec .
we fit the emission from the foreground cluster using a @xmath34 model for the cluster brightness profile combined with a uniform background of 0.005 events per pixel . prior to performing the fit we binned the image in 1 pixels and smoothed this with a gaussian ( @xmath6 = 3 ) .
the fits were performed with the cxc software package ` sherpa ` .
we find that the cluster center is @xmath35 = 1502 @xmath16 05 east and @xmath36 = 523 @xmath16 05 north of image d. the smoothed intensity distribution is nearly round with an ellipticity of @xmath37 0.10 @xmath16 0.3 .
the best - fit values for @xmath34 and the core radius of the cluster are @xmath34 = 0.4 @xmath16 0.2 and r@xmath38 = 42@xmath39 ( 8 kpc ) , respectively .
we extracted the spectrum of this @xmath40 cluster from a 50 arcsec radius circle centered on the x - ray cluster center .
this spectrum was fit with a simple model consisting of an emission spectrum from hot diffuse gas based on the xspec model mekal modified by galactic absorption .
we obtain best - fit values for the temperature and metal abundances of 1.2@xmath41 kev and a = 0.5@xmath42 , respectively ( both errors are at the 90% confidence level ) .
the 210 kev luminosity of this cluster of galaxies is 1.7 @xmath4 10@xmath43 ergs sec@xmath13 .
these are consistent with the morgan et al .
( 2006 ) estimates and the values for the temperature , @xmath34 and the core radius are consistent with the observed correlations of clusters ( e.g. jones & forman 1999 ) .
we also stacked the images of the six observations of he 1104@xmath01805listed in table 1 for a total exposure time of 71.8 ksec .
figure 4 shows the adaptively smoothed image of he 1104@xmath01805 including only photons with energies from 0.4 to 3.0 kev .
we find no significant sources of extended emission .
the 0.22 kev @xmath29 upper flux and luminosity limits are @xmath44 ergs @xmath19 s@xmath13 and @xmath45 ergs s@xmath13 , respectively , assuming a cluster with a temperature of @xmath32 kev at the lens redshift with an abundance of @xmath33 solar and an extraction radius of 30 arcsec .
the 0.210 kev time - delay corrected light curves of the images of rx j1131@xmath01231 and he 1104@xmath01805 are shown in figures 5a and 5b respectively .
significant x - ray flux variability is detected in all images .
rx j1131@xmath01231 shows a mixture of intrinsic and microlensing variability .
if we normalize the first epoch for image d to match the optical light curves , we find that images b , c and d roughly track the optical light curves and show similarities in their time variability . in figure 6a
we show the flux ratios f@xmath46/f@xmath47 , f@xmath48/f@xmath47 , and f@xmath49/f@xmath47 in the 0.210 kev and @xmath50-bands of rx j1131@xmath01231 .
the flux ratio f@xmath48/f@xmath47 is almost constant in both the x - ray and optical bands indicating that the observed variability in images b and c is correlated and therefore mostly intrinsic in origin .
image a shows a completely different behavior .
it is a factor of @xmath1 4 fainter in the x - rays than the optical relative to c / d in 2004 , comparable to the optical in the spring of 2006 , and a factor of @xmath1 5 brighter by the fall of 2006 .
since the 2006 observations come in clusters with temporal separations of order the a / b / c time delays , we can be confident that these differences are not due to intrinsic variability modulated by the time delay .
the significant rise of the f@xmath46/f@xmath47 ratio in the x - ray band compared to the one measured in the @xmath50-band supports our microlensing interpretation of the flux enhancement in image a of rx j1131@xmath01231 .
we note that the flux ratios reported in our analysis for the 2004 observation of rx j1131@xmath01231 are different from those reported by blackburne et al .
( 2006 ) because of the correction for the loss of counts due to pile - up and the correction for the spectral distortion due to pile - up .
we see no strong evidence for variations in the spectral hardness , but this is a weak statement given the limited count rates .
he 1104@xmath01805is also affected by intrinsic x - ray variability . in particular , the peak in the optical flux near 4100 days seems likely to be associated with a peak in the x - ray flux . in figure 6b
we show the flux ratios f@xmath46/f@xmath48 in the 0.210 kev and @xmath50-bands of he 1104@xmath01805 .
the february / march and october / november epochs are roughly separated by the time delay , so we can reasonably conclude that the x - ray flux ratio near 4050 days ( observed a ) is approximately f@xmath46/f@xmath48@xmath51 and very different from the optical flux ratio of f@xmath46/f@xmath48@xmath52 . since there is no evidence for significant soft x - ray or dust absorption ( see 2.3 , dai et al .
2006 , poindexter et al .
2008 ) this difference in flux ratios must be due to microlensing . in
3 we proceed with fitting the light - curves of he 1104 - 1805 with a sophisticated microlensing model .
the microlensing analysis of the light - curves of rx j 1131 - 1231 will be presented in a companion paper by dai et al .
( 2009 ) . as with rx j1131@xmath01231
, we see no evidence for variations in the spectral hardness , largely due to the large uncertainties .
we modeled the microlensing in he 1104@xmath01805 as in morgan et al .
( 2008a ) and poindexter et al .
( 2008 ) using a fixed 162 day time delay , the @xmath50-band light curve constructed from our smarts data ( poindexter et al . 2007 ) , ogle ( wyrzykowski et al . 2003 ) , and ofek & maoz ( 2003 ) , and the delay - corrected x - ray flux measurement . the main free parameters of our lensing model for he 1104@xmath01805are the x - ray and uv ( rest - frame ) scale lengths , a microlens mass scale , a mass fraction of the local surface density comprised of stars , and a velocity vector describing the motion of the agn emission region across the microlensing caustics .
we used the lens model sequence from poindexter et al .
( 2007 ) , where we start from a constant mass - to - light ratio model for the mass distribution , defined by @xmath53 , and then reduce its mass in 10% increments while adding an nfw halo , where @xmath54 would correspond to a pure halo model .
the measured time delay requires @xmath55 .
we made 8 random realizations of the star fields near each image and then generated @xmath56 pixel magnification maps using the methods of kochanek ( 2004 ) .
the maps had an outer scale of @xmath57 cm and an inner scale of @xmath58 cm . based on its emission line widths , peng et al .
( 2006 ) estimated that the black hole mass in he 1104@xmath01805is @xmath59 , corresponding to a gravitational radius of @xmath60 cm that is well - resolved in the magnification maps .
we modeled the surface brightness of the emission regions as a face - on , thin disk ( shakura & sunyaev 1973 ) without the central temperature depression , @xmath61^{-1 } \label{eqn : fnu}\ ] ] where the scale length @xmath62^{1/3}\ ] ] @xmath63 is the radius at which the disk temperature matches the rest wavelength of the observations , @xmath64 , @xmath65 is the accretion rate , @xmath66 is the luminosity in units of the eddington luminosity and @xmath67 is the accretion efficiency .
we scale the results assuming a radiative efficiency of @xmath68 . while the available energy of accretion is larger , the division of that energy between radiative , advective and kinetic losses is not directly calculable at present .
studies of the growth of black holes tend to find that the average radiative efficiency of accretion is less than 10% ( e.g. shankar et al . 2007 ) , and low radiative efficiencies are also needed to reconcile thin disk theory with microlensing accretion disk size estimates ( e.g. , morgan et al .
2007 ) . for the uv ( rest - frame ) emission we can neglect the central temperature depression provided @xmath69 where @xmath70 is the inner edge of the disk .
while this thermal accretion disk model is not appropriate for the non - thermal x - ray emission , mortonson et al .
( 2005 ) have shown that microlensing essentially measures the half - light radius of the emission region , so we used the same profile for the x - rays to simplify the computations but regard it as a measurement of the half - light radius @xmath71 .
the microlenses were given a power - law mass distribution @xmath72 with a dynamic range in mass of a factor of 50 , roughly matching the galactic disk mass function of gould ( 2000 ) , and we use a uniform prior on the mean microlens mass of @xmath73 . we modeled the probability distribution for the physical velocities as in poindexter et al .
( 2008 ) where our projected velocity onto the lens plane is @xmath74 km / s , the velocity dispersion of the lens is @xmath75 km / s based on the critical radius of the mass model , and the lens and source have rms peculiar velocities of @xmath76 and @xmath77 km / s respectively .
the data were then modeled using the bayesian monte carlo method of kochanek ( 2004 ) .
the microlensing data constrains the source velocity and size in `` einstein '' length units of @xmath78 cm .
we can determine the true sizes using the prior on the microlens masses , the prior on the physical velocities or both .
figure 7 shows the probability distribution found for @xmath79 assuming the prior on the physical velocities . while the distribution peaks at somewhat lower mass ( @xmath80 ) than the range of our prior ( @xmath81 to @xmath82 ) , the distribution is so broad ( @xmath83 to @xmath84 at 1-@xmath6 ) that we can view them as statistically consistent .
for example , none of the results depend significantly on the inclusion of the mass prior .
since the mass scales as @xmath85 where @xmath86 is the physical velocity from the prior and @xmath87 is the einstein - unit velocity estimated from the microlensing models , it is difficult to tightly constrain the mean mass with a single lens given the unknown peculiar velocities of the lens and source even with perfect knowledge of the einstein - unit velocity . figure 8 shows the raw estimates of the source sizes @xmath88 in einstein units , where @xmath89 . both of these are scaled source sizes @xmath90 defined by the emission profile presented earlier in this section . the consequence of adding the prior on the microlens masses is to drive the source sizes @xmath88 to be modestly smaller .
in essence , a given level of microlensing variability can be created by a large source moving rapidly ( big @xmath88 and @xmath87 ) or a small source moving more slowly ( small @xmath88 and @xmath87 ) .
but high effective velocities correspond to small mass scales given a fixed physical velocity , so imposing the mass prior given the distribution seen in figure 7 will favor lower velocities @xmath87 and smaller sources @xmath88 .
it will have less effect on the physical source size @xmath89 because the smaller size @xmath88 is balanced by the larger mass scale @xmath91 ( see kochanek 2004 ) . while figure 8 truncates the probability distribution for the x - ray source at the pixel scale of the magnification patterns
, it would not converge to zero if we used magnification patterns with smaller pixel scales to extend the distribution to smaller source sizes .
this is a consequence of not having a well - sampled x - ray light curve .
the source must be small enough to allow the x - ray flux ratio to differ from the optical , but still smaller sources are disfavored because even more discrepant flux ratios become more likely .
however , reducing the size still further has no effect on the probabilities because it converges to the likelihood of the data points avoiding precise caustic crossings .
unfortunately , this does mean that we will obtain robust upper limits on the x - ray source size but not lower limits .
figure 9 shows the results for the rest - frame uv and x - ray source sizes @xmath90 assuming either uniform or logarithmic priors for the size .
we used @xmath90 for both so that the sizes are easily compared .
if we regard the source size at 2000 as measuring a projected area , then @xmath90 scales as @xmath92 for inclination angle @xmath93 of a thin inclined accretion disk .
similarly , the x - ray size in figure 9 should be increased by a factor of @xmath94 to convert from @xmath90 to the half - light radius @xmath95 .
the disk size at 2000 ( @xmath50-band in the rest frame at @xmath96 ) is @xmath97 for a logarithmic prior on the size and @xmath98 for a linear prior on the size .
this agrees well with our earlier estimates in poindexter et al .
( 2008 ) and morgan et al .
( 2008a ) .
the x - ray size is less certain and more prior - dependent , with an estimated half - light radius @xmath99 for a logarithmic prior on the size and @xmath100 for the linear prior on the size .
we note that the difference in the sizes is entirely driven by the significant differences between the x - ray and optical flux ratios at the time of the x - ray observation .
the microlensing data also prefer lens models with low stellar mass fractions , as shown in figure 10.the probability distribution peaks at @xmath101 , in relatively good agreement with the value of @xmath102 that agrees best with the measured time delay and a hubble constant of @xmath103 km s@xmath13 mpc@xmath13 . while this preliminary result does not lead to a microlensing constraint on @xmath104 that is tight enough to break the degeneracies between the radial mass distribution and the hubble constant ( see kochanek 2002 ) , it is encouraging that there is agreement with the time delays given the expected value of @xmath105 and that the microlensing requires a dark matter - dominated mass distribution .
a detailed microlensing analysis of rx j1131@xmath01231 that provides similar constraints on the structure of the lens and the structure of the quasar is presented in dai et al .
( 2009 ) .
figures 11 and 12 show an example of a trial that fits the data well . in this trial
the einstein unit source sizes are @xmath106 and @xmath107 for the x - ray and uv rest - frame sources respectively and the effective velocity corresponds to a sensible microlens mass scale .
we chose this solution mainly because of the curious origin of the initial @xmath50-band peak , where what might be interpreted as a single heavily smoothed caustic crossing is actually a pair of competing caustic crossings in the a and b images , where the a image dominates because of its shallower magnification gradients .
the x - ray epoch then corresponds to time when the source is in a valley between caustics in both images .
we note that this particular solution probably would not reproduce the wavelength dependence of the optical and ir flux ratios in he 1104@xmath01805 discussed and modeled poindexter et al .
while there is a consensus that the origin of the kev x - ray continuum emission from quasars is inverse compton scattering ( e.g. , see the review of reynolds & nowak 2003 ) , the extent and geometric configuration of the emission region is less clear . in the standard disk - corona model ( e.g. , haardt & maraschi 1991 ; merloni 2003 ) , it is produced in a hot , extended corona surrounding the disk . in the general relativistic mhd simulations of hirose et al .
( 2004 , also machida & matsumoto 2003 ) the region of the corona with the highest current densities lies close to the inner edge of the disk . under the assumption that these are also the regions with the highest dissipation rates for heating the corona ( see hirose et al .
2004 ) , they will dominate the x - ray emission and we would expect the x - ray source size to track the inner edge of the disk .
microlensing of the x - ray source allows us to observationally measure the size of the x - ray emission region .
we start with well - sampled optical light curves that allow us to determine the size of the accretion disk .
morgan et al .
( 2008a ) found a clear scaling of the disk size with black hole mass that is consistent with the @xmath108 scaling of simple thin disk theory for a fixed eddington accretion rate ( shakura & sunyaev 1973 ) .
we can estimate the disk size by two routes other than microlensing .
first we can estimate it from the observed flux at some wavelength .
size estimates from the observed flux do not depend on the black hole mass but depend only on the assumed temperature profile and the assumption that the disk radiates as a black body at that temperature .
second , thin disk theory predicts the scale length through the expression we presented in the beginning of 3 .
size estimates from thin disk theory depend on the black - hole mass , the eddington factor and the wavelength .
the disk size from our microlensing analysis is roughly consistent with thin disk theory if the accretion efficiency is relatively low , but both the microlensing and theoretical sizes are significantly larger than that estimated from the observed optical flux of the quasar and the same disk model .
this mismatch between the sizes estimated from microlensing and from the flux was first pointed out by pooley et al .
one solution to this problem may be to use a shallower temperature profile than the @xmath109 of simple thin disk theory , and poindexter et al .
( 2008 ) found some evidence for a shallower profile in he 1104@xmath01805 by examining the optical / near - ir wavelength dependence of microlensing in this system .
our x - ray monitoring observations of rx j1131@xmath01231 and he 1104@xmath01805 show significant flux variability in all images of both quasars .
the factor of @xmath110 x - ray flux variability in image a of rx j1131@xmath01231 is interesting not only because of the high amplitude but also because it is a highly magnified saddle point image .
schechter & wambsganss ( 2002 ) pointed out that in a dark matter dominated galaxy , microlensing is in a regime where images at minima and saddle points of the time - delay surface behave differently . in particular , the saddle point images should show significantly higher variability amplitudes , just as we observe in rx j1131@xmath01231 .
the relatively large sizes of accretion disks in the optical relative to the typical einstein radius of the stars seems to have masked much of this effect in the optical signatures of microlensing for example , where the x - ray flux ratio @xmath111 changed by a factor of 4 during our monitoring period , the optical flux ratio changed by only a factor of 1.4 .
the differences between the x - ray and optical flux ratios of he 1104@xmath01805 also imply the presence of x - ray microlensing .
neither system shows evidence for soft x - ray or dust absorption that could explain the differences .
while we measure the source sizes by modeling the data in detail , the essence of the measurement is comparing the amplitude of the x - ray microlensing variability to that in the optical .
if the amplitude for the x - ray microlensing is larger than in the optical , then the x - ray source must be more compact , and this is what we observed in both rx j1131@xmath01231 and he 1104@xmath01805 . given the estimated black hole mass of @xmath112 ( peng et al .
2006 ) , our fits to the x - ray and optical light curves using the microlensing model described in 2.5 indicate that the x - ray emission in he 1104@xmath01805 comes from a region smaller than @xmath113 and the uv emission region has a size of @xmath114 .
we note that the source sizes of he 1104@xmath01805 were derived using a bayesian analysis which requires a prior .
as shown in figures 8 and 9 our upper limit on the x - ray source size is robust to prior selection but not the lower limit .
obtaining a more precise x - ray size estimate requires a more extensive x - ray light curve for the system .
this is somewhat challenging for he 1104@xmath01805 because the long time delay of he 1104@xmath01805 ( 162 days , morgan et al .
2008a ) and the intrinsic variability we observe here means that this lens requires observations separated by the time delay in order to correct for intrinsic variability .
this is less of a problem for almost all other candidates for x - ray monitoring because they have far shorter time delays .
for example , in rx j1131@xmath01231 we observe some intrinsic variability but it is significant only on time scales much longer than the delays between the a / b / c images .
given the morgan et al .
( 2008b ) estimates for pg 1115 + 080 , the dai et al .
( 2009 ) results for rx j1131@xmath01231 , and our present results for he 1104@xmath01805 , the x - ray microlensing results suggest that the size of the x - ray emission arises from close to the inner edge of the disk .
thus , our estimated sizes for the x - ray and uv emission regions place significant constraints on agn models that assume extended coronal envelopes around accretion disks . reducing the uncertainties in these measurements and measuring the scaling of the size with the estimated masses of the black holes
requires longer x - ray time series and more systems , but there is clearly no technical barrier to making the measurements and extending them to the dependence of the size on x - ray energy .
combining the data from the time series also allows us to search for extended emission from groups or clusters near the lens , which can be important for lens models ( e.g. williams et al .
2006 , fassnacht et al .
2007 ) . here
we analyzed the x - ray data for the foreground cluster found near rx j1131@xmath01231 by morgan et al .
( 2006 ) in more detail , and set limits on the presence of any other clusters near both systems ( see 2.4 ) .
we also used the combined images to search for central or odd images , although this is challenging given the resolution of _ chandra _ and the expected faintness of the images . while we reached 1@xmath6 limits on the flux ratios of @xmath115 , significant constraints on the central images
require limits better by another order of magnitude ( keeton 2003 ) .
agol , e. , & krolik , j. 1999 , apj , 524 , 49 + anguita , t. , schmidt , r.w . ,
turner , e.l . , wambsganss , j. , webster , r.l . ,
loomis , k.a .
, long , d. , & mcmillan , r. , 2008 , a&a , 480 , 327 + arnaud , k. a. 1996 , asp conf .
ser . 101 : astronomical data analysis software and systems v , 5 , 17 + blackburne , j. a. , pooley , d. , & rappaport , s. 2006 , apj , 640 , 569 + blandford , r. d. , & mckee , c. f. 1982 , apj , 255 , 419 + chartas , g. , et al . 2000 , , 542 , 655 + chartas , g. , agol , e. , eracleous , m. , garmire , g. , bautz , m. w. , & morgan , n. d. 2002a , apj , 68 , 509 + chartas , g. , gupta , v. , garmire , g. , jones , c. , falco , e. e. , shapiro , i. i. , & tavecchio , f. 2002b , , 565 , 96 + chartas , g. , dai , x. , gallagher , s. c. , garmire , g. p. , bautz , m. w. , schechter , p. l. , & morgan , n. d. 2001 , apj , 558 , 119 + chartas , g. , eracleous , m. , agol , e. , & gallagher , s. c. 2004 , apj , 606 , 78 + chartas , g. , eracleous , m. , dai , x. , agol , e. , & gallagher , s. 2007 , apj , 661 , 678 + chartas , g. , falco , e. , forman , w. , jones , c. , schild , r. , & shapiro , i. 1995 , apj , 445 , 140 + dai , x. , kochanek , c.s . , chartas , g. , & mathur , s. , 2006 , apj , 637 , 53 + dai , x. , & kochanek , c.s .
, 2008 , apj submitted [ arxiv:0803.1679 ] + dalal , n. , & kochanek , c. s. 2002 , , 572 , 25 + dickey , j. m. , & lockman , f. j. 1990 , , 28 , 215 + fassnacht , c.d . ,
kocevski , d.d .
, auger , m.w . ,
lubin , l.m .
, neureuther , j.l . , jeltema , t.e . ,
mulchaey , j.s . , & mckean , j.p . , 2007 ,
apj submitted [ arxiv:0711.2066 ] + garmire , g. p. , bautz , m. w. , ford , p. g. , nousek , j. a. , & ricker , g. r. 2003 , , 4851 , 28 + george , i. m. , & fabian , a. c. 1991 , , 249 , 352 + gould , a. , & gaudi , b. s. 1997 , apj , 486 , 687 + gould , a. , 2000 , apj , 535 , 928 + grieger , b. , kayser , r. , & refsdal , s. 1988 , a&a , 194 , 54 + grieger , b. , kayser , r. , & schramm , t. 1991 , a&a , 252 , 508 + haardt , f. , & maraschi , l. , 1991 , apj , 380 , 51 + hirose , s. , krolik , j. , de villiers , j .- p . , & hawley , j. , 2004 , apj , 606 , 1083 + jones , c. , & forman , w. 1999 , , 511 , 65 + jovanovi , p. , zakharov , a. f. , popovi , l. . , & petrovi , t. 2008 , , 386 , 397 + keeton , c.r .
, 2003 apj , 582 17 + kochanek , c. s. , dai , x. , morgan , c. , morgan , n. , poindexter , s. c. g. & chartas , g. , 2007 , statistical challenges in modern astronomy iv , 371 , 43 + kochanek , c.s . , morgan , n.d . , falco , e.e . , et al . , 2006 , apj , 640 , 47 + kochanek , c. s. 2002 , , 578 , 25 + kochanek , c. s. 2004 , , 605 , 58 + lehr , j. , et al . 2000 , , 536 , 584 + lewis , g. f. , irwin , m. j. , hewett , p. c. , & foltz , c. b. 1998 , , 295 , 573 + machida , m. , & matsumoto , r. , 2003 , apj , 585 , 429 + mao , s. , & schneider , p. 1998
, , 295 , 587 + merloni , a. , 2003 , mnras , 341 , 1051 + mineshige , s. , & yonehara , a. 1999 , pasj , 51 , 497 + morgan , n. d. , chartas , g. , malm , m. , bautz , m. w. , burud , i. , hjorth , j. , jones , s. e. , & schechter , p. l. 2001 , apj , 555 , 1 + morgan , n. d. , kochanek , c. s. , falco , e. e. , & dai , x. 2006 , arxiv astrophysics e - prints , arxiv : astro - ph/0605321 + morgan , c. w. , eyler , m. e. , kochanek , c. s. , morgan , n. d. , falco , e. e. , vuissoz , c. , courbin , f. , & meylan , g. , 2008 , apj , 676 , 80 + morgan , c. w. , eyler , m. e. , kochanek , c. s. , morgan , n. d. , falco , e. e. , vuissoz , c. , courbin , f. , & meylan , g. 2008a , , 676 , 80 + morgan , c. w. , kochanek , c. s. , morgan , n. d. , & falco , e. e. 2007 , arxiv e - prints , 707 , arxiv:0707.0305 + morgan , c. w. , kochanek , c. s. , dai , x. , morgan , n. d. , & falco , e. e. 2008b , arxiv e - prints , 802 , arxiv:0802.1210 + mori , k. , tsunemi , h. , miyata , e. , baluta , c. , burrows , d. n. , garmire , g. p. , & chartas , g. 2001 , in asp conf .
251 , new century of x - ray astronomy , ed .
inoue & h. kunieda ( san francisco : asp ) , 576 + mortonson , m.j . , schechter , p.l .
, & wambsganss , j. , 2005 , apj , 628 , 594 + mushotzky , r. f. , done , c. , & pounds , k. a. 1993 , , 31 , 717 + netzer , h. , & peterson , b. m. 1997 , astronomical time series , 218 , 85 + ofek , e.o . & maoz , d. , 2003 , apj , 594 , 101 + peng , c.y . , impey , c.d . ,
rix , h .- w . ,
kochanek , c.s . ,
keeton , c.r . ,
falco , e.e .
, lhar , j. , & mcleod , b.a . , 2006 ,
apj 649 , 616 + peterson , b. m. 1993 , pasp , 105 , 247 + poindexter , s. , morgan , n. , kochanek , c.s . , & falco , e.e . , 2007 , apj , 660 , 146 + pointdexter , s. mogan , n. , & kochanek , c.s . , 2008 , apj , 673 , 34 + pooley , d. , blackburne , j. a. , rappaport , s. , schechter , p. l. , & fong , w .- f .
2006 , apj , 648 , 67 + pooley , d. , blackburne , j. a. , rappaport , s. , & schechter , p. l. 2007 , apj , 661 , 19 + popovi , l. . , & chartas , g. 2005 , mnras , 357 , 135 + popovi , l. . ,
mediavilla , e. g. , jovanovi , p. , & muoz , j. a. 2003 , a&a , 398 , 975 + reynolds , c.s . & nowak , m.a . , 2003 , phr , 377 , 389 + risaliti , g. , elvis , m. , fabbiano , g. , baldi , a. , zezas , a. , & salvati , m. 2007 , apjl , 659 , l111 + schechter , p. l. , & wambsganss , j. 2002 , , 580 , 685 + schneider , p. , ehlers , j. , & falco , e. e. 1992 , gravitational lenses , xiv , springer - verlag berlin heidelberg new york . also astronomy and astrophysics library .
+ shankar , f. , weinberg , d. h. , & miralda - escude , j. 2007 , arxiv e - prints , 710 , arxiv:0710.4488 + tsunemi , h. , mori , k. , miyata , e. , baluta , c. , burrows , d. n. , garmire , g. p. , & chartas , g. 2001 , , 554 , 496 + williams , k.a . , momchemva , i. , keeton , c.r . , zabludoff , ai . , & lhar , j. , 2006 , apj , 646 , 85 wise , m. w. , davis , j. e. , huenemoerder , houck , j. c. , dewey , d. flanagan , k. a. , and baluta , c. 1997 , _ the marx 3.0 user guide , cxc internal document _ available at http://space.mit.edu/asc/marx/ + wryzkwoski , l. , et al . , 2003 acta astron .
53 229 + yonehara , a. , mineshige , s. , fukue , j. , umemura , m. , & turner , e. l. 1999 , a&a , 343 , 41 + young , a. j. , & reynolds , c. s. 2000 , apj , 529 , 101 + clcccccc + + & & & & & & & + & & _ chandra _ & exposure & & & & + epoch & observation & observation & time & @xmath116@xmath117 & @xmath118@xmath117 & @xmath119@xmath117 & @xmath120 @xmath117 + & date & i d & ( ks ) & counts & counts & counts & counts + & & & & & & & + + 1&2004 april 12 & 4814 & 10 . & 425 @xmath16 22 & 2950 @xmath16 54 & 839 @xmath16 29 & 211 @xmath16 15 + 2&2006 march 10 & 6913 & 4.9 & 393 @xmath16 20 & 624 @xmath16 25 & 204 @xmath16 14 & 103 @xmath16 10 + 3&2006 march 15 & 6912 & 4.4 & 381 @xmath16 20 & 616 @xmath16 25 & 233 @xmath16 15 & 93 @xmath16 10 + 4&2006 april 12 & 6914 & 4.9 & 413 @xmath16 20 & 507 @xmath16 23 & 146 @xmath16 12 & 131 @xmath16 12 + 5&2006
november 10 & 6915 & 4.8 & 3708 @xmath16 61 & 1411 @xmath16 38 & 367 @xmath16 19 & 155 @xmath16 13 + 6&2006 november 13 & 6916 & 4.8 & 3833 @xmath16 62 & 1618 @xmath16 40 & 415 @xmath16 20 & 115 @xmath16 11 + & & & & & & & + + + 1&2000 june 10 & 375 & 47.4 & 1024 @xmath16 32 & 547 @xmath16 23 + 2&2006 february 16 & 6918 & 5.0 & 38 @xmath16 6 & 86 @xmath16 9 + 3&2006 march 15 & 6917 & 4.6 & 40 @xmath16 6 & 59 @xmath16 8 + 4&2006 april 09 & 6919 & 4.9 & 33 @xmath16 6 & 61 @xmath16 8 + 5&2006 october 31 & 6920 & 5.0 & 60 @xmath16 8 & 40 @xmath16 6 + 6&2006 november 08 & 6921 & 4.9 & 52 @xmath16 7 & 57 @xmath16 8 + @xmath121background - subtracted source counts for events with energies in the 0.210 kev band . the counts for images b and c for the 2004 april 12 observation of rx j1131@xmath01231 are corrected for pile - up by factors of 37% and 16% , respectively .
the counts for images a and b for the 2006 november 10 observation of rx j1131@xmath01231 are corrected for pile - up by factor of 32% and 13% , respectively .
the counts for images a and b for the 2006 november 13 observation of rx j1131@xmath01231 are corrected for pile - up by factors of 33% and 17% , respectively . + ccccc + + & & & & + & & & & + & & & & + & & & & + + hst & 0,0 & [email protected],[email protected] & [email protected],[email protected] & @[email protected] , [email protected] + _ chandra _ & 0,0 & [email protected] , [email protected] & [email protected] , [email protected] & @[email protected] , [email protected] + & & & & + + hst & & 0,0 & [email protected],@[email protected] & + _ chandra _ & & 0,0 & [email protected] , @[email protected] & + cccccc + + & & & & & + & & & & & + & & image a@xmath122 & image b@xmath122 & image c@xmath122 & image d@xmath122 + & & & & & + 1 & @xmath123 & 1.44@xmath124 & 1.41@xmath125 & 1.47@xmath126 & 1.95@xmath127 + & 0.22 kev flux ( 10@xmath128 erg s@xmath13 @xmath19 ) & 1.4@xmath129 & 9.1@xmath130 & 3.1@xmath131 & 0.9@xmath129 + & 210 kev flux ( 10@xmath128 erg s@xmath13 @xmath19 ) & 2.7@xmath132 & 15.7@xmath133 & 3.8@xmath134 & 0.7@xmath129 + & @xmath135 & 556/787 & 661/787 & 583/787 & 368/787 + & @xmath136 & 25/31 & 105/97 & 28/41 & 15.4/15 + & @xmath137@xmath138 & 0.76 & 0.27 & 0.95 & 0.42 + & & & & & + 2 & @xmath123 & 1.58@xmath139 & 1.63@xmath140 & 1.62@xmath141 & 1.66@xmath142 + & 0.22 kev flux ( 10@xmath128 erg s@xmath13 @xmath19 ) & 2.9@xmath143 & 5.0@xmath130 & 1.3@xmath144 & 0.8@xmath132 + & 210 kev flux ( 10@xmath128 erg s@xmath13 @xmath19 ) & 4.5@xmath145 & 7.0@xmath146 & 2.3@xmath147 & 1.1@xmath148 + & @xmath135 & 531/838 & 653/838 & 435/838 & 300/838 + & @xmath136 & 33/28 & 61/52 & 14.4/16 & 2.4/6 + & @xmath137@xmath138 & 0.24 & 0.18 & 0.57 & 0.9 + & & & & & + 3 & @xmath123 & 1.61@xmath139 & 1.73@xmath140 & 1.57@xmath149 & 1.60@xmath150 + & 0.22 kev flux ( 10@xmath128 erg s@xmath13 @xmath19 ) & 3.1@xmath151 & 5.6@xmath152 & 1.8@xmath131 & 0.8@xmath39 + & 210 kev flux ( 10@xmath128 erg s@xmath13 @xmath19 ) & 4.6 @xmath145 & 6.7@xmath153 & 3.0@xmath154 & 1.2@xmath134 + & @xmath135 & 521/838 & 555/838 & 425/838 & 292/838 + & @xmath136 & 26/27 & 42/48 & 18.5/17 & 3.8/6 + & @xmath137@xmath138 & 0.52 & 0.73 & 0.36 & 0.7 + & & & & & + 4 & @xmath123 & 1.53@xmath155 & 1.64@xmath126 & 1.70@xmath156 & 1.62@xmath157 + & 0.22 kev flux ( 10@xmath128 erg s@xmath13 @xmath19 ) & 3.0@xmath143 & 3.9@xmath158 & 1.2@xmath132 & 1.0@xmath132 + & 210 kev flux ( 10@xmath128 erg s@xmath13 @xmath19 ) & 5.1@xmath159 & 5.4@xmath160 & 1.5@xmath134 & 1.5@xmath134 + & @xmath135 & 541/838 & 548/838 & 333/838 & 350/838 + & @xmath136 & 27/29 & 40/41 & 7.7/10 & 13/9 + & @xmath137@xmath138 & 0.58 & 0.52 & 0.66 & 0.15 + & & & & & + 5 & @xmath123 & 1.59@xmath125 & 1.82@xmath161 & 1.90@xmath139 & 1.54@xmath162 + & 0.22 kev flux ( 10@xmath128 erg s@xmath13 @xmath19 ) & 29@xmath163 & 13@xmath164 & 3.4@xmath131 & 1.1@xmath144 + & 210 kev flux ( 10@xmath128 erg s@xmath13 @xmath19 ) & 35@xmath165 & 11@xmath166 & 2.8@xmath147 & 1.9@xmath134 + & @xmath135 & 771/838 & 626/838 & 502/838 & 402/838 + & @xmath136 & 72/59 & 66/63 & 34/28 & 7/11 + & @xmath137@xmath138 & 0.1 & 0.37 & 0.19 & 0.7 + & & & & & + 6 & @xmath123 & 1.51@xmath125 & 1.82@xmath124 & 1.76@xmath155 & 2.28@xmath167 + & 0.22 kev flux ( 10@xmath128 erg s@xmath13 @xmath19 ) & 30@xmath163 & 15@xmath163 & 3.2@xmath131 & 1.4@xmath131 + & 210 kev flux ( 10@xmath128 erg s@xmath13 @xmath19 ) & 36@xmath168 & 13@xmath163 & 3.6@xmath154 & 0.6@xmath169 + & @xmath135 & 768/838 & 658/838 & 489/838 & 276/838 + & @xmath136 & 65/59 & 62/70 & 29/33 & 9/8 + & @xmath137@xmath138 & 0.29 & 0.76 & 0.66 & 0.36 + & & & & & + @xmath121 model 1 consists of a power - law modified by galactic absorption .
+ @xmath170all errors are for 90% confidence unless mentioned otherwise with all parameters taken to be of interest except absolute normalization .
+ @xmath171@xmath137 is the probability of exceeding @xmath172 for @xmath173 degrees of freedom if the model is correct .
+ cccc + + & & & + & & & + & & image a@xmath122 & image b@xmath122 + & & & + 1 & @xmath123 & 1.59@xmath124 & 1.86@xmath126 + & 0.22 kev flux ( 10@xmath174 erg s@xmath13 @xmath19 ) & 6.9@xmath143 & 4.28@xmath143 + & 210 kev flux ( 10@xmath174 erg s@xmath13 @xmath19 ) & 10.6@xmath175 & 3.97@xmath176 + & @xmath135 & 623/790 & + & & & + 2 & @xmath123 & 1.4@xmath132 & 1.6@xmath144 + & 0.22 kev flux ( 10@xmath174 erg s@xmath13 @xmath19 ) & 2.6@xmath177 & 6.5@xmath178 + & 210 kev flux ( 10@xmath174 erg s@xmath13 @xmath19 ) & 5.8@xmath179 & 10.2@xmath180 + & @xmath135 & 192/838 & 284/1060 + & & & + 3 & @xmath123 & 1.53@xmath181 & 2.4@xmath132 + & 0.22 kev flux ( 10@xmath174 erg s@xmath13 @xmath19 ) & 3.4@xmath182&8.4@xmath183 + & 210 kev flux ( 10@xmath174 erg s@xmath13 @xmath19 ) & 5.9@xmath184 & 3.0@xmath185 + & @xmath135 & 186/838 & 195/1060 + & & & + 4 & @xmath123 & 1.76@xmath186 & 2.03@xmath187 + & 0.22 kev flux ( 10@xmath174 erg s@xmath13 @xmath19 ) & 2.9@xmath188 & 6.5@xmath189 + & 210 kev flux ( 10@xmath174 erg s@xmath13 @xmath19 ) & 3.2@xmath190 & 4.3@xmath191 + & @xmath135 & 160/838 & 224/1060 + & & & + 5 & @xmath123 & 1.59@xmath192 & 2.33@xmath193 + & 0.22 kev flux ( 10@xmath174 erg s@xmath13 @xmath19 ) & 4.5@xmath194 & 5.0@xmath195 + & 210 kev flux ( 10@xmath174 erg s@xmath13 @xmath19 ) & 6.9@xmath196 & 1.8@xmath197 + & @xmath135 & 236/838 & 165/1060 + & & & + 6 & @xmath123 & 1.61@xmath198 & 1.81@xmath199 + & 0.22 kev flux ( 10@xmath174 erg s@xmath13 @xmath19 ) & 4.4@xmath200 & 4.6@xmath201 + & 210 kev flux ( 10@xmath174 erg s@xmath13 @xmath19 ) & 6.4@xmath202 & 4.7@xmath203 + & @xmath135 & 230/838 & 222/1060 + & & & + @xmath121 model 1 consists of a power - law modified by galactic absorption .
the x - ray fluxes represent unabsorbed values .
+ @xmath170all errors are for 90% confidence unless mentioned otherwise with all parameters taken to be of interest except absolute normalization . + | we present results from a monitoring campaign performed with the _ chandra x - ray observatory _ of the gravitationally lensed quasars rx j1131@xmath01231 and he 1104@xmath01805 .
we detect significant x - ray variability in all images of both quasars . the flux variability detected in image a of rx
j1131@xmath01231 is of particular interest because of its high amplitude ( a factor of @xmath1 20 ) .
we interpret it as arising from microlensing since the variability is uncorrelated with that of the other images and the x - ray flux ratios show larger changes than the optical as we would expect for microlensing of the more compact x - ray emission regions .
the differences between the x - ray and optical flux ratios of he 1104@xmath01805 are less dramatic , but there is no significant soft x - ray or dust absorption , implying the presence of x - ray microlensing in this system as well .
combining the x - ray data with the optical light curves we find that the x - ray emitting region of he 1104@xmath01805 is compact with a half - light radius @xmath2 , where the gravitational radius is r@xmath3 = 3.6 @xmath4 10@xmath5 cm , thus placing significant constraints on agn corona models .
we also find that the microlensing in he 1104@xmath01805 favors mass models for the lens galaxy that are dominated by dark matter .
finally , we better characterize the massive foreground cluster near rx j1131@xmath01231 , set limits on other sources of extended x - ray emission , and limit the fluxes of any central odd images to be 3050 ( 3@xmath6 ) times fainter than the observed images .
# 1 # 1 |
when @xcite compiled a list of relatively bright stars visible in the national geographic - palomar observatory sky survey ( poss ) that are associated with reflection nebulosity , he noted that the distance and luminosity for one such star , the @xmath11 cepheid su cas , might be estimated using spectroscopic and photometric observations of the nearby b - type stars hd 17138 and hd 17443 , which appear to illuminate a portion of the same dust complex .
his discovery was very important , given the complete lack in existing surveys of galactic clusters lying within several degrees of the cepheid that might serve as distance indicators .
curiously , the field of su cas was not surveyed in van den bergh s earlier search of the poss for previously - undetected star clusters @xcite .
the sparse group of b stars was subsequently designated as cas r2 , and used by @xcite with other r - associations to map local spiral structure in the galaxy . an initial estimate for the distance to the dust complex by @xcite yielded a value of @xmath12 pc from spectroscopic distance moduli for the three b - type stars
hd 17327 , hd 17443 , and hd 17706 , and the bright m giant hd 23475 .
a later study by @xcite using strmgren and h@xmath13 photometry implied that hd 17327 and hd 17443 were closer than hd 17706 , at a mean distance of @xmath14 pc , versus @xmath15 pc for the latter .
schmidt s result found support in an independent earlier survey by @xcite of b - type stars lying within @xmath16 of su cas .
only a few were found to lie at distances comparable to that expected for su cas , provided it was assumed to be either a fundamental mode or overtone pulsator .
a more detailed study of cas r2 stars by @xcite included an additional star , hd 16893 , associated with reflection nebulosity and an a - type companion to hd 17327 , yielding a small group of five stars , with five additional candidates , lying at an estimated distance of @xmath17 pc .
it was noted that background b - type stars were plentiful in the field , with most , including hd 17706 , congregating at distances of @xmath18 pc .
the inferred luminosity for su cas requires an estimate of its reddening , but the immediate field of the cepheid was surveyed by ( * ? ? ?
* see also @xcite ) , and the field reddening appeared to be fairly well established , although with some dependence on localized dust obscuration .
several of the surveyed companions to su cas were found to have intrinsic distance moduli @xmath19@xmath20 near 8.0 , but there were no suspicions at the time that the cepheid might be embedded in an anonymous open cluster .
the situation changed dramatically a few years ago when bruno alessi discovered a sparse cluster of stars surrounding su cas and centred at j2000.0 co - ordinates 02:52:15.1 , + 68:53:19 @xcite , as shown in fig .
the cluster , designated alessi 95 , has nuclear and coronal radii of @xmath21 and @xmath22 , respectively , and was discovered from an analysis of on - line databases ( see * ? ? ?
* ; * ? ? ?
the colour image ( fig . [ fig1 ] ) emphasises the large number of blue stars ( b and a - type ) dominating the central regions of the cluster , and also provides a clear picture of the relationship between su cas and its reflection nebula with the surrounding dust .
this study was initiated in order to provide additional information about alessi 95 its reddening , distance , and age based upon available observational material , and to discuss what it reveals about su cas .
for example , the revised _ hipparcos _
parallax of @xmath23 mas for su cas @xcite implies a distance of @xmath24 pc to the cepheid , consistent with intrinsic distance moduli near 8.0 found for several of its optical companions
. it may therefore be possible to establish the distance to su cas via trigonometric , open cluster , and pulsation parallaxes , which would solidly anchor the short - period end of the galactic calibration of the cepheid period - luminosity relation and strengthen the relationship established by @xcite from open clusters and associations .
an informative set of observations for bright b - type stars in the field of su cas was obtained four decades ago by @xcite , who made photoelectric _
_ measures for 31 of 33 stars lying within @xmath16 of the cepheid on two nights in september 1968 with the 0.9 m telescope at stewart observatory .
spectroscopic observations for 25 of the stars ( up to 3 spectrograms ) were obtained at a dispersion of 63 @xmath25 mm@xmath26 with the no . 1 , 0.9 m telescope at kitt peak national observatory ( kpno ) during the summer of 1968 .
the spectra were measured for radial velocity and classified by @xcite using the facilities of kpno , and also remeasured for radial velocity using the pds microdensitometer at the university of toronto , and reclassified by the lead author , for use in the study by @xcite .
part of the latter study included strmgren and h@xmath13 photometry for the stars obtained by forbes using the automatic photometer on the 0.76 m telescope of the behlen observatory at the university of nebraska , the same facility used in the study by @xcite .
.photometric and spectroscopic data for stars in the region of su cas . [ cols="<,^,^,^,<,^,^",options="header " , ] aro 3 = tyc 4313 - 355 , sp.t .
= f1 v.@xmath27 aro 4 , sp.t .
= a3 v. the colours of stars in the su cas field are plotted in fig .
[ fig3 ] , with filled circles used to denote stars lying in relatively close proximity to su cas .
reddenings have also been derived for four additional stars lying near the cepheid from _ bv _ data and spectral types by @xcite along with two stars from table [ tab2 ] below , four of the six appearing to lie at similar distances to su cas .
the derived mean reddening for the collection of 13 stars lying in close proximity to the cepheid , and not projected against an obvious dust cloud , is @xmath28 s.e .
( @xmath29 s.d . ) , which also appears to apply to a few stars in the table [ tab1 ] collection ( see mean reddening adopted in fig . [ fig3 ] ) .
there is noticeable differential reddening in the field according to the observations , that near su cas being associated with the visible dust clouds around the cepheid .
the mean reddening of the central regions of alessi 95 can be solidly established as @xmath28 s.e .
, however , with @xmath8 s.e . inferred for the space reddening of a star with the observed colours of su cas ( see * ? ? ?
the inferred reddening for the cepheid is a close match to an estimate of @xmath30 : established by @xcite from published spectrophotometric _
photometry for su cas in conjunction with intrinsic values established from cepheids of well - established space reddening .
the reddening also agrees reasonably well with an estimate of @xmath31 obtained by @xcite from stellar atmosphere model fitting , in this case linked to derived intrinsic colours and effective temperatures for bright stars of little to no reddening .
a reddening of @xmath32 was derived by @xcite from _ bvi_@xmath33 photometry for su cas linked to a calibration based on space reddenings for cepheids , but that included the earlier estimate of @xmath34 s.e . by @xcite .
given that su cas is the shortest period cepheid in the sample of pulsators with established space reddenings , the small offset of the @xcite reddening with the present result is presumably linked to the original underestimate of reddening for su cas obtained by @xcite .
potential members of alessi 95 were assembled from stars lying within the cluster boundaries : photoelectrically - observed stars ( table [ tab1 ] ) , stars in the _ hipparcos / tycho _ database @xcite , stars brighter than @xmath35 in the _ nomad _ database @xcite , recalibrated to the johnson _
system using faint stars from @xcite as reference standards , and stars near the cepheid with new ccd _ bv _ observations ( see below ) .
obvious `` ringers '' among the faint stars in _ nomad _ were omitted from the analysis , and the resulting data are plotted in fig . [ fig4 ] , which represents the colour - magnitude diagram for alessi 95 uncorrected for differential reddening or detailed membership selection . included
are a best - fitting zero - age main sequence ( zams , see below ) and a model isochrone from @xcite for @xmath36 , which appears to fit the data for cluster stars and the cepheid su cas reasonably well .
note that late - type dwarfs are only encountered in this direction at the distance of alessi 95 and beyond , so the cluster can not be less distant than implied from the zams fit , for example at the distance of @xmath37 pc derived for the foreground dust complex @xcite .
ccd observations of the central @xmath38 region surrounding su cas were made through johnson system _ bv _ filters in september 2011 with the sbig st8xme camera on the celestron 35-cm telescope of the robotic abbey ridge observatory ( see * ? ? ?
the observations were calibrated using previously published photoelectric _
photometry for stars in the field ( see * ? ? ?
* ; * ? ? ?
* and this paper ) , and are included in fig .
[ fig4 ] .
the reddening for individual stars was established using the reddening relation found from the spectroscopic observations ( fig .
[ fig2 ] ) through standard dereddening techniques ( see * ? ? ?
* ; * ? ? ?
* ) . for stars with _ bv _ data only
, colour excesses were inferred from the spatial trend of reddening across the field , except for those objects closely associated with the opaque dust clouds near the cepheid . in all cases the b0-star reddenings averaged for spatially - adjacent stars were adopted for individually dereddened stars , then adjusted for the colour dependence of reddening to that appropriate for the inferred intrinsic colour of each object ( see * ? ? ?
in a few cases it was possible to infer an independent reddening from the spectral classification for the star . each star was corrected for extinction using its inferred b0-star reddening in conjunction with the adopted value of @xmath39 .
the resulting data are plotted in fig .
[ fig5 ] along with similarly - derived data for six stars in table [ tab1 ] that appear to share comparable space motions and parallaxes with su cas .
observed radial velocities ( see * ? ? ?
* ; * ? ? ?
* ) , proper motions @xcite , and parallaxes @xcite were used to identify only _ bona fide _ potential outlying cluster members on the basis of similarity of the values to those for su cas .
cccc 11633 & bd+68@xmath40 170 & 2.31 & 1.46 + 12434 & hd 16228 & 2.80 & 1.23 + 12567 & hd 16393 & 3.03 & 0.52 + 12924 & hd 16841 & 3.08 & 1.01 + 13138 & bd+70@xmath40 205 & 1.04 & 1.02 + 13208 & hd 17327 & 0.98 & 0.77 + 13219 & bd+69@xmath40 181 & 1.82 & 1.20 + 13367 & su cas & 2.53 & 0.32 + 13465 & bd+68@xmath40 201 & 0.89 & 1.34 + 13595 & bd+69@xmath40 186 & 0.82 & 1.47 + 13728 & hd 17929 & 2.84 & 0.64 + 13956 & bd+69@xmath40 189 & 2.36 & 1.40 + 14442 & bd+70@xmath40 224 & 0.60 & 1.10 + 14483 & bd+67@xmath40 243 & 0.97 & 1.99 + 14714 & hd 19287 & 1.02 & 1.32 + 15138 & hd 19856 & 1.13 & 1.09 + 15959 & hd 20710 & 4.19 & 0.61 + 16633 & hd 21725 & 0.86 & 1.25 + [ tab3 ] the reddening for individual members of alessi 95 is well enough established for 26 likely zams members of the cluster to derive a mean intrinsic distance modulus of @xmath3 s.e .
( @xmath4 s.d . ) . the zero - age main sequence ( zams ) adopted here is that of @xcite , and the fit corresponds to a distance of @xmath5 pc , with the cited uncertainty representing the standard error of the mean .
a model isochrone for @xmath41 from @xcite fits the data reasonably well , with a likely uncertainty in @xmath42 no larger than @xmath43 .
it is possible to use alternate evolutionary isochrones from the literature , but they do not match the adopted zams nearly as well .
the isochrone fit for alessi 95 is not ideal for su cas , but that problem could be resolved by accounting for the opacity effects of cno mixing in the envelopes of post - supergiant stars ( see * ? ? ?
a more typical solution to the problem of a compressed blue loop for core helium - burning stars is to adopt a metallicity for the isochrone that is smaller than the solar value , but that does not appear to be justified in the case of alessi 95 , given that the derived metallicity of su cas ( and two associated stars ) from stellar atmosphere models is close to solar ( @xmath44[fe / h]@xmath45 = 0.12 , + 0.02 , 0.01 , and + 0.06 according to @xcite , respectively ) .
three of the bright members of alessi 95 , as well as 15 other stars lying within a few degrees of su cas sharing similar space motions ( proper motions and in a few cases radial velocities ) with the cepheid and with colours and magnitudes consistent with the @xmath36 isochrone in fig .
[ fig4 ] , are catalogued in the _ hipparcos _ catalogue @xcite .
the stars are collected in table [ tab3 ] , along with their cited parallaxes and uncertainties .
the weighted mean parallax for the group is @xmath46 mas , corresponding to a distance of @xmath47 pc .
an attempt was also made to include the less precise parallaxes from the _ tycho _ catalogue @xcite in the result .
a further fifteen stars were considered in such fashion , but the resulting mean parallax and distance remained unaffected , since the extra stars add negligible weight to the overall solution because of their large parallax uncertainties . in both cases
the parallax solution is in excellent agreement with the cluster distance of @xmath48 pc derived from zams fitting .
a further consistency check on the results was made using _
jhk_@xmath7 photometry @xcite from the two micron all sky survey ( 2mass , * ? ? ?
* ) for galactic star fields .
stars lying within @xmath22 of the centre of alessi 95 with proper motions similar to that of su cas are plotted in colour - colour and colour - magnitude diagrams in fig .
[ fig6 ] in the manner adopted by @xcite . also included are similar data for the stars from table [ tab1 ] and fig .
[ fig5 ] selected on the basis of comparable space motion with su cas .
most of the scatter in the observations can be attributed to photometric uncertainties , yet the data are a reasonably close match to the results inferred from the _ ubv _ analysis , namely the implied reddening and distance modulus .
a program was also completed in january 2012 to obtain _
jhk_@xmath7 photometry of greater precision for the field of alessi 95 using queue observing with the near - infrared imager ( cpapir ) of lobservatoire du mont - mgantic ( omm * ? ? ?
the details of that study , to be presented elsewhere , provide much stronger confirmation of the optical results than is the case for the less precise 2mass data , and also provide information on the extreme lower end of the cluster main sequence .
the nuclear and coronal radii for alessi 95 implied by its distance derived from zams fitting are 3 pc and @xmath49 pc , respectively .
both values are reasonable , although a larger tidal radius is expected , which would explain the presence of outlying cluster members a few degrees from su cas .
in the studies by @xcite and @xcite of stars associated with the dust complexes near su cas it was noted that the dust clouds lie at different distances .
the two main clouds at the declination of su cas and south of it are at distances of @xmath37 pc and @xmath50 pc according to the @xcite study .
su cas was assumed to lie at the distance of the former , according to a perceived connection on palomar observatory sky survey ( poss ) plates of the su cas dust complex with that associated with hd 16893 and hd 17443 , which were calculated to be 258 pc distant .
the deeper poss images used for the production of fig .
[ fig1 ] indicate that such an assumption can not be correct .
the reflection nebulae illuminated by su cas that lie south of the cepheid display no association with the opaque dust cloud east of it that continues further to the northeast .
strands of the same opaque cloud are seen to the west and south of su cas , connecting with the clouds illuminated by hd 16893 and hd 17443 , but the main cloud east of su cas displays no evidence for associated reflection nebulosity , so must lie foreground to the cepheid .
the true geometry of the stars and dust is revealed by those members of alessi 95 that , from their large reddenings , must be viewed _ through _ the dust extinction of the main cloud .
su cas and alessi 95 therefore lie beyond the main dust complex located at 258 pc .
indeed , the derived distance of @xmath51 pc to alessi 95 agrees closely with the previous estimate @xcite for the distance to the further dust complex in the field .
the derived luminosity for su cas as a member of alessi 95 , including the uncertainty in its field reddening , is @xmath9 s.e . , where the cepheid s magnitude @xcite has been adjusted for contamination by an unseen b - type companion 4.2 magnitudes fainter @xcite .
the result is more than a magnitude more luminous than expected for a classical cepheid with a mean period of @xmath52 days , as used in studies of the star s period changes @xcite .
the cepheid must therefore be an overtone pulsator , as argued previously ( e.g. , * ? ? ?
* ; * ? ? ?
according to the empirical relationship between fundamental mode and first overtone pulsation periods established for double - mode cepheids by @xcite , the undetected mean period for fundamental mode pulsation in su cas must be @xmath53 days .
an independent distance estimate for su cas is possible from its radius and inferred mean effective temperature , using the infrared surface brightness variant of the baade - wesselink method , as noted by @xcite .
the estimated distance to su cas in that study is @xmath54 pc for a pulsation period of @xmath55 . with the revised space reddening found here for su cas and a local ratio of total - to - selective extinction of @xmath39 ,
the derived distance becomes @xmath56 pc for fundamental mode pulsation ( @xmath57 ) .
the distance estimates are not completely independent , but the methodologies are .
the close agreement in the pulsation parallax , trigonometric parallax , and cluster parallax estimates for the distance to alessi 95 and su cas @xmath58 pc , @xmath59 pc , and @xmath51 pc , respectively , all of which agree to within their derived uncertainties provides strong confirmation of their validity .
the implied mean radius of su cas according to the cepheid period - radius relation of @xcite , which is tied to the almost identical slopes for the cepheid period - radius relation in studies by @xcite , @xcite , @xcite , and @xcite , is 25.0 @xmath60 . by comparison
, the surface brightness technique yields a mean radius of @xmath61 @xmath60 , while an independent baade - wesselink analysis using infrared colours by @xcite produced an estimate of @xmath62 @xmath60 .
the last study also summarizes previous estimates for the mean radius of su cas derived from variants of the baade - wesselink method , most of which lie in the range 3045 @xmath60 .
systematic effects can not be discounted , given the small amplitude of the light variations in su cas and the fact that @xcite adopted a projection factor of @xmath63 in their study .
@xcite find a value of @xmath64 to be more suitable for su cas , which would reduce the @xcite value to @xmath65 @xmath60 .
the _ vjk _
photometry used by @xcite leads to a mean radius of @xmath66 @xmath60 for su cas , while @xcite obtained a radius of @xmath67 @xmath60 using a baade - wesselink analysis tied to _
_ photometry .
the last three estimates are closer to the value predicted by the cepheid period - radius relation , and it may be that the star s small pulsation amplitude and contamination by its b - type companion limit further improvement . with the period of fundamental mode pulsation in su cas indicated by its membership in alessi 95 , one can estimate independently the age of the cepheid from existing period - age relations . a model - based relationship derived by @xcite yields an age of @xmath68 for a solar metallicity cepheid with @xmath57 , while a relationship by @xcite calibrated by isochrone fits to open clusters yields a similar age of @xmath69 .
both cited relationships display a dispersion in @xmath42 of @xmath70 , so are consistent with the cluster age inferred from the model isochrones of @xcite .
additionally , the implied age of @xmath41 for su cas from its membership in alessi 95 provides a key point at the short - period end of a semi - empirical period - age relationship for cluster cepheids developed by @xcite , where again the dispersion is no larger than @xmath43 in @xmath42 and the relationship is generated by isochrone fits to clusters containing galactic cepheids .
the luminosity of su cas inferred from cluster membership , namely @xmath9 , can be compared with a value of @xmath71 predicted from its inferred radius and effective temperature via the methodology described by @xcite and @xcite .
the @xmath72 offset from the empirical estimate is comparable to the offsets observed for other cluster cepheids @xcite , and can be partially explained by the nature of su cas : a small - amplitude cepheid lying near the high temperature side of the instability strip , as also argued by its overtone pulsation .
su cas has an unseen b9.5 v companion detected by iue @xcite , and the luminosity inferred for the cepheid according to the implied magnitude difference between su cas and the b - star in ultraviolet spectra is @xmath73 @xcite , closely coincident with the result from cluster membership . the companion , which may also be an ap star @xcite , has otherwise only a small effect on the overall visual brightness and colours of su cas .
confirmation of the location of su cas towards the blue edge of the cepheid instability strip is provided by its observed rate of period change of + 0.024 s yr@xmath26 @xcite .
that implies a rate of evolution for the cepheid that is more than twice as rapid as what is typical of fundamental mode pulsators with periods of @xmath74 near the centre of the instability strip @xcite .
the consistency in the implied parameters for su cas found from such diverse observational material provides further validation of the present results .
this first detailed photometric and spectroscopic study of alessi 95 , the sparse open cluster surrounding the cepheid su cas , produces empirical estimates for the reddening and luminosity of a _ bona fide _ cepheid calibrator .
estimates for the distance to su cas and alessi 95 from zams fitting , _ hipparcos _ parallaxes @xcite , and the cepheid s pulsation parallax ( * ? ? ?
* adjusted here ) agree closely to within a few percent , to our knowledge the first such instance of a tight consensus in distance estimates by these diverse methods .
future improvement to the cluster zams fit might be possible , for example through a more detailed analysis that accounts implicitly for the metallicity of su cas and alessi 95 members .
that affects zams fitting through small offsets in the zero - point , which is currently calibrated for solar metallicity stars .
given present knowledge of the near - solar chemical composition for alessi 95 stars @xcite , however , only a very minor improvement would be expected .
the inferred field reddening of @xmath8 found here for su cas agrees reasonably well with other independent estimates @xcite , and the effective temperature of 6620k for su cas inferred from its colour via the semi - empirical technique described by @xcite and @xcite is only slightly greater than the values obtained by @xcite from stellar atmosphere models .
any discrepancies between the present results and those of previous studies appear to be minor , except for the inferred luminosity of @xmath9 s.e . , which differs from that found in the earlier study by @xcite , primarily because of an erroneous selection of reflection nebulosity stars associated with the cepheid in that paper .
future work may improve the situation , since _
parallaxes , pulsation parallaxes , and cluster parallaxes are all susceptible to systematic effects that are the subject of ongoing study
. it should be evident , however , that there is no longer a need to rely on the cepheid s membership in cas r2 to derive its intrinsic parameters ( e.g. , * ? ? ?
* ; * ? ? ?
alessi b. s. , 2006 , unpublished data alessi b. s. , moitinho a. , dias w. s. , 2003 , a&a , 410 , 565 andrievsky s. m. , kovtyukh v. v. , luck r. e. , et al . , 2002 , a&a , 381 , 32 artigau .
m. , lamontagne r. , doyon r. , malo l. , 2010 , in silva d. r. , peck a. b. & soifer b. t. , eds .
, observatory operations : strategies , processes , and systems iii , proceedings of the spie , 7737 , pp .
1o 1 - 7 aveni a. f. , hunter j. h. , 1972 , aj , 77 , 17 berdnikov l. n. , 2007 , http://www.sai.msu.ru/groups/ cluster / cep / phe berdnikov l. n. , ignatova v. v. , pastukhova e. n. , turner d. g. , 1997 , astl , 23 , 177 berdnikov l. n. , mattei j. a. , beck s. j. , 2003 , jaavso , 31 , 146 bono g. , marconi m. , cassisi s. , et al . , 2005 , apj , 621 , 966 cutri r. m. et al . , 2003 , the irsa 2mass all - sky point source catalog of point sources , nasa / ipac infrared science archive efremov yu . n. , elmegreen b. g. , 1998 , mnras , 299 , 588 esa , 1997 , _ the hipparcos and tycho catalogues _ , esa sp-1200 evans n. r. , arellano ferro a. , 1987 , in stellar pulsation , springer - verlag , berlin , lecture notes in physics , 274 , p. 183
feltz k. a. , jr . , mcnamara d. h. , 1976 , pasp , 88 , 699 fernie j. d. , 1963 , aj , 68 , 780 fitzgerald m. p. , 1970 , a&a , 4 , 234 gieren w. , 1976 , a&a , 47 , 211 gieren w. , 1982 , pasp , 94 , 960 gieren w. p. , barnes t. g. , iii , moffett t. j. , 1989 , apj , 342 , 467 gieren w. p. , fouqu p. , gmez m. , 1998 , apj , 496 , 17 johnson h. l. , 1966 , ara&a , 4 , 193 kovtyukh v. v. , andrievsky s. m. , usenko i. a. , klochkova v. g. , 1996 , a&a , 316 , 155 kovtyukh v. v. , soubiran c. , luck r. e. , et al . , 2008 , mnras , 389 , 1336 kronberger m. , teutsch p. , alessi b. , et al . , 2006 , a&a , 447 , 921 havlen r. j. , 1971 , ph.d .
thesis , univ .
arizona herbst w. , sawyer d. l. , 1981 , apj , 243 , 935 lane d. j. , 2008 , jaavso , 36 , 143 laney c. d. , stobie r. s. , 1995 , mnras , 274 , 337 laney c. d. , caldwell j. a. r. , 2007 , mnras , 377 , 147 laney c. d. , joner m. d. , 2009 , in guzik j. a. & bradley p. a. , eds .
, stellar pulsation : challenges for theory and observation , melville , new york , aip conf .
proc . , 1170 , p. 93
luck r. e. , andrievsky s. m. , fokin a. , kovtyukh v. v. , 2008 , aj , 136 , 98 meynet g. , mermilliod j .- c .
, maeder a. , 1993 , a&as , 98 , 477 milone e. f. , wilson w. j. f. , volk k. , 1999 , aj , 118 , 3016 racine r. , 1968 , aj , 83 , 960 racine r. , van den bergh s. , 1970 , in becker w. & kontopoulos g. i. , eds .
, the spiral structure of our galaxy , reidel , dordrecht , iau symp . , 38 , p. 219
schild r. , romanishin w. , 1976 , apj , 204 , 493 schmidt e. g. , 1978 , aj , 73 , 588 skrutskie m. f. et al .
, 2006 , aj , 131 , 1163 storm j. , gieren w. , fouqu p. , et al . , 2011 , a&a , 534 , a94 szabados l. , 1988 , in kovcs g. , szabados l. & szeidl b. , eds . , multimode stellar pulsations , konkoly obs . , budapest , p. 1
turner d. g. , 1976a , aj , 81 , 97 turner d. g. , 1976b , aj , 81 , 1125 turner d. g. , 1979 , pasp , 91 , 642 turner d. g. , 1984 , jrasc , 78 , 229 turner d. g. , 1989 , aj , 98 , 2300 turner d. g. , 1990 , pasp , 102 , 1331 turner d. g. , 1996 , in milone e. f. & mermilliod j .- c . , eds . , the origins , evolution , and destinies of binary stars in clusters , aspc , 90 , p. 382
turner d. g. , 2003 , in gray r. o. , corbally c. j & philip a. g. d. , eds . , the garrison festschrift , l. davis press , schenectady , n.y . , p. 101
turner d. g. , 2010 , ap&ss , 326 , 219 turner d. g. , 2011 , rmxaa , 47 , 127 turner d. g. , 2012 , jaavso , 40 , in press turner d. g. , evans n. r. , 1984 , apj , 283 , 254 turner d. g. , burke j. f. , 2002 , aj , 124 , 2931 turner d. g. , forbes d. w. , lyons r. w. , havlen r. j. , 1985 , in madore b. f. , ed .
, cepheids : theory and observations , iau colloq .
turner d. g. , leonard p. j. t. , english d. a. , 1987 , aj , 93 , 368 turner d. g. , abdel - sabour abdel - latif m. , berdnikov l. n. , 2006 , pasp , 118 , 410 turner d. g. , majaess d. j. , lane d. j. , et al . , 2010 , odessa astron . publ . , 23 , 119 usenko i. a. , kovtyukh v. v. , klochkova v. g. , et al . , 2001 , a&a , 367 , 831 van den bergh s. , 1957 , apj , 126 , 323 van den bergh s. , 1966 , aj , 71 , 990 van leeuwen f. , 2007 , _ hipparcos , the new reduction of the raw data _ , springer , heidelberg , astrophysics and space science library , 370 xu h. y. , li y. , 2004 , a&a , 418 , 213 zacharias n. , monet d. , levine s. , urban s. , gaume r. , wycoff g. , 2005 , _ the naval observatory merged astrometric dataset ( nomad ) _ , baas , 36 , 1418 | the parameters for the newly - discovered open cluster alessi 95 are established on the basis of available photometric and spectroscopic data , in conjunction with new observations .
colour excesses for spectroscopically - observed b and a - type stars near su cas follow a reddening relation described by @xmath0 , implying a value of @xmath1 for the associated dust .
alessi 95 has a mean reddening of @xmath2 s.e .
, an intrinsic distance modulus of @xmath3 s.e .
( @xmath4 s.d . ) , @xmath5 pc , and an estimated age of @xmath6 yr from zams fitting of available _ ubv _ , ccd _ bv _ , _ nomad _ , and 2mass _ jhk_@xmath7 observations of cluster stars .
su cas is a likely cluster member , with an inferred space reddening of @xmath8 and a luminosity of @xmath9 s.e . ,
consistent with overtone pulsation ( @xmath10 ) , as also implied by the cepheid s light curve parameters , rate of period increase , and _ hipparcos _ parallaxes for cluster stars .
there is excellent agreement of the distance estimates for su cas inferred from cluster zams fitting , its pulsation parallax derived from the infrared surface brightness technique , and _ hipparcos _ parallaxes , which all agree to within a few percent . |
the dark halo substructure has recently come to one of the most lively topics in cosmology .
although the standard cosmological paradigm based on the cold dark matter ( cdm ) concept generically predicts the presence of the substructure inside the dark matter halos , there are many questions yet to be answered associated with the dark halo substructures .
the intrinsic alignment effect of the dark halo substructure is one of those questions .
there are plenty of observational evidences that the major axes of the brightest cluster galaxies ( bcgs ) have a strong tendency to be aligned with that of their host clusters @xcite .
the most popular theory for the bcg alignment is the anisotropic infall scenario based on the standard hierarchical clustering model @xcite : the initial density field of cdm is web - like , interconnected by the primordial filaments @xcite .
the gravitational collapse and merging to form structures occurs not in an isotropic way but in an anisotropic way along the large - scale filaments .
accordingly , the infall of materials into a cluster also occurs along the primordial filament , which will induce the alignment between the orientation of a host cluster and that of bcg embedded in it .
there are several reasons that the anisotropic infall theory became so popular : being simple and intuitive , it fits very well into the cold dark matter paradigm .
in addition , it has been supported by several numerical simulations ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) which demonstrated that the gravitational infall and merging of materials indeed occurs along the filaments . nevertheless , the theory is only qualitative and still incomplete .
recent observations indicate that not only the bcgs but also the less dominant cluster galaxies exhibit the alignment effect to a non - negligible degree @xcite . in the anisotropic infall model ,
the substructure alignment is a primordial effect , and would get damped away quickly by the subsequent nonlinear processes such as the violent relaxation , the secondary infall , and so on @xcite .
therefore , it is very unlikely that the cluster galaxies other than the bcgs keep the primordial alignment effect till the present epoch @xcite . here ,
we propose that the initial tidal interaction between the subhalos and the host halo will be responsible for the observed intrinsic alignment of the cluster galaxies .
when a subhalo forms inside a host halo , it acquires the angular momentum @xmath0 due to the tidal shear field @xmath1 generated by the gravitational potential of the host halo @xmath2 : @xmath3 .
@xcite proposed the following formula to quantify the mutual correlations between @xmath4 and @xmath5 : @xmath6 where @xmath7 $ ] is a correlation parameter to quantify the strength of the correlation between @xmath8 and @xmath5 , and @xmath9 is a unit traceless tidal shear tensor defined as @xmath10 with @xmath11 , and @xmath12 is a rescaled but not a unit angular momentum . if we replace the rescaled angular momentum by the unit angular momentum , @xmath13 in equation ( [ eqn : spin ] ) , then the correlation parameter @xmath14 is reduced by a factor of @xmath15 . note here that the lhs of equation ( [ eqn : spin ] ) represents a _
conditional _ ensemble average of @xmath16 provided that the unit traceless tidal shear tensor is given as @xmath17 . for the detailed explanations of equation ( [ eqn : spin ] ) ,
see appendix a in @xcite .
it is naturally expected that @xmath14 decreases with time as the correlation between @xmath8 and @xmath5 must decrease after the moment of the turn - around due to the subsequent nonlinear process .
@xcite found @xmath18 at present epoch by analyzing the data from the tully galaxy catalog and the point source redshift catalog redshift survey ( in their original work , they used a reduced correlation parameter @xmath19 and found @xmath20 ) .
strictly speaking , equation ( [ eqn : spin ] ) holds only if @xmath8 and @xmath5 are defined at the same positions @xcite . here ,
they do nt : @xmath8 and @xmath5 are defined at the centers of the mass of the host halo and the subhalo , respectively .
for simplicity , here we just assume that equation ( [ eqn : spin ] ) still holds , ignoring the separation between the centers of the mass of the subhalos and that of the host halo .
the distribution of @xmath5 under the influence of the tidal field is often regarded as gaussian @xcite : @xmath21^{1/2 } } \exp\left[-\frac{l_{i}(m^{-1})_{ij}l_{j}}{2}\right],\ ] ] where the covariance matrix @xmath22 is related to the tidal shear field @xmath8 by equation ( [ eqn : spin ] ) . in the principal axis frame of @xmath8 ,
let us express @xmath5 in terms of the spherical coordinates : @xmath23 where @xmath24 , and @xmath25 and @xmath26 are the polar and the azimuthal angles of @xmath5 .
note that the polar angle @xmath25 represents the angle between the direction of @xmath5 of the subhalo and the minor principal axis of @xmath8 of its host .
the probability density distribution of the cosines of the polar angle @xmath25 can be obtained by integrating out equation ( [ eqn : ldis ] ) over @xmath27 and @xmath26 @xcite : @xmath28 here the polar angle @xmath25 is forced to be in the range of @xmath29 $ ] satisfying @xmath30 , since we care about the relative spatial orientation of the subhalo axis , but not its sign . here
the three @xmath31 s ( @xmath32 ) are the eigenvalues of @xmath8 in a decreasing order satisfying the following two conditions : ( i ) @xmath33 ; ( ii ) @xmath34 . if @xmath4 is a gaussian random field which is true in the linear regime , one can show that @xmath35 and @xmath36 @xcite .
we adopt the following two assumptions : ( i ) on cluster scale , the principal axes of the inertia shape tensor @xmath37 of a host halo is aligned with its tidal shear tensor @xmath38 with the eigenvalues being in an opposite order . in other words , the major principal axis of @xmath37 is the minor principal axis of @xmath38 , and vice versa .
note that in @xcite , it was erroneously stated that it the major axis of @xmath37 is the major axis of @xmath38 ( trujillo , carretero , & juncosa 2004 in private communication ) ; ( ii ) the minor axis of a subhalo is in the direction of its angular momentum .
a justification of the first assumption is given by the zeldovich approximation @xcite which predicts a prefect alignment between the principal axis of @xmath37 and that of @xmath38 .
since the cluster - size halos are believed to be in the quasi - linear regime where the zeldovich approximation is valid , the first assumption should provide a good approximation to the reality .
moreover , recent n - body simulations indeed demonstrated that @xmath37 and @xmath38 are quite strongly correlated @xcite . regarding the second assumption
, there is an established theory that the spin axis of an ellipsoidal object in the gravitational tidal field is well correlated with its minor axis @xcite , which was also confirmed by several n - body simulations ( e.g. , * ? ? ?
* ) now that the minor principal axis of @xmath8 is the major axis of the host halo , and @xmath5 is aligned with the minor axis of the subhalo , the polar angle @xmath25 in equation ( [ eqn : vtdis ] ) actually equals _ the angle between the minor axis of the subhalo and the major axis of the host halo_. putting @xmath39 , @xmath40 , and @xmath41 , we simplify equation ( [ eqn : vtdis ] ) into @xmath42^{-3/2}d\phi,\ ] ] in the asymptotic limit of @xmath43 , equation ( [ eqn : tdis ] ) can be further simplified into the following closed form : @xmath44 equations ( [ eqn : tdis ] ) and ( [ eqn : tdisa ] ) imply that @xmath45 increases as @xmath25 increases .
that is , the minor axis of a subhalo has a strong propensity to be _ anti - aligned _ with the major axis of its host halo .
hence , it explains the observed alignment effect between the subhalo and the host halo major axes as a consequence of the intrinsic anti - alignment between the subhalo minor axis and the host halo major axis .
the value of @xmath14 in equations ( [ eqn : tdis ] ) and ( [ eqn : tdisa ] ) should depend on the distance from the host halo center ( @xmath46 ) , the subhalo mass ( @xmath47 ) and redshift ( @xmath48 ) : @xmath49 .
what one can naturally expect is that @xmath14 should decrease with @xmath46 since the tidal interaction must be strongest in the inner part of the host halo , and that @xmath14 should increase with @xmath47 and @xmath46 since the alignment effect gets reduced in the nonlinear regime .
unfortunately , it will be very difficult to find thefunctional form of @xmath50 as @xmath14 contains all the nonlinear informations of galaxy evolution .
we do not attempt to find @xmath51 here since it is beyond the scope of this letter .
instead , we simplely assume that @xmath14 is a constant , and determine the value of @xmath14 empirically by fitting equation ( [ eqn : tdis ] ) to the numerical results in @xmath52 .
the data we use in this letter is the high resolution halo simulations of @xcite .
first they selected dark matter halos from their previous cosmological p@xmath53 m n - body simulations with @xmath54 particles in a @xmath55mpc cube @xcite .
the halos were identified using the standard friend - of - friend ( fof ) algorithm , among which four halos on cluster - mass scales ( with mass around @xmath56 ) were then re - simulated using the nested - grid p@xmath53 m code which was designed to simulate high - resolution halos .
the force resolution is typically @xmath57 of the virial radius , and each halo is represented by about @xmath58 particles within the virial radius .
we then use the subfind routine of @xcite to identify the disjoint self - bound subhalos within these halos , and include those subhalos containing more than 10 particles in the analysis .
these simulations adopted the `` concordance '' @xmath59cdm cosmology with @xmath60 , @xmath61 , and @xmath62 . using this numerical data , we first compute the inertia tensors as @xmath63 for each host halo and its subhalos in their respective center - of - mass frames .
then , we find the directions of the eigenvectors corresponding to the largest and the smallest eigenvalues of the host halo and the subhalo inertia tensors , respectively , by rotating the inertia tensors into the principal axes frame , and determine the major axes of each host halo and the minor axis of its subhalos .
then , we measure the cosines of the angles , @xmath25 , between the major axis of the host halo and the minor axes of their subhalos by computing @xmath64 where @xmath65 and @xmath66 represent the major and the minor axes of the host halo and its subhalos , respectively . finally , we find the probability density distribution , @xmath67 , by counting the number density of the subhalos .
when computing the probability density distribution , we use all the subhalos in the host halo linked by the fof algorithm .
we perform the above procedure at four different redshifts : @xmath68 and @xmath69 .
the total number of the subhalos @xmath70 at each redshift is @xmath71 , @xmath72 , @xmath73 , and @xmath74 , respectively .
figure [ fig : dis1 ] plots the final numerical distributions ( solid dots ) with the error bars .
the error bar at each bin is nothing but the poisson mean for the case of no alignment given as given as @xmath75 where @xmath76 is the number of the subhalos at each bin .
as one can see , the numerical distribution @xmath67 increases as @xmath25 increases , revealing that the minor axes of substructures really tend to be anti - aligned with the major axes of their host halos , as predicted by the analytic model ( eq.[[eqn : tdis ] ] ) of @xmath77 .
figure [ fig : dis1 ] also plots the analytic predictions ( solid line ) and the approximation formula ( dashed line ) derived in @xmath77 .
the horizontal dotted line represents the uniform distribution of @xmath78 for the case of no alignment .
we fit the analytic distributions to the numerical data points to determine the best - fit values of the correlation parameter @xmath14 .
we find @xmath79 at @xmath80 , respectively .
the errors involved in the determination of @xmath14 is given as the standard deviation of @xmath14 for the case of no alignment effect given as @xmath81 where @xmath70 is the number of all subhalos used to compute @xmath14 .
the value of @xmath14 increases with redshifts @xmath48 , as expected .
in fact the value of @xmath82 gives quite a good fit , if not the best , not only at the present epoch of @xmath83 but also at all earlier epochs of @xmath84 , which implies that the initially induced anti - alignment effect is more or less conserved , reflecting the fact that the directions of the subhalo angular momentum are fairly well conserved .
to understand the dependence of the alignment effect on the subhalo mass , we derive the same probability distribution but by using only the most massive 30 subhalos in each cluster ( @xmath85 for each redshift ) , and find the corresponding best - fit values of @xmath14 .
we find @xmath86 at @xmath80 , respectively .
figure [ fig : dis2 ] plots the results .
the approximation formula ( eq.[[eqn : tdisa ] ] ) is excluded in this figure since the best - fit values of @xmath14 for this case is pretty close to unity .
although the large error bars prevent us from making a quantitative statement , it is obvious that the anti - alignment effect is stronger for the case massive subhalos .
although the currently popular anisotropic merging and infall scenario has provided a qualitative explanation for the bcg - cluster and cluster - cluster alignments ( e.g. , * ? ? ?
* ) , no previous approach based on this scenario was capable of making a quantitative prediction for the alignment effect of cluster galaxies other than bcgs with the host clusters .
we constructed an analytic model where the substructure alignment is a consequence of the tidal field of the host halo at least on the scale of cluster galaxies , and predicted quantitatively the strength of the alignment effect , by comparing the model with the results from recent high - resolution n - body simulations .
however , it is worth noting that we have yet to completely rule out the anisotropic infall model . an idealistic way to discriminate our analytic model from the anisotropic infall scenario would be to measure directly the correlation of the directions of the subhalo angular momentum with the host halo orientations
unfortunately , it is very difficult to determine the directions of the subhalo angular momentum vectors in current simulations . because the rotation speed of a dark halo in simulations is only a few percent of the virial motion of the particles , one needs more than @xmath87 particles to determine accurately the direction of the angular momentum vector of a subhalo with an average rotation speed . in current simulations the subhalos have much fewer particles than @xmath87
this is why we used the minor axes of the subhalos rather than the directions of the subhalo angular momentum vectors to investigate the intrinsic alignments of substructures .
nevertheless , the strong alignment between the halo minor axes and angular momentum vectors @xcite should indicate indirectly that the anti - alignments between the subhalo minor axes and the host halo major axes observed in our simulations are likely to be caused by the host halo tidal field as our model predicts .
many n - body simulations @xcite have already proved that the dark matter halos rotate and their angular momentum vectors are aligned with their minor axes .
furthermore , @xcite demonstrated evidently that there are good internal alignments between the halo minor axes and angular momentum vectors measured at different radii .
therefore , the alignments between the halo angular momentum vectors and the minor axes are expected to be hold even when the outer parts of the halos get disrupted when they fall into larger halos as substructures .
indeed , we ourselves check this effect in our simulations : we measure the angular momentum vectors of several very massive subhalos with more than @xmath88 particles , and find that the subhalos have _ non - zero _ angular momentum and that the subhalo minor axes are strongly aligned the directions of their angular momentum : the cosines of all alignment angles turn out to be bigger than @xmath89 .
anyway , it will be definitely necessary to investigate directly the correlations of the directions of the subhalo angular momentum with the orientations of their host halos in the future with larger simulation data , where the number of particles belonging to subhalos should be large enough . using larger simulation data
, it will be also possible to determine the functional form of the correlation parameter @xmath90 .
our future work is in thisdirection .
we thank the anonymous referee who helped us improve the original manuscript .
j.l . wishes to thank the shanghai astronomical observatory for a warm hospitality during the workshop on cosmology and galaxy formation where this collaboration began .
is supported by the research grant of the korea institute for advanced study .
are partly supported by nkbrsf ( g19990754 ) , by nsfc(nos .
10125314 , 10373012 ) , and by shanghai key projects in basic research ( no . 04jc14079 ) adams , m. t. , strom , k. m. , & strom , s. f. 1980 , , 238 , 445 bailin , j. , & steinmetz , m. 2004 , , in press , ( astro - ph/0408163 ) bailin , j. et al . 2005 , , in press , astro - ph/0505523 bardeen , j. m. , bond , j. r. , kaiser , n. , & szalay , a. s. 1986 , , 304 , 15 barnes , j. , & efstathiou , g. 1987 , , 319 , 575 binggeli , b. 1982 , , 107 , 338 binney , j. , & tremaine , s. 1987 , galactic dynamics , ( princeton : princeton univ . press ) binggeli , b. 1982 , , 107 , 338 bond , j. r. 1987 , in nearly normal galaxies , ed .
s. faber ( new york : springer ) , 388 bond , j. , r. , kofman , l. , & pogosyan , d. 1996 , nature , 380 , 603 carter , d. , & metcalfe , j. 1980 , , 191 , 325 catelan , p. , & theuns , t. 1996 , , 282 , 436 coutts , a. 1996 , , 278 , 87 dubinski , j. 1992 , 401 , 441 dubinski , j. 1998 , 502 , 141 faltenbacher , a. , kerscher , m. , gottloeber , s. , & mueller , m. 2002 , , 395 , 1 fuller , t. m. , west , m. j. , & bridges , t. j. 1999 , , 519 , 22 hopkins , p. f. , bahcall , n. , & bode , p. 2005
, in press jing , y. p. & suto , y. 1998 , apj , 503 , l9 jing , y. p. & suto , y. 2000 , apj , 529 , l69 kim , r. s. j. , annis , j. , strauss , m. a. , lupton , r. h. , bahcall , n. a. , cunn , j. e. , kepner , j. v. , & postman , m. 2002 , asp conf.ser .
268 , 393 klypin , a. , gottl@xmath91ber , s. , kravtsov , a. v. , & khokhlov , a. m. 1999 , , 516 , 530 lee , j. & pen , u. l. 2000 , , 532 , 5 lee , j. & pen , u. l. 2001 , , 555 , 106 lee , j. & pen , u. l. 2002 , , 567 , l111 lee , j. 2004 , , 614 , l1 pen , u. l. , lee , j. , & seljak , u. 2000 , , 543 , l107 pereira , m. j. , & kuhn , j. r. 2004 , preprint [ astro - ph/0411710 ] plionis , m. , barrow , j. d. , & frenk , c. s. 1991 , , 249 , 662 plionis , m. 1994 , apjs , 95 , 401 plionis , m. , & basilakos , s. 2002 , , 329 , l47 plionis , m. , benoist , c. , & maurogordato , s. , ferrai , c. , basilakos , s. 2003 , , 594 , 153 porciani , c. , dekel , a. , & hoffman , y. 2002 , , 332 , 325 quinn , t. , & binney , j. 1992 , , 255 , 729 rhee , g. f. r. n. , & katgert , p. 1987
, , 183 , 217 sastry , g. n. 1968 , pasp , 83 , 313 springel , v. , white , s. d. m. , tormen , g. , kauffmann , g. 2001 , , 328 , 726 struble , m. f. , & peebles , p. j. e. 1985 , , 99 , 743 van haarlem , m. , & van de weygaert , r. 1993 , , 418 , 544 van kampen , e. , & rhee , g. f. r. n. 1990 , , 237 , 283 warren , m. s. , quinn , p. j. , salmon , j. k. , & zurek , w. h. 1992 , , 399 , 405 west , m. j. 1989 , , 347 , 610 west , m. j. 1994 , , 268 , 79 west , m. j. , villumsen , c. , & dekel , a. 1991 , , 369 , 287 west , m. j. , jones , c. , & forman , w. 1995 , , 451 , l5 zeldovich , y. b. 1970 , a & a , 5 , 84 | we investigate the intrinsic alignments of dark halo substructures with their host halo major - axis orientations both analytically and numerically .
analytically , we derive the probability density distribution of the angles between the minor axes of the substructures and the major axes of their host halos from the physical principles , under the assumption that the substructure alignment on galaxy scale is a consequence of the tidal fields of the host halo gravitational potential .
numerically , we use a sample of four cluster - scale halos and their galaxy - scale substructures from recent high - resolution n - body simulations to measure the probability density distribution .
we compare the numerical distribution with the analytic prediction , and find that the two results agree with each other very well .
we conclude that our analytic model provides a quantitative physical explanation for the intrinsic alignment of dark halo substructures .
we also discuss the possibility of discriminating our model from the anisotropic infall scenario by testing it against very large n - body simulations in the future . |
time reversal ( t ) is a fundamental symmetry in particle physics and observation of t - asymmetry plays a crucial role to investigate the property of theoretical aspects .
babar collaboration announced their result@xcite that they observed the t - asymmetry through b - meson mixing system .
the idea of taking t - asymmetry is based on @xcite-@xcite .
babar stated that the observation is the first evidence of t - violation in b - system , and babar observable naively seems to be t - asymmetry .
however , discussion in ref.@xcite shows that babar observable slightly deviates from a t - odd quantity . in this article
, we construct an asymmetry which consists of @xmath4 meson system . the difference between babar asymmetry and the one in this article
is stated .
the mixing induced cp violation of kaon system contributes to the asymmetry so that one needs to take account of such mixing effect .
the contribution from @xmath3 to an observable is extracted .
the formulation in this article is independent from phase redefinition of quark fields . for further detail
, one can refer the paper@xcite .
in this section , an event number asymmetry is constructed and one can check how it behaves under t - transformation . using processes which come from @xmath5 decay ,
the following processes in @xmath4 system are considered , @xmath6 where @xmath7 indicates the cp odd state of @xmath4 meson .
the processes ( a ) and ( b ) are naively related with time reversal since the time direction is opposite to each other . in fig.[fig1 ] , the entire processes of ( a ) and ( b ) are shown respectively .
( lower ) process ( b ) in eq.(2.1),title="fig : " ] ( lower ) process ( b ) in eq.(2.1),title="fig : " ] hereafter , we explain how b mesons in ( a ) are identified . in ( a )
, cp tagging is conducted at time @xmath8 to determine @xmath9 ( cp even b meson ) , and one also identify @xmath7 due to einstein podolsky rosen ( epr ) correlation from @xmath5 decay . at time
@xmath10 , b meson decays into @xmath11 and the flavor tagging method enables us to determine @xmath12 .
hence , one can select the event of @xmath13 .
similarly , one can obtain events of ( b ) with flavor tagging and cp tagging .
time dependent asymmetry is defined as , @xmath14 where @xmath15 and @xmath16 are time dependent decay rates for the processes ( a ) and ( b ) in eq.([2.1 ] ) .
hence , the asymmetry above is naively thought to be t - odd quantity . here , time dependent decay rate is given as , @xmath17,&\label{2.3}\\ & t = t_2-t_1,\quad \gamma=\displaystyle\frac{\gamma_h+\gamma_l}{2},\quad x=\displaystyle\frac{m_h - m_l}{\gamma},\quad y=\displaystyle\frac{\gamma_h-\gamma_l}{2\gamma},&\nonumber\end{aligned}\ ] ] where the coefficients @xmath18 and @xmath19 are given in @xcite@xcite and we do not specify the expression of these coefficients in this article . here
the following quantities are introduced , @xmath20 where @xmath21 in eq.([2.5 ] ) is ratio of overall factors for the time dependent decay rates in eq.([2.3 ] ) .
@xmath22 shows the effect of non - zero difference of the overall factors in eq.([2.3 ] ) .
@xmath22 consists of small parameters such like mixing induced cp violation of @xmath4 system and cp violation in @xmath23 decay and @xmath22 is evaluated in ref.@xcite .
babar collaboration measured the processes : @xmath24 and @xmath25 .
babar - constructed asymmetry eliminates the effect of overall factors so that the definition of babar asymmetry is not identical to the event number asymmetry in eq.([2.2 ] ) . in ref.@xcite ,
babar asymmetry@xcite is analyzed . in this article
, we investigate the event number asymmetry so that the contribution from @xmath22 in eq.([2.5 ] ) needs to be taken account and the contribution from @xmath3 is calculated . with the time dependent decay rate in eq.([2.3 ] ) , the asymmetry in eq.([2.2 ] ) is given as , @xmath26 the event number asymmetry in eq.([asym ] ) is expanded as , @xmath27 hereafter , we discuss the coefficient of @xmath28 and the other coefficients in eq.([2.6 ] ) are evaluated and discussed in ref.@xcite .
the following quantities are introduced , @xmath29 where @xmath4 decay amplitude is denoted @xmath30 for @xmath31 and @xmath32 for @xmath33 , while inverse decay amplitudes are written as @xmath34 for @xmath35 and @xmath36 for @xmath37 . mixing parameters in b and k systems
are denoted as , @xmath38 one can check@xcite that @xmath39 and @xmath40 in eqs.([para]-[b ] ) are odd or even under t - transformation .
the transformation property is exhibited in tab.[tab1 ] .
.transformation properties of the parameters under t - transformation .
these parameters are to express the asymmetry so that one can show the transformation property of each term . [ cols="^,^,^,^,^,^,^",options="header " , ] [ tab1 ] the coefficient @xmath41 in eq.([2.6 ] ) is parameterized as , @xmath42 in eq.([2.8 ] ) , @xmath41 is written in terms of parameters which transform odd or even under t - transformation as shown in tab.[tab1 ] .
@xmath43 in eq.([2.8 ] ) denotes wrong sign semi - leptonic decays of @xmath4 meson such like @xmath44 and @xmath45 .
formulation of @xmath43 is given in ref.@xcite .
one can find that the quantities in eq.([2.8 ] ) include not only t - odd terms but also t - even contribution .
this slight deviation from t - odd quantities is pointed out in ref.@xcite regarding to babar asymmetry@xcite .
the event number asymmetry coefficients in eq.([2.8 ] ) slightly deviate from t - odd quantity as well as babar asymmetry .
there exists contribution from mixing induced cp violation in kaon system since the final states in fig.[fig1 ] include @xmath46 and @xmath47 which are different from exact cp eigenstates . extracting @xmath3 in eq.([2.8 ] )
, one expresses , @xmath48 under phase redefinition of a quark such like @xmath49 and @xmath50 , @xmath51 transforms as @xmath52 .
therefore , @xmath53 in eq.([2.14 ] ) is shown as a phase convention independent parameter .
one also sees that @xmath54 is phase convention independent and the quantities in eq.([2.13 ] ) are phase convention independent .
we find that @xmath55 arises to contribute to @xmath41 in eq.([2.13 ] ) and @xmath41 becomes a quantity of @xmath56 @xcite .
in the previous section , t - even parts in eq.([2.8 ] ) are found and the coefficients slightly deviate from t - odd quantities . hereafter , we discuss conditions for the coefficients in eq.([2.6 ] ) to be t - odd ( t - violating ) quantities .
there are two types of such a condition .
the first one is ( 1 ) b - meson state equivalence which is originally suggested and calculated in ref.@xcite .
this condition is regarding to b - meson states which appear in fig.[fig1 ] and it requires that b meson states should be equivalent to the states which reveal in diagrams whose time direction is totally opposite to fig.[fig1 ] .
( therefore such diagram contains @xmath57 as a final state and @xmath58 as an initial state ) this is diagrammatically discussed in ref.@xcite .
the second type of the conditions is ( 2 ) @xmath59 ; t - even part of the ratio of the overall normalizations in eq.([2.5 ] ) should vanish . since the event number asymmetry in eq.([2.2 ] )
is evaluated with non - zero @xmath22 , @xmath22 can give rise to t - even effect .
this is not the case for babar asymmetry@xcite since it eliminates the effect of overall normalization difference between the two processes ( a ) and ( b ) in eq.(2.1 ) .
however , when @xmath59 is satisfied , @xmath60 is a t - odd quantity while @xmath61 is proven@xcite to be a t - even quantity .
therefore , eq.([asym ] ) reads as , @xmath62 and a(t ) turns out to be t - odd quantity .
one can show that the event number asymmetry coefficients in eq.([2.6 ] ) become t - odd quantities when the two types of the conditions are simultaneously satisfied@xcite .
in this article , the event number asymmetry is investigated for the two processes : @xmath63 and @xmath64 . the constructed asymmetry in eq.([2.2 ] ) appears to include not only t - odd part but also t - even part , as shown in ref.@xcite with respect to babar asymmetry@xcite .
the mixing induced cp violation of kaon system , i.e. , @xmath3 contributes to the asymmetry since the final states given in @xmath65 and @xmath66 contain @xmath46 and @xmath47 .
we extracted the effect of @xmath3 and showed the contribution to an observable in eq.([2.13 ] ) .
formulation in this article is phase convention independent so that the coefficients in eqs.([2.13 ] ) are invariant under phase redefinition of quarks .
furthermore , we discussed the conditions for the asymmetry to be a t - odd quantity .
the two types of such a condition are investigated ; ( 1 ) b - meson state equivalence and ( 2 ) vanishment of t - even part of overall normalizations ratio of the decay rates which form the asymmetry ( denoted as @xmath59 ) .
the condition ( 1 ) requires @xcite that the b - meson state which appears in the process in fig.[fig1 ] should be equivalent to the b - meson which reveals in a process whose time direction is totally opposite to fig.[fig1 ] . the condition ( 2 ) is regarding to the difference of overall normalization of the rates .
we consider the event number asymmetry so that one should take into account the condition ( 2 ) unlike babar asymmetry @xcite .
one found that the t - even part of @xmath22 yields t - even contribution to the coefficients of the asymmetry .
we conclude that the coefficients of the event number asymmetry become t - odd quantities when the conditions ( 1 ) , ( 2 ) are simultaneously satisfied .
we would like to appreciate the organizers of fpcp2015 for the well - organized conference .
+ h. u. is supported by hiroshima univ .
alumni association .
99 j. p. lees _ et al .
_ [ babar collaboration ] , phys .
lett . * 109 * , 211801 ( 2012 ) [ arxiv:1207.5832 ] .
m. c. banuls and j. bernabeu , phys .
b * 464 * , 117 ( 1999 ) [ hep - ph/9908353 ] .
e. alvarez and a. szynkman , mod .
a * 23 * , 2085 ( 2008 ) [ hep - ph/0611370 ] .
j. bernabeu , f. martinez - vidal and p. villanueva - perez , jhep * 1208 * , 064 ( 2012 ) [ arxiv:1203.0171 ] .
e. applebaum , a. efrati , y. grossman , y. nir and y. soreq , phys .
d * 89 * , no . 7 , 076011 ( 2014 ) [ arxiv:1312.4164 [ hep - ph ] ] .
t. morozumi , h. okane and h. umeeda , jhep * 1502 * , 174 ( 2015 ) [ arxiv:1411.2104 [ hep - ph ] ] . | through @xmath0 mixing system , one can construct an asymmetry which naively seems to be a time reversal ( t ) odd quantity . in this article ,
the two processes @xmath1 and @xmath2 are used to construct the event number asymmetry .
the cp violation of kaon system denoted as @xmath3 contributes to observables and we evaluate the contribution from @xmath3 explicitly .
the asymmetry is formulated with phase convention independent parameters which are invariant under redefinition of phase of quark fields .
the overall factors of the time dependent decay rates are taken into account in this article .
furthermore , we suggest conditions for the asymmetry to be a t - odd quantity .
the one of such conditions arises due to the difference of overall factors which form the asymmetry . #
1/#1 0m_^0 |
the computational exfoliation protocol starts from a comprehensive initial set of bulk 3d crystal structures .
these are extracted from databases of experimental compounds , and for this study we focused on the icsd @xcite and cod @xcite , containing respectively @xmath4 ( icsd 2015.1 ) and @xmath5 entries ( cod rev .
# 171462 ) .
entries containing at most @xmath6 different species are retrieved in the form of cif files @xcite using routines distributed within the aiida platform @xcite , and passed through successive cod - tools filters @xcite to correct typical syntax errors that prevent the successful reconstruction of the intended crystal structures ( see methods for more details ) .
entries with partial occupancies in the unit cell are then filtered out , together with cif files that do not provide the explicit position of one or several atoms , can not be parsed , or are obviously wrong .
purely theoretical structures are also discarded .
the cif files are then converted into aiida structures using the pymatgen @xcite parser .
the spglib software @xcite is subsequently used to find the primitive cells and to refine the atomic positions , in order to avoid unwanted loss of crystal symmetries due to round - off errors ( see methods for more details ) .
finally , duplicate structures ( found using the pymatgen structure matcher@xcite - see methods ) are removed from our initial set of 3d structures .
the removal of duplicates is done first independently on each of the icsd and cod sets .
then , every structure of the icsd set for which there exists a similar structure inside the cod is also removed .
this procedure provides a starting set of @xmath7 unique 3d crystal structures , to be further investigated . the 3d structures classified as unique
can then be analysed to find promising candidates for exfoliation into one or several 2d materials .
the first step consists in identifying composite structures , i.e. , structures that can be decomposed into disconnected units held together by weak interactions .
a pioneering attempt in this direction has been put forward by lebgue and coauthors @xcite even if limited to high - symmetry compounds with layers perpendicular to the @xmath8 $ ] crystal axis .
a versatile algorithm is however needed in order to cope with the following general and realistic situations : * layered structures belonging to any crystal system ( obvious exceptions are the cubic systems @xcite ) ; * orientation of the stacking direction not along the @xmath8 $ ] axis ( see , e.g. , fig .
[ fig:2dlayerfinder]b ) ;
* layers apparently outside the 3d primitive unit cell , requiring the use of sufficiently large supercells to check for connectivity ( see , e.g. , fig .
[ fig:2dlayerfinder]c ) ; * layers that are interpenetrating , i.e. with overlapping projections along the stacking direction ( see , e.g. , fig .
[ fig:2dlayerfinder]d ) ; * composite structures that include units with different dimensionalities ( see , e.g. , fig .
[ fig:2dlayerfinder]e ) ; * different chemical environment affecting the typical bond distance between atoms , requiring some tolerance in the screening parameters . here
we provide a protocol that is able to deal with all the scenarios above and identify layers that can then be fed into first - principles calculations .
the goal is to automatically find chemically - bonded manifolds in a given structure , and among them select those that are periodic in two independent directions .
the fundamental steps are reported schematically in fig .
[ fig:2dlayerfinder]a .
first , from the bulk primitive cell of the 3d candidate a @xmath9 supercell is created .
all interatomic distances are evaluated and chemical bonds are identified as those for which @xmath10 where @xmath11 is the distance between two atoms @xmath12 and @xmath13 , @xmath14 is the van - der - waals radius of atom @xmath12 as obtained from alvarez ( ref . ) , and @xmath15 is a parameter that we specify below .
the criterion in is based on the statistical analysis of the distribution of over five million interatomic distances , according to which alvarez pointed out that covalent bonds are likely to form only when the distance between atoms is shorter than the sum of the vdw radii by more than a quantity @xmath15 ( on average equal to 1.3 ) .
thus , if the interatomic distance is smaller than @xmath16 as required by eq . , then the two atoms should be chemically connected . on the contrary , if the interatomic distance is larger than this , then any kind of strong chemical bonding should be absent and only vdw interactions remain ( some special case involving hydrogen or metallic bonding are still possible and discussed in ref . , but we will not consider them here ) . in our analysis
we have considered five equally spaced values for @xmath15 between @xmath17 and @xmath18 to account for possible uncertainties in the values of the vdw radii and of @xmath15 itself .
we believe that a criterion based on vdw radii could be more robust and versatile than one based on covalent radii , as it allows to take into account the unavoidable bond - length fluctuations associated with different chemical environments .
as soon as all bonds are identified , chemically connected groups are constructed .
@xmath9 supercells are needed to correctly describe cases when a group of bonded atoms spans a total of three unit cells in a given direction , as illustrated in fig .
[ fig:2dlayerfinder]c . in principle
, even such supercells might not be large enough in some rare cases , although the algorithm could then be applied to even larger supercells .
the dimensionality of each connected group is obtained from the rank of the set of vectors linking an atom to all of its chemically connected periodic images ( periodic according to the 3d lattice vectors ) .
this approach is general and does not assume any specific orientation of the 2d plane with respect to the crystal or cartesian axes .
moreover , it also identifies and returns other low - dimensional units , such as 1d chains or 0d clusters . applying the full protocol to all @xmath7 unique 3d structures , we found @xmath19 layered materials that can be considered as potential parents of 2d monolayers .
the geometrical screening described in the previous section provides an exhaustive list of potential candidates for exfoliation .
although an important first step , this geometrical characterization is not sufficient and has to be complemented with accurate calculations in order to identify materials that do actually exhibit weak interlayer bonding and are easily exfoliable .
here we use vdw dft simulations to proceed with such a refinement ( computational details can be found in the methods ) .
the first step is the optimization of the 3d geometry ( cell vectors and atomic positions ) of the layered structures found .
this procedure is completely automated within aiida .
preference is given to the simplest structures , both for their higher relevance and for computational efficiency .
as a result , at least all unary compounds containing less than @xmath20 atoms in the unit cell , all binaries and ternaries with less than @xmath21 atoms , and all other quaternary , quinary , and senary compounds with less than @xmath22 atoms are considered ; in total , @xmath23 layered compounds have been studied with first - principles calculations .
these structures are relaxed using two different non - local vdw - compliant functionals : the vdw - df2 functional @xcite with c09 exchange @xcite ( df2-c09 ) and the revised vydrov - van voorhis @xcite ( rvv10 ) functional . to assess the accuracy in predicting the equilibrium structure for the functionals adopted
, we compare the experimental interlayer distance ( defined as the distance between the geometrical centres of neighbouring layers ) of a subset of layered materials against computational predictions .
the results are shown in fig .
[ fig : volume_vs_exp ] , complemented also by the semi - local perdew - burke - ernzerhof ( pbe)@xcite and the revpbe@xcite functionals . since revpbe and pbe are not designed to predict vdw interactions , discrepancies with experiments are larger , as seen in fig .
[ fig : volume_vs_exp]a - b . on the contrary ,
the two vdw - compliant functionals provide interlayer distances closer to experimental values and with a much smaller dispersion around the average ( see fig .
[ fig : volume_vs_exp]c - d ) , and are able to correctly predict the structural properties of non - closely - packed materials ( including vdw - bonded structures ) .
this can be quantified by looking at the mean absolute percentage errors ( @xmath24 for both vdw - df2-c09 and rvv10 , against @xmath25 and @xmath26 for revpbe and pbe , respectively ) .
moreover , the low mean relative errors of vdw - df2-c09 and rvv10 predictions ( @xmath27 and @xmath28 , respectively ) show that these functional have no bias in determining the equilibrium interlayer distance of layered structures , while revpbe and pbe have a tendency to underbind that is emphasised at smaller packing ratios . to go further in the selection of 2d materials , we need to determine if the interlayer interactions are weak enough for the parent compound to be a good candidate for exfoliation .
after the structural optimization described previously , one can proceed with exfoliation , i.e. , separation into several substructures as described earlier .
if the exfoliation produces one or several two - dimensional structures , the ground - state energy of each isolated unit is computed .
note that , due to the change in atom positions and thus of bond lengths , dft optimization leads in certain cases to a non - exfoliable structure , even if the initial 3d parent directly extracted from the icsd or cod was classified as layered .
the ground - state energy difference between the 3d bulk and all of its substructures ( of any dimensionality , including also 1d chains or 0d clusters ) gives the binding energy @xmath29 , closely related to the exfoliation energy , i.e. , the energy needed to exfoliate one layer out of a multilayer system @xcite .
here @xmath30 is chosen to be positive when the bulk is more stable than its individual low - dimensional constituents .
we show in fig .
[ fig : binding]a the binding energies per layer and per unit surface obtained for well - known layered materials including graphite , hexagonal boron nitride and several transition - metal dichalcogenides .
the values obtained with the two vdw functionals are in very good agreement with accurate reference calculations performed in ref . using the random phase approximation ( rpa ) .
for all these compounds ( except pdte@xmath31 ) the binding energies range between @xmath32 and @xmath21 mev@xmath0 .
we then show in fig .
[ fig : binding]b the distribution of the binding energies for the @xmath33 systems studied in this work . for both vdw functionals , a majority of compounds ( resp . @xmath34 for df2-c09 and @xmath35 for rvv10 )
exhibit a binding energy smaller than @xmath20 mev@xmath0 , giving rise to a first very high peak in the distribution ( notice the logarithmic scale ) .
a second broader peak appears between @xmath36 and @xmath37 mev@xmath0 and is associated with structures containing also lower dimensionality units ( i.e. , 0d or 1d ) in addition to the 2d layer(s ) .
the larger binding energy is to be expected for structures with mixed dimensionality ( typically 0d and 2d ) where charge transfer between intercalated units and 2d layers can take place giving rise to ionic bonds , as in the case of nasb@xmath31as@xmath31 @xcite .
this means that these additional 0d or 1d substructures could have been erroneously separated from the 2d layers in the geometrical screening .
the binding energies distribution is quite continuous . as a result ,
a criterion to select layered materials solely based on the binding energy is difficult to define without a level of arbitrariness . to lift this ambiguity
, we correlate the binding energy with the relative discrepancy in interlayer distance obtained in vdw and non - vdw dft calculations , using for the non - vdw simulations the revpbe functional @xcite that , contrary to standard pbe , does not seem to display bonding interaction in the exchange energy @xcite .
we plot in fig .
[ fig : binding]c this binding energy against the interlayer - distance difference for @xmath38 compounds that have been studied with the revpbe and df2-c09 functionals .
we observe that most compounds with low binding energy exhibit a large relative discrepancy in interlayer distance calculated with the two functionals , which is expected since the interlayer interaction is very sensitive to vdw forces .
this allows us to identify which layered compounds are characterised by vdw interactions between layers .
we set the threshold for these ( shown in blue in fig .
[ fig : binding]c ) at @xmath39 mev@xmath0 for binding energies computed with the df2-c09 functional , and @xmath40 mev@xmath0 for those computed using rvv10 .
we classify two - dimensional systems exfoliated from compounds belonging to this group as `` easily exfoliable '' ( ee ) . as shown in fig .
[ fig : binding]c , this choice recovers 2d materials commonly exfoliated in experiments @xcite , validating the approach . at the top
left of fig .
[ fig : binding]c , a number of compounds exhibit high binding energies and a similar interlayer distance when comparing the revpbe and vdw functionals ; many also contain substructures of mixed dimensionalities , as discussed before .
this group can be clearly separated from the rest , and is shown in yellow in fig .
[ fig : binding]c , the boundary being set at @xmath41 mev@xmath0 for both functionals .
above this value , compounds are not considered exfoliable and they are discarded from the database . between these two regions , compounds with intermediate binding energies ( in green in fig .
[ fig : binding]c ) exhibit a relatively weak , possibly non - vdw , bonding . as an example
, the compound pdte@xmath31 discussed above belongs to this region , and is metallic out - of - plane .
we classify two - dimensional structures exfoliated from parents belonging to this group as `` potentially exfoliable '' ( pe ) .
finally , materials for which the binding energy has been computed with both vdw functionals are classified according to the most optimistic prediction . using these criteria , and after removal of duplicates ( see methods ) , we obtain @xmath42 easily exfoliable and @xmath43 potentially exfoliable compounds , for a total of @xmath44 candidates ( table [ tab : stat ] summarises all different steps ) . in fig .
[ fig:2dstats ] we plot some statistics for this portfolio of promising materials , as well as for their 3d parents : we show the distribution of packing ratios in the icsd and cod structures , in the group of 3d layered compounds , and in the ee and pe ones , using covalent radii @xcite to compute the atomic volume .
while the initial databases contain many closely packed structures with packing ratios of 0.7 or more , the layered structures are peaked around low packing ratios ( 0.3 ) but with a spread extending up to 0.8 , even in the final set of exfoliable compounds .
these considerations highlight how the packing ratio alone is not necessarily a good criterion to filter 2d structures .
( note that the peak of cod at a packing ratio of 0.15 which goes beyond the upper limit of the graph is due to the fact that cod is populated with a large number of molecular crystals , microporous networks , metal - organic frameworks , and zeolites . ) then , in fig .
[ fig:2dstats]b we show the distribution of point groups , arranged by crystal system , in all 3d structures , compared to that of the exfoliable 3d compounds .
most point groups are represented similarly in both sets , with the notable exception of all cubic point groups that are obviously absent from the set of layered compounds .
moreover , the @xmath45 point group is much rarer in the layered compounds than in the icsd and cod databases , while the @xmath46 m , @xmath47 , @xmath47 m and @xmath6 mm point groups are conversely more present among 3d exfoliable compounds .
it is interesting to note that while most of the 2d materials studied in the last ten years belong to the hexagonal crystal system , these structures represent only a minority of the full database of exfoliable materials . finally , in fig . [ fig:2dstats]c we show how the layered and 2d compounds are distributed in terms of the number of species and number of atoms in the unit cell . as apparent from the graph
, there are hundreds of 2d structures with few atoms in the unit cell ( @xmath48 with less than @xmath6 atoms , @xmath49 with less than @xmath50 atoms ) , that can be used as a starting set for screening 2d materials with desired properties .
we find 18 2d structures containing a single species , including all known exfoliable 2d unaries such as graphene and phosphorene .
novel unaries belong to the class of pe materials and mostly arise from intercalation layers in the parent 3d structures ; they are typically not mechanically stable . interestingly , ternary 2d materials , which have been so far largely unexplored ,
represent a significant fraction of structures with less than 10 atoms in the unit cell , demonstrating that it may be worth inspecting these more versatile compounds beyond the current realm of unary / binary 2d materials . in order to provide a more precise overview ,
we further classify the 2d materials of the ee group into different prototypes , according to their spacegroups and their structural similarity , when the species are all considered undistinguishable and only the stoichiometry matters ( for this we use the pymatgen structure matcher see methods for more details ) .
@xmath51 prototypes are found , among which the ten most common ones represent a total of @xmath52 structures and are shown in fig .
[ fig:2dprototypes ] .
the most common structural prototype is that of cdi@xmath31 , which groups together @xmath53 similar structures , including many transition - metal dichalcogenides and dihalides .
contrary to the general distribution of point groups in fig .
[ fig:2dstats]b , most of these prototypes have high - order rotation axes , with the only exception being the rectangular feocl .
although some of these structures have already been suggested as exfoliable @xcite , the abundance of some relatively complex prototypes involving more than two species is new and compelling .
moreover , we emphasise that the ndte@xmath54 prototype has not been reported before and it is common to many rare - earth tritellurides . lcccr & unique & unique & common & total + & to icsd & to cod & to both & sum + experimental data & & & & + cif inputs & 99212 & 87070 & & 186282 + unique 3d structures & 34548 & 60354 & 13521 & 108423 + layered 3d structures & 3257 & 1180 & 1182 & 5619 + + + layered 3d , relaxed & 2165 & 175 & 870 & 3210 + binding energies & 1795 & 126 & 741 & 2662 + 2d easily exfoliable ( ee ) & 672 & 81 & 300 & 1053 + 2d pot .
exfoliable ( pe ) & 521 & 36 & 234 & 791 + total & 1193 & 117 & 534 & 1844 +
we identified an extensive database of exfoliable materials by systematic analysis and screening of structures from both the icsd and the cod structural databases .
we developed a geometrical algorithm to select potentially layered materials that is general enough to recognise substructures independently of their crystallographic orientation , dimensionality and environment .
the list of 2d materials obtained by geometrical considerations is then tested against accurate , validated vdw - dft computations of the binding energies for all layers identified .
compounds with sufficiently small binding energies and where dispersion relations affect interlayer distances are deemed exfoliable .
the materials identified are in particular classified into groups of easily ( @xmath42 ) or potentially ( @xmath43 ) exfoliable compounds , showing how only a very small fraction of possible 2d materials has been considered up to now .
in particular , many opportunities might arise from materials with reduced symmetry or involving more than two atomic species but still with relatively simple structures . from the identification of the most common structural prototypes
, we suggest that materials displaying promising properties like ferromagnetic insulators or tunable superconductors actually belong to larger families of structures that potentially contain novel materials with even more compelling characteristics .
all results can be fully reproduced thanks to the deployment of the aiida infrastructure that tracks the provenance of each data entry . to date , this is the largest available database of 2d compounds .
it is an often repeated tenet that results of scientific research must be reproducible .
this objective is , however , challenging , especially in high - throughput research , due to the large number of simulations involved and the complex sequence of logical steps needed in the study . to ensure reproducibility we use aiida @xcite as a materials informatics infrastructure to implement the ades model of automation , data , environment , and sharing , as discussed in ref . .
all input information for each calculation is stored in the aiida repository before calculations are launched and actual code inputs are created by aiida using only data already stored in the repository . however , even the ability to replicate single calculations is not sufficient , since the majority of results are obtained via long sequences of simulations .
it is thus essential to persist the full provenance of each data item .
this means , for instance , that for any 2d material identified in this work one can track from which 3d parent structure it originated ( including its source database ) , how the structure was processed to obtain its coordinates from the wyckoff positions , how the low - dimensional units were identified , and which parameters were used to compute the dft results .
the cif files provided in crystallographic databases sometimes contain syntax errors or unnecessary information that needs to be corrected and/or removed before parsing them .
to do so , we use cod - tools@xcite , in particular the ` cif_filter ` and ` cif_select ` scripts , in order to : * fix common cif syntax errors , * parse and reformat the summary chemical formula ( from the tag ` _ chemical_formula_sum ` ) , * set explicit @xmath55 degrees cell angles if they are not given ( following the standard cif specifications ) , * remove empty non - loop tags , * remove cif data blocks without coordinates , * convert non - standard cif tag name capitalization into the capitalization specified by the cif dictionary , for example ` _ atom_site_cartn_x ` would be replaced by ` _
atom_site_cartn_x ` .
we also remove the tags ` _
publ_author_name ` and ` _
citation_journal_abbrev ` , that often contain problematic characters for , e.g. , the pymatgen cif parser ( this information is still stored in the aiida database and therefore fully searchable ) . as already mentioned in the main text , in this step we consider only structures coming from experimental measurements , while the source databases are partially populated by purely theoretical structures .
this is done using the flags set on the database entries by their curators .
we also implemented some heuristics to detect clearly wrong cif files .
for example , we discard structures where the chemical formula provided in the file is inconsistent with the elements in the unit cell . regardless of all these efforts , it is possible that some incomplete or incorrect structures are still not filtered out from the original databases .
the crystal structures parsed from cif files are often subject to round - off errors in the atomic positions and/or cell parameters that artificially lower the number of symmetries .
we use the spglib software @xcite to refine each structure and recover the maximum number of symmetries , using the algorithm described in ref . .
interatomic distances and cell parameters are modified by at most @xmath56 with respect to the initially parsed structure . at the same time ( and with the same software ) , a standardised crystallographic primitive cell is extracted @xcite . for the duplicate filtering procedure of both 3d and 2d structures , as well as for the prototyping of 2d structures , we have used extensively the `` structure matcher '' of pymatgen @xcite , which , thanks to the algorithm described in ref . , compares crystal structures .
the goal is to 1 ) reduce the two structures to be compared to their primitive cells , 2 ) rescale the volumes ( if 3d ) , or the in - plane areas ( if 2d ) , 3 ) find , when it exists , an affine map between the two cells , within certain tolerances on the cell lengths and angles , and 4 ) compute the maximum distance between atomic sites when the two lattices are mapped onto a common average lattice .
if this maximum is lower than a certain fraction of the average free length per atom , then the two structures are classified as similar .
when comparing 3d structures , the tolerances on cell lengths , cell angles , and distances between paired atomic sites were chosen to be @xmath57 , @xmath58 and @xmath59 of the average free length per atom , defined as the cubic root of the volume per atom . for 2d structures , due the artificial vacuum added to the cell in the out - of - plane direction
the relative tolerances needed to be set tighter , and we chose @xmath25 , @xmath60 and @xmath25 , respectively . to filter out duplicates of a set of structures , all compounds with same composition ( i.e. , the same species in the same relative proportion ) are compared among them using the structure - matcher algorithm , forming groups of matching structures .
one unique structure is then chosen in each group as a representative for further calculations .
the same tool is used to prototype 2d structures , but with a number of modifications : 1 ) two structures must have the same space - group to be considered as members of the same prototypical class ( we used spglib with a precision parameter of @xmath61 to get the space - group , see above ) , 2 ) the comparison is made irrespective of the specific species on each atomic site ( e.g. , mos@xmath31 and wte@xmath31 are both considered as ab@xmath31 ) , 3 ) when the two primitive cells do not contain the same number of sites , one also attempts to generate a supercell of the smaller lattice before the mapping , and 4 ) the tolerances on respectively the cell lengths , cell angles and maximum distance between sites are set to @xmath25 , @xmath58 and @xmath57 respectively .
again all structures of the same kind of composition ( e.g. , ab@xmath31 , a@xmath31b@xmath54 , etc . ) are compared in pairs , to classify them into distinct prototypical classes .
we use the quantum - espresso distribution@xcite with the sssp efficiency pseudopotentials library@xcite ( version @xmath62 ) .
this library of extensively tested pseudopotentials from various sources@xcite provides to date the best overall agreement with respect to all - electron calculations@xcite .
wavefunction and charge - density cutoffs are chosen according to convergence tests with respect to cohesive energies , stresses and phonon frequencies performed for each individual element@xcite .
two different van - der - waals functionals are employed , namely the vdw - df2-c09@xcite and the revised vv10@xcite , in addition to the pbe@xcite and revpbe@xcite functionals . @xmath63-point sampling of the brillouin zone is set using a @xmath64centred monkhorst - pack grid , with an even number of @xmath63 points in each direction of the reciprocal lattice such that the spacing between two consecutive @xmath63 points along each direction is as close as possible to @xmath65 @xmath66 .
a marzari - vanderbilt cold smearing@xcite of 0.02 ry is used .
all calculations are done without considering spin polarization and spin - orbit coupling .
when isolating 2d substructures , a vacuum space of 20 is set along the orthogonal direction to remove any fictitious interaction between periodic images of the 2d layers .
this work was supported by the marvel national centre of competence in research of the swiss national science foundation .
simulation time was provided by the swiss national supercomputing centre ( cscs ) under project ids mr0 and ch3 , amounting to @xmath67 million core hours to complete this study .
the authors would also like to thank davide campi for his help in evaluating the stability of unaries , as well as francesco ambrosio , antimo marrazzo and thibault sohier for many useful discussions . | we search for novel two - dimensional materials that can be easily exfoliated from their parent compounds . starting from 108423 unique ,
experimentally known three - dimensional compounds we identify a subset of 5619 that appear layered according to robust geometric and bonding criteria .
high - throughput calculations using van - der - waals density - functional theory , validated against experimental structural data and calculated random - phase - approximation binding energies , allow to identify 1844 compounds that are either easily or potentially exfoliable , including all that are commonly exfoliated experimentally .
in particular , the subset of 1053 easily exfoliable cases layered materials held together mostly by dispersion interactions and with binding energies in the range of few tens of mev@xmath0contains several hundreds of entries with few atoms per primitive cell ( 273 with less than 6 atoms , 606 with less than 12 ) , revealing a wealth of new structural prototypes , simple ternary compounds , and a large portfolio to search for optimal electronic , optical , magnetic , topological , or chemical properties .
= 1 two - dimensional ( 2d ) materials provide novel opportunities to venture into largely unexplored regions of the materials properties space . on one hand
, their ultimate thinness makes them extremely promising for applications in electronics ( e.g. , for field - effect transistors , where a reduced device size is beneficial to improve performance and reduce short - channel effects between contacts).@xcite on the other hand , the physical properties of monolayers often change dramatically from those of parent 3d materials , providing a new degree of freedom @xcite in the applications ( from thermoelectrics to spintronics ) while also unveiling novel physics ( e.g. , valley hall effect and composite excitations such as trions ) .
moreover , van - der - waals ( vdw ) heterostructures @xcite have recently emerged as an additional avenue to engineer novel properties by stacking 2d materials in a specifically designed fashion .
still , to date only a few dozens of 2d materials have been experimentally synthesised or exfoliated .
progress in this area would be strongly accelerated by the availability of a broader portfolio of potential realistic 2d materials . to illustrate this point , we can compare the current situation for known 3d crystals , for which the knowledge accumulated in the past century ( both the crystal structure and the measured physical properties ) has been collected in databases such as the pauling file @xcite , the inorganic crystal structure database @xcite ( icsd ) or the crystallographic open database @xcite ( cod ) ( the latter two , combined , contain to date over half a million entries ) . in comparison ,
2d materials databases are still scarce and limited in size : a first systematic scan of the icsd database providing @xmath1 two - dimensional compounds has been provided in ref . ,
followed by @xmath2 compounds selected among specific classes of materials @xcite ; finally a recent study focused on transition - metal dichalcogenides and oxides , identifying @xmath3 of them as stable @xcite . in order to fill this knowledge gap ,
high - throughput computational methods represent a promising tool@xcite that has proven to be extremely effective in screening materials without the need to synthesise them first@xcite .
for instance , these techniques have been successfully employed in the search for novel materials for li - air and li - ion batteries@xcite , for hydrogen storage@xcite , electrocatalysis@xcite , or to accelerate the discovery of novel light - absorbing@xcite and light - harvesting@xcite materials . in this work ,
we start from geometric and bonding criteria to identify layered materials among the compounds contained in the icsd and cod databases , independently of the layers shape , crystallographic orientation or embedding .
we then use first - principles vdw density - functional theory ( dft ) simulations to validate these results .
in particular , we compute the binding energy of all prospective layered structures and identify those that are held together by weak interactions and ready for mechanical@xcite or liquid - phase@xcite exfoliation@xcite .
this results into a database of 1844 exfoliable 2d materials .
the reproducibility of the complete study is ensured by the aiida @xcite materials informatics infrastructure , which keeps track of the provenance of each calculation and result . |
nova scuti 2009 ( v496 sct ) was discovered by nishimura on 2009 november 8.370 ut at @xmath3 = 8.8 ( nakano et al .
2009 ) on two 10s unfiltered ccd images .
the low resolution spectra obtained during the period 2009 november 9.73 ut to 10.08 ut which showed prominent h@xmath4 and h@xmath5 emission lines with p cygni components , along with the strong fe ii multiplets and o i lines indicating that v496 scuti is an fe ii class nova near maximum light ( munari et al . 2009a , balam & sarty 2009 ) .
the optical observations by munari et al .
( 2009b ) showed a post - discovery brightening for about 10 days before the onset of fading with maximum brightness @xmath6 = 7.07 around 2009 november 18.716 ut .
the ir observations by rudy et al .
( 2009 ) on 2009 november 27.08 ut showed strong first overtone co emission bands - an extremely short lived feature that is seen in only a few novae .
they also predicted that dust formation in v496 sct is almost certain .
the first result by raj et al
. 2009 showed the continuation of co emission during the period 2009 december 3.55 ut to 8.55 ut .
subsequent observations by russell et al .
( 2010 ) after v496 sct came out of the solar conjunction showed dust formation on 2010 february 10 .
the ir and optical observations for nova v496 sct are taken from mt .
abu ir observatory of prl in india , at asiago observatory of university of padova and inaf astronomical observatory of padova and schiaparelli observatory in italy , respectively . on both the days while the temperature of the gas t@xmath7 is 4000 k and 3600 k respectively .
the time from optical maximum are given for each spectrum .
, width=384,height=192 ]
from the optical light curve we estimate @xmath8 to be 59 @xmath2 5 d ( raj et al . 2012 ) which makes v496 sct as one of the moderately fast fe ii class of novae in recent years .
v496 sct is one of the large amplitude novae observed in recent years with @xmath9 @xmath10 13.5 magnitudes ( guido & sostero 2009 ) .
these observed values of the amplitude and @xmath8 for v496 sct put it above the upper limit in the observed spread of the amplitude versus decline rate plot for classical novae presented by warner ( 2008 , figure 2.3 ) which shows @xmath11 = 8 - 11 for @xmath12 = 59 days .
the height from the galactic plane is estimated to be @xmath13 = 89 @xmath2 3 pc by using the value of the ditance @xmath14 = 2.9 @xmath2 0.3 kpc ( raj et al .
2012 ) to the nova .
the outburst luminosity of v496 sct as calculated from m@xmath1 is l@xmath15 @xmath16 5.1 @xmath17 10@xmath18 l@xmath19 .
a small plateau is also seen in the light curve between 128 and 215 days after the outburst .
the best fit model spectra to the observed data are obtained by varying the input parameters m@xmath7 , @xmath4 , t@xmath7 ( see fig .
the increase in m@xmath7 enhances the absolute level of the co emission while the increase in t@xmath7 changes the relative intensities of different vibrational bands in addition to changing the absolute level of the emission .
the co emission is assumed to be optically thin .
the c i lines at 2.2906 and 2.3130 @xmath21 m and na i lines at 2.3348 and 2.3379 @xmath21 m are also likely to be present in the spectral region covered by the co emission giving rise to some deviations between the best model fit and the observed spectra .
the typical errors to the formal model fits are @xmath2 500 k. the model spectra with a reasonably similar range in mass of m@xmath7 = 1.5@xmath222 @xmath17 10@xmath23 m@xmath19 fit the observed spectra .
the model calculations also show that the @xmath24 = 2 - 0 bandhead of @xmath25co at 2.3130 @xmath21 m becomes discernibly prominent if the @xmath26c/@xmath25c ratio is @xmath271.5 . as this spectral feature
is not clearly detected in our observed spectra , we place a lower limit of 1.5 for the @xmath26c/@xmath25c ratio .
the various phases of the spectral evolution of v496 sct have been identified using the tololo classification system for novae ( williams et al .
( 1991 ) ; williams et al .
( 1994 ) ) .
the permitted lines of fe ii were the strongest non - balmer lines in the pre - maximum as well as the early decline phase indicating p@xmath28 class for the nova .
the nova had evolved to the auroral phase a@xmath29 in 2010 march as the [ n ii ] 5755 auroral line was the strongest non - balmer line .
we note the absence of [ fe x ] 6375 coronal emission line in the spectra taken as late as 2011 april 19 ( see fig .
thus the optical spectra show that the nova evolved in the p@xmath28a@xmath29 spectral sequence .
we have also tried to estimate the ejecta mass by using recombination line analysis of h i lines for 2009 december 6 ( fig .
3 ) . however , we find that the strengths of these lines , relative to each other , deviate considerably ( specially for br@xmath33 , fig .
3 ) from case b values for 2009 december 6 indicating that the lines are optically thick .
hence we are unable to estimate the ejecta mass from recombination analysis . from the optical spectra , we estimated the mass of oxygen m@xmath34 in the range 1.18@xmath1710@xmath35 - 2.28@xmath1710@xmath36 m@xmath19 .
the mass of hydrogen m(h ) in the ejecta is ( 6.3 @xmath2 0.2)@xmath1710@xmath35m@xmath19 .
we obtain m@xmath37 = 1 - 5@xmath1710@xmath38m@xmath19 for 2010 april 30 from the best fit value t@xmath37 @xmath39 1500 @xmath2 200 k ( with @xmath40@xmath41 minimization ) for @xmath14 = 2.9 kpc . hence the gas to dust ratio
is found to be m@xmath42/m@xmath37 @xmath16 1.3 - 6.3 @xmath1710@xmath43 indicating that a small amount of dust was formed in v496 sct comparable to 3 @xmath1710@xmath43 observed in the case of v2362 cyg by munari et al .
( 2008 ) .
the research work at physical research laboratory is funded by the department of space , government of india .
we would like to thank a. frigo , v. luppi , l. buzzi , a. milani , g. cherini , a. maitan , l. baldinelli ( ans collaboration ) . | we present near - infrared and optical observations of moderately fast feii - class nova scuti 2009 ( v496 sct ) covering various phases ; pre - maximum , early decline and nebular , during the first 10 months after its discovery followed by limited observations up to 2011 april . in the initial phase
the nova spectra show prominent p cygni profiles and later all the lines are seen in emission .
the notable feature of the near - ir spectra in the early decline phase is the rare presence of the first overtone bands of carbon monoxide ( co ) in emission .
the ir spectra show clear dust formation in the expanding ejecta at later phase about 150 days after the peak brightness .
the presence of lines of elements with low ionization potentials like na and mg in the early ir spectra and the detection of co bands in emission and the dust formation in v496 sct represents a complete expected sequence in the dust formation in nova ejecta .
the light curve shows a slow rise to the maximum and a slow decline indicating a prolonged mass loss .
this is corroborated by the strengthening of p cygni profiles during the first 30 days .
the broad and single absorption components seen in many lines in the optical spectra at the time of discovery are replaced by two sharper components in the spectra taken close to the optical maximum brightness .
these sharp dips seen in the p cygni absorption components of fe ii and h i lines during the early decline phase show increasing outflow velocities .
the onset of the nebular phase is evident from the optical spectra in 2010 march . during the nebular phase
, several emission lines display saddle - like profiles . in the nebular stage ,
the observed fluxes of [ o iii ] and h lines are used to estimate the electron number densities and the mass of the ejecta .
the optical spectra show that the nova is evolved in the @xmath0 spectral sequence .
the absolute magnitude and the distance to the nova are estimated to be m@xmath1 = 7.0 @xmath2 0.2 and d = 2.9 @xmath2 0.3 kpc , respectively . |
the galaxy m33 ( ngc 598 ) is one of the closest spiral galaxies of the local group .
the measured distance to m33 ranges from 730 kpc ( christian & schommer 1987 ) to 910 kpc ( kim et al .
2002 ) , with the most recent estimates of sarajedini et al .
@xcite and bonanos et al .
@xcite in the mid - range .
the proximity of m33 , together with its large angular size ( optical size 53@xmath783 , holmberg 1958 ) , and its intermediate inclination ( @xmath8=53 ) , allows detailed studies of its stellar populations and ionized nebulae .
m33 is a galaxy rich in both pne and regions .
the population of pne in m33 was early investigated using an objective - prism survey by ford @xcite and lequeux et al .
the advent of wide - field ccd cameras allowed deeper surveys .
magrini et al .
@xcite identified 131 candidate pne .
more recently , ciardullo et al .
@xcite confirmed a large number of previously detected pne and identified new candidates , leading to the current number of spectroscopically confirmed pne : 138 in the m33 disk and 2 in its halo .
due the closeness of m33 and the brightness of the giant regions , their spectroscopy was obtained since the 70s .
searle @xcite presented spectrophotometry of eight regions , and further spectroscopic studies were carried on by smith @xcite , kwitter & aller @xcite , and by vlchez et al .
several catalogs of the m33 regions have been published , such as those by courtes et al .
@xcite , calzetti et al .
@xcite , wyder et al .
@xcite , and hodge et al .
recently , the local group census ( lgc , corradi & magrini 2006 ) and the local group survey ( lgs , massey et al .
2007 ) have provided deep narrow- and broad - band @xmath92 square degree coverage of m33 , disclosing a conspicuous number of new emission - line objects located at all galactocentric distances .
given the wealth of information on its pn and region populations , m33 is indeed an ideal candidate to test chemical evolution predictions .
the aim of the spectroscopic studies of pne and regions is devoted both to the individual studies of these objects , and to draw the radial metallicity gradient across the disk .
in particular , pne and regions have two very different formation ages , and the comparison of these two populations provides insight on the evolution of the host galaxy .
pn progenitors are low- and intermediate - mass stars ( lims ) , with masses between 1 and 8 m@xmath10 , which must have been formed between @xmath93@xmath11 yr and 10 gyr ago ( maraston 2005 ) .
instead regions are very young . during their evolution , lims do not modify , at least at zeroth approximation , the composition of @xmath4- elements such as oxygen , neon , argon , and sulfur .
these elements are produced mainly from the nucleosynthesis of type ii supernovae and maintain their capability to testify their original presence in the interstellar cloud that gave birth to the pn progenitor . while there are indications that for extremely low metallicities both oxygen and neon could be modified , as it has been observed both in the smc and in other dwarf galaxies ( leisy and dennefeld 2006 ; magrini et al .
2005 ; kniazev et al .
2008 ) , at the metallicity of m33 one does not expect any nucleosynthesis activity involving these elements ( marigo 2001 ) . on the other hand , the helium , nitrogen , and carbon abundances measured in pne do not correspond to those at the time of the progenitor s formation , since these elements are synthesized in lims .
these elements hence give information on lims evolution accordingly to their initial mass and metallicity , and , at given metallicity , are helpful to constrain the pn progenitor mass and age .
the m33 metallicity gradient derived from @xmath4-element abundances has been the focus of several studies already , both involving pne and regions .
the advantage of m33 with respect to our galaxy , as a playground for gradients , is that pne in m33 have well determined galactocentric distances ( with relative errors within 5@xmath0 ) compared to the large indetermination of galactic pn distances ( stanghellini et al .
magrini et al .
@xcite using pn spectroscopy in m33 determined the elemental abundances of 11 pne , and magrini et al .
( 2007b ) ( hereafter m07b ) derived an oxygen radial gradient of @xmath12(o / h ) / @xmath12r = [email protected] dex kpc@xmath6 , where r is the galactocentric distance .
the sample of 11 pne was too small for a definite answer to the pn gradient of m33 , especially given the paucity of pne observed at large galactocentric distances .
the metallicity gradient of m33 using regions has been obtained by many authors , with broadly different results .
the first studies by smith @xcite , kwitter & aller @xcite , and vlchez et al . @xcite agreed on a steep oxygen gradient .
garnett et al .
@xcite also obtained a steep gradient , with @xmath12(o / h ) / @xmath12r = [email protected] dex kpc@xmath6 by homogeneously compiling published data .
recent determinations , such as those by crockett et al .
@xcite , magrini et al .
@xcite , and rosolowsky et al .
@xcite ( hereafter m07a and rs08 , respectively ) seem to converge to a much shallower gradient , respectively deriving @xmath12(o / h ) / @xmath12r = [email protected] , [email protected] , and [email protected] dex kpc@xmath6 .
the most recent result by rubin et al .
@xcite , based on ne / h and s / h , gives [email protected] and [email protected] dex kpc@xmath6 , respectively .
the rs08 sample is the largest to date , and should be used preferentially , since the results from small samples might emphasize the scatter rather than the slope .
metallicity gradients in m33 have also been estimated form young giant stars ( herrero et al .
1994 , mccarthy et al .
1995 , venn et al .
1998 , monteverde et al .
1997 , 2000 , urbaneja et al .
2005 ) , and from agb ( cioni et al . 2008 ) , rgb stars ( stephens et al . 2002 , kim et al 2002 , galletti et al .
2004 , tiede et al .
2004 , brooks et al .
2004 , barker et al .
2006 ) , and cepheids ( beaulieu et al .
2006 ) .
one of the most discussed questions about the metallicity gradient in disk galaxies is how it evolves with time .
chemical evolution models ( see e.g. m07b for a review of m33 models ) predict different temporal behaviors of the metallicity gradient depending on assumptions such as gas inflow and outflow rate , and star and cloud formation efficiencies .
observations are needed to constrain these theoretical scenarios , but so far they have been insufficient , especially for the old populations . comparing different sets of results for the young stellar populations , such as regions , and
the old population , such as agb and rgb stars , is also delicate , since the techniques of observing and analyzing nebulae and stars are very different , each with its own collection of uncertainties .
the idea behind the observations leading to this paper is to study the chemical and physical properties of a large number of pne and regions using the same set of observations , the same data reduction and analysis techniques , and identical abundance determination methods , in order to avoid all biases due to the stellar vs. nebular analysis .
the chemical properties of pne and regions will provide us with snapshots at two epochs in the life - time of m33 , in particular of its metallicity gradient .
in the present paper we show the results obtained from the pn data .
we have obtained the first sizable sample of m33 pne with uniformly derived abundances in order to build the first sound gradient determination from pne . in a forthcoming paper
we will present our own region results , and we will compare them with the pn results presented here .
this paper is organized as follows : in @xmath13 [ sect_obs ] we present the observations and data reduction , in @xmath13 [ sect_abu ] we discuss the plasma diagnostics used and how we determine the abundances of the pne , and in @xmath13 [ sect_pne ] we describe the pn properties . in @xmath13 [ sec_comp ] we compare the results with the pn populations in other galaxies .
the chemical abundance gradients and their implications in the evolution of m 33 are discussed in @xmath13 [ sect_grad ] and in @xmath13 [ sect_discu ] .
finally in @xmath13 [ sect_conclu ] we summarize our results , and present the conclusions of this work .
we obtained spectra of 102 pne and 48 regions in m33 using the mmt hectospec fiber - fed spectrograph ( fabricant et al .
2005 ) which is equipped with an atmospheric dispersion corrector .
the spectrograph was used with a single setup : 270 mm@xmath6 grating at a dispersion of 1.2 pixel@xmath6 .
the resulting total spectral coverage ranged from approximately 3600 to 9100 , thus including the basic emission - lines necessary for the determination of their physical and chemical properties .
the instrument deploys 300 fibers over a 1diameter field of view and the fiber diameter is @xmath9 1.5 ( 6 pc using a distance of 840 kpc to m 33 ) . the 102 pne were selected from the catalog of ciardullo et al .
@xcite , including four new pne discovered therein at large galactocentric radii . in table [ tab_pn ] we present the list of the observed pne .
column ( 1 ) gives the identification number from ciardullo et al .
@xcite except for pne 153 to 156 which are from lgc observations ( corradi & magrini 2006 ) ; columns ( 2 ) and ( 3 ) give the equatorial coordinates , ra and dec at j2000.0 , and column ( 4 ) gives the magnitude from ciardullo et al .
@xcite following the definition by jacoby @xcite .
cccc pn001 & 1:32:09.04 & 30:22:05.700 & 23.51 + pn002 & 1:32:26.54 & 30:25:49.800 & 23.53 + pn003 & 1:32:38.03 & 30:24:00.603 & 21.82 + [ tab_pn ] the regions of our sample were chosen either based on their brightness , or among those whose chemical abundances had already been measured by others so we can use them as a control sample . in this paper
we only use the data of the regions to test our procedures by comparing our fluxes with those in the literature .
we used a large number of fibers to take sky spectra for sky - subtraction . in order to optimize the sky - subtraction we selected low diffuse - emission areas on the face of m33 from the int h@xmath4 and -continuum frames ( magrini et al .
since the continuum from the pn central star can not be detected at the m33 distance any continuum emission in the spectra has to be arising from an underlying unresolved stellar population .
however , the continuum emission in our spectra is in most cases negligible , and therefore the balmer absorption lines coming from the stellar background are unimportant .
five 1800 s exposures were taken on the night of october 13 , 2007 , and three additional 1800 s exposures of the same field were obtained on november 12 , 2007 .
the airmass during the observations ranged between 1.07 and 1.4 during the night of october 13 , and from 1.3 to 1.5 on november 12 .
the seeing was 1.2 and 1.7 , respectively .
several dome - flat and sky - flat exposures were obtained during the nights of observations to perform the data reduction .
arc exposures with the calibration lamp he - ne - ar were taken for wavelength calibration .
the spectra were reduced using the hectospec package .
all observations were bias subtracted , overscan corrected , and trimmed .
the science exposures were flat - fielded and combined together to eliminate cosmic rays , and the one - dimensional spectra were extracted and wavelength calibrated .
the relative flux calibration was done observing the standard star hiltm600 ( massey et al .
1988 ) during the nights of october 15 and november 27 .
the standard star was observed with airmass @xmath91.1 .
the emission - line fluxes were measured with the package splot of iraf .
errors in the fluxes were calculated taking into account the statistical error in the measurement of the fluxes , as well as systematic errors of the flux calibrations , background determination , and sky subtraction .
the observed line fluxes were corrected for the effect of the interstellar extinction using the extinction law of mathis @xcite with @xmath14=3.1 .
we derived , the logarithmic nebular extinction , by using the weighted average of the observed - to - theoretical balmer ratios of h@xmath4 , h@xmath15 , and h@xmath16 to h@xmath17 ( osterbrock & ferland 2006 ) .
table 2 gives the results of our line measurements and extinction corrections .
column ( 1 ) gives the pn name ; column ( 2 ) gives the nebular extinction coefficient with its error ; columns ( 3 ) and ( 4 ) indicate the emitting ion and the rest - frame wavelength in ; columns ( 5 ) , ( 6 ) , and ( 7 ) give the measured line fluxes ( f@xmath18 ) , their absolute errors ( @xmath12(f@xmath18 ) , and finally the extinction corrected fluxes ( i@xmath18 ) . both f@xmath18 and i@xmath18
are normalized to = 100 .
ccccrrr pn001 & [email protected] & hi & 4340 & 47.6 & 3.7 & 52.4 + & & heii & 4686 & 58.0 & 4.4 & 60.1 + & & hi & 4861 & 100.0 & 4.8 & 100.0 + & & [ oiii ] & 4959 & 304.2 & 7.4 & 298.9 + & & [ oiii ] & 5007 & 917.8 & 12.2 & 894.0 + & & hi & 6563 & 362.2 & 8.0 & 289.0 + pn002 & [email protected] & hi & 4340 & 42.3 & 3.9 & 46.9 + & & hi & 4861 & 100.0 & 4.8 & 100.0 + & & [ oiii ] & 4959 & 304.1 & 7.5 & 298.5 + & & [ oiii ] & 5007 & 953.0 & 12 . &
927.04 + & & hi & 6563 & 368.3 & 8 .
& 289.70 + pn003 & [email protected] & [ oii ] & 3727 & 101.8 & 1.6 & 137.2 + & & hi & 3835 & 7.7 & 0.8 & 10.1 + & & hei & 3889 & 9.1 & 0.8 & 11.7 + & & [ neiii]/hi & 3968 & 7.6 & 0.6 & 9.6 + & & hi & 4100 & 17.3 & 0.8 & 21.1 + & & hi & 4340 & 37.5 & 1.2 & 43.1 + & & [ oiii ] & 4363 & 7.4 & 0.7 & 8.5 + & & hei & 4471 & 4.4 & 0.7 & 4.9 + & & hi & 4861 & 100.0 & 1.8 & 100.0 + & & [ oiii ] & 4959 & 178.5 & 2.0 & 174.2 + & & [ oiii ] & 5007 & 540.1 & 3.0 & 520.2 + & & [ nii ] & 5755 & 3.7 & 0.7 & 3.0 + & & hei & 5876 & 22.4 & 1.0 & 18.0 + & & [ nii ] & 6548 & 43.5 & 1.0 & 31.6 + & & hi & 6563 & 384.8 & 2.4 & 280.0 + & & [ nii ] & 6584 & 128.2 & 1.6 & 92.7 + & & hei & 6678 & 5.2 & 0.7 & 3.7 + & & [ sii ] & 6717 & 7.2 & 0.6 & 5.1 + & & [ sii ] & 6731 & 10.3 & 0.8 & 7.3 + & & hei & 7065 & 9.7 & 0.8 & 6.5 + & & [ ariii ] & 7135 & 14.8 & 0.8 & 9.9 + [ tab_flux ] in order to confirm the goodness of our spectroscopic calibration we compare our emission - line flux measurements of both regions and pne with previously published fluxes of the same objects . in figure 1 we plot our measured fluxes against the ones from the literature , where the two dashed lines mark differences of @xmath50.15 dex between the two sets .
we found a good agreement between the sets , especially for bright emission lines , giving us confidence of a sound spectral calibration . in figure [ fig_spectra ]
we show , as an example of the data quality , the spectra of three pne with different excitation .
we have used the extinction - corrected intensities to obtain the pn electron densities and temperatures . in order to calculate the electronic densities we used the doublet of the sulfur lines @xmath196716,6731 , while for the electron temperatures we used the ratios @xmath204363/(@xmath205007 + @xmath204959 ) and @xmath205755/(@xmath206548 + @xmath206584 ) .
we performed plasma diagnostics by using the 5-level atom model included in the _ nebular _ analysis package in iraf / stsdas ( shaw & dufour 1994 ) .
we have also used the standard forbidden line diagnostics iraf routines in _ nebular _ to determine the electron temperatures .
we have determined the low- and medium - excitation temperatures from the and the line ratios , respectively ( see also osterbrock & ferland 2006 , @xmath135.2 ) .
the temperature uncertainties have been estimated by formal error propagation of the absolute errors on the line fluxes ( see also table 4 ) .
the average relative uncertainties in the determination of and are of the order of 5@xmath0 and 15@xmath0 , respectively .
the and temperature diagnostics are available , respectively , for 32 and 8 pne of our sample . only for six pne we can use both temperature diagnostics . in order to improve the size of the pn sample with temperature determinations we looked for correlations between the electron temperature and various diagnostics of the nebular excitation , similarly to the approach followed by kaler @xcite for a galactic pn sample . in figure [ fig_3 ] we plot the electron temperature against the intensity of the i@xmath21 line , scaled for i@xmath22=100 and corrected for extinction .
both quantities are plotted with their formal error bars .
planetary nebulae in this plot are those hot enough for helium to be doubly ionized .
we find a clear correlation between the electron temperature and the intensity of the i@xmath21 line , an effect that is expected as the central stars heat the medium - high excitation nebulae . in order to use this correlation in our temperature calculation
we have applied the routine _ fitexy _ in numerical recipes ( press et al .
1992 ) to fit the relation between the two variables , taking into account their errors , and minimizing @xmath23 . for i@xmath240
we find : @xmath25}=(63\pm8)~{\rm i_{\lambda4686}}+(11,000\pm200),\eqno(1)\ ] ] where is given in k and i@xmath21 is scaled to = 100 .
note that using a different fitting routine , such as _ sixlin _ ( isobe et al .
1990 ) , we determine temperatures that agree with the ones of eq .
( 1 ) at the 1% level .
we infer that we are able to estimate the temperatures to better than 10% with the above formulation . if i@xmath21=0 we assume a constant value of t@xmath26[o iii ] , as in kaler @xcite .
the average t@xmath26[o iii ] for i@xmath21=0 is 12100 @xmath51500 k , which is compatible with the continuity of eq .
note that both this value and eq .
( 1 ) give temperatures that are consistently higher than those found by kaler @xcite in galactic pne , an effect that is probably due to the lower metallicity that characterizes the m33 stellar populations ( see stanghellini et al .
@xcite for a discussion on how metallicity affects the nebular temperatures ) . in figure [ fig_2 ]
we show against i@xmath21 , where both quantities are plotted with their formal errors .
the line represents the mean value of = 13600 @xmath52100 k , which has been derived using the ( inverse squared ) relative temperature errors as weights . once again , while the physical behavior of the temperatures is similar to that found in kaler @xcite , the actual values are higher , possibly reflecting the lower efficiency of the cooling process at low metallicity . for i@xmath270
we adopt = 13,600 k. if i@xmath21=0 , we expect to decrease with the stellar temperature . the only way to frame this decline is to use the ratio between the fluxes of the singly - ionized to the doubly - ionized oxygen . by correlating the oxygen line strengths , corrected for extinction , to the electron temperatures for the pne with measured lines we find : @xmath28}=(11,700\pm2800)-(2800\pm4000)~{\rm log}{{{\rm i_{\lambda\lambda3727 - 29}}}\over{{\rm i_{\lambda 4959}}+{\rm i_{\lambda5007}}}}. \eqno(2)\ ] ] note that in our spectra the @xmath29 lines are always blended . if the oxygen lines are not available , and i@xmath21=0 we assume = 14,100 @xmath52800 k , which , in this case , is the weighted mean of the temperatures .
cll [ tab_allabu ] pn001 & & + & te(oiii ) & 14650 + & te(nii ) & 13600 + & heii / h & 0.063 + & he / h & 0.063 + & oiii / h & 7.586 10@xmath30 + & icf(o ) & 1.000 + & o / h & 7.586 10@xmath30 + pn002 & & + & te(oiii ) & 14650 + & te(nii ) & 13600 + & oiii / h & 7.884 10@xmath30 + & icf(o ) & 1.000 + & o / h & 7.884 10@xmath30 + pn003 & & + & te(oiii ) & 14000 + & te(nii ) & 15200 + & te(oiii ) & 12060 + & te(nii ) & 14120 + & ne(sii ) & 2540 + & hei / h & 0.125 + & he / h & 0.125 + & oii / h & 1.590 10@xmath30 + & oiii / h & 6.500 10@xmath30 + & icf(o ) & 1.000 + & o / h & 8.090 10@xmath30 + & nii / h & 7.168 10@xmath31 + & icf(n ) & 5.087 + & n / h & 3.646 10@xmath30 + & ariii / h & 4.47 10@xmath32 + & icf(ar ) & 1.245 + & ar / h & 8.359 10@xmath32 + & sii / h & 1.736 10@xmath32 + & icf(s ) & 1.276 + & s / h & 1.665 10@xmath31 + we computed the ionic abundances using the _ nebular _ analysis package in iraf / stsdas ( shaw & dufour 1994 ) . the elemental abundances were then determined by applying the ionization correction factors ( icfs ) following the prescriptions by kingsburgh & barlow @xcite for the case where only optical lines are available .
as discussed in the previous @xmath13 , the @xmath204363 emission line was measured with a sufficiently high signal to noise ratio in 32 pne , affording the direct determination of the .
we could also directly measure in 8 pne .
for the other targets we adopted the temperatures derived from the relations we have described in the previous section . in the abundance analysis
we have used for the calculation of the n@xmath33 , o@xmath33 , s@xmath33 abundances , while was used for the abundances of o@xmath34 , s@xmath34 , ar@xmath34 , he@xmath35 , and he@xmath34 .
we calculated the abundances of and using the equations of benjamin et al .
@xcite in two density regimes , i.e. @xmath31000 @xmath36 and @xmath371000 @xmath36 .
the clegg s collisional populations were taken into account ( clegg 1987 ) . due to the effect of double collisions
not being properly corrected , the he@xmath33/h ionic abundance from the @xmath207065 emission - line is quite different from the ionic abundance computed from @xmath194471 , 5876 , and 6678 .
we thus omit the abundance derived from the former line from the average .
the computed individual ionic and elemental abundances of each pn are shown in table 3 , where column ( 1 ) gives the pn identification name ; columns ( 2 ) and ( 3 ) show the plasma diagnostic and abundances available for each pn , where in column ( 2 ) we label each diagnostic , and in column ( 3 ) we give the relative value obtained from our analysis .
lllllllll 20.60 - 21.00 & l & 0.03 & 0.025 & 0.06 & 0.05 & 0.06 & 0.09 & 8
+ & c & - & 0.17 & 0.40 & 0.43 & 0.34 & 0.43 & 1 + 21.01 - 21.50 & l & 0.04 & 0.03 & 0.06 & 0.06 & 0.11 & 0.11 & 9 + & c & 0.09 & 0.20 & 0.34 & 0.43 & 0.30 & 0.30 & 8 + 21.51 - 22.00 & l & 0.04 & 0.03 & 0.04 & 0.08 & 0.09 & 0.09 & 5 + & c & 0.08 & 0.21 & 0.34 & 0.43 & 0.34 & 0.30 & 8 + 22.01 - 22.50 & l & 0.08 & 0.04 & 0.07 & 0.15 & 0.17 & 0.11 & 6 + & c & 0.11 & 0.21 & 0.40 & 0.43 & 0.40 & 0.30 & 16 + 22.51 - 23.00 & l & 0.13 & 0.09 & 0.16 & 0.16 & 0.17 & 0.11 & 4 + & c & 0.17 & 0.20 & 0.40 & 0.43 & 0.43 & 0.30 & 9 + 23.51 - 24.00 & l & 0.09 & 0.07 & 0.09 & 0.13 & 0.18 & 0.16 & 1 + & c & 0.21 & 0.23 & 0.40 & 0.43 & 0.43 & 0.30 & 12 + 24.01 - 24.50 & c & 0.11 & 0.21 & 0.34 & 0.43 & 0.43 & 0.43 & 5 + 24.51 - 25.00 & c & - & 0.45 & - & - & - & - & 2 + 25.01 - 25.50 & c & - & 0.5 & - & - & - & - & 1 + [ tab_error ] the formal errors on the ionic and total abundances were computed taking into account the uncertainties in the observed fluxes , in the electron temperatures and densities , and in .
the errors ( in dex ) of the final abundances are given in table [ tab_error ] .
typical errors are given for pne in different magnitude ranges , taking also into account whether the electron temperatures were derived from emission - line diagnostic or from the relations described in @xmath13 [ sec_temp ] . in the latter case , we assumed a percentage error on the electron temperature of @xmath910% , which is the dispersion of the relation used to derive the temperature .
the first column of table [ tab_error ] gives the magnitude range ; column ( 2 ) gives the method used for the electron temperature determination , where _
l _ means that the temperature has been measured based on an emission line and _ c _ means that we used the correlations described in @xmath133 ; columns ( 3 ) to ( 8) give the typical errors in dex of the total abundances .
the last column gives the number of pne on each bin .
ccccccccc m33 pne , disk ( 91 ) & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + m33 pne , non - type i ( 72 ) & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + m33 pne , type i ( 19 ) & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + m33 pne , halo ( 2 ) & 0.120 : & [email protected] & 0.84 : & [email protected] & [email protected] & 6.45 : & 0.31 : & [email protected] + m33 hii regions & [email protected](a ) & [email protected](b ) & [email protected](a ) & [email protected](c ) & 1.31 @xmath50.45(a ) & [email protected](a ) & [email protected](a)&[email protected](c ) + solar value(d ) & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + lmc pne & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & 0.17 @xmath50.09 + smc pne & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & 0.17 @xmath50.08 + galactic pne & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & & [email protected] & [email protected] + [ tab_avabu ]
most of the pne discovered in m33 belong to its disk , with only two pne having radial velocities compatible with the halo population ( ciardullo et al .
therefore , we have the opportunity to analyze in detail a pure disk pn population by excluding from our analysis the two suspected halo pne ( namely pn067 and pn024 ) .
we have also excluded from our analysis the pn039 since its physical conditions do not allow us to obtain abundance diagnostics .
pn039 has a wr - nucleus and will be discussed in detail in the appendix .
the plot of the n / o ratio vs the he / h one is a classical diagnostic diagram used to discriminate pne of different types , i.e. with different mass progenitors .
the n / o ratio provides information about the stellar nucleosynthesis during the agb phase of lims . on the one hand ,
nitrogen is produced in agb stars in two ways : by neutron capture , during the cno cycle , and by hot - bottom burning .
hot - bottom burning produces primarily nitrogen but occurs only if the base of the convective envelope of the agb stars is hot enough to favor the conversion of @xmath38c into @xmath39n . thus nitrogen
is expected to be mostly enriched in those pne with the most massive progenitors , i.e. with turnoff mass larger than @xmath93 m@xmath10 ( van der hoek & groenewegen 1997 ; marigo 2001 ) .
the he / h ratio gives also an indication of the initial mass of the progenitor , the nebula is enriched progressively for more massive stars , it reaches a plateau between 3 and 4 m@xmath2 , and then it increases again toward the higher masses ( marigo 2001 ) . in order to make a selection of the pne with high - mass progenitors , in figure [ fig_heno ] , we plot the m33 disk pne on the he
- n / o plane .
peimbert & torres - peimbert @xcite found that galactic pne that were nitrogen- and helium - enriched also had bipolar shape , and were located closer to the galactic plane .
it looked like these pne formed a younger population , which they defined as those with @xmath40(n / o@xmath41 and he / [email protected] , and called them type i pne .
kingsburgh & barlow @xcite analyzed a much larger sample and redefine the galactic type i pne as those having log(n
/ o)@xmath43 - 0.1 , with 18@xmath0 of the pne in their sample being of type i. dopita @xcite analyzed lmc pne and noted that the type i definition needed a revision to allow for the lower metallicity of the lmc with respect to the galaxy , setting the limit of type i pne to log(n / o)@xmath43 - 0.5 for the lmc .
leisy & dennefeld @xcite basically confirmed this limit .
magrini et al .
@xcite also noted that the definition of type i pne depends on the metallicity , because the amount of nitrogen that can be produced by hot bottom burning is dependent on the amount of carbon present , and also because the oxygen abundance depends on the metallicity of the galaxy .
since the oxygen metallicity of m33 is very similar to that of the lmc , we use the type i limits as in dopita @xcite , namely , type i pne are those with log(n
/ o)@xmath3 - 0.5 , independent of helium abundance . with this definition we found 19 pne being
type i in the disk of m33 ( @xmath920% of the whole disk population ) and one in the halo .
the study of the chemical evolution of galaxies needs strong constraints regarding the past composition of the ism , and pne can supply them , in particular with their soundly - determined oxygen abundances .
the questions is whether oxygen ( and neon ) is modified in lims while in their agb phase . a possible way to verify
this is to study the relationship between neon and oxygen .
these elements derive both from primary nucleosynthesis , mostly in stars with m@xmath310 m@xmath10 .
if the o / h and ne / h abundances are really independent of the evolution of the pn progenitors through the agb phase , a tight correlation between their abundance should be observed . in figure [ fig_oxyne ]
we show o / h against ne / h for the 55 m33 pne where both abundances were available .
the slope of the correlation is close to unity , [email protected] , with a correlation coefficient r@xmath44=0.81 , pointing at a locked variation of these two elements .
recent results from wang & liu @xcite indicate that oxygen and neon could be manufactured with similar yields also in lims , but only at very low metallicities , 12 + @xmath45(o / h ) @xmath1 8 .
these results are however still not explained by current nucleosynthesis theories that predict different channels , and thus different yields , for ne and o production in lims at low metallicity ( cf . ,
e.g. , karakas & lattanzio 2003 , marigo et al .
our findings confirm that in m33 , as in lmc , and the galaxy , there is no evidence of enhancement of oxygen and neon in pne .
similar correlations are expected also for sulphur versus oxygen , since sulphur is manufactured by massive stars . in figure [ fig_oxysu ]
we plot the sulfur versus oxygen abundances of the 38 m33 pne where these abundances are available .
their relationship has a slope of [email protected] and a correlation coefficient of 0.51 , showing moderately good correlation .
we refrain from showing and analyzing the argon to oxygen relation . as seen in stanghellini et al .
@xcite , these quantities might not correlate due to their different origin .
even if both elements are manufactured in massive stars , argon is synthesized in very different @xmath4 -element processes than oxygen , thus a lockstep of these elements is not completely expected .
the average chemical abundances of our m33 pne are reported in table [ tab_avabu ] where the first column describes the selected sample , columns ( 2 ) to ( 7 ) give the average total abundance of he / h , o / h , ne / h , ar / h , and s / h , whereas the last two columns , ( 8) and ( 9 ) , give the mean values of n / o and ne / o .
the average abundances were computed excluding those abundances from upper limit detections .
when the emission lines for temperature diagnostics were not detected , we adopted the electron temperatures discussed in @xmath13[sec_temp ] .
their larger uncertainties were taken into account by larger formal errors in the total chemical abundances as described in table [ tab_error ] .
we group the pne in m33 into four populations : _ i _ ) the whole pn disk population ; _ ii _ ) the disk population excluding those pne which are believed to be type i ; _ iii _ ) the type i pne , and _ iv _ ) the halo pn population as classified by ciardullo et al . @xcite . for the sake of comparison we add to the table the average chemical abundances of the m33 regions , the solar abundances , and the average abundances of the pne in lmc , smc ( stanghellini 2008 ) and the galaxy ( stanghellini et al .
note , however , that the population of the type i and non - type i classes remains somewhat uncertain , since about a dozen pne could belong to either class based on their formal abundance errors .
however , the chemical abundances and distributions of type i and non - type i pne in m33 are very similar , so that the displacement of a few pne from a group to the other would not change our results significantly .
an inspection of table 5 affords several interesting clues on the chemical evolution of galaxies .
first , if we compare the average abundances of the @xmath4-elements of the m33 disk pne with those of the regions , we do not note significant enrichment in the regions , leading to the conclusion that these elements are not substantially modified in the lims evolution .
we substantially found a flat age - metallicity relationship in m33 all the way to the time of pn formation .
second , we infer that the metallicity of the pne in m33 , based on oxygen abundances , is sub - solar . if we consider the entire sample of disk pne in m33 we see that the averages of @xmath4-elements are virtually identical to those of pne in the lmc , with pn metallicity about 1/2 that of the galaxy , and twice the smc on average , both in oxygen and neon . from table 5
we see that the average ne / o ratio in the general population of m33 pne is identical to that of the lmc , but lower than the mean ne / o in galactic pne .
this finding is right on the mark with the recent results of wang & liu @xcite .
in fact wang & liu @xcite found that the ne / o ratio increases with increasing oxygen metallicity both in pne and in regions .
this suggests a different enrichment history of neon and oxygen in the ism and thus probably different production mechanisms of these two @xmath4-elements in massive stars , in agreement with current theoretical calculations by kobayashi et al . @xcite . while we should expect the solar ne / o ratio to be consistent with that of galactic disk pne and regions
, we note that the asplund et al.s @xcite value , as well as grevesse & sauval s @xcite , is slightly discrepant in this scenario .
both pne and regions in the galactic disk give a consistent ne / o ratio of @xmath90.25 , higher than the solar value .
our results on m33 pne seem thus to be consistent with the recent suggestions of a revision of the solar ne / o ratio and the absolute neon ( e.g. wang & liu 2008 , rubin et al .
2008 ) .
the n / o average shown in table 5 for m33 pne is lower than the galactic and higher than the smc values .
the comparison to the lmc average has limited importance , giving the large range of lmc n / o values .
in figure [ fig_oxy ] , we show the oxygen abundance as a function of galactocentric distance for our sample of disk pne .
the galactocentric distances were computed by adopting a distance to m33 840 kpc , as determined by freedman et al .
@xcite , an average and well - accepted distance estimate . a weighted linear least - square fit to the complete disk sample gives a gradient : @xmath46 where r@xmath47 is the de - projected galactocentric distance in kpc , computed assuming an inclination of 53 , and a position angle of 22 . in figures
[ fig_ne ] , and [ fig_su ] we plot the radial metallicity gradients of ne / h and s / h .
the solid lines are the weighted linear fits to the complete sample of disk pne ( first row of table [ tab_grad ] for each element ) . in all these figures
the two symbols refer to type i ( filled circles ) or non - type i ( empty circles ) pne . in table
[ tab_grad ] we show the slopes ( column 2 ) and zero - points ( chemical abundances in the centre of the galaxy , column 3 ) of the metallicity gradients considering different samples of pne .
the sample used and the number of pne included are given in columns ( 4 ) and ( 5 ) respectively . for each element ,
the first row report the gradient obtained considering the whole sample of disk pne ( thus excluding the two possible halo pne ) , the second row gives the gradient computed using non - type i pne , and the third row gives the gradient obtained for type i pne .
the gradients are computed with a weighted least mean square fit .
for each element the slope of the three different samples are consistent within the errors .
note that the metallicity gradient of an old stellar population could be affected by the radial migration of stars during cosmic times .
the present - time location of a pn could not necessarily be the place where it was born .
radial mixing of stars is believed to be due to several mechanisms , among them the diffusion of stars on their orbits because of various irregularities in the galactic potential ( cf .
wielen et al . 1996 ) , to the passage of spiral patterns ( cf .
minchev & quillen 2006 ) , to changes in the angular momentum due to non - asymmetric forces due to molecular clouds ( spitzer & schwarzzchild 1953 ) .
upper limits to the radial migration rate were estimated by de simone et al .
@xcite and by haywood @xcite , giving similar values around 1.5 - 3.7 kpc gyr@xmath6 , limited to a radial region @xmath52 kpc from the birth place .
the model of sellwood & binney @xcite show that spiral waves in galaxy disks churn both the interstellar medium and the stars , affecting their metallicity gradients .
the effect of stirring of the entire disk due to the spiral waves is a flattening of the metallicity gradient .
however , the fact that metallicity gradients survive and are well observed in disk galaxies seem to indicate that the effectiveness of the migration processes is moderate .
ccccc o / h & [email protected] & [email protected] & all pne & 91 + o / h & [email protected] & [email protected] & non - type i pne & 72 + o / h & [email protected] & [email protected] & type i pne & 19 + ne / h & [email protected] & [email protected] & all pne & 55 + ne / h & [email protected] & [email protected] & non - type i pne & 37 + ne / h & [email protected] & [email protected] & type i pne & 18 + s / h & [email protected] & [email protected] & all pne & 38 + s / h & [email protected] & [email protected] & non - type i pne & 19 + s / h & [email protected] & [email protected] & type i pne & 19 + [ tab_grad ]
the metallicity gradient of pne allows to analyze how was the metal distribution in the past epochs of m33 . in m33
we have to possibility to compare the metallicity gradient of pne with a good number of other measurements of abundance gradients , in particular from regions .
the advantage of comparing pne to regions is the similarity of their emission - line spectra in spite of their different evolutionary states that allows to use the same observation techniques , analysis , and abundance determinations . in order to compare our pn results with the leading samples of m33 regions
we build the largest , most reliable sample available to date .
as already noted by e.g. rs08 and m07a , the size and quality of the sample is fundamental to derive reliable metallicity gradients . in regions
there is a substantial intrinsic scatter of 0.11 dex in the metallicity at any given distance from the m33 center , which imposes a fundamental limit on the accuracy of gradient measurements that rely on small samples of objects ( rs08 ) .
also , in the past , only giant regions have been studied , and those might have a steeper metallicity gradient than small and compact regions .
we have re - calculated the regions metallicity gradients considering a cumulative sample , including that by rs08 , consisting of 61 regions with accurate abundance determinations , and that by m07a that included new and literature spectroscopy performed to a set of 28 regions .
the cumulative sample of 89 regions gives , @xmath48 when compared to the best gradient determined using m33 pne ( eq .
3 ) , we confirm that these gradients are consistent within the errors .
we infer that the metallicity gradient of m33 has not changed from the epoch of the formation of the pn progenitors to the present time .
we can use for comparison the gradient of the best measured element , oxygen , or the weighed average of the slope of o / h , ne / h , and s / h gradient , and shown in fig .
[ fig_grad ] , [email protected] dex kpc@xmath6 . in table
[ tab_grad_lit ] we compare our metallicity gradient with the literature values derived using different stellar populations .
@xmath4-element and iron - peak gradients are separated by a double horizontal line since their different origin ( snii vs. snia ) suspects a different metallicity distribution through the disk and possibly a different gradient . in each group ,
@xmath4-elements or fe / h , gradients are ordered by the increasing age of the corresponding stellar population .
we considered only metallicity gradients computed with a statically significant sample , thus , e.g. , excluding the metallicity gradient with cepheid stars ( beaulieu et al .
2006 ) .
it is remarkable that both the oxygen and neon metallicity gradients of m33 do not change within the errors even considering populations of different ages , though they are slightly different , the neon gradient having a slightly steeper slope . in the following we discuss which evolutionary scenario and chemical evolution models
are excluded or favored by our results . the first chemical evolution model attempting to the describe the evolution of m33 was that by diaz & tosi @xcite .
their model assumes a constant and uniform infall rate and fails to reproduce the metallicity gradient observed in m33 .
other models that describe the origin and evolution of metallicity gradients are listed in table 7 , where we report the reference ( column 1 ) , the model type ( column 2 ) , the resulting oxygen gradient ( column 3 ) , and its variation with time ( column 4 ) , where a plus sign indicates a flattening of the metallicity gradient with time .
the first three models listed in table 7 are a family of multiphase models , differing mainly from the origin and rate of the infalling material .
a multiphase model considers the galaxy divided in two or more zones , typically the halo and the disk , composed by baryonic material in different phases : diffuse gas , clouds , stars , and stellar remnants .
these models evolve with time ( see e.g. ferrini et al .
1992 , 1994 for a complete description ) , and are characterized by a flattening of the radial metal distribution with time , due to the inside - out formation of the disk .
moll et al .
@xcite specifically analyzed the time - evolution of the metallicity gradient in several spiral galaxies , where the disk is formed by the primordial gas of the spheroidal protogalaxy which collapses onto the galactic plane and forms out the disk .
for m33 they obtained that @xmath12(o / h)/@xmath12(r)@xmath9 -0.21 dex kpc@xmath6 , flattening with time at a rate of + 0.005 dex kpc@xmath6 gyr@xmath6 .
moll & diaz @xcite applied their results to m33 and found metallicity gradients similar to those by moll et al .
gradient evolution is not computed explicitly , but a clear flattening of the gradient is evident in their figures .
m07b built the so - called `` accretion '' model reproducing the slow formation of the disk of m33 from the inflow of intergalactic gas , which predicts @xmath12(o / h)/@xmath12(r)@xmath9 -0.067 dex kpc@xmath6 , also flattening with time at a rate of + 0.003 dex kpc@xmath6 gyr@xmath6 .
other types of models , such as those by chiappini et al .
( 1997 , 2001 ) , were not applied directly to m33 , but are notable since they predict steepening of the metallicity gradients with time .
these models assume two main accretion episodes for the formation of the galaxy , the first forming the halo , the bulge , and the thick disk in a short timescale , and the second forming the thin disk , with a timescale that is an increasing function of the galactocentric distance .
different assumptions on the star formation threshold and on the timescales of the infall episodes produce the gradients slopes and evolutions reported in table [ tab_model ] ( see chiappini et al .
2001 for a detailed description ) .
note that the behavior of model d is different from the other models by the same authors in that the assumed slower formation of the outer halo does not influence the disk formation , resulting in steeper gradients .
if for example we assume that the age of all pne is 5 gyr , then according to the models by molla et al .
we would expect a pn gradient which is -0.025 dex kpc@xmath6 steeper than that of the regions . using the m07b models for the same pne age ,
the pn gradient is -0.015 dex kpc@xmath6 steeper than that of the regions . on the other hand , with models a
, b , c by chiappini et al .
( 2001 ) we would observe pn gradients that are flatter that those of regions by + 0.04 , + 0.02 , + 0.03 dex kpc@xmath6 respectively .
these variations could be even larger for older pn populations . in summary , only the models by m07b , and model b by chiappini et al .
@xcite , are marginally consistent with the observations . to match the small metallicity gradient variation with time we observed with the models of m33 we could invoke a lower accretion rate from the ism than that used in the m07b models .
the resulting simulation should produce a flat metallicity gradient , and a small slope variation in the last 8 - 10 gyr , as observed .
detailed analysis of galactic chemical evolution models are out of the scope of this work .
cccc moll et al . (
1997 ) & multiphase , infall from collapse of the halo & -0.21 & + 0.005 + moll & diaz ( 2005 ) & multiphase , infall from collapse of the halo & -0.20 & ... + m07b & multiphase , accretion model & -0.067 & + 0.003 + chiappini et al.(2001 ) & model a & -0.065 & -0.008 + & model b & -0.045 & -0.004 + & model c & -0.070 & -0.006 + & model d & -0.20 & + 0.05 + [ tab_model ] cccccc o / h & [email protected] & b giants & few myr & 30 & urbaneja et al .
@xcite + o / h & [email protected] & regions & few myr & 89 & rs08 + m07a + o / h & [email protected] & young pne & @xmath10.3 gyr & 19 & this paper + o / h & [email protected] & all pne & @xmath110 gyr & 91 & this paper + o / h & [email protected] & old pne & 0.3 - 10 gyr & 72 & this paper + ne / h & [email protected] & regions & few myr & 30 & willner & nelson - patel @xcite + ne / h & [email protected] & regions & few myr & 25 & rubin et al .
@xcite + ne / h & [email protected] & young pne & @xmath10.3 gyr & 18 & this paper + ne / h & [email protected] & all pne & @xmath110 gyr & 55 & this paper + ne / h & [email protected] & old pne & 0.3 - 10 gyr & 37 & this paper + fe / h & -0.06 & agb & @xmath94 - 6 gyr & - & cioni et al .
( 2008 ) + fe / h & [email protected] & rgb & @xmath38 gyr & - & kim et al .
( 2002 ) + fe / h & [email protected] & rgb & @xmath38 gyr & - & tiede et al .
( 2004 ) + [ tab_grad_lit ]
we present spectroscopic observations of a large sample ( 102 ) of pne in the spiral galaxy m33 .
the observations were secured with the spectrograph hectospec on the 6.5 m telescope mmt . for 32 pne
the electron temperature was directly computed .
empirical relationships among electron temperatures and bright emission lines as @xmath204686 , and @xmath203727 were used to derive electron temperatures of the remaining pne .
abundance diagnostic diagrams were built to study the pn population . from the plot of n / o versus he / h we found that the pne that fulfill the dopita @xcite definition of type i pn are 19 .
most pne are non type i , implying a population mainly composed by pne from old progenitors , with @xmath13m@xmath2 and ages @xmath30.3 gyr .
a tight relationship between the o / h and ne / h abundances was found , excluding the modification of both elements by pne progenitors and ensuring oxygen to be a good tracer of the galaxy metallicity .
the average chemical abundances of the pne were compared to the pn population of the lmc , the smc , and the galaxy .
generally speaking , the elemental abundances of the @xmath4-elements of the pne in the disk of m33 are very similar to those of the lmc .
the comparison between m33 pne and regions indicates a negligible global enrichment of the m33 disk from the epoch of the formation of the pn progenitors to the present time .
the radial metallicity gradients of those elements which are not modified during the lifetime of lims oxygen , neon , sulphur were derived .
the best measured element , oxygen , has a slope -0.031 ( @xmath5 0.013 ) dex kpc@xmath6 in agreement within the errors with the same gradient derived from regions -0.032 ( @xmath5 0.009 ) dex kpc@xmath6 .
since the metallicity gradients of pne and regions are practically indiscernible and the mean abundances of the pn and region populations are very close , we conclude that the chemical enrichment in m33 from the time of the formation of the pn progenitors to the present - time has been nearly negligible .
pn039 has a wr central star .
its spectrum shows prominent and wide features of he ii at @xmath20 4686 , 4541 , 5411 and of he i at @xmath20 4471 , 5876 , 6678 , 7065 .
in addition other stellar lines of [ ciii ] at @xmath20 4056 and 4649 , and [ cii ] at 7107 and 7124 were detected .
the fwhm of these lines is 15 , to be compared instrumental resolution of about 6 .
thus , following the classification of acker & neiner @xcite pn039 has a late - type [ wc]-9 or [ wc]-10 nucleus due to the presence of very bright [ ciii ] and [ cii ] lines respect to the [ civ ] lines .
the ratio of the sulphur doublet indicates a low density limit @xmath49100 @xmath36 , and , 4363/5007 ratio is completely anomalous for a normal pn , giving an electron temperature @xmath9 30,000 k. the reddening of the nebula is null , within the errors , indicating a little presence of dust .
we thank an anonymous referee for his / her valuable comments and suggestions that have improved the paper .
we thank d. fabricant for making hectospec available to the community and the tac for awarding us the observing time .
we thank the hectospec instrument team and mmt staff for their expert help in preparing and carrying out the hectospec observing runs .
we thank n. caldwell , d. ming and their team for the help during the data reduction .
thanks to katia cunha for enlightening scientific discussion , and richard shaw for his help with the _ nebular _ package .
acker , a. , neiner , c. 2003 , , 403 , 659 asplund , m. , grevesse , n. , sauval , a. j. 2005 , aspc , 336 , 25 barker , m. k. , sarajedini , a. , geisler , d. , harding , p. , schommer , r. 2006 , in press , astro - ph/0611892 beaulieu , j. -p . , buchler , j. r. , marquette , j. -b .
, hartman , j. d. , schwarzenberg - czerny , a. , 2006 , , 653 , 101 benjamin , r. a. , skillman , e , d. , smits , d. p. , 1999
, , 514 , 307 bonanos , a. z. , et al .
2006 , , 652 , 313 brooks , r. s. , wilson , c. d. , harris , w. e. 2004 , , 128 , 237 calzetti , d. , kinney , a. l. , ford , h. , doggett , j. , long , k. s 1995 , , 110 , 2739 chiappini , c. , matteucci , f. , & gratton , r. 1997 , , 477 , 765 chiappini , c. , matteucci , f. , & romano , d. 2001 , , 554 , 1044 christian , c. a. , & schommer , r. a. 1987 , aj , 93 , 557 ciardullo , r. , durrell , p. r. , laychak , m. b. , herrmann , k. a. , moody , k. , jacoby , g. h. , feldmeier , j. j. , 2004 , , 614,167 cioni , m .-
l. , irwin , m. , ferguson , a. m. n. , mcconnachie , a. , conn , b. c. , huxor , a. , ibata , r. , lewis , g. , tanvir , n. , 2008 , , 487 , 131 clegg , r. e. s. , 1987 , , .229 , 31 corradi , r. l. m. , magrini , l. , 2006 , in planetary nebulae beyond the milky way , eso astrophysics symposia , 36 courtes , g. , petit , h. , petit , m. , sivan , j .-
p . , dodonov , s. 1987 . , 174 , 28 crockett , n. r. , garnett , d. r. , massey , p. , jacoby , g. 2006 , , 637 , 741 davidge , t. j 2003 , , 125 , 304 de simone , r. , wu , x. , tremaine , s. 2004 .
, 350 , 627 devereux , n. , duric , n. , scowen , p. a. 1997 , , 113 , 236 diaz , a. i. , & tosi , m. 1984 , , 208 , 365 dopita , m. a. 1991 , iaus , 148 , 393 fabricant , d. , et al .
2005 , , 117 , 1411 ferrini , f. , matteucci , f. , pardi , c. , & penco , u. 1992 , , 387 , 138 ferrini , f. , molla , m. , pardi , m. c. , & diaz , a. i. 1994 , , 427 , 745 freedman , w. l. , wilson , c. d. ; madore , b. f. 1991 , , 372 , 455 ford , h. c. , 1983 , in planetary nebulae ; proceedings of the symposium , dordrecht , d. reidel publishing co. , 1983 , p. 443
- 460 galletti , s. , bellazzini , m. , ferraro , f. , 2004 , , 423 , 925 garnett , d. r. , shields , g. a. , skillman , e. d. , sagan , s. p. , dufour , r. j. 1997 , , 489 , 63 grevesse , n. , sauval , a. j. 1998 , space science reviews , v. 85 , issue 1/2 , p. 161
- 174 haywood , m. 2008 , , 388 , 1175 herrero , a. , lennon , d. j. , vlchez , j. m. , kudritzki , r. p. , humphreys , r. h. 1994 , , 287 , 885 hodge , p. w. , balsley , j. , wyder , t. k. , skelton , b. p. , 1999 , pasp , 111 , 685 holmberg , e. , 1958 , lund medd .
obs . ser .
ii , 136 , 1 kaler , j. b. , 1986 , , 308 , 322 karakas , a , lattanzio , j.c .
, 2003 , publ .
austr.20 , 393 kim , m. , kim , e. , lee , m. g. , sarajedini , a. , geisler , d. 2002 , , 123 , 244 kingsburgh , r. l. , barlow , m. j. , 1994 , , 271 , 257 kniazev , a. y. , et al .
2008 , , 388 , 1667 kobayashi , c. , umeda , h. , nomoto , k. , tominaga , n. , ohkuo , t , 2006 , , 653 , 1145 kwitter , k. b. , aller , l. h. , 1981 , 195 , 939 jacoby , j. h. , 1989 , , 339 , 39 isobe , t. , feigelson , e. d. , akritas , m. g. , & babu , g. j. 1990 , , 364 , 104 leisy , p. , dennefeld , m. 2006 , , 456 , 451 lequeux , j. , meyssonnier , n. , azzopardi , m. , 1987 , a&as , 67 , 169 magrini , l. , corradi , r. l. m. , mampaso , a. , perinotto , m. 2000 , , 355 , 713 magrini , l. , perinotto , m. , corradi , r. l. m. , mampaso , a. 2003 , , 400 , 511 magrini , l. , perinotto , m. , mampaso , a. , corradi , r. l. m. 2004 , , 426 , 779 magrini , l. , leisy , p. , corradi , r. l. m. , perinotto , m. , mampaso , a. , & vlchez , j. m. 2005 , , 443 , 115 magrini , l. , vilchez , j. m. , mampaso , a. , corradi , r. l. m. , leisy , p , 2007a , , 470 , 865 ( m07a ) magrini , l. , corbelli , e. , galli , d. 2007b , , 470 , 843 ( m07b ) maraston , c. 2005 , , 362 , 799 marigo , p. , 2001
, , 370 , 194 marigo , p. , bernard - salas j. , et al .
2003 , , 409 , 619 massey , p. , strobel , k. , barnes , j. v. , anderson , e. , 1988 , , 328 , 315 massey , p. , olsen , k. a. g. , hodge , p. w. , jacoby , g. h. , mcneill , r. t. , smith , r. c. , strong , s. b. , 2007 , , 133 , 2393 mathis , j. s. , 1990 , ara&a , 28 , 37 mccarthy , j. k. , lennon , d. j. , venn , k. a. , kudritzki , r .- p .
1995 , , 455 , 135 minchev , i. , quillen , a. c. 2006 . ,
368 , 623 moll , m. , ferrini , f. , & diaz , a. i. 1997 , , 475 , 519 moll , m. , & daz , a. i. 2005 , , 358 , 521 monteverde , m. i. , herrero , a. , lennon , d. j. , kudritzki , r .- p .
1997 , , 474 , 107 monteverde , m. i. , herrero , a. , lennon , d. j. 2000 , , 545 , 813 osterbrock , d. e. , ferland , g. j. 2006 , _ astrophysics of gaseous nebulae and active galactic nuclei _
2nd . ed . by d.e .
osterbrock and g.j .
sausalito , ca : university science books , sky and telescope , vol .
78 peimbert , m. , torres - peimbert , s. , 1983 , in proceedings of the symposium , dordrecht , d. reidel publishing co. , 1983 , p. 233 - 241 press , w. h. , teukolsky s. a. , vetterling w.t . , flannery , b.f .
, numerical recipes in fortran : the art of scientific computing .
( 1992 ) rosolowsky , e. , simon , j. d. , 2008 , , 675 , 1213 rubin , r.h .
et al . , 2008 , , 387 , 45 sarajedini , a. , barker , m. k. ; geisler , d. , harding , p. schommer , r. , 2006 , , 132 , 1361 searle , l. , 1971 , , 168 , 327 sellwood , j. a. , binney , j. j 2002 , , 336 , 785 shaw , r. a. , dufour , r. j , 1994 , aspc , 61 , 327 smith , h. e. 1975 , , 199 , 591 spitzer , l. , jr . ,
schwarzschild , m. 1953 , , 118 , 106 stanghellini , l. , villaver , e. , shaw , r. , mutchler , m. 2003 , , 598 , 1000 stanghellini , l. , guerrero , m. a. , cunha , k. , manchado , a. , villaver , e. , 2006 , , 651 , 898 stanghellini , l. , 2008 , in iau symp .
256 , in press stephens , a. w. , frogel , j. a. 2002 , 124 , 2023 tiede , g. p. , sarajedini , a. , barker , m. k. 2004 , 128 , 224 urbaneja , m. a. , herrero , a. , kudritzki , r .-
, najarro , f. , smartt , s. j. , puls , j. , lennon , d. j. , corral , l. j. 2005 , , 635 , 311 van den hoek , l. b. , groenewegen , m. a. t. , 1997 , , 123 , 305 venn , k. a. , mccarthy , j. k. , lennon , d. j. , kudritzki , r. p. 1998 aspc , 147 , 54 vlchez , j. m. , pagel , b. e. j. , diaz , a. i. , terlevich , e. , edmunds , m. g. , 1988 , 235 , 633 wang , w. , liu , x .- w .
, 2008 , , 389 , 33 wielen , r. , fuchs , b. , dettbarn , c. 1996 , 314 , 438 willner , s. p. , nelson - patel , k. 2002 , , 598 , 679 wyder , t. k. , hodge , p. w. , skelton , b. p. 1997
, pasp , 109 , 927 | the planetary nebula ( pn ) population of m33 is studied via multi - fiber spectroscopy with hectospec at the mmt . in this paper
we present the spectra of 102 pne , whereas plasma diagnostic and chemical abundances were performed on the 93 pne where the necessary diagnostic lines were measured .
about 20@xmath0 of the pne are compatible with being type i ; the rest of the sample is the progeny of an old disk stellar population , with main sequence masses m@xmath13m@xmath2 and ages [email protected] gyr . by studying the elemental abundances of the pne in the m33 disk we were able to infer that : ( 1 )
there is a tight correlation between o / h and ne / h , broadly excluding the evolution of oxygen ; ( 2 ) the average abundances of the @xmath4-elements are consistent with those of regions , indicating a negligible global enrichment in the disk of m33 from the epoch of the formation of the pn progenitors to the present time ; ( 3 ) the radial oxygen gradient across the m33 disk has a slope of [email protected] dex kpc@xmath6 , in agreement , within the errors , with the corresponding gradient derived from regions .
our observations do not seem to imply that the metallicity gradient across the m33 disk has flattened considerably with time .
we report also the discovery of a pn with wolf - rayet features , pn039 , belonging the class of late [ wc ] stars # 1.#2 # 1#2 |
black holes in the universe have either masses less than about 20 m@xmath2 ( orosz 2003 ) , or masses exceeding about @xmath3 ( see kormendy & richstone , 1995 , for some exceptions see fillipenko & ho 2003 ; greene & ho 2004 ) .
the first category of black holes are typically the member of a binary system in a normal galactic environment , the other type of supermassive black holes can be very bright in x , but always live in the nuclei of galaxies .
the discovery of bright x - ray sources which are not in the nuclei of galaxies has led to the suggestion that these sources may hide black holes , with a mass intermediate between 20 and @xmath3 ; we call them intermediate mass black holes ( imbh ) .
bright non - nuclear x - ray sources have been known to exist since the 1980s ( watson , stanger & griffiths 1984 ) , recently with the advent of chandra and xmm these sources have attracted considerably more attention . the conventional definiton of an ultraluminous x - ray source is @xmath4 ergs / s , but for the purpose of this paper we consider only luminosities of @xmath5 ergs / s as true ulx .
this definition corresponds to the eddington luminosity of an accreting compact object of about 80 times the chandrasekhar mass . the brightest galactic x - ray binary , grs 1915 + 105 , occasionally reaches @xmath6 ergs / s , and should then , according to the earlier definition be classified as a ulx .
grs 1915 + 105 , however , contains a rather ordinary @xmath7 black hole ( harlaftis & greiner 2004 ) , which has an eddington limit of 3.5@xmath8 ergs / s if it accretes helium , and a factor of two lower if it accretes hydrogen rich material .
we therefore argue that the term ultraluminous is inflationary , and opt for a more stringent definition ( see also irwin , bregman & athey 2004 ; irwin et al .
2004 ) . the definition with @xmath9erg / s is thus chosen because it is difficult to explain these sources with conventional x - ray binaries containing stellar mass black holes ( see also kalogera et al . 2004 ) .
just the x - ray luminosity is not sufficient to identify a ulx as an imbh .
a few studies have claimed spectral evidence for intermediate mass black holes , in the form of relatively low temperature accretion disks ( i.e. @xmath10@xmath11 kev ) which can not produce the observed high x - ray luminosities unless the accretion radius is quite large , and hence flowing into massive black holes ( miller et al .
2003a , b ) . on the other hand , such spectral results are generally based on fitting the disk emission ( shakura & sunayev 1973 ) using the models of mistuda et al . ( 1984 ) and gierlinski et al . ( 1999 ) including a power law , which presumably fits the compton up - scattered emission at high energies ( thorne & price 1975 ; shapiro , lightman & eardley 1976 ; sunyaev & trmper 1979 ) . inferring the black hole mass from the inner disk radius measured by spectral fits typically works well for sources whose
x - ray fluxes are dominated by a quasi - thermal component , but less well for sources dominated by the power law component ( see e.g. sobczak et al .
as the ulxs for the spectral mass measurement technique has been applied have spectra dominated by the power law component , these results should be taken with some caution . a small subset of extragalactic ulxs show properties that are unlike x - ray sources in the milky way .
these objects provide some indication for the existence of either intermediate mass black holes or for unusual modes of accretion .
many of these bright x - ray sources are variable , which helps provide confidence that the emission comes from a single accreting object , and some also show evidence of associations with counterparts in other wavelengths , such as star clusters ( zezas & fabbiano 2002 ; smith & wilson 2003 ) or optical emission nebulae ( shih , iwasawa & fabian 2003 ) or even bright stars ( liu , bregman , seitzer 2004 ) , which makes them unlikely to be background objects .
there are a number of theoretical models for these sources which do not require an intermediate mass black hole , but rather use the fact that the eddington limit applies strictly to spherical accretion .
the non - imbh explanations include mechanical beaming ( king et al .
2001 ) , relativistic beaming from a jet ( krding , falcke & markoff 2002 ) or photon bubble instabilities ( begelman 2002 ) ; many of these configurations can be obtained from a rather ordinary 10 black hole accreting from a donor during case b or c mass transfer ( podsiadlowski et al .
a review of the theory of ulxs can be found in king ( 2003 ) .
one particular source , m 82 x-1 , represents an especially compelling candidate for hosting an intermediate mass black hole ( ebisuzaki et al 2001 ; portegies zwart et al .
it is located sufficiently close to a massive young star cluster that a chance superposition is unlikely , so it can be claimed to truly reside in m 82 with considerable confidence .
it has an x - ray luminosity of about 5@xmath12 ergs / s ( i.e. well in excess of the luminosity of any galactic x - ray binary ) , and shows a 54 mhz quasi - periodic oscillation ( qpo ) as well as a broad iron line ( strohmayer & mushotzky 2003 ) .
they argue that in a mechanical beaming scenario , the qpo should be broadened by scatterings which increase the photons light travel times , while the iron line should be narrow , since lines are typically broadened due to projected rotation effects . in a relativistic beaming scenario , neither qpos nor iron lines
are expected .
furthermore , they note that the observed qpo and noise spectrum are reminiscent of the galactic x - ray binaries in their very high state ( see van der klis 1995 for a review ) , except that the qpo frequency is a factor of about 10 - 100 lower than those seen in the galactic x - ray binaries . applying a simple linear mass scaling to the qpo frequency implies then a mass of about 100 - 1000 @xmath13 for the black hole . for a characterization of the bright x - ray sources it is important to identify models for their origin and evolution .
for the remainder of the paper we adopt the intermediate mass black hole hypothesis and study the characteristics of an x - ray binary in which such a black hole accretes from a stellar companion .
our assumptions are motivated by the compelling case of m 82 x-1 .
the configuration in mind can be compared to low - mass x - ray binaries in which a main - sequence star or giant fills its roche - lobe to a black hole , in this case , however , the accreting object is a black hole o(1000 ) . in this paper
we discuss the consequence of a @xmath14 black hole that accretes from a stellar companion .
we envision that such systems may form in the dense cores for young star clusters .
we discuss the formation of such systems in sect.[sect : fande ] .
we further evolve the binary system through several phases of roche - lobe overflow ( rlof ) .
for this purpose we perform detailed binary evolution calculations for such systems and demonstrate that if the binary can form with sufficiently small orbital period a stable low - mass x - ray binary forms in which the donor is slowly consumed by the black hole . the x - ray emission while the @xmath15 donor is on the main - sequence reaches luminosities above the eddington limit for a 1 accreting compact object . in the shorter post main - sequence phase
less massive donors also exhibit super - eddington luminosities , while the higher mass donors may become ultraluminous for a few million years . for larger orbital periods , rlof starts while the donor is on the giant branch , which can lead to very high x - ray luminosities for a short while .
intermediate mass black holes with a lower mass donor may contribute to the emission of gravitational waves in the lisa band and simultaneously be visible as bright x - ray sources .
several theoretical models exist for producing black holes of intermediate mass , ( between @xmath16 ; portegies zwart & mcmillian 2002 ; miller & hamilton 2002 ; madau & rees 2001 ) .
young and dense star clusters may experience an evolutionary phase in which the stellar density in the core becomes so high that physical collisions between stars become frequent ( quinland & shapiro 1987 ; portegies zwart et al 1999 ; rasio , freitag & grkan , 2003 ) .
portegies zwart & mcmillan ( 2002 ) derive a mass accumulation rate due to collisions of @xmath17 , with @xmath18 the mass of the star cluster and its half - mass relaxation time .
the requirement for this mass coagulation is that the initial relaxation time of the cluster is less than @xmath19myr , otherwise the most massive stars which drive the collision runaway experience supernovae before the star cluster experiences core collapse .
if the cluster is born with sufficiently concentrated , a longer relaxation time up to about 100myr can still lead to a phase of collision runaway ( portegies zwart et al 2004 ) .
the final mass of such a episode of coagulation is about 0.1 percent of the initial cluster mass portegies zwart & mcmillan ( 2002 , see also grkan et al 2004 ) . following this scenario , intermediate mass black holes populate the central regions of massive and dense star clusters . but the presence of an intermediate mass black hole is no guarantee for x - ray emission .
the ambient gas density in such clusters is generally rather low and is insufficient to power an ultraluminous x - ray source
. another possibility would be the tidal disruption of a main sequence star that smashes into the black hole or grazes it near the tidal radius . as a result
, the star may be completely destroyed and accreted by the black hole .
tidal disruption with a supermassive black hole leads to short flares ( lasting about a year , rees , 1988 ; ulmer , 1999 ; ayal , livio & piran , 2000 ) . for lower mass black holes the duration of accretion will longer inversely proportional with the mass of the compact object . from this point of view
, such a scenario is consistent with the properties of ulx , but the event rate must be impractically high to explain the number of ultraluminous x - ray sources which have been observed , like m82-x1 .
a more attractive model was recently proposed by hopman , portegies zwart & alexander ( 2004 ) .
they argue that in a young and dense cluster a star may be captured by the intermediate mass black hole in a tidal event .
further tidal interaction between the black hole and the captured star then circularize the orbit .
they further assume that the captured star is on the main - sequence , but the same argumentation can be made for evolved stars . as a orbit has been fully circularized the captured star under - fills its roche - lobe only slightly ( hopman et al 2004 ) . during the remainder of the main - sequence lifetime of the captured star it grows in size by about a factor of two and
gravitational wave radiation reduces the orbital separation .
ultimately the star fills its roche lobe and mass transfer to the intermediate mass black hole starts .
an alternative scenario for the formation of a binary is by the dynamical capture of a main - sequence star by the intermediate mass black hole .
such events will typically result in rather wide ( 100 1000 day ) binaries ( baumgardt et al . 2004 ) .
the companion to the intermediate mass black hole is then be among the most massive stars in the cluster . at young age ( @xmath20myr ) the companion and may well be a massive ( @xmath21 ) main - sequence star , but for older cluster the companion is more likely to be another black hole .
the star clusters in which either tidal or dynamical capture can take place are very young ( @xmath22myr ) , extremely dense @xmath23pc@xmath24 and rich @xmath25 stars ( portegies zwart et al .
the relatively recently identified class of young dense clusters ( sometimes pronounced as yodec ) tend to populate active star forming regions in interacting galaxies , like m82 and the antennae ( zezas & fabbiano 2002 ) .
hopman et at .
( 2004 ) argue that in about half of the star clusters with are able to produce an intermediate mass black hole via collision runaway , the back hole is able to capture a companion star in a sufficiently short orbital period binary that roche - lobe overflow starts some time on the main sequence .
together with the possibility of dynamical capture , wide binaries with an intermediate mass black hole may be comparatively common as short period binaries .
we do not further consider the details of dynamical or tidal capture , nor do we assess the probability of forming the type of binary systems discussed in this paper , as we mainly concentrate on the phase in the lifetime of the intermediate mass black hole at which it is visible in x - rays or as a gravitational wave source .
here we assume that an intermedate mass black hole was formed and has acquired a stellar companion via one of the mechanisms discussed above .
mass transfer from a lower mass ( secondary ) star to a much more massive ( primary ) star is driven by the expansion of the donor and the loss of angular momentum from the binary system . in our evolutionary model ,
we initialize the binary systems with a black hole mass ( typically 1000 ) and a stellar mass secondary of 2 5 , 10 or 15with some selected orbital period at birth ( @xmath26 ) . apart from the previous model description in
[ sect : formation ] we do not further discuss how such parameters are obtained .
the evolution of each selected secondary mass is calculated with an initial orbital period such that roche - lobe overflow starts at zero - age , which could be representative if the system was formed via tidal caputre .
this choice seems somewhat arbitrary , as it leaves no time for the imbh to form and to acquire roche - lobe contact ; to first order however , this is a good assumption as it takes less than about three million years to form the imbh ( portegies zwart et al 2004 ) , and the capture of a companion star followed by tidal secularization takes only a few million years ( hopman et al , 2004 ) . for our 2 donors this time lag is negligible , while for the 15 donor it is roughly half the main - sequence lifetime .
we still argue , however , that the exact moment of rlof has little effect for the further evolution of the binary system , as long as it starts somewhere on the main - sequence .
some binaries start roche - lobe overflow at a later evolutionary stage , on the giant or super giant branches
. such larger initial orbital period could be representative in the case that the binary was formed by a dynamical capture .
an overview of the selected initial conditions is presented in tab.[tab : models ] .
once initialized , we evolve the binary through various stages using the binary evolution code of eggleton ( see pols et al .
1995 and references therein for a description ) , assuming a population i chemical composition ( y = 0.98 , z = 0.02 ) and mixing - length parameter of @xmath27 .
convective overshooting is taken into account using an overshooting constant @xmath28 ( pols et al .
1998 ) .
the orbital evolution of the system is affected by the emission of gravitational waves ( landau & lifshitz 1958 ) , roche - lobe overflow by non - conservative mass transfer or mass loss via a wind ( soberman et al .
the evolution of the orbital separation , @xmath29 , can then be computed by the sum of three terms @xmath30}{1 + q } \frac{\dot{m}_{{\rm d}}}{m_{{\rm d } } } \nonumber \\ & & - 2 \frac{\dot{m}_{{\rm bh}}}{m_{{\rm bh } } } + \frac{\dot{m}_{{\rm t}}}{m_{{\rm t}}}. \label{doublens : eq : orbit } \end{aligned}\ ] ] here @xmath31 is the constant of gravity , @xmath32 is the speed of light in vacuum , @xmath33 and @xmath34 are the masses of the donor star and the black hole , respectively , @xmath35 and @xmath36 .
the parameter @xmath37 gives the fraction of mass lost from the donor star in the form of fast isotropic wind ( de jager et al .
1988 ) and @xmath38 is the fraction of the wind mass accreted by the black hole . during the detached phase , we assume that @xmath39 and that the black hole does not accrete mass from the stellar wind ( @xmath40 ) .
matter leaves the system carrying the specific angular momentum of the donor . during mass - transfer
we assume that the black hole accretes matter up to its eddington limit ( see e.g. king 2000 ) as @xmath41 the remaining mass is lost from the system with the specific angular momentum of the black hole .
a more sophisticated eddington mass accretion is given by podsiadlowski et al .
( 2003 , see their eq.9 ) , which depends on the hydrogen abundance of the donor and the initial and present mass of the black hole .
because in our calculations @xmath42 , throughout the evolution the present mass of the black hole does not differ much from its initial mass , and therefore our adopted eddington limit is about a factor @xmath43 4.3 lower than that of podsiadlowski et al .
as the consequence , the black holes in our case accrete less mass and the orbital period expands more strongly .
table[tab : models ] gives an overview of the initial conditions and results of our calculations . in fig .
[ fig : tmdotedd_zams ] we present the evolution of mass - transfer rate for our calculations where mass transfer started at the zero - age main sequence . binaries which start roche - lobe contact at the zams undergo case a mass transfer .
this prolongs their main - sequence lifetime considerably .
the most peculiar evolution in our sample is arguably that of the 2 donor . after transferring about 1
the star starts to develop a convective envelope which initiates chemical mixing , which again leads to magnetic activity . as a result orbital evolution
will become dominated by magnetic braking .
this mechanism ceases when the whole star becomes convective with a homogeneous composition of @xmath44 .
some time later , the central temperature drops below the hydrogen burning limit and thermonuclear fusion stops : this happens at about @xmath45myr . at that point
the system remains detached for about 8myr before it undergoes another phase of mass transfer . at @xmath46gyr
the code breaks while the donor still fills its roche lobe ; at this point @xmath47 840 k and @xmath48 0.011 ( about 10 @xmath49 ) , which can be considered a brown dwarf . under the influence of gravitational - wave radiation alone
, the donor spirals in toward the black hole until they merge after about 74 myr ( peters 1964 ) . for the more massive stars ,
case a mass transfer is followed by mass transfer when the star ascends the giant branch during hydrogen shell burning ( case ab ) , lasting for 8.3 myr ( for @xmath50 ) , 0.26myr ( for 10 ) and 0.05myr ( for @xmath51 ) .
the mass - transfer rate increases dramatically in this phase to 5.2 @xmath52 , @xmath53 and 2.3 @xmath54 /yr for the 5 , 10 and 15 donor , respectively ; as shown by the first spike in the evolution in fig.[fig : tmdotedd_zams ] .
these mass transfer rates result in x - ray luminosities in excess of @xmath55 erg / s .
the 10 donor experiencing case a mass transfer ends as a helium star . this star experiences helium shell burning , lasting for about 40000 years , during which mass transfer starts again at a rate of @xmath56 /yr , regardless of the black - hole mass .
these phases of mass transfer are shown by the second spikes in figs.[fig : tmdotedd_zams ] for 5 and 10 donors .
though seemingly unresolved , some of these spikes last for several million years and are well resolved in our simulations ( see also tab.[tab : models ] ) .
binaries initialized with a larger orbital period experience mass transfer when the donor is evolved to a giant ( case b ) or supergiant ( case c ) .
the case b systems remain detached for only 0.29myr ( 5 )
0.05myr ( 10 ) at mass - transfer rate of 2.4 @xmath57 ( 5 ) 7 @xmath54 /yr ( 10 ) .
the 10 donors which experience case b mass transfer ascends the super giant branch at a later instant , leading to case bc mass transfer .
the mass - transfer rate during this stage is of the same order as in the preceding phase .
the evolution of the 10 donor in a @xmath58 day orbit around a 1000 black hole is quite peculiar ; mass transfer , in this system , ensues in a relatively advanced stage ( late case b ) such that only about 0.5 is transferred before the hydrogen shell burning is terminated , causing the star to shrinks inside its roche lobe .
another 2 is transferred during case bc mass transfer before the donor turns into a helium star , which experiences helium shell burning , during which it expands quickly to overfill its roche lobe , initiating a short phases of high ( @xmath59 /yr ) mass transfer .
also the 5 donor ultimately turns in a degenerate co core and experiences several carbon flashes during its remaining lifetime during which the mass transfer rate varies between @xmath60 /yr and @xmath59 /yr . only one of our simulations experienced case c mass transfer , which occurred for the 5 donor with a 600 day orbital period .
this resulted in a short phase of mass - transfer at a rates of 1.4 @xmath57 /yr . only the binaries with a 10 donor for case b , and 15 for case a ,
leave neutron stars at the end of their evolution ; all other evolutions leave carbon - oxygen white dwarfs . [
cols="<,<,<,<,<,<,>,>,>,>,>,^,^ " , ] [ tab : models ] the durations of mass transfer are presented in columns 710 of tab.[tab : models ] . only if roche - lobe overflow occurs at an early stage , near the zams , can mass transfer be sustained at high rates for a long time .
high accretion rates , however are only achieved on the giant ( case b ) and supergiant ( case c ) branches .
we also find that mass transfer caused by helium shell burning can result in mass transfer rates as high as @xmath59 /yr , which is sufficient to power an @xmath9 erg / s x - ray source , but only for a short while .
in this section we apply and discuss an empirical observational model based on our understanding of low - mass x - ray binaries in the galaxy , but then applied to a system where the donor is a 215 star and the accretor is an intermediate mass black hole .
we then use the results of our binary evolution calculations as input for the empirical model . in appendix[sect : appendix ] we present the empirical evidence on which this model is based .
the binary evolution simulations provide us with the donor mass , accretor mass , orbital period and a secular mass accretion rate , all as a function of time .
we use these information to calculate the x - ray luminosity of the simulated binaries ( below ) ; the gravitational wave signatures are discussed in [ sect : gwr ] .
our binary evolution calculations provide secular accretion rates , rather than instantaneously observable ones .
this difference can be quite important , considering that a large fraction of the stallar mass accreting black holes are transients .
since the theoretical understanding of disk instabilities is not presently well enough established to make quantitative predictions , we develop an _ ad hoc _ model based on phenomenology ( discussed in appendix[sect : appendix ] ) .
the parameters we require in order to compare our simulations with observations are the recurrence time ( @xmath61 ) and the decay time ( @xmath62 ) , via which we calculate the time @xmath63 that the source is brighter than a certain limiting x - ray luminosity @xmath64 .
we then compute the duty cycle ( @xmath65 ) . after applying the empirical model for x - ray emission , we study the observational characteristics of these sources .
we require that to recognize the source as a ulx it has to be brighter than @xmath64 , for which we adopt @xmath66 erg / s .
this is done in the following way : first we calculate the total energy from a fit to the empirical data in fig.[fig : porb_e ] and the decay time from fig.[fig : ltr ] .
the recurrence time is subsequently a function of the total energy @xmath67 and the rate of mass transfer via , @xmath68 . the bright time is subsequently calculated by integrating the luminosity curve , for which we assume an exponential decaying function ( see appendix[sect : appendix ] ) from the start to the moment it drops below @xmath64 .
this results in @xmath69 , where we adopt the factor two in the denominator as seems reasonable from fig.[fig : ltr ] .
the duty cycle is subsequently determined as @xmath70 . in figures[fig
: bright_time ] and[fig : duty_cycle ] we present the resulting bright time ( @xmath63 ) and duty cycle ( @xmath71 ) for a selection of binaries with 2 , 5and 10 donors which start rlof at the zams . the binaries which start at wider orbits and the simulations for a 15 donor which starts rlof at the zams are all persistent x - ray source with luminosities well above @xmath64 .
the binaries with a relatively wide initial orbit , however , remain bright for only a short while .
our empirical prescription is supported by the qualitative similarities with the proposed limit for transient and persistent sources of dubus et al .
( 1999 ) , which yields that binaries with accretion rates exceeding @xmath72 are persistent x - ray emitters . with this prescription ,
lower mass donors ( @xmath73 ) show transient behavior .
we have analyzed the results of evolutionary calculations in which a donor star of 2 15 transfers mass to a 100 2000 black hole . in this section
we discuss our simulation results with respect to the observed bright x - ray source m82-x1 in the starburst galaxy m82 .
this source is one of the best candidate intermediate mass black holes , and we argue that such a black hole may produce x - rays as it receives mass from a companions star . binaries which start rlof at the zams remain bright only for a few weeks continuously . near the end of the main sequence
the bright time increases to several years only for the 5 and 10 donors .
our calculation for a 2 donor only experiences a short ( few days ) bright ( @xmath74 ) flash every few months .
since the ultraluminous x - ray source m82-x1 was observed for an extended period ( over 20 years ) we conclude that it is unlikely to host a low mass ( @xmath75 ) donor . of course
, it could be that the x - ray source was caught near peak luminosity each time it was observed , in which case a lower mass donor would prove satisfactory , but this is unlikely .
our simulations for the 5 and 10 donor on the main - sequence exhibit similar episodes of bright time but with shorter duty cycle , making these sources still rather unlikely candidates for the ulx in m82 .
however , near the end of the main sequence , @xmath63 increases considerable to several years .
the binaries with a 10 donor then become steady sources , making these suitable candidates for the ulx in m82 .
the 15 donor which starts rlof at the zams is also persistently bright . strictly speaking , the systems with an evolved donor are transients , but for as long as rlof is sustained the bright time and recurrence time become so long that these sources could rather easily be identified as persistent ultraluminous x - ray emitters . and
though these sources are rather short lived they should not be excluded as suitable candidates ( see also kalogera et al 2004 ) .
based on the above analysis and the figures[fig : bright_time ] and [ fig : duty_cycle ] we conclude that within the limitations of our model the most likely configuration for the bright x - ray source m82-x1 is a 1000 with a 10 - 15 donor , probably near the end of its main - sequence lifetime .
note that for the star cluster mgg-11 this would also be consistent with the observed age of the cluster , being 7 - 12myr .
the turn - off age of a 15 star is about 11.5myr . in a phase of mass transfer
the stellar lifetime can be extended considerably possibly allowing even more massive donors ( see section[sect : binev ] ) .
a binary emits gravitational waves , and as long as the stars are far apart we can assume that they behave as point masses .
in the case we consider here , the donor star fills its roche - lobe and can hardly be described as a point mass .
the amplitude of the gravitational wave signal is given by ( evens et al .
1987 ) @xmath76 here @xmath77 is the chirp mass .
the gravitational wave frequency is twice the orbital frequency .
the resulting gravitational wave signal is presented in fig.[fig : tmdotedd_gwr ] , for binaries which started roche - lobe overflow on the zero - age main - sequence .
binaries which do not start rlof during the main - sequence phase of the donor will not become detectable gravitational wave sources .
figure[fig : tmdotedd_gwr ] presents the evolution of the gravitational wave signal for several systems which start rlof at the zams .
the last column of tab.[tab : models ] presents the time that the binary is visible above the noise curve of the lisa gravitational wave detector . assuming a standard distance of 10kpc , the binaries which start mass transfer at birth emit gravitational waves at a strain of about @xmath78 ( almost independent of the donor mass ) .
the frequency of the gravitational waves is barely detectable for the _ lisa _ space based antennae ( see figure[fig : tmdotedd_gwr ] ) . during mass transfer the binary moves out of the detectable frequency band , as the orbital period increases . once the donors has turned into a white dwarf , as happens for the 5 and 10 donors , the emission of gravitational waves brings the two stars black into the relatively high frequency regime and it becomes detectable again for the _ lisa _ antennae .
this process , however takes far longer than a hubble time , unless the orbit has an eccentricity @xmath79 .
such high eccentricities , after the phase of mass transfer , can only be achieved in the cases where the initial donor collapses to a compact object in a supernova explosion , or if the binary is perturbed by external influences .
we conclude that the here discussed class of binaries with relatively massive donors @xmath15 are probably not important sources of gravitational waves .
the binary in which a 2 main - sequence star starts to transfer to a 1000 black hole , however , remains visible as a bright source of gravitational waves for its entire lifetime ( see figure[fig : tmdotedd_gwr ] ) .
the reason for this striking result is the curious evolution of the donor ( see sect.[sect : binev ] ) .
mass transfer in this evolutionary stage is rather slow causing the x - ray source to be transient .
we argue that such a transient could result in an interesting synchronous detection of x - rays and gravitational waves .
the known ulxs are at mpc distance and unlikely to be seen with lisa .
we performed detailed simulations of the evolution of close binaries with an intermediate mass ( 1001000 ) black hole and a stellar companion .
if the initial orbital period is of the order of a day , the main - sequence star starts filling its roche lobe before the terminal age . in this case
the binary remains in contact for it s remaining main - sequence lifetime , which is considerably extended compared to the main - sequence lifetime of an isolated star of the same mass . relatively low mass ( @xmath80 ) donors which start mass transfer at zero age are likely to be transient , with high ( @xmath81erg / s ) peak luminosity .
the recurrence time for these sources , however , is generally rather long ( several weeks ) and the duration of the bright phase is short ( few days ) .
more massive donors are bright for a longer time and have shorter recurrence time .
the sources become persistent for donor stars of about 10 .
our simulation with a 15 zams donor resulted in a persistent bright x - ray source for its entire main - sequence lifetime .
the duty cycle also increases near the end of the main - sequence .
stars which start roche - lobe overflow after the main - sequence phase remain persistent for their mass transferring episode , which , for a 5 donor lasts for @xmath82 million years . based on these findings , and within the limitations of our analysis and assumptions
, we argue that low mass ( @xmath80 ) donors and donors near the zero - age main - sequence are unlikely candidates for the ultraluminous x - ray source m82-x1 . a 1015 donor near the terminal - age main sequence or on the early giant branch produces recurrence times and luminosities consistent with the observed ultra luminous x - ray source in m82 . an interesting possibility is provided by a @xmath83 donor which starts to transfer mass to an intermediate mass black hole at birth .
such a binary is likely to be visible as a bright transient x - ray source and simultaneously as gravitational wave source in the _ lisa _ band to a distance of several mpc .
we are grateful to chris belczynski , clovis hopman , peter jonker , vicky kalogera , emmi meyer - hofmeister , onno pols , ed van den heuvel , lev yungelson , andrzej zdziarski , steve zepf and andreas zezas for discussions and for providing useful data .
this work was supported by the royal netherlands academy of sciences ( knaw ) , netherlands research school for astronomy ( nova ) and the dutch organization for scientific research ( nwo ) for the spinoza grant 08 - 0 to e. p. j. van den heuvel and a talent fellowship .
we are grateful to northwestern university for hospitality .
in this appendix we discuss the characteristics of low - mass black hole x - ray binaries in the galaxy , in order to compare them with the x - ray binaries we have been simulating in the main body of the paper .
the main reasons to use the galactic population of soft x - ray transients is because of the similarities in the geometry and composition of these systems with accreting intermediate mass black holes .
most accreting black holes in the milky way are transient systems . in fact
, only cygnus x-1 is a persistent emitter among the dynamically confirmed black hole candidates ( see mcclintock & remillard 2003 for a review ) , and the only other persistent emitters considered to be plausible black hole candidates are 4u 1957 + 115 , ss 433 and cygnus x-3 , along with the large magellanic cloud sources lmc x-1 and lmc x-3 .
of these , only 4u 1957 + 115 has a low mass ( i.e. @xmath84 ) donor star , and the case for its being a black hole is by far the weakest ( see nowak & wilms 1999 ) . on the other hand ,
there are many persistent neutron star accretors . since standard accretion disk theory suggests that disk instabilities occur due to changes in the disk s hydrogen ionization state , we expect stable disks only when the disk is hot enough for its outer edge to be above the ionization temperature of hydrogen .
given that more massive accretors have lower inner disk temperatures ( shakura & sunyaev 1973 ) and larger orbital separations ( and hence bigger disks ) for the same orbital period ( kepler 1609 ) , it is then not too surprising that neutron stars can more easily be persistent accretors than black holes .
it thus becomes rather likely that intermediate mass black holes will be transient accretors ( see the discussion in dubus et al .
1999 ) .
the most comprehensive study to date of the outburst properties of soft x - ray transients is that of chen , shrader & livio ( 1997 , hereafter csl97 ) .
this work compiled the outburst durations , total fluencies , and peak luminosities for all the observed x - ray transient outbursts in the literature at the time of its writing , considering both optical and x - ray properties of the outbursts .
we will borrow a few key results from that paper , while also making a few simplifications , to allow for a prescription for converting secular accretion rate into x - ray luminosities that is suitable for porting to accreting intermediate mass black holes .
we will supplement these data with some more recent work , including that by garcia et al .
( 2003 ) who tabulated the peak outbursts luminosities of several x - ray transients , and other data sets which will be specified in more detail .
the frequencies of outbursts are poorly constrained both by observations and by theory . a reasonable approximation , then , is to assume that the outbursts we see are `` typical '' , and hence that roughly periodic , identical outbursts will be seen for sources , with the periodicity and the properties of the outburst varying from source to source . to estimate the recurrence timescale ( @xmath61 ) , we compare the secular mass accretion rate from our binary evolution simulations ( [ sect : binev ] ) with an estimate of the total amount of mass accreted during an outburst based on the energy emitted during the outburst , @xmath67 , plus an estimate of the radiative efficiency .
we then apply the observational characteristics to our binary evolution calculations , which is done in [ sect : observations ] .
the first assumption we make is that the outburst follow a fast - rise , exponential decay ( fred ) luminosity law as a function of time .
this approximation is supported by the shape of the light curves for about half the outbursts studied in chen , shrader & livio ( 1997 ) .
the other half , and in particular grs 1915 + 105 the galactic system most analogous to a ulx , show rather `` flat - topped '' outburst light curves ( chen , shrader & livio 1997 ) .
the next assumption is that the orbital period of a system is the key parameter in determining the properties of its outbursts .
the radial size of the accretion disk is likely to be the actual driver of the outburst properties , but it is hard to measure . given simple scaling laws , though , and the assumption that the compact accretor dominates the mass of the binary system , which is here the case , it can be shown that the mass content of an accretion disk depends primarily on the orbital period , with a linear dependence on mass of the accretor . the outer radius of the disk , @xmath85 then scales with the orbital separation , @xmath29 , since @xmath86@xmath87 ; and the mass of the disc is proportional to @xmath88 ;
then the mass of the accretion disk is proportional to the mass of the accretor .
if we then scale all observational quantities in eddington units , the mass of the accretor has relatively weak effects on the properties of the outbursts .
fig.[fig : ltr ] shows the peak luminosity at outbursts as a function of the orbital period for several black hole x - ray binaries ( data from garcia et al .
there are relatively few systems , but the systems in short orbital periods ( i.e. less than about 10 hours ) fit approximately to the relation @xmath89 , with the sources above 10 hours fitting approximately to @xmath90 , which is also supported by theoretical studies , which suggests that the amount of mass accumulated in the disc in an outburst cycle is a function of the orbital period with a cutoff around 10 hours ( see meyer - hofmeister & meyer , 2000 ; meyer - hofmeister , 2004 ) .
we note that the systems used to make the correlations between outburst properties and orbital periods appear to be prediminantly at most only mildly evolved , so that the applications of these results to highly evolved systems ( i.e. giant donors ) could be dangerous .
the two outliers are v4641 sgr , at 67.6 hours orbital period , and @xmath91 and xte j 1118 + 480 , with @xmath92 hours and @xmath93 .
since v4641 sgr shows a radio jet with an apparent proper motion of more than 10 times the speed of light ( hjellming et al . 1999
; orosz et al .
2001 ) its jet must be pointed very close to the line of sight , and if its jet has a substantial x - ray spectral component , then this beamed component may be responsible for the strong x - ray emission , in contrast to what is expected for the other sources .
the other outlier , xte j 1118 + 480 , is one of the x - ray transients that never reached a high luminosity state ( see e.g. brocksopp , bandyopadhyay & fender , 2004 , for a discussion of such sources ) .
since these low luminosity systems are thought to be radiatively inefficient , it may be that the instantaneous mass transfer rate did reach levels comparable to the level of the correlation ( indicating that the disk outburst physics was essentially the same for this source as the others ) .
alternatively , since the outbursts of individual recurrent x - ray transients are not all identical , it may simply be that we saw a fainter than typical outburst of xte j 1118 + 480 .
given that the primary goal of this paper is to develop a model for _ bright _ x - ray binaries , this outlier is not particularly troublesome .
next , we consider the outburst durations . taking the data from csl97 for black hole binaries outburst energies
, we find a similar relation with orbital period , that @xmath94 , as shown in fig.[fig : porb_e ] .
we note that the strong outlier is gro j 1655 - 40 , which has shown multiple outbursts with a short recurrence time , highlighting the limitations of the `` periodic outburst '' hypothesis , which we use here as a simplification .
we also note that the highest energy , longest period system , grs 1915 + 105 has been in outburst continuously since about 1992 , but the data from chen et al .
( 1997 ) obviously covers only a fraction of that timespan . using the value from chen et al .
( 1997 ) , it falls a factor of about 3 below the relation we propose between orbital period and outburst energy . combining the two empirical relations from the data , with the added assumption of a `` fred '' lightcurve shape
, we then find that the exponential decay timescale @xmath62 for an outburst should be approximately @xmath95 we emphasize that the prescription presented here is a crude approximation to the poorly understood phenomenology of transient disks ; however , we feel it is a better approximation that the standard ones made in converting secular accretion rates into instantaneous mass accretion rates in the x - ray producing regions of the accretion flow .
the theoretical model recently proposed by meyer - hofmeister ( 2004 ) for outbursts also predict a correlation between orbital period and peak output luminosity .
we note especially that the fred shape may be a poor approximation for the ulxs , considering that grs 1915 + 105 , the galactic system most analogous to the ulxs , has shown a rather `` flat - topped '' outburst light curve . based on this phenomenological analysis
we conclude that the peak luminosity for a @xmath1black hole binaries easily exceed @xmath55 erg / s , but these high luminosities are generally sustained for a short while only ( see also [ sect : observations ] ) .
next , we provide a brief review of spectral states of x - ray binaries , which will be used to justify bolometric corrections to the data .
between three and five distinct spectral states exist for black hole x - ray binaries , depending on the classifying author .
these states are essentially accretion modes with distinct spectral and variability characteristics .
the state of an object is , to first order , correlated with its mass accretion rate , with often some hysteresis ( miyamoto & kitamoto , 1995 ; nowak , wilms & dove 2002 ; smith , heindl & swank 2002 ; maccarone & coppi 2003 ) . for a more detailed review , we refer to nowak ( 1995 ) which concentrates more on the x - ray spectroscopic features and van der klis ( 1995 ) which concentrates more on the variability features .
the three canonical spectral states are the low / hard state ( typically seen above below 2% of the eddington luminosity
maccarone 2003 ) , the high / soft state ( typically seen between about 2 and 30% of the eddington luminosity ) , and the very high state ( typically seen above about 30% of the eddington luminosity
nowak 1995 ) .
the two additional states are the quiescent state , which may just be the low luminosity extension of the low / hard state , and the intermediate or transition state , which shows spectral and timing properties essentially the same as the very high state , but occurs at lower luminosities , most notably as the source moves between the low / hard state and the high / soft state ( see homan et al .
2001 for a discussion of the intermediate state ) .
active galactic nuclei seem to show similar spectral states as black hole binaries at the same fractions of the eddington luminosity ( maccarone , gallo & fender 2003 ) , so the extension of these results to intermediate mass black holes seems to be a reasonable assumption . for illustrative purposes we present figure[fig : hardsoft ] , which shows the x - ray spectrum of cygnus x-1 in the soft state ( top solid curve for low @xmath22kev and high energies @xmath96kev ) and in the hard state ( bottom curve in the _ chandra _ band ) .
the dashed curves are corrected for absorption .
the two lowest luminosity states are the quiescent state and the low / hard state .
no clear indications exist that there are fundamental differences in the accretion mode between these two states , and it has been suggested that the quiescent state may be merely the lowest luminosity low / hard state .
the x - ray spectra of low / hard states are typically well fit by thermal comptonization models with photon indices , @xmath97 ( defined such that the number of photons per unit energy , @xmath98 of 1.52.0 and roughly exponential cutoffs at about 200 kev .
the quiescent state objects are typically so faint that their spectral states can not be well constrained but they seem to have power law slopes consistent with those of the low / hard state objects ( kong et al .
additionally , since their spectra have been measured only with soft x - ray imaging telescopes , no spectral cutoffs have been seen from the quiescent black holes .
taking typical black hole low / hard state spectral parameters ( e.g. nowak et al .
2002 ; mcclintock et al .
2001 ) the bolometric corrections from the _ chandra _ band are factors of @xmath43 515 . if a x - ray source is in the low / hard state an observed luminosity of @xmath99ergs@xmath100 in the _ chandra _ bandpass indicates that the total luminosity should exceed @xmath101ergs@xmath100 ; the black hole then should exceed about 2000 in order for this luminosity to be less than 2% of the eddington luminosity .
another key issue for converting mass accretion rate into observed luminosity for the low luminosity states is the radiative efficiency correction .
geometrically thin accretion disks are generally believed to be radiatively efficient . defining the radiative efficiency @xmath102 , it is generally found that @xmath103 for non - rotating black holes accretion disks , and a factor of @xmath104 higher for disks around maximally rotating kerr black holes . on the other hand , thick accretion disks , such as those thought to be responsible for low / hard states ,
may be quite radiatively inefficient . in the context of advection
dominated accretion disks , it is found that @xmath105 ( esin , mcclintock & narayan 1997 ) , where @xmath106 is the mass accretion rate at the luminosity of the state transition between the low / hard and high / soft states .
the correlation between radio and x - ray luminosity in low / hard state black hole binaries ( gallo , fender & pooley 2003 ) has been interpreted that the radio jet takes kinetic power from the accretion flow in such a manner as to also produce a @xmath107@xmath108 ( fender , gallo & jonker 2003 ) , although the coefficient of proportionality was not well constrained by the data .
the other two ( or three ) states all have similar spectral properties .
they show dominant strong thermal emission at low energies , consistent with a standard thermal accretion disk spectrum with a temperature of @xmath109 kev at the inner edge , and a weaker power law component with a spectral index of @xmath110 and typically no observable cutoff .
the very high state and the intermediate state typically show a slightly higher fraction of their luminosity coming out in the power law component , but in reality the difference between the two states is best seen from the variability properties , as a strong @xmath111 hz qpo is seen in the very high state , while the high / soft state typically shows root - mean square fractional variability less than 5% .
we take typical source parameters from miller et al .
( 2001 ) who presented tabulated fits to the rxte spectra for an entire outburst of the x - ray binary xte j1748 - 288 .
we then impose a low energy cutoff to the power law at 2 kev , where the flux is dominated by the thermal component , in order to keep the spectral model from diverging at low energies ( the exact location of this cutoff is unimportant so long as it is in the photon energy range where the luminosity is dominated by the thermal component ) .
we then compare the bolometric luminosity from the spectral model to the bolometric luminosity in the _ chandra _ bandpass and find that about 75% of the luminosity is within the _ chandra _ bandpass for the very high state and about 90% for the high / soft state .
bolometric corrections are therefore not a major source of uncertainty in converting between accretion rates and luminosity for the high luminosity sources .
additionally , we wish to take the opportunity to make a few points about the spectral states of x - ray binaries that may help explain certain phenomenology of ulxs , and especially some points which seem to contradict this picture , but in fact do not .
in particular , there has been discussion of low / soft and high / hard spectral state phenomenology .
we note first that with the _ chandra _ bandpass , a low / hard state will usually fit to a steeper power law spectral index than will a high / soft state ( e.g. liu et al .
this is because the peak of a 1 - 2 kev disk blackbody spectrum in power per decade will be at about 3 - 6 kev .
the exponential cutoff of the spectrum will be in the part of the _ chandra _ bandpass where the effective area is relatively low ; with a few hundred total photons , as is typical for ulx studies , the statistical weight of the spectral fit will be dominated by the softer x - rays , where the spectrum is quite flat . with very high signal - to - noise spectra , where one can actually rule out a power law model for the high / soft states or a disk blackbody model for the low / hard states
; only then is the spectral state identification useful for comparison with the black hole binaries .
we note that kubota , done & makishima ( 2002 ) have also pointed out that ulx spectra should not be compared with the low / hard states of accreting black holes , but rather with the more high luminosity states .
a second key related point is that in the high / soft state , the accretion disk seems to extend in to a constant inner radius ( see e.g. sobczak et al .
2000 ) , which is consistent with being the marginally stable orbit around the black hole .
the luminosity variations are then determined almost entirely by changes in the disk temperature . in the very high state ,
the disk may change size a bit more , but still the luminosity rises are correlated primarily with increases in the temperature of the disk and increases in the ratio of the power law component to the blackbody component .
that is to say , even in normal x - ray binaries , at high x - ray luminosities , the spectra frequently harden as the luminosity rises , but without the qualitative change in behavior associated with state transitions .
some recent work on bright well - studied ulxs show slight deviations from the simplistic picture presented for ulxs .
in particular , the hardness ratios of a bright source in the antennae galaxies are sometimes positively correlated with flux and sometimes inversely correlated with flux , with some indications that loops are traced out in a hardness versus intensity diagram ( zezas et al .
this is consistent with behavior seen in the galactic microquasar grs 1915 + 105 ( fender & belloni 2004 ) , where within the very high state , such loops are quite often seen .
this may be revealing some quite interesting physics , as discussed in the reference above , but is unlikely to have a major affect on the bolometric corrections we have described here .
we note further that in the outbursts of bright x - ray transients , the bulk of the mass accretion will occur during the high luminosity phases of the outburst .
computations of the outburst duty cycle , then can assume radiatively efficient accretion , and can assume the exponential decay continues down to arbitrarily small luminosity ( in reality , soft x - ray transients seem not to drop below about @xmath112 of the eddington limit
garcia et al .
2001 ) without seriously affecting the parameter values .
angelini , loewenstein & mushotzky , 2001 , apjl , 557 , l35 , s. , livio , m. , piran , t. 2000 , , 545 , 772 baumgardt , h. , portegies zwart , s.f . , mcmillan , s.l.w . ,
makino , j. , ebisuzaki , t. , 2004 , to appear in the proceedings of the `` formation and evolution of masive young clusters '' conference , cancun , mexico , 17 - 21 november 2003 ( astro - ph/0403140 ) .
begelman , m.c .
, 2002 , apjl , 568 , 97 brocksopp , c. , bandyopadhyay , r. m. , fender , r. p. , 2004 ,
newa , 9 , 249 chen , w. , shrader , c.r . ,
livio , m. , 1997 , apj , 491 , 312 coppi , p.s .
, 1999 , asp conf .
161 , p. 375
( poutanen & svenson editors ) , c. , nieuwenhuijzen , h. , van der hucht , k. a. 1988 , , 72 , 259 dubus , g. , lasota , j - p . ,
jameury , j - m . ,
charles , p. , 1999 , mnras , 303 , 139 , t. , makino , j. , tsuru , t. g. , funato , y. , portegies zwart , s. , hut , p. , mcmillan , s. , matsushita , s. , matsumoto , h. , kawabe , r. 2001 , , 562 , l19 esin , a.a . ,
mcclintock , j. e. , narayan , r , 1997 , apj 489 , 865 , g. , baldi , a. , king , a. r. , ponman , t. j. , raymond , j. , read , a. , rots , a. , schweizer , f. , zezas , a. 2004 , , 605 , l21 fabbiano , g. , 1998 , apj , 325 , 544 fender , r.p .
, gallo , e. jonker , p.g . , 2003 , mnras , 343 , l99 fender , r.p . ,
belloni , t. , 2004 , araa in press ( astro - ph/0406483 ) fillipenko & ho 2003 , apjl , 588 , 13 foschini , l. et al . ,
2002,a&a , 392 , 817 gallo , e. , fender , r.p . , pooley , g.g .
, 2003 , mnras , 344 , 60 garcia , m. r. , mcclintock , j. e. , narayan , r. , callanan , p. , barret , d. , murray , s. s. 2001 , apj , 553 , l47 garcia , m. r. , miller , j. m. , mcclintock , j. e. , king , a. r. , orosz , j. , 2003 , apj , 591 , 388 gierlinski , m. , zdziarski , a.a . ,
poutanen , j. , coppi , p.s . ,
ebisawa , k. & johnson , w.n . , 1999 ,
mnras , 309 , 496 green , j. , ho , l , 2004 , submitted to apj ( astro - ph/0404110 ) , m. a. , freitag , m. , rasio , f. a. 2003 , arxiv astrophysics e - prints harlaftis , e.t . & greiner , j. , 2004 , a&a let . , 414 , 13 hjellming , r. m. , rupen , m. p. , mioduszewski , a. j , et al . , 1999 , apj , 514 , 383 liu , j - f , bregman , j.n . , seitzer , p. , 2004 ,
apj , 602 , l249 , c. , portegies zwart , s. f. , alexander , t. 2004 , , 604 , l101 irwin , j.a . , bregman , j.n . &
athey , a. , 2004 , apjl , 601 , 143 jeltema , t.e .
, canizares , c.r . , buote , d.a . & garmire , g.p . , 2003 , apj , 585 , 756 , v. , henninger , m. , ivanova , n. , king , a. r. , 2004 , apj l41 , a. r. 2000 , , 312 , l39 king , a.r . , 2003 , in _ compact stellar x - ray sources _ king , a.r.,davies , m. b. , ward , m. j. , fabbiano , g. , elvis , m. , 2001 , apj , 552 , l109 krding , e. , falcke , h. , markoff , s. , 2002 , a&a , 382 , l13 kormendy , j. & richstone , d. 1995 , araa 33 , 581 , l. d. , lifshitz , m. 1958 , , pergamon press , oxford , london , new york , paris , p. , rees , m. j. 2001 , , 551 , l27 maccarone , t.j .
, gallo , e. , fender , r.p . , 2003 , mnras , 345 , l19 maccarone , t.j . ,
coppi , p.s . & taam , r. , 2000 , head , 32.2507 maccarone , t.j . , 2001 , ph.d .
thesis , yale university masetti , n. , foschini , l. , ho , l.c . ,
dadina , m. , di cocco , g. , malaguti , g. & palazzi , e. , 2003 , a&a let . , 406 , 27 mcclintock , j. e. and remillard , r. a. , 2003 astro - ph/0306213 meyer - hofmeister , e. , meyer , f. , 2000 , a&a 355 , 1073 meyer - hofmeister , e. , 2004 , submitted to a&a .
mcconnell , m. et al .
, 2000,apj , 543 , 928 miller , j.m . , fabbiano , g. , miller , m.c . &
fabian , a.c .
, 2003a , apjl , miller , j.m . , fabian , a.c . & miller , m.c . , 2003b , apj , submitted ( astro - ph/03 ... ) , m. c. , hamilton , d. p. 2002
, , 330 , 232 mitsuda , et al . , 1984 , pasj , 36 , 741 muno , m.p . , morgan , e.h .
& remillard , r.a . , 1999 ,
apj , 527 , 321 nowak , m.a . , 1995 , pasp , 107 , 1207 nowak , m.a . , wilms , j , 1999 , apj , 522 , 476 nowak , m.a . , wilms , j , dove , j.b . , 2002 , mnras , 332 , 856 nowak , m.a . , heinz , s. , begelman , m. , 2002 , apj , 573 , 778 miyamoto , s. , kitamoto , s. , 1999 , an , 320 , 355 orosz , j. , kuulkers , e. , van der klis , m. , mcclintock , j.e . , garcia , m.r . , callanan , p.j . , bailyn , c.d . , jain , r.k . ,
remillard , r.a .
, 2001 , apj , 555 , 489 pakull , m. , 2003 , iauc 194 peters , p. c. 1964 , phys .
, 136 , 1224 podsiadlowsky , ph . , rappaport , s. , han , z. , 2003 mnras 341 , 385 podsiadlowsky , ph . , rappaport , s. , pfahl , e. , 2002 , apj 565 , 1107 , o. r. , schroder , k. , hurley , j. r. , tout , c. a. , eggleton , p. p. 1998
, , 298 , 525 , o. r. , tout , c. a. , eggleton , p. p. , han , z. 1995 , , 274 , , s. f. , makino , j. , mcmillan , s. l. w. , hut , p. 1999 , , 348 , 117 , s. f. , mcmillan , s. l. w. 2002 , , 576 , 899 , s. f. , baumgardt , h. , hut , p. , makino , j. , mcmillan , s. l. w. 2004 , nature , 428 , 724 , g. d. , shapiro , s. l. 1987 , , 321 , 199 , m. j. 1988 , , 333 , 523 shakura , n.i . & sunyaev , r.a . , 1973 , a&a , 24 , 337 shapiro , s.l . , lightman a.b . & eardley , d.p . , 1976 , apj , 204 , 187 shih , d.c . , iwasawa , k. , fabian a.c , 2003 , mnras , 341 , 973 smith , d.m . , heindl , w.a . , swank j.h . , 579 , 2002 , apj , l129 smith , d.m . ,
wilson , 2003 , apj , 591,138 , g. e. , phinney , e. s. , van den heuvel , e. p. j. 1997 , , 327 , 620 spergel , d. et al .
, 2003 , apjs , 148 , 175 strohmayer , t.e .
mushotzky , r.f . , 2003 ,
apj , 586 , l61 sunyaev , r.a . & trmper , j. 1979 , nature , 279 , 506 thorne , k.s .
& price , r.h .
1975 , apjl , 195 , 101 , a. 1999 , , 514 , 180 van der klis , m. , 1995 , proceedings of the iau colloquium no .
151 ; held in sonneberg ; germany ; 5 - 9 december 1994 ; xxii ; 477 pp 321 , m. g. , stanger , v. , griffiths , r. e. 1984 , , 286 , 144 white , n.e .
& angelini , l. , 2001 , apjl , 561 , 101 , m. g. , stanger , v. , griffiths , r. e. 1984 , , 286 , 144 zdziarski , a.a . , grove , j.e . , poutanen , j. , rao , a.r . & vadawale , s.v . , 2001 ,
apjl , 554 , 45 zdziarski , a.a . &
gierlinski , m. , 2004 , astro - ph/0403683 zezas , a. , fabbiano , g. , 2002 , apj , 577 , 723 zezas , a. , fabbiano , g. , schweizer , f. , king , a. r. , baldi , a. , ponman , t. j. , raymond , j. c. , read , a. m. , rots , a. h , 2003 , baas 203 , 126.04
in the paper `` intermediate mass black holes in accreting binaries : formation , evolution and observational appearance '' by simon f. portegies zwart , jasinta dewi and tom maccarone ( mnras october [ 2004 ] ) contains an error causing the frequency of the gravitational wave sources to be too large by a factor @xmath113 .
this affects the discussion in 4.1 and the numbers in the right most column of tab .
1 . if instead of a distance of 10kpc one assumes a distance of 1kpc the new numbers in the right most column of tab .
1 become 1200myr , 0 , 10 , 0 , 0 , 0 , 0 , 0 , 2 , 0 , 0 , 0 , 0 , 1myr , from the top row to the bottom row . | we study the origin of the ultraluminous x - ray source m82-x1 in the nearby starburst galaxy m82 .
this x - ray source is of particular interest as it is currently the best candidate intermediate mass black hole ; it is associated with a 54mhz quasi periodic oscillations with a relatively low ( @xmath0kev ) black - body temperature .
we perform detailed binary evolution calculations of 215 stars which transfer mass to a 1002000 black hole and present an empirical model for the x - ray characteristics expected for such binaries . based on the binary evolution calculations and the assumption in our simulations
we conclude that the most likely candidate for the bright x - ray source m82-x1 is a 10 - 15 star near the end of its main - sequence or slightly evolved , which transfers mass to a @xmath1 black hole .
we expect the system to be in the high / soft state . in that case
the binary will not be visible as a source of gravitational wave radiation , but other transient x - ray binaries with lower mass donors way be rather bright sources of gravitational wave radiation .
[ firstpage ] black hole physics
binaries : close galaxies : individual : m82 x - ray : binaries x - rays : individual m82-x1 |
for some time flux tube models @xcite and wilson loop expansion calculations @xcite have been used to understand hadron states .
the usual assumption is that the color field can be taken to lie on straight lines connecting the quarks .
this `` straight line '' or `` rigid '' flux configuration is in the spirit of an adiabatic or born - oppenheimer approximation that is assumed to hold at least for slowly moving quarks .
nesterenko @xcite points out that this adiabatic approximation can not be valid if the quark has angular acceleration .
the argument is based on a classic theorem of ruled surfaces .
the nambu - goto string action requires the string to sweep out a surface of minimal area .
catalan s theorem @xcite states that if the string is straight there is only one minimal surface swept out ; a helicoid .
this surface describes a meson state consisting of uniformly rotating quarks .
as we will observe , a straight string with the end moving radially also sweeps out a minimal surface as long as the angular velocity is constant .
this solution does not correspond to the motion of an actual meson since the quark plus string angular momentum is not conserved .
however , one may take the point of view that the string shape can be studied with arbitrary end motion , such as caused by an arbitrary external force acting on the ends .
the motion of the string with quarks at the ends is then a special case . in this paper
we use the rigidly rotating solution as our starting point and consider angular acceleration of the endpoints as a perturbation .
a rigidly rotating string with its endpoint quarks moving in perfectly circular orbits is not a realistic model for meson because quantum mechanics requires there to be some radial motion , and by angular momentum conservation , some angular acceleration as well .
the string must be curved if it is to sweep out a minimal area while its endpoints undergo angular acceleration .
we will discuss here the shape of the curved string and conclude that small curvatures do not change its dynamics .
this result holds even for relativistic string motion . in section [ sec : ng ]
we define our notation for the nambu - goto action and obtain the string wave equation , angular momentum and energy relations .
we also establish exact straight string solutions with arbitrary radial motion but constant angular velocity .
an intuitive picture of a non - relativistic string is developed in section [ sec : nrstring ] .
the relativistic shape equation and solutions for small angular acceleration are established in section [ sec : relstring ] .
we demonstrate numerically in section [ sec : validity ] that size of the string deformations is small and therefore that the perturbative approach is sound .
the angular momentum and energy for the curved string are considered in section [ sec : constants ] and our conclusions are given in section [ sec : conclusion ] .
the string action is proportional to the string tension @xmath0 and to the area swept out by the string .
it is conventionally written in polyakov @xcite form as @xmath1 where @xmath2 is a two - dimensional metric auxiliary field whose indices run over @xmath3 and @xmath4 , and @xmath5 .
@xmath6 is the string position and @xmath7 .
the action is invariant under reparametrization of the coordinates @xmath8 and @xmath9 , and under weyl scaling of the metric @xmath10 because the metric @xmath2 has no dynamics , we may solve its field equations and substitute the result back into the action without changing the equations of motion for the other fields .
the equations of motion for @xmath2 , @xmath11 and invariance under weyl scaling ( [ weyl ] ) allow us to set @xmath2 equal to the induced metric @xmath12 from which we see that ( [ action ] ) reduces to , and is therefore classically equivalent to , the standard square - root nambu - goto action .
we will fix the coordinates @xmath3 and @xmath4 by choosing @xmath3 to be equal to coordinate time @xmath13 , and @xmath4 to run linearly along the length of the string and have a fixed interval @xmath14 regardless of the physical length of the string .
this simplifies the description of the end - points of the string , but unfortunately precludes the use of orthonormal coordinates . a more detailed explanation of the nambu - goto - polyakov string , its geometrical aspects , and its relation to qcd strings can be found in ref
. @xcite .
the boundary conditions we take on @xmath6 are that one end is fixed and the other is arbitrarily forced , @xmath15 where @xmath16 is a prescribed function of time .
this simplifies our presentation because we are only examining the piece of string from the center of mass to one of the quarks .
it is a trivial generalization to include quarks at both ends . to include the dynamics of a ( spinless ) quark on the end of the string
, we would use its equation of motion , @xmath17 which is essentially newton s second law with the tension in the string providing the force .
the equation of motion ( [ quarkeom ] ) arises from the variation of the boundary of the string when quarks are included from the beginning .
our purpose here is not to solve for the actual motion of the string
quark system , which is a difficult problem , but rather to find the shape of a string given a prescribed motion of its end .
knowing the shape of the string , and using eq .
( [ quarkeom ] ) , we are in principle able to find the motion of the string quark system . in practice , we use the expressions for the total energy and angular momentum and quark momentum in terms of the quark velocity to find the spectrum for the quantized string quark system numerically .
we demonstrate in appendix [ app : cons ] that eq .
( [ quarkeom ] ) is equivalent to the conservation of energy , angular momentum and the quark mass - shell relation .
planar motion is best described in terms of complex coordinates instead of vector notation . in complex coordinates , the string position @xmath18 can be written as @xmath19 with @xmath20 the metric @xmath2 , from eq .
( [ embmetric ] ) , is @xmath21 where we adopt the notation @xmath22 . for simplicity
, we consider here the fixed end at @xmath23 and the moving end at @xmath24 .
variation of the action with respect to the position of the string yields the equations of motion @xmath25 once @xmath6 is known , the string four - momentum and angular momentum are @xmath26 } = -\ , 2 a \int_0 ^ 1 d\sigma \,\sqrt{-h}\,\ , h^{\tau a}\ , { \rm im}\,(x^ * x_{,a})\ . \label{j}\end{aligned}\ ] ] the equations ( [ eom ] ) have an exact solution in which a straight string rotates uniformly ( constant angular velocity ) , but has an arbitrarily changing length .
this is the solution that we will perturb to find the string shape when its end undergoes angular acceleration .
our ansatz is @xmath27 from eq .
( [ metric ] ) , we find the metric tensor @xmath28 where we define @xmath29 substituting the above into the string equations ( [ eom ] ) , we find that they are exactly satisfied .
this result is the realization of catalan s theorem mentioned earlier , where the straight line to the quark sweeps out a helicoid of fixed pitch but arbitrarily varying radius .
again we emphasize that this is a solution of the string equations ( [ eom ] ) only . in a realistic meson , for which eq .
( [ quarkeom ] ) also holds for the quarks on the ends , we must perturb eq .
( [ ss ] ) to find a solution that takes into account the necessary angular acceleration of the ends . for future reference
, we give here expressions for @xmath30 and @xmath31 of the straight string . from eq .
( [ j ] ) we find that the angular momentum is given as @xmath32 \nonumber \\ & = & a\omega r^3 \int_0 ^ 1 { \sigma^2 \ , d\sigma \over\sqrt{1-\sigma^2v_\perp^2 } } = { a r^2\over 2 v_\perp}\left[{\arcsin(v_\perp)\over v_\perp } - \sqrt{1 - v_\perp^2}\right]\ . \label{jstraight}\end{aligned}\ ] ] the corresponding energy and spatial momentum from eq .
( [ pmu ] ) are @xmath33 where @xmath34\ . \label{pperp}\end{aligned}\ ] ] we observe that @xmath35 is purely transverse in direction . it is also interesting to note the small velocity limits for these quantities .
if we define the moment of inertia of a uniform string of `` mass '' @xmath36 rotating about one end as @xmath37 then the low velocity limits of @xmath30 , @xmath38 , and @xmath39 are @xmath40 as one might expect , the string can be thought of as a rod of mass @xmath36 .
in the above we have demonstrated that a uniformly rotating string can remain straight even if the length changes .
when the end has angular acceleration the string must curve . to gain intuition we consider a quasi - newtonian string of `` mass '' density @xmath0 which rotates with instantaneous angular velocity @xmath41 about one end . in the rotating frame
an element at distance @xmath42 is assumed to be in equilibrium under two forces , the tension force @xmath43 and the angular acceleration fictitious force @xmath44 as shown in fig .
[ fig : one ] .
the only other possible transverse force is the coriolis force due to the radial motion of the right end .
however , the element does not experience a coriolis force since motion of the end only creates more string and the notion of longitudinal velocity has no meaning .
this type of motion can be thought of as `` adiabatic '' since the resulting shape depends only on the end acceleration .
the force equilibrium condition yields @xmath45 in terms of dimensionless variables , @xmath46 and @xmath47 , the above equation is @xmath48 using the end condition @xmath49 , we find the non - relativistic string shape @xmath50 which is illustrated in fig . [
fig : one ] .
in order to find the relativistic string shape , in this section we consider the nature of small string deflections .
we again consider adiabatic solutions generalizing the non - relativistic concept of the previous section to the nambu - goto case .
we straightforwardly perturb about the straight string solution ( [ ss ] ) to take @xmath51}\ , \label{perturbx}\end{aligned}\ ] ] where the function @xmath52 is assumed complex .
we also assume that @xmath52 , @xmath53 , @xmath54 , and @xmath55 are small , and therefore we drop terms like @xmath56 , @xmath57 , @xmath58 , etc . using eq .
( [ metric ] ) , we find in this approximation @xmath59 the assumed string position ( [ perturbx ] ) and the metric @xmath2 above must satisfy the equations of motion ( [ eom ] ) for the time and spatial components @xmath60 upon substitution , and with considerable but straightforward algebra , we find that each equation is satisfied when @xmath52 satisfies @xmath61 there are no constraints on the real part of @xmath62 , which is a consequence of the reparametrization invariance of the nambu - goto - polyakov action ( [ action ] ) . in terms of a dimensionless quantity @xmath63
we find that the displacement from the straight string satisfies @xmath64 for small rotational velocities @xmath65 , the string shape equation ( [ stringshape ] ) reduces to @xmath66 which is identical to the non - relativistic result , eq .
( [ nrstringshape ] ) .
the nambu - goto shape must then reduce to the previous result , eq .
( [ cubic ] ) . with the choice of independent variable @xmath67 the shape equation ( [ stringshape ] )
reduces to the simpler form @xmath68 whose exact analytic solution , @xmath69\arcsin\xi + c_1 \xi + c_2\left(\sqrt{1-\xi^2 } + \xi\ , \arcsin\xi\right)\ , \ ] ] is discussed in appendix [ app : diffeq ] .
the constants @xmath70 and @xmath71 are fixed by the end conditions that @xmath72 vanish at @xmath73 and @xmath74 .
the final result is @xmath75\ . \label{shapefcn}\end{aligned}\ ] ] by comparison of the above expression with ( [ cubic ] ) we see that the only difference from the non - relativistic result is in the shape function . from the power series expansion it is straightforward to verify that @xmath76 and , hence , @xmath72 reduces to the non - relativistic limit . in fig .
[ fig : two ] we show @xmath77 for non - relativistic , intermediate , and fully relativistic speeds .
one can observe that even for a very rapid rotation , @xmath78 , the string shape does not change dramatically with respect to the non - relativistic result .
as we have seen from eq .
( [ fpert ] ) , the actual size of the displacement from the straight string is controlled by a factor of @xmath79 .
its magnitude for meson states can be estimated using the heavy - light version of the relativistic ( straight ) flux tube ( rft ) model @xcite . since the numerical solution of the model provides us with a matrix representation for the @xmath80 operator in a particular basis , it is convenient to rewrite @xmath81 as @xmath82 the above classical expression can be quantized by the appropriate symmetrization procedure , and by promoting @xmath80 and @xmath83 to quantum - mechanical operators , for which one can use @xmath84 $ ] . in this way , once the model is solved and matrix representations for @xmath80 and the hamiltonian @xmath85 are found , it is straightforward to compute the expectation value of @xmath81 for a given quantum state .
figures [ fig : mdep ] and [ fig : ldep ] show the dependence of the quantity @xmath86 on the light quark mass and angular momentum , respectively .
these results indicate that the actual size of the displacement from the straight string in real mesons is smaller than @xmath87 , which illustrates the validity of the straight string approximation .
note that numerical estimates shown in those figures were obtained using only the confining part of the heavy - light rft model with string tension @xmath88 .
addition of the short - range one gluon exchange interaction would further reduce our results .
we now compute the effects of the deformation away from straightness on string dynamics . to this end
we compare the angular momentum , energy , and linear momentum of the actual string with that of the straight string .
the similarities and differences are quite interesting .
we proceed in each case by substituting the perturbed string form ( [ perturbx ] ) and the consequent perturbed metric into the desired dynamical quantities . the expression ( [ j ] ) for the angular momentum to first order in small quantities becomes @xmath89 where we refer to eq .
( [ jstraight ] ) for the straight string result . to compare with the straight string angular momentum we must have strings of the same length and with the same end velocity which requires that both @xmath90 and @xmath91 vanish .
we then conclude that for small deviations from straightness the angular momentum of the curved string is the same as that of the straight string , @xmath92 similarly , when we evaluate eq .
( [ pmu ] ) with @xmath93 we obtain @xmath94 again requiring the accelerating string to have the same length and angular velocity as the straight string , @xmath90 = @xmath95 = @xmath96 , we have @xmath97 finally we compute the linear momentum , @xmath98 of the curved string using the spatial part of eq .
( [ pmu ] ) .
the transverse momentum component is @xmath99 with the usual end conditions , we obtain @xmath100 where @xmath101 is given in eq .
( [ pperp ] ) .
the curved string also has radial momentum @xmath102 which , after integration by parts , becomes @xmath103 it is worth noting that the expression for radial momentum ( [ radp ] ) ( and also @xmath39 ) can be directly read off of fig .
[ fig : one ] . referring back to our explicit solution ( [ shapefcn ] ) for @xmath72
, we find the analytic solution @xmath104 to be @xmath105
\arcsin v_\perp\ .\ ] ] for small orbital velocities we have @xmath106 to understand the size of the radial string momentum we will compare it to other `` relativistic corrections '' that arise in meson dynamics . in a meson with a quark mass @xmath107 large enough
that the quark velocity is small , the quark s angular momentum dominates that of the string . in this case @xmath108 @xmath109
constant and we have @xmath110 . the radial momentum is primarily due to the non - relativistic quark with corrections from relativity and the string ; @xmath111\ .\ ] ] the string radial momentum is smaller by a factor of @xmath112 than the first relativistic correction . in this section we have observed that to leading order the angular momentum , energy , and transverse momentum of the curved string are unchanged by small deviations from a straight string .
the bending of the string will induce a radial momentum but it is of higher order than the leading relativistic corrections . in this way , we agree with the work of brambilla et al .
@xcite who showed that in the wilson loop formalism relativistic corrections are correctly computed assuming a straight path between the quarks .
we have considered here the shape of a qcd string with one end fixed and the other moving with arbitrary velocity , but with a small angular acceleration .
we calculated the deviation from a straight string and found it to reduce to a physically reasonable non - relativistic limit .
the solutions we have obtained are adiabatic in the sense that they depend only on the end condition , and would be static in the end rest frame . to find the shape of the curved string ,
we perturbed an exact straight solution to the nambu - goto - polyakov string equations and solved exactly the resulting equations of motion to leading order in the perturbation .
as long as the angular acceleration is sufficiently small this should provide an accurate picture of the string shape .
we numerically solved a straight string model to estimate the angular acceleration that occurs in actual mesons .
the result was that the string deviation from a straight line is never very large , justifying the use of our perturbative approach , and showing the validity of the straight string approximation .
in addition we have computed the angular momentum , energy , and linear momentum of the curved string . in each case but one , the perturbation drops out and the perturbed string behaves identically to the straight one . only for the radial momentum
does the deviation from straightness have an effect . in the semi - relativistic approximation
this radial momentum is smaller by a factor of @xmath113 than the first relativistic correction .
the straight string approximation is then justified for heavy quark mesons as previously pointed out from a different point of view @xcite . for relativistic mesons
the string radial momentum may have small , but perhaps interesting and calculable consequences .
this work was supported in part by the us department of energy under contract no .
de - fg02 - 95er40896 .
fermilab is operated by ura under doe contract de - ac02 - 76ch03000 .
here we consider the energy , momentum and angular momentum of a string with a quark at each end .
the relevant actions are the nambu - goto - polyakov action , eq .
( [ action ] ) , and the quark action @xmath114 in which @xmath115 is the position of the i@xmath116 quark of mass @xmath117 , and the auxiliary field @xmath118 is a one - dimensional metric density analogous to the string s auxiliary field metric @xmath2 .
the boundary condition relating the two actions ( [ action ] ) and ( [ quarkaction ] ) is that the quarks sit at the ends of the string , @xmath119 variation of the sum of the actions ( [ action ] ) plus ( [ quarkaction ] ) with respect to @xmath18 yields the equations of motion for the quarks @xmath120 where @xmath121 is the momentum of the quark at @xmath122 , related to the quark velocity by @xmath123 and @xmath124 is the momentum current on the string worldsheet , which is related to the canonical string momentum @xmath125 by @xmath126 .
it is easy to show that the total momentum of the string quark system , @xmath127 is conserved under @xmath3 evolution , @xmath128 the equations of motion @xmath129 are used to go from the first to the second line , and the final result vanishes by the quark equations of motion ( [ quarkeoma ] ) .
the conservation of the angular momentum , @xmath130\ , \tau } + \sum_{i=1,2 } x_{i}^{[\mu}p^{\nu]}_{i}\ , \ ] ] is only slightly more complicated to demonstrate .
the @xmath3 derivative of @xmath131 is @xmath132\ , \tau } + x^{[\mu } \dot{{\cal p}}^{\,\nu]\ , \tau}\right ) + \sum_{i=1,2 } \left(\dot{x}_{i}^{[\mu}p^{\nu]}_{i } + x_{i}^{[\mu}\dot{p}^{\nu]}_{i}\right ) \nonumber \\ & = & \int_{\sigma_1}^{\sigma_2}d\sigma\ , \left(\dot{x}^{[\mu}{\cal p}^{\,\nu]\ , \tau } - x^{[\mu}{{\cal p}}^{\,\nu]\ , \sigma\,\prime}\right ) + \sum_{i=1,2 } x_{i}^{[\mu}\dot{p}^{\nu]}_{i}\nonumber \\ & = & \int_{\sigma_1}^{\sigma_2}d\sigma\ , \left(-{x}^{[\mu\,\prime } { \cal p}^{\,\nu]\ , \sigma } - x^{[\mu}{{\cal p}}^{\,\nu]\ , \sigma\,\prime}\right ) + \sum_{i=1,2 } x_{i}^{[\mu } \dot{p}^{\nu]}_{i}\nonumber \\ & = & -{x}^{[\mu}{\cal p}^{\,\nu]\ , \sigma } \big|_{\sigma=\sigma_2 } + x^{[\mu } { \cal p}^{\,\nu]\ , \sigma } \big|_{\sigma=\sigma_1 } + \sum_{i=1,2 } x_{i}^{[\mu}\dot{p}^{\nu]}_{i}\ .\end{aligned}\ ] ] in going from the second to the third line , we have used @xmath133\ , \tau } = - a \sqrt{-h}\ , h^{\tau a}\ , \dot{x}^{[\mu } { x}^{\nu]}\mathstrut_{,a } = - a \sqrt{-h}\ , h^{\tau \sigma}\ , \dot{x}^{[\mu } { x}^{\nu]\,\prime } = -{x}^{[\mu\,\prime } { \cal p}^{\,\nu]\ , \sigma}$ ] . using the boundary conditions ( [ bc ] ) that place the quark at the end of the string , we find @xmath134}_{1 } + x_{1}^{[\mu}{\cal p}^{\nu]\ , \sigma}\big|_{\sigma=\sigma_1}\right ) + \left(x_{2}^{[\mu}\dot{p}^{\nu]}_{2 } - x_{2}^{[\mu}{\cal p}^{\nu]\ , \sigma}\big|_{\sigma=\sigma_2}\right ) = 0\ .\label{angmomcons } \end{aligned}\ ] ] finally , the quark mass - shell condition , @xmath135 which follows identically from the relation of the quark momenta to quark velocities , @xmath136 must be preserved by the equations of motion ( [ quarkeom ] ) .
we find that @xmath137 the three conditions ( [ momcons ] ) , ( [ angmomcons ] ) , and ( [ shellcons ] ) together imply that the conservation of energy momentum , angular momentum , and the quark mass - shell relation are equivalent to the quark equations of motion , eqs .
( [ quarkeom ] ) and ( [ quarkeoma ] ) .
we consider here the details of the solution of ( [ shapediffeq ] ) @xmath138 to obtain the central result ( [ shapefcn ] ) of the paper .
the solution method follows standard procedures .
first we define a new function , @xmath139 , so that @xmath140 giving @xmath141 now define @xmath142 to obtain @xmath143 = \beta\,\xi\ .
\label{eqn : a5}\ ] ] the integrating factor @xmath144 is chosen to make the coefficient of @xmath145 vanish and by quadrature , @xmath146 substituting into ( [ eqn : a5 ] ) , we find @xmath147 and , by ( [ eqn : a4 ] ) , @xmath148 finally by quadrature we have @xmath149 and @xmath150 is @xmath151 99 y. nambu , _ `` quark model and the factorization of the veneziano amplitude , '' _ in _ symmetries and quark models , _
chand , ( gordon and breach , 1970 ) ; t. goto , _ prog . theor . phys . _
* 46 * , 1560 ( 1971 ) ; l. susskind , _
nuovo cimento _ * 69a * , 457 ( 1970 ) ; a.m. polyakov , _ phys .
lett . _ * 103b * , 207 ( 1981 ) . | it is well known that a straight nambu - goto string is an exact solution of the equations of motion when its end moves in a circular orbit . in this paper
we investigate the shape of a confining relativistic string for a general motion of its end .
we determine analytically the shape of the curved string to leading order in deviation from straightness , and show that it reduces to an expected non - relativistic result .
we also demonstrate numerically that in realistic meson models this deviation is always small .
we further find that the angular momentum and energy are the same as for the straight string , but that the curved string has a small radial momentum not present in a straight string .
our results justify the common assumption of straight strings usually made in hadron models .
epsf 6.5 in 8.5 in 0.0 in -20 pt |
this note considers a stochastic growth model in the _ kpz anisotropic class _ in .
this model was introduced in @xcite and studied in depth for a specific initial condition , the case considered here .
this model describes the evolution of particles subject to an _ interlacing property_. the model can also be thought of as a two - dimensional _ stochastically growing interface _ , or as a _ random lozenge tiling model_. another tiling model which shares similar features of the dynamical perspective is _ domino tilings of the aztec diamond _ using _ the shuffling algorithm _ ; see @xcite . for
the model considered in this paper , the evolution of the interface , under hydrodynamic scaling , grows deterministically according to a pde . at the microscopic level in the bulk ,
the specified boundary conditions of the system are _ forgotten _ : in the bulk of the system one sees an invariant measure which depends only on the normal direction of the macroscopic surface @xcite . from the lozenge tiling perspective ,
these limiting measures are determinantal and they are parameterized by the relative proportions of lozenges .
these measures match up with the dimer model on the infinite honeycomb graph which are the unique translation invariant stationary measures for any given normal direction @xcite .
stochastic growth models have been studied in many different guises .
many of these studies have focussed around the 1 + 1 kpz universality class ; see for example @xcite for surveys .
the @xmath0 anisotropic case has not been as extensively studied @xcite but there are , however , numerous results in connection with random tiling models @xcite . one feature of these models which is of particular interest for this note is the _
limit shape _
this is the average profile which the stochastic interface fluctuates around .
in particular , we focus on giving an elementary approach to computing the _ speed of growth _ , which is the growth rate of the stochastic growing interface under the prescribed dynamics .
this is an important quantity since it determines the limit shape of the system . in the work by borodin and ferrari @xcite ,
the speed of growth was obtained by computing the _ infinitesimal current _ and then taking the bulk scaling limit .
the computation was relatively straightforward , but requires the knowledge of the correlations of particles or lozenges at two different times .
recently toninelli in @xcite , considered the same model and a generalized version of the dynamics for the infinite honeycomb graph using a _ bead _ perspective ; beads and the _ bead model _ were introduced in @xcite .
he shows that the model is well - defined for the stationary measure .
this requires some effort since the dynamics allow , a - priori , infinitely long - range interactions , but under the stationary measure these interaction probabilities decay exponentially with the distance .
the speed of growth was not determined in @xcite . by appealing to the underlying combinatorics of the model and the kasteleyn approach for dimer models ( e.g. see @xcite ) , we are able to determine the speed of growth in the infinite honeycomb case , thus establishing the conjecture in ( * ? ? ?
* eq . ( 3.6 ) ) ; see theorem [ thminfinitesystem ] . from @xcite ,
the speed of growth @xmath1 is given by a specific ( unsigned ) off - diagonal entry of the inverse of the kasteleyn matrix , where the kasteleyn matrix is a type of ( possibly signed ) adjacency matrix @xcite .
the approach used is to first find a recursive formula for this particular entry for the honeycomb graph embedded on a torus which leads to a series , and then take the toroidal exhaustion limit .
the resulting limiting series matches with the speed of growth defined through the dynamics under the stationary measure . to prove theorem [ thminfinitesystem ]
, we found a relation for the specific off - diagonal entry of the inverse of the kasteleyn matrix for any _ tileable _ finite honeycomb graph with arbitrary edge weights in terms of single times . as a consequence
, we use this relation to motivate an algebraic proof that the expression for the speed of growth of the model in @xcite and the speed of growth computed from the prescribed dynamics are the same ( the latter is a series with entries given by determintants of increasing size ) . in principle
it seems to be feasible to obtain theorem [ thminfinitesystem ] from theorem [ thmfinitesystem ] by a precise asymptotic analysis and a careful manipulation of sums , but we did not pursue this since we are primarily interested in understanding combinatorial structures behind the identity .
* acknowledgments . *
the authors are grateful for discussions with f. toninelli about his work and to both icerm and the galileo galileo institute which provided the platform to make such discussions possible .
the work is supported by the german research foundation via the sfb 1060b04 project .
we first describe the interacting particle system model introduced in @xcite .
it is a model in the @xmath2-dimensional kpz anisotropic class .
it is a continuous time markov chain on the state space of interlacing variables @xmath3 @xmath4 is interpreted as the position of the particle with label @xmath5 , but we will also refer to a given particle as @xmath4 .
we consider fully - packed initial conditions , namely at time moment @xmath6 we have @xmath7 for all @xmath8 ; see figure [ figureinitialconditions ] .
[ b]@xmath9 [ b]@xmath10 [ b]@xmath11 [ cl]@xmath12 [ cl]@xmath13 [ cl]@xmath14 [ cl]@xmath15 [ cl]@xmath16 [ cl]@xmath17 [ cl]@xmath18 [ cl]@xmath19 [ cl]@xmath20 [ cl]@xmath21 [ cl]@xmath22 [ cl]@xmath23 [ cl]@xmath24 [ cl]@xmath25 [ cl]@xmath26 the particles evolve according to the following dynamics .
each particle @xmath4 has an independent exponential clock of rate one , and when the @xmath4-clock rings the particle attempts to jump to the right by one .
if at that moment then the jump is blocked . otherwise , we find the largest @xmath27 such that , and all @xmath28 particles in this string jump to the right by one .
we illustrate the dynamics using figure [ figureintro ] , which shows a possible configuration of particles obtained from the fully - packed initial condition . in this state of the system , if the @xmath14-clock rings , then the particle @xmath14 does not move , because it is blocked by particle @xmath13 .
if the @xmath17-clock rings , then the particle @xmath17 moves to the right by one unit , but to respect the interlacing property , the particles @xmath21 and @xmath24 also move by one unit to the right at the same time . this aspect of the dynamics is called _ pushing_. [ b]@xmath9 [ b]@xmath10 [ b]@xmath11 [ cl]@xmath12 [ cl]@xmath13 [ cl]@xmath14 [ cl]@xmath15 [ cl]@xmath16 [ cl]@xmath17 [ cl]@xmath18 [ cl]@xmath19 [ cl]@xmath20 [ cl]@xmath21 [ cl]@xmath22 [ cl]@xmath23 [ cl]@xmath24 [ cl]@xmath25 [ cl]@xmath26 [ remlozenge ] the positions of the particles uniquely determine a lozenge tiling in the region bordered by the thick line in figure [ figureintro ] . [ remmeasureunif ] as shown in @xcite , the measure at time @xmath29 generated by the dynamics starting from the fully - packed initial condition has the property that , conditioned on the measure of the particles at level @xmath30 , the other particles are uniformly distributed on @xmath31 with fixed level @xmath30 configuration .
further , it is proved in theorem 1.1 of @xcite that the correlation function of the particles are determinantal on a subset of space - time @xcite ( see @xcite for information on determinantal point processes ) .
denote by @xmath32 the random variable that is equal to @xmath33 if there is a particle at @xmath34 at time @xmath29 and @xmath35 otherwise .
[ thmdetstructure ] for any given , consider , and @xmath36 .
then @xmath37=\det{[{\cal k}(x_i , n_i , t_i;x_j , n_j , t_j)]}_{i , j=1}^m,\ ] ] where @xmath38 }
\\ + \frac{1}{(2\pi i)^2}\oint_{\gamma_0}{\mathrm{d}}w\oint_{\gamma_1}{\mathrm{d}}z \frac{e^{t_1/w}}{e^{t_2/z } } \frac{(1-w)^{n_1}}{(1-z)^{n_2 } } \frac{w^{x_1}}{z^{x_2 + 1}}\,\frac{1}{w - z},\end{gathered}\ ] ] the contours @xmath39 , @xmath40 are simple positively oriented closed paths that include the poles @xmath35 and @xmath33 , respectively , and no other poles ( hence , they are disjoint ) . here
we used the notation @xmath41 given a configuration of particles , we define the height @xmath42 as @xmath43 in particular , the growth rate of the height at a position @xmath34 is given by the ( infinitesimal ) particle current at @xmath34 , denoted by @xmath44 with @xmath45.\ ] ] this quantity was computed in the proof of lemma 5.3 of @xcite with the result @xmath46 where here we use the notation @xmath47 . on the other hand , at any time @xmath29 , the height function @xmath11 at @xmath34 increases by one whenever there is a particle at position @xmath34 which jumps to . by definition of the dynamics , at rate 1 a particle at @xmath34 could jump to the right provided that there is no particle at @xmath48 . however , a particle at @xmath34 can also be pushed if there is a column of particles directly below @xmath34 .
more explicitly , for @xmath49 a particle at @xmath34 is pushed to the right by the move of a particle at @xmath50 .
this happens at rate @xmath33 provided that @xmath51 is empty and @xmath52 for @xmath53 are occupied . in the case
@xmath54 , the constraint that @xmath55 is empty does not exist as there are no particles at level @xmath35 .
therefore the growth rate obtained from the dynamics , denoted by @xmath56 , is given by @xmath57+{\mathbbm{e}}\left[\prod_{k=0}^{n-1 } \eta(x , n - k , t)\right].\ ] ] using theorem [ thmdetstructure ] and the complementation principle for determinantal point processes ( see appendix of @xcite ) , the expected value in the right side of ( [ eqspeed ] ) is given by @xmath58_{i , j=0}^{\ell-1 } & \left[-{\cal k}_t(x , n - i;x+1,n-\ell)\right]_{i=0}^{\ell-1}\\ \left[{\cal k}_t(x+1,n-\ell;x , n - j)\right]_{j=0}^{\ell-1 } & 1-{\cal k}_t(x+1,n-\ell;x+1,n-\ell ) \end{array } \right].\ ] ] in this paper we show directly that indeed @xmath1 and @xmath59 are the same . [ thmfinitesystem ] it holds @xmath60 this finite - time result can be easily generalized to the case of level - dependent jump rates .
this system is still determinantal when starting with fully - packed initial conditions , with correlation kernel obtained in proposition 3.1 of @xcite ; see corollary 2.26 of @xcite too .
the height function , defined in , gives a three dimensional surface with cartesian coordinate axis . in a projection in the @xmath61-direction ,
each unit square of the surface becomes a lozenge , while in the projection of figure [ figureintro ] there are three types of parallelograms , still referred to as lozenges ; see figure [ figuretilings ] .
[ c]i - @xmath62 [ c]ii - @xmath63 [ c]iii - @xmath28 facets types and their associated lozenges and angles .
the gray circle is at position @xmath34 .
for dimers , their coordinate is given by the black site .
more precisely , in this case we say that dimer of type i is at @xmath34 , of type ii is at @xmath48 , while of type iii at @xmath64.,title="fig:",height=226 ] another representation useful for the proofs of the theorems in this paper , is through _ perfect matchings _ or _ dimers _ via the dual graph associated to the underlying graph of the lozenge tilings . for a lozenge tiling on a finite graph ,
the dual graph is a subgraph of the _ honeycomb _ ( or _ hexagonal _ ) graph , with each lozenge representing an edge , called a dimer , on this dual graph ; see figure [ figuretilings ] .
each lozenge tiling represents a _ dimer covering _ of the dual graph ; see figure [ figuredimers ] for an example .
[ r]@xmath34 [ l]@xmath65 [ r]@xmath48 [ l]@xmath66 for the purpose of this paper , we denote @xmath67 to be the infinite ( bipartite ) honeycomb graph ; see the left side of figure [ figcoordinates ] for a finite snapshot .
we set @xmath68 , that is , the restriction of length - size @xmath69 of @xmath67 with periodic boundary conditions .
we use the standard terminology for the dimer model ; see for example section 1 in @xcite for details .
now we consider the case where the state space are particles on @xmath70 satisfying interlacing between levels @xmath10 and @xmath71 , for any @xmath72 , i.e. , @xmath73 formally one would like to consider the dynamics on @xmath74 as described for the model in section [ sectfinitet ] .
however , the dynamics are not well - defined for all elements in @xmath75 due to the possibility of pushing from an infinitely long stack of particles .
it is shown in @xcite , that the dynamics are almost surely well - defined starting from the stationary measure because under this measure , the probability of having a stack of particles of length @xmath76 decays exponentially in @xmath76 .
the translation invariant stationary measures form a two - parameter family , uniquely determined by the `` two - dimensional slope '' of the height function . in terms of dimers
, we give to each type of dimer a weight , say @xmath77 as indicated in figure [ figuretilings ] .
the probability of a dimer configuration is then proportional to the product of the weights of each dimer configuration , thus we effectively have only two free parameters .
one nice property of stationary measures on @xmath70 is that their correlation functions are determinantal with a correlation kernel given only in terms of the slope ( which is determined by @xmath77 ) .
correlations functions , at different times for the model on @xmath70 with stationary initial conditions , are not explicitly known ( and for the asymmetric version studied by toninelli @xcite they are unlikely to be determinantal ) and therefore one can not use ( [ eq2.6 ] ) to determine the speed of growth . however , the speed of growth is defined by the dynamics via ( [ eqspeed ] ) with @xmath10 replaced by @xmath78 .
sheffield showed that indeed the translation invariant measures are uniquely determined by the slope of the height function .
[ thmsheffield ] for each @xmath79 with @xmath80 with @xmath81 there is a unique translation - invariant ergodic gibbs measure @xmath82 on the set of dimer coverings of @xmath83 , for which the height function has average normal @xmath84 .
this measure can be obtained as the limit as @xmath85 of the uniform measure on the set of those dimer coverings of @xmath86 , whose proportion of dimers in the three orientations is @xmath87 , up to errors tending to zero as @xmath85 .
moreover every ergodic gibbs measure on @xmath83 is of the above type for some @xmath84 .
the correlation functions of the stationary measure @xmath82 are determinantal and they are given in terms of the so - called inverse kasteleyn matrix , denoted by @xmath88 ( see more details in section [ sect3.2 ] and the lecture notes @xcite for a complete treatment of the subject ) .
it is given as follows ( it is obtained from eq .
( 4 ) in @xcite with appropriate change of coordinates ) @xmath89 with @xmath90 as shown by kenyon , okounkov and sheffield in @xcite , the latter is the limiting inverse kasteleyn matrix obtained in the toroidal exhaustion limit where the edge weights @xmath91 and @xmath28 are depicted in figure [ figuretilings ] .
kenyon introduced a very useful mapping from @xmath92 to the upper - half complex plane @xmath93 illustrated in figure [ figgeometry ] .
[ l]@xmath94 [ r]@xmath95 [ c]@xmath96 [ c]@xmath35 [ c]@xmath33 associated with @xmath92 is given by constructing the triangle with basis @xmath97 and angles @xmath98 , @xmath99 as indicated .
the length of the segment @xmath100 is @xmath95 and the length of the segment of @xmath101 is @xmath94.,title="fig:",height=151 ] by change of variables @xmath102 and @xmath103 in eq .
one obtains @xmath104 a simple computation ( see appendix ) leads to the following representation of the kernel @xmath105 where for @xmath106 the integration contour crosses @xmath107 , while for @xmath108 the contour crosses @xmath109 .
this is consistent with theorem 5.1 and proposition 3.2 of @xcite , where the correlation kernel was obtained by taking the finite system considered in theorem [ thmfinitesystem ] and followed by taking the large time / space limit in such a way that the local normal direction of the surface is given by @xmath92 : this introduced a conjugation factor and a shift in the @xmath9 by @xmath10 .
finally , the point in the complex plane @xmath110 in @xcite equals @xmath111 here . ] .
[ thminfinitesystem ] consider the particle model on @xmath70 distributed according to @xmath82 , that is , with determinantal correlation functions given by the correlation kernel ( [ eqapp1 ] ) .
then the speed of growth is given by @xmath112 the speed of growth of the model defined in @xcite is then given by , where @xmath113 are the two parameters in the model of @xcite .
the totally asymmetric dynamics is the one given by @xmath114 and @xmath115 .
the algebraic proof presented here is strongly inspired by the combinatorial proof for finite graphs obtained beforehand that will be used to prove theorem [ thminfinitesystem ] . in this proof
we will suppress all the @xmath29 indices and have reasonably sized matrices , we use the notation @xmath116 we use the coordinate system illustrated in figure [ figcoordinates ] .
we derive a series expansion of @xmath117 by expanding it step - by - step .
the idea behind the expansion is that @xmath117 is ( intuitively ) suggestive of an `` edge '' between @xmath118 and @xmath119 .
if this `` edge '' is covered by a dimer , then either there is a dimer covering @xmath120 or not . in the latter case
, then there are dimers covering @xmath121 and @xmath122 .
this can be repeated for @xmath123 and for @xmath124 .
the latter case will not occur and therefore the series naturally ends .
this idea is exploited in depth for the finite graph lozenge tiling ; see proposition [ proprecursiongraph ] .
we start with two algebraic identities satisfied by the kernel ( [ startingkernel ] ) .
[ lemidentity ] it holds @xmath125 it is quite trivial .
one uses linearity of the integrals .
[ lemidentityb ] it holds @xmath126 it is quite trivial .
one uses linearity of the integrals .
we start with a proposition that will be used recursively .
[ proprecursionfinitet ] consider any set of @xmath127 @xmath128 black / white couples of ( disjoint ) vertices that do not include black vertices at @xmath129 and white at @xmath130
. then @xmath131=\det\left[\begin{array}{lll } { \cal k}_{x_i , m_i;x_j',m_j ' } & { \cal k}_{x_i , m_i;x , m } & { \cal k}_{x_i , m_i;x+1,m}\\ { \cal k}_{x , m;x_j',m_j ' } & { \cal k}_{x , m;x , m } & { \cal k}_{x , m;x+1,m } \\ { \cal k}_{x , m-1;x_j',m_j ' } & { \cal k}_{x , m-1;x , m } & { \cal k}_{x , m-1;x+1,m } \end{array}\right]\\ & + \det\left[\begin{array}{llll } { \cal k}_{x_i , m_i;x_j',m_j ' } & { \cal k}_{x_i , m_i;x , m } & { \cal k}_{x_i , m_i;x+1,m-1 } & { \cal k}_{x_i , m_i;x+1,m}\\ { \cal k}_{x , m;x_j',m_j ' } & { \cal k}_{x , m;x , m } & { \cal k}_{x , m;x+1,m-1 } & { \cal k}_{x , m;x+1,m}\\ { \cal k}_{x , m-1;x_j',m_j ' } & { \cal k}_{x , m-1;x , m } & { \cal k}_{x , m-1;x+1,m-1 } & { \cal k}_{x , m-1;x+1,m}\\ { \cal k}_{x+1,m-1;x_j',m_j ' } & { \cal k}_{x+1,m-1;x , m } & { \cal k}_{x+1,m-1;x+1,m-1 } & { \cal k}_{x+1,m-1;x+1,m } \end{array}\right ] , \end{aligned}\ ] ] where , whenever there is a @xmath132 or @xmath59 index , this means a block - matrix with @xmath132 and/or @xmath59 from @xmath33 to @xmath127 .
schematically this is represented in figure [ figdimerrec ] , where ( [ eq3.3 ] ) equals ( a)=(c1)+(c2 ) .
[ c]@xmath133 [ c]@xmath134 [ c]@xmath135 [ c](a ) [ c](b1 ) [ c](b2 ) [ c](c1 ) [ c](c2 ) .
the link in ( a ) is between @xmath136.,title="fig:",height=113 ] the left side of ( [ eq3.3 ] ) is represented in figure [ figdimerrec](a ) .
the scheme of figure [ figdimerrec](b1 ) is given by @xmath137.\ ] ] the scheme of figure [ figdimerrec](b2 ) is given by @xmath138,\ ] ] we subtract from the last column the previous two . using the identity in lemma [ lemidentity ]
we then obtain @xmath139\\ & = \det\left[\begin{array}{lll } { \cal k}_{x_i , m_i;x_j',m_j ' } & { \cal k}_{x_i , m_i;x+1,m } & { \cal k}_{x_i , m_i;x , m}\\ { \cal k}_{x , m;x_j',m_j ' } & { \cal k}_{x , m;x+1,m } & { \cal k}_{x , m;x , m } \\ { \cal k}_{x , m-1;x_j',m_j ' } & { \cal k}_{x , m-1;x+1,m } & { \cal k}_{x , m-1;x , m } \end{array}\right ] . \end{aligned}\ ] ] the determinants in ( [ eq3.5 ] ) and ( [ eq3.8 ] ) differs only by the last row . thus summing them up and using the identity in lemma [ lemidentityb ] one immediately sees that the last row and the second - last row are identical except for an extra @xmath140 term in the last matrix entry .
therefore we have recovered left side of ( [ eq3.3 ] ) .
finally one has to verify the equality between the schemes of figure [ figdimerrec](b1)/(b2 ) and figure [ figdimerrec](c1)/(c2 ) .
this is trivial since it corresponds to permuting the position of one column and take care of the signature of the permutation .
now we are ready to finish the proof of theorem [ thmfinitesystem ] . the schematic representation of the proof is in figure [ figprooffinitet ] .
[ c]@xmath133 [ c]@xmath134 [ cb]@xmath141 .,title="fig:",height=113 ] recall that @xmath142 is the random variable of a particle located at @xmath34 .
similarly , denote by @xmath143 the random variable of having a white lozenge ( type iii in figure [ figuretilings ] ) at @xmath34 , where the position is given by the one of the black triangle .
using proposition [ proprecursionfinitet ] repeatedly by starting with the case @xmath144 and @xmath145 and recalling the correspondence between dimers and lozenges ( see figure [ figuretilings ] ) we get @xmath146.\ ] ] [ l]forced [ l]only two choices ) and ( [ eqspeed]).,title="fig:",height=113 ] the conditions of theorem [ thmfinitesystem ] stipulate that the system is bounded from below by level @xmath33 and this is the reason why the above series is finite .
finally we have to see that ( [ eq3.11 ] ) and ( [ eqspeed ] ) match .
first consider @xmath147 .
as illustrated in figure [ figprooffinitetb ] , by observing that there are particles at @xmath148 , it implies that the square lozenges at occur with probability one and thus we remove them from the right side of ( [ eq3.11 ] ) .
further , if we have a particle at @xmath50 , then either there is a particle at @xmath51 or there is a white lozenge at @xmath51 since the other type of lozenge do not fit .
therefore we can replace in ( [ eq3.11 ] ) @xmath149 with @xmath150 .
finally , for @xmath54 , @xmath55 is forced to be white and therefore we can remove it from the product too .
this finishes the proof of theorem [ thmfinitesystem ] .
the strategy is first to obtain a recursion relation analogue of proposition [ proprecursionfinitet ] for a finite honeycomb graph with generic weights , then use this result to extend the recursion relation to @xmath86 with @xmath77 weights . using theorem [ thmsheffield ]
we take the toroidal exhaustion limit and have the same recursion relation for the infinite honeycomb graph @xmath151 .
the proof of theorem [ thminfinitesystem ] ends applying recursively the recursion relation .
let @xmath152 be a finite subgraph of the honeycomb graph which is tileable by dimers .
@xmath153 is bipartite and has the same number of white and black vertices .
the geometry of the honeycomb graph is as illustrated in figure [ figcoordinates ] so that the black vertices are in ( a subset of ) @xmath70 and the white vertices are in ( a subset of ) @xmath154 . to avoid using half - integer coordinates ,
we adopt a notation so that the white vertices are also on @xmath70 , more precisely , let @xmath155 , @xmath156 , and @xmath157 , then @xmath158 edges of the graphs are of the form @xmath159 to @xmath160 , @xmath161 , and .
we denote by @xmath162 and @xmath163 the set of the black and white vertices respectively with the above coordinates . assign @xmath164 to be the edge weights and denote the kasteleyn matrix , the matrix whose rows are indexed by all the white vertices and whose columns are indexed by black vertices , by @xmath165 for all @xmath166 and @xmath167 .
the above formulation defines a valid kasteleyn orientation of the graph , that is , the number of counterclockwise edges in any face is odd ; see figure [ figkast1 ] .
[ c](a ) [ c](b ) the kasteleyn matrix was originally introduced by kasteleyn @xcite to count the number of dimer covers of a graph as the latter is given by @xmath168 , but its use goes beyond the uniform weight case .
the probability @xmath169 on dimer configurations is defined as follows . for a given dimer configuration ,
we associate a weight to be the product of all the weights of the dimers present in that configuration .
the partition function @xmath170 is then the sum of all weighted dimer configurations of @xmath153 .
then , the probability of a dimer configuration is given by its weight divided by @xmath170 .
in particular , given any disjoint set of edges @xmath171 , the probability of seeing dimers on the edges @xmath171 is given by @xmath172 kenyon in @xcite showed the following .
[ thmkenyon97 ] consider a set of @xmath173 disjoints edges of @xmath153 , , @xmath174 .
then , @xmath175=\det\left[{\mathbb{k}}_g^{-1}(b_i , w_j)\right]_{i , j=1}^m \prod_{i=1}^m { \mathbb{k}}_g(w_i , b_i ) , \ ] ] where @xmath176 represents the inverse of @xmath177 . in other words
, the dimers form a determinantal point process with correlation kernel @xmath178 given by @xmath179 for @xmath180 and @xmath181 .
the kasteleyn matrix approach has been used with some success for computing combinatorial and asymptotics of random tiling models ; see @xcite for domino tiling models and @xcite for the honeycomb case .
the first result is the one that inspired proposition [ proprecursionfinitet ] . to state it , we introduce some notations . for @xmath182 set @xmath183 and @xmath184 .
[ proprecursiongraph ] assume that the set of vertices @xmath185 belong to the graph @xmath153 for @xmath186 and let the edges @xmath187 for @xmath188
. then @xmath189+r_g(n),\ ] ] with @xmath190 here @xmath170 denotes the partition function of @xmath153 and @xmath191 denotes the partition function of the graph obtained from removing @xmath192 from @xmath153 .
we add an _ auxiliary edge _
@xmath193 , which is an edge not present in the graph but helpful for computations ; a similar idea to that used in @xcite .
we assign a weight @xmath33 to the auxiliary edge @xmath193 . to preserve the kasteleyn orientation of the new graph , this edge is directed from @xmath159 to @xmath194 .
since @xmath159 and @xmath194 are on the same face , removing them from the graph preserves the kasteleyn orientation ( see figure [ figkast1 ] ) .
therefore , each matching of @xmath195 has the same sign .
cramer s rule gives @xmath196}{\det[{\mathbb{k}}_g]}= \frac{z_{g\setminus \sigma_0}}{z_g}.\ ] ] for @xmath197 , @xmath160 is either matched to @xmath198 or @xmath199 . if @xmath160 is matched to @xmath198 , then the edge @xmath200 is forced to be matched too .
notice that @xmath201 the edge weight of is equal to and the edge weights of @xmath202 and @xmath200 are equal to @xmath203 and @xmath204 respectively .
notice that the product of the latter two matrix entries is exactly @xmath205 .
therefore , @xmath206 now we proceed by induction .
consider the graph @xmath207 , which can be thought as the graph where the vertices in @xmath208 are matched as in figure [ figkast2](a ) .
[ r]@xmath34 [ r]@xmath209 [ c]@xmath210 [ c]@xmath211 [ c]@xmath212 [ c]@xmath213 .
( a ) for @xmath208 , ( b ) for @xmath214 , ( c ) for @xmath215 , and ( d ) for @xmath216.,title="fig:",height=188 ] the vertex @xmath217 is either matched to or to .
for the former , then the vertex is incident to only one vertex , which means that also is matched .
since @xmath218 then we have @xmath219 define @xmath220 the vertex set @xmath216 is equal to @xmath214 .
therefore ( [ eq3.21 ] ) reads @xmath221 we iterate ( [ eq3.22 ] ) and find that @xmath222 now , from , we have @xmath223}{c_2(m+1)},\ ] ] because @xmath224 is the product of the edge weights of the edges @xmath225 .
therefore , dividing ( [ eq3.23 ] ) by @xmath170 , we find the right side of ( [ eq3.15 ] ) , while the left side follows from . in order to prove theorem [ thminfinitesystem ] we start by considering a finite honeycomb graphs with @xmath226 weights ( see figure [ figuretilings ] ) .
as above , define @xmath227 to be the random variable of having a lozenge of type i at @xmath34 .
that is @xmath228 then proposition [ proprecursiongraph ] gives the following .
[ corfiniteabc ] for any @xmath226-weighted honeycomb graph @xmath153 ( satisfying the assumptions of proposition [ proprecursiongraph ] ) , @xmath229 + \widetilde r_g(n)\ ] ] with @xmath230 .
the @xmath226-weighting means that @xmath231 see figure [ figuretilings ] .
then @xmath232 as well as @xmath233 so that proposition [ proprecursiongraph ] gives immediately @xmath234 + r_g(n)\ ] ] with @xmath235 .
next notice that having a chain of particles @xmath236 which corresponds to the edges @xmath237 and no particle at @xmath238 corresponding to the edge @xmath239 , then the dimer configurations incident to the edge @xmath240 for @xmath241 are forced since each vertex becomes incident to one edge by increasing @xmath132 . hence the dimers for @xmath242 are forced .
this implies @xmath243={\mathbbm{p}}[e_0 ^ 0,\ldots , e_m^0,e_m^1]={\mathbbm{e}}\bigg[(1-\eta(x+1,n - m))\prod_{i=0}^m \eta(x , n - i)\bigg].\ ] ] this ends the proof of the corollary .
we focus on the @xmath226-weighting of the honeycomb graph . recall that @xmath151 denotes the infinite honeycomb graph and @xmath86 denotes the honeycomb graph embedded on a torus of size @xmath69 .
the first result is the extension of the recursion relation to the infinite honeycomb graph .
[ proprecursioninfinite ] on the infinite honeycomb graph with @xmath226-weights it holds @xmath244+r_n\ ] ] with @xmath245\ ] ] for some finite constant @xmath246 .
on @xmath86 there are four parity classes corresponding to vertical / horizontal height changes when crossing the torus ( see @xcite ) : @xmath247 where @xmath248 is obtained from the kasteleyn matrix which has even horizontal and vertical height change and so on .
we apply corollary [ corfiniteabc ] to each parity class : let @xmath249 be the kasteleyn matrix associated with these four parity classes , i.e. , @xmath250=z_{{\mathcal{h}}_l}^{ij}$ ] .
then , we have @xmath251\end{gathered}\ ] ] where @xmath252 and @xmath253 represents the particle probability measure for the parity class @xmath254 .
we sum ( [ eq3.32 ] ) over the parity classes with the correct sign . after taking the thermodynamic limit @xmath85 ( which does exist because in @xcite
is was shown that it converges to the infinite gibbs measure characterized by the inverse kasteleyn matrix @xmath255 ) we obtain ( [ eq3.40 ] ) . to prove ( [ eq3.32b ] ) , notice that by the extra restrictions imposed on the number of dimer covers by reducing the number of vertices to be covered and by comparing allowed dimer configurations on the graphs @xmath256 , @xmath257 and @xmath258 , we have @xmath259 for some constant @xmath260 . we divide the above equation by @xmath261 which gives @xmath262 this means that by using we have @xmath263 \leq \tilde c bc \,\frac{(bc)^n z^{ij}_{{\mathcal{h}}_l\setminus\sigma_n}}{z^{ij}_{{\mathcal{h}}_l}}. \end{aligned}\ ] ] using ( [ eq3.30 ] ) , setting @xmath264 and summing over the parity classes gives the desired bound . with the result of proposition [ proprecursioninfinite ] we can now easily end the proof of theorem [ thminfinitesystem ] . notice that the left side of ( [ eq3.40 ] ) does not depend on @xmath30 and is finite , while the right side of ( [ eq3.40 ] ) is a sum of positive numbers .
therefore the last terms in the sum with @xmath265 goes to zero faster than @xmath266 , which in turns implies that @xmath267 as @xmath268 ( in reality , the decay is exponential as shown in @xcite ) .
therefore , by taking @xmath268 in ( [ eq3.40 ] ) we obtain @xmath269.\ ] ] finally notice that the prefactor in ( [ eq2.12 ] ) for @xmath270 and @xmath271 is exactly @xmath272 , i.e. , @xmath273 the other equalities are easy to compute and were already obtained in @xcite .
it seems plausible that theorem [ thminfinitesystem ] could also be verified using proposition [ proprecursionfinitet ] recursively for the kernel @xmath274 giving an algebraic proof .
however , a bound for the remainder , analogous to ( [ eq3.32b ] ) , would still be required .
this bound seems mysterious from the algebraic expression and even the positivity of the remainder is not obvious from the determinant expression .
here we derive a single integral representation for the inverse kasteleyn matrix @xmath274 given in ( [ eq2.14 ] ) .
let us do the change of variables @xmath275 so that @xmath276 _ case @xmath106_. in this case when @xmath277 is not in the arc of the circle of radius @xmath95 ( anticlockwise oriented ) from @xmath278 to @xmath96 , then no poles of @xmath279 lies inside its integration contour and the contribution is @xmath35 .
if @xmath277 is in the arc of circle from @xmath278 to @xmath96 , then there is a simple pole at @xmath280 , which leads to @xmath281 here the integration path can be any path crossing the real axis on @xmath107 .
_ case @xmath108_. in this case when @xmath277 is in the arc of the circle of radius @xmath95 ( anticlockwise oriented ) from @xmath278 to @xmath96 , then outside the integration contour of @xmath279 there are no poles and the integral over @xmath279 gives @xmath35 .
if @xmath277 is in the arc of circle from @xmath96 to @xmath278 , then there is a simple pole at @xmath280 , which leads to @xmath282 this is equal to the expression ( [ eqapp1 ] ) if the paths are chosen to cross real axis on @xmath109 .
a. borodin , _ determinantal point processes _ , the oxford handbook of random matrix theory ( g. akemann , j. baik , and ph . di francesco ,
eds . ) , oxford university press , 2011 , pp . 233252 [ arxiv:0911.1153 ] . k. johansson , _ random matrices and determinantal processes _ , mathematical statistical physics , session lxxxiii : lecture notes of the les houches summer school 2005 ( a. bovier , f. dunlop , a. van enter , f. den hollander , and j. dalibard , eds . ) , elsevier science , 2006 , pp . | the speed of growth for a particular stochastic growth model introduced by borodin and ferrari in @xcite , which belongs to the kpz anisotropic universality class , was computed using multi - time correlations .
the model was recently generalized by toninelli in @xcite and for this generalization the stationary measure is known but the time correlations are unknown . in this note , we obtain algebraic and combinatorial proofs for the expression of the speed of growth from the prescribed dynamics . |
the 8-meter diameter planetarium at the university of washington ( uw ) is primarily used for formal education as a part of our introductory astronomy classes .
graduate students also organize and coordinate informal educational presentations to k-12 students and astronomy clubs as well as programs for the general public . with an ms-8 star - ball projector and three vga projectors , the range of content of the presentations was limited by the lack of visualizations or at least constrained by the canned videos available . worldwide telescope ( wwt ) as a full dome digital projection system effectively eliminates the lack of visualizations , and with this software available to students and presenters every user can create their own tours and presentations ; removing the constraints of pre - made media .
we present our methods to digitize the uw planetarium on a tight budget i.e. , $ 40,000 total ( $ 10,000 for computers and $ 30,000 for projectors ) . in section [ proj ]
we discuss our choice of projectors and the need for conversion lenses and mounts , in section [ comp ] we explain the decisions involved in choosing the computer system . with the projectors and computers
nominally set up , we explain the alignment procedure in section [ alignm ] and in section [ conc ] we conclude by describing the expected integration of this digital dome within the educational curriculum .
with the ms-8 projector in the middle of our dome and a shelf already built into the perimeter of the planetarium about a foot under the spring - line ( see figure [ venus ] ) , we had a natural choice to place four projectors around the perimeter , with two projectors aimed to the top of the dome ( see figure [ jupiter ] .
we considered alternative designs including placing the projectors at ground - level in the center of the dome , and a single projector for the dome cap .
each of these scenarios had significant drawbacks .
these include : space limitations that restrict the number of projectors we can mount in the center of the dome , a concern that the the projectors would need to be shielded to block the projector beams from the eyes of the patrons , and few projectors in our price range with the features we desire are designed to be positioned at a steep angle .
ignoring this final limitation could at worst burn out a projector ( since the fans would not be functioning properly ) or at best void any warrantee .
the goals in choosing the projectors were to provide a rich and immersive environment with as high a resolution as possible .
as projectors are the primary cost of this upgrade we sought machines that would have longevity , good dynamic range , excellent black response and that would not be obsolete on a timescale of a few years .
we limited our search such that our projectors budget would be under $ 30,000 , we would need the maximum cost of a projector to be $ 5,000 .
in the summer of 2009 , it seemed impossible to find a projector at that cost such that when we placed four of them around the circumference of our planetarium , they filled the entire width including some extra space for the projected images to overlap .
in fact , there seemed to be a jump in projector prices from `` home theater '' ( $ 1,000-$4,000 ) to professional grade ( @xmath0 ) .
limited by home theater projectors , while ensuring the richest immersive experience that would also stay `` cutting edge '' for the longest time , we limited our search to projectors with the highest resolution available in 2009 , 1920x1080p .
this resolution usually implied an aspect ratio ( width of projected image @xmath1 to height of projected image ) of 16:9 . in a dome of a given diameter , @xmath2 ,
each projector must fill a width of @xmath3 . in the terms used in the projector industry ,
the home theater projectors have throw ratios ( distance to screen , @xmath4 , to width of image ) of about 1.8 - 3 .
so , as an estimate of the throw ratio necessary for a dome environment , our projectors would be placed about the diameter of the dome away from the image .
this makes the throw ratio @xmath5 , and by our approximation , @xmath6 .
this means the throw ratio would need to be 1.27 if there were no overlap , and less than 1.27 to have any overlap .
we overcame this `` short throw '' problem by using conversion lenses .
navitar , a company that makes conversion lenses , makes the screensaver 0.65 conversion lenses listed at $ 2,500 each , thus defining the same cost as the upper limit to the cost of a projector on our budget . with a 0.65 conversion lens
, the projected image width will be increased by a factor of 1.5 and we can move the projector closer to the screen by about 33% . in other words
the throw ratio will change to @xmath7 . with our demand of a @xmath8
, we need to find a projector with a throw ratio of less than 2.9 , an easy feat in the home theater market .
there are many home theater projectors available for under $ 2,500 that have the resolution we require , and with the constraint of the throw ratio relaxed , the next feature we looked into was the contrast ratio . for our purposes ,
the contrast ratio is a red herring .
it is supposed to compare the brightest pixel to the darkest pixel and thus giving the consumer an idea of how dark the projection when the pixel is defined to be black . with some projectors having automatically adjusting irises ( aka a dynamic iris , an iris that will block light instead of projecting dark ) , the ratio is not comparable from projector to projector . the contrast ratio becomes more of a distraction since it takes the control away from us to make what we want dim or bright .
since we wo nt be using the dynamic iris features on the projector , we looked for low lumen projectors with good dark levels with the assumption our audience s eyes will adjust to a very dark room .
this lead us to look at the projection type , that is liquid crystal display ( lcd ) or ti s digital light processing ( dlp ) .
lcd projectors have less pixellation issues but are known to have hot or dead pixels ( when lcd panels have defective transistors ) , while dlp projectors tend to do better with dark levels ( since tiny mirrors point the light away from the projection screen ) , thus we chose to limit ourselves to a dlp projector with around 1600 lumen brightness . to nail down the brightness levels , we contacted avidex , a seattle based audio / visual design company , to compare different projector models . based on these tests we selected the sharp xv - z15000 but we believe that the appropriate choice for a given planetarium will depend on the dome surface and the configuration of the planetarium itself . after picking out the locations of the projectors in the planetarium
, we needed to allow small amounts of freedom in rotation and inclination so we can control the overlap areas but lock down the focus .
the need for conversion lenses added extra complications .
we needed to optically align the lenses and projectors and find rigid placements for the mounted systems .
to do this , mounts were designed by one of us ( lc ) , an optical engineer with the uw telescope engineering group . once the physical location ( x , y and z ) of the projectors had been defined , the remaining two degrees of freedom that required adjustability for final positioning the projector were azimuthal rotation and rotation in altitude .
azimuthal rotation is handled using the two fasteners that attach the base of the adjustable mount to the selected mounting surface of the planetarium .
the mounting base plate is designed with one mounting hole that is bolted down to the mounting surface in the planetarium at the desired `` x , y , and z '' location for the projector , providing a pivot point for the base plate and plane for azimuthal rotation .
the second fastener fits into a curved slot that is centered on the pivot point , allowing for azimuthal rotation of the mount and projector assembly around the first fastener .
rotation in altitude is controlled by hinging , on a horizontal axis , the mounting plate on which the projector sits , relative to the base plate that is mounted to the planetarium s mounting surface .
the altitude adjustment is controlled using spherical washers and nuts on a fine thread screw that can fix the projector mounting plate in the desired position using lock nuts .
the plate on which the projector is mounted also provides the mounting surface for the lens ( mounted in front of the projector ) that is needed to re - size the projector image for the planetarium dome .
the custom designed hinges and mounting plates were manufactured and assembled by uw astronomy students ( bt and cs ) in the uw physics student machine shop ( for engineering drawings please contact the authors ) .
computer technology changes faster than the speed of conference proceedings so we will limit the amount of detail in this section .
our goal was to maximize the video display possible while minimizing the cost of the system .
we chose to use seven computers , six of them attached to the projectors in client mode and the seventh , the master mode computer .
for our video card , we chose nvidia s gtx275 .
the seven computers all have windows 7 os with a 1 tb hard drive and 6 gb ram . to communicate from one computer to the another ( for projector alignment , software instillation , and running wwt ) an ethernet switch is required and a virtual network computing ( vnc ) program is recommended .
windows 7 has a home network setup with remote connections possible , making it easy to install a program from the master computer on a client , but the software logs the user out of the client computer , while a vnc program will allow the master user to see their actions on the client monitor ( or in this case , projected on the dome ) .
another consideration when purchasing multiple computers is the option for `` wake on lan '' in the system bios of the motherboard chosen .
this would allow the master computer to turn on the other machines instead of having to physically switch on each computer at every use .
once the projectors , lenses , and mounts are in place , the software must know which pixels go where , how they are distorted , and which pixels are in the overlapped sections of the projection .
projection designer , a manual set up tool to correct for distortion and edge blending , is a free and open source way to align the projectors in your planetarium or any multi - projector environment .
it is not a trivial procedure and takes some time , we found that it is easiest with more than one person . starting with the projector aimed at the `` sweet spot '' the place on the dome that will be the apex of any traveling - like movement
, we aligned a projected full dome grid to points and lines existing on our dome or projected from our ms-8 star - ball .
the output of projection designer is a set of blend and distortion maps that wwt has been designed to read . at this point , see the wwt documentation for instructions on incorporating projection designer s output files .
we ve shown how we ve digitized the university of washington s planetarium with off - the - shelf equipment and wwt software .
table [ t - costs ] contains a summary of our costs . we will be incorporating this new technology into formal and informal education at all levels .
in addition to the planetarium experience we have all grown up with , wwt can currently project more than thirteen all - sky data sets onto the night sky covering nearly every detectable band of the electromagnetic spectrum even allowing cross fading from one to another .
wwt also has 3-d interactive visualization with the hipparcos catalogue and sloan galaxy map . with a click of a mouse or a twist of joy stick or xbox controller ,
the presenter or the student from home can fly through the observed universe : look at the planets orbit the sun from the vantage point of pluto , see how the night sky appears from a star far from the sun , watch a swarm of millions of minor planets in their trajectories around the sun , and zoom through the coma cluster of galaxies .
this is a huge visualization upgrade from a geocentric star - ball . for upper level undergraduate work , wwt can make real time calls to astronomical databases to access research grade data on ads , simbad , seds , and ned , it can download existing fits or jpg image files , and even import fits files that students in an observing class would take with a telescope and place them correctly in the sky . at the uw ,
a 16 " telescope for undergraduate projects is very near our planetarium , and future capabilities would drive it from the planetarium displaying the raw image taken on the dome .
students are to search the virtual observatory registry and overlay data or output .xml files to be plotted or analyzed in a myriad of ways . for general information they can simply right click on the object in question and look it up directly in wikipedia .
instructors wishing to duplicate or save their own classroom planetarium shows or astronomy public outreach talks can create a tour to run an automated planetarium show that would record their own voice , incorporate images and music from external sources , and write text as if it were a dome filled powerpoint presentation . the tour
could then be saved to disc for the instructor to use each quarter , for a ta to play for the class , or to upload for the wwt online community .
a slew of guided tours are already available , created by astrophysicists from the adler planetarium in chicago , the smithsonian center for astrophysics , the space telescope science institute and more .
with wwt in a full digital dome , all levels of undergraduates in astronomy could begin classroom labs in the planetarium and finish their assignments on their home computers or laptops .
graduate students and advanced undergraduate students will have a unique opportunity to create truly interactive shows for the public .
we ve outlined how we put it together costing less than $ 40,000 . | digital planetariums can provide a broader range of educational experiences than the more classical planetariums that use star - balls .
this is because of their ability to project images , content from current research and the 3d distribution of the stars and galaxies . while there are hundreds of planetariums in the country the reason that few of these are full digital is the cost . in collaboration with microsoft research ( msr )
we have developed a way to digitize existing planetariums for approximately $ 40,000 using software freely available .
we describe here how off the shelf equipment , together with msr s worldwide telescope client can provide a rich and truly interactive experience .
this will enable students and the public to pan though multi - wavelength full - sky scientific data sets , explore 3d visualizations of our solar system ( including trajectories of millions of minor planets ) , near - by stars , and the sdss galaxy catalog . |
the optical hall effect ( ohe ) is a physical phenomenon , which describes transverse and longitudinal magnetic field - induced birefringence , caused by the non - reciprocal magneto - optic response of electric charge carriers .
the term ohe is used in analogy to the classic , electrical hall effect@xcite , since the electrical hall effect and certain cases of ohe observation can be explained by extensions of the classic drude model for the transport of electrons in matter ( metals).@xcite for the ohe , drude s classic model is extended by a magnetic field and frequency dependency , describing the electron s momentum under the influence of the lorentz force . as a result
an antisymmetric contribution is added to the dielectric tensor @xmath3 , whose sign depends on the type of the free charge carrier ( electron , hole ) .
the non - vanishing off - diagonal elements of the dielectric tensor reflect the magneto - optic birefringence , which lead to conversion of _
p_-polarized into _
s_-polarized electromagnetic waves , and vice versa .
the ohe can be quantified in terms of the mueller matrix , which characterizes the transformation of an electromagnetic wave s polarization state.@xcite experimentally the mueller matrix is measured by generalized ellipsometry ( ge).@xcite during a ge measurement different polarization states of the incident light are prepared and their change upon reflection from or transmission through a sample is determined .
an ohe instrument conducts ge measurements on samples in high , quasi - static magnetic fields , and detects the magnetic field induced changes of the mueller matrix.@xcite though several instruments with partial ohe instrument characteristics were described in the literature , most instruments did not fulfill all criteria for an ohe instrument .
nederpel and martens developed in 1985 a single wavelength ( 444 nm ) magneto - optical ellipsometer for the visible spectral range , but the instrument provided only low magnetic fields ( @xmath4 mt).@xcite in 2003 erne _
_ presented a magneto - polarimetry instrument ( @xmath5 t ) for the mid - infrared spectral range ( spectral lines of co@xmath6 laser)@xcite , and in 2004 padilla _
et al . _ developed a terahertz - visible ( 6 to 20000 @xmath0 ) magneto - reflectance and -transmittance instrument ( @xmath7 t)@xcite .
while both instruments provide high magnetic fields , and contain polarizers and photo - elastic - modulators , these instruments were not designed to record mueller matrix data ( ge ) .
the full @xmath2 mueller matrix in the terahertz - mid - infrared spectral range ( 20 to 4000 @xmath0 ) can be measured by an instrument described in 2013 by stanislavchuk _
et al._,@xcite but here the instrument is not designed for experiments with the sample in magnetic fields . the first full ohe instrument developed in 2006 by hofmann _
et al . _ for the far - infrared ( fir ) spectral range ( 30 to 650 @xmath0 ) provided magnetic fields up to @xmath8 t and allowed sample temperatures between @xmath9 k and room temperature.@xcite this first ohe instrument has since been successfully used to determine free charge carrier properties@xcite , including effective mass parameters for a variety of material systems.@xcite later , ohe experiments were conducted in the terahertz ( thz ) spectral range,@xcite but were limited to room temperature and low magnetic fields .@xcite since the magnitude of the ohe depends on the magnetic field strength , higher magnetic fields facilitate the detection of the ohe . furthermore , the sensitivity to the ohe is greatly enhanced by phonon mode coupling,@xcite surface guided waves@xcite and fabry - prot interferences.@xcite since these effects appear from the thz to the mid - infrared ( mir ) spectral range , depending on the structure and material of the sample , it is necessary to extend the spectral range covered by ohe instrumentation .
an ohe instrument for the mir , for example , can detect the magneto - optic response of free charge carriers enhanced by phonon modes present in the spectral range above 600 @xmath0 , which applies to many substrate materials , _
e.g. _ , sic,@xcite al@xmath6o@xmath10,@xcite or gan,@xcite as well as to many materials used for thin films , _
e.g. _ , iii - v nitride semiconductors al@xmath11ga@xmath12n,@xcite al@xmath11in@xmath12n@xcite or in@xmath11ga@xmath12n@xcite . in addition , inter - landau - level transitions can be studied in the mir spectral range@xcite with a mir ohe instrument.@xcite the extension to the thz spectral range enables the detection of the ohe in samples with low carrier concentrations.@xcite furthermore , the strongest magneto - optic response can be observed at the cyclotron resonance frequency , which typically lies in the microwave / thz spectral range for moderate magnetic fields ( few tesla ) and effective mass values comparable to the free electron mass . in this article , we present an ohe instrument , covering an ultra wide spectral range from 3 @xmath0 to 7000 @xmath0 ( 0.1210 thz or 0.4870 mev ) , which combines mir , fir and thz magneto - optic generalized ellipsometry in a single instrument .
this integrated mir , fir and thz ohe instrument incorporates a commercially available , closed cycle refrigerated , superconducting 8 tesla magnet - cryostat sub - system , with four optical ports , providing sample temperatures between @xmath13 k and room temperature .
the ellipsometer sub - systems were built in - house and operate in the rotating - analyzer configuration , capable of determining the normalized upper @xmath14 block of the sample mueller matrix .
the operation of the integrated mir , fir and thz ohe instrument is demonstrated by three sample systems . combined experimental data from the mir , fir and thz spectral range of a single epitaxial graphene sample , grown on a 6__h__-sic substrate by thermal decomposition , are shown .
the mir ohe data of the same epitaxial graphene sample is used to demonstrate the operation of the mir ellipsometer sub - system of the integrated mir , fir and thz ohe instrument,@xcite over the full available magnetic field range of the instrument .
the magneto - optic response of free charge carriers and quantum mechanical inter - landau - level transitions is observed , and their polarization selection rules obtained therefrom are briefly discussed .
a te - doped , n - type gaas substrate serves as a model system for the fir spectral range of the fir / thz ellipsometer sub - system .
the ohe signal originating from conduction band electrons in a bulk material is discussed , and the concentration , mobility , and effective mass of the conduction band electrons is determined . finally , ohe data from an algan / gan high electron mobility transistor structure ( hemt ) from the thz spectral range of the fir / thz ellipsometer sub - system is presented and analyzed.@xcite the data was recorded at different temperatures between @xmath15 k and room temperature , representing the full sample temperature range of the instrument .
the manuscript is organized as follows , in section [ sec : theory ] dielectric and magneto - optic dielectric tensors are introduced , a brief theoretical overview on mueller matrices and ge data - acquisition is given , and general ge data analysis procedures are introduced . section [ sec : experimental setup ] gives a detailed description of the experimental setup , while in section [ sec : acq - anal ] data acquisition and data analysis procedures for ohe data are discussed .
examples of experimental results , demonstrating the operation of the integrated mir , fir and thz ohe instrument , are presented and discussed in section [ sec : samplesystems ] , which is followed by a short summary in section [ sec : summary ] .
the evaluation of physical relevant parameters from the ohe requires the experimental observation and quantification of the ohe , and a physical model to analyze ohe data .
experimentally , the ohe is quantified in terms of the mueller matrix @xmath16@xcite by employing generalized ellipsometry ( ge ) .
the physical model which is used to analyze the observed transverse and longitudinal magneto - optic birefringence of the ohe is based on the magneto - optic dielectric tensor @xmath17 , which is a function of the slowly varying external magnetic field @xmath18 . if , among other parameters , the magneto - optic dielectric tensor of a sample is known , experimental mueller matrices @xmath16 can be modeled from @xmath17 using the relation @xmath19 this relation is in general not invertible analytically , but can be used to determine the magneto - optic dielectric tensor from experimental mueller matrix data through non - linear model regression analysis@xcite .
dielectric tensors , mueller matrix calculus , generalized ellipsometry including data acquisition , as well as data analysis will be addressed in this section .
the optical response of a sample is here described by the dielectric tensor @xmath20 .
if the dielectric tensor of the sample without a magnetic field is given by @xmath21 and the change of the dielectric tensor induced by a magnetic field @xmath18 by @xmath22 , the magneto - optic dielectric tensor describing the ohe , can be expressed as @xmath23 the magneto - optic response of the sample described by @xmath22 usually originates from bound and unbound charge carriers subjected to the magnetic field and is caused by the action of the lorentz force .
the magneto - optic response is anisotropic , and non - reciprocal in time.@xcite thus , the corresponding magneto - optic contributions @xmath24 and @xmath25 to the permittivity tensor @xmath26 , where @xmath27 is the @xmath1 identity matrix , originate from the interaction of right- and left - handed circularly polarized light with the sample , respectively.@xcite without loss of generality , if the magnetic field @xmath18 is pointing in the _
z_-direction , the polarization vector can be described by arranging the electric fields in their circularly polarized eigensystem by @xmath28 , where @xmath29 is the imaginary unit.@xcite transforming @xmath30 back into the laboratory system the change of the dielectric tensor induced by the magnetic field takes the form:@xcite is not part of the magneto - optic dielectric tensor @xmath22 since it is already incorporated in the non - magnetic part of the dielectric tensor @xmath21 . ]
@xmath31 note , under field inversion @xmath32 , the polarizabilities for left- and right - handed circularly polarized light interchange .
@xmath22 is only diagonal if @xmath33 , and otherwise is non - diagonal with anti - symmetric off diagonal elements .
charged carriers , subject to a slowly varying magnetic field obey the classical newtonian equation of motion ( lorentz - drude model)@xcite @xmath34 where @xmath35 , @xmath36 , @xmath37 , @xmath38 and @xmath39 represent the effective mass tensor , the electric charge , the mobility tensor , the spatial coordinate of the charged carrier and the eigenfrequency of the undamped system without external excitation and magnetic field , respectively . for a time harmonic electromagnetic plane wave with an electric field @xmath40 with angular frequency @xmath41 , the time derivative of the spatial coordinate of the charge carrier is @xmath42 , where @xmath43 is the velocity of the charge carrier . with the current density @xmath44 eq .
( [ eqn : fcc - newton ] ) reads @xmath45\;,\ ] ] where @xmath46 is the charge carrier density . with the levi - cevita - symbol @xmath47 , the conductivity tensor @xmath48 , the dielectric constant @xmath49 , and using and the dielectric tensor for charge carriers subject to the external magnetic field @xmath18 can be expressed as @xmath50^{-1}\;.\ ] ] [ [ sec : phonondf ] ] polar lattice vibrations ( lorentz oscillator ) +
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + for isotopic effective mass tensors the cyclotron frequency @xmath51 can be defined . for the mass of the vibrating atoms of polar lattice vibrations ,
the cyclotron frequency is several orders of magnitude smaller than for effective electron masses , and can be neglected for the magnetic fields and spectral ranges discussed in this paper .
therefore , the dielectric tensor of polar lattice vibrations @xmath52 can be approximated using eq .
( [ eqn : fulldrude ] ) with @xmath53 .
when assuming isotropic effective mass and mobility tensors , the result is a simple harmonic oscillator function with lorentzian - type broadening.@xcite for materials with orthorhombic symmetry and multiple optical excitable lattice vibrations , the dielectric tensor can be diagonalized to @xmath54 where @xmath55 ( @xmath56 ) is given by@xcite @xmath57 where @xmath58 , @xmath59 , @xmath60 , and @xmath61 denote the component of the frequency and the broadening values of the @xmath62 longitudinal optical ( lo ) and transverse optical ( to ) phonon modes , respectively , while the index @xmath63 runs over @xmath64 modes .
further details can be found in refs . , and a detailed discussion of the requirements to the broadening parameters , such as im@xmath65 , in ref . .
[ [ sec : drudedf ] ] free charged carriers ( extended drude model ) + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + for free charged carriers no restoring force is present and the eigenfrequency of the system is @xmath66 . for isotropic effective mass and conductivity tensors , and magnetic fields aligned along the _
z_-axis eq .
( [ eqn : fulldrude ] ) can be written in the form @xmath67 , with the drude dielectric tensor for @xmath68 @xmath69 where @xmath70 is the plasma frequency , and @xmath71 is the isotropic drude dielectric function .
the magneto - optic contribution to the dielectric tensor @xmath72 for isotropic effective masses and conductivities can be expressed , using eq .
( [ eq : llteps ] ) , through polarizability functions for right- and left - handed circularly polarized light @xmath73 where @xmath74 is the isotropic cyclotron frequency .
the dielectric tensor @xmath75 describing a series of inter - landau - level transitions can be approximated by a sum of lorentz oscillators .
the quantities @xmath76 in eq .
( [ eq : llteps ] ) are then expressed by @xmath77 where @xmath78 , @xmath79 , and @xmath80 are amplitude , transition energy , and broadening parameter of the @xmath81 inter - landau - level transition , respectively , which in general depend on the magnetic field .
the phase factor @xmath82 was introduced empirically here to describe the experimentally observed line shapes of all mueller matrix elements . for inter - landau - level transitions in graphite or bi - layer graphene
we find @xmath83 , and for inter - landau - level transitions in single layer graphene @xmath84 .
note that for @xmath84 , the polarizabilities for left and right handed circularly polarized light are equal , and @xmath75 is diagonal .
the real - valued stokes vector @xmath85 has four components , carries the dimension of an intensity , and can quantify any polarization state of plane electromagnetic waves . if expressed in terms of the _ p_- and _ s_-coordinate system , its individual components can be defined by @xmath86 , @xmath87 , @xmath88 , and @xmath89 , with @xmath90 , @xmath91 , @xmath92 , @xmath93 , @xmath94 , and @xmath95 being the intensities for the _ p_- , _
s_- , + 45@xmath96 , -45@xmath96 , right- and left - handed circularly polarized light components , respectively.@xcite the real - valued @xmath2 mueller matrix @xmath97 describes the change of electromagnetic plane wave properties ( intensity , polarization state ) , expressed by a stokes vector @xmath85 , upon change of the coordinate system or the interaction with a sample , optical element , or any other matter@xcite @xmath98 where @xmath99 and @xmath100 denote the stokes vectors of the electromagnetic plane wave before and after the change of the coordinate system , or an interaction with a sample , respectively .
note that all mueller matrix elements of the ge data discussed in this paper , are normalized by the element @xmath101 , therefore @xmath102 and @xmath103 .
the mueller matrix can be decomposed in 4 sub - matrices , where the matrix elements of the two off - diagonal - blocks @xmath104 only deviate from zero if _
p_-_s_-polarization mode conversion appears , while the matrix elements in the two on - diagonal - blocks @xmath105 mainly contain information about _ p_-_s_-polarization mode conserving processes .
_ p_-_s_-polarization mode conversion is defined as the transfer of energy from the _
p_-polarized channel of an electromagnetic plane wave to the _
s_-polarized channel , or vice versa .
polarization mode conversion can appear when the _ p_-_s_-coordinate system is different for @xmath100 and @xmath99 rotation of the coordinate system around the beam path . ] , or when a sample shows birefringence , for example .
in particular , polarization mode conversion appears if the dielectric tensor of a sample possesses non - vanishing off - diagonal elements .
therefore , in mueller matrix data from optically isotropic samples , ideally all off - diagonal - block elements vanish , while , for example , magneto - optic birefringence can cause non - zero off - diagonal - block elements in the mueller matrix . here
, we define ohe data as mueller matrix data from an ohe experiment [ eq .
( [ eqn : mm - df ] ) ] with magnetic field
@xmath106 @xmath107 furthermore , we define the derived ohe datasets @xmath108 as difference data between the mueller matrix datasets , measured at the magnetic field @xmath106 and the corresponding zero field dataset @xmath109 where @xmath110 is the mueller matrix of the zero field experiment and @xmath111 is the magnetic field induced change of the mueller matrix .
this form of presentation is in particular advantageous in case the magnetic field causes only small changes in the mueller matrix , and provides improved sensitivity to magnetic field dependent model parameters during data analysis .
another form of presentation for derived ohe data is @xmath112 which can be used to inspect symmetry properties of magneto - optic mueller matrix data , and can help to improve the sensitivity to magnetic field dependent model parameters during data analysis .
rotating element ellipsometers can be classified into two categories:@xcite ( i ) rotating analyzer ellipsometers ( rae)@xcite in polarizer - sample - rotating - analyzer ( @xmath113 ) or rotating - polarizer - sample - analyzer ( @xmath114 ) configuration , capable to measure the upper left @xmath1 block of the mueller matrix ; ( ii ) rotating compensators ellipsometers ( rce)@xcite in polarizer - sample - rotating - compensator - analyzer ( @xmath115 ) or polarizer - rotating - compensator - sample - analyzer ( @xmath116 ) configuration , capable to measure the upper left @xmath117 or @xmath118 block of the mueller matrix , respectively .
the mueller matrices of a polarizer @xmath119 , analyzer @xmath120 , compensator @xmath121 with phase shift @xmath122 , coordinate rotation along beam path @xmath123 by an angle @xmath124 , and of the sample @xmath97 are given by @xmath125{\(-\)}\sin \delta\\ \hspace{8pt}0&\hspace{17pt}0&\hspace{10pt}\sin \delta&\hspace{7pt}\cos \delta \end{bmatrix}\;,\\ & \mathbf{r}\left(\theta\right ) & = \ \ \ & \begin{bmatrix } \hspace{8pt}1&\hspace{6pt}0 & \hspace{-12pt}0 & \hspace{1pt}0&\hspace{5pt } { } \\
\hspace{8pt}0&\hspace{4pt}\cos 2\theta_j & \hspace{-1pt}\sin 2\theta_j & \hspace{1pt}0&\hspace{5pt } { } \\
\hspace{8pt}0&\hspace{-2pt}\scalebox{0.6}[1.0]{\(-\)}\hspace{1pt}\sin 2\theta_j & \hspace{-1pt}\cos 2\theta _ j & \hspace{1pt}0&\hspace{5pt } { } \\ \hspace{8pt}0&\hspace{7pt}0 & \hspace{-12pt}0 & \hspace{1pt}1&\hspace{5pt } { } \end{bmatrix}\;,\\ & \mathbf{m } &
= \ \ \ & \begin{bmatrix } \hspace{3pt}m_{11}&\hspace{4pt}m_{12}&\hspace{4pt}m_{13}&\hspace{4pt}m_{14}&\hspace{-2pt } { } \\ \hspace{3pt}m_{21}&\hspace{4pt}m_{22}&\hspace{4pt}m_{23}&\hspace{4pt}m_{24}&\hspace{-2pt } { } \\ \hspace{3pt}m_{31}&\hspace{4pt}m_{22}&\hspace{4pt}m_{33}&\hspace{4pt}m_{34}&\hspace{-2pt } { } \\ \hspace{3pt}m_{41}&\hspace{4pt}m_{42}&\hspace{4pt}m_{43}&\hspace{4pt}m_{44}&\hspace{-2pt } { } \end{bmatrix}\ ; , \end{aligned}\ ] ] respectively .
execution of the matrix multiplication characteristic for the corresponding ellipsometer type@xcite shows that , due to the rotation of optical elements , the measured intensity at the detector is typically sinusoidal .
fourier analysis of the detector signal provides fourier coefficients , which are used to determine the mueller matrix of the sample ( see sec .
[ sec : ohe - acq ] ) .
ellipsometry is an indirect experimental technique .
therefore , in general , ellipsometric data analysis invokes model calculations to determine physical parameters in dielectric tensors or the thickness of layers , for instance.@xcite sequences of homogeneous layers with smooth and parallel interfaces are assumed in order to calculate the propagation of light through a layered sample , by the 4@xmath1264 matrix formalism.@xcite to best match the generated data with experimental results , parameters with significance are varied and mueller matrix data is calculated for all spectral data points , angles of incidence and magnetic fields . during the mean square error ( mse ) regression , the generated mueller matrix data @xmath127 is compared with the experimental mueller matrix data @xmath128 and their match is quantified by the mse @xmath129 where @xmath130 , @xmath131 and @xmath132 denotes the total number of data points , the total number of parameters varied during the non - linear regression process , and the standard deviation of @xmath128 , obtained during the experiment , respectively . for fast convergence of the mse regression , the levenberg - marquardt fitting algorithm is used.@xcite the mse regression is interrupted when the decrease in the mse is smaller than a set threshold and the determined parameters are considered as best model parameters .
the sensitivity and possible correlation of the varied parameters is checked and , if necessary , the model is changed and the process is repeated.@xcite
figure [ fig : cad - overview ] shows ( side view ) the integrated mir , fir and thz ohe instrument with its four sub - systems : ( a ) the mir ellipsometer sub - system , ( b ) the fir / thz ellipsometer sub - system , ( c ) the magneto - cryostat sub - system , and ( d ) the magneto - cryostat transfer sub - system . in order to utilize the magneto - cryostat sub - system with the mir or the fir / thz ellipsometer sub - system the magneto - cryostat transfer sub - system was installed .
the integrated mir , fir and thz ohe instrument contains multiple light sources and detectors , and covers a spectral range from 3 @xmath0 to 7000 @xmath0 ( 0.1210 thz or 0.4870 mev ) .
both ellipsometer sub - systems can be operated without the magneto - cryostat sub - system , in a variable angle of incidence ellipsometry mode ( @xmath133 ) .
figure [ fig:2dsetup ] shows a schematic overview ( top view ) of all major components in the integrated mir , fir and thz ohe instrument . the upper part of fig .
[ fig:2dsetup ] shows a schematic drawing ( top view ) of the optical configuration of the mir sub - system of the integrated mir , fir and thz ohe instrument .
the mir ellipsometer sub - system is composed of ( i ) the mir source unit , ( ii ) the polarization state preparation unit , ( iii ) the mir goniometer unit , and ( iv ) the polarization state detection unit . to minimize absorption due to water vapor ,
the complete beam path of the mir ellipsometer sub - system is purged with dried air.@xcite due to the high magnetic stray - fields ( see fig . [
fig : cad ] ) , all opto - mechanical components in the polarization state preparation and detection units were designed and manufactured without ferromagnetic materials ( with exception of the stepper motors ) .
the mir source unit of the mir ellipsometer sub - system is a bruker vertex 70 fourier - transform - infrared spectrometer ( fig .
[ fig:2dsetup ] : mir - ftir ) with a silicon carbide globar light source ( spectral range 5807000 @xmath0 ) . after being collimated
, the light beam passes the interferometer ( potassium bromide ( kbr ) beam splitter ) , is reflected by a plane mirror , and exits the mir source unit .
the beam enters the polarization state preparation unit . inside the polarization state preparation unit
the beam passes a beam steering plane mirror assembly ( m@xmath134 ) , a source selection and beam focusing assembly ( r@xmath134 ) and a rotation stage assembly ( p@xmath134-b@xmath134-st@xmath134 ) , respectively .
the beam steering plane mirror assembly ( m@xmath134 ) is composed of an opto - mechanic mount and a plane first surface gold mirror@xcite .
the source selection and beam focusing assembly ( r@xmath134 ) comprises two sub - assemblies ( detailed drawing : fig .
[ fig : beamlift ] a ) , the rotatable plane mirror sub - assembly ( stepper motor , axis extension , mechanic mount for axis extension , opto - mechanic mount for mirror , plane first surface gold mirror@xcite ) and the beam focusing off - axis paraboloid stage sub - assembly ( opto - mechanic mount for paraboloid , gold surface @xmath135 off - axis paraboloid with an effective focal length of @xmath136 mm is the distance between the focal point and the center of the off - axis paraboloid mirror . for 90@xmath137 off - axis paraboloids
the effective focal length is twice the focal length @xmath138 . ] ) , which focuses the beam onto the sample position .
the focused beam then reaches the rotation stage assembly ( p@xmath134-b@xmath134-st@xmath134 ) which has a nominal angular resolution of 0.045@xmath96 . the rotation stage assembly ( detailed drawing : fig .
[ fig : beamlift ] b ) contains a krs-5 substrate based wire grid polarizer ( p@xmath134 ) , which is mounted in a hollow tube , which is fitted into two polymer bearings with glass balls , which are held by an aluminum block .
a 48-tooth polymer gear is mounted to the hollow tube , and is connected via a kevlar timing belt ( b@xmath134 ) to a 12-tooth polymer gear on a stainless steel shaft ( gear ratio : 1:4 ) , leading to the stepper motor ( st@xmath134 ) .
after passing the rotation stage assembly , the polarized and focused beam leaves the polarization state preparation unit .
the beam is then reflected by , or transmitted through the sample ( s@xmath134 ) . the sample can be mounted on a sample holder , attached to the mir goniometer unit ( g@xmath134 ) ( commercially available 2-circle goniometer 415 , huber diffraktionstechnik ) , or inside the magneto - cryostat sub - system ( m1 ) .
if the magneto - cryostat sub - system is used , reflection type measurements can only be conducted at a @xmath139 angle of incidence .
a detailed description of the magneto - cryostat sub - system , its sample mount , and the optical window configuration is given in section ( [ sec : magnet ] ) .
the beam then enters the polarization state detection unit , which is mounted to the rotatable arm of the mir goniometer unit ( g@xmath134 ) .
the polarization state detection unit contains a rotation stage assembly ( a@xmath134-b@xmath6-st@xmath6 ) , a beam collimation and detector selection assembly ( r@xmath6 ) , and three beam focusing / detection assemblies ( o@xmath134-mct , o@xmath6-dtgs@xmath134 and o@xmath10-b@xmath134 ) .
the rotation stage assembly ( a@xmath134-b@xmath6-st@xmath6 ) is equivalent to the one in the polarization state preparation unit ( p@xmath134-b@xmath134-st@xmath134 , fig .
[ fig : beamlift ] b ) , but in the polarization state detection unit the krs-5 substrate based wire grid polarizer serves as the analyzer ( a@xmath134 ) of the mir ellipsometer sub - system .
the beam collimation and detector selection assembly ( r@xmath6 ) is composed of two sub - assemblies : the rotatable plane mirror sub - assembly and the collimating off - axis paraboloid stage sub - assembly ( gold surface @xmath135 off - axis paraboloid , @xmath136 mm ) .
both sub - assemblies are equivalent to the source selection and beam focusing assembly in the polarization state preparation unit ( fig .
[ fig : beamlift ] a ) , but are used in reverse order to first collimate and then redirect the beam to one of the three beam focusing / detection assemblies using the rotatable mirror@xcite .
each beam focusing / detection assembly contains a beam focusing off - axis paraboloid stage sub - assembly and a detector sub - assembly .
all detector sub - assemblies contain an opto - mechanic mount and a detector .
the beam focusing off - axis paraboloid stage sub - assemblies ( o@xmath140 ) are composed of an opto - mechanic mount and a 90@xmath137 off - axis paraboloid with an uncoated gold surface .
the focal lengths of the off - axis paraboloids are matched to the corresponding detector .
the off - axis paraboloids ( o@xmath141 , o@xmath142 ) for the liquid nitrogen cooled hgcdte - detector sub - assembly ( mct ) and the pyroelectric , solid state deuterated triglycine sulfate detector sub - assembly ( dtgs@xmath141 ) , each have an effective focal length of 38 mm ( 1.5 in ) .
the off - axis paraboloid o@xmath143 for the liquid helium cooled bolometer detector sub - assembly ( b@xmath141)@xcite has an effective focal length of 190.5 mm ( 7.5 in ) . the signal of the used detector is fed back into the mir - ftir - spectrometer to record interferograms . for more information on the data acquisition and processing
see section [ sec : data analysis ] . the lower part of fig .
[ fig:2dsetup ] shows a schematic drawing ( top view ) of the optical configuration of the fir / thz ellipsometer sub - system of the integrated mir , fir and thz ohe instrument .
the fir / thz ellipsometer sub - system can be divided in five units : ( i ) the fir source unit , ( ii ) the thz source unit , ( iii ) the polarization state preparation unit , ( iv ) the fir / thz goniometer unit , and ( v ) the polarization state detection unit . for measurements in the fir spectral range
, the fir / thz ellipsometer sub - system is operated in analyzer - step mode,@xcite while for measurements in the thz spectral range the fir / thz ellipsometer sub - system is operated in continuously rotating analyzer mode . to minimize absorption due to water vapor
, the complete beam path of the fir / thz ellipsometer sub - system can be purged with dried air.@xcite due to the high magnetic stray - fields ( see fig . [
fig : cad ] ) , all opto - mechanical components in the thz source unit , polarization state preparation unit and polarization state detection unit were designed and manufactured without ferromagnetic materials ( with exception of the stepper motors and the thz source ) .
the fir source unit of the fir / thz ellipsometer sub - system is a bruker vertex v-70 ftir - spectrometer ( fig .
[ fig:2dsetup ] : fir - ftir ) .
the spectrometer is equipped with a silicon beam splitter but otherwise identical to the spectrometer used as mir source unit in the mir ellipsometer sub - system .
+ the thz source unit comprises five assemblies : the thz source and thz - beam collimation assembly ( fig .
[ fig:2dsetup ] : bwo - o@xmath144 ) , the optical chopper assembly ( ch ) , the beam steering plane mirror assembly ( m@xmath143 ) , polarization state rotator assembly ( pr - b@xmath143-st@xmath143 ) , and the beam steering plane mirror assembly ( m@xmath144 ) .
the thz source and thz - beam collimation assembly contains a thz source sub - assembly and a thz - beam collimating off - axis paraboloid stage sub - assembly . the thz source sub - assembly
is composed of an opto - mechanic mount and the backward wave oscillator ( bwo ) thz source ( microtech ) .
the bwo source emits thz radiation with a high brilliance ( bandwidth @xmath1452 mhz ) and a high output power ( @xmath1450.10.01 w ) . the thz radiation is almost perfectly linearly polarized and the orientation of the polarization is fixed in space .
the base frequency range of the bwo is 107177 ghz , which can be converted to higher frequency bands using gaas schottky diode frequency multipliers .
the spectral range accessible by the bwo can be expanded to 220350 ghz ( @xmath146 multiplier ) , 330525 ghz ( @xmath147 multiplier ) , 6501040 ghz ( @xmath146 and @xmath147 multiplier ) and 9801580 ghz ( double @xmath147 multiplier ) . for further details on the bwo based thz source and thz ellipsometry
are given in ref . and
references therein .
the thz radiation from the bwo is collimated by the thz - beam collimating off - axis paraboloid stage sub - assembly , composed of an opto - mechanic mount and a 90@xmath137 off - axis paraboloid ( o@xmath144 ) with an uncoated gold surface and an effective focal length @xmath148 mm .
the thz - beam then reaches the optical chopper assembly ( ch ) , which contains an opto - mechanic mount and a 3 bladed optical chopper , driven by a linear motor .
the 3 bladed optical chopper is rotated with a frequency of @xmath149 hz , resulting in a optical chopping frequency of @xmath150 hz , which is close to the optimal frequency response of the golay cell detector ( @xmath151 1215 hz ) .
after interaction with thz - beam steering plane mirror assembly ( m@xmath143 ) ( opto - mechanic mount and plane first surface gold mirror@xcite ) , the thz - beam is redirected to the polarization state rotator assembly ( pr ) . the polarization state rotator assembly ( fig .
[ fig : pr ] ) is an odd - bounce image rotation system.@xcite the polarization state rotator is designed to rotate the polarization state of an incoming electromagnetic beam azimuthally in a non - deviating , non - displacing fashion ( with respect to the incoming electromagnetic beam direction ) , and is used to pre - align the polarization direction of the thz - beam with the polarizing axis of the wire - grid polarizer in the polarization state preparation unit .
the polarization state rotator assembly is composed of a stepper motor ( st@xmath10 ) with a 12-tooth polymer gear , which is connected to a 48-tooth polymer gear ( gear ratio : 1:4 ) via a kevlar timing belt ( b@xmath10 ) , rotating a peek cage ( pr ) which contains three opto - mechanic mounts with plane first surface gold mirrors@xcite ( rotation axis parallel to the incoming and outgoing thz - beam ) .
after reflection on a thz - beam steering plane mirror assembly ( m@xmath144 ) ( opto - mechanic mount and plane first surface gold mirror@xcite ) , the thz - beam leaves the thz source unit .
+ the polarization state preparation unit contains four assemblies : a fir - beam steering plane mirror assembly ( m@xmath152 ) , a thz - beam steering plane mirror assembly ( m@xmath153 ) , a source selection and beam focusing assembly ( r@xmath143 ) , and a rotation stage assembly ( p@xmath142-b@xmath144-st@xmath144 ) .
the fir- and thz - beam steering plane mirror assemblies ( m@xmath154 ) are identical and both are composed of an opto - mechanic mount and a plane first surface gold mirror.@xcite the plane first surface gold mirrors redirect , depending on the spectral range the fir / thz ellipsometer sub - system is operated in , the fir- or thz - beam to the source selection and beam focusing assembly . the source selection and
beam focusing assembly ( r@xmath143 ) comprises two sub - assemblies , the rotatable plane mirror sub - assembly and the beam focusing off - axis paraboloid stage sub - assembly , which are equivalent to those in the source selection and beam focusing assembly in the polarization state preparation unit of the mir ellipsometer sub - system ( fig .
[ fig : beamlift ] a ) .
depending on the orientation of the plane first surface gold mirror@xcite in the rotatable plane mirror sub - assembly , either the fir- or thz beam is directed to beam focusing off - axis paraboloid stage sub - assembly ( gold surface @xmath135 off - axis paraboloid , @xmath136 mm ) .
the focused beam is then routed through the rotation stage assembly , which contains two polyethylene substrate based wire - grid polarizers ( p@xmath142 ) , but is otherwise identical to the rotation stage assembly in the polarization state preparation unit of the mir ellipsometer sub - system ( fig .
[ fig : beamlift ] b ) .
the beam then leaves the polarization state preparation unit .
the beam is then reflected by , or transmitted through the sample ( s@xmath142 ) . the sample can be mounted on a sample holder , attached to the fir / thz goniometer unit ( g@xmath142 ) ( commercially available , 2-circle goniometer 415 , huber diffraktionstechnik ) , or inside the magneto - cryostat sub - system ( m2 ) .
if the magneto - cryostat sub - system is used , reflection type measurements can only be conducted at @xmath139 angle of incidence .
a detailed description of the magneto - cryostat sub - system , its sample mount , and the optical window configuration is given in section [ sec : magnet ] .
the beam then enters the polarization state detection unit , which comprises a rotation stage assembly ( a@xmath142-b@xmath153-st@xmath153 ) , a beam collimation and detector selection assembly ( r@xmath144 ) , and three beam focusing / detection assemblies ( m@xmath155-o@xmath153-g , o@xmath152-dtgs@xmath142 and o@xmath155-b@xmath142 ) .
the beam is routed through the rotation stage assembly for the analyzer of the fir / thz ellipsometer sub - system , which contains two polyethylene substrate based wire - grid polarizers ( a@xmath142 ) , but is otherwise identical to the rotation stage assembly in the polarization state detection unit of the mir ellipsometer sub - system ( fig .
[ fig : beamlift ] b ) .
the rotation stage assembly for the analyzer of the fir / thz ellipsometer sub - system can be operated in step mode for fir measurements@xcite or in continuous rotation mode for thz measurements .
the beam is then collimated ( gold surface @xmath135 off - axis paraboloid , @xmath136 mm ) and redirected to the selected detector by the beam collimation and detector selection assembly ( r@xmath144 ) , identical to the beam collimation and detector selection assembly in the polarization state detection unit of the mir ellipsometer sub - system ( fig .
[ fig : beamlift ] a ) .
the beam focusing and golay - cell - detector assembly ( m@xmath155-o@xmath153-g ) contains a beam steering plane mirror sub - assembly ( m@xmath155 ) ( opto - mechanic mount , plane first surface gold mirror@xcite ) , a beam focusing off - axis paraboloid stage sub - assembly ( o@xmath153 ) ( opto - mechanic mount , gold surface @xmath135 off - axis paraboloid , @xmath156 mm ) , and a golay - cell detector sub - assembly ( g ) ( opto - mechanic mount , golay - cell detector ) .
the beam focusing and dtgs detector assembly ( o@xmath152-dtgs@xmath142 ) comprises a beam focusing off - axis paraboloid stage sub - assembly ( o@xmath152 ) ( opto - mechanic mount , gold surface @xmath135 off - axis paraboloid , @xmath157 mm ) , and a solid state deuterated triglycine sulfate detector sub - assembly ( dtgs@xmath142 ) ( opto - mechanic mount , bruker vertex v-70 dtgs detector ) .
alternatively , the beam focusing and bolometer detector assembly ( o@xmath155-b@xmath142 ) , composed of a beam focusing off - axis paraboloid stage sub - assembly ( o@xmath155 ) ( opto - mechanic mount , gold surface @xmath135 off - axis paraboloid , @xmath158 mm ) and the bolometer detector sub - assembly ( b@xmath142 ) ( opto - mechanic mount ; commercially available , liquid helium cooled bolometer detector , infrared laboratories inc . ) , can be used . for thz measurements the bolometer or the golay cell detector can be chosen , while for fir measurements only the bolometer or the dtgs detector provide frequency responses fast enough to record interferograms .
the central piece of equipment of the integrated mir , fir and thz ohe instrument is the commercially available , superconducting , closed cycle magneto - cryostat sub - system ( 7t - spectromagpt , oxford instruments ) with four optical ports ( fig .
[ fig : cad ] ) .
the design of the integrated mir , fir and thz ohe instrument allows the usage of the magneto - cryostat sub - system with the mir and the fir / thz ellipsometer sub - system by employing the magneto - cryostat transfer sub - system .
the magneto - cryostat sub - system can be subdivided into the magnet head , the primary cooling cycle , the secondary cooling cycle , the sample holder and the optical windows .
the magnet head contains two magnet coils , which are mounted around the sample position and are fabricated in a split - coil pair design . in a spherical volume of 10 mm diameter around the sample , magnetic fields up to @xmath159 t with an inhomogeneity of less than 0.3 % can be achieved .
the magnetic field can be reversed and points towards one of the optical windows . therefore , for reflection type ohe measurements , the magnetic field lies within the plane of incidence and forms an angle of @xmath160 with the sample normal .
this leads to a magnetic field component @xmath161 perpendicular to the sample surface .
the primary cooling cycle comprises a pulse tube cooler ( srp-082 , shi cryogenics ) , high pressure helium lines and a helium compressor ( f-70 , sumitomo heavy industries ) .
ultra - high - purity helium ( uhp - he ) gas at high pressure is provided by the helium compressor , and guided by a high pressure helium line to the pulse tube cooler .
the pulse tube cooler is thermally coupled to the superconducting magnet coils and allows to cool the magnet coils to temperatures of @xmath162 k. the pulse tube cooler also pre - cools the uhp - he in the secondary cooling cycle .
the uhp - he gas is then guided back to the helium compressor by a high pressure helium line , and is reused .
the secondary cooling cycle uses uhp - he gas to cool the sample .
the uhp - he gas in the closed cycle , is circulated by an oil - free , dry scroll pump ( xds-10 , edwards ) .
when the sample is at base temperature , the uhp - he gas leaves the outlet of the scroll pump at a pressure of @xmath145 0.5 bar , and is pumped through a zeolite and a liquid nitrogen trap , in order to extract possible contaminants leaking into the closed cycle .
the uhp - he gas then passes a distiller spiral , which is thermally coupled to the pulse tube cooler ( primary cooling cycle ) , and condenses the uhp - he gas into liquid helium ( lhe ) .
the lhe flow is controlled by a needle valve , and the lhe is then injected into a heat exchanger .
the heat exchanger is attached to the double - walled , hollow variable temperature inset ( vti ) cryostat ( fig .
[ fig : sampleholder ] ) .
finally , the scroll pump reduces the gas pressure above the lhe to 1.5 - 3 mbar and thereby cools the vti to a minimal temperature of @xmath13 k. the sample is thermally coupled to the vti by a static exchange gas ( uhp - he ) .
a resistive heater allows to warm the sample up to room temperature , without bringing the temperature of the superconducting magnet coils above the critical temperature . the sample holder of the magneto - cryostat sub - system ( fig .
[ fig : sampleholder ] ) can hold up to two samples at a time . if two samples are mounted the optimal sample size is , while the maximum sample size is ( only one sample can be mounted ) .
the sample position is adjustable in the vertical direction ( @xmath163 mm ) and rotationally around the vertical axis of the vti by 360@xmath137 ( angle of incidence alignment ) .
all other degrees of freedom necessary for sample alignment ( linear motion , rotation , tip / tilt ) can only be accessed by moving the sample together with the magneto - cryostat sub - system .
the magneto - cryostat sub - system can be moved parallel to the incoming beam , using the magneto - cryostat transfer sub - system .
the alignment perpendicular to the incoming beam and the rotational alignment of the magneto - cryostat sub - system is achieved by sliding the magneto - cryostat sub - system in the magneto - cryostat holding frame of the magneto - cryostat transfer sub - system ( see sec . [
sec : magnet_transfer ] ) .
four screws in the same magneto - cryostat holding frame of the magneto - cryostat transfer sub - system are used for tip / tilt alignment .
after successful alignment , in order to minimize motion , the magneto - cryostat sub - system is clamped to the magneto - cryostat holding frame of the magneto - cryostat transfer sub - system .
when the light beam is routed through the magnet head , it passes a set of exterior and interior optical windows . for measurements in the fir / thz spectral range ,
the exterior optical windows are made of 0.27 mm thick homo - polypropylene films , while for the mir spectral range potassium bromide ( kbr ) windows are used .
all four exterior windows are purged on the exterior side with dried air , to prevent condensation of moisture from the ambient air .
the exterior optical windows of the magnet can be exchanged and arranged for both transmission- and reflection - type measurements .
the latter window configuration allows reflection - type measurements over the full spectral range of the ohe instrument , by simply moving the magneto - cryostat between the ellipsometer sub - systems without warming up the superconducting coils or the sample .
the interior four optical windows on the vti are made of polished diamond , grown by chemical vapor deposition ( cvd ) , with a thickness of 0.35 mm , a diameter of 14 mm and an average surface roughness of @xmath164 nm ( arithmetic average ) .
to reduce fabry - prot interferences in the interior optical windows the interior windows were wedged by an angle of @xmath1450.5@xmath137 .
the interior windows were glued into stainless steel window frames by a cryogenic two component epoxy , leaving a clear aperture of @xmath14512 mm ( see cutout in fig . [
fig : sampleholder ] ) . in general ,
optical windows affect experimental ellipsometry data , especially if the optical window material is birefringent
. therefore all optical windows were characterized by transmission ge on a commercial mir ellipsometer ( j.a.woollam co. , inc . ) .
no birefringence was observed in the mir spectral range . nevertheless , through mounting of the optical windows , and in particular through stress due to the vacuum in the magneto - cryostat sub - system , strain induced birefringence can not be completely excluded and therefore window effects are included in the analysis of the ohe data ( see sec .
[ sec : ohe - analysis ] ) . the magneto - cryostat transfer sub - system ( fig .
[ fig : cad - overview ] ) contains the magneto - cryostat transfer frame , the mir source unit frame and the fir / thz source unit frame . the magneto - cryostat transfer frame comprises the magneto - cryostat transfer assembly , the mir goniometer unit platform assembly and fir / thz goniometer unit platform assembly . the magneto - cryostat transfer assembly is built from alumina extrusions .
the grooves of the alumina extrusions are used as guides for a teflon roll based rail sub - assembly . on top of the teflon roll based rail sub - assembly the magneto - cryostat holding frame ( black alumina extrusions in fig .
[ fig : cad - overview ] and [ fig : cad ] ) is mounted , which is used for alignment purposes ( see sec .
[ sec : magnet ] ) .
the mir and fir / thz goniometer unit platform assemblies are mounted into the magneto - cryostat transfer frame and are platforms for the mir and fir / thz goniometer unit of the ellipsometer sub - systems .
the mir and fir / thz source unit frames are built from alumina extrusions , equipped with a 19-inch rack mounting system , and provide platforms for the source units of the mir and fir / thz ellipsometer sub - systems , respectively .
all spectroscopic ellipsometer sub - systems of the integrated mir , fir and thz ohe instrument are operated in a @xmath113 configuration , capable to measure the upper left @xmath1 block of the mueller matrix .
note that this does not affect the ability to determine certain sample properties related to anisotropy , which can be obtained from the off - diagonal - block elements @xmath165 , @xmath166 , @xmath167 and @xmath168 . for rotating - analyzer based ellipsometers with lossless and ideal polarizing optical elements , the stokes vector of a beam of light at the detector @xmath169 can be described within the mueller matrix formalism by@xcite @xmath170 where @xmath171 is the stokes vector of the light leaving the source , and @xmath119 , @xmath120 , @xmath172 and @xmath97 are the mueller matrices of a polarizer , analyzer , coordinate rotation by the angle @xmath173 ( polarizer : @xmath174 ; analyzer : @xmath175 ) and sample , respectively . if the analyzer is rotated with a constant angular frequency @xmath176 , and both , light source and detector exhibit no polarization dependency , _
e.g. _ the source emits unpolarized light with intensity @xmath177 and the detector is only sensitive to the total intensity @xmath178 , and @xmath179 the ratio of these quantities is @xmath180\ ; , \end{split}\ ] ] with the time harmonic fourier coefficients @xmath181 individual mueller matrix elements @xmath182 are determined from fourier coefficients measured at different polarizer orientations @xmath183 on the input side ( see subroutine mueller matrix regression in fig .
[ fig : flow_aqu ] ) .
the complete sequence of operations executed by the ohe instrument during the ge and ohe data acquisition is summarized in fig .
[ fig : flow_aqu ] .
since the intensity of light at the detector and the fourier coefficients ( employed to determine the mueller matrix elements ) depend on the absolute angular positions of all rotating optical elements [ eq . ( [ eqn : fouriercoef ] ) and ( [ eqn : ncsparameter ] ) ] , the knowledge of these absolute angular positions is crucial for the operation of every ellipsometer .
furthermore , all spectroscopic ellipsometers have to account for non - ideal polarization characteristics of sources and detectors , as well as commonly present non - idealities of mir and fir optical elements ( such as polarizers ) , in their calibration routine .
the absolute positions of the optical components used in the ohe instrument , with respect to the _ p_- and _ s_-coordinate system ( the _ s_-axis is parallel to the goniometer axis , the _ p_-axis is perpendicular to the beam ) , as well as the non - idealities of the optical elements are carefully calibrated prior to the ohe data acquisition .
all ellipsometers sub - systems are calibrated prior to the experiments . during ohe data analysis , additional non - idealities ,
not included in the ellipsometer calibration , introduced by the magneto - cryostat sub - system have to be considered .
in particular , three effects are considered : ( i ) in- and - out - of - plane anisotropy in the optical window@xcite , ( ii ) changes in the alignment of mirrors and/or off - axis paraboloids in the polarization state preparation and detection units due to magnetic forces on ferromagnetic components ( ellipsometric coordinate system change ) , and ( iii ) imperfect sample alignment ( ellipsometric coordinate system and angle - of - incidence change ) . in order to account for sample misalignment in the model based data analysis described in section [ sec : data analysis ] , the angle of incidence and the sample tilt angle are included as model parameters . in order to model
the combined non - idealities due to strain induced birefringence in the optical windows and minute changes in the alignment of optical elements on the input ( output ) side , an additional mueller matrix @xmath184 ( @xmath185 ) was included in the model based data analysis ( sec .
[ sec : data analysis ] ) .
the mueller matrices @xmath184 and @xmath185 are assumed to have no dispersion ( wavelength independent ) .
the best - model mueller matrix @xmath186 , used for mse regression as described in section [ sec : data analysis ] reads @xmath187 where @xmath188 represents the sample mueller matrix calculated by the @xmath2 matrix formalism . of the in- and output matrices @xmath189 and @xmath190 is smaller then 0.02 over the full magnetic field range . ] for the analysis of ohe data , two strategies may be used exploiting either model - free or model - based approaches .
model free analysis provides semi - quantitative results , by studying trends in amplitudes or spectral positions of features in ohe or derived ohe data vs. the magnitude of the magnetic field @xmath191
. the model free analysis can provide insight into the symmetry properties of magneto - optic dielectric tensors,@xcite and an example for the model free analysis of derived ohe data @xmath192 vs. @xmath191 is given in section [ sec : graphene ] for the case of epitaxial graphene .
the model - based data analysis approach provides more quantitative parameters than the model - free data analysis approach , and can be used to determine model parameters such as the free charge carrier concentration or the effective mass parameters . during the data analysis the ge , ohe and derived ohe datasets are analyzed simultaneously to determine physical model parameters of @xmath193 and @xmath194 by a single , consistent optical model ( fig .
[ fig : flow ] ) .
* in the first step , @xmath110 ( ge data obtained at @xmath68 t ) is analyzed only . during the analysis , all model parameters _
independent _ of the magnetic field , such as those describing , for example , the polar lattice resonances or layer thickness are varied until the calculated data match the measured data as closely as possible . * in the second step , ohe data @xmath195 for @xmath196 t is included in the analysis , and all model parameters _ dependent _ on the magnetic field , such as , for example , effective mass or inter - landau - level transition parameters are varied .
in addition , derived ohe data @xmath108 and/or @xmath197 can be included into the simultaneously analyzed dataset .
these derived ohe datasets do not increase the amount of information collected during the experiment , but help ( _ a _ ) to visualize magnetic field induced changes , and ( _ b _ ) improve the sensitivity to magnetic - field dependent model parameters during ohe data analysis .
t , as for example in case of inter - landau - level transitions in epitaxial graphene , discussed in section [ sec : graphene ] . ]
the parameters determined in the first analysis step are used as starting values for the second analysis step and are varied if necessary . note that an optical model , describing the ge dataset ( @xmath68 t ) sufficiently well , might not be capable to describe the complete dataset including ohe and/or derived ohe data correctly . if necessary the optical model has to be adjusted , and the data analysis procedure has to be repeated .
additional information on the data analysis strategies can be found in ref . .
in this section , experimental data from the integrated mir , fir and thz ohe instrument is presented .
first , a combined dataset of derived ohe data from the mir , fir and thz spectral range from 7.5 @xmath0 to 4000@xmath198 ( 0.22120 thz or 0.9500 mev ) , and a magnetic field of @xmath1994 t , for an epitaxial graphene sample grown on 6__h__-sic is shown ( fig .
[ fig : full_range ] ) . in the spectral range below approximately 500 - 600 @xmath0
the experimental data reveals an ohe signal enhanced by fabry - prot interferences in the 6__h__-sic .
an ohe signal , enhanced by coupling with the 6__h__-sic phonon mode , is observed near 1000 @xmath0 .
furthermore , between 500 and 4000 @xmath0 inter - landau - level transitions in single layer graphene are observed in the on - diagonal block elements of the mueller matrix
. then derived ohe data from the individual mir , fir and thz spectral ranges of the integrated mir , fir and thz ohe instrument , and the corresponding best - model calculated data are shown exemplarily .
we present results from ohe experiments on an epitaxial graphene sample grown on 6__h__-sic , a te doped n - type gaas substrate and an algan / gan high electron mobility transistor structure ( hemt ) , representing the mir , fir , and thz spectral range of the integrated mir , fir and thz ohe instrument , respectively .
the selected experimental datasets demonstrate the full spectral , magnetic field and temperature range of the integrated mir , fir and thz ohe instrument , as well as analysis strategies . effects from free charge carriers in bulk and in two dimensional confinement as well as quantum mechanical effects ( inter - landau - level transitions ) are observed and discussed .
exemplarily , epitaxial graphene on 6__h__-sic was investigated to demonstrate the mir ellipsometer sub - system of the ohe instrument ( sec .
[ sec : mir ] ) .
the epitaxial graphene sample was grown in an argon atmosphere at 1400 @xmath96c , by sublimating si from the polar _
c_-face ( 000@xmath200 ) of a semi - insulating 6__h__-sic substrate . from previous measurements on c - face 4__h__-sic
@xcite , the number of graphene layers is estimated to be 10 - 20 .
further details on growth conditions are beyond the scope of this manuscript , and can be found in ref . .
the ohe experiment was conducted at an angle of incidence of @xmath201 = 45@xmath96 , while the magnetic field was aligned along the reflected beam .
experimental data was recorded in the spectral range from 600 to 4000 @xmath0 with a spectral resolution of 1 @xmath0 , using the hgcdte detector , while the sample was kept at a temperature of @xmath15 k. figure [ fig : graphene1 ] shows derived ohe data @xmath192 ( green , dotted line ) for @xmath159 t
( effective field parallel to sample normal @xmath202 t ) and results from best model calculations ( red , solid line ) .
the model free analysis approach provides valuable information .
the ohe data shows several resonances , which can be divided in three groups : a peak near @xmath203 @xmath0 labeled fcc ( triangles , pink ) , a set of features labeled ll@xmath204 ( rectangles , blue ) and a set of peaks labeled ll@xmath205 ( diamonds , orange ) .
while the resonances labeled fcc and ll@xmath204 are present in all mueller matrix elements , the features labeled ll@xmath205 are only present in the on - diagonal - block elements of the mueller matrix
. since non - vanishing off - diagonal - block elements in the mueller matrix are inherently tied to non - vanishing off - diagonal elements in the underlying dielectric tensor , the dielectric tensor has to be diagonal @xmath206 .
furthermore , the magneto - optic contributions of ll@xmath205 to the permittivity tensor in its representation for circularly polarized light [ eq .
( [ eq : llteps ] ) ] must satisfy @xmath33 .
in other words , the physical processes leading to the fingerprints labeled ll@xmath205 are polarization conserving - while the processes leading to fcc and ll@xmath204 are polarization mixing ( polarization selection rules ) . the origin of the individual processes can be determined by field - dependent measurements . figure [ fig : graphene2 ] ( c ) displays the magnetic field dependence of a representative on - diagonal - block mueller matrix element @xmath207 of the derived ohe dataset for @xmath208 in @xmath209 increments (
the corresponding representative off - diagonal - block mueller matrix element is omitted here and the interested reader is referred to ref . ) . at a first glance it can be noted that the resonance labeled fcc increases in amplitude while its spectral position is not affected .
this indicates that the resonance labeled fcc is the sic phonon mode ( best model parameter : @xmath210 , , , , , , , , , , and ) enhanced , magneto - optic fingerprint from free charge carriers .
in contrast , the spectral positions of the resonances labeled ll@xmath205 shift to higher energies and their amplitudes increase with increasing magnetic field strength .
a detailed analysis of the peaks reveals a @xmath211-dependence of their spectral positions , indicative for inter - landau - level transitions in the dirac - type band structure of single - layer graphene with the fermi level close to the charge neutrality point @xcite .
landau level energies in single layered graphene are given by @xmath212 with @xmath213 , where @xmath214 is the average velocity of dirac fermions in graphene .
optical selection rules for transitions between levels @xmath215 and @xmath46 demand @xmath216 .
the fermi velocity is determined as @xmath217 m/s , in very good agreement with refs . .
the analysis of the magnetic field dependence of the resonances labeled ll@xmath204 reveals their origin from the bi - layer branch of inter - landau - level transitions in aba - stacked ( bernal ) multi - layered graphene .
inter - landau - level transitions from bi - layer and tri - layer graphene were identified . for further details the interested reader
is referred to ref . .
the _ n_-type gaas substrate investigated here , was moderately doped with tellurium ( @xmath218 @xmath219 ) and is opaque to fir radiation .
all measurements were conducted at an angle of incidence of @xmath139 , while the magnetic field was parallel to the reflected beam .
figure [ fig : gaas_8 t ] shows experimental data from the fir spectral range ( 50650 @xmath0 ) obtained at a sample temperature of @xmath220 k. the sub - figures ( a ) and ( b ) display ge data ( @xmath68 t ) and derived ohe data @xmath221 for @xmath159 t , respectively . without magnetic field , gaas is optically isotropic , i.e. , for @xmath68 t all off - diagonal - block elements ( @xmath165 , @xmath166 , @xmath167 , and @xmath168 ) in ( a ) are zero .
in contrast , due to the underlying magneto - optic dielectric tensor symmetry , all on - diagonal - block elements in ( b ) are zero , indicating the anti - symmetry of the off - diagonal elements of the dielectric tensor , as well as their sign change under field inversion .
note that for a bulk material and a field orientation that is not perpendicular to the sample surface , all off - diagonal elements of the dielectric tensor are different from zero .
this also leads to the asymmetry between corresponding mueller matrix elements with interchanged indices ( @xmath222 and @xmath223 ) .
the optical model for the gaas substrate consists of a single , semi - infinite layer which contains two contributions , the optical response of polar lattice vibrations and free charge carriers described by the drude model . due to the crystal symmetry of gaas ,
the dielectric tensor for the lattice vibration [ eq . ( [ eqn : phonon2 ] ) and ( [ substrate ] ) ] has identical diagonal elements , _
e.g. _ @xmath224 and possesses only one to - lo resonance in the fir spectral range .
best model parameter for the lattice vibrations were determined as @xmath225 , @xmath226 @xmath0 , @xmath227 @xmath0 and @xmath228 @xmath0 .
the model parameters for the free - charge - carrier contribution @xmath229 @xmath219 , @xmath230 @xmath231/vs and @xmath232 m@xmath233 are in good agreement with results in previous publications.@xcite the hemt structure investigated here , was grown on a semi - insulating 4__h__-sic substrate by metal - organic chemical vapor deposition at temperatures of 1050 @xmath137c.@xcite first , a nominally 100 nm thick aln nucleation layer was grown , followed by a 1800 nm thick gan buffer and a 20 nm thick al@xmath234ga@xmath235n electron barrier layer .
figure [ fig : hemt ] displays the temperature dependence of non - vanishing off - diagonal - block mueller matrix elements of derived ohe data @xmath236 and @xmath237 for @xmath238 t in the spectral range from 0.22 to 0.32 thz . the experimental ohe data is depicted as dotted lines ( green ) , while best - model calculated ohe data is plotted as solid lines ( red ) for different temperatures between 1.5 k and 300 k. the observed ohe is caused by free charge carriers in the high mobility channel of the hemt structure , which is enhanced by the fabry - prot interference in the @xmath239 @xmath240 m thick sic substrate .
the optical model is composed of a sic substrate layer , a aln nucleation layer , a gan buffer layer , a layer for the gan hemt channel and a algan layer .
the thickness of the gan hemt layer was set to @xmath241 nm and not varied during data analysis .
the hemt layer thickness was used to calculate the sheet charge density @xmath242 , where @xmath243 is the bulk charge density .
derived thz ohe data was analyzed simultaneously with ge data recorded with a commercial mir ellipsometer.@xcite during model based ohe data analysis the parameters for the sheet charge density @xmath244 , the mobility @xmath240 and the effective mass parameter @xmath245 were determined .
the sheet carrier density was found to be constant within the error limits at a value of @xmath246 @xmath247 for all temperatures .
the mobility and the effective mass parameters were determined as @xmath248 ( @xmath249 ) and @xmath250 @xmath231/vs ( @xmath251 @xmath231/vs ) at @xmath15 k ( @xmath220 k ) , where @xmath252 is the mass of the free electron .
further details are omitted here and the interested reader is referred to an in depth discussion in ref . .
in this article , we have given an overview over theoretical and experimental aspects of the ohe , and have successfully demonstrated the operation of an integrated mir , fir and thz optical hall effect instrument .
two in - house built ellipsometers sub - systems , operating in the rotating - analyzer configuration , were employed to determine the upper @xmath1 block of the normalized mueller matrix in an ultra wide spectral range from 3 @xmath0 to 7000 @xmath0 ( 0.1210 thz or 0.4870 mev ) .
different aspects of the integrated mir , fir and thz optical hall effect instrument , such as the cryogen - free , superconducting 8 tesla magneto - cryostat sub - system , and the optical setup of the integrated mir , fir and thz optical hall effect instrument , as well as the data acquisition and data analysis strategies were discussed in detail . for demonstration purposes of the operation of the integrated mir , fir and thz optical hall effect instrument in the mir , fir and thz spectral range ,
three sample systems were studied : an epitaxial graphene sample ( mir ) , a gaas substrate ( fir ) and a algan / gan hemt structure ( thz ) .
the presented data covered the full magnetic field range ( @xmath253 t ) and temperature range ( @xmath254 k ) of the system .
we have demonstrated that the integrated mir , fir and thz optical hall effect instrument can be used to determine the parameters of free charge carriers in bulk , and in two - dimensional confinement .
furthermore , it was shown that the integrated mir , fir and thz optical hall effect instrument is capable to investigate the symmetry properties of magneto - optic , dielectric tensors of quantized systems , and was able to determine the polarization selection rules of inter - landau - level transitions .
the integrated mir , fir and thz optical hall effect instrument has a wide range of applications , from determining free charge carrier properties contact - free , over the access to free charge carrier properties in buried layers , to magneto - optic quantum phenomena , polarization selection rules , to symmetry properties of magneto - optic dielectric tensors .
this work was supported by the army research office ( d. woolard , contract no .
w911nf-09-c-0097 ) , the national science foundation ( grant nos .
mrsec dmr-0820521 , dmr-0907475 , eps-1004094 ) , with primary support under mri dmr-0922937 , the university of nebraska - lincoln , the j.a .
woollam foundation and the j.a .
wollam company .
the authors thank dr .
gaskill and his workgroup at the u.s .
naval research laboratory , washington dc for providing epitaxial graphene samples .
the authors would like to acknowledge dr .
v. darakchieva and dr .
e. jansen and their workgroups at the university of linkping , sweden for providing the hemt structure samples .
thanks to c. rice , c. briley , s. m. slone and a. boosalis for their help building and testing mechanical and electrical components .
special thanks to the instruments shop at the department of physics at unl , and in particular to l. marquart for his invaluable advice .
102ifxundefined [ 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 * * , ( ) link:\doibase 10.1002/andp.19003060312 [ * * , ( ) ] @noop _ _ ( , , ) @noop _ _ ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) in @noop _ _ , ( , , ) pp . and , eds .
, @noop _ _ ( , , ) @noop _ _ , , vol .
( , , ) @noop * * , ( ) \doibase http://dx.doi.org/10.1063/1.1138206 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1619582 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1805252 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4789495 [ * * , ( ) ] link:\doibase 10.1063/1.2209968 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1063/1.1578162 [ * * , ( ) ] @noop * * , ( ) \doibase http://dx.doi.org/10.1016/j.tsf.2003.11.215 [ * * , ( ) ] link:\doibase 10.1063/1.4772625 [ * * , ( ) ] link:\doibase 10.1002/pssc.200565467 [ * * , ( ) ] link:\doibase 10.1063/1.2168258 [ * * , ( ) ] link:\doibase 10.1063/1.2735669 [ * * , ( ) ] link:\doibase 10.1007/s11664 - 008 - 0385 - 8 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1063/1.3556617 [ * * , ( ) ] \doibase http://dx.doi.org/10.1557/opl.2013.811 [ * * ( ) , http://dx.doi.org/10.1557/opl.2013.811 ] \doibase http://dx.doi.org/10.1063/1.4833195 [ * * , ( ) ] link:\doibase 10.1063/1.4765351 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.111.077402 [ * * , ( ) ] link:\doibase 10.1103/physrevb.61.8187 [ * * , ( ) ] link:\doibase 10.1103/physrevb.62.7365 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.20.1292 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.155409 [ * * , ( ) ] http://0-search.ebscohost.com.library.unl.edu/login.aspx?direct=true&db=inh&an=10096425&site=ehost-live&scope=site [ * * , ( ) ] http://0-search.ebscohost.com.library.unl.edu/login.aspx?direct=true&db=inh&an=9600466&site=ehost-live&scope=site [ * * , ( ) ] @noop _ _ ( , , ) @noop _ _ , optical engineering - marcel dekker ( , ) , ed .
, @noop _ _ , vol .
( , , ) @noop _ _ ( , , ) @noop * * , ( ) @noop _ _ ( , , ) in @noop _ _ , ( , , ) @noop _ _ ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevb.66.195204 [ * * , ( ) ] @noop _ _ ( , , ) \doibase http://dx.doi.org/10.1016/0039-6028(80)90297-6 [ * * , ( ) ] link:\doibase 10.1063/1.1141417 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1364/josa.66.000949 [ * * , ( ) ] link:\doibase 10.1364/josa.68.001519 [ * * , ( ) ] @noop * * , ( ) @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop `` , '' ( ) http://www.google.de/patents/download/methods_for_uncorrelated_evaluation_of_p.pdf?id=daieaaaaebaj&output=pdf&sig=acfu3u3it9tbri83bleybiqov23nj6awpa [ `` , '' ] ( ) link:\doibase 10.1364/josaa.5.000773 [ * * , ( ) ] link:\doibase 10.1063/1.4732159 [ * * , ( ) ] link:\doibase 10.1063/1.3442903 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.266405 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.125302 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.166401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.087403 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.216603 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.267601 [ * * , ( ) ] http://stacks.iop.org/0022-3719/12/i=12/a=014 [ * * , ( ) ] link:\doibase 10.1063/1.331665 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.jcrysgro.2009.01.045 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.jcrysgro.2006.10.242 [ * * , ( ) ] | we report on the development of the first integrated mid - infrared , far - infrared and terahertz optical hall effect instrument , covering an ultra wide spectral range from 3 @xmath0 to 7000 @xmath0 ( 0.1210 thz or 0.4870 mev ) .
the instrument comprises four sub - systems , where the magneto - cryostat - transfer sub - system enables the usage of the magneto - cryostat sub - system with the mid - infrared ellipsometer sub - system , and the far - infrared / terahertz ellipsometer sub - system .
both ellipsometer sub - systems can be used as variable angle - of - incidence spectroscopic ellipsometers in reflection or transmission mode , and are equipped with multiple light sources and detectors .
the ellipsometer sub - systems are operated in polarizer - sample - rotating - analyzer configuration granting access to the upper left @xmath1 block of the normalized @xmath2 mueller matrix .
the closed cycle magneto - cryostat sub - system provides sample temperatures between room temperature and 1.4 k and magnetic fields up to 8 t , enabling the detection of transverse and longitudinal magnetic field - induced birefringence .
we discuss theoretical background and practical realization of the integrated mid - infrared , far - infrared and terahertz optical hall effect instrument , as well as acquisition of optical hall effect data and the corresponding model analysis procedures .
exemplarily , epitaxial graphene grown on 6__h__-sic , a tellurium doped bulk gaas sample and an algan / gan high electron mobility transistor structure are investigated .
the selected experimental datasets display the full spectral , magnetic field and temperature range of the instrument and demonstrate data analysis strategies .
effects from free charge carriers in two dimensional confinement and in a volume material , as well as quantum mechanical effects ( inter - landau - level transitions ) are observed and discussed exemplarily . |
the solution of the fundamental dynamical equations is an interesting phenomenon because of its importance in quantum - field theory , molecular physics , solid - state and statistical physics . to obtain the exact @xmath4-state solutions of the schrdinger equation ( se ) to various quantum mechanical problems are possible only for few potentials and hence approximation methods
are used to obtain their solutions . according to the schrdinger formulation of quantum mechanics , a total wave function provides implicitly all relevant information about the behaviour of a physical system .
hence if it is exactly solvable for a given potential , the wave function can describe such a system completely . until now
, many efforts have been made to solve the stationary se with a sextic anharmonic oscillator and perturbed coulomb potentials in one to three dimensions through the hill determinant matrix method [ 1 - 4 ] .
the study of the se with these potentials provides us with insight into the physical problem under consideration .
further , the study of se with some of these potentials in the arbitrary dimensions @xmath0 is presented [ 3 ] .
the purpose of this paper is to carry out the analytical solutions of the @xmath0-dimensional radial se with exactly - solvable coulomb plus linear plus harmonic ( clh ) @xmath5 and sextic anharmonic oscillator ( aho ) @xmath6 potentials through an appropriate ansatz to the wave function . in cases of exactly solvable models ,
the wave function can be expressed as a finite power series polynomial multiplied by an appropriate reference function ( usually , the asymptotic form ) to reproduce the exact solutions .
the analytical solution of the schrdinger equation for the energy levels with a class of confining potentials of type @xmath7 have been studied by datta and mukherjee [ 5 ] by using kato - rellich perturbation theory for linear operations .
it is well known that this confinement potential has been used for calculation of @xmath8 bound - state masses [ 6 ] .
killingbeck [ 7 ] has calculated the energy eigenvalues of the confinement potential by using hypervirial relations .
exact solutions [ 8 ] of potentials of type @xmath7 and @xmath9 are obtained by a number of authors in three - dimensional space when the coupling parameters satisfy certain relations .
chaudhuri and mondal [ 3 ] studied the @xmath0-dimensional sextic aho and clh problems within the framework of supersymmetric quantum mechanics ( susyqm ) and hill determinant method [ 4,9 - 11 ] .
chaudhuri and mondal have shown that susyqm yields exact solutions for a single state only for the quasi - exactly - solvable potentials of type @xmath7 and @xmath9 in @xmath0-dimensions with some constraints on the coupling constants [ 3 ] .
they have also obtained numerical results throughout the hill determinant method .
the ideas of supersymmetric quantum mechanics have been used for the study of atomic systems [ 11 ] , the evaluation of the eigenvalues of a bistable potential [ 12 ] , the improvement of the large - n expansion [ 13 ] , the analysis of all known shape invariant potentials [ 10,14 ] , and the development of a more accurate wkb approximation [ 14 ] .
tymczak et al [ 4 ] devised a highly accurate quantization procedures for the inner product representation both in configuration and momentum space for various aho potentials in one and two dimensions . additionally , dobrovolska and tutik [ 15 ] studied the bound - state problem within the framework of the se through the logarithmic perturbation theory .
recently , they also extended the formalism to the bound - state problem for spherical oscillator of type @xmath9 with its subsequent application to the doubly anharmonic oscillator [ 16 ] .
furthermore , a simple formalism [ 17 ] based on a suitable choice of the wave function ansatz has been proposed for reproducing exact bound - state energy eigenvalues and eigenfunctions for exactly solvable model within the framework of the se .
very recently , this simple approach has also been applied [ 18 ] with remarkable success , to various molecular quantum mechanical problems in @xmath0-dimensions [ 19,20 ] .
the object of this paper is to extend the above simple approach to reproduce bound - state exact solutions for potentials of type @xmath7 and @xmath9 in @xmath0-dimensions with some constraints on the coupling constants . for a certain choice of parameters
the method provides exactly - solvable potentials for the @xmath0-dimensional sextic aho and clh problems .
we then compare our results with the exact ones obtained from the transformation of lho into sextic aho problem .
this paper is organized as follows . in section
[ tdd ] , we solve analytically the @xmath0-dimensional radial schrdinger equation for the sextic aho and clh problems by a suitable choice of a wave function ansatz to each exactly - solvable problem . on the other hand , the exact energy eigenvalues of sextic aho
are obtained by transforming radial wave se in @xmath0-dimensions with angular momentum @xmath4 to another problem in @xmath10-dimensions with angular momentum @xmath11 the results and conclusion will be given in section [ cr ] .
in the @xmath0-dimensional hilbert space , the reduced radial wave schrdinger equation ( with @xmath12 units ) for a spherically symmetric potential @xmath13 takes the form [ 21 ] @xmath14 \psi ( r)=0,\ ] ] where the interaction potential is chosen to be of type @xmath7 or @xmath9 and @xmath1 stands for its eigenvalues .
further , equation ( 1 ) can be simply transformed to the form [ 21 ] @xmath15 } { 4r^{2}}+2\left ( e - v(r)\right ) \right\ } r(r)=0,\ ] ] where @xmath16 the reduced radial wave function , is defined by @xmath17 and @xmath18 which is a parameter depends on a linear combination of the spatial dimensions @xmath0 and the angular momentum quantum number @xmath4 [ 21]@xmath19 we substitute @xmath20 and @xmath21 to transform eq.(1 ) into another schrdinger - type equation in ( @xmath22-dimensional space with angular momentum @xmath23 @xmath24 \chi ( \rho ) = 0,\ ] ] where @xmath25 it is seen that through this transformation , the @xmath0-dimensional radial wave schrdinger equation ( 1 ) with angular momentum @xmath4 can be transformed to a ( @xmath26-dimensional problem with new angular momentum @xmath27 in particular , under this transformation , clh problem of type @xmath7 with eigenvalue @xmath1 can be transformed to a sextic aho problem of type @xmath9 with eigenvalue @xmath28 as @xmath29 with coupling constants given by @xmath30 and eigenvalue @xmath31 we attempt to solve the wave equation ( 2 ) of reduced radial wave @xmath32 in the @xmath0-dimensions for a spherically symmetric potential of confining clh form [ 3,22 ] : @xmath33 in particular , for @xmath34 and @xmath35 such a potential reduces to the well known quarkonium cornell potential ( cf .
[ 21 ] and references therein ) .
apart from its releavance in heavy quarkonium spectroscopy ( cf [ 21 ] and references therein ) , this class of potentials with @xmath34 has important applications in atomic physics . for exactly solvable problems such as clh ,
the representation of the radial portion of wave function ansatz , containing an appropriate reference function ( usually , asymptotic form ) , is @xmath36 \sum_{n=0}a_{n}r^{2n+(k-1)/2},\ ] ] provided that the power of the reference function has the following selection : @xmath37 should fall faster than the asymptotic form of the wave function . upon substituting eq . (
11 ) into eq .
( 2 ) and equating the coefficients of @xmath38 to zero , we readily arrive at the following relation @xmath39 where @xmath40 @xmath41 @xmath42 and the value of the parameters for @xmath43 can be evaluated as @xmath44 to obtain a well - behaved solution at the origin and infinity , it is more convenient to take @xmath45 and @xmath46 which ensure that wave function ansatz representation in ( 11 ) , be finite for all @xmath47 and convergent at @xmath48 further , for a given @xmath49 if @xmath50 but @xmath51 we then obtain @xmath52 from eq .
( 14 ) , i.e. , @xmath53 carrying through a parallel analysis to refs [ 17,18 ] , @xmath54 @xmath55 and @xmath56 must satisfy the determinant relation for a nontrivial solution @xmath57 to utilize this method showing the simplicity of this approach , we pursue determinant analysis to present the exact solution for @xmath58 as follows . case ( 1 ) : when @xmath59 we get from eq .
( 18 ) , the exact bound - state solution of the clh problem in @xmath0-dimensions .
so that the eigenvalue ( ground state ) is given by [ 17,18 ] @xmath60 further , it is shown from eq .
( 19 ) that @xmath61 , which leads to the following constraint on the coupling parameters as @xmath62 which consequently , from eqs .
( 20 ) and ( 21 ) , the perturbed coulomb potential admits an exact ground eigen energy@xmath63 @xmath64 , \ ] ] which is consistent with refs [ 3,22 ] .
further , the corresponding wave function ( unormalized ) : @xmath65 .\ ] ] case ( 2 ) : when @xmath66 the exact energy spectrum becomes @xmath67 , \ ] ] and the corresponding wave function ( unnormalized ) can be readily found as : @xmath68 , \ ] ] where @xmath69 and @xmath70 are normalization constants . the relation between them
can be determined through the relation @xmath71 to be @xmath72 a_{0}.\ ] ] furthermore , we have the following recurrence relation from eq .
( 19 ) , that is , @xmath73 which consequently provides the following constraint on the coupling constants of potential : @xmath74 following this approach , we can further generate a class of exact solutions through setting @xmath75 etc . generally speaking , if @xmath76 @xmath77 from which we can obtain the energy spectra ( cf .
determinant ( 19 ) ) . for the generalization
, one needs to use the shape invariance property and the relation between supersymmetric partners [ 22,23 ] to find the general solution of energy eigenvalues as @xmath78 , \ ] ] with the corresponding wave functions @xmath79 , \ ] ] where @xmath80 @xmath81 are normalization constants . finally , considering the following exactly solvable potentials .
in particular : ( i ) harmonic oscillator : when @xmath82 and @xmath83 are inserted to eq .
( 10 ) , giving @xmath84 and @xmath85 thus , we can readily obtain the energy eigenvalues through using eqs .
( 14)-(16 ) as [ 16,18,24 ] @xmath86 and the corresponding radial wave function becomes @xmath87 .\ ] ] where @xmath80 with @xmath88 are normalization constants .
\(ii ) coulomb problem : when @xmath89 and @xmath90 implies that @xmath91 and @xmath92 from which we readily obtain the exact eigenvalues as [ 25 ] @xmath93 and radial wave function becomes @xmath94 , \text { } \ ] ] where @xmath80 with @xmath88 are normalization constants this version of sextic aho potential @xmath95 has been studied in the @xmath0 dimensions through hill determinant method [ 3 ] .
we want to solve the radial se , eq .
( 2 ) , with eq . ( 34 ) by selecting the following representation of ansatz to the radial portion of wave function @xmath36 \sum_{n=0}a_{n}r^{2n+(k-1)/2},\ ] ] provided that the power of the reference function has the following selection : @xmath96 implementing the present method on the representation ( 35 ) , and taking the coefficients of @xmath97 to zero , we obtain the relation @xmath98 where @xmath99 @xmath100 @xmath101 and the value of the parameters for @xmath43 can be evaluated as @xmath102 to obtain a well - behaved solution at the origin and infinity , we must set @xmath103 and @xmath104 which ensures that wave function ansatz , eq .
( 35 ) , be finite for all @xmath47 .
further , for a given @xmath49 if @xmath76 but @xmath51 we then obtain @xmath52 from eq .
( 38 ) , which leads to the following constraint on the coupling parameters of sextic aho problem as @xmath105 case ( 1 ) : when @xmath59 it is shown from eq .
( 19 ) that @xmath61 , which leads to the following energy eigenvalue ( ground state ) : @xmath106 and the corresponding wave function ( unormalized ) : @xmath107 , \ ] ] case ( 2 ) : when @xmath66 it is shown from eq . ( 19 ) that @xmath108 which leads to the constraints on the coupling parameters of the potential as @xmath109 } .\ ] ] in particular , an important version of the above sextic aho is the harmonic oscillator problem @xmath110 selecting the following representation of the wave function ansatz : : @xmath111 \sum_{n=0}a_{n}r^{2n+(k-1)/2},\ ] ] with @xmath112 and itirating the previous steps , one gets : @xmath113 @xmath100 @xmath101 which leads to choosing @xmath114 case ( 1 ) : when @xmath59 it is shown from eq . ( 19 ) that @xmath61 , which leads to the following energy eigenvalue ( ground state ) : @xmath115 and the corresponding wave function ( unormalized ) : @xmath116 , \ ] ] case ( 2 ) : when @xmath66 it is shown from eq . ( 19 ) that @xmath73 which leads to the restriction on @xmath117 and the parameters of the potential as @xmath118 generally speaking , the energy eigenvalues are @xmath119 and the corresponding wave function can be read @xmath120 , \ ] ] where @xmath80 @xmath121 are normalization constants .
we applied the wavefunction ansatz method to the confining clh and sextic aho interactions .
table 1 shows the calculated energies of the clh type potential together with those obtained by the susyqm and hill determinant method for high values of parameters in three- and four - dimensions .
we know from ( 8) and ( 9 ) that the clh problem with exactly solvable potentials @xmath122 in four - dimensions can be transformed to the sexrtic aho problem ( 7 ) in four - dimensions with the exact eigenvalues given through ( 9 ) .
we compute the energy eigenvalues of the following conjugate aho potentials : @xmath123 r^{4}$ ] @xmath124 r^{6},$ ] @xmath125 r^{4}$ ] @xmath126 r^{6}$ ] and @xmath127 r^{4}$ ] @xmath128 r^{6}$ ] in two - dimensions by the present simple method and compute our results in table 2 with the exact values given by ( 9 ) .
a class of aho may be constructed from ( 43 ) that admits exact solutions .
we also compute the energy eigenvalues of the following conjugate aho potentials : @xmath129 r^{4}$ ] @xmath130 r^{6},$ ] @xmath131 r^{4}$ ] @xmath132 r^{6}$ ] and @xmath133 r^{4}$ ] @xmath134 r^{6}$ ] in four - dimensions by the method and compute our results in table 3 with the exact values given by ( 9 ) .
these eigenvalues are checked by other methods .
this method yields exact solutions for a single state only for a potential of type @xmath9 and many states of type @xmath9 with some constraints on the coupling parameters .
our method is applicable to many general clh or aho and produces excellent results for the low - lying states .
it gives the exact solutions of the couloimb and the harmonic oscillator in @xmath0-dimensions .
the clh and sextic aho in @xmath0-dimensions are related through eq.(9 ) and are verified in table 2 by this method .
a class of conjugate aho having exact eigenvalues may be constructed from the transformation of clh potential .
it is found that the eigenvalues of central potential @xmath13 are identical for @xmath135,@xmath136 @xmath137 and @xmath138 states .
this is because @xmath3 remains unaltered for these states .
s. m. ikhdair and r. sever _ , _ z. phys .
c 56 ( 1992 ) 155 ; c 58 ( 1993 ) 153 ; d 28 ( 1993 ) 1 ; hadronic j. 15 ( 1992 ) 389 ; int .
phys . a 18 ( 2003 ) 4215 ; a 19 ( 2004 ) 1771 ; a 20 ( 2005 ) 4035 ; a 20 ( 2005 ) 6509 ; a 21 ( 2006 ) 2191 ; a 21 ( 2006 ) 3989 ; a 21 ( 2006 ) 6699 ; [ arxiv : hep - ph/0504176 ] ( in press ) int
e ; s. ikhdair , o. mustafa and r. sever , tr .
j. phys . 16
( 1992 ) 510 ; 17 ( 1993 ) 474 .
s. m. ikhdair and r. sever , int .
phys . a 21 ( 2006 ) 6465 ; [ arxiv : quant - ph/0604073 ] : doi 10.1007/s10910 - 006 - 9080 - 2 , to appear in j. math .
[ arxiv : quant - ph/0604078 ] : doi 10.1007/s10910 - 006 - 9115 - 8 , to appear in j. math .
[ arxiv : quant - ph/0508009 ] to appear in j. math .
chem ; [ arxiv : quant - ph/0603205 ] j. mol .
( theochem ) ( 2007 ) ( doi:10.1016/j.theochem.2007.01.019 ) . | utilizing an appropriate ansatz to the wave function , we reproduce the exact bound - state solutions of the radial schrdinger equation to various exactly solvable sextic anharmonic oscillator and confining perturbed coulomb models in @xmath0-dimensions . we show that the perturbed coulomb problem with eigenvalue @xmath1 can be transformed to a sextic anharmonic oscillator problem with eigenvalue @xmath2 we also check the explicit relevance of these two related problems in higher - space dimensions .
it is shown that exact solutions of these potentials exist when their coupling parameters with @xmath3 appearing in the wave equation satisfy certain constraints .
keywords : sextic anharmonic oscillator problem , perturbed coulomb problem , bound states , wave functions , schrdinger equation pacs number : 03.65.-w ; 03.65.fd ; 03.65.ge . |
fullerene molecules are a highly stable form of nanoscopic carbon allotrope that can exist at room temperature .
therefore , they are routinely attractive candidates for spectroscopic studies in understanding aspects of fundamental physics both in their vapor and condensed matter phases . technologically also , fullerenes hold the prospect of exciting applications in solid state quantum computations @xcite , improving the superconducting ability of materials @xcite , biomedical fields @xcite , contrast - enhancement research for magnetic resonance imaging ( mri ) , and improving organic photovoltaic devices @xcite . therefore , investigations of the response of these materials to radiations are valuable . one direction of these studies is to understand the collective response of fullerene electrons to relatively low - energy photons . in an infinite system like graphite , the incoming oscillatory electric field induces plane - wave type plasma oscillations in the electron cloud within the system s translational symmetry .
this can only quantize a surface plasmon quasi - particle , but not the longitudinal ( compressional ) volume plasmon , since light is a transverse wave .
but when the medium has a boundary , the broken translational symmetry enables the plasma wave to reflect and induce other eigen modes of oscillations , including the volume quantization .
in particular , for finite systems with boundaries in all directions , such as fullerenes and metallic nanoclusters , photospectroscopy reveals multiple plasmons that were measured @xcite .
the photoelectron angular distribution asymmetry @xcite and the emission time delay @xcite at the surface plasmon of c@xmath0 also predicted interesting behaviors .
the other direction of fullerene studies involves the response to photons whose energy is higher than the plasmon excitation energies .
these photons with their shorter wavelengths begin to resolve the fullerene molecular geometry , entering the spectral region of photoelectron diffraction .
this effect results into the occurrence of a series of cavity minima observed in the ionization spectra as the integer multiples of the photoelectron half wavelength fit the molecular redii at certain energies @xcite .
the effect also accompanies a beating modulation in the ionization spectra as a signature of c@xmath0 molecular width @xcite .
emission delay spectroscopy predicted structures at these minima @xcite .
since the first observation of c@xmath0 giant plasmon resonance @xcite , theoretical studies with various levels of approximation and success formed a large body of published research , an account of which up until 2008 can be found in ref .
after 2008 , there have been mainly two lines of theoretical calculations that attempted to account for the atomistic details of the fullerene carbon - core on a truncated icosahedral geometry .
one involves the geometric optimization of the c@xmath0 structure by the commercially available dmol3 software followed by the calculation of kohn - sham ground state and then its linear response to the incoming radiation @xcite .
the other uses the general access octopus software to directly solve the time - dependent density functional equations for excited c@xmath0 to subsequently fourier transform the density fluctuation to obtain the dynamical structure factor utilized to derive the electron energy - loss signal @xcite .
however , in spite of these important new developments , the jellium approximation to c@xmath2 ion - core , a model based on which we have developed a linear - response density functional methodology known as the time - dependent local density approximation ( tdlda ) , has seen a significant range of success over last several years and continues to remain relevant @xcite .
this is because of the ease and transparency of this model to capture the primary , robust observable effects and to access the key physics that underpins the photo - dynamics and related spectroscopy .
let us cite two sets of results from our methods that directly connected the experiments : ( i ) our calculations have predicted the photoionization of a second plasmon at a higher energy whose first observation was reported in our joint publications @xcite with the experimental group for gas phase c@xmath0 anions ; a subsequent experiment accessed this new plasmon even for the neutral c@xmath0 @xcite .
( ii ) another experiment - theory joint study of ours revealed oscillations in c@xmath0 valence photoemissions providing trains of diffraction minima mentioned above @xcite .
besides these pivotal results , our jellium - based study also extended to the photoionization of several atomic endofullerene molecules @xcite and the c@xmath0@c@xmath1 buckyonion @xcite .
for some of these fullerene systems , tdlda investigations of the photoemission time - delay @xcite and multitudes of resonant inter - coulombic decay processes @xcite were also carried out with reasonable success .
one limitation of the kohn - sham density functional method is its approximate treatment of the electron exchange .
this is because the exchange interaction can only be fully treated in a non - local theory such as hartree - fock ( hf ) that exactly cancels out all self - interactions , restoring correct @xmath3 behavior at @xmath4 . in most of our previous calculations involving c@xmath0 and its derivative endo - c@xmath0 compounds we used a widely utilized approximate scheme @xcite of exchange - correlation ( xc ) functional augmented by an orbit - by - orbit elimination of self - interaction @xcite originally proposed by perdew and zunger @xcite .
a different scheme is to use the gradient - corrected xc potential of van leeuwen and baerends @xcite that intrinsically approximates the correct long distance properties . while we adopted the latter in some of our most recent works , no detailed study on the comparative abilities between the two schemes
has yet been made .
this is the primary objective of the current work that considers the photoionization of the fullerene molecule .
along with c@xmath0 , a larger spherical fullerene , c@xmath1 , has also been considered to further broaden the scope of the comparison .
significant differences from the choice of the xc treatment , both in ground and photoionization descriptions , are uncovered .
improved agreement of c@xmath0 plasmonic spectrum from the gradient - corrected xc approach with the measured data is found .
this paper is structured as follows .
section ii includes three subsections : a ) the description of jellium core ground state structures with brief accounts of two xc parametrization schemes , b ) comparison of ground state numerical results between two schemes and between two fullerenes , and c ) the essentials of the method that incorporates electron correlations in responding to the radiation ; section iii compares the results of the valence ( subsection a ) and total ( subsection b ) photoemissions , as well as a comparisons with available measurements for c@xmath0 ( subsection c ) .
conclusions are presented in section iv .
the details of the method follow the framework as described in ref .
the jellium potentials , @xmath5 , representing 60 and 240 c@xmath2 ions , respectively for c@xmath0 and c@xmath1 , are constructed by smearing the total positive charge over spherical shells with radius @xmath6 and thickness @xmath7 .
@xmath6 is taken to be the known radius of each molecule : 3.54 for c@xmath0 and 7.14 for c@xmath1 .
a constant pseudopotential @xmath8 is added to the jellium for quantitative accuracy @xcite .
the kohn - sham equations for systems of 240 and 960 electrons , made up of four valence ( @xmath9 ) electrons from each carbon atom , are then solved to obtain the single electron ground state orbitals in the local density approximation ( lda ) .
the parameters @xmath8 and @xmath7 are determined by requiring both charge neutrality and obtaining the experimental value ( for c@xmath0 ) and the known theoretical value ( for c@xmath1 ) of the first ionization thresholds .
the values of @xmath7 and the binding energies of the highest occupied molecular orbital ( homo ) and homo-1 levels of both systems are given in table1 .
using the single - particle density @xmath10 the lda potential can be written as , @xmath11,\ ] ] where the 2nd and 3rd terms on the right are the direct and xc components . in one scheme
, @xmath12 is parametrized directly from @xmath13 by the following formula @xcite : @xmath14 & = & -\left(\frac{3\rho(\mathbf{r})}{\pi}\right)^{1/3 } \nonumber \\ & - & 0.0333\log\left[1 + 11.4\left(\frac{4\pi\rho(\mathbf{r})}{3}\right)^{1/3}\right],\end{aligned}\ ] ] in which the first term on the right is exactly derivable by a variational approach from the hf exchange energy of a uniform electron system with a uniform positively charged background and the second term is the so called correlation potential , a quantity not borne in hf formalism . in addition
, we include an appropriate correction to eliminate unphysical electron self - interactions for the @xmath15-th subshell that renders the lda potential orbital - specific @xcite , @xmath16 - v_{\mbox{\scriptsize xc}}[\rho_i({\bf r})]).\end{aligned}\ ] ] this correction approximately captures the electron s long range properties .
we use the acronym sic to refer this model .
the other alternative account for xc - functional that utilizes eq .
( [ gl ] ) but further refines it by adding a parametrized potential @xcite in terms of the density and its gradient @xmath17 as follows , @xmath18^{1/3}}\frac{x^2}{1 + 3\beta x \sinh^{-1}(x)},\ ] ] where @xmath19 is adjustable and @xmath20/\rho^{4/3}$ ] .
this scheme , termed as lb94 , is known to have lead to a considerable improvement in the asymptotic behavior of the electron when compared to the exact kohn - sham potentials calculated from correlated densities .
consequently , this model is expected to also significantly improve the quality of both the excited and continuum spectra .
we show the ground state radial potentials of c@xmath0 and c@xmath1 obtained via both sic and lb94 in fig.1(b ) where the sic curves , labeled as sic - av , are occupancy - weighted average over all the subshells . this particular shape of the potentials earlier interpreted multiple frequencies in the fourier transform of the measured photoelectron spectra of c@xmath0 @xcite .
yet , note the differences in details from sic to lb94 : ( i ) for c@xmath0 , to retain the exact same configuration of occupied states optimized earlier @xcite based on a number of experimental findings @xcite , the lb94 potential gets slightly narrower ( see table1 ) and deeper but with more widening of the wings on either side of the shell .
( ii ) while these general shapes also hold good for c@xmath1 , we note the following . in the absence of enough experimental information
, c@xmath1 ground states were optimized by requiring identical widths and similar first ionization energies ( table1 ) for both sic and lb94 .
this alters some properties of occupied configuration that includes the lb94 homo to be of @xmath21 character ( a level with no radial node ) @xmath22 with a very high angular momentum @xmath23 as opposed to a @xmath24 level ( with one node ) @xmath25 in sic of a lower @xmath26 [ see fig.1(a ) ] .
the direct repercussion of this change on their photoionization cross sections will be discussed below in section iii a. fig.1(a ) illustrates the general differences , sic versus lb94 , of some valence radial wavefunctions .
( iii ) finally , the potential depth decreases from c@xmath0 to c@xmath1 even though the latter accommodates four times more electrons than the former . why does this happen ? to answer , we need to bear in mind that the effective radial potential also includes the angular momentum dependent centrifugal barrier part @xmath27 which varies slower as a function of @xmath28 over the c@xmath1 shell region that is radially farther from c@xmath0 , creating more energy - room " for larger c@xmath1 .
indeed , a far denser angular momentum manifold of @xmath24 and @xmath21 energy bands are generally found for c@xmath1 as seen for lb94 bands presented in fig.1(b ) .
c|ccccc & @xmath7 ( ) & homo & be@xmath29 ( ev ) & homo-1 & be@xmath30 ( ev ) + c@xmath0 & 1.50 ( 1.30 ) & @xmath31 & @xmath32 & @xmath33 & @xmath34 + c@xmath1 & 1.50 & @xmath25 ( @xmath22 ) & @xmath35 ( @xmath36 ) & @xmath37 ( @xmath25 ) & @xmath38 ( @xmath39 ) + our sic and lb94 descriptions of c@xmath0 also produced static dipole polarizability ( sdp ) values of respectively 92.8 @xmath40 and 114 @xmath40 which are reasonably close to the measured value of 76.5 @xmath41 8 @xmath40 @xcite , particularly given that the jellium model disregards the molecular core vibration
. likewise , our calculated values of sdp for c@xmath1 are 565 @xmath40 and 638 @xmath40 , respectively for sic and lb94 .
the slight increase in sdp from sic to lb94 for both fullerenes is due to a somewhat higher spill - out electron density in lb94 .
this spill - out can be recognized by noticing the lb94 potential in fig.1 being a bit wider at the top causing slight outward spreads of the radial wavefunctions .
a time - dependent lda ( tdlda ) approach @xcite is used to calculate the dynamical response of the compounds to the external dipole field @xmath42 . in this method ,
the photoionization cross section corresponding to a bound - to - continuum dipole transition @xmath43 is @xmath44 where the matrix element @xmath45 , with @xmath46 being the independent - particle lda matrix element ; obviously , @xmath46 solely yields the lda cross section . here
@xmath47 represents the complex induced potential that accounts for electron correlations . in the tdlda , @xmath48 is proportional to the induced frequency - dependent changes in the electron density @xcite .
this change is @xmath49 where the full susceptibility @xmath50 builds the dynamical correlation from the lda susceptibilities , @xmath51 via the matrix equation @xmath52^{-1}$ ] involving the variation of the ground - state potential @xmath53 with respect to the ground - state density @xmath54 .
the radial components of the full green s functions in eq .
( [ suscep ] ) are constructed with the regular ( @xmath55 ) and irregular ( @xmath56 ) solutions of the homogeneous radial equation @xmath57 as @xmath58}\ ] ] where @xmath59 represents the wronskian and @xmath60 .
obviously , tdlda thus includes the dynamical correlation by improving upon the mean - field lda basis .
( a ) and c@xmath1 ( b).,width=302 ] ( a ) and c@xmath1 ( b).,width=302 ]
the photoionization cross sections of the homo level calculated in tdlda , both in lb94 and sic schemes , are presented in fig.2 .
let us first note that the host of narrow spikes that appears represents single - electron autoionizing resonances .
the positions and shapes of these resonances largely vary between two xc schemes this happens mainly because of their significantly different descriptions of the unoccupied excited states ( that depend on the potential s asymptotic behavior ) , even though their occupied spectra are by and large similar . in fact , it is expected that owing to the better long - range accounts of electronic properties , lb94 resonances are more accurate in all current results . neglecting these single - electron features , broad build - ups of the oscillator strength above 10 ev are due to the two collective plasmon resonances .
the general shape of the curves is qualitatively similar between lb94 and sic for c@xmath0 [ fig.2(a ) ] , largely because the homo levels are of the same @xmath24 symmetry in both the schemes [ fig.1(a ) ] .
in contrast , due to the different symmetries of homo for c@xmath1 [ fig.1(a ) ] , the broad shapes of the lb94 and sic curves in fig.2(b ) noticeably differ from each other . in all the curves ,
there appear some imposing oscillation - type structures above 30 ev that somewhat mask the second ( 40-ev ) plasmon .
further , comparing fig.1(a ) with ( b ) , we note a general shift of the plasmonic enhancements toward lower energies for larger c@xmath1 similar to the known trend in the size dependence of plasmons in noble metal cluster studies @xcite .
tdlda cross sections for homo-1 level are shown in fig.3 . since for each fullerene
the homo-1 level retains the same @xmath24 symmetry going from sic to lb94 , the broad shapes of the curves obtained from these approximations compare well , barring the mismatch in details including in the single - electron resonances .
we also note here the superposed oscillatory structures at higher energies and the red - shift of the plasmon resonances in c@xmath1 compared to c@xmath0 as in the case of homo .
cleaner shapes of the plasmon resonances are more readily captured in the total cross sections that we discuss in the next subsection .
we address at this point a photoelectron diffraction - driven phenomenon that begins to surface from the waning region of the higher energy plasmon where the collective effect starts to weaken .
an interference between photoelectron waves , predominantly produced at the boundaries of the fullerene shell , underpins this process .
this essentially single - electron effect is the root cause of the oscillations seen at higher energies in figs.2 and 3 that has been observed before in photoelectron spectroscopy @xcite and theoretically discussed at great lengths @xcite .
following ref .
[ ] , one can simply model these oscillations in a @xmath61-level cross section by @xmath62,\end{aligned}\ ] ] where @xmath63 is a steady energy - dependent part , @xmath64 and @xmath65 are the values of the radial bound wavefunction at inner ( @xmath66 ) and outer ( @xmath67 ) radii of the fullerene shell , @xmath68 , @xmath69 and @xmath70 are respectively proportional to the derivatives of the radial potential [ fig.1(b ) ] at @xmath67 and @xmath66 , and @xmath71 . obviously , the oscillations in photoelectron momentum ( @xmath72 ) depend on the potential shape that also includes the angular momentum dependent centrifugal barrier
. therefore , it is not surprising that the higher energy sub - structures in fig.2(a ) and fig.(3 ) qualitatively match between lb94 and sic which have identical angular momentum symmetry . in fig.2(b ) , however , this matching worsens .
this is the consequence of increased centrifugal barrier from much higher angular momentum of lb94 homo level @xmath22 for c@xmath1 that obliterates the inner radius in the effective potential to effectuate @xmath73 in eq .
( [ cross_osc ] ) , qualitatively altering the net shape of the oscillations .
the details of this angular momentum effect were discussed earlier @xcite .
we must also note in figs.(2 ) and ( 3 ) that these higher energy oscillations are in general smaller for c@xmath1 as a consequence of the larger radius of this system leading to higher oscillation frequencies " in eq .
( [ cross_osc ] ) . equation ( [ cross_osc ] ) unravels some further insights .
note that the first three oscillatory terms in this equation carry a constant phase shift @xmath74 , where the dipole selected final angular momentum @xmath75 .
the implication is that each of these oscillations for ionization from two neighboring @xmath76 states will be 180@xmath77 out - of - phase to each other @xcite .
however , the oscillation from @xmath78 in eq .
( [ cross_osc ] ) is independent of @xmath76 .
but note that between the ionization of a @xmath24 and a @xmath21 electron this oscillation is roughly opposite , since the product @xmath79 is negative for a @xmath24 radial wave , but positive for a @xmath21 an implication of which will be discussed in the following subsection .
( a ) and c@xmath1 ( b ) .
the corresponding single - electron ( lda ) results using lb94 are also shown.,width=302 ] figure 4 presents the total tdlda photoionization cross sections and compares them with respective single - electron lda results ( shown only for lb94 ) for both the fullerenes .
the sum over @xmath76 largely cancels out oscillations due to the reason discussed above ( in the last paragraph of subsection iii a ) and makes the broad higher energy plasmon ( hep ) emerge clearly .
in fact , both the plasmon resonances in tdlda stand out in fig.4 against the relatively smooth lda curves .
unlike to the lower energy plasmon ( lep ) resonances , heps exhibit far weaker effects of single - electron resonances but rather long decay tails .
energy red - shifts of the resonances in c@xmath1 compared to those in c@xmath0 are noted along with the fact that c@xmath1 plasmons are utilizing significantly higher oscillator strength density due to its much larger electron pool to collectivize . for each fullerene , significant differences in the resonance widths between lb94 and sic are also noted .
values of various resonance parameters are given in table ii .
from a non - perturbative many body theory , the emergence of plasmon resonances can be thought of as originating from the formation of collective excited states under the influence of external electromagnetic field @xcite .
since the collective excitations are energetically embedded in the single - electron ionization channels , they provide alternative ionization pathways degenerate with single - electron channels .
thus , the _ autoionization _ of these collective excited states induce resonant enhancements in the subshell cross sections as shown in figs.2 and 3
. however , from a perturbative approach the plasmon mechanism can be best modeled by fano s interchannel coupling formalism @xcite . to include the effects of channel - coupling upon the final state wave function of each of the perturbed dipole matrix elements @xmath80
one can write @xcite , @xmath81 is the unperturbed ( lda ) @xmath61 matrix element , @xmath82 s are the unperturbed final continuum channel wave functions of the single - electron channels , and the sum is over all of the photoionization channels except the @xmath61 channel .
the matrix element within the integral of eq.([mat_element ] ) is known as the interchannel coupling matrix element ; the fact that each of @xmath61 initial state orbital overlaps strongly with all other fullerene - orbitals insures that these interchannel coupling matrix elements will be strong .
further , this also justifies the existence of both low and high energy plasmons at exactly the same energies for all the subshells for a given fullerene and implies the various dipole matrix elements are `` in phase '' over the two energy regions ( bands ) of each fullerene [ fig.1(a ) ] .
consequently , the various terms in the sum in eq.([mat_element ] ) will add up _ coherently _ , leading to the dramatic enhancement .
-band cross sections in sic and lb94 for c@xmath0 ( a ) and c@xmath1 ( b ) .
curves are scaled at higher energies to illustrate a strong minimum ( see text ) .
the total cross sections are also displayed for comparisons.,width=302 ] -band cross sections in sic and lb94 for c@xmath0 ( a ) and c@xmath1 ( b ) .
the total cross sections are also displayed for comparisons.,width=302 ] equation ( [ mat_element ] ) reveals one further important correlation feature . since a @xmath24 ( or a @xmath21 ) bound orbital will have near - perfect overlaps with other @xmath24 ( or @xmath21 ) orbitals due to their almost identical shape and spatial extent , the interchannel coupling matrix element in eq .
( [ mat_element ] ) will be stronger for a @xmath24-@xmath24 or a @xmath21-@xmath21 self - coupling than a @xmath24-@xmath21 cross - coupling .
therefore , it is expected that the @xmath24 electrons will show a preferred participation for building lep and the @xmath21 for hep . figs.5 and 6 respectively show the only-@xmath24 and only-@xmath21 band cross sections in tdlda calculated in both lb94 and sic . for each fullerene , if we compare the @xmath24 band result in lb94 with the total cross section in lb94 ( also shown ) , a dominant contribution of the @xmath24-cloud to lep and of the @xmath21-cloud to hep are indeed noted . in general , however , it is also obvious from these comparisons that either of lep and hep in a fullerene are of both @xmath24 and @xmath21 mixed character , it is just that one is dominant on the other . a discussion on the red - shift of tdlda plasmon resonances from c@xmath0 to c@xmath1 [ fig.4 ] may now be in order .
classical plasmon - model @xcite of a spherical dielectric shell with symmetric and antisymmetric vibrations between the inner and outer surfaces suggests that the midpoint energy between the two resonances to be about the same for c@xmath0 and c@xmath1 , since they have approximately same initial electron densities @xcite .
according to this model the plasmons are then formed below and above this midpoint energy shifted equally both ways , and this shift grows with the increasing radius , suggesting that the plasmons will be more separated out for c@xmath1 @xcite .
clearly , that is not seen in fig.4 , in which both the plasmons red - shift for c@xmath1 and in fact move close to each other compared to their c@xmath0 results ( see table .
ii for the actual values ) , suggesting that quantum effects play an important role .
one possible way to understand this phenomenon quantum mechanically is to recall in fig.1(b ) that the c@xmath1 ground state potential is shallower while accommodating a number of electrons four times that of c@xmath0 producing far compact energy levels .
this suggests a decrease of the _ average _ ground state binding energy for c@xmath1 .
therefore , since in the spirit of eq .
( [ mat_element ] ) the plasmons can be interpreted as the coherence in close - packed single - electron excitations , it is only expected that the plasmons will begin to excite at lower photon energies causing their early onsets for c@xmath1 , as seen in fig.4 .
in fact , this trend of red - shifting plasmons with increasing fullerene size should be rather generic , at least in the jellium based quantum calculations .
an insight in the phenomenon can be motivated by perceiving a collective mode as having a natural oscillation frequency @xmath83 of a mass density ( @xmath54 ) on a spring of stiffness @xmath84 @xcite .
thus , a shallower binding potential with higher electron population for c@xmath1 translates to the loosening of the spring decreasing @xmath84 and thereby its resonant frequencies .
let us now compare between the predictions of lb94 and sic for the band - cross sections .
for the @xmath24-band , lb94 retains a contribution approximately similar to that of sic at lep , but shows depletion at hep which is more prominent for c@xmath0 [ fig.5(a ) ] than c@xmath1 [ fig.5(b ) ] . for the @xmath21-band , on the other hand , a notably higher contribution to lep and some increase at hep by lb94 for both
the systems are found [ fig.6 ] .
there are more .
our discussions following eq .
( [ cross_osc ] ) indicate that the @xmath76-sum over @xmath24 or @xmath21 cross sections will significantly weaken the diffraction oscillations coming from the first three oscillatory terms in eq .
( [ cross_osc ] ) , while the fourth oscillation @xmath85 will survive being free of @xmath76 . as a result , in the band - cross sections this @xmath7-dependent oscillation will dominate .
since @xmath7 slightly shortened in lb94 than sic for c@xmath0 [ fig.1(b ) ] , @xmath24-band lb94 curve in fig.5(a ) produces a longer wavelength in @xmath72 to induce its first minimum above 100 ev at an energy higher than that in sic .
the equality of @xmath7 in lb94 versus sic for c@xmath1 , on the other hand , justifies the occurrence of these minima at about the same energy as in @xmath24-band results for this fullerene [ fig.5(b ) ] .
however , this effect is not so intuitive for the @xmath21-band case . as seen , while the minimum in lb94 for c@xmath0 [ fig.6(a ) ]
does appear at higher energies than the sic minimum , they do not seem to coincide for c@xmath1 [ fig.6(b ) ] as they did for c@xmath0 .
the latter is due to the fact that the @xmath21 states for c@xmath1 , reaching very high in @xmath76 values compared to their counterparts in c@xmath0 , produce such strong centrifugal repulsions that the effective potentials for high @xmath76 considerably deform rendering the role of @xmath7 less meaningful @xcite .
measurements @xcite of plasmon resonances in the photoionization of neutral and ionic c@xmath0 produced relatively smooth curves without any evidence of autoionizing resonances , which exist in our theoretical tdlda results .
as discussed in details in our earlier study @xcite , this is likely because the coupling of electronic motions with the temperature - induced vibration modes of the core @xcite and the fluctuation of the cluster shape around the shape at absolute zero @xcite .
in addition , the inherent over - delocalization of jellium models predicts autoionizing resonances that are too narrow , as seen in our results .
therefore , in fig.7 , we fit the non - spiky background parts of our tdlda total cross sections obtained via lb94 for both the fullerenes using a formula that includes two lorentzian line profiles .
we further present in fig.7 , two similar fitting curves for the sic results of the fullerenes . for both fullerene systems , table 2
presents the positions ( @xmath86 ) and full - widths at half - maxima ( fwhm ) , @xmath87 , and oscillator strength densities ( osd ) corresponding to each plasmon resonances calculated in lb94 and sic ; lb94 results are displayed in parenthesis .
c|ccc & @xmath86 ( ev ) & @xmath87 ( ev ) & osd + c@xmath0 lep & 15.8 ( 16.8 ) & 2.5 ( 3.5 ) & 136 ( 184 ) + c@xmath0 hep & 37.5 ( 38.5 ) & 10.0 ( 13.0 ) & 35 ( 30 ) + c@xmath1 lep & 11.9 ( 12.4 ) & 0.9 ( 2.0 ) & 642 ( 601 ) + c@xmath1 hep & 33.8 ( 31.5 ) & 10.5 ( 9.5 ) & 281 ( 241 ) + from table 2 and fig.7 , going from sic to lb94 , both lep and hep of c@xmath0 move up in energy by 1 ev , while c@xmath1 lep by a half ev .
we recall the spirit of a classical oscillator model of dielectric shell that the plasmon frequencies are proportional to the square - root of the ratio of rigidity to density ( @xmath88 , in analogy to the oscillation frequency of a mass on a spring of stiffness constant @xmath84 ) introduced in the previous subsection .
note , lb94 radial waves being slightly more spread out than their sic counterparts occupy a larger space effectively decreasing the density .
this explains the blue - shift of lb94 plasmons .
this trend in lb94 is an improvement , since jellium based prediction of c@xmath0 plasmon resonance energies are known to be below their measured values @xcite .
however , this trend is reversed for c@xmath1 hep where lb94 moves this plasmon lower in energy by more than 2 ev from its sic prediction , ascertaining the importance of quantum effects to capture the details of these resonances .
furthermore , the lb94 width of c@xmath0 lep is found to be 3.5 ev , which is an increase of 40% over its sic value of 2.5 ev , while this increase is 30% for c@xmath0
hep . more than a double increase of width for c@xmath1 lep is found going from sic to lb94 , while again , this trend reverses by a small amount for c@xmath1 hep .
significant variations in the osd utilized by each plasmons for either system between two xc approximations are also noted in table 2 , accounting for the detailed differences that the two calculation schemes generate .
comparisons of the results between the two fullerenes in fig.7 as well as in table 1 indicate a generic red - shift of plasmon energies for the larger fullerene c@xmath1 , as noted and discussed earlier .
we also find in table 2 a general trend of the width @xmath87 to decrease with the increasing size of fullerene , except for c@xmath1 hep in sic .
further note that while for c@xmath0 lep the osd value increases from sic to lb94 , the trend is found opposite for this resonance of c@xmath1 . for the hep , either fullerene exhibits decrease in osd going from sic to lb94 .
fig.7 further includes two sets of experimental measurements for c@xmath0 , where the data from hertel et al @xcite are red - shifted by 3 ev and those from reinkster et al @xcite by 1 ev to match respectively with the energies of lep and hep calculated in lb94 . as evident ,
the modifications in @xmath87 and osd , as brought about by the lb94 scheme , indicate an improved agreement with experimental results compared to what sic achieves .
we must also note that in a jellium model , the plasmon resonances only decay via the degenerate single - electron channels . in the real system , however , there would be additional effects from the independent local ion sites positioned based on an appropriate atomistic symmetry , at least for relatively more tightly bound electrons .
as shown in detailed with sic results in ref .
[ ] , in order to account for these additional decay channels , the theoretical cross section in a jellium frame must be convoluted with a small width in order for a more meaningful comparison with measurements . with the already improved agreement of the current zero - width " results of lb94 , it is only expected that such a convolution will further better the agreement with the experiment .
in conclusion , the work accounts for various robust similarities but detailed differences between the results obtained via two standard xc schemes , sic and lb94 , in the framework of density functional description of delocalized valence electrons of the fullerene molecule where the ionic core is treated as a jellium shell .
the focus has been applied to understand both the ground state and single - photoionization properties of the system . for the ionization study , the ultraviolet energy range of plasmon activities and above - plasmon soft x - ray range were considered . the comparison between the results of two prototypical spherical fullerenes , c@xmath0 and c@xmath1 , further unravels the scopes of validity of these two theoretical schemes .
a natural next step is to consider the influence of xc formalism on the photospectroscopy of non - spherical fullerenes , which , however , is a topic for our future research . to this end , within the known limitation of the jellium description of the molecular ion - core , the gradient corrected lb94 formalism seems to bring the results closer to the measurements on c@xmath0 over the plasmon resonance energy region .
we hope that with possible future experiments with c@xmath1 the success of lb94 scheme can be verified for larger fullerene systems as well .
t. barillot , c. cauchy , p -a .
hervieux , m. gisselbrecht , s. e. canton , p. johnsson , j. laksman , e. p. mansson , j. m. dahlstrm , m. magrakvelidze6 , g. dixit , m. e. madjet , h. s. chakraborty , e. suraud , p. m. dinh , p. wopperer , k. hansen , v. loriot , c. bordas , s. sorensen and f. lpine , phys .
rev . a * 91 * , 033413 ( 2015 ) .
chakraborty and m. magrakvelidze , in _ from atomic to mesoscale : the role of quantum coherence in systems of various complexities _ , ed .
s. malinovoskaya and i. novikova ( world scientific , singapore , 2015 ) , pp .
221 - 237 .
scully , e.d .
emmons , m.f .
gharaibeh , r.a .
phaneuf , a.l.d .
kilcoyne , a.s .
schlachter , s. schippers , a. mueller , h.s .
chakraborty , m.e .
madjet , and j.m .
rost , phys .
lett . * 98 * , 179602 ( 2007 ) . | we study the photoionization properties of the c@xmath0 versus c@xmath1 molecule in a spherical jellium frame of density functional method .
two different approximations to the exchange - correlation ( xc ) functional are used : ( i ) the gunnerson - lundqvist parametrization [ phys .
rev .
b * 13 * , 4274 ( 1976 ) ] with an explicit correction for the electron self - interaction ( sic ) and ( ii ) a gradient - dependent augmentation of ( i ) by using the van leeuwen and baerends model potential [ phys .
rev .
a * 49 * , 2421 ( 1994 ) ] , in lieu of sic , to implicitly restore electrons asymptotic properties .
ground state results from the two schemes for both molecules show differences in the shapes of mean - field potentials and bound - level properties .
the choice of a xc scheme also significantly alters the dipole single - photoionization cross sections obtained by an _ abinitio _
method that incorporates linear - response dynamical correlations .
differences in the structures and ionization responses between c@xmath0 and c@xmath1 uncover the effect of molecular size on the underlying physics .
analysis indicates that the collective plasmon resonances with the gradient - based xc - option produce results noticeably closer to the experimental data available for c@xmath0 . |
scattering of waves in disordered media has proved to be a nontrivial topic possessing intriguing and still not completely understood features.@xcite accordingly to the strength of disorder , one observes a variety of propagation regimes ranging from ballistic transport , through single scattering and wave diffusion , to the anderson localization . in this paper
we are interested in the regime of wave diffusion , corresponding to a relatively strong disorder , which is , however , still largely insufficient to bring the system to the localization transition ( @xmath0 , where @xmath1 is a wave number in the medium , and @xmath2 is a mean free path ) .
it is well - known , that multiple scattering of coherent wave in a disordered medium results in a complicated spatial intensity distribution @xmath3 known as a `` speckle pattern '' .
the speckle pattern is highly irregular and appears random to the eye .
it is now well established that the speckle pattern exhibits large intensity fluctuations@xcite @xmath4 , originating from the interference of partial waves arriving at @xmath5 with completely randomized phases . here
the angular brackets @xmath6 denote ensemble averaging , and @xmath7 . besides , the speckle pattern possesses nontrivial long - range spatial correlation @xmath8 even for @xmath9 .
this correlation is due to interaction of diffusing modes.@xcite if the points @xmath10 are far enough from the boundaries of the medium , @xmath11 .
recently , it has been shown that in a particular case of a point source of waves embedded inside a disordered medium , there exists an infinite - range contribution to @xmath12 originating from scattering events taking place in the immediate neighborhood of the source.@xcite this contribution is highly sensitive to the short - distance properties of disorder , as well as to the source size and shape.@xcite if the scatterers in the medium are allowed to move , @xmath3 fluctuates with time , and the statistics of these fluctuations is also a subject of active research .
a wave propagating in a disordered , multiple - scattering medium undergoes a large number of scattering events , and hence the scattered intensity is highly sensitive to displacements of scatterers.@xcite consequently , the decay of the intensity autocorrelation function @xmath13 is considerably faster than in the single - scattering case.@xcite recently , long - range autocorrelation function of intensity fluctuations has been measured,@xcite and the existence of the universal conductance fluctuations ( analogous to that in disordered conductors ) has been demonstrated@xcite for optical waves .
theoretical analysis of the temporal correlation function of multiple - scattered waves has been extended to amplifying disordered media,@xcite as well as to the case of intense incident waves producing flows of scatterers in the disordered medium.@xcite an additional contribution to @xmath14 , originating from scattering in the immediate neighborhood of source and/or detector , decaying much slower than all the previously known contributions , is predicted to exist.@xcite high sensitivity of multiple - scattering speckle patterns to scatterer motion gave rise to a new technique for studying the scatterer dynamics in disordered , turbid media , the so - called `` diffusing - wave spectroscopy '' ( dws).@xcite the latter is now widely applied in concentrated colloidal suspensions,@xcite foams,@xcite emulsions,@xcite granular@xcite and biological@xcite media . besides , the dws has been extended to macroscopically heterogeneous turbid media , providing a tool for imaging of dynamic heterogeneities@xcite and visualization of scatterer flows@xcite in the bulk of the medium .
a generalization of dws technique has been also accomplished for anisotropic disordered media.@xcite recently , the dws approach has been extended to nonergodic turbid media.@xcite the above - mentioned , extensive studies of temporal fluctuations of multiple - scattered waves , as well as the numerous application of dws , are all restricted to _ linear _ disordered media . in general , little information is available on the subject of multiple scattering in _ nonlinear _ disordered media .
meanwhile , the question concerning the way in which the nonlinearity affects the multiple - scattering speckle pattern still remains open and continue to attract the researches .
considerable efforts have been made to understand the phenomena of coherent backscattering in disordered media with kerr - type nonlinearity:@xcite a narrow dip has been predicted to appear on the top of the backscattering peak .
weak localization effects are shown to exist in the radiation of second harmonic and difference frequency,@xcite though their experimental observation failed.@xcite also studied with account for disorder is the optical phase conjugation.@xcite more recently , correlations in transmission and reflection coefficients of second harmonic waves have been investigated both theoretically and experimentally,@xcite and the angular correlation functions of fundamental wave in a disordered medium with kerr - type nonlinearity have been calculated.@xcite despite the fact that theoretical description of wave scattering in nonlinear media is complicated by the simultaneous presence of both disorder and nonlinearity , the standard diagram technique for impurity scattering has been extended to the case of disordered medium with nonlinearity of kerr type.@xcite very recently , it has been shown that the speckle pattern resulting from the multiple scattering of coherent wave in a _ nonlinear _ disordered medium with kerr - type nonlinearity , should be extremely sensitive to changes of scattering potential,@xcite i.e. much more sensitive than the _ linear _ speckle pattern . this high sensitivity has been explained by the multiplicity of solutions of nonlinear wave equation.@xcite the multiplicity of solutions has been then shown to lead to the temporal _ instability _ of the multiple - scattering speckle pattern in nonlinear medium , resulting in _ spontaneous _ fluctuations of scattered wave with time.@xcite an important prediction of ref .
is that the nonlinearity should exceed some threshold value for the instability to develop .
the threshold value is principally determined by the absorption length @xmath15 , or by the sample size @xmath16 , if @xmath17 .
the striking feature of the established result is that the threshold value of nonlinearity tends to zero in an unbounded medium without absorption .
purely elastic , unbounded nonlinear multiple - scattering systems are therefore always unstable .
the physical origin of the instability is easy to understand.@xcite nonlinearity modifies the phases of partial waves propagating in the medium .
the phase modifications are proportional to the intensity @xmath3 and affect the mutual interference of partial waves . as it is this interference which is responsible for @xmath3 , a sort of feedback establishes in the medium .
a small modification of @xmath3 causes modifications of phases of partial waves which , in their turn , produce changes of @xmath3 , and so on .
it is well - known that nonlinear wave systems with sufficiently strong , positive feedback may become unstable.@xcite as an example , we cite a family of nonlinear optical systems with two - dimensional feedback,@xcite where spontaneous formation of complicated spatial structures is observed . despite the absence of disorder , such systems exhibit transition to seemingly chaotic dynamics with increasing nonlinearity.@xcite an analogy can be drawn between the nonlinear optical systems with two - dimensional feedback and nonlinear disordered media by considering the scattering as a ( three - dimensional ) feedback mechanism . in the case of disordered media , however , the feedback is of random nature and it is therefore hopeless to expect regular spatial structures to form .
meanwhile , the instability can manifest itself in spontaneous fluctuations of speckle pattern . in order to clarify the issue of instability of speckle patterns in nonlinear disordered media
, we consider the following questions : * can the multiple scattering provide a positive feedback mechanism for waves propagating in a nonlinear disordered medium ? *
if `` yes '' , how strong the nonlinearity should be for the instability to develop ? a general announcement of our principal answers to the above questions has been given in our recent letter.@xcite in the present paper , we discuss and justify the assumptions and approximations underlying our conclusions , provide the missing details of calculations , and give a comprehensive discussion of results . also developed and discussed
is the perturbation approach to the calculation of the temporal autocorrelation function of multiple - scattered wave in a nonlinear disordered medium .
it is important that the validity condition of the perturbation theory coincides with the condition for the instability threshold as obtained by using the self - consistent approach .
in addition , we give a detailed consideration to an experimentally important case of moving scatterers , when the decrease of the time autocorrelation function is due to a combined effect of spontaneous and scatterer - motion - induced fluctuations of the speckle pattern .
the remainder of the paper is arranged as follows . in sec .
[ nweq ] , we introduce the nonlinear wave equation , and discuss how the path integral approach can be applied for its analysis .
we also formulate the basic models and approximations used throughout the paper .
section [ linear ] is devoted to linear disordered media . in this section ,
we provide the expressions for the spatio - temporal intensity correlation functions .
although correlations of multiple - scattered waves in linear media are well studied nowadays , we present their first , to our knowledge , treatment with a simultaneous account for absorption , boundary conditions at the sample surface , and scatterer motion .
the results of sec . [ linear ] serve as a base for further calculations . in sec .
[ dephasing ] , we present a calculation of dephasing of waves in a nonlinear disordered medium .
our calculation takes into account the fluctuations of the local refractive index due to nonlinear effects , as well as the long - range spatial correlation of this fluctuations .
three `` nonlinear '' contributions to the dephasing are identified in addition to the usual , `` linear '' term originating directly from the motion of scatterers .
further , in sec .
[ perturb ] we develop a perturbation theory for calculation of the temporal autocorrelation function of a multiple - scattered wave , and show its failure at short correlation times , for sufficiently weak absorption .
a condition of validity of the perturbation theory is established by comparing the linear and nonlinear contributions to the dephasing found in sec .
[ dephasing ] .
section [ self ] presents an alternative , self - consistent approach to the calculation of the temporal autocorrelation of scattered wave .
development of self - consistent theory requires some additional assumptions which are also discussed in this section . in sec .
[ disc ] , the main results of our self - consistent approach are presented and discussed . the multiple - scattering speckle pattern is shown to exhibit spontaneous fluctuations even in the absence of scatterer motion , which we interpret as a signature of its instability .
a comparison of self - consistent and perturbative results is given , and the condition of the speckle pattern instability is shown to coincide with the condition of validity of the perturbation theory . finally , concluding remarks are presented in sec .
[ concl ] . in order to maintain the text of the paper readable , we have chosen to collect the technical details of calculations in three appendices .
appendix [ appa ] is devoted to the derivation of the field - field spatio - temporal correlation function . in appendix [ appb ]
we compute the spatio - temporal long - range intensity correlation function .
appendix [ appc ] provides the details of calculations of path distributions @xmath18 and @xmath19 defined in sec .
[ dephasing ] .
we consider a scalar monochromatic wave of frequency @xmath20 propagating in a random medium with kerr - type nonlinearity .
the wave amplitude @xmath21 obeys a nonlinear wave equation:@xcite @xmath22 \right\ } \psi({{\bf r } } , t ) \nonumber \\ & & \hspace{4.4 cm } = 0 .
\label{weq}\end{aligned}\ ] ] here @xmath23 is the free - space wave number , @xmath24 is the average ( complex ) dielectric function , @xmath25 is the fluctuating part of the dielectric function , @xmath26 is the nonlinear susceptibility@xcite ( the two latter quantities are assumed to be real ) .
( [ weq ] ) is valid only if @xmath27 does not change significantly on the time scale of @xmath28 .
the expression in the square brackets of eq .
( [ weq ] ) can be considered as some `` effective '' dielectric function of the medium .
general analysis of eq .
( [ weq ] ) for arbitrary relation between various terms comprising this function constitutes a formidable task , and is not a purpose of this paper .
we assume the following hierarchy : @xmath29 in other words , we assume that the role of nonlinearity is less significant than that of disorder , and that absorption is weak allowing multiple scattering of waves in the medium .
it is then convenient to define the effective refractive index @xmath30 , the absorption length @xmath31 , and the nonlinear coefficient @xmath32 , which determines the nonlinear correction to the ( linear ) refractive index of the medium : @xmath33 , where @xmath34 is the wave intensity . in this paper
, we study the fluctuations of the solution of eq .
( [ weq ] ) with time @xmath35 . in a linear medium ( @xmath36 )
, these fluctuations can only be due to random fluctuations of @xmath25 with time .
the fluctuations of @xmath21 are commonly characterized by the autocorrelation function @xmath37 .
we assume that this autocorrelation function is independent of @xmath35 which implies that for fixed @xmath5 , @xmath21 represents a stationary random process ( this is obviously true if the sample geometry and the source distribution do not change with time , and if @xmath25 is a stationary random process ) .
me take @xmath25 to be a gaussian random field with zero mean and the correlation function @xmath38 . for a medium composed of point - like scatterers undergoing brownian motion with a diffusion coefficient @xmath39,@xcite @xmath40 where @xmath41 , and the mean
free path @xmath42 is introduced ( a weak scattering limit @xmath0 is assumed ) . a natural time scale for scattering of waves in the medium described by eq .
( [ cde ] ) is set by the characteristic time needed for a scatterer to move a distance of order of the wavelength : @xmath43 . from here on , we will be interested in short correlation times @xmath44 . in the linear case ( @xmath36 ) ,
several approaches have been elaborated to analyze eq .
( [ weq ] ) .
we mention the diagrammatic techniques,@xcite theory of radiative transfer,@xcite and the method of path integrals.@xcite the three approaches are known to give equivalent results for @xmath45 at @xmath44 . in the present paper , we adopt the method of path integrals that was originally proposed in the framework of quantum electrodynamics,@xcite but later has been successfully used in various areas of physics,@xcite and , in particular , for the analysis of wave scattering problems.@xcite the method is based on the fact that the solution @xmath21 of the wave equation ( [ weq ] ) can be written in a form of a functional integral , with integration performed over all possible trajectories ( paths ) going from the source of waves to @xmath5.@xcite since in the weak scattering limit ( @xmath0 ) different trajectories can be considered independent , it appears that the correlation function @xmath45 reduces to the following integral:@xcite @xmath46 d s , \label{path}\end{aligned}\ ] ] where @xmath47 is the average intensity in a nonabsorbing medium , @xmath48 is a weight coefficient of paths of length @xmath49 , and @xmath50 denotes the squared phase difference @xmath51 , averaged over various realizations of disorder , and over all possible paths of the same length @xmath49 . from here on , we denote such an averaging by @xmath52 .
note that @xmath53 for the model of brownian point - like scatterers .
meanwhile,@xcite @xmath54 where the superscript @xmath55 denotes the linear case .
it is worth noting that @xmath56 does not depend neither on the sample geometry , nor on the source and detector positions .
its value is only determined by the scatterer dynamics ( through the single - scattering correlation time @xmath57 ) , and the path length @xmath49 .
in contrast , @xmath48 can only be calculated if the sample geometry , source distribution , and detector position @xmath5 are specified . in what follows
, we restrict our analysis to a semi - infinite medium occupying the half - space @xmath58 , and illuminated by a plane monochromatic wave incident at @xmath59 . for @xmath60 , @xmath48 becomes@xcite @xmath61 once the field correlation @xmath62
is known , the autocorrelation of intensity fluctuations @xmath63 can be found applying the factorization approximation : @xmath64)([pn ] ) , one obtains the normalized autocorrelation function of multiple - scattered wave in a semi - infinite disordered medium : @xmath65 \frac{z}{\ell } \right\ } , \label{glin}\end{aligned}\ ] ] where the superscript @xmath66 denotes the linear case , @xmath67 , and @xmath68 . for the diffusely
reflected wave , we assume @xmath69 and get @xmath70 from here on , we will use @xmath71 to denote the normalized autocorrelation function at a point @xmath5 _ inside _ the medium , while @xmath72 for the normalized autocorrelation function of _ diffusely reflected _ wave .
the superscripts @xmath66 and @xmath73 will be used to distinguish between linear and nonlinear cases .
now we turn to the nonlinear medium . strictly speaking , the method of path integrals
can not be applied for the analysis of eq .
( [ weq ] ) , once @xmath74 .
the failure of the path - integral technique follows from the fact that this approach relies on the superposition principle which is not valid for waves in nonlinear media .
however , if the nonlinearity is weak [ which is ensured by the first two inequalities of eq . ( [ hierarchy ] ) ] , eq .
( [ path ] ) is still approximately valid provided that its main ingredients @xmath75 and @xmath50 are computed with account for nonlinear effects . to simplify such a calculation
, we assume that the nonlinearity is sufficiently weak to validate the following two assumptions : * propagation of waves in a weakly nonlinear disordered medium is diffusive with a mean free path @xmath2 unaffected by the nonlinearity .
this implies that nonlinear refraction is negligible at distances of order @xmath2 , and consequently , that @xmath76 , where @xmath77 , and @xmath47 is the average intensity in the absence of absorption .
this assumption is an alternative formulation of the fact that the role of nonlinearity is much less significant than that of disorder [ see also the first two inequalities of eq .
( [ hierarchy ] ) ] . *
intensity of the third harmonic remains always much smaller than the intensity of the fundamental wave .
this implies either that @xmath21 is considered as a complex quantity [ in this case , eq .
( [ weq ] ) is a nonlinear schrdinger equation , @xmath78 is time - independent , and the third harmonic is not generated at all ] , or that the medium has a sufficient degree of dispersion for the phase matching condition@xcite to be violated : @xmath79 with @xmath80 and @xmath81 being the wave number at frequency @xmath82 .
assumption ( _ i _ ) allows us to consider the path distribution @xmath75 being unaffected by nonlinearity .
the only object to be recalculated with account for nonlinear effects is then @xmath50 . before going into an explicit calculation of @xmath50
, we devote the next section to a brief derivation of some important results for linear medium .
as indicated above , we consider a monochromatic plane wave incident at the surface @xmath59 of a semi - infinite medium occupying the half - space @xmath58 . the average intensity at @xmath83 can be then found in the diffusion approximation:@xcite @xmath84 .
the spatio - temporal correlation function of the field is given by a solution of the bethe - salpeter equation ( see appendix [ appa ] for details of the calculation ) : @xmath85 let us now consider the correlation functions of intensity fluctuations .
in addition to @xmath86 , which is the deviation of intensity from its average value , it is convenient to define @xmath87 , which is the change of the local intensity during the time interval @xmath88 . while @xmath89 , the correlation functions @xmath90 and @xmath91 for @xmath92 can be found in the factorization approximation : @xmath93 . \label{tint2}\end{aligned}\ ] ] both correlation functions ( [ tint1 ] ) and ( [ tint2 ] ) decrease exponentially with @xmath94 , and thus become negligible for @xmath9 .
intensity correlation persists , however , even for two points separated by a distance @xmath9 .
this correlation is due to the diffusive nature of wave transport in the medium and can be found either using the langevin approach@xcite or applying diagrammatic methods@xcite .
we give the details of calculations in appendix [ appb ] , the final results are @xmath95 \nonumber \\ & & \times\ , j_0\left(k \delta r/\ell \right ) , \label{tint3 } \\ & & c_{\delta i}({{\bf r } } , \delta{{\bf r } } , \tau ) = \frac{6}{(k \ell)^2 } i_0 ^ 2 \int_0^{\infty } d k\ , k \nonumber \\ & & \times\ , \delta q \left[k , \sqrt{k^2 + \ell^2/l_a^2 } , \frac{z}{\ell } , \frac{\delta z}{\ell } , \alpha\left ( \tau \right ) , \alpha(0 ) \right ] \nonumber \\ & & \times\ , j_0\left(k \delta r/\ell \right ) .
\label{tint4}\end{aligned}\ ] ] here we use the cylindrical coordinates : @xmath96 , @xmath97 is the bessel function of zeroth order , the function @xmath98 is defined in appendix [ appb ] , and @xmath99 & = & q \left[\cdots , \alpha(0 ) \right ] \nonumber \\ & - & q \left[\cdots , \alpha\left ( \tau \right ) \right ] .
\label{dqdef}\end{aligned}\ ] ] due to a rather complicated structure of the function @xmath98 ( see appendix [ appb ] ) , further calculations can be done only approximately . in the case of @xmath100
, we get @xmath101 \nonumber \\ & \times & \exp\left[-2\alpha\left ( \tau \right ) \frac{z}{\ell } \right ] , \label{lrs}\end{aligned}\ ] ] where @xmath102 is the geometrical average of @xmath103-coordinates of the two points for which the correlation is computed : @xmath104 , and @xmath105 . for @xmath106 , eq . ( [ lrs ] ) is exact . if @xmath107 , the correlation behaves essentially as @xmath108 , while for @xmath109 it becomes proportional to @xmath110 . for the correlation function of eq .
( [ tint4 ] ) we find : @xmath111 ( z/\ell ) c_{\delta i}({{\bf r } } , \delta{{\bf r } } , 0),\cr \hspace{4 cm } 2 \alpha(\tau ) ( z/\ell ) \ll 1\cr 2 c_{\delta i}({{\bf r } } , \delta{{\bf r } } , 0 ) , \hspace{1.75 cm } 2 \alpha(\tau ) ( z/\ell ) \gg 1 } \label{didi}\end{aligned}\ ] ]
consider a single wave path of length @xmath49 going from the source of waves to some point @xmath5 .
the phase acquired by a wave traveling along such a path can be written as @xmath112\ , d s_1 , \label{ph}\end{aligned}\ ] ] where the integration is along the path , and @xmath33 .
the squared difference @xmath113 , averaged over various realizations of disorder , and over all possible paths of length @xmath49 , is found directly from eq .
( [ ph ] ) : @xmath114 where the four contributions corresponding to @xmath115 originate from different physical processes .
below , we give explicit expressions of these terms and discuss their origin . the first term in eq .
( [ dphnl ] ) , @xmath56 is the linear term given by eq .
( [ dphi0 ] ) .
the next three terms , namely the terms corresponding to @xmath116 , are absent in the linear case and only appear because of nonlinear nature of wave interaction with the medium .
explicit expressions for these terms are @xmath117 here the integrations are assumed along wave paths of length @xmath60 [ in eq .
( [ dphi3 ] ) , both integrals are along the same path ] .
( [ dphi2 ] ) originates from the short - range correlation of intensity fluctuations [ see eqs .
( [ tfield])([tint2 ] ) ] : @xmath118 ^ 2 \exp\left(-\frac{\delta s}{\ell } \right ) , \label{dphi2p}\end{aligned}\ ] ] where the wave path is assumed to be ballistic at distances shorter than @xmath2 .
( [ dphi2p ] ) reduces to ( [ dphi2 ] ) for @xmath0 .
next , the term given by eq .
( [ dphi3 ] ) is due to the long - range correlations of intensity fluctuations [ see eq .
( [ tint4 ] ) ] .
the averages entering into the right - hand sides of eqs .
( [ dphi1])([dphi2p ] ) are @xmath119 where the integrations are over the volume of the disordered medium , and @xmath120 . in eqs .
( [ im])([didim ] ) , @xmath18 is the probability density for a path of length @xmath49 , going from the source of waves to the point @xmath5 , to pass through the vicinity of @xmath5 , and @xmath121 is the probability density for the path to pass consequently through the vicinities of @xmath122 and @xmath123 .
these two `` path distributions '' should be calculated with account for a particular geometry of disordered sample and source of waves . once the geometry is fixed , the calculation is straightforward . for the case of a plane wave incident upon a semi - infinite disordered medium ,
the calculations of @xmath18 and @xmath121 are presented in appendix [ appc ] .
let us discuss briefly the physical origins of nonlinear contributions to the dephasing given by eqs .
( [ dphi1])([dphi3 ] ) .
@xmath124 describes the change of the effective wave number in the nonlinear medium : @xmath125 $ ] .
this contribution can be either positive or negative , depending on the sign of @xmath126 , but its absolute value is always much smaller than @xmath56 , as long as @xmath127 .
@xmath124 can therefore only cause a small correction to the linear correlation function .
the next contribution , @xmath128 , originates from fluctuations of the local intensity , while @xmath129 is due to the long - range correlation of these fluctuations . an important difference between the linear term ( [ dphi0 ] ) , the first nonlinear term ( [ dphi1 ] ) , and the two last nonlinear terms ( [ dphi2 ] ) , ( [ dphi3 ] ) , is that the latter terms do not depend explicitly on @xmath57 .
the terms given by eqs .
( [ dphi2 ] ) and ( [ dphi3 ] ) are determined by the _ intensity fluctuations , _ and not by the scatterer displacements .
this might seem to be a meaningless statement , as the intensity fluctuations are , in their turn , caused by the scatterer motion .
the important point is that the scatterer motion is not the only possible reason for the fluctuations of intensity with time .
weak , spontaneous fluctuations of @xmath3 ( due to thermal fluctuations of various parameters , vibrations in the experimental setup , fluctuations of the incident wave , etc . )
are inevitable in real physical systems .
( [ dphi2 ] ) and ( [ dphi3 ] ) provide a mechanism for this spontaneous and generally weak fluctuations to affect the dephasing @xmath50 and , consequently , the temporal correlation function of scattered wave .
as stated in the title , the present paper is devoted to _ weakly _ nonlinear disordered media .
we limit ourselves to a weak nonlinearity , as otherwise the problem becomes too involved . above
, we have already mentioned that we assume @xmath130 , and that this condition allows us to consider the transport of average intensity to remain unaffected by the nonlinearity [ assumption ( _ i _ ) of sec .
[ nweq ] ] .
this allows us to use `` linear '' results , @xmath131 and eqs .
( [ rhoa ] ) , ( [ rhob ] ) of appendix [ appc ] , for @xmath132 , @xmath18 , and @xmath121 in eqs .
( [ im])([didim ] ) .
it seems then natural to assume that @xmath133 and @xmath134 are also close to their linear values .
we can therefore replace these correlation functions in eqs .
( [ dim ] ) and ( [ didim ] ) by the expressions found in sec .
[ linear ] . then , making use of eqs .
( [ tint1])([tint4 ] ) , and performing necessary integrations , we obtain from eqs .
( [ dphi1])([dphi3 ] ) : @xmath135 \right\ } \frac{s}{\ell } , \label{dphi11 } \\
\left < \delta\varphi^2(\tau ) \right>_s^{(2 ) } & = & 2 \pi k_0 \ell \delta n^2 \left\ { h\left [ \alpha(\tau ) \sqrt{\frac{s}{3 \ell } } \;\right ] \right . \nonumber \\ & - & \left .
[ \alpha(0 ) \sqrt{\frac{s}{3 \ell } } \;\right ] \right\ } \frac{s}{\ell } , \label{dphi22}\\ \left <
\delta\varphi^2(\tau ) \right>_s^{(3 ) } & = & 6 \delta n^2 s\left [ \alpha(\tau)\sqrt{s/\ell } , \alpha(0)\sqrt{s/\ell}\ ; \right ] \nonumber \\ & \times & \left(\frac{s}{\ell}\right)^{3/2}. \label{dphi33}\end{aligned}\ ] ] here @xmath136 $ ] and @xmath137 is assumed for simplicity .
the function @xmath138 in eq.([dphi33 ] ) is @xmath139 \nonumber \\ & - & \frac{2 z + \sqrt { r^2 + 4 z^2}}{\sqrt { r^2 + 4 z^2 } } \nonumber \\ & \times & \exp\left [ -\frac{3}{4}\left(2 z + \sqrt { r^2 + 4 z^2 } \right)^2 \right ] . \label{f}\end{aligned}\ ] ] unfortunately , integrations in eq .
( [ s ] ) can not be performed in the general case .
we find , however , the following approximate results : @xmath140 here numerical factors of order unity are omitted before each of the four asymptotic expressions .
while approximate , the above results enable one to compute the temporal autocorrelation function @xmath141 of diffusely reflected wave numerically using eqs .
( [ path ] ) , ( [ pn ] ) , and ( [ dphnl ] ) : @xmath142/w[0 ] , \label{g1nl } \\ & & w\left [ \left
< \delta \varphi^2(\tau ) \right>_s \right ] \nonumber \\ & & = \int_{0}^{\infty } p(\ell ,
s ) \exp\left [ - \frac12 \left <
\delta \varphi^2(\tau ) \right>_s \right ] d s. \label{pathnl}\end{aligned}\ ] ] we present the results of the calculation in fig .
[ fig1 ] for fixed @xmath143 , @xmath144 , and the values of the inverse absorption length @xmath145 indicated near each curve .
the `` linear '' correlation functions , corresponding to the same three values of the absorption length , and to @xmath146 , are shown by dashed lines . for weak absorption (
@xmath147 , @xmath148 ) , the initial ( i.e. at short correlation delay times @xmath88 ) decrease of the `` nonlinear '' autocorrelation function is much faster than that of the linear one .
hence , our perturbation approach fails at short times . indeed , the results of this section are based on the assumption that @xmath133 and @xmath134 in the nonlinear medium are close to their values in the linear one . applying the factorization approximation
, we find that this implies that @xmath149 .
although this condition seems to hold well for sufficiently large @xmath88 , it can be violated at small @xmath88 , where , as follows from fig .
[ fig1 ] , @xmath150 can become much larger than @xmath151 , if absorption is sufficiently weak . to estimate the region of validity of our perturbation approach , we require that the linear contribution to the dephasing @xmath152 given by eq .
( [ dphi0 ] ) , should be considerably greater than the sum of nonlinear contributions [ eqs .
( [ dphi11])([dphi33 ] ) ] . as the longest path length contributing to the integral of eqs.([path ] ) and ( [ pathnl ] ) is @xmath153
, we obtain the conditions of validity of the perturbation theory in the form : @xmath154 \ll 1 .
\label{validity}\end{aligned}\ ] ] since @xmath155 is an increasing function of its argument , and @xmath156 , condition @xmath157 \ll 1 $ ] ensures eq .
( [ validity ] ) at any @xmath88 .
it is the case for the upper curve of fig .
[ fig1 ] , corresponding to @xmath158 .
for such a strong absorption , the perturbation theory is valid at any @xmath88 , and the nonlinear autocorrelation function is close to the linear result .
by contrast , if @xmath157 > 1 $ ] , the perturbation approach can be applied only for sufficiently long correlation times @xmath159 , where the critical time @xmath160 is determined by eq .
( [ validity ] ) . for the lower curve of fig .
[ fig1 ] , corresponding to @xmath147 , we find @xmath161 .
it is worthwhile to note that in the absence of absorption , even an infinitely small nonlinearity ( @xmath162 ) suffices for the perturbation theory to fail at @xmath163 .
our condition of validity of the perturbation approach ( [ validity ] ) is consistent with the result of spivak and zyuzin,@xcite who have shown that the perturbation analysis of the sensitivity of speckle pattern in a nonlinear disordered medium to changes of scattering potential fails for @xmath164 , where @xmath165 is the typical size of the medium .
as demonstrated in the previous section , the perturbation theory fails to describe the temporal autocorrelation function of wave diffusely reflected from a nonlinear medium for @xmath166 , where @xmath160 is defined by eq .
( [ validity ] ) . to calculate @xmath141 at such short times
, one has to use a non - perturbative , self - consistent analysis .
however , the possibility of performing such an analysis is considerably limited by the mathematical complexity of the considered problem . to make the self - consistent analysis possible , we adopt the following two additional assumptions : * the statistics of a wave field scattered in a weakly nonlinear disordered medium , is close to gaussian .
consequently , the factorization approximation holds in a weakly nonlinear medium : @xmath167 . *
the functional form of the @xmath5-dependence of @xmath168 is the same as that of @xmath169 : @xmath170 $ ] , where @xmath171 is some unknown function , which can depend not only on @xmath88 but also on other parameters of the problem ( namely , on @xmath145 , @xmath144 , @xmath143 ) . strictly speaking ,
the above assumptions define a sort of perturbation theory , but now we do not limit the values of deviations of intensity and field correlation functions from their values in the linear case . instead , we assume that the nonlinearity does not cause significant modifications of the _ statistics _ of scattered waves [ assumption ( _ iii _ ) ] and of the _ functional form _ of the field correlation function [ assumption ( _ iv _ ) ] .
note that now the autocorrelation function @xmath168 can deviate significantly from @xmath169 .
condition ( _ iv _ ) fixes the functional form of this deviation , but implies no constraints on its absolute value .
obviously , both the assumptions ( _ iii _ ) and ( _ iv _ ) require the nonlinearity to be weak .
assumption ( _ iii _ ) is justified under the same conditions as ( _ i _ ) ( see sec . [ nweq ] ) , since the gaussian statistics of the _ total _ scattered wave field @xmath21 is a consequence of the complete randomization of phases of _ partial _ waves arriving at @xmath5 .
the reason for the randomization is that the typical distance @xmath2 between individual scattering events in a multiple - scattering sequence is much larger than the wavelength ( @xmath0).@xcite obviously , such a mechanism of phase randomization is equally effective in both linear and weakly nonlinear media , as long as ( _ i _ ) holds . to justify the ansatz of assumption ( _ iv _ ) , we apply eq .
( [ path ] ) and write @xmath172 as @xmath173 where we assumed that @xmath174 increases monotonically with @xmath49 and becomes of order unity at @xmath175 . after performing integration , eq .
( [ g1ps ] ) can be approximately rewritten as @xmath176^{1/2}\ } \simeq \exp[-\beta(\tau)\ , z/\ell]$ ] with @xmath177 . in the opposite limit of @xmath178
, @xmath168 vanishes and the functional form of its @xmath103-dependence is of no importance .
anyway , it will be seen from the following that the exact functional form of the @xmath5-dependence of @xmath71 is not of crucial importance , since @xmath179 is integrated over the whole medium during the calculation of the correlation function of diffusely reflected wave . making use of eqs .
( [ dphi1])([dphi3 ] ) and relying on the assumption ( _ iii _ ) , ( _ iv _ ) , we recalculate the nonlinear contributions to the dephasing @xmath50 . due to the assumption ( _ iv _ ) , the result can be obtained from eqs .
( [ dphi11])([dphi33 ] ) by a simple substitution : @xmath180 .
recalling that @xmath156 , we obtain : @xmath181 \frac{s}{\ell } , \label{d11 } \\
\delta\varphi^2(\tau ) \right>_s^{(2 ) } & = & 2 \pi k_0 \ell \delta n^2 \left\ { h\left [ \left(\beta + \frac{\ell}{l_a } \right ) \sqrt{\frac{s}{3 \ell } } \;\right ] \right . \nonumber \\ & - & \left .
h\left [ \frac{\ell}{l_a } \sqrt{\frac{s}{3 \ell } } \;\right ] \right\ } \frac{s}{\ell } , \label{d22 } \\ \left
< \delta\varphi^2(\tau ) \right>_s^{(3 ) } & \simeq & 6 \delta n^2 \nonumber \\ & \times & \cases { \beta \times ( s/\ell)^2 , \cr \hspace{5 mm } \beta \sqrt{s/\ell } \leq 1 , ( \ell / l_a ) \sqrt{s/\ell } \leq 1\cr ( s/\ell)^{3/2 } , \cr \hspace{5 mm } \beta \sqrt{s/\ell } > 1 , ( \ell / l_a ) \sqrt{s/\ell } \leq 1 \cr \beta \times ( l_a/\ell)^3 \times \sqrt{s/\ell } , \cr \hspace{5 mm } \beta \sqrt{s/\ell } \leq 1 , ( \ell / l_a ) \sqrt{s/\ell } > 1 \cr ( l_a/\ell)^3 , \cr \hspace{5 mm } \beta \sqrt{s/\ell } > 1 , ( \ell / l_a ) \sqrt{s/\ell } > 1 } \label{d33}\end{aligned}\ ] ] as follows from eqs . ( [ d11])([d33 ] ) , the nonlinear dephasing depends on the unknown function @xmath182 .
since the temporal autocorrelation function of diffusely reflected wave @xmath183 $ ] is determined by @xmath182 as well , eqs .
( [ g1nl ] ) and ( [ pathnl ] ) allow us to formulate a self - consistent equation for @xmath182 : @xmath184 = f\left [ \beta\left ( \tau \right ) \right ] , \label{beq}\end{aligned}\ ] ] where @xmath185/w[0]$ ] , @xmath186 $ ] is defined by eq .
( [ pathnl ] ) , and @xmath50 is a sum of terms given by eqs .
( [ dphi0 ] ) , ( [ d11])([d33 ] ) .
( [ beq ] ) is the main result of our self - consistent analysis .
although the functional form of @xmath187 is rather complicated , and eq .
( [ beq ] ) can not be solved analytically , the numerical solution is straightforward and can be carried out for any values of @xmath143 , @xmath145 , @xmath144 and @xmath88 .
( [ beq ] ) can be considered as a self - consistent equation for the autocorrelation function of diffusely reflected wave , since @xmath171 and @xmath141 are directly related .
it is worthwhile to note that in the absence of nonlinearity ( @xmath146 ) , eq .
( [ beq ] ) yields @xmath188 , and hence eq .
( [ glin0 ] ) is recovered for @xmath189 .
we start the analysis of eq .
( [ beq ] ) from the case of immobile scatterers , taking a limit of @xmath163 .
we denote the autocorrelation functions corresponding to this limit by @xmath190 in order to distinguish them from @xmath191 , which correspond to @xmath192.@xcite obviously , @xmath193 . in a nonlinear medium , eqs .
( [ d22 ] ) and ( [ d33 ] ) can still contribute to the dephasing even for @xmath163 .
these contributions are insensitive to the sign of @xmath143 .
a corresponding value of @xmath194 and , consequently , of the field autocorrelation function @xmath195 can be found by solving eq .
( [ beq ] ) numerically . in fig .
[ fig2 ] , we plot the left - hand and the right - hand sides of eq .
( [ beq ] ) for fixed @xmath196 , @xmath197 , and for several values of @xmath145 .
if absorption is strong ( @xmath198 for considered @xmath199 and @xmath144 ) , eq .
( [ beq ] ) has a unique solution @xmath200 which corresponds to @xmath201 .
however , a second solution @xmath202 appears for sufficiently weak absorption ( @xmath203 ) .
the point of appearance of the second solution is a _ bifurcation _ point of eq .
( [ beq ] ) . to choose the solution realizable in a physical system
, we note that the first solution [ @xmath204 exists only for @xmath205 , and disappears at finite @xmath206 , since @xmath207 for @xmath208 .
this solution is therefore inaccessible by continuity , and `` unstable '' with respect to small scatterer displacements .
the physically realizable solution should represent the limit of @xmath171 for @xmath209 , which is given by the second solution of eq .
( [ beq ] ) .
it is therefore this solution which one expects to be realized in a real physical system .
the fact that eq .
( [ beq ] ) can have a positive solution @xmath202 is a very important issue , since @xmath202 leads to @xmath210 < 1 $ ] .
a value of the temporal autocorrelation function which is less than unity is commonly associated with temporal fluctuations of scattered waves . in the considered case , however , the reason for these fluctuations is not the motion of scatterers , as the limit of @xmath211 corresponds to immobile scatterers .
the fluctuations are _ spontaneous _ and represent a clear signature of _ instability _ of the multiple - scattering speckle pattern.@xcite despite a rather complicated structure of the function @xmath187 in eq .
( [ beq ] ) , a relation between the parameters of the problem corresponding to the onset of the speckle pattern instability can be found analytically . as follows from fig .
[ fig2 ] , the initial [ at @xmath204 decay of @xmath212 $ ] should be faster than the decay of @xmath213 $ ] , for the second solution of eq .
( [ beq ] ) to appear . a surface in a three - dimensional space of the problem parameters @xmath143 , @xmath145 , @xmath144 , separating the stable [ @xmath204 and unstable [ @xmath202 ] regions , is therefore given by the equation @xmath215 \bigg|_{\beta(0^+ ) = 0 } = -1 .
\label{surface}\end{aligned}\ ] ] recalling that @xmath216 , @xmath217 is assumed , we obtain from eq .
( [ surface ] ) : @xmath218 \simeq 1 , \label{stbl}\end{aligned}\ ] ] where we introduce a control parameter @xmath219 , and a numerical factor of order unity is omitted .
if @xmath220 , the multiple - scattering speckle pattern is stable [ @xmath201 ] , while for @xmath222 , an instability shows up leading to @xmath223 . a striking feature of eq . ( [ stbl ] )
is that @xmath219 can become larger than unity even for very small @xmath199 , provided that the extensive parameter @xmath224 is large enough .
our condition of the speckle pattern stability @xmath220 agrees with the condition of validity of the perturbation theory developed in sec.[perturb ] [ eq .
( [ validity ] ) ] , evaluated at @xmath225 .
this readily explains the failure of the perturbation theory for short correlation times and weak absorption : perturbation approach is not suitable for description of unstable regimes .
moreover , the fact that the condition of validity of the perturbation theory and the speckle pattern stability condition agrees indicates that the additional assumptions ( _ iii _ ) and ( _ iv _ ) of sec . [ self ] are not essential for obtaining unstable regimes . to illustrate our self - consistent theoretical framework , we solve eq .
( [ beq ] ) numerically for @xmath163 , and plot the resulting temporal autocorrelation function of diffusely reflected wave , @xmath195 , in fig .
[ fig3 ] . as discussed above
, @xmath223 corresponds to spontaneous fluctuations of scattered waves which is a manifestation of the speckle pattern instability .
it follows from fig .
[ fig3 ] , that in accordance with eq .
( [ stbl ] ) , an infinitely small @xmath199 is sufficient to make the speckle pattern unstable in the absence of absorption , while a certain threshold degree of nonlinearity is required to destabilize the speckle pattern in a dissipative medium . in the absence of absorption , eq .
( [ beq ] ) always has two solutions : @xmath226 and @xmath202 , corresponding to @xmath201 and @xmath223 , respectively . as discussed above , it is the second solution , shown by a dashed line in fig .
[ fig3 ] , which is the physically realizable one .
in contrast , if @xmath227 , @xmath199 should be greater than some threshold value for the second solution @xmath228 to appear
. threshold values of @xmath199 following from eq .
( [ stbl ] ) are shown in fig .
[ fig3 ] by arrows . for a nonabsorbing ,
elastically scattering medium ( @xmath147 ) , the value of @xmath195 can be estimated analytically . indeed , at @xmath163 the principal contribution to @xmath229
is given by @xmath230 for @xmath231 , and by @xmath232 for @xmath233 .
this allows us to put @xmath234 if @xmath235 , and @xmath236 if @xmath237 .
integration in eq .
( [ pathnl ] ) can be then carried out , and eq .
( [ beq ] ) is easily solved , yielding @xmath238 this result agrees well with the numerical solution of eq .
( [ beq ] ) presented in fig [ fig3 ] .
the physical origin of the instability of speckle pattern for @xmath222 can be revealed by realizing that the system `` coherent wave + nonlinear disordered medium '' has a positive three - dimensional feedback . in a nonlinear medium , an infinitely small perturbation of the wave intensity
@xmath3 produces a change of the local refractive index , which alters the phases of waves propagating in the medium and , consequently , affects their mutual interference . since it is this interference which determines @xmath3 , the loop of the feedback is closed .
for @xmath239 , the feedback is sufficiently strong to compensate for the ( diffusive on average ) spreading of the initial intensity perturbation , and the speckle pattern @xmath3 is unstable .
it is worthwhile to note that unstable regimes are not exceptional in nonlinear wave systems and , in particular , in optical systems ( see , e.g. , refs . ) . our eq .
( [ beq ] ) is in no way limited to the case of @xmath211 , and can be used to compute the temporal autocorrelation function of diffusely reflected wave at any @xmath208 . in the latter case , one should take into account all the four dephasing terms given by eqs .
( [ dphi0 ] ) , ( [ d11])([d33 ] ) .
the results of the numerical solution of eq .
( [ beq ] ) are shown in fig .
[ fig4 ] . for weak absorption ( @xmath240 ) ,
@xmath241 and the speckle pattern is unstable [ @xmath223 ] . as @xmath206 increases , the difference between the `` nonlinear '' and `` linear '' curves becomes less pronounced . for strong absorption ( @xmath158 ) , @xmath219 becomes less than unity and stability of the speckle pattern is recovered [ @xmath201 ] .
we remind that the temporal autocorrelation function @xmath189 , corresponding to a linear medium , always equals @xmath243 for @xmath205 , as shown by dashed lines in fig .
[ fig4 ] .
it is instructive to compare the results obtained using the perturbation theory of sec .
[ perturb ] and the self - consistent approach of sec . [ self ] .
such a comparison is shown in fig .
the two upper curves , corresponding to relatively strong absorption ( @xmath158 , @xmath244 ) , are almost indistinguishable , which means that for @xmath220 , the perturbation theory works very well . in contrast , for @xmath222 ( see the two lower curves of fig
. [ fig5 ] , corresponding to @xmath147 ) , the perturbation and the self - consistent curves are close only at the right from the dotted vertical line , showing the minimum time @xmath160 at which the perturbation theory is valid [ see eq .
( [ validity ] ) ] . for @xmath245 , the perturbation and the self - consistent results disagree significantly , which confirms our conclusion about the failure of the perturbation approach at short correlation delay times .
we are now in a position to answer the two central questions formulated in the introductory section : * the phenomenon of multiple scattering is capable of providing a positive feedback for a coherent wave propagating in a nonlinear disordered medium . *
the onset of the speckle pattern instability occurs when the control parameter @xmath246 \label{stbl1}\end{aligned}\ ] ] becomes of order or larger than unity .
the speckle pattern is stable for @xmath244 .
the instability of the multiple - scattering speckle pattern manifests itself in spontaneous fluctuations of the scattered wave field and intensity .
the following features are characteristic for the development of the instability .
first , the development of the instability is independent of the sign of nonlinearity .
this is not common for nonlinear waves since the instability is often due to self - focusing phenomena , which only occur for @xmath247.@xcite the instability of waves in a _ disordered _ nonlinear medium has nothing to do with the self - focusing , and occurs at relatively weak nonlinearities , when the self - focusing can be neglected .
second , in the absence of absorption , the speckle pattern is unstable for any ( even infinitely small ) value of the nonlinearity strength @xmath199 , while in a dissipative medium @xmath199 should exceed a certain threshold value for the instability to show up .
finally , the instability results in a value of the autocorrelation function @xmath141 of scattered wave , which is smaller than @xmath243 for @xmath163 ( i.e. , in the absence of scatterer motion ) .
this is a clear signature of spontaneous fluctuations of the multiple - scattered speckle pattern , and should be observable in experiments .
the author is indebted to r. maynard for numerous discussions and continuous support of this work .
in this appendix , we provide a derivation of the spatio - temporal correlation function , @xmath248 , of a random field @xmath21 in the bulk of disordered medium . starting from the linear wave equation [ eq . ( [ weq ] ) with @xmath36 ]
, we obtain the bethe - salpeter equation in the form:@xcite @xmath250 , \label{bethe}\end{aligned}\ ] ] where the integrations are over the volume of disordered medium , @xmath251 is the average green function of the linear wave equation , and @xmath252 is the irreducible four - point vertex . far enough from the medium boundaries
, we can replace @xmath251 by its value in the infinite medium : @xmath253 .
\label{grfunc}\end{aligned}\ ] ] now we assume that point scatterers in the medium undergo brownian motion , and that the correlation function of the dielectric function fluctuations is given by eq .
( [ cde ] ) . for @xmath44
, we can neglect the @xmath88-dependence of @xmath252 in eq .
( [ bethe ] ) , which in the limit of @xmath0 becomes @xmath254 eq .
( [ bethe ] ) then reduces to @xmath255 where we assumed the coherent field @xmath256 to be negligible .
now we remind that @xmath257 is a slow - varying function of @xmath258
. we can therefore pull it out of the integral in eq .
( [ bethe2 ] ) , taking its value at @xmath5 . performing the remaining integral , we find @xmath259 this equation relates the spatio - temporal correlation of the field with its purely temporal correlation .
( [ cf1 ] ) holds for any sample geometry and source distribution , far enough from boundaries and sources , and for @xmath44 . in the case of a plane wave incident at the surface @xmath59 of a semi - infinite disordered medium , occupying the @xmath260 half - space , @xmath261 \label{cf1t}\end{aligned}\ ] ] with @xmath262 , leading to eq .
( [ tfield ] ) in the main text . in the factorization approximation ( [ tint1 ] ) , the square of eq .
( [ cf1 ] ) gives the short - range correlation function of intensity fluctuations .
to calculate the spatio - temporal correlation function of intensity fluctuations , @xmath263 , for @xmath9 we generalize the langevin approach,@xcite which has been initially developed for calculation of purely spatial correlations .
we assume that the spatio - temporal correlation function of langevin random sources is determined by the short - range correlation function of intensity fluctuations . in the factorization approximation , the latter
is given by the square of the field - field correlation function [ eq . ( [ cf1 ] ) ] .
hence , the generalized langevin equations read @xmath264 \delta i({{\bf r } } , t ) = { \rm div } { { \bf j}}_{ext}({{\bf r } } , t ) , \label{lgvn1 } \\ & & \left < j_{ext}^{(i)}({{\bf r}}-\delta{{\bf r}}/2 , t)\ , j_{ext}^{(m)}({{\bf r}}+\delta{{\bf r}}/2 , t+\tau ) \right > \nonumber \\ & & = \frac{1}{3 } \delta_{i m } \frac{c^2 2\pi\ell}{k^2 } \left| c_{\psi}({{\bf r } } , \tau ) \right|^2 \delta(\delta{{\bf r } } ) , \label{lgvn2}\end{aligned}\ ] ] where @xmath265 is the diffusion coefficient , and @xmath266 is the speed of wave in the medium . in the case of a plane wave incident upon a boundary @xmath59 of a semi - infinite disordered medium ,
it is convenient to make a two - dimensional fourier transform of eq .
( [ lgvn1 ] ) in the @xmath267 plane .
then the equations corresponding to @xmath268 and @xmath269 are multiplied , and the product is ensemble averaged .
further transformations , which are equivalent to those discussed in ref . ,
yield @xmath270 \exp\left[-2 \alpha(\tau ) \frac{z^{\prime}}{\ell } \right ] , \label{ik}\end{aligned}\ ] ] where @xmath271 , and @xmath272 is the green function of eq .
( [ lgvn1 ] ) .
evaluating eq .
( [ ik ] ) and transforming the result back to the real space , after lengthy but straightforward algebra we obtain @xmath273 \nonumber \\ & \times & j_0\left ( k \frac{\delta r}{\ell } \right ) , \label{ir}\end{aligned}\ ] ] where @xmath274/\left[4 \alpha \exp(2 \alpha \zeta_1)(\alpha^2 - p^2 ) p^2 \right ] \right ] \nonumber \\ & & + \left ( 1 + \frac{k^2}{p^2 } \right ) \frac { \sinh(p \zeta_1 ) \sinh ( p \zeta_2)}{2 \exp\left [ 2 ( \alpha + p ) \zeta_2 \right ] ( \alpha + p ) } \nonumber \\ & & + \frac{\exp(-p \zeta_2 ) \sinh(p \zeta_1)}{4 \alpha p^2 ( \alpha+p ) } \nonumber \\ & & \times \left\ { p(p^2 + 2\alpha p - k^2 ) \left[\cosh(p \zeta_2 ) \exp[-(2\alpha + p ) \zeta_2 ] \right.\right.\nonumber \\ & & - \left.\left . \cosh(p
\zeta_1 ) \exp[-(2\alpha + p ) \zeta_1 ] \right ] \right .
\nonumber \\ & & \left.+(p^3 - 2\alpha k^2 - pk^2 ) \left[\sinh(p \zeta_2 ) \exp[-(2\alpha + p ) \zeta_2 ] \right.\right.\nonumber \\ & & \left.\left.- \sinh(p \zeta_1 ) \exp[-(2\alpha + p ) \zeta_1 ] \right ] \right\ } , \label{q}\end{aligned}\ ] ] where @xmath275 and @xmath276 .
equations ( [ ir ] ) , ( [ q ] ) can not be evaluated in a general form . in the limits of @xmath277 we can , however , approximately replace @xmath98 in eq .
( [ ir ] ) by @xmath278 \nonumber \\ &
\times & \exp[-2 \alpha(\tau ) \zeta ] , \label{q0}\end{aligned}\ ] ] which yields eq . ( [ lrs ] ) in the main text .
propagation of waves in a disordered medium can be interpreted in terms of partial waves traveling along various paths inside the medium .
the spatial distribution of such paths and their relative weights depend on the scattering properties of the medium , and on the geometry of the sample . in the case of multiple scattering ,
the simplest and , at the same time , sufficiently accurate model of wave propagation is the _ diffusion _ model . according to this model ,
wave paths in the medium coincide with trajectories of brownian particles .
the probability @xmath279 for a path of length @xmath49 to pass from @xmath122 to @xmath123 is then given by a solution of the diffusion equation , which in the absence of absorption reads:@xcite @xmath280 where @xmath2 is the mean free path .
commonly used boundary conditions for eq .
( [ de ] ) consist in putting @xmath281 at open boundaries and @xmath282 at reflecting boundaries of the sample ( where @xmath283 denotes the normal derivative of @xmath284 ) .
@xmath279 is called the green function , or the propagator . for a semi - infinite medium occupying the half - space @xmath260
, one finds @xmath285 \right.\nonumber \\ & - & \left .
\exp\left[-\frac{3}{4 \ell s } \left(\delta r^2 + z^2 \right ) \right ] \right\ } , \label{green}\end{aligned}\ ] ] where cylindrical coordinates are used : @xmath286 , @xmath287 , @xmath288 , and @xmath289 . following ref .
, we introduce @xmath290 , the _ density distribution _ of paths of length @xmath49 , as a number of visits of a given site @xmath123 inside @xmath291 in the ensemble of paths of length @xmath49 starting at @xmath122 and ending at @xmath292 , over the total length of the ensemble distinct paths : @xmath293 @xmath290 describes the probability density for a path of a given length @xmath49 , starting at @xmath122 and ending at @xmath292 , to pass through @xmath123 . this quantity is normalized : @xmath294 where the integration is performed over the volume of disordered medium . as the green function
@xmath284 is known [ eq . ( [ green ] ) ] , the calculation of @xmath290 is straightforward . for diffusely reflected paths , assuming that the first and the last scattering events take place at @xmath295 , we obtain @xmath296 where @xmath297 is the surface of the semi - infinite medium . equation ( [ rhoa ] ) defines the probability density for a diffusely reflected path of length @xmath49 to pass through a vicinity of some point @xmath298 .
generalizing definition of @xmath299 , we define the probability density for a path of length @xmath49 starting at @xmath122 and ending at @xmath300 to pass consequently through @xmath123 and @xmath292 : @xmath301 the normalization of eq .
( [ rhobdef ] ) is @xmath302 for a semi - infinite medium , we get @xmath303 \right .
\nonumber \\ & - & \left.\frac{z + \sqrt{\delta r^2 + z^2}}{\sqrt{\delta r^2 + z^2 } } \right .
\nonumber \\ & \times & \left.\exp\left [ -\frac{3}{4 \ell
s}\left(z + \sqrt{\delta r^2 + z^2 } \right)^2 \right ] \right\ } , \label{rhob}\end{aligned}\ ] ] where @xmath96 , @xmath304 , @xmath305 , @xmath306 .
although eqs .
( [ rhoa ] ) and ( [ rhob ] ) have been found for a nonabsorbing medium , it is easy to show that these results hold in the presence of _ spatially - homogeneous _ absorption as well .
this stems from the fact that the attenuation of wave in a homogeneously absorbing medium depends only on the path _ length _ , while being independent of the path _ geometry .
_ as a consequence , the green function [ eq . ( [ green ] ) ] should be multiplied by a factor @xmath307 , where @xmath308 is the absorption length .
this factor , however , disappears after the substitution of the green function ( [ green ] ) in eqs .
( [ rhoadef ] ) and ( [ rhobdef ] )
. consequently , @xmath18 and @xmath309 are independent of @xmath308 and remain unchanged .
d. j. pine , d. a. weitz , g. maret , p. e. wolf , e. herbolzheimer , and p. m. chaikin , in _ scattering and localization of classical waves in random media , _ edited by p. sheng ( worls scientific , london , 1989 ) , p. 312 . for the sake of concreteness
, we use the `` optical '' language throughout the paper .
our results are , however , general , and apply to all types of scalar waves ( in the case of optical waves , polarization effects are neglected ) .
although we take the limit of @xmath163 , we assume @xmath310 , where @xmath311 is the typical time required for the system to `` jump '' from one solution of eq .
( [ weq ] ) to another .
the time autocorrelation function @xmath141 is independent of @xmath88 for @xmath310 and @xmath163.@xcite obviously , eq . ( [ weq ] ) can not be used to study the _ dynamics _ of spontaneous fluctuations . in order to do this ,
it is necessary to keep the time derivatives of @xmath21 in the wave equation .
our analysis only allows to reveal the presence of such fluctuations , and to calculate their time autocorrelation function @xmath141 for @xmath312 . | * abstract * we consider the multiple scattering of a scalar wave in a disordered medium with a weak nonlinearity of kerr type .
the perturbation theory , developed to calculate the temporal autocorrelation function of scattered wave , fails at short correlation times .
a self - consistent calculation shows that for nonlinearities exceeding a certain threshold value , the multiple - scattering speckle pattern becomes unstable and exhibits spontaneous fluctuations even in the absence of scatterer motion .
the instability is due to a distributed feedback in the system `` coherent wave + nonlinear disordered medium '' .
the feedback is provided by the multiple scattering .
the development of instability is independent of the sign of nonlinearity .
= 0 cm |
this paper describes the initial configuration of stacee , a new ground - based gamma - ray detector . before addressing the detector itself
we provide a brief summary of the current experimental situation in gamma - ray astronomy and show the motivation for stacee and similar detectors .
gamma - ray astronomy has recently become a very exciting area of research . during the lifetime of nasa s compton gamma ray observatory ( cgro ) from april , 1991 to may , 2000 and following the construction of ground based detectors during the 1990 s , the field experienced rapid growth .
the amount and quality of data increased and theoretical understanding of the related astrophysics improved greatly .
two instruments that were aboard the cgro are of special interest to high energy astrophysics .
the burst and transient source experiment ( batse ) amassed a large data set of enigmatic gamma - ray bursts ( grbs ) and the energetic gamma ray experiment telescope ( egret ) produced a catalog of over 200 high energy point sources @xcite .
six of these sources have been identified with pulsars within our galaxy and more than 70 have been found to be active galactic nuclei ( agns ) at great distances .
optical or radio counterparts for the remaining sources have yet to be identified .
the egret telescope detected gamma rays by converting gamma ray to @xmath0 pairs in a spark chamber tracking device surrounded by an anti - coincidence shield which vetoed charged particles .
this latter feature ensured an excellent signal - to - background ratio .
egret could operate in this way because it was in orbit above the earth s atmosphere .
thus it was necessarily a small detector and could only look at sources below about 10 gev .
this upper limit was given by statistics ; the exact value was defined by intrinsic source strength , the steepness of its spectrum and the observing time allocated .
most ground - based gamma - ray detectors use the atmospheric cherenkov technique .
most resemble the whipple telescope @xcite , which was the first experiment to convincingly detect gamma - ray sources .
typical cherenkov telescopes detect gamma rays by using large , steerable mirrors to collect and focus cherenkov light produced by the relativistic electrons in air showers resulting from interactions of high energy gamma rays in the upper atmosphere .
this cherenkov light is distributed on the ground in a circular pool with a diameter of 200 - 300 m. the cherenkov telescopes need only capture a small part of the total pool to detect a gamma ray so the telescopes have very large effective collection areas relative to satellite detectors , albeit at higher energy thresholds .
the energy threshold for cherenkov telescopes is the result of a competition between collecting a small number of photons from a low energy shower and rejecting a large number of photons from night sky background .
it dictated by a number of important parameters , including photon collection area ( @xmath1 ) , detector field of view ( @xmath2 ) , integration time ( @xmath3 ) and efficiency for getting a photoelectron from a photon hitting the primary collection mirror ( @xmath4 ) .
it can be summarized in the following approximate formula : @xmath5 all of the parameters except @xmath1 are limited by current technology ( @xmath6 @xmath4 depends partly on quantum efficiency of photocathodes ) or by the physics of the air shower ( @xmath6 @xmath3 can not be less than the time over which shower photons arrive at the detector . ) the only parameter which is practical to control is the collection area of the instrument . for present generation imaging detectors , this is less than 100 m@xmath7 and thresholds are typically greater than 300 gev .
the energy range between egret and the cherenkov telescopes remained unexplored until recently since there were no detectors sensitive to the region from 10 gev to 300 gev .
there are , however , new detectors being built or commissioned that use the solar tower concept .
this concept is a variant of the air cherenkov technique whereby the collecting mirror is synthesized by an array of large , steerable mirrors ( heliostats ) at a central - tower solar energy installation .
the large effective size of the collecting mirror allows one to trigger at lower photon densities and therefore lower primary gamma - ray energies .
it should be noted that future satellite detectors such as glast @xcite will explore some of this region but will be statistics - limited above some energy which will depend on the source strength .
ground - based detectors will be able to complement low energy satellite measurements with data taken over shorter time intervals .
this is important for detecting flaring activity in agns as well as pulsed emission from pulsars . with longer integration times they will be sensitive to fainter sources .
stacee ( solar tower atmospheric cherenkov effect experiment ) is designed to lower the threshold of ground based gamma - ray astronomy to approximately 50 gev , near the upper limit of satellite detectors .
three other projects of a similar nature , celeste @xcite , @xcite , solar-2 @xcite and graal @xcite have also recently been built or are under construction .
stacee will investigate established and putative gamma - ray sources .
one of its principal aims is to follow the spectra of agns out to energies beyond that of egret measurements to determine where the spectra steepen drastically .
these cut - offs are expected since , although whipple - type cherenkov telescopes have the sensitivity to see many of the egret agns if their spectra continue without a break , they have not detected them .
this effect could be due to cut - off mechanisms intrinsic to the source or to absorption effects between the source and the detector .
a likely mechanism is pair production wherein the high energy gamma ray combines with a low energy photon ( optical or infrared ) from the extragalactic background radiation field .
the infrared ( ir ) component of this field is difficult to measure directly but can provide information on the early universe since the photons are red - shifted relics of light from early galaxies .
correlating spectral cut - offs with source distance will help elucidate the nature of the ir field .
the field of high energy gamma - ray astronomy has recently been reviewed in three comprehensive articles @xcite , @xcite , @xcite .
the stacee detector has been under development since 1994 .
stacee-32 represents an intermediate step in the progress towards the design detector which is very similar in concept but has twice as many heliostats and more sophisticated electronics for triggering and pulse measurement .
earlier steps in the development of stacee are described in @xcite and @xcite .
stacee-32 was installed at the national thermal solar test facility ( nsttf ) , a solar energy research facility built by the us department of energy in the mid 1970 s .
it is situated at sandia national laboratories in albuquerque , new mexico ( 34.96@xmath8 n , 106.51@xmath8 w ) , at an altitude of 1700 m. as shown in figure [ nsttf ] , the nsttf consists of a 60 m concrete tower and an array of 212 heliostats , each of which has a mirror area of 37 m@xmath7 .
thirty two of these heliostats were used during clear , moonless nights to collect cherenkov light from air showers and direct it onto two secondary mirrors located near the top of the concrete tower .
the secondaries focussed the cherenkov light onto banks of photomultiplier tubes ( pmts ) such that each pmt viewed a single heliostat .
signals from the pmts were amplified and routed to a control room where they were processed by electronics and recorded if an appropriate trigger condition was satisfied .
the essential concept is depicted in figure [ stacee - concept ] . in the following subsections
we describe in more detail the elements of the detector .
the locations of the 32 heliostats used in stacee-32 , shown in figure [ h - layout ] , were chosen to sample relatively uniformly the cherenkov light pool expected from a shower impacting near the center of the array .
crowding of pmts in the image plane also played a role in the selection of heliostats .
each heliostat consists of 25 square facets mounted on a steel framework , as shown in figure [ heliostat ] .
each facet is a 4 foot by 4 foot square of back - surfaced aluminized glass glued onto a thin metal sheet .
a metal suspension is welded to the back of the facet and bolts attached to this suspension allow it to be fixed to the heliostat frame .
a bolt at the center of the facet and four at the corners are used to distort the facet into an approximately parabolic shape , the focus of which is set to be equal to the distance between the heliostat and the tower .
other bolts precisely set the orientation of the facet with respect to the heliostat frame .
each facet can be separately aligned so that their beams overlap at the tower .
the adjustment of the facet focusing was done during nsttf commissioning . since
then the facet alignment was checked and tuned using images of the sun projected onto the tower near solar noon .
a well tuned heliostat produced a small , bright `` sunspot '' composed of 25 overlapping images of the sun .
facets out of alignment caused distortions or broadening of this spot and were brought into line using the image quality as a diagnostic .
note that optimizing the image in this way does not assure that the optics are optimized for all locations of the sun ( or a cherenkov source ) on the sky .
it is , however , a practical compromise .
we have tracked individual stars to check optical alignment .
this provides some useful information but the sun is better as a diagnostic , partly because its angular size ( 0.5 degrees ) is very similar to that of a cherenkov shower .
the point source nature of stars makes the procedure overly sensitive to small errors in individual facet alignment or imperfections in their surfaces .
an extended source like the sun smoothes these out and provides a better estimate of average alignment .
the entire heliostat is mounted on a y - shaped yoke structure which allows motion in azimuth and elevation .
this motion is achieved with two electric motors , each of which is controlled by the nsttf central computer using 13 bit encoders .
the absolute pointing of the heliostats were determined using sunspots also .
the corresponding encoder readings , known as `` bias values '' were important parameters in the operation of the heliostats .
the important optical properties of the heliostats are their reflectivities and the size of their sun - spots .
both of these directly affect how many cherenkov photons will be directed onto a pmt .
the heliostat facets are made from back - surfaced glass to withstand decades of weathering in a desert environment .
the double transit of light through the glass results in a cut - off in the reflectance curve , as shown in figure [ reflect ] .
as previously mentioned , the focus and facet alignment of the heliostats were checked and adjusted using the image of the sun or moon on the nsttf tower . to quantify the size and brightness of each heliostat s sunspot
, a photograph was taken using a ccd camera and the data were analyzed to give information on the diameter of each spot and the amount of light contained with circles of specified diameter .
this information was used in the simulations used to calculate the optical throughput of the detector .
the heliostats used for stacee-32 were all inspected after being chosen and were found to be of uniformly high quality . although no systematic tests were carried out on the individual facets , visual inspection , sunspot images and performance in the experiment are consistent with the heliostats being approximately identical .
cherenkov photons were reflected by the heliostats onto two secondary mirrors located near the top of the central tower on a platform located 48 m above the base of the tower .
one mirror viewed the 16 heliostats in the eastern part of the array while the other viewed the 16 heliostats in the west .
the secondary mirrors were spherical , with a nominal diameter of 1.9 m and a focal length of 2.0 m. they focussed the light from the heliostats , which arrived as a wide beam , onto phototube cannisters fixed in position at the focal plane .
the optics were such that each heliostat was mapped onto its corresponding pmt channel .
this one - to - one mapping was vital for pattern recognition which was used in trigger formation and background suppression .
the off - axis geometry and spherical optics gave rise to coma aberrations in the heliostat images .
the coma caused a certain amount of cross - talk since light from one heliostat s tail could end up in the wrong pmt .
however , it was usually possible to reject this light because in almost all cases it arrived at the wrong time as a result of the different time - of - flight from its heliostat .
for certain pointing geometries , cross - talk could arrive close enough in time to be included in the adc gate but this would not significantly degrade the analog measurement which was already quite contaminated with night sky background integrated during the long ( 37 ns ) gate . to degrade the timing measurement , cross - talk photons would need to arrive closer in time to be accepted by the trigger and by offline cuts . on any given event
only one or two channels would be affected .
typically , the coma tails contained less than 10% of a channel s light so the net effect of timing and flux levels meant that cross - talk was not an important effect .
each secondary mirror was composed of seven identical hexagonal facets made from front - surfaced aluminized glass covered with a silicon oxide protective coating .
the reflectance curve , as provided by the manufacturer @xcite , is shown in figure [ reflect ] .
the facets were mounted on a spider , as shown in figure [ spider ] , with the center facet fixed in position and the outer six adjustable .
the outer six were aligned by placing a point source of light , surrounded by a small white screen , at the center of curvature of the mirror and adjusting the facets until all reflected images overlapped . with the facets aligned the resultant spot was approximately 1 mm in size .
the spider itself was attached at the pivot point of a jib crane mounted on a stationary forklift , or stacker .
the stacker allowed the raising of the mirror from its daytime stow position to its nighttime operational position .
jack screw actuators connecting the spider and the jib crane assembly controlled the pitch and yaw orientation of the mirror and were set once , upon installation .
the axes of the secondaries made an angle of @xmath9 with respect to the horizontal .
the final stage in the stacee optics chain is the camera . in stacee-32
there were two cameras , one for each secondary mirror .
a photograph of the mirrors is shown in figure [ 2mirrors ] and a photograph of one of the cameras and mirror assemblies is shown in figure [ rene_photo ] .
each camera consisted of 16 pmts and light concentrators enclosed in cannisters mounted on a slotted aluminum plate attached to a trolley .
the trolley rolled on clamping wheels , controlled by a threaded rod along a skid made from welded steel i - beams .
the skid extended from under the stacker and ran parallel with the horizontal component of the secondary mirror s axis .
about two thirds of the skid rested on the platform of the solar tower and the remaining third extended out from the tower .
the trolley was rolled in for stowage and out for observing . mounted on the trolley was a 1.0 m x 1.7 m aluminum plate machined with 25 mm slots on a 75 mm pitch .
this plate swivelled in elevation and was set to @xmath10 with respect to vertical .
this angle was prescribed by the optics of the system in the following way .
a line from the center of the group of heliostats observed by the camera to the center of the secondary mirror defines the primary optical path .
this line should reflect from the secondary to the center of the camera when it is in its correct position on the skid .
that position is defined by the focal length of the secondary mirror .
the tilt angle of the slotted plate is defined by requiring its normal to point to the center of the secondary .
the plate was tilted to this angle using a digital protractor and secured using jam nuts .
the pmt cannisters were mounted in cylindrical sleeves attached to an azimuth - elevation mounting system secured to the slotted plate .
a drawing of this assembly is shown in figure [ az - el ] . with this system it was possible to position a pmt cannister anywhere laterally on the slotted plate and to adjust the orientation of the cannisters such that they pointed to the center of the secondary mirror .
shims in the bottom of the sleeves provided adjustment along the axis of the cannister . this longitudinal degree of freedom was necessary to accommodate the curved focal plane of the system
. an exploded view of a pmt cannister is shown in figure [ cannister ] .
it contained the pmt ( not shown ) a silicone rubber optical coupling disk ( also not shown ) and a light concentrator , known as a dtirc .
the pmt base was fitted into the recess in the disk at the back of the cannister tube with its hv and signal connectors protruding through holes provided .
the 3 mm thick silicone disk was sandwiched between the pmt and the dtirc which was held in place by a locator device .
the entire assembly was held in longitudinal compression by a thin ring screwed in from the front .
the light concentrators used in stacee were dielectric total internal reflection concentrators ( dtircs ) @xcite .
these are non - imaging devices which use total internal reflection to transport light from the front surface to the exit aperture ; the light from a circular area of 11 cm diameter was focussed to an exit diameter of less than 4 cm .
only light from a given angular range could reach the exit aperture so the dtircs had the added feature of being able to define the field of view of the pmt .
functionally , a dtirc is very much like the more familiar winston cone @xcite , used extensively in particle physics detectors .
in fact , it is a generalization .
the dtirc characteristics , defined by the shape of the front surface and the side profile , can be varied over a range of possibilities . for compactness and ease of manufacture
the stacee design incorporated a spherical front face and straight , tapering sides .
a photograph of a stacee dtirc is shown in figure [ fig.dtirc ] .
the dtircs were made from solid uv - transparent acrylic manufactured for the sudbury neutrino observatory @xcite .
this material has transmission well into the ultraviolet ( attenuation length greater than 10 cm for wavelengths greater than 320 nm ) and has a refractive index of 1.49 .
the transmittance @xmath11 wavelength for a typical pathlength in a dtirc is shown in figure [ reflect ] .
three types of dtircs were used , each with the same front aperture but differing cone angles and therefore different acceptance angles .
these were used to roughly equalize the fields of view of different channels .
the field of view of a given heliostat is given by the effective diameter of the secondary mirror divided by the distance from the secondary to the heliostat .
to equalize the fields of view we used dtircs with a narrow angular acceptance ( which therefore view a smaller part of the secondary ) to view nearby heliostats .
large acceptance dtircs were used to view more distant heliostats . for practical reasons
, we limited the number of different dtirc sizes to three rather than make a different size for each channel .
table [ tab.dtircs ] summarizes the properties of the three dtirc types .
.dtirc properties .
[ tab.dtircs ] [ cols= " < ,
< , < , < , < , < " , ] the pmts were supplied with high voltage from a lecroy 4032a high voltage supply located on the detector level and controlled by a lecroy 2132 camac interface in the control room via a long ribbon cable .
voltages were typically in the neighborhood of -1600 v. the operating voltages for the pmts were determined by measuring the gain of each channel at several high voltage settings .
these measurements were fit to the function : @xmath12 and the function was inverted to get the voltage required to obtain the desired gain .
to measure the absolute gain for a pmt , we placed it in a dark box and obtained a spectrum of single photoeletron pulses using a 2 gs / s digital oscilloscope in self - triggered mode , adding up the samples over a 20 ns range .
a sample pulse height spectrum is shown in figure [ singlepe ] .
the gains were adjusted slightly , after installation of the pmts in the detector , using a collimated led to produce light pulses ( approximately 50 photo - electrons each ) in each pmt separately .
the high voltage values were changed to equalize the response of all channels to the led signal .
the stacee-32 electronics could be grouped into subsystems corresponding to their function in the experiment , as shown in figure [ electronics ] .
signals from the phototubes were filtered and amplified near the cameras then sent to the stacee control room , located 20 m below the detector in the tower .
here they were discriminated and used in trigger logic and timing measurements .
analog pulses from selected channels were digitized .
the electronics were an assembly of commercially available nim and camac modules which served to define a trigger and measure the times and charges of pulses from the pmts .
as such , the electronics were very similar to those found in a typical high - energy physics experiment .
a major difference , however , was the requirement of a dynamic delay system . due to the earth s rotation , the gamma - ray source appeared to move across the sky during the course of a night s observations .
this effect continuously changed times of arrival of cherenkov photons at each heliostat . in order to maintain tight coincidences , signals from different channels
were required to be delayed by amounts which had to be incremented in nanosecond steps every few seconds .
the range of these delays was of order 250 ns .
this was one of the design challenges faced and it influenced the configuration presented here .
the front - end analog electronics were physically close to the pmts ; they were installed in a small hut next to the cameras on the detector level of the nsttf tower .
the pmt signals arrived at the hut via 11 m long rg58 cables which were bundled , in groups of 4 , inside ground braids of the same length . in the hut
, the signals passed through a high - pass rc filter having a time constant of 75 ns .
this filter blocked any dc component of the pmt signal and removed slow pmt transients which were not associated with cherenkov signals .
the dc component of the photocurrent developed a voltage across the resistor in the filter and this voltage was sent by ribbon cable to a joerger adc-32 scanning adc module in the control room .
typical currents were between 20 and 40 @xmath13a .
the pulsed components of the signals exiting the filters were amplified by two cascaded fixed gain ( x10 ) wide - band ( 275 mhz ) amplifiers ( phillips scientific 776 ) .
this amplification factor of 100 allowed us to keep the pmt gain to @xmath14 which prolonged the life of the pmts in an environment of high night sky background light levels .
the filtered and amplified signals were routed through 40 m low - loss coaxial cables ( rg213 ) from the detector level to the control room level of the tower where they were fed into linear fanouts ( phillips scientific 748 ) .
the outputs of these fanouts were passed to the discriminators and adcs .
the effect of these components on the pmt pulse was to broaden and attenuate it .
the electronics had a negligable effect compared with the cables .
a typical pulse was about 4 ns wide at the pmt and 6 ns wide after the 40 m cables .
( additional cables were used to delay the pulses which were sent to the adcs .
these are described later ; their effect was to broaden the pulses to approximately 25 ns . )
a schematic of the stacee-32 trigger electronics is shown in figure [ trigger ] .
all analog signals from the cameras were discriminated in 16 channel discriminators ( lecroy 4413 ) operating with a common threshold .
this threshold was set according to a rate @xmath11 threshold curve like the one shown in figure [ rate_thresh ] . in
this plot one sees a `` break point '' , defined as the threshold below which the rate climbs exponentially . the location of the break point depended on the individual channel rates , the widths of the discriminator pulses and the number of channels required to form a trigger . at thresholds below the break point ,
the rate was dominated by accidental coincidences and at very low values it flattened due to dead - time effects . at thresholds above
the break point the rate decreased slowly with threshold . here , the experiment was triggering largely on cherenkov light from cosmic ray showers .
in running stacee-32 we set the threshold 15 - 20 mv above the break point , which means there was very little background from accidental triggers .
the accidental rate @xmath15 for a trigger requiring @xmath16 channels to fire , from a group of @xmath17 , each firing at rate @xmath18 and with pulse widths of @xmath3 is given by :
@xmath19 this formula shows that the accidental rate varies with r@xmath3 to a large power . to reduce the contribution from accidentals , one can either lower r , by raising the discriminator threshold or use a narrower pulse width , thus making @xmath3 smaller . since the discriminator threshold is directly related to the minimum number of photoelectrons detected ( and thereby the energy threshold of the detector ) , it is advantageous to reduce @xmath3 as much as possible . with stacee-32
it was not possible to make @xmath3 less than approximately 20 ns .
this width is much longer than the physical limit given by the intrinsic width of the cherenkov wavefront ( about 4 ns ) so there is obvious room for improvement .
the final version of stacee will have new delay units that do not impose this limitation and we will be able to run at a lower threshold .
stacee-32 had a two - level trigger .
two levels were necessary because of the limited ranges of the programmable delays described below .
the 32 heliostats were divided into four local subclusters of eight heliostats each , as shown in figure [ h - layout ] .
the discriminator outputs from the eight channels in each subcluster were routed through delays programmed to bring in - time hits into coincidence .
the in - time signals were added in a passive summing circuit .
the output of this circuit was sent to a discriminator with a threshold set corresponding to a certain multiplicity , for example 5/8 .
this multiplicity requirement constituted the `` level 1 '' ( l1 ) or subcluster trigger .
the programmable delays were accomplished by commercial modules ( lecroy 4518 ) .
these delays compensated for the changing timing of the wave - front caused by the apparent motion of the source across the sky .
they were supplemented by fixed delays , made of low loss rg213 cables , which accounted for the different distances of heliostats from the tower .
signals emerging from the delays were re - shaped by a discriminator with a 12 ns output width .
typical pmt rates were 1 - 5 mhz and typical l1 rates were 1 - 10 khz .
each l1 trigger was routed through a second delay unit ( caen c469 ) which adjusted for the time differences among the four subclusters .
these differences were order 200 ns , larger than the intracluster delays owing to the longer distances involved .
the lecroy 4518 delays used tapped delay line technology so they were effectively deadtimeless .
the pmt discriminators contributed an effective deadtime that was rate dependent , typically less than 5% per channel .
individual channel deadtime was studied for on - source and off - source data for each source tracked , to ensure that its effect canceled in the on - off subtraction .
the caen units used digital counting techniques to set the delay so were dead during each delay period .
the level 2 ( l2 ) trigger was formed by requiring a minimum multiplicity from the delayed l1 triggers , typically 3/4 .
this multiplicity requirement was implemented using a standard logic unit ( phillips scientific 755 ) with input pulses of 12 ns width .
trigger formation was inhibited whenever an external veto signal was asserted .
a by - product of the two level trigger ( the l2 requirement ) is that the light pool is required to be more spread out over the entire detector .
this is a feature expected of showers due to gamma - rays .
the multiplicity requirements for the l1 and l2 triggers were selected to give the lowest possible energy threshold consistent with good gamma - ray sensitivity over the entire array . a figure of merit which was the crab gamma - ray rate divided by the square root of the rate for charged cosmic rays was used in simulations to quantify the studies @xcite .
l2 triggers were combined with fake trigger signals generated at a rate of 0.5 hz .
fake triggers were used for determining individual channel rates and pedestal values . whenever either type of trigger occurred ( l2 or fake )
, a common stop signal was sent to the tdc , a gps clock time was latched and adc integration gates were sent to the adcs ( see next section ) .
event read - out was then initiated . during readout ,
a veto was asserted to prevent occurrence of additional triggers .
the veto was cleared by the data acquisition program ( daq ) at the conclusion of readout .
the typical readout time was 40 ms per event .
the stacee-32 delay system had sufficient range to trigger on cherenkov showers coming from any region of the sky within @xmath20 of zenith .
individual delay settings could be controlled with nominal precision of 1 ns for the l1 delays and 2.5 ns for the l2 delays . to ensure precise timing ,
every channel was calibrated with test pulses and the calibrations were employed in a lookup table .
the delays were updated every few seconds as stacee-32 tracked sources across the sky .
accidental rates due to random night sky hits in the pmts could be directly measured by using random settings for delays , thereby imposing unphysical coincidence requirements . under normal operating conditions the accidental rate was 5 - 10 triggers per hour .
analog pulses from the pmts were digitized using wilkinson - type integrating adcs ( lecroy 2249sg ) which are 11 bit devices with 0.25 pc / count resolution .
there are 12 channels per double width camac module and each channel has its own gate .
this latter feature was required because of the different and changing times between signals in different pmts .
signals routed to the adc inputs passed through long ( 1000 - 1300 ns ) low - loss rg213 cables which delayed the signals while the trigger was formed , a task that took of the order of a microsecond .
the adc gates were made by fanning out a digital pulse from the trigger and delaying each copy by the appropriate amount in a caen c469 programmable delay module .
the gates arrived at the adc module 5 ns before the pulse and , as with the trigger , the delays were updated as the source was tracked across the sky .
this scheme is depicted in figure [ adc ] . due to budget constraints ,
only 24 of the 32 channels in the detector were instrumented with adcs .
routing the pmt signals through long cables had the effect of distorting the pulses .
the main effect was to make each pulse longer and smaller in amplitude .
( this is discussed more fully in the section on adc modeling . ) to accommodate the lengthened pulse , we set the gate width to 37 ns . this long gate allowed a significant amount of night sky noise to be added to the signal . assuming a typical single photoelectron counting rate of 1.5 ghz on each pmt , an average of 56 night sky background photoelectrons arrived during the integration gate . since the pmts were ac - coupled , the error on the pulse measurement was the square root of this or about 8 photoelectrons .
this meant that the charge resolution of the adc system was poor near the threshold of about 6 photoelectrons .
the utility of the adc system was largely in allowing time slewing corrections to be made to the data .
as part of the upgrade to the final detector we are equipping all channels with 1 gs / s digitizers ( fadcs ) . the digitizers will obviate the need for delay cables since the pmt pulses can be retrieved from the appropriate location in the digitizer memory .
stacee-32 was equipped with a laser calibration system comprising a @xmath21j nitrogen laser and dye cell feeding a network of optical fibres through a system of adjustable neutral density filters .
the fibres delivered light to the pmt s by exciting small wavelengthshifter plates attached to the center of the secondary mirrors .
the intensity of each laser shot was measured independently using 4 pin photodiodes .
a very similar system is described more fully in @xcite .
the system was used for measuring the relative time differences between pmt channels . in the final detector
, it will also be used for monitoring gain stability of the pmts .
a key application in stacee-32 was the parameterization of time - slewing effects .
time slewing is the phenomenon whereby pulses of different amplitude exceed the discriminator threshold at systematically different times .
this effect can be studied by sending a series of laser pulses covering a range of intensities to the pmts . by plotting the measured time for a given channel against the corresponding charge , as in figure [ slewing ]
, one can fit a parameterization of the slewing effect to the data .
( the parameterization chosen is an empirical function which describes the data ; it has no physical significance . )
fits were performed for all 24 adc - equipped channels and the parameters obtained were used in data analysis .
the correction can be up to about 3 ns , which is large relative to the desired timing resolution of 1 ns .
the timing resolution was estimated by examining distributions of residuals to shower fits .
it was better than 1 ns for all channels and did not depend on pointing angle .
it was stable over time .
stacee-32 had a gps clock ( truetime gps ii ) which was used to provide a time - stamp , accurate to @xmath22s , for all recorded events .
these time - stamps were necessary for pulsed emission searches . as shown in figure [ electronics ]
, counters were used to measure the fraction of time that the experiment was live .
whenever a trigger occurred , a flip - flop was set which asserted a veto signal that inhibited further triggers .
this veto was required to prevent additional triggers from interfering with the readout , and it was cleared by the data acquisition computer once the readout was complete .
a 10 khz clock was sent to two scaler channels to measure the deadtime .
one scaler counted a vetoed copy of the clock and the other counted an unvetoed copy .
the ratio of the vetoed scaler count to the unvetoed scaler count gave the livetime fraction of the experiment .
additional scalers were used to monitor the rates of all the phototubes as well as the l1 subcluster triggers .
the l1 deadtime fractions during normal running were negligable .
the l2 combined l1 rates that were typically of order 10 khz and used delays of order 200 ns .
thus the deadtime introduced by the non - pipelined caen delays was not important .
the overall deadtime was dominated by the readout time and varied between 8% and 12% depending on the trigger rate .
the data acquisition ( daq ) system for stacee-32 was based on a silicon graphics indy computer which was coupled to two camac crates via two hytec ethernet crate controllers .
data were read out after each event trigger and stored on a local disk . at intervals
the data were written to dlt tapes for archiving and off - site analysis .
processes running on the daq computer included run control ( starting , stopping and clearing of vetoes ) , monitoring of pmt currents and rates , generation of alarms , and readout of the adc s , tdc s , scalers and the gps clock
. the daq system also calculated and set trigger delays using the camac delay modules .
it was also used to set discriminator thresholds and high voltage values .
in stacee-32 the experimental setup involved the atmosphere as part of the detector so its behavior can not be verified by a test beam .
instead , it is modelled by monte carlo programs and certain aspects of these programs can be checked with data . in this section
we report on results of some of these calculations and tests .
the modeling of stacee-32 can be divided into three parts .
they are : * simulation of extensive air - showers * simulation of the optical throughput of stacee-32 * simulation of the electronics in the following subsections we deal with each part in turn . the design and understanding of stacee-32 was aided by the mocca air shower simulation package @xcite .
mocca simulates the entire development of an extensive air shower , starting with the first interaction of the primary particle in the upper atmosphere and following all generated secondary particles until they reach the ground or their energy falls below the point where they no longer contribute to shower development .
note that at the energies relevant to stacee-32 ( e @xmath23 ev ) we are able to follow all particles since the multiplicities are small enough . with user - specified energy , incident direction and particle type of the initial primary , mocca s output consists of wavelengths , coordinates and directions of cherenkov photons , along with their time of impact at ground level .
the intervening processes accounted for by the program include ionization , bremsstrahlung and pair production , as well as effects such as deflection by the geomagnetic field and coulomb scattering in the atmosphere .
our chief goal for stacee-32 was the detection of gamma rays so the understanding of the detector s response to electromagnetic showers is important .
the precise response to hadrons is less important since it is largely needed for understanding backgrounds from charged cosmic rays .
mocca , like all shower monte carlo programs , has little difficulty with electromagnetic showers since they are well described by qed calculations .
uncertainties arise with hadronic showers , which can not be calculated from first principles .
even these uncertainties are not expected to be very large since , for collisions in the stacee energy range , most cross sections have been measured at accelerators or can be reliably extrapolated from lower energies .
mocca uses simplified descriptions of these cross sections to model hadronic interactions .
there are other programs that simulate air showers .
another leading package is corsika @xcite , developed by the kaskade collaboration .
corsika makes use of packages such as egs4 @xcite and gheisha @xcite , which are widely used in nuclear and particle physics .
the more detailed particle interaction models in corsika tend to make it slower than mocca . to check the robustness of mocca predictions of key observables such as total yield of cherenkov photons we have compared it with corsika .
the photon densities from the two simulators agree at the @xmath24 level for primary gamma - ray energies of 50 gev and 100 gev and for zenith angles from @xmath25 to @xmath20 .
this agreement is reassuring but not surprising since an electromagnetic shower is a well understood and well modelled phenomenon .
the agreement for hadronic showers is not as good , reflecting the greater uncertainties inherent in modeling hadronic interactions .
the total photon densities from mocca agree with those from corsika to within @xmath26 .
an important input to the calculations is the assumed atmospheric profile .
the development of the shower depends on the density profile of the atmosphere .
the rate and angle of production of cherenkov photons by particles in the shower depend on the local refractive index .
this index can be computed from the local temperature and pressure .
in addition , the attenuation of cherenkov light due to rayleigh and mie scattering and absorption by oxygen allotropes are important effects .
the calculations for stacee-32 used a profile in which the atmosphere was divided into five zones .
the density profile followed an exponential distribution with a characteristic scale height in each zone .
the model parameters were determined from linsley s parameterization of the u.s .
standard atmosphere @xcite .
rayleigh scattering is the most important mechanism for attenuating the cherenkov light of interest to stacee-32 .
it scatters photons outside the field of view of the detector with negligible chance of scattering back into the acceptance .
the @xmath27 absorption length for this process is given by : @xmath28 where @xmath29 is the number density of air molecules , atoms and ions ( but not electrons ) , @xmath30 is the photon wavelength , and @xmath16 is the index of refraction . for vertical air showers simulated by mocca , the transmission probability for cherenkov photons as a function of wavelength is shown in figure [ trans ] .
the second part of the stacee-32 simulation chain traced the optical path of cherenkov photons through the detector elements .
for this part , a custom - made ray tracing package called `` sandfield '' ( sandia field simulator ) was developed .
sandfield followed the path of every cherenkov photon through the optical elements ( heliostats , secondary mirrors and dtircs ) onto the pmt photocathodes , folding in transfer efficiencies at every stage .
the end result was a list , for each channel , of photo - electrons and their times of arrival .
these lists were passed to the electronics simulator for further processing . in the following ,
we describe in more detail some features included in modeling the different optical elements of stacee-32 . as described earlier , the stacee-32 heliostats were compound structures consisting of 25 individually adjustable facets mounted on a frame which itself could be oriented in two directions .
each facet was a 4-foot square of back - aluminized glass which could be deformed to a shape which was , to first order , a paraboloid .
the facets were co - aligned to bring the image of the sun to a tight spot on the central tower .
sandfield modelled a heliostat as a collection of 25 parabolic reflectors , each with a focal length equal to the distance from the heliostat to the tower .
the orientation of each facet with respect to the heliostat frame was individually adustable within the program .
these orientations were set to their correct values ( those that would produce the smallest possible sunspot ) then given a random gaussian error in heading .
this error was expressed in terms of a parameter called @xmath31 which represented the rms of the linear displacement of the facet image on the tower .
the displacement was measured radially from the nominal target position .
typically @xmath32 m .
the value of @xmath31 for each heliostat was determined by comparing simulated sunspots , made with different values of @xmath31 , with the real sunspot , as measured using a ccd camera . in figure
[ sunspot ] , simulated and measured sunspot curves for a sample heliostat are shown .
plotted are the normalized integrals of light as a function of radius , @xmath33 fractions of total light contained within a given radius .
the simulations have been made with a value of @xmath31 that provides the best fit to the measured data . typically a @xmath34 m radius secondary captured about @xmath35 of the light .
cherenkov photons reflecting from the heliostats were traced to the tower where they encountered the secondary mirrors .
the photons were then directed towards the cameras .
some of the photons from heliostats near the tower were occulted by the camera structures and did not survive to hit the secondaries ; this effect was accounted for in the simulation .
the secondaries themselves were modelled as single - piece spherical mirrors with reflectance as shown in figure [ reflect ] and were oriented as determined during a survey and alignment procedure during installation .
the faithful treatment of the spherical optics automatically produced the spherical and off - axis aberrations which were important features of this setup .
sandfield included ray tracing routines which simulated the dtircs and the silicone rubber disks which coupled them to the pmts .
dielectric reflection and refraction effects were included at all medium interfaces .
absorption inside the dtirc and attenuation effects in the silicone were also included .
the last step in the optical modeling of stacee-32 was at the pmt photocathode . here
the photons were converted to electrons with a probability equal to the quantum efficiency of the pmt .
this quantity depended on the wavelength of the incident photon and composition of the photocathode . additionally , it depended on the angle of incidence and point of impact of the photon on the photocathode .
the angular dependence has been modelled and verified experimentally by the celeste collaboration @xcite who use the same pmts as stacee-32 .
their model was incorporated into sandfield ; the effect of the angular dependence was to enhance the reponse of the system at angles near to the geometric cut - off , where photons were impacting the photocathode at very oblique angles .
the enhancement can be seen in figure [ enhance ] .
stacee-32 pmts were bombarded with a high flux of photons either from air showers or from night sky background .
the elevated rates caused pile - up effects which needed to be understood quantitatively , so a detailed electronics simulation package was essential . in this section
we describe hermes ( helpful electronics reconstruction and measurement endcap simulation ) , the package developed for stacee-32 .
hermes simulated the behavior of the stacee electronics starting from the analog pulses at the output of the pmts .
it made use of a library of 2200 single photoelectron pulses which were obtained by recording the memory buffer of a digital oscilloscope ( tektronix 640a - 2.0 gs / s sample rate ) responding to single photoelectron pulses from a stacee pmt , set to a gain of @xmath36 .
simulated pmt pulses were constructed from library pulses by linear superposition .
the process began by combining photoelectrons from the air shower ( generated by the sandfield program ) with random night sky background photoelectrons generated uniformly in time according to poisson statistics .
the rate for night sky photoelectrons was an adjustable parameter , and was determined for each pmt from the ratio of its photocurrent to its gain .
the signal and background combinations were used to generate a simulated analog waveform by placing randomly chosen library pulses at the arrival times of each photoelectron .
the sum of these components was sampled in 50 ps steps by interpolating between the coarse grained points to define the pulse amplitude at each point in time .
this process was repeated for all channels resulting in a simulated event for the experiment .
different gains on the channels were simulated by appropriate scaling of the pulse amplitudes .
each waveform was 85 ns long , including a pre - pulse interval of 25 ns ; a sample is shown in figure [ fakepulse ] . to show how well hermes simulated stacee-32 pulses
, we display , in figure [ realpulse ] , a pulse waveform from one channel of the detector as captured by a 2 gs / s digital oscilloscope while running with the standard trigger conditions .
there is a strong similarity between the waveforms shown in figures [ fakepulse ] and [ realpulse ] . at this point , the simulated data consisted of 32 waveforms .
the waveforms were passed through simulated discriminators which modelled the behavior of the lecroy 4413 .
features such as threshold , double pulse resolution , updating behavior , and hysteresis effects were simulated .
deadtime ( order 10 ns ) after firing was automatically accounted for . to check the validity of the electronics simulation we compared certain key quantities .
for example , the widths and characteristically asymmetric shapes of the tdc distributions for real and simulated data were compared and have been found to be in good agreement .
a more powerful check is to reproduce the absolute rate @xmath11 threshold curves for various pmt channels .
the pmt rate at low threshold was due to the random pile - up of single photoelectrons exceeding the discriminator threshold .
it was directly related to the photocurrent , both quantities depending on the single photoelectron rate , the pmt gain , and the average pulse width .
small variations in any of these quantities produced large changes in the rate .
thus , reproducing actual rate @xmath11 threshold curves was a sensitive test of the validity of the simulation , particularly the calibration of the discriminator threshold level in terms of photoelectron equivalents .
this calibration directly affected the energy threshold calculation .
a representative simulation is shown in figure [ ratecurve ] where the absolute rate , measured @xmath37 , is plotted against the discriminator threshold .
the pmt current divided by the calibrated gain gives a single photoelectron rate of 1.21 ghz . a simulation curve , assuming this single photoelectron rate , fits the data remarkably well whereas curves made with @xmath38 changes to the single photoelectron rate fail to describe the data .
( note that this rate value is peculiar to this channel .
there are channel - to - channel variations in the exact rate and these are well described by our simulations .
@xcite ) as previously stated , pmt pulses were sent through delay cables before being digitized in the stacee-32 adcs .
the cables had the effect of considerably lengthening the pulses , leading to the requirement of a long ( 37 ns ) adc gate .
the pulse lengthening was modelled in hermes using a simple transform function , @xmath39 , which had the following effect on a delta function : @xmath40 the transform of any arbitrary waveform is found by convolving the waveform with this function .
this model was tested using a sample of digitized pmt pulses made with an led flasher before and after they had passed through a long cable .
the empirically determined values for @xmath30 varied linearly from 0.071 ns@xmath41 for a 1000 ns cable to 0.048 ns@xmath41 for a 1300 ns cable .
the model reproduced the observed amplitude attenuation and the charge collected within the adc integration gate .
the most important application of the simulation programs was the determination of stacee-32 s response to gamma ray and cosmic ray air showers . to determine flux levels from a source one needs to know the effective area of the detector as a function of energy .
we used the simulation to estimate stacee-32 s energy response .
a good end - to - end test of the simulation chain is the calculation of the rate due to cosmic ray triggers . the cosmic ray spectrum in the energy range relevant to stacee-32 is well known .
many cosmic ray runs were taken by stacee-32 so a comparison between calculated and measured cosmic ray rates was straight - forward .
the cosmic ray data were obtained by collecting showers from the zenith .
a suite of 30,000 proton showers scattered over the spatial , angular and energy range of the detector was generated .
the showers ranged in energy from 100 gev to 1 tev , following an @xmath42 differential spectrum , and were uniformly distributed in solid angle within @xmath43 of zenith .
each shower was used 10 times by placing its impact point at a different random place inside a 500 m diameter circle positioned at the center of the heliostat array .
the cherenkov photons in the simulated showers were propagated through the sandfield program and the resultant photoelectrons were analyzed by hermes .
night sky background rates from a zenith run were used to calculate the correct number of random photons to add to the signal photons .
the trigger criteria were applied and the ratio of accepted showers to generated showers was the trigger fraction .
this fraction was multiplied by the known flux of cosmic rays ( protons and helium nuclei only - heavier nuclei do not contribute significantly to the trigger rate ) to get the predicted trigger rate .
the rate predicted from the simulation , @xmath44 hz , was in reasonable agreement with the measured rate . for a run near the beginning of the 1998 - 99 campaign ,
the measured rate was @xmath45 hz . this provided a measure of confidence in the validity of the entire simulation chain .
it is worth noting that the observed cosmic ray rate did not remain constant throughout the campaign .
there was a drop from 2.9 hz in november , 1998 to 2.5 hz in december .
the rate remained at this level for the rest of the campaign .
it is suspected that bias values on some of the heliostats shifted .
these shifts degraded the optical alignment of the detector , thereby increasing the energy threshold .
a system of periodic bias checks has since been implemented to prevent this from re - occurring .
another end - to - end check of the simulation package was the prediction of mean adc values for cosmic ray triggers . using the proton showers described in the previous subsection ,
the mean charge as measured by each adc was computed and compared with that obtained from cosmic ray runs . as can be seen in figure [ adcmean ] , there was good agreement between simulations and data at the @xmath46 level .
the channel to channel structure , an artifact of the detector geometry , was well modelled .
air shower energies are usually reconstructed by fitting the measured pmt adc values to a lateral distribution function and by varying the normalization of the function .
this procedure could be done with data from stacee-32 but the resulting resolution was expected to be poor , especially near threshold , due to the large amount of night sky background included in the adc gates .
for example , for proton induced showers , a simple sum over adc channels gave an energy estimate with a resolution of @xmath47 .
the situation was better for gamma - ray induced showers coming from a point source at the center of the detector s field of view .
these showers were smoother and the lateral distribution of cherenkov light was more regular than for hadronic showers . using a simple sum over adc channels gave an energy resolution of @xmath48 for 200 gev gamma rays , averaged over showers whose
cores were distributed uniformly across the heliostat field @xcite .
a key parameter of any gamma ray detector is its effective area .
this area is , in general , an energy dependent quantity which depends on the details of the detector elements and the trigger criteria .
we define the effective area as follows : @xmath49 where @xmath50 is the fraction of gamma rays , the extrapolated trajectories of which land at position @xmath51 on the array , which trigger the experiment and are accepted for analysis .
the rate of accepted gamma rays is then : @xmath52 where @xmath53 is the differential energy spectrum . to trigger the detector
, one needs a minimum density of cherenkov photons on the ground .
thus , for vertical showers , the effective area rises from zero at low energy , where the photon density is below trigger threshold , to the area of the light pool @xmath54(100 m)@xmath7 ) at high energy . for non - vertical showers ,
the situation is more complicated . here , geometric effects cause the light pool to be spread out over a larger area , thus reducing the photon density . also , the photons must pass through a larger slant depth of atmosphere which subjects them to increased attenuation .
these effects cause the energy threshold to increase with zenith angle , which introduces an angular dependence into the effective area .
in stacee there are angular effects in the light collection optics ( for example projected areas of heliostats ) which further complicate the dependence of area on angle .
the effective area as a function of energy for a series of angles was computed using the simulation .
gamma rays with energies between 50 gev and 1 tev were generated for three different arrival directions corresponding to positions along the trajectory of the crab nebula .
these showers were scattered within a 200 m radius circle centered on the heliostat array , each shower being used for 20 different impact positions .
the standard trigger for the 1998 - 99 campaign ( level 1 = 5/8 , level 2 = 3/4 , discriminator = @xmath55 mv ) was applied and the trigger fraction was determined .
this fraction was multiplied by the area over which showers were thrown ( @xmath56(200 m)@xmath7 ) to get the effective area .
the resulting curves are shown in figure [ eff_area ] .
the acceptance is essentially zero below 100 gev but turns on rapidly and rises to a plateau value of @xmath57 28,000 m@xmath7 at high energy .
the energy at which the turn - on occurs depends on angle , as expected , as does the plateau value .
note that the uncertainties shown in the figure are statistical only .
the largest systematic errors are due to the understanding of the detectors , especially the optics , which are under continuing study .
( knowledge of these has been included in our published science results @xcite ) the energy threshold of a detector has to be defined before it can be calculated . because most detectors , especially those measuring air showers , do not have a sharp turn - on in sensitivity , assigning an energy threshold is a subtle task .
here we use the concept of the spectral threshold energy which is the energy at which the gamma - ray detection rate per unit energy reaches its peak .
it is defined as the energy at which @xmath58 is at a maximum . as such it depends on @xmath59 , the differential spectral flux of the gamma - ray source , and is therefore source dependent . for this paper
we assume a differential spectral flux : @xmath60 which is similar to that measured for the crab nebula in the @xmath61 gev region .
convolving this spectrum with an average of the effective area curves shown in figure [ eff_area ] gives 190 + /-
60 gev as the spectral threshold energy .
these numbers do not change significantly for spectral indices between 2.3 and 2.5 .
another important feature of a gamma - ray detector is its hadron rejection capability .
stacee achieves hadron rejection at the trigger level ( a factor of 50 - 75 ) essentially by an on - line topology cut given by the l1/l2 criteria .
offline , another factor of between 2 and 3 is achieved by more precise timing cuts .
these figures are estimated by comparing the event rates obtained ( mostly due to hadrons ) with those expected in the event of no rejection ( calculated from known cosmic ray fluxes ) .
one figure of merit used to quantify a detector s capability is the sensitivity to the crab . using our published result on this source @xcite we can say that stacee-32 could detect the crab at a 5@xmath31 significance level in 23 hours .
stacee-32 was a first - generation heliostat cherenkov detector .
it used 32 heliostats from an array of 212 heliostats to collect cherenkov light from air showers generated by the impact of high energy gamma rays on the upper atmosphere .
the large collection area allowed operation at a lower energy threshold than previously obtained by ground based detectors .
this device was operated for a complete observing season ( 1998 - 99 ) during which time a strong gamma - ray signal from the crab nebula was obtained @xcite , @xcite .
stacee-32 served as a proof - of - concept experiment .
we are now completing an upgrade of the detector to its design configuration .
features of the upgrade include : * the heliostat field ( asphalt ) has been darkened to reduce background from albedo of night sky light , * new custom trigger electronics with less dead - time allow operation at lower thresholds , * twice as many heliostats are used , for a greater light collection area , * a new linux - based data acquisition system with much reduced dead - time has been installed , and * state - of - the - art 1 gs / s fadcs have been introduced to improve timing and energy measurements on all channels .
we are grateful to the staff at the nsttf for their excellent support .
we acknowledge the loan of electronic equipment from the physics division of los alamos national laboratory .
we also thank the sno collaboration for providing us with acrylic for light concentrators , and the machine shop staffs at mcgill and chicago for their expert assistance .
many thanks to tom brennan , katie burns , jaci conrad , cathy farrow , william loh , anthony miceli , gora mohanty , alex montgomery , marta lewandowska , heather ueunten , and franois vincent . this work was supported in part by the national science foundation , the natural sciences and engineering research council , fcar ( fonds pour la formation de chercheurs et laide la recherche ) , the research corporation , and the california space institute . | we describe the design and performance of the solar tower atmospheric cherenkov effect experiment detector in its initial configuration ( stacee-32 ) .
stacee is a new ground - based gamma ray detector using the atmospheric cherenkov technique . in stacee ,
the heliostats of a solar energy research array are used to collect and focus the cherenkov photons produced in gamma - ray induced air showers .
the large cherenkov photon collection area of stacee results in a gamma - ray energy threshold below that of previous detectors . |
it is well - known that two generic cubics @xmath10 and @xmath11 in @xmath3 intersect each other in nine points @xmath12 . by constructing the corresponding _ pencil _ of curves
@xmath13 \in \cpo \}\ ] ] one can define a map @xmath14 @xmath15 .
blowing up @xmath3 at @xmath16 one can extend @xmath17 to a lefschetz fibration @xmath18@xmath5 @xmath19@xmath20 whose generic fiber is an elliptic curve . our aim in this paper is to describe this construction in terms of differential topology .
it turns out that many 4manifold topologists were curious about such a construction .
( for instance this was posed explicitly as a question in @xcite ) .
let @xmath21 denote the mapping class group of a compact connected orientable genus @xmath22 surface with @xmath0 boundary components and @xmath23 marked points , so that diffeomorphisms and isotopies of the surface are assumed to fix the marked points and the points on the boundary .
( we will drop @xmath0 if the surface is closed and drop @xmath23 if there are no fixed points . ) a product @xmath24 of right - handed dehn twists in @xmath25 provides a genus-@xmath22 lefschetz fibration @xmath26 over the disk with closed fibers . if @xmath27 in @xmath25 then the fibration closes up to a fibration over the sphere @xmath28 . a lift of the relation @xmath29 to @xmath30 shows the existence of @xmath0 disjoint sections of the induced lefschetz fibration .
the self - intersection of the @xmath31-th section is @xmath32 if @xmath33 in @xmath34 , for some positive integers @xmath35 , where @xmath36 s are right - handed dehn twists along circles parallel to the boundary components of the surface at hand ( cf .
@xcite ) . on the other hand ,
an expression @xmath37 in @xmath34 naturally describes a lefschetz pencil : the relation determines a lefschetz fibration with @xmath0 disjoint sections , where each section has self - intersection @xmath38 , and after blowing these sections down we get a lefschetz pencil ( cf .
conversely , by blowing up the base locus of a lefschetz pencil we arrive to a lefschetz fibration which can be captured ( together with the exceptional divisors of the blow - ups , which are all sections now ) by a relation of the above type . in this paper
we find relations of the form @xmath39 in @xmath40 for @xmath41 , generalizing the well - known cases @xmath42 .
a relation of this type naturally induces a lefschetz pencil and by blowing up we get an elliptic lefschetz fibration with @xmath0 sections .
moreover by taking the @xmath7-th power of our relation ( for @xmath43 ) we get @xmath44 for @xmath1 .
once again this relation induces an elliptic lefschetz fibration @xmath8 with @xmath45 singular fibers and @xmath0 disjoint sections , where the self - intersection of each section is equal to @xmath46 .
the reader is advised to turn to @xcite , @xcite and @xcite for background material on lefschetz fibrations and pencils . to simplify the notation in the rest of the paper
we will denote a right - handed dehn twist along a curve along @xmath47 also by @xmath47 .
a left - handed dehn twist along @xmath47 will be denoted by @xmath48 .
we will multiply the dehn twists from right to left , i.e. , @xmath49 means that we first apply @xmath50 then @xmath51 . * acknowledgement *
: the authors would like to thank john etnyre , david gay and andras stipsicz for their interest in this work .
] consider the four - holed sphere depicted in figure [ lantern ] .
then there is the relation @xmath52 in @xmath53 which was discovered by dehn @xcite .
it was rediscovered by johnson @xcite and named as the lantern relation .
we will freely use this relation in any subsurface ( of another surface ) which is homeomorphic to a sphere with four holes .
the particular depiction of the lantern relation on the four - holed sphere in figure [ lantern ] will be convenient in the subsequent discussions .
in this section we will generalize the well known one - holed torus relation to a relation on the @xmath0-holed torus for @xmath54
. we will give all the details in each case since the relation for @xmath55-holes is derived by using the relation for @xmath0-holes for @xmath56 .
the relations in the cases @xmath57 are also known , but we compute them anyway for the sake of completeness and to show our method . we note that if two circles are disjoint , then the corresponding dehn twists commute . also , if two circles @xmath47 and @xmath58 intersect transversely at one point , then the corresponding dehn twists satisfy the braid relation ; @xmath59 .
we use these facts in our computations below .
if @xmath50 and @xmath51 are two circles on a torus with one boundary @xmath60 which intersect each other transversely at one point , then there is the relation @xmath61 we call it the one - holed torus relation
. it turns out that this relation was known to dehn @xcite in a slightly different form .
consider the two - holed torus depicted in figure [ 2-star ] . by the lantern relation
, we have @xmath62 one - holed torus relation is @xmath63 note that @xmath64
. then @xmath65 it follows that @xmath66 . ]
consider the three - holed torus depicted in figure [ 3-star ] . by the lantern relation , @xmath67 and by the two - holed torus relation , @xmath68 note that @xmath69 .
then @xmath70 it follows that @xmath71 .
this relation was called the star relation in @xcite . ]
the lantern relation for the sphere with boundary @xmath72 in figure [ 4-star ] is @xmath73 the relation on the three - holed torus with boundary @xmath74 given in section [ three - star ] is @xmath75 here we identify the curves @xmath76 in figure [ 3-star ] by the curves @xmath77 in figure [ 4-star ] .
combining we get @xmath78 * remark . *
although we will not need it in the rest of the paper , by plugging in @xmath79 it is easy to see that this relation may also be written in a more symmetric form as @xmath80 ] the lantern relation for the sphere with boundary @xmath81 in figure [ 5-star ] is @xmath82 the relation on the four - holed torus with boundary @xmath83 given in section [ four - star ] is @xmath84 here we identify the curves @xmath85 in figure [ 4-star ] with the curves @xmath86 in figure [ 5-star ] .
combining we get @xmath87 ] the lantern relation for the sphere with boundary @xmath88 in figure [ 6-star ] is @xmath89 the relation for the five - holed torus with boundary @xmath90 given in section [ five - star ] is @xmath91 we identify the curves @xmath92 in figure [ 5-star ] with the curves @xmath93 in figure [ 6-star ] .
combining we get @xmath94 where @xmath95 . ]
the lantern relation for the sphere with boundary @xmath96 in figure [ 7-star ] is @xmath97 the relation on the six - holed torus with boundary @xmath98 given in section [ six - star ] is @xmath99 where we use the identification @xmath100 to go from figure [ 6-star ] to figure [ 7-star ] .
combining we get @xmath101 where @xmath102 and @xmath103 . ]
the lantern relation for the sphere with boundary @xmath104 in figure [ 8-star ] is @xmath105 the relation on the seven - holed torus with boundary @xmath106 given in section [ seven - star ] is @xmath107 where we use the identification @xmath108 to go from figure [ 7-star ] to figure [ 8-star ] .
combining we get @xmath109 where @xmath110 , @xmath111 and @xmath112 . ]
the lantern relation for the sphere with boundary @xmath113 in figure [ 9-star ] is @xmath114 the relation on the eight - holed torus with boundary @xmath115 given in section [ eight - star ] is @xmath116 where we identify @xmath117 with @xmath118 to go from figure [ 8-star ] to figure [ 9-star ] .
combining we get @xmath119 where @xmath110 , @xmath111 and @xmath120 . ] * remark .
* the curious reader might wonder why we stopped at @xmath121 .
first of all our process will not allow us to go any further .
in fact , there is a good reason for that : an elliptic fibration @xmath122 admits at most nine disjoint sections ( all with negative self - intersections ) .
so we conclude that there is no such relation for @xmath0-holed torus with @xmath123 .
first we consider the case @xmath124 .
the relation @xmath125 in @xmath126 we derived in section [ four - star ] induces the word @xmath127 in @xmath128 which gives us an elliptic lefschetz fibration on the elliptic surface @xmath4@xmath5 . to draw a kirby diagram ( @xcite ) of this elliptic fibration we start with a 0-handle , attach two 1-handles ( see figure [ 4-section ] ) and attach a 2-handle which yields @xmath129 .
a torus fiber of the trivial fibration @xmath130 can be viewed in figure [ 4-section ] as follows : take the obvious disk on the page , attach two 2-dimensional 1-handles ( going through two 4-dimensional 1-handles ) and cap off by a 2-dimensional disk .
then we draw the curves which appear in the monodromy of the elliptic fibration on parallel copies of this fiber .
notice that these curves are the attaching curves of some 2-handles .
once we attach all twelve of these 2-handles with framing one less than the page framing we get an elliptic lefschetz fibration over @xmath131 with twelve singular fibers , which then can be closed off to an elliptic lefschetz fibration over @xmath28 .
we depicted the four disjoint sections @xmath132 of the induced fibration in figure [ 4-section ] .
( imagine replacing @xmath133 s in figure [ 4-section ] by holes where they intersect the page and embed the curves in figure [ 4-star ] into distinct fibers . ) for each @xmath134 , the curve @xmath133 bounds two disks one in the neighborhood of a regular fiber , one outside of that neighborhood which in turn gives a section of the elliptic lefschetz fibration when we glue these two disks along their common boundary @xmath133 . with four disjoint sections .
, width=226 ] similarly we can draw the kirby diagrams corresponding to the relations we derived for @xmath135 , and explicitly indicate the locations of the @xmath0 disjoint sections of @xmath122 in these diagrams .
we skip the cases @xmath136 and jump to the case @xmath121 .
the relation @xmath137 on the nine - holed torus induces the word @xmath138 in the mapping class group @xmath128 which gives us an elliptic lefschetz fibration on the elliptic surface @xmath4@xmath5 , where @xmath139 ( which is indeed a right - handed dehn twist ) .
note that we cyclicly permuted the curves in the equation we derived in section [ nine - star ] to obtain the relation above . with nine disjoint sections .
note that we just indicated the intersection points of sections with the regular fiber by the encircled numbers corresponding to the boundary components @xmath140 of the nine - holed torus.,width=264 ] finally , for @xmath141 , by taking the @xmath7-th power of our relation for the @xmath0-holed torus we can find @xmath0 disjoint sections of the corresponding elliptic fibration on the elliptic surface @xmath142 for any @xmath43 .
in this appendix , we outline how to prove the lantern relation in section [ lant ] and the one - holed torus relation in section [ one - holedtorus ] .
these are well known , but we include them here to have a complete presentation .
it is a classical fact that a self - diffeomorphism of a @xmath143-disc which is identity on the boundary is isotopic to the identity through such diffeomorphisms .
it follows that in order to show that two self - diffeomorphisms of a surface are isotopic to each other , it is sufficient to prove that their action on a set of curves whose complement is a disc are the same up to isotopy . after this remark ,
the lantern relation @xmath144 is proved by looking at the action of the diffeomorphisms on the both sides of the equality on the curves shown on the left in figure [ rel ] .
similarly , the one - holed torus relation @xmath145 is proved by looking at the action of the diffeomorphisms on the both sides of the equality on the curves shown on the right in figure [ rel ] . | we find a new relation among right - handed dehn twists in the mapping class group of a @xmath0-holed torus for @xmath1 .
this relation induces an elliptic lefschetz pencil structure on the four - manifold @xmath2 @xmath3 with @xmath0 base points and twelve singular fibers . by blowing up the base points we get an elliptic lefschetz fibration on the complex elliptic surface @xmath4@xmath5 @xmath6 with twelve singular fibers and @xmath0 disjoint sections .
more importantly we can locate these @xmath0 sections in a kirby diagram of the induced elliptic lefschetz fibration .
the @xmath7-th power of our relation gives an explicit description for @xmath0 disjoint sections of the induced elliptic fibration on the complex elliptic surface @xmath8 for @xmath9 . |
over - the - counter ( otc ) derivatives represent a sizable fraction of financial transactions worldwide , encompassing a wide variety of contracts and asset classes . figure [ notionals_fig ]
depicts the development in the otc derivatives markets for different asset classes since 1998 .
although the increase in notionals has stopped since the peak of the financial crisis the overall growth is impressive . the contraction in some otc derivatives asset classes since the beginning of the crisis
is clearly seen from table [ notionals.table ] , which shows gross notional values in different asset classes as of june 2007 and june 2010 .
[ c]lrr asset class & 2007 & 2010 + commodity & 8,255 & 3,273 + equity linked & 9,518 & 6,868 + foreign exchange & 57,604 & 62,933 + interest rate & 381,357 & 478,093 + credit derivatives & 51,095 & 31,416 + other & 78 & 72 + total & 507,907 & 582,655 + [ notionals.table ] most bilateral otc contracts involve exchange of collateral between counterparties , to mitigate counterparty risk .
however , there is evidence that collateral requirements in bilateral credit support annexes ( csas ) lead to non - negligible residual counterparty exposures .
this is particularly the case in credit derivatives contracts where jump - to - default risk is typically not covered by bilateral collateral requirements @xcite .
in fact , the collapse of aig in september 2008 was triggered by cds collateral requirements it could not meet .
@xcite and @xcite both show that even in the extreme case of continuously collateralized cds contracts , counterparty risk is present , e.g. jump - to - default risk
. @xcite find evidence of counterparty risk in the interdealer cds markets .
@xcite finds that most large banks active in the otc derivatives market are under - collateralized .
one of the solutions proposed for mitigating counterparty risk is the central clearing of otc derivatives .
the dodd - frank wall street reform and consumer protection act of 2010 has mandated central clearing for standardized over - the - counter ( otc ) derivatives .
many classes of otc derivatives are already being partially cleared through central counterparties ( ccps ) , e.g. lch clearnet for interest rate swaps ( irss ) , ice clear and cme for credit default swaps ( cds ) , fixed income clearing corp ( ficc ) for fixed - income otc derivatives and many others .
the volume of otc derivatives cleared through central counterparties will most likely increase with the implementation of the dodd - frank act and increased capital requirements in basel 3 for non - cleared otc transactions . with a ccp , a bilateral otc derivative trade between two counterparties is replaced by two symmetric trades between the ccp and each counterparty . trading otc derivatives through a ccp has some desirable effects : in a bilateral market default of one entity can spread throughout the system leading to a chain of contagious defaults @xcite .
a ccp can break this chain of contagion .
it has been argued that clearing trades through a ccp leads to a gain in multilateral netting among market participants in the asset class being cleared , higher transparency , risk sharing among members of the clearinghouse , no duplicative monitoring , mitigation of counterparty risk as counterparties are insulated from each other s default , a reduction in commitment frictions @xcite .
moreover , by centralizing information , supervision and transparency for regulators are facilitated while retaining trade anonymity , and from a moral hazard perspective it is less problematic to bail out a ccp compared to bailing out individual banks .
while a ccp leads to higher multilateral netting it comes at the cost of reduced bilateral netting : if only a subset of otc contracts is cleared , netting _ across _
asset classes due to hedging is not accounted for , e.g. if a fixed income position is hedged with an otc interest rate derivative , separate clearing of the derivative leads to larger collateral requirements .
[ c]lrrrr dealer & forwards & options & swaps & credit + jp morgan chase & 8,422 & 10,633 & 51,221 & 7,495 + bank of america & 9,132 & 6,908 & 50,702 & 5,649 + goldman sachs & 1,631 & 6,754 & 30,958 & 6,601 +
morgan stanley & 1,127 & 3,530 & 26,112 & 6,307 + citigroup & 4,743 & 5,868 & 15,199 & 2,950 + wells fargo & 1,217 & 543 & 2,748 & 286 + hsbc & 595 & 185 & 1,565 & 913 + taunus & 667 & 20 & 162 & 144 + bank of new york & 371 & 304 & 404 & 1 + state street & 571 & 45 & 24 & 0 + total & 28,476 & 34,792 & 179,094 & 30,348 + [ indivnotionals2009.table ] [ c]lrrrr dealer & forwards & options & swaps & credit + jp morgan chase & 11,807 & 8,899 & 49,332 & 5,472 + bank of america & 10,287 & 5,848 & 43,482 & 4,367 + citigroup & 6,895 & 7,071 & 28,639 & 2,546 + goldman sachs & 3,805 & 8,568 & 27,392 & 4,233 + morgan stanley & 5,459 & 3,855 & 27,162 & 4,648 + wells fargo & 1,081 & 463 & 1,806 & 93 + hsbc & 758 & 127 & 1,901 & 700 + bank of new york & 420 & 367 & 555 & 1 + taunus & 848 & 21 & 199 & 33 + state street & 599 & 76 & 79 & 0 + total & 41,959 & 35,295 & 180,547 & 22,093 + [ indivnotionals2010.table ] tables [ indivnotionals2009.table ] and [ indivnotionals2010.table ] show the otc derivatives notionals in four different asset classes of the 10 largest us derivatives dealers
. inspection of the tables reveals a decrease in the notional of the credit derivatives class between 2009 and 2010 .
this is primarily a consequence of increased trade compression and the increased use of central clearing in the cds market @xcite : the first cds clearing facility , operated by ice clear , started its operations in 2009 and currently clears a sizable fraction of interdealer trades in index and single name corporate cdss .
comparing table [ indivnotionals2010.table ] with table [ indivnotionals2009.table ] suggests that the central clearing of cdss initiated in 2009 and the resulting compression of bilateral trades led to an overall decrease of the magnitude of dealer exposures . to understand why central clearing of otc trades can lead to such a decrease in the sum of bilateral exposures ,
consider a stylized market of four participants with bilateral exposures like shown in the left part of figure [ bilateral_fig ] . in this market
the sum of bilateral exposures amounts to 350 , taking into account bilateral netting . introducing a ccp
enables _ multilateral netting _ of the exposures ( figure [ bilateral_fig ] , right ) which reduces the total net exposure to 180 .
the above example highlights the benefits of multilateral netting through a ccp .
but there are also situations where central clearing of a single asset class may actually increase overall net exposures . if only a subset of otc contracts is cleared , netting _ across _
asset classes due to hedging is not accounted for .
for instance , if a fixed income position is hedged with an otc interest rate derivative , separate clearing of the derivative leads to larger exposure , which in return leads to larger collateral requirements .
a relevant question is thus to compare the impact of these two competing effects : * the reduction of net exposures due to multilateral netting across counterparties through central clearing , and * the increase in net bilateral exposures through the loss of netting across asset classes . in an influential paper
widely disseminated to regulators , @xcite argue that if only a subset of otc contracts is cleared in a clearinghouse , then there may be a loss of netting _ across _
asset classes which exceeds the gain in multilateral netting across market participants .
this conclusion has been interpreted by some to mean that ccps increase net exposure. [ [ contribution ] ] contribution + + + + + + + + + + + + in the present work , we reconsider the impact of central clearing of over - the - counter ( otc ) transactions on counterparty exposures in a market with otc transactions across several asset classes with heterogeneous characteristics .
we find that the impact , on total expected interdealer exposure , of introducing a central counterparty for a single class of otc derivatives is highly sensitive to assumptions on heterogeneity of asset classes in terms of ` riskyness ' of the asset class as well as correlation of exposures across asset classes .
in particular , while a model with independent , homogeneous exposures suggests that central clearing is efficient only if one has an unrealistically high number of participants , these conclusions are not robust to departures from the assumptions of independent and homogeneity . in fact , we find that the opposite conclusion is reached if differences in riskyness across asset classes are realistically taken into account .
empirically plausible specifications of these parameters lead to the conclusion that the gain from multilateral netting in a ccp overweighs the loss of netting across asset classes in bilateral netting agreements .
when a ccp exists for interest rate derivatives , adding a ccp for credit derivatives is shown to decrease overall exposures .
this result distinguishes our study from @xcite where the marginal effect of clearing credit derivatives is very limited when interest rate swaps are already being cleared .
these findings are shown to be robust to the choice of distribution for otc credit derivatives exposures .
[ [ outline ] ] outline + + + + + + + the remainder of the paper is structured as follows . section [ model_section ] introduced a stylized model of an otc market with @xmath0 different asset classes and examines how the separate or joint clearing of one or more asset classes in a ccp affects bilateral exposures .
section [ sec : heterogeneity ] emphasizes the key role played by assumptions on heterogeneity of various parameters heterogeneity of riskyness across asset classes , heterogeneity of exposure sizes across dealers and proposed a flexible parameterization to account for this heterogeneity .
section [ analysis_sec ] uses this framework to study how the introduction of ccps affects otc derivatives exposures , using three different risk measures . finally , section [ conclusion_sec ] concludes .
to compare the impact on exposures of multilateral netting across counterparties , as opposed to cross - asset netting in otc markets with bilateral transactions , we consider a market where @xmath1 participants trade over - the - counter derivatives in @xmath0 different asset classes , labeled @xmath2 .
collateral requirements , whether bilateral or in ccps , are typically based on an estimation of exposures over a horizon which we denote by @xmath3 ; the net value of the total positions at some future time point that @xmath4 owns of derivatives belonging to asset class @xmath5 with @xmath6 as the counterparty . the exposure of @xmath4 to @xmath6 in asset class @xmath5 is then defined as @xmath7 we want to compare total exposures across three configurations : 1 . the total exposure @xmath8 of company @xmath4 under bilateral netting ( no central clearing ) : @xmath9 2 .
the exposure @xmath10 assuming asset class @xmath0 is cleared through a central counterparty @xmath11 where @xmath12 denotes the fraction of asset class @xmath5 that is being cleared . hence in this case we have that , @xmath13 for @xmath14 .
the exposure @xmath15 when two asset classes @xmath16 and @xmath0 are separately cleared through a central counterparty @xmath17 where @xmath13 for @xmath18 . in equation
all netting is done bilaterally across all @xmath0 asset classes and the total net exposure of dealer @xmath4 is the sum of the net exposure with each counterparty @xmath6 .
when a ccp is introduced in the market , as in equations and , the total net exposure is the sum of the bilateral net exposures to all counterparties and the exposures to the ccps .
hence , the introduction of clearinghouses transfers exposure reductions from bilateral netting to multilateral netting .
whether the introduction of a clearinghouse increases or decreases net exposures depends on the particular market , e.g. the notional sizes of the asset classes , riskyness of the asset classes , correlation between the asset classes , the number of asset classes , the number of banks etc . in terms of dealer risk management
our analysis underestimates the risk reduction from central clearing since we put the same risk weights on bilateral dealer exposures and exposures to a ccp . in most cases
the riskyness of a ccp will be lower than that of a dealer - and hence have a smaller risk weight .
the reader is referred to @xcite for a model to quantify the ccp risk a dealer faces when participating in a ccp . in the sequel
we consider different distributional choices for the exposures , and we compare expected net exposures , value at risk ( var ) , expected shortfall and the mean of the maximum of realized exposures across dealers across various clearing scenarios .
the stylized model above requires assumptions on the joint distribution of exposures .
let us first consider the baseline case where @xmath19 are normally distributed for all @xmath20 and allowed to be correlated across asset classes . as in @xcite , we assume that the standard deviation of @xmath3 is proportional to the credit exposure @xmath21 [ c]lr asset class & exposure + commodity & 457 + equity linked & 706 + foreign exchange & 2,524 + interest rate & 17,533 + cds & 1,666 + other & 1,788 + total & 24,673 + [ exposures.table ] table [ exposures.table ] shows the gross credit exposures @xmath22 in six different asset classes as of june 2010 .
otc interest rate derivatives clearly account for the major portion of this value .
consider first a market where only one single asset class is being cleared .
in @xcite it is shown that if @xmath23 for all @xmath24 and if the exposures are uncorrelated , the number of entities required to participate in the central clearinghouse has to be equal to or greater than 461 in order to see a reduction in expected exposures : this is an unrealistically high number , so it would imply in practice that existing clearing schemes do not result in reduction of exposures .
but , altering the assumption on risk per dollar notional to e.g. @xmath25 in the same model , the picture is changed dramatically and the threshold number drops to 54 .
if furthermore we assume the correlation between the asset class exposures is , e.g. , @xmath26 , then the number of clearing members needed to achieve compression of exposures is down to ... 17 ! table [
participants.table ] reports the required number of participants for different scenarios .
figure [ clearingmembers_fig ] depicts the minimal number of members required to participate in the clearinghouse in order to see expected exposure reductions as a function of correlation and riskyness of the cleared asset class .
the figure clearly reveals that the base case considered in @xcite .
@xmath27 and @xmath28 , is a singularity : a slight deviation from these parameters alters the conclusions of the model dramatically .
[ c]lr assumptions & @xmath1 + duffie & zhu s example : & + @xmath29 & 461 + @xmath30 & + @xmath31 & 54 + @xmath32 & + @xmath31 & 17 + @xmath33 & + @xmath31 & 11 + [ participants.table ] required to participate in the ccp in order to see a reduction in the expected exposures as a function of riskyness of the asset class being cleared @xmath34 and the correlation between asset classes @xmath35.,scaledwidth=90.0% ] these observations illustrate that , even in a simple setting with ( gaussian ) identically distributed exposures , the tradeoff between netting across asset classes versus multilateral netting across market participants is highly sensitive to assumptions on the correlation of exposures and the heterogeneity of riskyness across asset classes .
conclusions based on a model with homogeneous risk across asset classes are not robust to departures from this homogeneity assumption .
we now introduce a model which allows to parametrize this heterogeneity of exposure sizes across dealers and asset classes and differentiates various asset classes in terms of their risk characteristics , as measured by the risk ( or collateral requirement ) per dollar notional . instead of credit exposures we now use notional values to determine the distribution of the @xmath3s and
assume that they are proportional to the notional values via @xmath36 where @xmath37 is a random variable with zero mean and unit variance , and @xmath38 is the total notional size of dealer @xmath4 in asset class @xmath5 . in this way
the exposure of @xmath4 to @xmath6 in class @xmath5 is a weighted fraction of company @xmath4 s total
notional in derivatives class @xmath5 , where the weight equals @xmath6 s notional size in asset class @xmath5 relative to the total notional size of the dealers in asset class @xmath5 .
we first study the case where @xmath37 is a standard normally distributed variable . in the gaussian case
we can get closed formulas for the expected net exposures @xmath39 = & \frac{1}{\sqrt{2\pi}}\sum_{j\neq i } \sqrt{\sum_{k=1}^k \sum_{m=1}^k \rho_{km } \beta_k z_i^k \frac{z_j^k}{\sum_{h\neq i } z_h^k } \beta_m z_i^m \frac{z_j^m}{\sum_{h\neq i } z_h^m}}\\ \mathbf{e}\left[e_i^1\right ] = & \frac{1 } { \sqrt{2\pi } } \sum_{j\neq i } \sqrt{\sum_{k=1}^{k } \sum_{m=1}^{k } \left ( 1 - w_{k } \right)\left ( 1 - w_{m } \right ) \rho_{km } \beta_k \frac { z_i^kz_j^k}{\sum_{h\neq i } z_h^k } \beta_m \frac { z_i^mz_j^m}{\sum_{h\neq i } z_h^m}}\nonumber\\ & + \frac{1 } { \sqrt{2\pi } } \beta_k w_k z_i^k \frac{\sqrt { \sum_{j\neq i } \left(z_j^k \right)^2}}{\sum_{h\neq i } z_h^k } \\ \mathbf{e}\left[e_i^2\right ] = & \frac{1 } { \sqrt{2\pi } } \sum_{j\neq i } \sqrt{\sum_{k=1}^{k } \sum_{m=1}^{k } \left ( 1 - w_{k } \right)\left ( 1 - w_{m } \right ) \rho_{km } \beta_k \frac{z_i^k z_j^k}{\sum_{h\neq i } z_h^k } \beta_m \frac{z_i^m z_j^m}{\sum_{h\neq i } z_h^m}}\nonumber\\ & + \frac{1 } { \sqrt{2\pi } } \beta_{k-1 } w_{k-1 } z_i^{k-1 } \frac{\sqrt{\sum_{j\neq i } \left(z_j^{k-1 } \right)^2}}{\sum_{h\neq i } z_h^{k-1 } } \nonumber\\ & + \frac{1 } { \sqrt{2\pi } } \beta_k w_k z_i^k \frac{\sqrt{\sum_{j\neq
i } \left(z_j^k \right)^2}}{\sum_{h\neq i } z_h^k } \end{aligned}\ ] ] where @xmath40 for @xmath41 .
we compute the value at risk and expected shortfalls at the 99% level by simulation as these quantities are not available in closed form .
we also assume that the value of the cds positions are given by @xmath43 where @xmath44 is a normalized @xmath42-distribution with unit variance and 3 degrees of freedom @xcite .
the other asset classes are kept normally distributed as in equation .
correlation between the classes with different distributional assumptions is introduced via a normal copula with correlation parameter @xmath35 . in the case of the @xmath42-distribution
the risk measures are not known in analytical closed form so we will use simulation in order to compute them .
another important feature which affects the analysis is the difference of asset classes in terms of riskyness .
for example , an interest rate swap and a cds with same notional and maturity typically will not lead to the the same collateral requirements , since their risk is not considered to be identical .
a simple metric for this is the exposure per dollar notional for a given asset class . to obtain an estimate of this parameter for various asset classes ,
a simple approach is to examine the standard deviation of the historical daily profit / loss , in terms of percentage of contract notional . for the cds asset class
, we estimate @xmath45 as the mean of the standard deviation of the daily profit - loss of 5-year credit default swaps on the names constituting the cdx na ig hvol series 12 in the period july 1st , 2007 to july 1st , 2009 . in the estimation
the notionals of the cdss are assumed equal to 1 and the recovery rate equal to 0.4 .
for the same period , we compute the @xmath46 for the interest rate class as the standard deviation of the historical daily profit - loss from holding a 5-year interest rate swap with notional of 1 and we use this estimate for all asset classes other than the cds class .
the estimated values are @xmath47 i.e. the risk per dollar notional is around three times higher for the cds asset class .
using the setting described in the previous section , we now compare the magnitude of inter - dealer exposures under different market clearing scenarios : 1 . no central clearinghouse .
2 . one central clearinghouse for irss .
3 . one central clearinghouse for cdss .
4 . two central clearinghouses : one for irss and one for cdss . 5 . one central clearinghouse for both irss and cdss . given that , in each of these scenarios , the total interdealer exposure is a random variable , we need a criterion for evaluating the magnitude of exposures .
@xcite focus on the _ expected exposures _ ; to check for robustness of results with respect to this choice we have computed various risk measures for the total interdealer exposure in each scenario : * expected exposure * @xmath48 quantile of the total exposure * @xmath48 tail conditional expectation these risk measures are reported below in terms of the ratio to their value under the base scenario where no clearinghouse exists .
currently some markets for otc derivatives are already being cleared . following a meeting on january 27th , 2010 at the federal reserve new york ,
14 major dealers say in a letter of march 1st , 2010 that they will aim to centrally clear @xmath49 of new and historical credit derivatives trades . in interest rate derivatives
, dealers are working to clear @xmath50 of new eligible trades .
we will use these clearing fractions in equations - when evaluating the above scenarios . in the case where one single ccp clears both irss and
cdss the classes can be considered to belong to the same class and we use equation we use the notional values reported in table [ indivnotionals2009.table ] and assume that there exist 10 identical european dealers , hence the total number of participants is 20 .
this is not a high number , at the time of writing the number of participants in swapclear at lch clearnet is 38 . for the case of the gaussian cds exposures we use the analytical formulas to compute the expected exposures and @xmath51 simulated scenarios to compute the other two risk measures .
when @xmath42-distributed exposures are involved we approximate the three measures using @xmath51 simulations and correlate the exposures across asset classes via a gaussian copula .
tables [ riskmeasures2009rhozero.table ] and [ riskmeasures2009rho01.table ] report the results for the cases where the exposures are decorrelated and correlated with @xmath52 , respectively . from table
[ riskmeasures2009rhozero.table ] , comparing the case of no central clearing in the market to the case where the otc interest rate swaps are cleared ( columns 1 , 5 and 9 ) , we see that the overall exposures are reduced significantly for almost all dealers and in all risk measures .
the notional positions of the dealers who have no reduction are very small compared to the total notionals of the 10 biggest dealers .
the rows with totals report the total expected exposure of all 20 dealers relative to the total expected exposures in the scenario 0 with no ccp .
we only compute the
total " for the expected exposures as the other two risk measures are not additive . while it is debatable whether the introduction of a clearinghouse for cdss will reduce exposures when there is no existing clearinghouse in the market ( columns 2 , 6 and 10 ) , the results are more definitive when the interest rate class is already being cleared ( columns 3 , 7 and 11 ) . for both distributional choices of the cds class , adding a ccp for
the cds class decreases the three different risk measures . depending on the assumptions made ,
the total expected exposures are decreased with additional 6 - 11% compared to the case when only interest rate swaps are cleared . in absolute values
this corresponds to reductions in the order of usd 41 - 76 billions in total expected net exposure . in the case where a single ccp clears both irss and
cdss the net exposure reductions are higher since multilateral netting is possible across both types of derivatives classes .
we also note that the difference in exposure reductions between scenarios 1 ) and 3 ) are bigger than the difference between scenarios 3 ) and 4 ) , and especially for the risk measures var and expected shortfall .
hence , the decision to clear the cds asset class is the most important , whether it is done by the same ccp that clears irss or not .
as expected , the results reported in table [ riskmeasures2009rho01.table ] confirm that in the presence of positive asset class correlation , introduction of central clearinghouses increases the benefit from clearing compared to the zero correlation case .
all the fractions are slightly smaller with a correlation parameter of @xmath52 . when the asset classes are positively correlated the exposure reductions due to bilateral netting across classes are reduced .
interestingly , we note that using @xmath42-distributed cds exposures instead of gaussian exposures does not change the conclusions . for the different scenarios , table [ maxexposures.table ] reports the average of the maximum realized net exposure across dealers , which can be interpreted as a worst case measure of dealer exposure . yet
again we observe that scenarios 1),3 ) , and 4 ) result in significant exposure reductions .
[ c]lrrrrrscenario & bilateral & irs & cds & two & joint + & & ccp & ccp & ccps & ccp + gaussian cds , @xmath27 & 136,460 & 114,340 & 138,830 & 108,110 & 103,190 + gaussian cds , @xmath52 & 145,040 & 120,440 & 144,850 & 113,130 & 109,160 +
t - distributed cds , @xmath27 & 136,020 & 111,630 & 138,480 & 107,770 & 103,070 + t - distributed cds , @xmath52 & 144,140 & 117,500 & 144,370 & 112,650 & 108,820 + [ maxexposures.table ] for the different clearinghouse scenarios , figure [ histograms_fig ] depicts the histograms of the reduction in exposures across counterparties @xmath53 where @xmath54 corresponds to the clearing scenario , in a market with 20 participants and independent gaussian exposures .
positive values correspond to cases where central clearing reduces the exposure .
hence , scenarios 1),3 ) , and 4 ) are generally giving rise to positive reductions in exposures .
in particular it is observed how the cds clearing scenarios 3 ) and 4 ) significantly reduce the tail of the distribution of differences compared to scenario 1 ) where only irs are cleared .
using a stylized model of otc exposures with multiple asset classes , we have compared the effects of bilateral netting across classes with those of multilateral netting across counterparties as allowed in a ccp .
we find that the impact , on total expected exposure , of introducing a central clearing facility for a given class of otc derivatives is highly sensitive to assumptions of heterogeneity across asset classes in terms of riskyness , as measured by the exposure per dollar notional , as well as correlation of exposures across asset classes .
the base example ( iid , homogeneous gaussian exposures ) previously considered by @xcite suggested that ccps are inefficient in reducing exposures unless one has an unrealistically high number of participants , but the opposite conclusion is reached if plausible model parameters are used to take into account differences in riskyness across asset classes .
a second finding is that , when a ccp already exists for interest rate swaps , clearing of credit derivatives decreases exposures whether it is done by a new ccp or by the same ccp clearing irss .
this result distinguishes our study from @xcite where the marginal effect of clearing credit derivatives is very limited when interest rate swaps are already cleared .
the findings are robust to distributional assumptions on ( cds ) exposures .
we find , as in @xcite , that the exposure reduction is highest if one ccp clears all asset classes .
however , this scenario would lead to a high concentration of systemic risk in the clearinghouse and also expose it to a high level of operational risk , since simultaneous clearing of different asset classes requires a more sophisticated risk management technique .
a serious examination of the benefits and drawbacks of having multiple ccps can not be done solely based on expected exposures and requires a model where systemic risk can be quantified as part of the tradeoff . | we study the impact of central clearing of over - the - counter ( otc ) transactions on counterparty exposures in a market with otc transactions across several asset classes with heterogeneous characteristics . the impact of introducing a central counterparty ( ccp ) on expected interdealer exposure is determined by the tradeoff between multilateral netting across dealers on one hand and bilateral netting across asset classes on the other hand .
we find this tradeoff to be sensitive to assumptions on heterogeneity of asset classes in terms of ` riskyness ' of the asset class as well as correlation of exposures across asset classes .
in particular , while an analysis assuming independent , homogeneous exposures suggests that central clearing is efficient only if one has an unrealistically high number of participants , the opposite conclusion is reached if differences in riskyness and correlation across asset classes are realistically taken into account .
we argue that empirically plausible specifications of model parameters lead to the conclusion that central clearing does reduce interdealer exposures : the gain from multilateral netting in a ccp overweighs the loss of netting across asset classes in bilateral netting agreements .
when a ccp exists for interest rate derivatives , adding a ccp for credit derivatives is shown to decrease overall exposures .
these findings are shown to be robust to the statistical assumptions of the model as well as the choice of risk measure used to quantify exposures . |
among many propositions for the physics beyond the standard model ( sm ) , supersymmetry ( susy ) is generally considered most elegant and respected . it can accommodate or explain some of the outstanding problems of the sm .
in fact this is the only mathematically consistent uv completion of the sm that stabilizes the gap between electroweak and planck scales .
it provides the gauge coupling unification , radiative electroweak symmetry breaking , a candidate for dark matter ( dm ) , offers new ideas on the matter - antimatter asymmetry etc .
the simplest n=1 supersymmetric extension calls for each sm particle a sparticle that differs in spin by half a unit .
the lagrangian must be supplemented by susy breaking terms that keep unseen superpartners out of the current experimental reach while retaining all goodies of the model : renormalizability and perturbatively stable hierarchy of scales . with ever improving experimental constraints on susy breaking parameters , mainly from flavor and higgs physics ,
the model building of successful susy breaking scenarios becomes more and more difficult .
however , the successes of susy do not rely on its simplest realization .
in fact , non - minimal realizations may ameliorate some of the above problems .
for example , dirac gauginos ( in contrast to majorana in n=1 ) forbid some of the couplings and may lead to additional suppression of loop contributions with gauginos in flavor - changing processes .
such scenarios can be based on d - term supersymmetry breaking models @xcite or continuous r - symmetries @xcite .
a dirac gaugino requires additional fermionic degrees of freedom .
they can be provided by adding chiral super - multiplets in the adjoint representations of the corresponding gauge group , as realised in n=2 susy @xcite .
when extending to n=2 , the additional supersymmetry also requires the introduction of mirror matter superfields . in order to avoid chirality problems the n=2 mirror ( s)fermions have to be assumed to be heavy .
thus effectively we consider a hybrid model that expands to n=2 only in the gaugino sector @xcite . here
we will present elements of the phenomenology of the scalar partner of the dirac gluino ( sgluon ) , as worked out in a recent paper @xcite ; see also @xcite . for an independent analysis of sgluons at the lhc
, we refer to ref .
if gluinos are seen at the lhc one of the main experimental goals will be to verify if they are majorana or dirac fermions . to address properly this issue a theoretical model
is needed that allows a smooth dirac / majorana transition , as for example the hybrid model constructed in ref .
@xcite . in n=1 gluinos are majorana fields with two degrees of freedom to match gluons in the color - octet vector super - multiplet .
to provide the two additional degrees of freedom for dirac fields , the usual n=1 gluon / gluino @xmath0 super - multiplet @xmath1 ( @xmath2 ) is supplemented by an additional n=1 color - octet chiral super - multiplet @xmath3 of extra gluinos @xmath4 and scalar @xmath5 fields to form a vector hyper - multiplet of n=2 susy .
similarly , the electroweak sector , not to be discussed here , is supplemented by additional su(2)@xmath6 and u(1)@xmath7 super - multiplets @xcite .
standard @xmath8 and new gluinos @xmath9 couple minimally to the gluon field @xmath10 + \overline{\tilde{g } ' } \gamma^{\mu } [ { g}_{\mu } , { \tilde{g } ' } ] ) \ , , \label{eq : gluino}\ ] ] as required by the gauge symmetry . here
@xmath11 denotes the qcd coupling , the fields @xmath12 are color - octet matrices ( for the gluon @xmath13 with the gell - mann matrices @xmath14 ) ; @xmath8 and @xmath9 are two 4-component majorana spinor fields .
quark and squark fields interact only with the standard gluino , @xmath15 since only their mirror partners ( assumed to be heavy ) couple to @xmath9 , as required by n=2 susy .
soft supersymmetry breaking generates masses for the gluino fields @xmath8 and @xmath9 , @xmath16 diagonal terms are induced by the individual majorana mass parameters @xmath17 and @xmath18 while an off - diagonal term corresponds to the dirac mass .
it can arise from the d term , @xmath19 when a hidden sector spurion field @xmath20 gets a vacuum expectation value , @xmath21 , the dirac mass is generated , @xmath22 .
diagonalization of ( [ eq : gluinomass ] ) gives rise to two majorana mass eigenstates , @xmath23 and @xmath24 , with masses @xmath25 and @xmath26 .
there are two limiting cases of interest : in the limit @xmath27 the standard majorana gluino is recovered ; in the limit of vanishing majorana mass parameters @xmath17 and @xmath18 with @xmath28 , the mixing is maximal and the two majorana gluino states are degenerate .
one can draw a smooth path interpolating these two limiting cases by introducing an auxiliary parameter @xmath29 $ ] and parameterizing @xmath30 , @xmath31 . with @xmath32
kept fixed , the parameter @xmath33 allows for a continuous transition from @xmath34 , where the standard limit is reached with one majorana gluino ( the second being infinitely heavy ) , to @xmath35 that corresponds to two degenerate majorana fields combined to a dirac gluino @xmath36 the phenomenology of dirac gluinos is characteristically different from majorana .
the detailed discussion can be found in ref .
@xcite , from which we borrow an example for illustration .
[ fig : d / m ] shows the partonic cross sections @xmath37 ( left ) and @xmath38 ( right ) for different - flavor squark production , in both cases mediated by the gluino @xmath39-channel exchange , plotted as a function of a dirac / majorana control parameter @xmath33 ( for partonic center - of - mass energy @xmath40 gev , @xmath41 gev and @xmath42 gev ) . we see that @xmath43 $ ] and @xmath44 are non - zero in the majorana limit but vanish in the dirac , while @xmath45=\sigma[q \bar q ' \to { \ensuremath{\tilde{q}}}_r { \ensuremath{\tilde{q}}}_r'^\ast]$ ] and @xmath46 are non - zero for any @xmath33 and reach the same values in both limits . for equal - flavor quark - antiquark scattering
the additional gluino @xmath47-channel exchange must be added to the @xmath39-channel exchange diagrams .
the gluinos @xmath4 are accompanied by a color - octet complex scalar field @xmath5 . in parallel to the case of degenerate majorana gluinos combined to the dirac , the real and imaginary components of the scalar field @xmath48 will be assumed degenerate with the sgluon mass denoted by @xmath49 @xcite .
although sgluons are r - parity even , they can not be singly produced at tree level in gluon - gluon or quark - antiquark collision .
this is so because they couple only in pairs to gluons , @xmath50 and @xmath51 , as required by the gauge symmetry , and singly only to dirac gluino pairs via the yukawa - type gauge interaction @xmath52 where @xmath53 are the su(3)@xmath54 structure constants .
it is interesting to note that even the loop - induced @xmath55 coupling due to gluino exchange vanish as a consequence of bose symmetry since the coupling is even in momentum space but odd , @xmath56 , in color space . ) , while gluons always couple to the same pair , eq .
( [ eq : gluino ] ) , the coupling of the octet sgluon to any number of gluons via the gluino loop is forbidden . ]
when susy is broken spontaneously , the dirac gluino mass generates , via the super - qcd @xmath57 term , a scalar coupling between @xmath48 and squark pair @xcite @xmath58 note that the @xmath59 and @xmath60 squarks contribute with opposite signs . since squarks couple to gluons and quarks , the loop diagrams with squark / gluino ( 300,60)(0,70 ) ( 0,95)[r]@xmath61 ( 15,95)[c ] @xmath48 ( 100,95)[c ] @xmath62 ( 25,95)(50,95)3 ( 50,95)(90,115)3 ( 90,75)(50,95)3 ( 90,115)(90,75)3 ( 125,115)[l ] @xmath63 ( 125,75)[l ] @xmath63 ( 90,115)(120,115)25 ( 90,75)(120,75)25 ( 150,95)[r ] @xmath48 ( 200,100)[c ] @xmath62 ( 155,95)(180,95)3 ( 200,95)(20,180,0)3 ( 200,95)(20,0,180)3 ( 220,95)(250,115)24 ( 220,95)(250,75)24 ( 255,115)[l ] @xmath63 ( 255,75)[l ] @xmath63 ( 300,60)(0,70 ) ( 0,95)[r]@xmath64 ( 15,95)[r ] @xmath48 ( 70,115)[c ] @xmath65 ( 95,95)[l ] @xmath62 ( 25,95)(50,95)3 ( 50,95)(90,115 ) ( 90,75)(50,95 ) ( 90,75)(90,115)3 ( 125,115)[l ] @xmath66 ( 125,75)[l ] @xmath67 ( 90,115)(120,115 ) ( 120,75)(90,75 ) ( 150,95)[r ] @xmath48 ( 200,115)[c ] @xmath62 ( 225,98)[l ] @xmath65 ( 155,95)(180,95)3 ( 180,95)(220,115)3 ( 220,75)(180,95)3 ( 220,75)(220,115 ) ( 220,115)(250,115 ) ( 250,75)(220,75 ) ( 255,115)[l ] @xmath66 ( 255,75)[l ] @xmath67 lines , fig . [
fig : loops ] , generate @xmath68 and @xmath69 couplings .
the interaction lagrangian however , eq . ( [ eq : sqq ] )
, implies that all @xmath59- and @xmath60-squark contributions to the couplings come with opposite signs so that they cancel each other for mass degenerate squarks .
in addition , the quark - antiquark coupling is suppressed by the quark mass as evident from general chirality rules .
since sgluons have a large color charge they might be more copiously produced than squarks at the lhc .
in addition they have quite spectacular decay modes , as discussed below .
sgluons can decay into a variety of different channels that include gluinos , squarks , gluons and quarks . at tree level
the @xmath48 particles can decay only to a pair of dirac gluinos @xmath70 or into a pair of squarks ( if kinematically open ) , @xmath71 & = & \frac{3 \alpha_s m_{\sigma}}{4 } \beta_{\tilde{g}}\ , ( 1+\beta^2_{\tilde{g}})\ , , \\ \gamma [ \sigma \to \tilde{q}_a \tilde{q}_a^ * ] & = & \frac { \alpha_s |m_3^d|^2 } { 4 m_\sigma } \beta_{\tilde{q}_a}\ , , \label{eq : tree2}\end{aligned}\ ] ] where @xmath72 are the velocities of @xmath73 ( @xmath74 ) . if @xmath75 , one or both of these sparticles can be virtual .
squarks and gluinos will further cascade to sm particles and lsp .
decays into gluon or quark - antiquark pairs can proceed only via loop - induced couplings . in the case of no @xmath76 squark mixing and @xmath77 of a given flavor mix
, the subscripts @xmath78 have to be replaced by @xmath79 labeling the squark mass eigenstates , and the contribution from this flavor is suppressed by the factor @xmath80 . ]
the decay widths take the form @xmath81|^2\,,\\ \gamma(\sigma \rightarrow q \bar q ) & = & \frac{9 \alpha_s^3 } { 128 \pi^2 } \frac{|m_3^d|^2 m_q^2}{m_\sigma}\ , \beta_q [ ( m^2_\sigma-4 m_q^2 ) |{\cal i}_s|^2 + m^2_\sigma \
, |{\cal i}_p|^2 ] \ , , \label{loop - qq}\end{aligned}\ ] ] where @xmath82 and @xmath83 is the standard function from a squark circulating in the loop @xcite , while the effective scalar ( @xmath84 ) and pseudoscalar ( @xmath85 ) @xmath86 couplings are given by ( @xmath87 ) @xmath88 \,.\end{aligned}\ ] ] in the above expression @xmath89 , @xmath90 , @xmath91 , @xmath92 , and the squark / gluino denominators are ( @xmath93 ) + @xmath94 , + @xmath95 .
+ because of the chirality , decays to quark - antiquark pairs are suppressed by the quark mass .
therefore the decays into top quark pairs will dominate .
in addition both loop - induced decays @xmath96 and @xmath97 are absent if @xmath59 and @xmath60 squarks are degenerate .
therefore squarks with substantial mixing ( mostly top squarks ) will contribute to the decay widths .
the hierarchy between the tree - level and loop - induced decay modes depends on the values of various soft breaking parameters .
level , two
body decays of eqs .
( [ eq : tree1],[eq : tree2 ] ) will dominate if they are kinematically allowed .
well above all thresholds the partial width into gluinos always dominates : it grows @xmath98 since the supersymmetric @xmath99 coupling is dimensionless , while the partial width into squarks asymptotically scales like @xmath100 due to dimensionfull susy - breaking @xmath101 coupling .
when the above decay channels are shut kinematically , even for small @xmath59-@xmath60 squark mass splitting the loop - induced sgluon decays into two gluons , and to a @xmath102 pair if kinematically allowed , always dominate over tree - level off - shell four body decays @xmath103 and @xmath104 . increasing the gluino mass increases the @xmath105 coupling . as a result ,
the partial width into two gluons ( due to pure squark loops ) increases , while the @xmath102 partial width ( due to mixed squark gluino loops ) decreases since the increase of the @xmath106 couplings is over compensated by the gluino mass dependence of the propagators .
the above qualitative features are verified numerically in fig .
[ fig : br ] , where the branching ratios for @xmath48 decays are shown for two different squark / gluino mass hierarchies . plots are done for moderate mass splitting between the @xmath59 and @xmath60 squarks of the five light flavors , and somewhat greater for soft breaking @xmath107 masses : @xmath108 with the off diagonal element of the squared @xmath107 mass matrix @xmath109 ; the gluino is taken to be a pure dirac state , i.e. @xmath110 . the signatures for single @xmath48 production are potentially very exciting
. they can be produced singly in gluon - gluon collisions via squark loops since the production via @xmath111 annihilation is negligible for light incoming quarks .
the partonic cross section is given by @xmath112 = \frac{\pi^2 } { m^3_\sigma } \gamma(\sigma \to gg)\,,\ ] ] where the partial width for @xmath113 decays is given in eq.@xmath114([loop - gg ] ) .
the expected cross section for single @xmath48 production can be quite large at the designed lhc energy , of the order 100 fb .
its dependence on the sgluon mass is shown by the curves ( c ) and ( d ) in fig.@xmath114[fig : lhcsigma ] ( with the lo cteq6l parton densities @xcite ) . in the scenario of the right frame of fig .
[ fig : br ] , the single @xmath48 cross section can exceed the @xmath48-pair production cross section for @xmath115 tev ( solid curve ( c ) ) ; the dashed one ( d ) is for the benchmark point sps1a@xmath116 @xcite with the gluino mass interpreted as a dirac mass .
since the latter scenario has a somewhat smaller gluino mass , it generally leads to smaller cross sections for single @xmath48 production .
it would be very exciting to observe the sgluon as an @xmath47-channel resonance .
this however will not be easy since ( a ) the 2-gluon decay channel must be discriminated from large background , ( b ) large production cross section in the gluon fusion implies diminished decay rates to other channels , which in addition do not allow a direct reconstruction of @xmath49 .
detailed experimental simulations are needed to see if the single @xmath48 production can be detected as a resonance above the sm and susy backgrounds .
sgluons can also be pair - produced in @xmath111 and @xmath117 processes at tree - level , @xmath118 & = & \frac{4 \pi \alpha_s^2}{9s } \,\beta^3_\sigma \ , , \\
\sigma [ gg \to \sigma\sigma^{\ast } ] & = & \frac{15 \pi \alpha_s^2\beta_\sigma}{8s } [ 1 + \frac{34}{5}\ , \frac{m_\sigma^2}{s}-\frac{24}{5 } ( 1-\frac{m_\sigma^2}{s})\frac{m_\sigma^2}{s}\ , l_\sigma ] \,,\end{aligned}\ ] ] where @xmath119 is the invariant parton - parton energy , @xmath120 and @xmath121 are the mass and center - of - mass velocity of the @xmath48 particle , and @xmath122 .
the cross section for @xmath48-pair production at lhc , @xmath123 , is shown by the solid curve ( a ) in fig.@xmath114[fig : lhcsigma ] for the @xmath48-mass range between 500 gev and 2 tev . with values from several picobarn downwards ,
a sizable @xmath124 event rate can be generated .
as expected , due to large color charge of the sgluon , the @xmath124 cross section exceeds stop or sbottom - pair production ( dashed line ( b ) ) , mediated by a set of topologically equivalent feynman diagrams , by more than an order of magnitude . with the exception of @xmath125 ,
all the @xmath48 decays give rise to signatures that should easily be detectable .
most spectacular ones come from @xmath126 decay followed by the gluino decays giving at the end at least four hard jets and two invisible neutralinos as lsp s .
the @xmath48-pair production then generates final states with a minimum of four lsp s and eight jets with transverse momenta distributions markedly different from the corresponding standard gluino or squark production with the same mass configurations @xcite .
other interesting final states are four top squarks @xmath127 , which can dominate if @xmath128 and @xmath59-@xmath60 mixing is significant in the stop sector , and @xmath129 if @xmath130 .
they also give rise to four lsp s in the final state and a large number of hard jets .
the @xmath131 and @xmath132 final states , which can be the dominant modes if the two body decays into squarks and gluinos are kinematically shut , might also allow the direct kinematic reconstruction of @xmath49 .
preliminary analyses of the @xmath131 channel show that pairing the four hardest jets in the central region of rapidity with no missing @xmath133 could reveal the signal and provide a means to reconstruct the @xmath48 mass directly @xcite .
in addition , the observation of @xmath134 final states would signal a substantial mixing in the up - type squark sector .
models with dirac gauginos offer an interesting alternative to the standard susy scenario .
they predict the existence of scalar particles in the adjoint representation of the gauge groups .
the color - octet scalars , sgluons , can be copiously produced at the lhc since they carry large color charge .
their signatures at the lhc are distinctly different from the usual topologies .
depending on the assumed scenario of masses and susy breaking parameters , either multi - jet final states with high sphericity and large missing transverse momentum are predicted , or four top quarks and/or four gluon jets should be observed in @xmath135 production . in the latter
the sgluon mass could be reconstructed directly .
if the mass splitting between @xmath59 and @xmath60 squarks is not too small , loop induced single @xmath48 production may also have a sizable cross section .
it remains to be checked by detailed experimental simulations whether such a single resonant @xmath48 production is detectable on the large sm and susy background .
we thank peter zerwas and manuel drees for a fruitful collaboration . work supported in parts by the korea research foundation grant krf-2008 - 521-c00069 , the polish ministry of science and higher education grant n n202 230337 , and the ec programmes mrtn - ct- 2006 - 035505 `` tools and precision calculations for physics discoveries at collider '' and mtkd
ct2005029466 `` particle physics and cosmology : the interface '' .
jk is grateful to the theory division for the hospitality extended to him at cern .
100 p. j. fox , a. e. nelson and n. weiner , jhep * 0208 * ( 2002 ) 035 .
y. nomura , d. poland and b. tweedie , nucl .
b * 745 * , 29 ( 2006 ) .
l. j. hall and l. randall , nucl .
b * 352 * , 289 ( 1991 ) ; g. d. kribs , e. poppitz and n. weiner , phys . rev .
d * 78 * ( 2008 ) 055010 .
p. fayet , nucl .
b * 113 * ( 1976 ) 135 , b * 149 * ( 1979 ) 137 ; f. del aguila , m. dugan , b. grinstein , l. j. hall , g. g. ross and p. c. west , nucl .
b * 250 * ( 1985 ) 225 ; l. lvarez - gaum and s. f. hassan , fortsch .
* 45 * ( 1997 ) 159 .
choi , m. drees , a. freitas and p.m. zerwas , phy .
d * 78 * ( 2008 ) 095007 .
s. y. choi , m. drees , j. kalinowski , j. m. kim , e. popenda and p. m. zerwas , phys .
b * 672 * ( 2009 ) 246 .
s. y. choi , m. drees , j. kalinowski , j. m. kim , e. popenda and p. m. zerwas , acta phys .
b * 40 * ( 2009 ) 1947 .
j. m. kim , talk at the 15th int .
symposium on particles , strings and cosmology pascos2009 , desy , hamburg .
e. popenda , talk at the grd terascale workshop , heidelberg 2009 . for a recent discussion of dark matter with dirac and majorana neutralinos ,
see g. belanger , k. benakli , m. goodsell , c. moura and a. pukhov , jcap * 0908 * ( 2009 ) 027 .
gunion and h.e .
haber , nucl .
b278 * ( 1986 ) 449 .
j. pumplin , d. r. stump , j. huston , h. l. lai , p. m. nadolsky and w. k. tung , jhep * 0207 * ( 2002 ) 012 .
j. a. aguilar - saavedra _ et al .
_ , eur .
j. c * 46 * ( 2006 ) 43 .
a. renaud , _ sgluons at the lhc : a first look _ , talk at the grd terascale workshop , heidelberg 2009 . see also c. kilic , s. schumann and m. son , jhep * 0904 * ( 2009 ) 128 . | the hybrid n=1/n=2 supersymmetric model predicts scalar gluons ( sgluons ) as susy partners of the dirac gluino .
their strikingly distinct phenomenology at the cern large hadron collider is discussed . |
to satisfy the increasingly higher data rate demands , visible light communication ( vlc ) has drawn tremendous interest from both industry and academia as a promising complement to traditional radio frequency communication ( rfc ) that suffers from spectrum saturation @xcite .
the maturing of led manufacturing techniques during the recent decade largely boosts the trend of replacing traditional lighting systems with led alternatives for both indoor and outdoor illumination purposes , and the resulting infrastructures are ready for deployment of vlc .
it is a low - cost technology where one can use the simple intensity modulation and direct detection ( im / dd ) techniques .
in addition , one can enjoy a bunch of additional advantages such as eye - safety , high security and causing no electromagnetic inference .
vlc is both unique from and similar to rfc . with regard to the uniqueness of vlc
, it only allows positive and real signals to drive the leds as intensities ( a non - informative dc - bias is typically used ) ; its channel especially for indoor environment is much more slower varying than a rfc counterpart ; baseband waveforms modulates the leds directly instead of being up - converted first , etc . as for the similarity between vlc and rfc ,
many existing rfc techniques can be applied , although possibly with non - straightforward modifications to vlc systems , e.g. optical multiple input multiple output ( o - mimo ) @xcite , optical orthogonal frequency division multiplexing ( o - ofdm ) @xcite , and other advanced signal processing techniques @xcite .
these ( and relevant ) works nicely take advantage of various diversities a vlc system provides , such as spatial diversity , frequency diversity , color diversity and adaptive dc - bias , to improve system performance .
it is worth noting that the color diversity and adaptive dc - bias configuration are specific to vlc .
the motivation behind this work is to exploit the benefits of various diversities jointly for a very power efficient vlc , while the focus of this paper is on the problem of constellation design in high dimensional space .
this space is formed by several dimensions of freedoms including adaptive dc - bias , baseband subcarriers and multiple wavelengths corresponding to r / g / b led lights . according to the fundamental idea that spheres ( i.e. , constellation points ) can pack more compactly in a higher dimensional space
, a constellation with larger minimum euclidean distance ( med ) can be expected in a higher dimensional space .. this med maximization problem is formulated in a non - convex optimization form , and is then relaxed to a convex optimization problem by a linear approximation method .
key practical lighting requirements are taken into account as constraints , e.g. , the optical power constraint , average color constraint , non - negative intensity constraint , color rendering index ( cri ) and luminous efficacy rate ( ler ) requirements @xcite . for rfc ,
one well - known shortcoming with utilizing multiple subcarriers is the excessive peak - to - average power ratio ( papr ) problem introduced , which can cause severe nonlinear distortion to degrade system performance .
plenty of methods have been proposed to reduce papr ( see @xcite and references therein ) .
in fact , when using multiple subcarriers for vlc , high papr is also a very severe issue , due to the limited linear dynamic range of amplifiers and leds .
this paper shows that such distortion can be avoided by formulating the dynamic range requirement as ( convex ) constraints of the optimization problem . in such way distortion control
becomes a offline process or a by - product of constellation design .
the remainder of this paper is organized as follows . in section
ii , we first provide an overview of dc - informative modulation schemes for optical communications .
dc - informative multicarrier modulation is introduced as a power efficient alternative to traditional non - dc - informative optical ofdm configurations . in section iii
, we propose the dci - jcfm method for systems with rgb led .
the key lighting constraints including dynamic range control are discussed .
the cases of `` balanced '' , `` unbalanced '' and `` very unbalanced '' systems are introduced . in section iv , we discuss the pros and cons of using dynamic range constraint , short time papr constraint and long time papr constraint . in section
v , we provide simulation results to verify the significant performance gains of the proposed method over a decoupled method for balance , unbalance , and very unbalanced systems .
finally , section vi provides conclusions .
optical communications based on im / dd has a unique feature of requiring all signals modulating the leds to be positive ( and real ) , so multiple schemes are proposed accordingly such as the well - know asymmetrically clipped optical ofdm ( aco - ofdm ) , dc - biased optical ofdm ( dco - ofdm ) and optical multisubcarrier modulation ( msm ) .
these schemes all discard the dc - bias at the receiver , which causes significant power loss .
the dc - informative modulation schemes were then proposed such that @xmath0 optical power is used for data transmission ( see @xcite for single carrier case and @xcite for multiple carrier selective fading case ) . to be specific , consider the channel model @xmath1 , \label{1}\ ] ] where @xmath2 is a symbol waveform mapped from @xmath3 that contains @xmath4 bits of information , @xmath5 denotes the convolution operator , @xmath6 is either flat - fading or selective - fading channel , @xmath7 denotes white noise , @xmath8 is the received signal , @xmath9 and @xmath10 are electrical - to - optical and optical - to - electrical conversion factors respectively , with out loss of generality ( w.o.l.g ) . ] and @xmath11 stands for the constellation size .
the key feature of a dc - informative modulation is that the following basis are used _
jointly _ to carry information @xmath12 @xmath13 @xmath14 where both i and q channels are used .
@xmath15 is the dc - bias basis , @xmath16 is the symbol interval , @xmath17 is the @xmath18-th subcarrier , @xmath19 is the total number of subcarriers , and a rectangular pulse - shaper is used @xmath20 the relationship between symbol waveforms @xmath2 and constellation points @xmath21 $ ] is @xmath22 and one of the reasonable goals is to minimize the ser subject to fixed electrical / optical power by properly design the constellation matrix @xmath23 where each column of @xmath24 is a constellation point , and the med of all columns should be maximized to reach the goal .
we typically stack the columns into a single vector instead , i.e. @xmath25^t,\ ] ] for simplification of the formulation of the optimization problem discussed in section ii .
based on the idea discussed in section ii , we propose two constellation design methods taking advantage of the informative dc - bias for a visible light communication system employing one rgb led as shown by fig.1 . if the inputs into r / g / b carry independent bit information as shown by fig.1(a ) , it is termed a decoupled scheme . in comparison ,
if the inputs into r / g / b modulators only carry information jointly , it is a joint scheme instead as shown by fig.1(b ) . in other words , although for a joint scheme still three modulators are used to create continuous domain waveforms to drive corresponding leds , information can not be estimated though recovery of a single ( or any pair ) of them .
it is observed that the joint scheme may utilize four types of diversities during per channel use , including frequency diversity , color(wavelength ) diversity , adaptive dc , and spatial diversity .
spatial diversity can be achieved by extending from employing only one rgb led to include @xmath26 ones , which is out of scope of this paper .
we only emphasize the first three diversities here .
this scheme is termed dc - informative joint color - frequency modulation ( dci - jcfm ) .
the decoupled system shown in fig.1(a ) works as follows : at the transmitter - side three independent bit stream @xmath27~i\in[1,n_c]$ ] , of length @xmath28 are mapped to corresponding constellation points @xmath29 of size @xmath30 first , which are modulated separately to generate continuous symbol waveform ( current ) @xmath31 by for each tunnel .
if cross - talks exist for any tunnel , a corresponding precoder @xmath32 needs to be applied before modulation .
waveforms @xmath31 are then electrical - to - optical converted to intensity signals @xmath33 to drive the leds . at the receiver - side ,
photo detectors of each tunnel collect the waveforms ( convoluted with channel and corrupted by noise ) .
the received signal is sent through red , green , and blue color filters respectively and after optical - to - electrical conversion waveforms @xmath34 are obtained .
then @xmath35 matched filters are employed for each tunnel to demodulate @xmath34 to obtain signal vector @xmath36 .
three symbol detectors follow to provide @xmath37 , estimates of the symbol vectors , which are de - mapped separately and the estimate of original bit sequences @xmath38 are finally obtained .
if cross - talks exist , post - equalizers @xmath39 are applied before the symbol detectors .
system using dci - jcfm as shown by fig.1(b ) works differently . a joint bit sequence @xmath40^t,\ ] ] is firstly mapped jointly to a constellation point @xmath41 of size @xmath42 .
then @xmath41 is converted by a joint modulator to the continuous domain to obtain @xmath43 through @xmath44 if cross - talks exist , a joint precoder @xmath45 is applied before modulation .
also we observe the expectation of @xmath46 as follows @xmath47&=\sum_{p=1}^3 s_{(p-1)(2k+1)+1,i}\phi_1(t),\end{aligned}\ ] ] since all non - dc basis has zero time averages .
therefore , both the average optical power and average color of system are determined solely by these three adaptive dc - bias . while the dynamic range of waveform , instead , is influenced by all subcarriers of all leds . for our design
, we will demonstrate with a line - of - sight ( los ) scenario when channel has cross - talks , due to the imperfectness of receiver color filters .
the discrete channel model can be written as follows @xcite @xmath48 where @xmath49 $ ] is termed the `` cross - talk index ( ci ) '' and @xmath50 . with working snrs for vlc which are typically medium - to - high , the minimum euclidean distance between constellation pairs governs ser .
therefore , we seek to minimize the system ser by maximizing the minimum euclidean distance , through carefully design the constellation vector @xmath51 subject to key lighting constraints . for a constellation containing @xmath52 points , distances of a total of @xmath53 pairs have to be constrained as follows @xcite @xmath54 where we define @xmath55 and @xmath56 where @xmath57 denotes kronecker product , @xmath58 is the @xmath59-th column of identity matrix @xmath60 , and @xmath61 where @xmath62 .
the distance constraints are nonconvex in @xmath51 .
we choose to use the follows linear approximation at point @xmath63 @xmath64 for our design problem , practical lighting issue considered include average optical power , average illumination color , ler and cri , non - negative intensity , and flickering - free requirements .
the first three requirements can be constraint using only one equation written as follow @xmath65 where @xmath66 is the average optical power of a rgb led , @xmath67 is a selection matrix ( containing only ones and zeros ) adding up r / g / b components in @xmath51 respectively by a multiplication of each row with it , @xmath68^t$ ] is termed the average color ratio vector and the follow equation holds @xmath69 thus the optical power and illumination color requirements are constrained together .
the luminous efficacy rate and color rendering index requirements can be satisfied by properly choosing @xmath70 . since the linear dynamic range of leds
are limited , the ranges of signal for each led have to be constrained to avoid nonlinear distortion , i.e. @xmath71 where @xmath72 is the highest current level and for simplicity we have assumed that red , green and blue leds have the same dynamic range .
we propose to constrain dynamic range of a sequence of sampled signal @xmath73 and @xmath74 is picked as @xmath75 where @xmath76 is the oversampling rate , @xmath77 , and @xmath78 is the total number of sample points .
it should be noted that although does not guarantee , which means the continuous signal waveforms designed subject to could result in negative amplitudes in between the sample instances , the negative peak is very small compare to the dynamic range of signal .
we can compensate this effect by adding a small post dc - bias after obtaining an optimized constellation .
therefore , we can formulate this point - wise dynamic range constraints as follows @xmath79 @xmath80 where @xmath81 selects the @xmath82-th constellation point and @xmath83 selects the corresponding coefficients for color @xmath84 , @xmath85^t$ ] , @xmath86 , @xmath87 , and @xmath88 . we first formulate the optimization problem when there is no cross - talks among different colored leds , i.e. @xmath89 , as follows @xmath90 which is convex in @xmath51 and @xmath91 , and specialized solver such as cvx toolbox for matlab can be utilized @xcite .
start from initial point @xmath63 , the scheme can iteratively converge to a local optima with each run .
the best constellation is chosen from local optimal constellation obtained from multiple runs .
when the channel suffers from cross - talks , we choose to deal with it by employing the well - known singular value decomposition ( svd ) based pre - equalizer @xmath92 and post - equalizer @xmath93 for our system , where @xmath94 .
constellation is designed by an optimization with transformed constraints as follow @xmath95 where @xmath96 is defined is the kronecker product . ] and apparently this optimization is convex as well . for the decoupled system ,
three independent problems can be formulated to find the meds for each color , i.e. @xmath97 where @xmath98 , @xmath99 , @xmath100 , @xmath101 , @xmath102 , and @xmath103 are defined in a similar manner with corresponding parameters in .
for brevity we omit the explicit definitions .
in fact , although a hard constraint on the dynamic range of symbol waveforms can help avoid non - linear distortion completely , it may bring with side effects such as excessive power efficiency decrease .
this is particularly true if only one or few symbol waveforms have notably larger dynamic range than the majority . in such case
, one can consider using certain papr constraint to replace the dynamic range constraint .
in other words , there is a tradeoff between allowable papr and power efficiency .
two types of papr constraints ( for each led light ) can be considered .
one is the so - called long - term papr ( l - papr ) , i.e. the ratio of the peak power of all waveforms and the time average of them . assuming no cross - talks , the l - papr constraint for led @xmath84
can be written as @xmath104 ^ 2 } { \mathbf{s}_j^t\mathbf{s}_j / n_c}\leq \beta_x\notag\\ & = \frac{n_c[\max_{i , n}(\mathbf{u}_n^t\mathbf{k}_x\mathbf{j}_i\mathbf{s}_j)]^2 } { \mathbf{s}_j^t\mathbf{s}_j}\leq\beta_x,\label{27}\end{aligned}\ ] ] where @xmath105 is the required l - papr for led @xmath84 .
thus , a set of constraints can be formulated as follows @xmath106 which is non - convex in @xmath51 .
a way to deal with this is to use a similar linear approximation as in at the same initial point @xmath63 .
the above constraints are thus transformed as follows @xmath107 the other is the individual papr ( i - papr ) , i.e. the ratio of the peak power of each waveform and the average power of it .
the i - papr for @xmath82-th waveform for led @xmath84 can be written as @xmath108 ^ 2 } { \mathbf{s}_j^t\mathbf{j}_i^t\mathbf{j}_i\mathbf{s}_j}\leq \beta_{x , i}\end{aligned}\ ] ] the corresponding constraint can be written as @xmath109 and a similar linear approximation process is applied to convert them to convex constraints .
to the best of our knowledge , there is no comprehensive comparison on performance of systems applying those three constraints available so far .
an interesting observation recently in @xcite shows that the non - linearity mitigation for an im / dd vlc is a more involved problem than expected .
the reason is that the low part of the baseband frequencies are causing larger non - linear distortion than the higher part .
this effect , if taken into account along with the three constraints discussed above , is expected to make the design problem even more worthwhile to look into .
in this section , we compare the performances of the decoupled scheme and dci - jcfm by assessing the maximum med and bit error rate ( ber ) under different channel cross - talk and color illumination assumptions . each constellation point
is assumed to have equal probability of transmission , and the union bound for ser of both the scheme can be written as ( * ? ? ?
* eq.25 ) @xmath110 where @xmath111 is the number of neighbor constellation pairs @xcite and @xmath112 denotes the gaussian q - function , @xmath113}$ ] stands for the med of the dci - jcfm and meds for decoupled schemes .
the bit error rate is thus calculated as @xmath114 where @xmath115 is the number of wrongly detected bits in each bit sequence , which can be minimized by employing the bsa mapper .
we first compare the dci - jcfm and the decoupled scheme when channel cross - talks do not exist . to guarantee a fair comparison ,
the following system parameters are chosen : the number of monte - carlo runs for each scheme @xmath116 , the length of bit sequence for each channel use is @xmath117 for dci - jcfm and @xmath118 respectively for each tunnel of the decoupled system , the number of subcarriers for each led is @xmath119 , the average optical power @xmath120 , the symbol interval @xmath121 is used is chosen , since the design is rate independent . ] , the upper bound of waveform amplitude @xmath122 , the average color ratio vector for a balanced system @xmath123^t,\ ] ] for an unbalanced system @xmath124^t,\ ] ] and for a very unbalanced system @xmath125^t.\ ] ] the meds of the two schemes for three systems obtained through picking the best constellatio from the @xmath126 local optimums are summarized by table i. from table i , key observations include : a. with the dci - jcfm , the `` joint med '' is much larger than the r / g / b `` decoupled meds '' , except for the cases with very unbalanced illumination . while for the very unbalanced case the blue and green tunnels could suffer from severe performance loss with the small meds , and therefore the dci - jcfm is still expected to work better .
b. larger meds are obtained with an increased number of subcarriers .
we are only listing the cases when @xmath127 and @xmath128 for brevity , while we have observed through additional simulations that this gain continue to grow with @xmath19 . c. the more balanced a system is
, the better performance is expected .
.med comparison with no cross - talk , dci - jcfm ( row 1@xmath1292 ) vs decoupled ( row 3@xmath1294 ) .
[ cols="^,^,^,^,^,^,^,^",options="header " , ] from table ii , key observations include : a. with increased channel cross - talks , the performances of all system degrade monotonously .
b. the performance of the balanced system remains the best with any level of channel cross - talks . c. the dci - jcfm is kind of robust with cross - talks , since with a severe cross - talk level , i.e. @xmath130 , the system performance is still comparable or even better than a decoupled counterpart .
we pick the optimized constellation @xmath131 designed for unbalanced system as an example in this section .
if cross - talks do not exist , the corresponding subcarrier symbol waveforms @xmath132 obtained with dci - jcfm are plotted in fig .
4 . with fixed optical power @xmath133 and varying noise power @xmath134 , fig . 5
includes bit error rate curves of the two scheme with different color illumination across selected working electrical snrs , which is defined as @xmath135 for the dci - jcfm and snr for the decoupled scheme is defined similarly .
significant power gains of the dci - jcfm over the decoupled scheme are observed .
also , the more unbalanced the system is , the worse performance is expected .
if cross - talks exist and @xmath136 , the corresponding subcarrier symbol waveforms are plotted in fig .
6 - fig . 8 . with each figure ,
the symbol waveforms are differentiated by color . in practice ,
the sampled version of these waveforms can be pre - stored in the memory of a high speed waveform generator .
, no cross - talks . ] , no cross - talks . ] , no cross - talks . ] , @xmath136 . ] , @xmath136 . ] , @xmath136 . ]
we have propose a joint constellation design scheme termed dci - jcfm taking advantage of the wavelength , frequency , and adaptive bias diversities at the same time for indoor visible light communication systems . by applying the dci - jcfm scheme , waveform symbols with a much larger med can be obtained than those from a decoupled scheme with or without channel cross - talks .
future works will include a comprehensive comparison among three systems : one applying dynamic range , one with long - term papr , and one with instantaneous papr constraint respectively ; comparison of power efficiency of the dci - jcfm and the popular dco / aco - ofdm schemes for multi - carrier multi - color vlc systems ; and advanced precoder design to replace the svd - based pre and post - equalizers utilized in this paper .
t. komine and m. nakagawa , `` fundamental analysis for visible light communication system using led lights , '' _ ieee trans . consumer electronics _ , vol .
100 - 107 , 2004 . s. rajagopal , r. roberts , and s. lim , `` ieee 802.15.7 visible light communication : modulation schemes and dimming support , '' _ optics express _ , vol .
21835 - 21842 , dec .
h. elgala , r. mesleh , and h. haas , `` indoor optical wireless communication : potential and state - of - the - art , '' _ ieee commun . mag .
56 - 63 , 2011 .
l. zeng , d. obrien , h. minh , g. faulkner , k. lee , d. jung , y. oh , and e. won , `` high data rate multiple input multiple output ( mimo ) optical wireless communications using white led lighting , '' journal on selected area in communications , vol .
9 , pp . 1654 - 1662 , 2009 .
r. you and j. kahn , `` average power reduction techniques for multiple - subcarriers intensity - modulated optical signals , '' _ ieee trans . commun . _ , vol .
49 , no . 12 , pp . 2164 - 2170 , dec .
s. teramoto and t. ohtsuki , `` multiple - subcarrier optical communication systems with subcarrier signal - point sequence , '' _ ieee trans .
_ , vol . 53 , no . 10 , pp .
1738 - 1743 , oct .
2005 j. karout , e. agrell , k. szczerba , and m. karlsson , `` optimizing constellations for single - subcarrier intensity modulated optical systems , '' _ ieee trans .
inf . theory _ , vol .
58 , no . 7 , pp . 4645 - 4659 , july 2012 .
e. monteiro and s. hranilovic , `` constellation design for color - shift keying using interior point methods , '' _ ieee owc - ws _ , pp .
1224 - 1228 , dec . 2012
. q. gao , j.h .
manton , g. chen , and y. hua , `` constellation design for a multicarrier optical wireless communication channel , '' _ ieee trans .
_ , pp . 214 - 225 , jan
. 2014 .
g. cossu , a.m. khalid , p. choudhury , and e. ciaramella , `` 3.4 gbit / s visible optical wireless transmission based on rgb led , '' _ optics express _ , vol .
b501 - b506 , 2012 .
b. bai , q. he , z. xu , and y. fan , `` the color shift key modulation with non - uniform signaling for visible light communication , '' the 1st international workshop on optical wireless communications in china , pp .
37 - 42 , 2012 .
j. vucic , c. kottke , k. habel , and k.d .
langer , `` 803 mbit / s visible light wdm link based on dmt modulation of a single rgb led luminary , '' _ optical fiber communication conference _ , los angeles , mar .
ieee 802.15.7 visible light communication task group .
( 2010 , aug .
`` ieee 802.15 documents '' [ online ] .
available : https://mentor.ieee.org/802.15/documents?is group=0007 .
cie ( 1999 ) , `` colour rendering ( tc 1 - 33 closing remarks ) , '' _ publication 135/2 , vienna : cie central bureau , isbn 3 - 900734 - 97 - 6_. a. stimson , `` photometry and radiometry for engineers , '' _ new york : wiley and son_. s. han and j. lee , `` an overview of peak - to - average power ratio reduction techniques for multicarrier transmission , '' _ ieee trans .
wireless commun .
523 - 527 , apr . 2005 .
m. beko and r. dinis , `` designing good multi - dimensional constellations , '' _ ieee wireless commun .
_ , vol . 1 , no . 3 , pp .
221 - 224 , 2012 .
f. schreckenbach , n. gortz , j. hagenauer , and g. bauch , `` optimization of symbol mappings for bit - interleaved coded modulation with iterative decoding , '' _ ieee commun .
_ , vol . 7 , no . 12 , dec . 2003 | in this paper , we consider the problem of constellation design for a visible light communication ( vlc ) system using red / green / blue light - emitting diodes ( rgb led ) , and propose a method termed dc - informative joint color - frequency modulation ( dci - jcfm ) .
this method jointly utilizes available diversity resources including different optical wavelengths , multiple baseband subcarriers , and adaptive dc - bias .
constellation is designed in a high dimensional space , where the compact sphere packing advantage over lower dimensional counterparts is utilized . taking into account multiple practical illumination constraints ,
a non - convex optimization problem is formulated , seeking the least error rate with a fixed spectral efficiency .
the proposed scheme is compared with a decoupled scheme , where constellation is designed separately for each led .
notable gains for dci - jcfm are observed through simulations where balanced , unbalanced and very unbalanced color illuminations are considered . * constellation design , visible light communication , dc - informative , joint color - frequency , im / dd . * |
the h@xmath13 molecular ion is a key molecule for the gas phase chemistry in interstellar clouds ( herbst & kemplerer 1973 ; watson 1973 ) . produced by the fast reaction of h@xmath0 with h@xmath14 , it reacts with almost all neutral atoms and molecules , and thus it is the precursor of a large number of complex neutral and ionic molecular species . initially suggested to be present in molecular clouds by martin et al .
( 1961 ) , there is a long history behind the search for this crucial molecule . due to its lack of permanent dipole moment
, h@xmath13 must be observed in the near ir through its ro - vibrational transitions ( oka 1980 ) .
the first searches only provided upper limits to the h@xmath13 column density ( geballe & oka 1989 ; black et al . ,
however , h@xmath13 has been finally detected toward gl2136 and the w33a clouds ( geballe & oka 1996 ) .
in addition , h@xmath13 has also been detected toward the diffuse interstellar medium ( ism ) by mccall et al .
( 1998 ) , and more recently in the line of sight toward several galactic center clouds ( oka et al . 2005 ) . unfortunately
, these observations only provide information on the h@xmath13 abundance in the most diffuse and warm external layers of molecular clouds , and are limited to lines of sight with bright near - ir background sources . a different approach to study h@xmath13 in denser regions obscured by dust extinction
is to observe the asymmetrical species h@xmath0d@xmath1 and hd@xmath14 , which do have a rotational spectrum .
these species play a crucial role in the deuteration enhancement of many molecular species in clouds with kinetic temperatures below @xmath1230 k ( see , e.g. , gerlich et al . , 2002 and references therein ) .
in particular , the h@xmath0d@xmath1/h@xmath13 abundance ratio is expected to dramatically increase above the elemental d / h ratio as the gas cools down .
the first calculation of the h@xmath0d@xmath1 rotational spectrum was reported by dalgarno et al .
( 1973 ) , who suggested that it could be detectable in dense and cold clouds . the h@xmath0d@xmath1 1@xmath15 - 1@xmath4 transition has been observed in the laboratory at 372.42134 ghz ( bogey et al .
1984 ; warner et al .
this precise measurement opened the quest for the detection of h@xmath0d@xmath1 in dense molecular clouds .
initial searches for the o - h@xmath0d@xmath1 1@xmath15 - 1@xmath4 line , and also the p - h@xmath0d@xmath1 1@xmath16 - 0@xmath17 ( at 1370 ghz ) , performed with the @xmath18 ( phillips et al . , 1985 ; pagani et al . , 1992a ; boreiko & betz , 1993 ) , and with the @xmath19 ( van dishoeck et al . , 1992
) were , however , unsuccessful .
fortunately , h@xmath0d@xmath1 has finally been detected toward the young low - mass protostar ngc1333-iras4a ( stark et al .
1999 ) , and in several dark clouds ( caselli et al . 2003 ; vastel et al . , 2004 , 2006a , b ; hogerheijde et al . , 2006 ; harju et al . ,
2006 ) , and tentatively in protoplanetary disks ( ceccarelli et al . , 2004 ; see also guilloteau et al . , 2006 ) .
observational evidence for the crucial role of h@xmath0d@xmath1 , hd@xmath14 ( vastel et al .
, 2004 ) , and even of d@xmath13 , in the deuteration processes of the ism have been widely probed by the detection of double and triple deuterated species such as d@xmath0co ( turner , 1990 ; ceccarelli et al . , 1998 ) ,
nhd@xmath0 ( roueff et al . , 2000 ; loinard et al . , 2001 ; gerin et al .
2006 ) , nd@xmath20 ( van der tak et al . , 2002 ; lis et al . ,
2002 ) , chd@xmath0oh ( parise et al . , 2002 ) , cd@xmath20oh ( parise et al . , 2004 ) , d@xmath0cs ( marcelino et al . , 2005 ) .
ground - based observations of h@xmath0d@xmath1 and hd@xmath14 have always been performed toward very dense and cold cores as the high einstein coefficients of the involved transitions require high volume densities to produce significant emission .
therefore , it is difficult to estimate the h@xmath0d@xmath1 abundance in lower density regions through submillimeter observations . in this letter
, we report on the first detection of h@xmath0d@xmath1 in the far - ir spectrum of sgr b2 observed with the _ infrared space observatory _ ( iso ) .
the observed transition connects the ground state , 1@xmath4 , with the 2@xmath5 level at 113.4 k , and has a wavelength of 126.853 @xmath3 m .
this detection opens the possibility to detect h@xmath0d@xmath1 in clouds with moderate volume density in which the 1@xmath16 - 1@xmath4 transition at 372 ghz will be subthermally excited and its intensity too weak to be detected with currently available radio telescopes .
it also opens the quest to detect h@xmath0d@xmath1 toward far - ir luminous galaxies and high redshift objects .
the data presented in this paper are based on the final calibration analysis of the high spectral resolution line survey of sgr b2 obtained with the long wavelength fabry - prot spectrometer ( lws / fp ) on board iso in its l03 mode ( polehampton 2002 ; polehampton et al . , 2007 ) .
the data for nh@xmath0 , which are necessary to interpret the h@xmath0d@xmath1 absorption reported in this paper , were previously presented by goicoechea et al .
the spectral resolution of the lws / fp spectrometer is @xmath12 35 km s@xmath6 .
the telescope was pointed toward @xmath21=17@xmath22 47@xmath23 21.75@xmath24 ; @xmath25=-28@xmath26 23 14.1 . the h@xmath0d@xmath1 and
nh@xmath0 data are presented in figs . 1 and 2 .
the procedure to reduce the lws / fp data have been described in detail by polehampton et al .
this provides a significant improvement over the previous reduction of goicoechea et al .
( 2004 ) , who did not apply any interactive processing to the standard pipeline data products except for the removal of glitches , averaging of individual scans and removal of baseline polynomials .
the improvements applied in the full spectral survey reduction include determination of accurate detector dark currents ( including stray light ) , improved instrumental response calibration ( including a correction to account for adjacent fp orders ) and interactive shifting of individual miniscans ( to correct for uncertainty in the angle of the lws grating ) .
the signal - to - noise ratio was also enhanced by including every available observation in the survey ( in this case 2 independent observations ) . the broad absorption feature shown in figure 1
was previously assigned only to p - nh2 ( goicoechea et al . , 2004 ; transition 2@xmath27 - 1@xmath15 j=5/2 - 3/2 ) .
however , the final reduction of the lws / fp survey toward sgr b2 shows that the broadening is real .
the feature appears in 2 independent observations : iso tdt number 50601112 , observed on 1997 april 5 with lws detector lw2 ; and iso tdt number 50800515 , observed on 1997 april 7 with lws detector lw3 .
the fact that the feature appears on 2 separate detectors rules out any spurious instrumental broadening of the p - nh2 line .
taking into account the latest data reduction , the broad feature detected at @xmath7126 @xmath3 m can not be due to p - nh@xmath0 alone .
in fact , nh@xmath0 far - ir lines are narrow and appear only at the velocity of sgr b2 ( see , van dishoeck et al . , 1993 , for the detection of nh@xmath0 in the submm ) .
m toward sgr b2 .
the upper panel shows the data and the expected contribution of p - nh@xmath0 ( see figure 2 ) .
the middle panel shows the o - h@xmath0d@xmath1 line profile after subtracting the contribution from p - nh@xmath0 .
finally , the lower panel shows the absorption of ch for comparison purposes . ]
only a few lines remain unidentified in the iso far - ir line survey of sgrb2 .
the feature discussed in this paper must correspond to a line arising from a low energy level or to a blend of lines .
however , molecules heavier than h@xmath0d@xmath1 could have other lines in the far - ir domain .
we have considered the possibility of b - type transitions from slightly asymmetric rotors like hnco and hoco@xmath1 .
both have one transition at similar wavelengths .
however , the line of hnco involves two high energy levels and can be easily ruled out .
hoco@xmath1 has several other lines in the 120 - 127 @xmath3 m range involving low energy levels .
none of which are detected in the far - ir line survey of sgr b2 ( polehampton et al . , 2007 ) .
hence , we are confident that the broad feature detected at 126.853 @xmath3 m corresponds to h@xmath0d@xmath1 .
the intensity scale in figures 1 and 2 was obtained by dividing the observed flux by a fitted continuum as described by polehampton et al .
hence , it is directly related with the line opacity convolved with the instrumental spectral profile of the lws / fp spectrometer .
the velocity scale in figure 1 has been obtained assuming @xmath28=126.853 @xmath3 m ( 2363.325 ghz ) , i.e. , the wavelength of the o - h@xmath0d@xmath1 2@xmath29 - 1@xmath30 transition ( amano & hirao , 2005 ) . the expected error on the frequency of this transition is @xmath12 5 mhz , i.e. , a velocity uncertainty @xmath31 2 km s@xmath6 ( 3@xmath32 ) , which is much lower than the spectral resolution of the lws / fp spectrometer .
the p - nh@xmath0 2@xmath27 - 1@xmath15 j=5/2 - 3/2 line is separated by @xmath12 - 123 km s@xmath6 with respect to the 2@xmath29 - 1@xmath30 line of o - h@xmath0d@xmath1 .
the upper panel of figure 2 shows the three spin rotational components of the o - nh@xmath0 2@xmath33 - 1@xmath4 transition at @xmath7117 @xmath3 m ( goicoechea et al .
2004 ) . in order to compute the expected contribution of the p - nh@xmath0 2@xmath27 - 1@xmath15 triplet at @xmath12126 @xmath3 m
we have modeled the nh@xmath0 absorption assuming a rotational temperature of 20 k , an intrinsic linewidth of 10 km s@xmath6 , a total nh@xmath0 column density of 1.2@xmath810@xmath34 @xmath10 ( goicoechea et al .
2004 ) , and an ortho / para ratio of 3 .
nh@xmath0 wavelengths and line strengths are from the cologne database for molecular spectroscopy ( mller et al . 2001 ) .
the nh@xmath0 hyperfine structure is unresolved by the lws / fp . the synthetic spectrum , convolved with the lws / fp spectral resolution , is shown in figure 2 .
the expected contribution from p - nh@xmath0 to the absorption feature is shown as a continuous line in the upper panel of figure 1 .
the middle panel of this figure shows the same data with the expected contribution from p - nh@xmath0 removed .
the remaining absorption feature is broad and has several absorption velocity peaks , which are similar to those found in unsaturated ground state transitions of light species such as the ( 2,3/2)-(1,1/2 ) line of ch ( see bottom panel of figure 1 ) .
therefore , in the following , we assign this unidentified broad feature to the o - h@xmath0d@xmath1 2@xmath5 - 1@xmath4 line .
h@xmath0d@xmath1 is an asymmetric molecule with @xmath35 @xmath120.6 d ( dalgarno et al . , 1973 ) .
it has two different species , ortho ( ka = odd ) , and para ( ka = even ) , with a spin weight ratio of 3:1 .
the three lowest energy levels , 1@xmath4 , 1@xmath15 and 2@xmath5 , which are at 0 , 17.9 and 113.4 k ( see insert in the middle panel of figure 1 ) , are enough to compute the partition function ( f@xmath36 ) at low temperatures .
hence , f@xmath37
@xmath12 3(1+e@xmath38 ) is a good approximation for t@xmath39 @xmath31 30 k , where t@xmath39 is the rotational temperature .
for example , f@xmath36(3 k)= 3.004 ( all molecules in the ground state ) , while f@xmath36(30 k)=4.8 .
although the 1@xmath15 level starts to be populated in this case , the 2@xmath5 one is still unpopulated .
consequently , if there is a strong continuum source behind or inside the source , we could expect to see the 2@xmath5 - 1@xmath4 transition of o - h@xmath0d@xmath1 in absorption .
the same behavior is observed in ch toward sgr b2 , where only the ground - state line is observed with the line profile shown in the bottom panel of figure 1 ( goicoechea et al .
2004 ; polehampton et al .
2007 ) .
the einstein coefficient of the o - h@xmath0d@xmath1 2@xmath5 - 1@xmath4 transition is large , 1.7@xmath810@xmath40 s@xmath6 , which implies that the rotational temperature of the line will be extremely low in most relevant cases ( or close to the dust temperature if continuum emission is optically thick ) , even when the 1@xmath15 - 1@xmath4 transition is already collisionally excited ( its einstein coefficient is two orders of magnitude lower ) .
therefore , in order to derive a column density from absorption measurements of the 2@xmath5 - 1@xmath4 line , we can assume that all molecules are in the 1@xmath4 ground level , and that the transition has a low rotational temperature in absence of infrared pumping .
the opacity of the 2@xmath5 - 1@xmath4 line can be written as @xmath41 0.067(n/10@xmath9)(10/@xmath42v)/f@xmath37 , where @xmath42v is the linewidth at half intensity in km s@xmath6 , and n is the column density of o - h@xmath0d@xmath1 in @xmath10 . if n(o - h@xmath0d@xmath1)=10@xmath9 @xmath10 and @xmath42v=10 km s@xmath6 , @xmath43(2@xmath5 - 1@xmath4 ) would be 0.024 for t@xmath39=3 k and 0.014 for t@xmath39=30 k respectively . for comparison purposes ,
the opacity of the 1@xmath15 - 1@xmath4 transition at 372 ghz would be 0.024 and 0.006 for t@xmath39=3 and 30 k respectively .
the interpretation of the absorption produced by o - h@xmath0d@xmath1 in sgr b2 requires knowledge of the rotational temperature of the transition .
the cloud has a large dust opacity in the far - ir , @xmath44(100@xmath3m)=3.5 ( cernicharo et al .
1997 , 2006 ; goicoechea et al .
hence , infrared photons cloud play an important role in the pumping of the molecular levels of o - h@xmath0d@xmath1 .
in addition , the fraction of the cloud depth that is `` seen '' at this wavelength corresponds only to its external layers , where @xmath44(127 @xmath3m)@xmath121 , i.e. , n(h@xmath0)@xmath123@xmath810@xmath45 @xmath10 , rather than the @xmath4610@xmath47 @xmath10 corresponding to the total column density along the line of sight ( @xmath44(127 @xmath3m)@xmath123 ) . in order to estimate the radiative and collisional effects on the excitation of the 2@xmath5 - 1@xmath4 transition
, we have assumed de - excitation collisional rates @xmath32(1@xmath15 - 1@xmath4)=@xmath32(2@xmath5 - 1@xmath15)= 10@xmath11 @xmath48 s@xmath6 , and @xmath32(2@xmath5 - 1@xmath4)=5@xmath810@xmath49 @xmath48 s@xmath6 . for a gas at 30 k purely excited by collisions with n(o - h@xmath0d@xmath1)= 10@xmath9 @xmath10 , and @xmath42v=10 km s@xmath6 ,
the rotational temperature varies from 9 to 12 k for h@xmath0 densities of 10@xmath50 and 10@xmath51 @xmath52 respectively .
reducing the collisional rates by a factor 10 produces rotational temperatures of 6 and 8 k for the same volume densities .
the role of radiative pumping by far
ir dust photons can be estimated assuming that the gas we observe in absorption , @xmath44(127@xmath3m)=1 , surrounds a region with @xmath44(127@xmath3m)=2.0 and t@xmath53=30 k. if the radius of the absorbing shell is twice that of the region emitting the bulk of the infrared emission , r@xmath54/r@xmath55=2 , we obtain an excitation temperature of 17 k , with a little dependency on the assumed collisional rates . for r@xmath54/r@xmath55=3 and 1.5 ,
the rotational temperature is 15 and 20 k respectively . in the following we assume t@xmath39=10 - 20 k , and we use these values to derive upper and lower limits to the opacity of the h@xmath0d@xmath1 2@xmath5 - 1@xmath4 line . assuming a linewidth of 10 km s@xmath6 we obtain @xmath43(2@xmath5 - 1@xmath4)=0.18 and 0.36 for t@xmath39=10 and 20 k respectively .
the corresponding column densities are @xmath12 9@xmath810@xmath9 and 2.3@xmath810@xmath56 @xmath10 , and the abundance of o - h@xmath0d@xmath1 , for a total gas column density of 3@xmath810@xmath45 @xmath10 , is @xmath12 3@xmath810@xmath11 for t@xmath39=10 k and 9@xmath810@xmath11 for t@xmath39=20 k. the latter can be ruled out from the upper limit obtained by pagani et al ( 1992a ) for the emission of the 1@xmath15 - 1@xmath4 line at 372 ghz , n(o - h@xmath0d@xmath1)@xmath311.1@xmath810@xmath56 @xmath10 .
hence , a low rotational temperature , t@xmath5710 k , and an abundance of 3@xmath810@xmath11 are the best fit to the o - h@xmath0d@xmath1 absorption at the velocity of sgr b2 . for the other absorbing clouds which are not directly associated to sgr b2 , it is reasonable to assume radiative thermalization with the cosmic background , t@xmath39 @xmath123 k , i.e. , no significant collisional excitation or infrared pumping .
assuming an intrinsic line width of 10 km s@xmath6 , the observed absorption features at -100 , 10 and 120 km s@xmath6 have opacities of of 0.1 , 0.15 and 0.06 and column densities of @xmath12 4.5 , 6.7 , and 2.7 10@xmath9 @xmath10 respectively .
these clouds have low volume densities and total gas column densities of a few 10@xmath58 - 10@xmath45 @xmath10 .
the associated o - h@xmath0d@xmath1 abundances are of a few 10@xmath1110@xmath59 .
the derived abundance of o - h@xmath0d@xmath1 in sgr b2 corresponds to the predictions of chemical models for t@xmath60 @xmath3120 k ( see pagani et al .
, 1992b ; walmsley
et al . , 2004 ; flower et al . , 2006 ) .
these models predict that the para species could be more abundant than the ortho one for that temperature and the expected density of the gas ( a few 10@xmath61 - 10@xmath50 @xmath52 ) .
hence , the fundamental line of p - h@xmath0d@xmath1 could produce significant absorption at 218.8 @xmath3 m in the direction of sgr b2 .
the abundances observed toward the other clouds in the line of sight of sgr b2 indicate lower kinetic temperatures , t@xmath6210 k. in these clouds the predicted ortho / para abundance ratio favors the ortho species .
due the high dependency of the ortho / para ratio with the gas temperature , simultaneous observation of both the 126.853 @xmath3 m ( o - h@xmath0d@xmath1 ) and the 218.8 @xmath3 m ( p - h@xmath0d@xmath1 ) lines will provide a direct measurement of the efficiency of the reactions of h@xmath0d@xmath1 with h@xmath0 and of the conversion of ortho to para h@xmath0d@xmath1 and vice - versa , i.e. , to check chemical models . due to the poor spectral and angular resolution of iso , and due to the limited sensitivity of the lws / fp instrument , the o
h@xmath0d@xmath1 line at 126.853 @xmath3 m could only be expected in sources with a large column density of cold gas with broad linewidths in the line of sight . as an example
, a broad absorption feature has been also reported in the bright ir galaxy arp220 at 127 @xmath3 m by gonzlez - alfonso et al .
( 2004 ) using the lws grating spectrometer ( spectral resolution more than 20 times lower than the lws / fp ) .
these authors assigned it to nh@xmath20 .
while several lines of nh and nh@xmath20 have been observed in this source , no lines have been assigned to nh@xmath0 so far .
hence , the broad absorption feature assigned to nh@xmath20 by gonzlez - alfonso et al . , could also contain some contribution from o - h@xmath0d@xmath1 .
the detection of h@xmath0d@xmath1 presented in this work , shows that far
ir ortho and para
h@xmath0d@xmath1 lines can be used to trace the deuteration fractionation of h@xmath13 in galaxies ( see also ceccarelli and dominik , 2006 ) . therefore , far
ir observations with the pacs instrument on board herschel satellite ( or in the future with sofia ) , will be an important tool to understand the cold ism chemistry in distant galaxies where the d / h ratio will be different and the gas temperature affected by the local cosmic background temperature . finally , the opacity of the 2@xmath5 - 1@xmath4 transition for a dark cloud with @xmath42v=0.5 km s@xmath6 ( see vastel et al . , 2004 ) ,
n(o - h@xmath0d@xmath1)=10@xmath63 @xmath10 , and t@xmath39= 3 k is @xmath120.05 .
while the expected emission in the 1@xmath15 - 1@xmath4 transition at 372 ghz will be negligible , the expected absorption at 126.853 @xmath3 m will be 5% , i.e. , easy to detect with the expected sensitivities of the high spectral resolution instruments planned for sofia .
dark clouds ( see , e.g. , nisini et al . , 1999 ) , prestellar cores , and protoplanetary disks ( see wyatt et al
. , 2003 ) have enough continuum emission in the far - ir , hence , the 2@xmath5 - 1@xmath4 transition can be used as a key observing tool to trace column densities of o - h@xmath0d@xmath1 larger than a few 10@xmath64 @xmath10 , and to understand the deuterium fractionation at large spatial scales in cold dark clouds .
observations of the fundamental lines of p - hd@xmath14 at 691.6605 ghz and o - hd@xmath14 203.03 @xmath3 m will provide additional constraints on these deuteration processes in molecular clouds and protoplanetary disks .
we would like to thank cecilia ceccarelli , serena viti , and our referee for useful comments and suggestions .
we also thank the spanish mec for funding support through grant esp2004 - 665 , aya2003 - 2785 , and `` comunidad de madrid '' government under pricit project s-0505/esp-0237 ( astrocam ) .
this study is supported in part by the european community s human potential programme under contract mcrtn - ct-2004 - 51230 , molecular universe .
jrg was supported by an individual marie curie fellowship under contract meif - ct-2005 - 5153340 .
amano t. , hirao t. , 2005 , j. mol .
, 233 , 7 black , j.h . ,
van dishoeck , e.f . ,
willner , s.p . ,
woods , r.c . , 1990 , , 358 , 459 bogey m. , demuynck c. , denis m. , et al . , 1984 , a.&a . , 137 , l15 boreiko , r.t . , betz , l. , 1993 , , 405 , l39 caselli , p. , van der tak , f.f.s
ceccarelli , c. , bacmann , a. , 2003 , a.&a .
, 403 , l37 ceccarelli , c. , castets , a. , loinard , l. , et al . , 1998 , a.&a . , 338 , l43 ceccarelli , c. , dominik , c. , lefloch , b. , et al .
, 2004 , , 607 , l51 ceccarelli , c. , dominik , c. , 2006 , , 640 , l131 cernicharo , j. , lim , t. , cox , p. , et al . , 1997 ,
a.&a . , 323 , l25 cernicharo , j. , goicoechea , j.r . , pardo , j.r . , asensio - ramos , a. , 2006 , , 642 , 940 dalgarno , a. , herbst , e. , novick , s. , kemplerer w. , 1973 , 183 , l131 flower , d. , pineau des frets , g. , walmsley , c.m . , 2006 , a.&a . , 449 , 621 geballe , t.r . ,
oka , t. , 1989 , , 342 , 855 geballe , t.r .
, oka , t. , 1996 , nature , 384 , 344 gerin , m. , lis , d. c. , philipp , s. , gusten , r. , roueff , e. & reveret , v. 2006 , a&a , 454 , l63 .
gerlich , d. , herbsst , e. , roueff , e. , 2002 , planetary and space science , 50 , 1275 goicoechea , j.r . ,
rodriguez - fernandez , n.j . , cernicharo , j. , 2004 , , 600 , 214 gonzlez - alfonso , e. et al . , 2004 , , 613 , 247 guilloteau , s. , pitu , v. , dutrey , a. , gulin , m. , 2006 , a.&a . , 448 , l5 harju , j. , haikala , l.k . ,
lehtinen , k. , et al .
, 2006 , a.&a . , 454 , l55 herbst , e. , kemplerer , w. , 1973 , , 185 , 505 hogerheide , m.r . ,
caselli , p. , emprechtinger , m. , et al .
( 2006 ) , a.&a . , 454 , l59 lis , d.c . ,
roueff , e. , gerin , m. et al . , 2002 , , 571 , l55 loinard , l. , castets , a. , ceccarelli , c. , tielens , a.g.g.m . , 2001 , , 552 , l163 marcelino , n. , cernicharo , j. , roueff , e. , et al . , 2005 , , 620 , 308 martin , d.w . ,
mcdaniel , e.w . , meeks , m.l .
, 1961 , , 134 , 1012 mccall b.j . ,
geballe , t.r . ,
hinkle , k.h . ,
oka , t. , 1998 , science , 279 , 1910 mller , h. s. p. , thorwirth , s. , roth , d. a. & winnewisser , g. 2001 , 370 , l49 .
nisini , b. , benedettini , m. , giannini , t. , et al . , 1999 ,
a.&a . , 343 , 266 oka , t. , phys . rev .
lett . , 1980 , 45 , 531 .
oka , t. , geballe , t.r . , goto , m. , usuda , t. , mccall , bj .
2005 , apj , 632 , 882 .
pagani , l. , wannier , p.g . , frerking , m.a . , et al . ,
1992a , a.&a .
, 258 , 472 pagani , l. , salez , m. , wannier , p.g . , 1992b , a.&a . , 258 , 479 parise , b. , ceccarelli , c. , tielens , a.g.g.m . , et al
, 393 , l49 parise , b. , castets a. , herbst , e. , et al .
, 416 , 159 polehampton , e.t . , 2002 , ph.d .
thesis , linacre college , oxford polehampton , e.t . , baluteau , j.p . ,
swinyard , b. , goicoechea , j.r .
2007 , mnras , submitted .
roueff , e. , tin , s. , coudert , l.h . ,
et al . , a.&a . , 354 , l63 stark , r. , van de tak , f.f.s . ,
van dishoeck , e.f . , 1999 , , 521 , l67 turner , b. , 1990 , , 362 , l29 van der tak , f.f.s . ,
schilke , p. , muller , h.s.p . ,
et al . , a.&a .
, 388 , l53 van dishoeck , e.f . , phillips , t.g . , keene , j. , blake , g.a .
, 1992 , a.&a . , 261 , l13 van dishoeck , e.f . ,
jansen , d.j .
, schilke p. , phillips , t.g .
, 1993 , , 416 , l83 vastel , c. , phillips , t.g . ,
yoshida , h. , 2004 , , 606 , l127 vastel , c. , caselli , p. , ceccarelli , c. , et al .
, 2006 , , 645 , 1198 vastel , c. , phillips , t. , caselli , p. , et al . , 2006 , proceedings of the royal society meeting physics , chemistry , and astronomy of h@xmath13 walmsley , c.m . ,
flower , d.r .
, pineau des frets , g. , 2004 , a.&a . , 418 , 1035 warner , h.e . , conner , w.t . ,
petrmichl , r.h . , woods , j. , 1984 , j. chem .
phys . , 81 , 2514 watson , w.d . , 1973 , , 183 , l17 wyatt , m.c . ,
dent , w.r.f . ,
greaves , j.s . , 2003 , mnras , 342 , 876 | we report on the first far ir detection of h@xmath0d@xmath1 , using the _ infrared space observatory _ , in the line of sight toward sgr b2 in the galactic center .
the transition at @xmath2=126.853 @xmath3 m connecting the ground level of o - h@xmath0d@xmath1 , 1@xmath4 , with the the 2@xmath5 level at 113 k , is observed in absorption against the continuum emission of the cold dust of the source .
the line is broad , with a total absorption covering 350 km s@xmath6 , i.e. , similar to that observed in the fundamental transitions of h@xmath0o , oh and ch at @xmath7179 , 119 and 149 @xmath3 m respectively . for the physical conditions of the different absorbing clouds the h@xmath0d@xmath1 column density ranges from 2 to 5@xmath810@xmath9 @xmath10 , i.e. , near an order of magnitude below the upper limits obtained from ground based submillimeter telescopes .
the derived h@xmath0d@xmath1 abundance is of a few 10@xmath11 , which agrees with chemical models predictions for a gas at a kinetic temperature of @xmath1220 k. |
masslessness often arises as a consequence of the existence of a symmetry . in quantum electrodynamics
the masslessness of the photon is normally attributed to gauge invariance , or symmetry under local changes of phase . in quantum chromodynamics ,
the theory of the strong interaction , masslessness of the gluons is likewise attributed to a gauge invariance , albeit a nonlinear one . in general relativity
, the masslessness of gravitons can be traced to symmetry under active diffeomorphisms : no diffeomorphism - invariant mass term exists .
in some circumstances , however , there exists an alternative reason why a field might be massless .
surprisingly , this alternative explanation involves the breaking of a symmetry rather than its existence . a general result , the nambu - goldstone theorem @xcite , states under mild assumptions that there must be a massless particle whenever a continuous global symmetry of an action is nt a symmetry of the vacuum . in this talk , based on ongoing work with alan kosteleck @xcite ,
we show that an alternative description of gravity can be constructed from a symmetric two - tensor without the assumption of masslessness . in this picture , masslessness is a consequence of symmetry breaking rather than of exact symmetry : diffeomorphism symmetry and local lorentz symmetry are spontaneously broken , but the graviton remains massless because it is a nambu - goldstone mode .
the cardinal object in the theory is a symmetric two - tensor - density , denoted by @xmath0 .
starting point is the lagrange density @xmath1 here , @xmath2 is the usual quadratic kinetic operator for a massless spin-2 field , @xmath3,\end{aligned}\ ] ] and @xmath4 is a potential which is built out of the four scalars @xmath5 we will suppose that @xmath4 has a local minimum at @xmath6 , and that @xmath0 acquires an expectation value @xmath7 .
note that this implies spontaneous breaking of lorentz invariance .
we can now decompose @xmath8 where @xmath9 are the fluctuations of @xmath0 . at low energy , the values of the scalars @xmath10
will be approximately fixed to their values @xmath11 in the local minimum .
then , the linearized form of the potential can be taken equivalent to a sum of lagrange multiplier terms that fix the values of the four scalars @xmath10 to the values @xmath11 : @xmath12 we obtain the linearized equations of motion @xmath13 together with the constraints @xmath14 noting that the left - hand side of eq .
( [ lin_eom ] ) equals the linear part of the ricci tensor , it follows that the low - energy linearized dynamics of this model is equivalent to linearized general relativity in an axial - type gauge defined by conditions ( [ gauge_conditions ] ) .
the four gauge conditions ( [ gauge_conditions ] ) reduce the original 10 @xmath15 modes to 6 degrees of freedom . they can be expressed
as the generators of the lorentz generators @xmath16 acting on the vacuum expectation value @xmath17 of the cardinal field : @xmath18 imposing the equations of motion imposes masslessness as well as the lorenz conditions : @xmath19 reducing the number of propagating degrees of freedom to two helicities .
in order that the cardinal model correctly describe gravity it needs to be coupled to the matter energy - momentum tensor . at the linear level , this can be done by including the term @xmath20 where @xmath21 is the trace - reversed energy - momentum tensor . as a result
, the equations of motion reduce to the linearized einstein equation @xmath22 ( where we have taken , for now , the values of the lagrange multipliers @xmath23 equal to zero ) .
the total energy - momentum tensor consists not only of contributions of matter .
there is a contribution of the gravitons themselves as well , quadratic in @xmath9 . as a consequence
, the inclusion of a cubic term in ( [ cardinal_flat ] ) is required .
this , in turn , implies a cubic contribution to the energy - momentum tensor , corresponding to a quartic term in the lagrangian .
this process continues indefinitely , yielding in the limit the full einstein - hilbert action @xcite .
it has been shown by deser @xcite that this `` bootstrap '' process can be carried out , for general relativity , in one step if one rewrites the free graviton action in first order ( palatini ) form . in order to implement this bootstrap procedure for the cardinal theory ,
we pass , as a first step , to the trace - reversed tensor @xmath24 as @xmath25 the kinetic term in ( [ cardinal_flat ] ) can be rewritten in palatini form as @xmath26 where the connection coefficients @xmath27 are to be considered as independent additional variables .
it can be shown that the bootstrap of the kinetic term then terminates after one step , yielding @xmath28 which is equivalent to the einstein - hilbert lagrangian @xmath29 from the latter form we conclude that @xmath30 is naturally interpreted as a curved - space metric density . the bootstrap procedure can also be applied to the matter interaction to determine the form of the matter coupling for cardinal gravity . for this purpose ,
the interaction ( [ lin_matter ] ) is conveniently expressed in terms of the trace - reversed energy - momentum tensor @xmath31 for the matter .
this tensor arises by variation of the lagrange density @xmath32 for the matter fields via @xmath33 in the usual way .
we can therefore write @xmath34 for the matter interaction with the cardinal field @xmath35 . applying the bootstrap procedure
can be shown to generate the lagrange density @xmath36 this expression corresponds to the usual curved - space matter lagrangian if we identify ( as above ) @xmath30 as the metric density .
for example , the bootstrap procedure applied to the flat - space electromagnetic energy - momentum tensor yields @xmath37 it is interesting to note at this point that the way we implemented the bootstrap procedure above is not completely unique . instead of writing the full trace - reversed tensor @xmath35 in the palatini form ( [ palatini ] )
, one could substitute its fluctuations @xmath38 ( defined analogously to @xmath9 above ) instead . while this makes no difference in the linearized expression ,
when applying the bootstrap to this form one ends up with a different result , yielding @xmath39 as the curved - space metric density . applying this alternative procedure to the matter coupling yields a curved - space matter lagrangian with an explicit reference to the expectation value @xmath17 , constituting lorentz - violating terms that can be related to the standard model extension @xcite .
its parameters for lorentz violation have been subject to numerous experimental measurements @xcite . for details
see ref .
@xcite .
the most interesting application of the bootstrap is to the scalar potential v , which we now express as a function of four scalars @xmath40 ( @xmath41 ) defined analogously to the @xmath42 defined in eqs .
( [ x1])([x4 ] ) but with @xmath0 replaced by the trace - reversed tensor @xmath35 . as it turns out , for the procedure to be able to be applied to @xmath4 , the latter needs to satisfy nontrivial integrability conditions that strongly restrict its functional form .
for instance , it follows that the unique lowest - order ( in terms of the total power of @xmath24 ) integrable polynomials are @xmath43 more generally , it follows that any polynomial obtained as the term at order @xmath44 in the series expansion of @xmath45|}$ ] is a solution @xmath46 . applying the bootstrap to @xmath46 yields @xmath47|}-\sum_{k=0}^{q-1}\fv_k.\ ] ] we postulate that any integrable potential can be obtained as a suitable linear combination of the @xmath46 polynomials . of particular interest
are scalar potentials @xmath48 that have a taylor expansion of the form @xmath49 with @xmath50 real constants .
if we impose that @xmath50 be positive definite , @xmath48 has a local minimum at @xmath51 ( @xmath41 ) , and it can represent a possibly stable vacuum . it can be shown that the integrability conditions impose constraints on the @xmath50 and on the higher order coefficients , but that a nontrivial solution space exists @xcite .
we have verified numerically that there is a non - empty subset of positive definite solutions , confirming the possibility of scalar potential with stable vacua @xcite .
it is reasonable to expect that , in the limit @xmath52 the integrable potentials @xmath48 of eq . [ potential - taylor ] correspond to integrable potentials in the linearized limit considered above @xmath53 fixing the value of @xmath40 to be @xmath54 .
we conclude that in this limit , the bootstrapped cardinal model corresponds to general relativity in the gauge defined by the constraints @xmath51 .
this is also the case for a potential with finite @xmath50 , if we only consider energy scales low enough such that any fluctuations away from the minimum in the potential can be neglected .
after applying the bootstrap procedure described above we end up with a lagrangian density of the form @xmath55 the linearized equations of motion become : @xmath56 where we defined @xmath57 we see from eq .
( [ full_lin_eqs ] ) that the first term on the right - hand side naturally takes the form of a ( trace - reversed ) energy momentum tensor .
explicitly , we can identify a `` vacuum energy - momentum tensor '' @xmath58 with @xmath59 it takes nonzero values whenever the scalar potential takes values away from the minimum .
explicit solutions of the linearized equations of motion can be obtained with nonzero vacuum energy - momentum tensor . in the latter case
, independent initial / boundary field values can be defined on maximally 4 suitably defined timelike / spacelike spacetime slices .
if the matter energy - momentum tensor is known to be independently conserved ( e.g. , by symmetry arguments ) , the same has to be true for the vacuum energy - momentum tensor . in such a case , choosing @xmath60 to be zero at the initial value spacetime slices ensures it is zero throughout spacetime
we showed it is possible to construct an alternative theory of gravity , the cardinal model , based on spontaneous breaking of lorentz symmetry .
the massless gravitons can be interpreted as nambu - goldstone modes of the spontaneously broken lorentz symmetry .
the full nonlinear form of the lagrangian is fixed by consistent coupling of gravity to the total energy - momentum tensor , and can be constructed by a bootstrap process , starting with an initial lagrangian for the cardinal field consisting of a quadratic kinetic term and a scalar potential . as it turns out , consistency of the bootstrap process imposes strong restrictions on the form of the scalar potential .
nevertheless , consistent potentials exist with local minima . at low energy ,
the lagrangian corresponds to the einstein - hilbert action , with the possible presence of a nonzero `` vacuum energy - momentum tensor '' . at high energy ,
four extra massive graviton modes appear that modify the dynamics of the theory .
an open problem remains the classification of all possible bootstrapped potentials and a study of their properties .
other issues that merit further study are the effect of the massive modes , in particular near singularities or at high temperatures , as well as a study of the quantization of the cardinal model .
it is a pleasure to thank alan kosteleck for collaboration . financial support by the fundao para a cincia e a tecnologia is gratefully acknowledged .
9 kraichnan r 1947 _ mit thesis _ ; kraichnan r 1955 _ phys .
rev . _ * 98 * 1118 ; papapetrou a 1948 _ proc .
irish acad .
_ * 52a * 11 ; gupta s n 1952 _ proc
. phys .
london _ * a65 * 608 ; feynman r p 1956 _ chapel hill conference _ | we present a model of gravity based on spontaneous lorentz symmetry breaking .
we start from a model with spontaneously broken symmetries for a massless 2-tensor with a linear kinetic term and a nonderivative potential , which is shown to be equivalent to linearized general relativity , with the nambu - goldstone ( ng ) bosons playing the role of the gravitons .
we apply a bootstrap procedure to the model based on the principle of consistent coupling to the total energy energy - momentum tensor . demanding consistent application of the bootstrap to the potential term
severely restricts the form of the latter .
nevertheless , suitable potentials exists that permit stable vacua .
it is shown that the resulting model is equivalent , at low energy , to general relativity in a fixed gauge . |
the narrow - line region ( nlr ) in active galactic nuclei ( agns ) is ideally suited to study the central region through its various emission lines .
it has the great advantage over the broad - line region ( blr ) to be spatially resolvable , at least for nearby agns . since the nlr is affected by various parameters such as the energy input from the central engine , the existence of a nuclear torus , radio jets , and star formation , it accesses a number of key questions of agn physics . [ ]
@xmath05007 ( hereafter [ ] ) narrow - band imaging is commonly used to study the nlrs of active galaxies , especially seyferts ( e.g. mulchaey et al .
1996a , falcke et al . 1998 ) . for qsos , this method was first applied by @xcite using an hst emission - line imaging survey of seven radio - quiet pg qsos with [email protected] . including seven seyferts
, they discovered a relation between nlr size and [ ] luminosity .
this newly found nlr size - luminosity relation connects the nlr of qsos and seyferts over three orders of magnitude .
this relation seems to continue the relation found between blr size and luminosity [ e.g. @xcite ] , indicating that both blr and nlr size are only determined by the agn luminosity .
indeed both sizes are related , providing the potential to estimate black hole ( bh ) masses from the nlr size @xcite .
still , the question remains whether the nlr size - luminosity relation holds at high luminosity or whether these objects lose their very large , dynamically unbound nlr @xcite .
however , the slope of the nlr size - luminosity relation has been controversial . while @xcite obtained a slope of 0.5 suggesting a self - regulating mechanism that determines the size to scale with the ionization parameter , @xcite report a slope of 0.33 for their sample of 60 seyfert galaxies imaged with hst .
such a slope is expected for a homogeneous gaseous sphere ionized by a central source ( strmgren sphere ) .
when comparing the two samples and calculating a fit to all type-1 and type-2 agns separately , the relation for type-2 agns is close to 0.33 , while type-1 objects are fit with a slope of 0.55 @xcite .
this result explains the different slopes found by the two groups : in the sample of @xcite , type-1 qsos determine the slope whereas the slope of the sample studied by @xcite is dominated by sy2s .
the difference can be explained by the observers viewing angle in the framework of the unified model and a receding torus @xcite .
one problem that remains when studying the nlr by [ ] imaging alone is that this emission can be contaminated by contributions from e.g. star formation or shock - ionized gas , altering the nlr size .
in addition , the measured size depends on the depth of the images : when comparing ground - based [ ] images of seyfert galaxies from @xcite with the hst snapshot survey of @xcite , the latter reveal , on average , six times smaller nlr sizes , probably due to the 15 to 20 times lower sensitivity .
these considerations question the determination of the `` nlr size '' from [ ] imaging alone .
spatially resolved spectroscopy is a valuable alternative approach as it can directly probe the size in terms of agn photoionization and discriminate the stellar or shock - ionized contribution by making use of selected emission - line ratios .
in addition , several physical parameters of the nlr such as ionization parameter , electron density , and velocity can be directly accessed and analysed as a function of distance from the nucleus .
therefore , we studied the nlr of six sy1 and six sy2 galaxies by means of optical longslit spectroscopy with vlt / fors1 and ntt / emmi . in the following ,
we summarize these results , focussing on the nearby seyfert-2 galaxy ngc1386 which can be considered as prototypical for the whole sample . for a detailed discussion of the sample properties , observations , data reduction and analysis , as well as results ,
see bennert et al .
( 2005 , 2006a , b , d ) . from spatially resolved spectral diagnostics
, we find a transition between central line ratios falling into the agn regime and outer ones in the -region regime for four objects in our sample ( two sy1s , two sy2s ; fig .
[ diag ] ( _ left panel _ ) . applying ` cloudy ` @xcite photoionization models , we show that the observed distinction between -like and agn - like ratios in ngc1386 represents a true difference in ionization source and can not be explained by variations of physical parameters such as ionization parameter , electron density or metallicity @xcite .
we interpret it as a real border between the nlr , i.e. the central agn - photoionized region , and surrounding regions .
it allows us to determine the nlr size free of stellar contamination , yielding an nlr radius of 6 .
this is twice as large as the one measured from the @xcite [ ] hst image , due to the low sensitivity of their snap - shot survey . on the other hand , it is half the value derived by @xcite from long - slit spectroscopy as their nlr size includes [ ] emission from surrounding regions . for almost all 12 seyfert galaxies ,
we find that both the electron density and the ionization parameter decrease with roughly @xmath2 ( fig . [ density ] ) .
this is in agreement with photoionization of the nlr by a central source , supporting our approach to determine the nlr size via diagnostic diagrams .
n.b . and s.k .
thank the organizers for invitation and hospitality .
n.b . acknowledges the support of the american astronomical society and the national science foundation in the form of an international travel grant .
b.j . acknowledges the support of the czech research project no .
is supported by `` nordrhein - westflische akademie der wissenschaften '' .
bennert , n. , falcke , h. , schulz , h. , et al .
2002 , apj , 574 , l105 bennert , n. , falcke , h. , shchekinov , y. , wilson , a. s. 2004 , iaus , 222 , 307 bennert , n. 2005 , phd thesis , university of bochum , germany bennert , n. , jungwiert , b. , komossa , s. , haas , m. , chini , r. 2006a , a&a , 446 , 919 bennert , n. , jungwiert , b. , komossa , s. , haas , m. , chini , r. 2006b , a&a , 456 , 953 bennert , n. , jungwiert , b. , komossa , s. , haas , m. , chini , r. 2006c , newar , 50 , 708 bennert , n. , jungwiert , b. , komossa , s. , haas , m. , chini , r. 2006d , a&a , in press falcke , h. , wilson , a. s. , & simpson , c. 1998 , apj , 502 , 199 ferland g. j. , korista k. t. , verner d. a. , et al . , 1998 , pasp , 110 , 761 fraquelli , h. a. , storchi - bergmann , t. , & levenson , n. a. 2003 , mnras , 341 , 449 kaspi , s. , smith , p. s. , netzer , h. , et al .
2000 , apj , 533 , 631 kewley , l. , dopita , m. , sutherland , r. , et al .
2001 , apj , 556 , 121 netzer , h. , shemmer , o. , maiolino , r. et al .
2004 apj , 614 , 558 mulchaey , j. s. , wilson , a. s. , & tsvetanov , z. i. 1996 , apjs , 102 , 309 schmitt , h. , donley , j. , antonucci , r. , et al .
2003 , apjs , 597 , 768 | we study the narrow - line region ( nlr ) of six seyfert-1 and six seyfert-2 galaxies by means of spatially resolved optical spectroscopy and photoionization modelling . from spatially resolved spectral diagnostics
, we find a transition between the agn - excited nlr and the surrounding star - forming regions , allowing us to determine the nlr size independent of stellar contamination .
cloudy photoionization models show that the observed transition represents a true difference in ionization source and can not be explained by variations of physical parameters . the electron density and ionization parameter decrease with radius indicating that the nlr is photoionized by the central source only .
the velocity field suggests a disky nlr gas distribution . |
one of the most basic models of evolution in finite populations is the moran process @xcite . in the moran process , a population consisting of two types , a mutant type and a wild type , is continually updated via a birth - death process until only one type remains .
the mutant and wild types are distinguished by only their reproductive fitness , which is assumed to be an intrinsic property of a player .
a mutant type has fitness @xmath0 relative to the wild type ( whose fitness relative to itself is @xmath1 ) , and in each step of the process an individual is selected for reproduction with probability proportional to fitness .
reproduction is clonal , and the offspring of a reproducing individual replaces another member of the population who is chosen for death uniformly at random .
eventually , this population will end up in one of the monomorphic absorbing states : all mutant type or all wild type . in this context , a fundamental metric of the success of the mutant type is its ability to invade and replace a population of wild - type individuals @xcite . in a population of size @xmath2 ,
the probability that a single mutant in a wild - type population will fixate in the moran process is @xmath3 in this version of a birth - death process , the members of the population are distinguished by only their types ; in particular , there is no notion of spatial arrangement , i.e. the population is _ well - mixed_. @xcite extend the classical moran process to graph - structured populations , which are populations with links between the players that indicate who is a neighbor of whom . in this structured version of the moran process
, reproduction happens with probability proportional to fitness , but the offspring of a reproducing individual can replace only a _
neighbor _ of the parent .
since individuals are now distinguished by both their types ( mutant or wild ) _ and _ locations within the population , a natural question is whether or not the fixation probability of a single mutant type depends on where this mutant appears in the population .
@xcite show that this fixation probability is independent of the location of the mutant if everyone has the same number of neighbors , i.e. the graph is _ regular _ @xcite .
in fact , remarkably , the fixation probability of a single mutant on a regular graph is the same as that of eq .
( [ fpmoran])an observation first made in a special case by @xcite .
this result , known as the isothermal theorem , is independent of the number of neighbors the players have ( i.e. the _ degree _ of the graph ) .
the moran process is _ frequency - independent _ in the sense that the fitness of an individual is determined by type and is not influenced by the rest of the population .
however , the moran model can be easily extended to account for frequency-_dependent _ fitness . a standard way in which to model frequency - dependence is through evolutionary games @xcite . in the classical setup ,
each player in the population has one of two strategies , @xmath4 and @xmath5 , and receives an aggregate payoff from interacting with the rest of the population .
this aggregate payoff is usually calculated from a sequence of pairwise interactions whose payoffs are described by a payoff matrix of the form @xmath6 each player s aggregate payoff is then translated into fitness and the strategies in the population are updated based on these fitness values . since a player s payoff depends on the strategies of the other players in the population , so does that player s fitness . traditionally , this population is assumed to be infinite , in which case the dynamics of the evolutionary game are governed deterministically by the replicator equation of @xcite .
more recently , evolutionary games have been considered in finite populations @xcite , where the dynamics are no longer deterministic but rather stochastic . in order to restrict who interacts with whom in the population
, these populations can also be given structure .
popular types of structured populations are graphs @xcite , sets @xcite , and demes @xcite .
we focus here on evolutionary games in graph - structured populations that proceed in discrete time steps .
such processes define discrete - time markov chains , either with or without absorbing states ( depending on mutation rates ) .
typically , in evolutionary game theory , one starts with a population of players and repeatedly updates the population based on some update rule such as birth - death @xcite , death - birth @xcite , imitation @xcite , pairwise comparison @xcite , or wright - fisher @xcite .
these update rules can be split into two classes : cultural and genetic ( see * ? ? ?
cultural update rules involve strategy imitation while genetic update rules involve reproduction and inheritance . without mutations ,
an update rule may be seen as giving a probability distribution over a number of strategy - acquisition scenarios : a player inherits a new strategy through imitation ( cultural rules ) or is born with a strategy determined by the parent(s ) ( genetic rules ) .
mutation rates disrupt these scenarios by placing a small probability of a player taking on a novel strategy .
the way in which strategy mutation rates are incorporated into an evolutionary process depends on both the class of the update rule and the specifics of the update rule itself . in a general sense
, we say that strategy mutations are _ homogeneous _ if they depend on neither the players themselves nor the locations of the players .
this notion of homogeneous strategy mutations is analogous to that of a symmetric game , which is a game for which the payoffs depend on the strategies played but are independent of the identities and locations of the players .
the isothermal theorem seems to indicate that populations structured by regular graphs possess a significant degree of homogeneity , meaning that different locations within the population appear to be equivalent for the purposes of evolutionary dynamics .
however , it is important to note that ( i ) fixation probability is just one metric of evolutionary success and ( ii ) the moran process is only one example of an evolutionary process .
for example , in addition to the probability of fixation , one could look at the _ absorption time _ , which is the average number of steps until one of the monomorphic absorbing states is reached .
moreover , one could consider frequency - dependent processes , possibly with different update rules , in which fitness is no longer an intrinsic property of an individual but is also influenced by the other members of the population .
we show that the isothermal theorem does not extend to arbitrary frequency - dependent processes such as evolutionary games .
furthermore , we show that this theorem does not apply to fixation _ times _ ; that is , even for the moran process on a regular graph , the average number of updates until a monomorphic absorbing state is reached can depend on the initial placement of the mutant .
given that the isothermal theorem does not extend to other processes defined on regular graphs , the next natural question is the following : what is the meaning of a spatially - homogeneous population in evolutionary game theory ?
in fact , we argue using asymmetric games @xcite that the term homogeneous " should apply to an evolutionary process _ as a whole _ rather than to just the population structure . even for populations that appear to be spatially homogeneous , such as populations on complete graphs , non - uniform distribution of resources within the population can result in heterogeneity of the overall process .
similarly , for symmetric games , heterogeneity can be introduced into the dynamics of an evolutionary process through strategy mutations .
therefore , a notion of homogeneity of an evolutionary game should take into account _ at least _
( i ) population structure , ( ii ) payoffs , and ( iii ) strategy mutations .
if the strategy - mutation rates are miniscule , then the population spends most of its time in monomorphic states . with small mutation rates
, one can define an embedded markov chain on the monomorphic states and use this chain to study the success of each strategy @xcite .
that is , when a mutation occurs , the population is assumed to return to a monomorphic state before another mutant arises .
thus , the states of interest are the monomorphic states and the states consisting of a single mutant in an otherwise - monomorphic population .
we say that an evolutionary game is _ homogeneous _ if any two states consisting of a single mutant ( @xmath4-player ) in wild - type population ( @xmath5-players ) are mathematically equivalent .
we make precise what we mean by mathematically equivalent " in [ sec : markov ] , but , informally , this equivalence means that any two such states are the same up to relabeling .
in particular , _ all _ metrics such as fixation probability , absorption time , etc .
are the same for any two states consisting of a single @xmath4-mutant in a @xmath5-population .
we show that an evolutionary game in a graph - structured population is homogeneous if the graph is vertex - transitive ( looks the same " from each vertex ) , the payoffs are symmetric , and the strategy mutations are homogeneous .
this result holds for any update rule and selection intensity .
finally , we explore the effects of population structure on asymmetric evolutionary games . in the weak selection limit
, we show that asymmetric matrix games with homogeneous strategy mutations can be reduced to symmetric games if the population structure is arc - transitive ( looks the same " from each edge in the graph ) .
this result is a finite - population analogue of the main result of @xcite , which states that a similar reduction to symmetric games is possible in sufficiently large populations .
thus , we establish that this reduction applies to any population size if the graph possesses a sufficiently high degree of symmetry .
our explorations , both for symmetric and asymmetric games , clearly demonstrate the effects of population structure , payoffs , and strategy mutations on symmetries in evolutionary games .
the evolutionary processes we consider here define discrete - time markov chains on finite state spaces .
the notions of symmetry and evolutionary equivalence that we aim to introduce for evolutionary processes can actually be stated quite succinctly at the level of the markov chain .
we first work with general markov chains , and later we apply these ideas to evolutionary games .
[ def : symmetry ] suppose that @xmath7 is a markov chain on a ( finite ) state space , @xmath8 , with transition matrix @xmath9 .
an automorphism of @xmath10 is a bijection @xmath11 such that @xmath12 for each @xmath13 .
two states @xmath13 are said to be _ symmetric _ if there exists @xmath14 such that @xmath15 . definition [ def : symmetry ] says that the states of the chain can be relabeled in such a way that the transition probabilities are preserved .
this relabeling may affect the long - run distribution of the chain since it need not fix absorbing states , so we make one further refinement in order to ensure that if two states are symmetric , then they behave in the same way : [ def : evolutionaryequivalence ] states @xmath16 and @xmath17 are _ evolutionarily equivalent _ if there exists an automorphism of the markov chain , @xmath14 , such that 1 . @xmath15 ; 2 .
if @xmath18 is a stationary distribution of @xmath10 , then @xmath19 . for a markov chain with absorbing states , the notions of symmetry and evolutionary equivalence of states need not coincide ( see example [ ex : swapstrategies ] of appendix b )
. however , if the markov chain has a unique stationary distribution ( as would be the case if it were irreducible ) , then symmetry implies evolutionary equivalence : [ prop : ergodicsymmetry ] if @xmath10 has a unique stationary distribution , then two states are symmetric if and only if they are evolutionarily equivalent .
we show in appendix b that a markov chain symmetry preserves the set of stationary distributions ( lemma [ lem : stationarydistribution ] ) , so if there is a unique stationary distribution , then condition ( ii ) of definition [ def : evolutionaryequivalence ] is satisfied automatically by any symmetry .
proposition [ prop : ergodicsymmetry ] is then an immediate consequence of this result .
if @xmath16 and @xmath17 are evolutionarily equivalent , then it is clear for absorbing processes that the probability that @xmath16 fixates in absorbing state @xmath20 is the same as the probability that @xmath17 fixates in state @xmath20 ( and similarly for fixation times ) .
if the process has a unique stationary distribution , then the symmetry of @xmath16 and @xmath17 implies that this distribution puts the same mass on @xmath16 and @xmath17 .
these properties follow at once from the fact that the states @xmath16 and @xmath17 are equivalent up to relabeling .
our focus is on evolutionary games on fixed population structures . if @xmath21 is a finite set of strategies ( or actions " ) available to each player , and if the population size is @xmath2 , then the state space of the markov chain defined by an evolutionary game in such a population is @xmath22 . for evolutionary games without random strategy mutations ,
the absorbing states of the chain are the monomorphic states , i.e. the strategy profiles consisting of just a single unique strategy .
thus , states @xmath16 and @xmath17 are evolutionarily equivalent if they are symmetric relative to the monomorphic states . on the other hand , evolutionary processes with strategy mutations are typically irreducible ( and have unique stationary distributions ) ; in these processes , the notions of symmetry and evolutionary equivalence coincide by proposition [ prop : ergodicsymmetry ] . in order to state the definition of a _ homogeneous _ evolutionary process , we first need some notation
. for @xmath23 , denote by @xmath24 the state in @xmath25 whose @xmath26th coordinate is @xmath27 and whose @xmath28th coordinate for @xmath29 is @xmath30 ; that is , all players are using strategy @xmath30 except for player @xmath26 , who is using @xmath27 .
[ def : homogeneousevolutionaryprocess ] an evolutionary process on @xmath25 is _ homogeneous _ if for each @xmath23 , the states @xmath24 and @xmath31 are evolutionarily equivalent for each @xmath32 .
an evolutionary process is _ heterogeneous _ if it is not homogeneous . in other words ,
an evolutionary process is homogeneous if , at the level of the markov chain it defines , any two states consisting of a single mutant in an otherwise - monomorphic population appear to be relabelings of one another . as noted in
[ subsec : generalmarkov ] , _ all _ quantities with which one could measure evolutionary success are the same for these single - mutant states if the process is homogeneous .
we consider evolutionary games in graph - structured populations .
unless indicated otherwise , a graph " means a directed , weighted graph on @xmath2 vertices .
a directed graph is one in which the edges have orientations , meaning there may be an edge from @xmath26 to @xmath28 but not from @xmath28 to @xmath26 . moreover , the edges carry weights , which we assume are nonnegative real numbers .
a directed , weighted graph is equivalent to a nonnegative @xmath33 matrix , @xmath34 , where there is an edge from @xmath26 to @xmath28 if and only if @xmath35 .
if there is such an edge , then the weight of this edge is simply @xmath36 .
since there is a one - to - one correspondence between directed , weighted graphs on @xmath2 vertices and @xmath33 real matrices , we refer to graphs and matrices using the same notation , describing @xmath34 as a graph but using the matrix notation @xmath36 to indicate the weight of the edge from vertex @xmath26 to vertex @xmath28 .
every graph considered here is assumed to be connected ( strongly ) , which means that for any two vertices , @xmath26 and @xmath28 , there is a ( directed ) path from @xmath26 to @xmath28 .
this assumption is not that restrictive in evolutionary game theory since one can always partition a graph into its strongly - connected components and study the behavior of an evolutionary process on each of these components .
moreover , for evolutionary processes on graphs that are not strongly connected , it is possible to have both ( i ) recurrent non - monomorphic states in processes without mutations and ( ii ) multiple stationary distributions in processes with mutations .
some processes ( such as the death - birth process ) may not even be defined on graphs that are not strongly connected .
therefore , we focus on strongly - connected graphs and make no further mention of the term connected . "
since our goal is to discuss symmetry in the context of evolutionary processes , we first describe several notions of symmetry for graphs .
the three types of graphs we treat here are _
regular _ , _ vertex - transitive _ , and _
symmetric_. informally speaking , a regular graph is one in which each vertex has the same number of neighboring vertices ( and this number is known as the _ degree _ of the graph ) .
a vertex - transitive graph is one that looks the same from any two vertices ; based on the graph structure alone , a player can not tell if he or she has been moved from one location to another .
a symmetric ( or _ arc - transitive _ ) graph is one that looks the same from any two edges .
that is , if two players are neighbors and are both moved to another pair of neighboring vertices , then they can not tell that they have been moved based on the structure of the graph alone .
we recall in detail the formal definitions of these terms in appendix b. the relationships between these three notions of symmetry , as well as some examples , are illustrated in fig .
[ fig : graphdiagram ] . ) and vertex - transitive but not symmetric ( fig .
[ fig : transitivesymmetric](a)).[fig : graphdiagram ] ] we turn now to evolutionary processes in graph - structured populations : consider the moran process on a graph , @xmath34 .
@xcite show that if @xmath34 is regular , then the fixation probability of a randomly placed mutant is given by eq .
( [ fpmoran ] ) , the fixation probability of a single mutant in the classical moran process .
this result ( known as the isothermal theorem ) proves that , in particular , this fixation probability does not depend on the initial location of the mutant .
( we refer to this latter statement as the weak " version of the isothermal theorem . ) our definition of homogeneity in the context of evolutionary processes ( definition [ def : homogeneousevolutionaryprocess ] ) is related to this independence of initial location and has nothing to do with fixation probabilities in the classical moran process .
naturally , the isothermal theorem raises the question of whether or not this location independence extends to absorption times ( average number of steps until an absorbing state is reached ) when @xmath34 is regular .
suppose that @xmath34 is the _ frucht graph _ of fig .
[ fig : fruchtgraph ] .
the frucht graph is an undirected , unweighted , regular ( but not vertex - transitive ) graph of size @xmath37 and degree @xmath38 @xcite .
the fixation probabilities and absorption times of a single mutant in a wild - type population are given in fig .
[ fig : moranbargraphs ] as a function of the initial location of the mutant .
the fixation probabilities do not depend on the initial location of the mutant , as predicted by the isothermal theorem , but the absorption times _ do _ depend on where the mutant arises .
in fact , the absorption time is distinct for each different initial location of the mutant .
the details of these calculations are in appendix c. therefore , even the weak form of the isothermal theorem fails to hold for absorption times . in particular , the moran process on a regular graph need not define a homogeneous evolutionary process . of the frucht graph .
in the snowdrift game , the probability that cooperators fixate depends on the initial location of this mutant on the frucht graph ( even if the intensity of selection is weak).[fig : fruchtgraph ] ] this setup involving two types of players , frequency - independent interactions , and a population structure defined by a single graph , can be generalized considerably : a powerful version of evolutionary graph theory uses _ two _ graphs to define relationships between the players : an _ interaction graph _ , @xmath39 , and a _ dispersal graph _ , @xmath34 @xcite .
these graphs both have nonnegative weights . as an example of how these two graphs are used to define an evolutionary process
, we consider a birth - death process based on two - player , symmetric interactions : [ interactiondispersalexample ] consider a symmetric matrix game with @xmath40 strategies , @xmath41 , and payoff matrix @xmath42 if @xmath43 , then the total payoff to player @xmath26 is @xmath44 if @xmath45 is the intensity of selection , then the fitness of player @xmath26 is @xmath46 in each time step , a player ( say , player @xmath26 ) is chosen for reproduction with probability proportional to fitness . with probability @xmath47 , the offspring of this player adopts a novel strategy uniformly at random from @xmath48 ; with probability @xmath49 , the offspring inherits the strategy of the parent .
next , another member of the population is chosen for death , with the probability of player @xmath28 dying proportional to @xmath36 .
the offspring then fills the vacancy created by the deceased neighbor , and the process repeats .
@xmath39 is called the interaction " graph since it governs payoffs based on encounters , and @xmath34 is called the dispersal " graph since it is involved in strategy propagation . we now explore the ways in which population structure and strategy mutations can introduce heterogeneity into an evolutionary process .
consider the snowdrift game with strategies @xmath50 ( cooperate ) and @xmath51 ( defect ) and payoff matrix @xmath52 suppose that @xmath39 and @xmath34 are both the ( undirected , unweighted ) frucht graph ( see fig .
[ fig : fruchtgraph ] ) .
if the selection intensity is @xmath53 , then the fixation probability of a single cooperator in a population of defectors in a death - birth process depends on the initial location of the cooperator ( fig .
[ fig : snowdriftbargraphs ] ) . since the frucht graph is regular ( but not vertex - transitive )
, this example demonstrates that the isothermal theorem does not extend to frequency - dependent games .
in particular , symmetric games on regular graphs can be heterogeneous , and regularity of the graph does not imply that the fixation probability of a randomly placed mutant " is well - defined .
this dependence of the fixation probability on the initial location of the mutant is not specific to the snowdrift game or the death - birth update rule ; one can show that it also holds for the donation game in place of the snowdrift game or the birth - death rule in place of the death - birth rule , for instance . with @xmath53 ,
the selection intensity is fairly strong , which raises the question of whether or not these fixation probabilities still differ if selection is weak . in fact
, our observation for this value of @xmath54 is not an anomaly : suppose that @xmath16 and @xmath17 are states ( indicating some non - monomorphic initial configuration of strategies ) , and that @xmath55 and @xmath56 are monomorphic absorbing states ( indicating states in which each player uses the same strategy ) .
let @xmath57 denote the probability that state @xmath26 is reached after starting in state @xmath16 , and let @xmath58 denote the average number of updates required for the process to reach an absorbing state after starting in state @xmath16 .
each of @xmath57 and @xmath58 may be viewed as functions of @xmath54 , and we have the following result : [ prop : functionofbeta ] each of the equalities @xmath59 holds for either ( i ) _ every _ @xmath45 or ( ii ) at most finitely many @xmath45 .
thus , if one of these equalities fails to hold for even a single value of @xmath54 , then it fails to hold for all sufficiently small @xmath60 . for a proof of proposition [ prop : functionofbeta ] , see appendix a. this result allows one to conclude that if there are differences in fixation probabilities or times for large values of @xmath54 ( where these differences are more apparent ) , then there are corresponding differences in the limit of weak selection .
even if a symmetric game is played in a well - mixed population , heterogeneous strategy mutations may result in heterogeneity of the evolutionary process .
consider , for example , the pairwise comparison process @xcite based on the symmetric snowdrift game , ( [ snowdriftgame ] ) , in a well - mixed population with @xmath61 players .
we model this well - mixed population using a complete , undirected , unweighted graph of size @xmath38 for each of @xmath39 and @xmath34 ( see fig .
[ fig : wellmixedthree ] ) . for @xmath62 ,
let @xmath63 $ ] be the strategy - mutation ( exploration " ) rate for player @xmath26 .
these strategy mutations are incorporated into the process as follows : at each time step , a focal player ( player @xmath26 ) is chosen uniformly at random to update his or her strategy . a neighbor ( one of the two remaining players )
is then chosen randomly as a model player .
if @xmath54 is the selection intensity , @xmath64 is the payoff of the focal player , and @xmath65 is the payoff of the model player , then the focal player imitates the strategy of the model player with probability @xmath66 and chooses to retain his or her strategy with probability @xmath67 with probability @xmath68 , the focal player adopts a new strategy uniformly at random from the set @xmath69 , irrespective of the current strategy .
provided at least one of @xmath70 , @xmath71 , and @xmath72 is positive , the markov chain on @xmath73 defined by this process is irreducible and has a unique stationary distribution , @xmath18 .
let @xmath74 and @xmath75 . since the mutation rate depends on the location , @xmath26 , the strategy mutations are _
heterogeneous_. if the selection intensity is @xmath53 , then a direct calculation ( to four significant figures ) gives @xmath76 where @xmath77 ( resp .
@xmath78 ) is the mass @xmath18 places on the state @xmath79 ( resp .
@xmath80 ) . therefore , by proposition
[ prop : ergodicsymmetry ] and definition [ def : homogeneousevolutionaryprocess ] , this evolutionary process is _ not _ homogeneous , despite the fact that the population is well - mixed and the game is symmetric . this result is not particularly surprising , but it clearly illustrates the effects of hetergeneous strategy - mutation rates on symmetries of the overall process .
the behavior of an evolutionary process sometimes depends heavily on the choice of update rule . as a result , a particular problem in evolutionary game theory
is often stated ( such as the evolution of cooperation ) and subsequently explored separately for a number of different update rules .
for example , consider the donation game ( an instance of the prisoner s dilemma ) in which cooperators pay a cost , @xmath81 , in order to provide the opponent with a benefit , @xmath82 .
defectors pay no costs and provide no benefits . on a large regular graph of degree @xmath83 , @xcite
show that selection favors cooperation in the death - birth process if @xmath84 , but selection _ never _ favors cooperation in the birth - death process .
therefore , the approach of exploring a problem in evolutionary game theory separately for several update rules has its merits . on the other hand
, one might expect that high degrees of symmetry in the population structure , payoffs , and strategy mutations induce symmetries in an evolutionary game for a variety of update rules . before stating our main theorem for symmetric matrix games
, we must first understand the basic components that make up an evolutionary game .
evolutionary games generally have two timescales : interactions and updates . in each ( discrete ) time step ,
every player in the population has a strategy , and this strategy profile determines the _ state _ of the population .
neighbors then interact ( quickly ) and receive payoffs based on these strategies and the game(s ) being played .
the total payoff to a player determines his or her fitness . in the update step of the process ,
the strategies of the players are updated stochastically based on the fitness profile of the population , the population structure , and the strategy mutations .
popular examples of evolutionary update rules are birth - death , death - birth , imitation , pairwise comparison , and wright - fisher .
since interactions happen much more quickly than updates , there is a separation of timescales .
the most difficult part of an evolutionary game to describe in generality is the update step .
if @xmath21 is the strategy set of the game and @xmath2 is the population size , then a state of the population is simply an element of @xmath25 , i.e. a specification of a strategy for each member of the population .
implicit in the state space of the population being @xmath25 is an enumeration of the players .
that is , if @xmath85 is an @xmath2-tuple of strategies , then this profile indicates that player @xmath26 uses strategy @xmath55 . for our purposes
, we need only one property to be satisfied by the update rule of the process , which we state here as an axiom of an evolutionary game : the update rule of an evolutionary game is independent of the enumeration of the players .
[ axiomremark ] as an example of what this axiom means , consider a death - birth process in which a player is selected uniformly at random for death and is replaced by the offspring of a neighbor .
a neighbor is chosen for reproduction with probability proportional to fitness , and the offspring of this neighbor inherits the strategy of the parent with probability @xmath49 and takes on a novel strategy with probability @xmath86 for some @xmath87 . if all else is held constant ( fitness , mutations , etc . ) , the fact that a player is referred to as the player at location @xmath26 is irrelevant : let @xmath88 be the symmetric group on @xmath2 letters . if @xmath89 is a permutation that relabels the locations of the players by sending @xmath26 to @xmath90 , then the strategy of the player at location @xmath90 after the relabeling is the same as the strategy of the player at location @xmath26 before the relabeling . in particular , if @xmath85 is the state of the population before the relabeling , then @xmath91 is the state of the population after the relabeling , where @xmath92 .
the probability that player @xmath90 is selected for death and replaced by the offspring of player @xmath93 after the relabeling is the same as the probability that player @xmath26 is selected for death and replaced by the offspring of player @xmath28 before the relabeling .
thus , for this death - birth process , the probability of transitioning between states @xmath16 and @xmath17 before the relabeling is the same as the probability of transitioning between states @xmath94 and @xmath95 after the relabeling . in this sense
, a relabeling of the players induces an automorphism of the markov chain defined by the process ( in the sense of definition [ def : symmetry ] ) , and the axiom states that this phenomenon should hold for _ any _ evolutionary update rule . in order to state our main result for symmetric games
, we note that an evolutionary graph , @xmath96 , in this setting consists of _ two _ graphs : @xmath39 and @xmath34 .
we say that @xmath96 is regular if both @xmath39 and @xmath34 are regular . for vertex - transitivity of @xmath96 ( resp .
symmetry of @xmath96 ) , we require slightly more than each @xmath39 and @xmath34 being vertex - transitive ( resp . symmetric ) ; we require them to be _ simultaneously _ vertex - transitive ( resp .
symmetric ) .
first of all , we need to define what an _ automorphism _ of @xmath96 is . for @xmath89 ,
let @xmath97 be the graph defined by @xmath98 for each @xmath26 and @xmath28 . using this action
, we define an automorphism of an evolutionary graph as follows : an _ automorphism _ of @xmath99 is an action , @xmath89 , such that @xmath100 and @xmath101 .
we denote by @xmath102 the set of automorphisms of @xmath96 .
we now have the definitions of vertex - transitive and symmetric evolutionary graphs : @xmath99 is _ vertex - transitive _ if for each @xmath26 and @xmath28 , there exists @xmath103 such that @xmath104 .
@xmath99 is _ symmetric _ if @xmath105 and @xmath39 is a symmetric graph . finally , using the notion of an automorphism of @xmath96
, we have our main result : [ thm : mainsymmetrictheorem ] consider an evolutionary matrix game on a graph , @xmath99 , with symmetric payoffs and homogeneous strategy mutations . if @xmath103 , then the states with a single mutant at vertex @xmath26 and @xmath106 , respectively , in an otherwise - monomorphic population , are evolutionarily equivalent .
that is , in the notation of definition [ def : homogeneousevolutionaryprocess ] , the states @xmath24 and @xmath107 are evolutionarily equivalent for each @xmath23 .
the proof of theorem [ thm : mainsymmetrictheorem ] may be found in appendix b. the proof relies on the observation that the hypotheses of the theorem imply that any two states consisting of a single @xmath4-player in a population of @xmath5-players can be obtained from one another by relabeling the players .
thus , in light of the argument in remark [ axiomremark ] , relabeling the players induces an automorphism on the markov chain defined by the evolutionary game .
since any relabeling of the players leaves the monomorphic states fixed , there is an evolutionary equivalence between any two such states in the sense of definition [ def : evolutionaryequivalence ] .
note that this theorem makes no restrictions on the selection strength or the update rule .
[ cor : mainsymmetriccorollary ] an evolutionary game on a vertex - transitive graph with symmetric payoffs and homogeneous strategy mutations is itself homogeneous
. by theorem [ thm : mainsymmetrictheorem ] , two mutants appearing on a graph might define evolutionarily equivalent states even if the graph is not vertex - transitive .
for example , the tietze graph ( see * ? ? ?
* ) , like the frucht graph , has @xmath37 vertices and is regular but not vertex - transitive
. however , unlike the frucht graph , the tietze graph has some nontrivial automorphisms . by theorem [ thm : mainsymmetrictheorem ]
, any two vertices in the tietze graph that are related by an automorphism have the property that the two corresponding single - mutant states are indistinguishable .
an example of two evolutionarily equivalent states on this graph is given in fig .
[ fig : tietze ] . in appendix c ,
for the snowdrift game with death - birth updating , we give the fixation probabilities and absorption times for all configurations of a single cooperator among defectors , which further illustrates the effects of graph symmetries on an evolutionary process .
[ remark : transitivesymmetric ] for a given population size , @xmath2 , and network degree , @xmath83 , there may be many vertex - transitive graphs of size @xmath2 with degree @xmath83 .
for each such graph , the fixation probability of a randomly occuring mutant is independent of where on the graph it occurs by theorem [ thm : mainsymmetrictheorem ] .
however , this fixation probability depends on more than just @xmath2 and @xmath83 ; it also depends on the configuration of the network .
for example , fig .
[ fig : transitivesymmetric ] gives two vertex - transitive graphs of size @xmath108 and degree @xmath109 . as an illustration ,
consider the snowdrift game ( with payoff matrix ( [ snowdriftgame ] ) ) on these graphs with birth - death updating .
if the selection intensity is @xmath110 , then the fixation probability of a single cooperator in a population of defectors is @xmath111 in ( a ) and @xmath112 in ( b ) ( both rounded to four significant figures ) .
these two fixation probabilities differ for all but finitely many @xmath45 by proposition [ prop : functionofbeta ] . until this point
, our focus has been on states consisting of just a single mutant in an otherwise - monomorphic population .
one could also inquire as to when any two states consisting of two ( or three , four , etc . )
mutants are evolutionarily equivalent .
it turns out that that the answer to this question is simple : in general , the population must be well - mixed in order for any two @xmath113-mutant states to be evolutionarily equivalent if @xmath114 . the proof that this equivalence holds in well - mixed populations follows from the argument given to establish theorem [ thm : mainsymmetrictheorem ] ( see appendix b ) .
on the other hand , if the population is not well - mixed , then one can find a pair of states with the first state consisting of two mutants on neighboring vertices and the second state consisting of two mutants on non - neighboring vertices . in general , the mutant type will have different fixation probabilities in these two states .
for example , in the snowdrift game on the graph of fig .
[ fig : transitivesymmetric](b ) , consider the two states , @xmath16 and @xmath17 , where @xmath16 consists of cooperators on vertices @xmath1 and @xmath115 only and @xmath17 consists of cooperators on vertices @xmath1 and @xmath38 only . if @xmath110 , then the fixation probability of cooperators under death - birth updating when starting at @xmath16 ( resp .
@xmath17 ) is @xmath116 ( resp .
@xmath117 ) . therefore , despite the arc - transitivity of this graph , it is not true that any two states consisting of exactly two mutants are evolutionarily equivalent . only in well - mixed populations
are we guaranteed that any two such states are equivalent .
one particular form of payoff asymmetry appearing in evolutionary game theory is _ ecological asymmetry _ @xcite .
ecological asymmetry can arise as a result of an uneven distribution of resources .
for example , in the donation game , a cooperator at location @xmath26 might provide a benefit to his or her opponent based on some resource derived from the environment .
both this resource and the cost of donating it could depend on @xmath26 , which means that different players have different payoff matrices .
these payoff matrices depend on both the location of the focal player and the locations of the opponents .
thus , payoffs for a player at location @xmath26 against an opponent at location @xmath28 in an @xmath40-strategy bimatrix " game @xcite are given by the asymmetric payoff matrix @xmath118 similar to eq .
( [ eq : symmetricmatrixpayoff ] ) , the total payoff to player @xmath26 for strategy profile @xmath43 is @xmath119 we saw in
[ subsec : frequencydependentgames ] an example of a heterogeneous evolutionary game in a well - mixed population with symmetric payoffs . rather than looking at a symmetric game with heterogeneous strategy mutations ,
we now look at an asymmetric game with homogeneous strategy mutations .
consider the ecologically asymmetric donation game on the graph of fig .
[ fig : wellmixedthree ] ( both @xmath39 and @xmath34 ) with a death - birth update rule . in this asymmetric donation game ,
a cooperator at location @xmath26 donates @xmath120 at a cost of @xmath121 ; defectors donate nothing and incur no costs . if @xmath110 , @xmath122 , @xmath123 , and @xmath124 , then the fixation probability of a single cooperator at location @xmath1 ( see fig .
[ fig : wellmixedthree](a ) ) is @xmath125 , while the fixation probability of a single cooperator at location @xmath115 ( see fig . [
fig : wellmixedthree](b ) ) is @xmath126 ( both rounded to four significant figures ) . therefore , even in a well - mixed population with homogeneous strategy mutations ( none , in this case ) , asymmetric payoffs can prevent an evolutionary game from being homogeneous .
asymmetric matrix games in large populations reduce to symmetric games if selection is weak @xcite . in the limit of weak selection , @xcite establish a selection condition for asymmetric matrix games in finite graph - structured populations that extends the condition ( for symmetric games ) of @xcite : [ asymmetricselectioncondition ] there exists a set of _ structure coefficients _ , @xmath127 , independent of payoffs , such that weak selection favors strategy @xmath128 if and only if @xmath129 where @xmath130 , @xmath131 , @xmath132 , and @xmath133 .
strictly speaking , theorem [ asymmetricselectioncondition ] is established for @xmath39 and @xmath34 undirected , unweighted , and satisfying @xmath105 .
however , the proof of theorem [ asymmetricselectioncondition ] extends immediately to the case with @xmath39 and @xmath34 directed , weighted , and possibly distinct , so we make no restrictive assumptions on @xmath39 and @xmath34 in the statement of this theorem here . in the simpler case @xmath134 ,
condition ( [ asymmetricnetwork ] ) takes the form @xmath135 for some collection @xmath136 . for the death - birth process with @xmath39 and @xmath34 the graph of fig .
[ fig : transitivesymmetric](a ) , we calculate _ exact _
values for all of these structure coefficients ( see appendix c ) . in particular , we find that @xmath137 and @xmath138 , so vertex - transitivity does not guarantee that the structure coefficients are independent of @xmath26 and @xmath28 . for the same process on the graph in fig . [
fig : transitivesymmetric](b ) , we find that @xmath139 for each @xmath26 and @xmath28 , so these coefficients do not depend on @xmath26 and @xmath28 .
( in general , even for well - mixed populations , @xmath140 and @xmath141 need not be the same ; for the same process studied here but on the graph of fig .
[ fig : wellmixedthree ] , @xmath142 and @xmath143 for each @xmath26 and @xmath28 . )
this lack of dependence on @xmath26 and @xmath28 is due to the fact that the graph of fig .
[ fig : transitivesymmetric](b ) is _ symmetric _ , and it turns out to be a special case of a more general result : [ thm : mainasymmetrictheorem ] suppose that an asymmetric matrix game with homogeneous strategy mutations is played on an evolutionary graph , @xmath99 . for each @xmath103 , @xmath144 , and @xmath145 , @xmath146 the proof of theorem [ thm : mainasymmetrictheorem ] may be found in appendix b. the following corollary is an immediate consequence of theorem [ thm : mainasymmetrictheorem ] : if @xmath105 and @xmath39 is a symmetric graph ( i.e. @xmath96 is a symmetric evolutionary graph ) , then the structure coefficients are independent of @xmath26 and @xmath28 .
since symmetric graphs are also regular , we have : [ cor : spatialaverage ] if @xmath105 and @xmath39 is a symmetric graph ( i.e. @xmath96 is a symmetric evolutionary graph ) , then strategy @xmath147 is favored in the limit of weak selection if and only if @xmath148 where @xmath149 is the spatial average of the matrices @xmath150 , i.e. @xmath151 where @xmath83 is the degree of the graph , @xmath96 .
( [ tarnitaselectioncondition ] ) is just the selection condition of @xcite for symmetric matrix games .
it follows from corollary [ cor : spatialaverage ] that asymmetric matrix games on arc - transitive ( symmetric ) graphs can be reduced to symmetric games in the limit of weak selection .
evolutionary games in finite populations may be split into two classes : those with absorbing states ( absorbing processes " ) and those without absorbing states . in absorbing processes ,
the notion of fixation probability has played a crucial role in quantifying evolutionary outcomes , but fixation probabilities are far from the only measure of evolutionary success .
much of the literature on evolutionary games with absorbing states has neglected other metrics such as the time to absorption or the time to fixation conditioned on fixation occurring ( conditional fixation time " ) .
this bias toward fixation probabilities has resulted in certain evolutionary processes appearing more symmetric than they actually are .
we have illustrated this phenomenon using the frequency - independent moran process on graphs : the isothermal theorem guarantees that , on regular graphs , a single mutant can not distinguish between initial locations in the graph if the only metric under consideration is the probability of fixation .
however , certain initial placements of the mutant may result in faster absorption times than others if the graph is regular but not vertex - transitive , and the frucht graph exemplifies this claim .
the same phenomenon also holds for conditional fixation times .
the frucht graph , which is a regular structure with no nontrivial symmetries ( there are no two vertices from which the graph
looks " the same ) , also allowed us to show that the isothermal theorem of @xcite does not extend to frequency - dependent evolutionary games .
that is , on regular graphs , the probability of fixation of a single mutant may depend on the initial location of the mutant if fitness is frequency - dependent .
this claim was illustrated via a death - birth process on the frucht graph , in which the underlying evolutionary game was a snowdrift game .
for @xmath53 ( strong selection ) , the fixation probability of a cooperator at vertex @xmath152 was nearly @xmath153 larger than the fixation probability of a cooperator at vertex @xmath154 .
moreover , we showed that if the fixation probabilities of two initial configurations differ for a single value of @xmath54 , then they are the same for at most finitely many values of @xmath54 .
in particular , these fixation probabilities differ for almost every selection strength , so our observation for @xmath53 was not an anomaly .
similar phenomena are observed for frequency - dependent birth - death processes on the frucht graph , for example , and even for frequency - dependent games with the equal gains from switching " property , such as the donation game .
theorem [ thm : mainsymmetrictheorem ] is an analogue of the isothermal theorem that applies to a broader class of games and update rules . the isothermal theorem is remarkable since regularity of the population structure implies that the fixation probabilities are not only independent of the initial location of the mutant , they are the _ same _ as those of the classical moran process . our treatment of homogeneous evolutionary processes is focused on when different single - mutant states are equivalent , not when they are equivalent to the corresponding states in the classical moran process . even if the fixation probability of a single mutant does not depend on
the mutant s location , other factors ( such as birth and death rates ) may affect whether or not this fixation probability is the same as the one in a well - mixed population @xcite .
remark [ remark : transitivesymmetric ] , which compares the fixation probabilities for the snowdrift game on two different vertex - transitive graphs of the same size and degree , shows that the fixation probability of a single mutant even if independent of the mutant s location can depend on the configuration of the graph . in light of these results , the symmetry phenomena for the moran process guaranteed by the isothermal theorem do not generalize and should be thought of as properties of the frequency - independent moran process and not of evolutionary processes in general .
theorem [ thm : mainsymmetrictheorem ] , and indeed most of our discussion of homogeneity , focused on symmetries of states consisting of just a _ single _ mutant . in many cases ,
mutation rates are sufficiently small that a mutant type , when it appears , will either fixate or go extinct before another mutation occurs @xcite .
thus , with small mutation rates , one need not consider symmetries of states consisting of more than one mutant . however ,
if mutation rates are larger , then these multi - mutant states become relevant .
our definition of evolutionary equivalence ( definition [ def : evolutionaryequivalence ] ) applies to these states , but , as expected , the symmetry conditions on the population structure guaranteeing any two multi - mutant states are equivalent are much stronger .
in fact , as we argued in
[ subsubsec : homogeneousevolutionarygames ] , the population must in general be well - mixed even for any pair of states consisting of _ two _ mutants to be evolutionarily equivalent .
consequently , our focus on single - mutant states allowed us to simultaneously treat biologically relevant configurations ( assuming mutation rates are small ) and obtain non - trivial conditions guaranteeing homogeneity of an evolutionary process . the counterexamples presented here could be defined on sufficiently small population structures , and thus all calculations ( fixation probabilities , structure coefficients , etc . ) are exact .
however , these quantities need not always be explicitly calculated in order to prove useful : in our study of asymmetric games , we concluded that an asymmetric game on an arc - transitive ( symmetric ) graph can be reduced to a symmetric game in the limit of weak selection .
( the graph of fig .
[ fig : transitivesymmetric](a ) demonstrates that vertex - transitivity alone does not guarantee that an asymmetric game can be reduced to a symmetric game in this way . )
this result was obtained by examining the qualitative nature of the structure coefficients in the selection condition ( [ asymmetricnetwork ] ) , but it did not require explicit calculations of these coefficients . therefore , despite the difficulty in actually calculating these coefficients , they can still be used to glean qualitative insight into the dynamics of evolutionary games . on large random regular graphs ,
the dynamics of an asymmetric matrix game are equivalent to those of a certain symmetric game obtained as a spatial average " of the individual asymmetric games @xcite .
corollary [ cor : spatialaverage ] is highly reminiscent of this type of reduction to a symmetric game .
for large populations , this result is obtained by observing that large random regular graphs approximate a bethe lattice @xcite and then using the pair approximation method @xcite to describe the dynamics .
the pair approximation method is _ exact _ for a bethe lattice @xcite , so , from this perspective , corollary [ cor : spatialaverage ] is not that surprising since bethe lattices are arc - transitive . of course , a bethe lattice has infinitely many vertices , and corollary [ cor : spatialaverage ] is a finite - population analogue of this result .
the term homogeneous " is used in the literature to refer to several different kinds of population structures .
this term has been used to describe well - mixed populations @xcite . for graph - structured populations , homogeneous graph " sometimes refers to vertex - transitive graphs @xcite . in algebraic graph theory , however , the term homogeneous graph " implies a much higher degree of symmetry than does than vertex - transitivity ( see * ? ? ?
homogeneous " has also been used to describe graphs in which each vertex has the same number of neighbors , i.e. regular graphs @xcite . in between regular and vertex - transitive graphs , homogeneous graph " has also referred to large , random regular graphs @xcite .
as we noted , large , random regular graphs approximate bethe lattices ( which are infinite , arc - transitive graphs ) , but these approximations need not themselves be even vertex - transitive . in many of the various uses of the term homogeneous , " a common aim is to study the fixation probability of a randomly placed mutant . our definition of _ homogeneous evolutionary game
_ formally captures what it means for two single - mutant states to be equivalent , and our explorations of the frucht graph ( in conjunction with theorem [ thm : mainsymmetrictheorem ] ) show that vertex - transitivity , and _ not _ regularity , is what the term homogeneous " in graph - structured populations should indicate .
we also demonstrated the effects of payoffs and strategy mutations on the behavior of these single - mutant states and concluded that the term homogeneous should apply to the entire process rather than to just the population structure .
the homogeneity ( theorem [ thm : mainsymmetrictheorem ] ) and symmetry ( theorem [ thm : mainasymmetrictheorem ] ) results given here do not depend on the update rule , in contrast with results such as the symmetry of conditional fixation times in the moran process of @xcite or the isothermal theorem of @xcite .
we now know that games on regular graphs are _ not _ homogeneous , and we know precisely under which conditions the fixation probability of a randomly placed mutant " is well - defined .
these results provide a firmer foundation for evolutionary game theory in finite populations and a basis for defining the evolutionary success of the strategies of a game .
the authors thank wes maciejewski for helpful conversations and the anonymous referees for their comments .
a. m. and c. h. acknowledge financial support from the natural sciences and engineering research council of canada ( nserc ) , grant rgpin-2015 - 05795 , and c.h . from the foundational questions in evolutionary biology fund ( fqeb ) , grant rfp-12 - 10 .
48 [ 1]#1 [ 1]`#1 ` urlstyle [ 1]doi : # 1 m. assaf and m. mobilia . metastability and anomalous fixation in evolutionary games on scale - free networks .
_ physical review letters _ , 1090 ( 18 ) , oct 2012 .
doi : 10.1103/physrevlett.109.188701 .
l. w. beineke , r. j. wilson , and p. j. cameron , editors .
_ topics in algebraic graph theory_. cambridge university press , 2004 .
doi : 10.1017/cbo9780511529993 .
b. bollobs .
_ random graphs_. cambridge university press , 2001 .
doi : 10.1017/cbo9780511814068 .
j. a. bondy and u. s. r. murty .
_ graph theory_. springer london , 2008 .
doi : 10.1007/978 - 1 - 84628 - 970 - 5 .
d. cheng , f. he , h. qi , and t. xu .
modeling , analysis and control of networked evolutionary games .
_ ieee transactions on automatic control _ , 2015 .
doi : 10.1109/tac.2015.2404471 .
f. r. k. chung .
_ spectral graph theory_. american mathematical society , 1996 .
f. dbarre , c. hauert , and m. doebeli .
social evolution in structured populations .
_ nature communications _ , 5 , mar 2014 .
doi : 10.1038/ncomms4409 .
w. j. ewens .
_ mathematical population genetics_. springer new york , 2004 .
doi : 10.1007/978 - 0 - 387 - 21822 - 9 .
r. frucht .
herstellung von graphen mit vorgegebener abstrakter gruppe .
_ compositio mathematica _ , 6:0 239250 , 1939 .
d. fudenberg and l. a. imhof .
imitation processes with small mutations . _ journal of economic theory _ , 1310 ( 1):0 251262 , nov 2006 .
doi : 10.1016/j.jet.2005.04.006 .
c. hauert and l. imhof .
evolutionary games in deme structured , finite populations .
_ journal of theoretical biology _
, 299:0 106112 , apr 2012 .
doi : 10.1016/j.jtbi.2011.06.010 .
l. hindersin and a. traulsen .
counterintuitive properties of the fixation time in network - structured populations .
_ journal of the royal society interface _ , 110 ( 99 ) , jul 2014 .
doi : 10.1098/rsif.2014.0606 .
j. hofbauer .
evolutionary dynamics for bimatrix games : a hamiltonian system ?
_ journal of mathematical biology _ , 340 ( 5 - 6):0 675688 , may 1996 .
doi : 10.1007/bf02409754 .
j. hofbauer and k. sigmund .
_ evolutionary games and population dynamics_. cambridge university press , 1998 .
doi : 10.1017/cbo9781139173179 .
l. a. imhof and m. a. nowak .
evolutionary game dynamics in a wright - fisher process .
_ journal of mathematical biology _ , 520 ( 5):0 667681 , feb 2006 .
doi : 10.1007/s00285 - 005 - 0369 - 8 .
k. kaveh , n. l. komarova , and m. kohandel .
the duality of spatial death - birth and birth - death processes and limitations of the isothermal theorem . _ royal society open science _ , 20 ( 4):0 140465140465 , apr 2015 .
doi : 10.1098/rsos.140465 .
n. l. komarova .
spatial stochastic models for cancer initiation and progression . _ bulletin of mathematical biology _ , 680 ( 7):0 15731599 , jul 2006 .
doi : 10.1007/s11538 - 005 - 9046 - 8 .
e. lieberman , c. hauert , and m. a. nowak .
evolutionary dynamics on graphs .
_ nature _ , 4330 ( 7023):0 312316 , jan 2005 .
doi : 10.1038/nature03204 .
t. maruyama .
a simple proof that certain quantities are independent of the geographical structure of population . _ theoretical population biology _ , 50 ( 2):0 148154 , apr 1974 .
doi : 10.1016/0040 - 5809(74)90037 - 9
. h. matsuda , n. ogita , a. sasaki , and k. sato .
statistical mechanics of population : the lattice lotka - volterra model . _
progress of theoretical physics _ , 880
( 6):0 10351049 , dec 1992 .
doi : 10.1143/ptp/88.6.1035 .
a. mcavoy and c. hauert .
structure coefficients and strategy selection in multiplayer games .
_ journal of mathematical biology _ ,
doi : 10.1007/s00285 - 015 - 0882 - 3 .
a. mcavoy and c. hauert .
asymmetric evolutionary games .
_ plos computational biology _ , 110 ( 8):0 e1004349 , aug 2015 .
doi : 10.1371/journal.pcbi.1004349 .
p. a. p. moran .
random processes in genetics . _
mathematical proceedings of the cambridge philosophical society _ , 540 ( 01):0 60 , jan 1958 .
doi : 10.1017/s0305004100033193 .
m. a. nowak .
_ evolutionary dynamics : exploring the equations of life_. belknap press , 2006 .
m. a. nowak , a. sasaki , c. taylor , and d. fudenberg .
emergence of cooperation and evolutionary stability in finite populations .
_ nature _ , 4280 ( 6983):0 646650 , apr 2004 .
doi : 10.1038/nature02414 .
h. ohtsuki .
stochastic evolutionary dynamics of bimatrix games .
_ journal of theoretical biology _ , 2640 ( 1):0 136142 , may 2010 .
doi : 10.1016/j.jtbi.2010.01.016 .
h. ohtsuki and m. a. nowak .
the replicator equation on graphs .
_ journal of theoretical biology _ , 2430 ( 1):0 8697 , nov 2006 .
doi : 10.1016/j.jtbi.2006.06.004 .
h. ohtsuki , c. hauert , e. lieberman , and m. a. nowak .
a simple rule for the evolution of cooperation on graphs and social networks .
_ nature _ , 4410 ( 7092):0 502505 , may 2006 .
doi : 10.1038/nature04605 .
h. ohtsuki , m. a. nowak , and j. m. pacheco .
breaking the symmetry between interaction and replacement in evolutionary dynamics on graphs . _ physical review letters _ , 980 ( 10 ) ,
doi : 10.1103/physrevlett.98.108106 .
h. ohtsuki , j. m. pacheco , and m. a. nowak .
evolutionary graph theory : breaking the symmetry between interaction and replacement . _ journal of theoretical biology _
, 2460 ( 4):0 681694 , jun 2007 .
doi : 10.1016/j.jtbi.2007.01.024 .
j. m. pacheco , f. l. pinheiro , and f. c. santos .
population structure induces a symmetry breaking favoring the emergence of cooperation . _
plos computational biology _ , 50 ( 12):0 e1000596 , dec 2009 .
doi : 10.1371/journal.pcbi.1000596 .
w. h. press and f. j. dyson .
iterated prisoner s dilemma contains strategies that dominate any evolutionary opponent .
_ proceedings of the national academy of sciences _ , 1090 ( 26):0 1040910413 , may 2012 .
doi : 10.1073/pnas.1206569109 .
c. p. roca , j. a. cuesta , and a. snchez .
evolutionary game theory : temporal and spatial effects beyond replicator dynamics . _ physics of life reviews _ , 60 ( 4):0 208249 , dec 2009 .
doi : 10.1016/j.plrev.2009.08.001 .
g. szab and g. fth .
evolutionary games on graphs . _ physics reports _ , 4460 ( 4 - 6):0 97216 , jul 2007 .
doi : 10.1016/j.physrep.2007.04.004 .
g. szab and c. tke . evolutionary prisoner s dilemma game on a square lattice .
_ physical review e _ , 580 ( 1):0 6973 , jul 1998 .
doi : 10.1103/physreve.58.69 .
a. szolnoki , m. perc , and m. mobilia .
facilitators on networks reveal optimal interplay between information exchange and reciprocity . _
physical review e _ , 890 ( 4 ) , apr 2014 .
doi : 10.1103/physreve.89.042802 . c. e. tarnita and p. d. taylor .
measures of relative fitness of social behaviors in finite structured population models .
_ the american naturalist _ , 1840 ( 4):0 477488 , oct 2014 .
doi : 10.1086/677924 . c. e. tarnita , t. antal ,
h. ohtsuki , and m. a. nowak .
evolutionary dynamics in set structured populations .
_ proceedings of the national academy of sciences _ , 1060 ( 21):0 86018604 , may 2009 .
doi : 10.1073/pnas.0903019106 . c. e. tarnita ,
n. wage , and m. a. nowak .
multiple strategies in structured populations .
_ proceedings of the national academy of sciences _ , 1080 ( 6):0 23342337 , jan 2011 .
doi : 10.1073/pnas.1016008108 . c. taylor , d. fudenberg , a. sasaki , and m. a. nowak .
evolutionary game dynamics in finite populations .
_ bulletin of mathematical biology _ , 660 ( 6):0 16211644 , nov 2004 .
doi : 10.1016/j.bulm.2004.03.004 . c. taylor , y. iwasa , and m. a. nowak .
a symmetry of fixation times in evolutionary dynamics .
_ journal of theoretical biology _
, 2430 ( 2):0 245251 , nov 2006 .
doi : 10.1016/j.jtbi.2006.06.016 .
p. d. taylor and l. b. jonker .
evolutionary stable strategies and game dynamics . _ mathematical biosciences _ , 400 ( 1 - 2):0 145156 , jul 1978 .
doi : 10.1016/0025 - 5564(78)90077 - 9
d. taylor , a. j. irwin , and t. day .
inclusive fitness in finite deme - structured and stepping - stone populations .
_ selection _ , 10 ( 1):0 153164 , jan 2001 .
doi : 10.1556/select.1.2000.1 - 3.15 .
d. taylor , t. day , and g. wild .
evolution of cooperation in a finite homogeneous graph .
_ nature _ , 4470 ( 7143):0 469472 , may 2007 .
doi : 10.1038/nature05784 .
a. traulsen , j. m. pacheco , and m. a. nowak .
pairwise comparison and selection temperature in evolutionary game dynamics .
_ journal of theoretical biology _ , 2460 ( 3):0 522529 , jun 2007 .
doi : 10.1016/j.jtbi.2007.01.002 .
a. traulsen , f. c. santos , and j. m. pacheco .
evolutionary games in self - organizing populations . in _ adaptive networks _ , pages 253267 .
springer science + business media , 2009 .
doi : 10.1007/978 - 3 - 642 - 01284 - 6@xmath15512 .
b. wu , c. s. gokhale , l. wang , and a. traulsen .
how small are small mutation rates ?
_ journal of mathematical biology _
, 640 ( 5):0 803827 , may 2011 .
doi : 10.1007/s00285 - 011 - 0430 - 8 .
j. zukewich , v. kurella , m. doebeli , and c. hauert .
consolidating birth - death and death - birth processes in structured populations . _
plos one _ , 80 ( 1):0 e54639 , jan 2013 .
doi : 10.1371/journal.pone.0054639 .
using a method inspired by a technique of @xcite , we derive explicit expressions ( in terms of the transition matrix ) for fixation probabilities and absorption times .
subsequently , we prove a simple lemma that says that markov chain symmetries preserve the set of a chain s stationary distributions .
suppose that @xmath156 is a discrete - time markov chain on a finite state space , @xmath8 , that has exactly @xmath157 ( @xmath158 ) absorbing states , @xmath159 .
moreover , suppose that the non - absorbing states are transient ( see * ? ? ?
the transition matrix for this chain , @xmath9 , may be written as @xmath160 where @xmath161 is the @xmath162 identity matrix and @xmath163 is the matrix of zeros ( in this case , its dimension is @xmath164 , where @xmath165 is the number of states in @xmath8 ) .
this chain will eventually end up in one of the @xmath157 absorbing states , and we denote by @xmath57 the probability that state @xmath55 is reached when the chain starts off in states @xmath166 .
let @xmath167 be the @xmath168 matrix of fixation probabilities , i.e. @xmath169 for each @xmath16 and @xmath26 .
this matrix satisfies @xmath170 which is just the matrix form of the recurrence relation satisfied by fixation probabilities ( obtained from a first step " analysis of the markov chain ) . consider the matrix , @xmath171 , defined by @xmath172 since @xmath173 for @xmath174 , we see that @xmath175 moreover , the matrix @xmath176 must have full rank since the non - absorbing states are transient ; that is , @xmath177 therefore , by cramer s rule , @xmath178 where the notation @xmath179 means the matrix obtained by replacing the column corresponding to state @xmath16 with the @xmath26th standard basis vector , @xmath180 . thus , eq . ( [ fixationprobabilityformula ] )
gives explicit formulas for the fixation probabilities .
let @xmath181 be the @xmath165-vector indexed by @xmath8 whose entry @xmath58 is the expected time until the process fixates in one of the absorbing states when started in state @xmath166 .
this vector satisfies @xmath182 for @xmath183 as well as the recurrence relation @xmath184 .
therefore , @xmath185 so , by cramer s rule , @xmath186 we now turn to markov chains defined by evolutionary games . before proving proposition [
prop : functionofbeta ] , we make two assumptions : 1 . the payoff - to - fitness mapping is of the form @xmath187 , @xmath188 , @xmath189 , or @xmath190 , where @xmath54 denotes the intensity of selection and @xmath191 denotes payoff .
( of course , fitness can be defined in one of the latter three ways only if the payoffs are such that @xmath192 . ) 2 .
the update probabilities are rational functions of the fitness profile of the population . assumptions ( i ) and ( ii ) are not at all restrictive in evolutionary game theory .
any process in which selection occurs with probability proportional to fitness will satisfy this rationality condition , and indeed all of the standard evolutionary processes ( birth - death , death - birth , imitation , pairwise comparison , wright - fisher , etc . )
have this property .
the four payoff - to - fitness mappings are standard as well . with assumptions ( i ) and
( ii ) in mind , we have : each of the equalities @xmath59 holds for either ( i ) _ every _ @xmath45 or ( ii ) at most finitely many @xmath45 . thus , if one of these equalities fails to hold for even a single value of @xmath54 , then it fails to hold for all sufficiently small @xmath60 .
suppose that @xmath13 and that @xmath55 and @xmath56 are absorbing states . by eq .
( [ fixationprobabilityformula ] ) , @xmath193 similarly , by eq .
( [ absorptiontimeformula ] ) , @xmath194 assuming ( i ) and ( ii ) , eqs .
( [ equalityconditionfixation ] ) and ( [ equalityconditionabsorption ] ) are equivalent to polynomial equations in either @xmath54 or @xmath195 . either way ,
since nonzero polynomial equations have at most finitely many solutions , we see that the equalities @xmath196 and @xmath197 each hold for either ( i ) every @xmath54 or ( ii ) finitely many values of @xmath54 .
thus , if @xmath198 ( resp .
@xmath199 ) for even a single selection intensity , then these fixation probabilities ( resp . absorption times ) differ for almost every selection intensity . in particular , they differ for all sufficiently small @xmath54 .
here we recall some standard notions of symmetry for graphs .
although we treat directed , weighted graphs in general , throughout the main text we give several examples of _ _ un__directed and _ _ un__weighted graphs , which are defined as follows : a graph , @xmath34 , is _ undirected _ if @xmath200 for each @xmath26 and @xmath28 .
[ def : unweightedgraph ] a graph , @xmath34 , is _ unweighted _ if @xmath201 for each @xmath26 and @xmath28 . since our goal is to discuss symmetry in the context of evolutionary processes ,
we first describe several notions of symmetry for graphs . in a graph , @xmath34 , the _ indegree _ and
_ outdegree _ of vertex @xmath26 are @xmath202 and @xmath203 , respectively . with these definitions in mind , we recall the definition of a _ regular _ graph : [ def : regulargraph ] @xmath34 is regular if and only if there exists @xmath204 such that @xmath205 for each @xmath26 . if @xmath34 is regular , then @xmath83 is called the _ degree _ of @xmath34 .
let @xmath88 denote the symmetric group on @xmath2 letters ; that is , @xmath88 is the set of all bijections @xmath206 .
each @xmath89 extends to a relabeling action on the set of directed , weighted graphs defined by @xmath207 . in other words , any relabeling of the set of vertices results in a corresponding relabeling of the graph .
the _ automorphism group of _
@xmath34 , written @xmath208 , is the set of all @xmath89 such that @xmath101 .
we now recall a condition slightly stronger than regularity known as vertex - transitivity : [ def : vertextransitivegraph ] @xmath34 is _ vertex - transitive _ if for each @xmath26 and @xmath28 , there exists @xmath209 such that @xmath104 .
informally , a graph is vertex - transitive if and only if it looks the same " from every vertex .
if a graph is vertex - transitive , then it is necessarily regular .
the strongest form of symmetry for graphs that we consider here is the following : [ def : symmetricgraph ] @xmath34 is _ symmetric _ ( or _ arc - transitive _ ) if for each @xmath210 with @xmath35 and @xmath211 with @xmath212 , there exists @xmath209 such that @xmath213 and @xmath214 .
a graph is symmetric if it looks the same " from any two directed edges .
arc - transitivity is typically defined for unweighted graphs , i.e. graphs satisfying @xmath215 . for the more general class of weighted graphs
, we require that @xmath88 act transitively on the set of edges of @xmath34 , where edge " means a pair @xmath216 with @xmath35 .
thus , all of the edges in a symmetric , weighted graph have the same weight : otherwise , if @xmath216 and @xmath217 are edges but @xmath218 , then there would exist no permutation , @xmath191 , sending @xmath26 to @xmath219 , @xmath28 to @xmath220 , and preserving the weights of the graph .
therefore , since the weights of a symmetric graph take one of two values ( @xmath221 or else the only nonzero weight ) , such a graph is essentially unweighted .
in [ subsec : generalmarkov ] , we defined two states , @xmath16 and @xmath17 , to be _ evolutionarily equivalent _ if ( i ) there exists an automorphism of the markov chain , @xmath14 , such that @xmath15 , and ( ii ) this automorphism satisfies @xmath19 for each stationary distribution , @xmath18 , of the chain . condition ( i ) , which means that @xmath16 and @xmath17 are _ symmetric _ , alone is not quite strong enough to guarantee that @xmath16 and @xmath17 have the same long - run behavior . to give an example of a symmetry of states that is not an evolutionary equivalence , we consider the neutral moran process in a well - mixed population of size @xmath61 : [ ex : swapstrategies ] in a well - mixed population of size @xmath61 , consider the ( frequency - independent ) moran process with two types of players : a mutant type and a wild type .
suppose that the mutant type is neutral with respect to the mutant ; that is , the fitness of the mutant relative to the wild type is @xmath1 .
since the population is well - mixed , the state of the population is given by the number of mutants it contains , @xmath222 .
consider the map @xmath11 defined by @xmath223 .
states @xmath221 and @xmath38 are absorbing , and , for @xmath224 , the transition probabilities of this process are as follows : @xmath225 it follows at once that @xmath226 preserves these transition probabilities , so @xmath226 is an automorphism of the markov chain .
let @xmath227 be the probability that mutants fixate given an initial abundance of @xmath26 mutants .
the states @xmath1 and @xmath115 are symmetric since @xmath228 , but it is _ not _ true that @xmath229 since @xmath230 and @xmath231 .
the reason for this difference in fixation probabilities is that states @xmath1 and @xmath115 , although symmetric , are not evolutionary equivalent since @xmath226 swaps the two absorbing states of the process .
in contrast to example [ ex : swapstrategies ] , processes with unique stationary distributions have the property that every symmetry of the markov chain is an evolutionary equivalence ( proposition [ prop : ergodicsymmetry ] ) .
the following lemma establishes proposition [ prop : ergodicsymmetry ] : [ lem : stationarydistribution ] if @xmath11 is a symmetry of a markov chain and @xmath18 is a stationary distribution of this chain , then @xmath232 is also a stationary distribution .
in particular , if @xmath18 is unique , then @xmath19 .
if @xmath9 is the transition matrix of this markov chain , then @xmath233_{\mathfrak{s } } & = \sum_{\mathfrak{s}'\in\mathcal{s}}\phi\left(\mu\right)_{\mathfrak{s}'}\mathbf{t}_{\mathfrak{s}',\mathfrak{s } } \nonumber \\ & = \sum_{\mathfrak{s}'\in\mathcal{s}}\mu_{\phi\left(\mathfrak{s}'\right)}\mathbf{t}_{\phi\left(\mathfrak{s}'\right ) , \phi\left(\mathfrak{s}\right ) } \nonumber \\ & = \sum_{\mathfrak{s}'\in\mathcal{s}}\mu_{\mathfrak{s}'}\mathbf{t}_{\mathfrak{s}',\phi\left(\mathfrak{s}\right ) } \nonumber \\ & = \left[\mu^{t}\mathbf{t}\right]_{\phi\left(\mathfrak{s}\right ) } \nonumber \\ & = \mu_{\phi\left(\mathfrak{s}\right ) } \nonumber \\ & = \phi\left(\mu\right)_{\mathfrak{s } } , \end{aligned}\ ] ] so @xmath234 , which completes the proof .
we turn now to the proofs of our main results ( theorems [ thm : mainsymmetrictheorem ] and [ thm : mainasymmetrictheorem ] ) : consider an evolutionary matrix game on a graph , @xmath99 , with symmetric payoffs and homogeneous strategy mutations . if @xmath103 , then the states with a single mutant at vertex @xmath26 and @xmath106 , respectively , in an otherwise - monomorphic population , are evolutionarily equivalent .
that is , in the notation of definition [ def : homogeneousevolutionaryprocess ] , the states @xmath24 and @xmath107 are evolutionarily equivalent for each @xmath23 .
the state space of the markov chain defined by this evolutionary game is @xmath25 , where @xmath21 is the strategy set and @xmath2 is the population size . for @xmath103 ,
let @xmath191 act on the state of the population by changing the strategy of player @xmath26 to that of player @xmath90 for each @xmath235 .
in other words , @xmath191 sends @xmath85 to @xmath236 , which is defined by @xmath237 for each @xmath26 .
therefore , for @xmath23 , we have @xmath238 for each @xmath235 .
since @xmath191 is an automorphism of the evolutionary graph , @xmath96 , we have @xmath100 and @xmath101 . moreover ,
@xmath191 preserves the strategy mutations since they are homogeneous .
since the payoffs are symmetric , @xmath191 just rearranges the fitness profile of the population : the payoff of player @xmath26 becomes the payoff of player @xmath90 ( see eq .
( [ eq : symmetricmatrixpayoff ] ) ) , so the same is true of the fitness values .
therefore , applying the map @xmath191 to @xmath25 is equivalent to applying the map on @xmath25 obtained by simply relabeling the players . since any such relabeling of the players results in an automorphism of the markov chain on @xmath25 that preserves the monomorphic absorbing states
, it follows that @xmath24 and @xmath107 are evolutionarily equivalent .
suppose that an asymmetric matrix game with homogeneous strategy mutations is played on an evolutionary graph , @xmath99 .
for each @xmath103 , @xmath144 , and @xmath145 , @xmath146 let @xmath9 be the transition matrix for the markov chain defined by this process .
since there are nonzero strategy mutations , this chain has a unique stationary distribution , @xmath18 .
the matrix @xmath9 defines a directed , weighted graph on @xmath239 vertices that has an edge from vertex @xmath16 to vertex @xmath17 if and only if @xmath240 .
if there is an edge from @xmath16 to @xmath17 , then the weight of this edge is simply @xmath241 .
the ( outdegree ) laplacian matrix of this graph , @xmath242 , is defined by @xmath243 ( see * ? ? ?
* ) . in terms of this laplacian matrix ,
@xcite show that for any vector , @xmath244 , the stationary distribution satisfies @xmath245 for each state , @xmath16 , where @xmath246 denotes the matrix obtained from @xmath247 by replacing the column corresponding to state @xmath16 by @xmath244 .
thus , if @xmath248 is the vector indexed by @xmath22 with @xmath249 being the frequency of strategy @xmath147 in state @xmath16 , then the _ average abundance _ of strategy @xmath147 is @xmath250 since @xmath9 is a function of the payoffs , @xmath251 , we may write @xmath252 .
( @xmath253 is just an ordered tuple defined by @xmath254 for each @xmath30 , @xmath255 , @xmath26 , and @xmath28 . ) moreover , since the entries of @xmath9 are assumed to be smooth functions of @xmath253 ( see * ? ? ?
* ) , @xmath256 is also a smooth function of @xmath253 by eq .
( [ eq : averageabundance ] ) and the definition of @xmath247 .
we will show that for each @xmath30 and @xmath255 , @xmath257 for each @xmath26 and @xmath28 .
the theorem will then follow from the derivations of @xmath258 , @xmath259 , and @xmath260 in @xcite since it is shown there that each @xmath261 is a function of the elements in the set @xmath262 for @xmath30 , @xmath255 , @xmath26 , and @xmath28 fixed and @xmath263 , let @xmath264 be the function of the vector with @xmath265 at entry @xmath266 and @xmath221 in all other entries . symbolically , if @xmath267 is defined as @xmath1 if @xmath268 and @xmath221 otherwise and @xmath269 then @xmath270
let @xmath89 and suppose that @xmath103 ; that is , @xmath100 and @xmath101 .
@xmath191 induces a map on the payoffs , @xmath253 , defined by @xmath271 .
let @xmath272 denote the orbit of @xmath253 under this action , and consider the enlarged state space @xmath273 . using the markov chain on @xmath25 coming from the evolutionary process
, we obtain a markov chain on @xmath274 via the transition matrix , @xmath275 , defined by @xmath276 for @xmath277 and @xmath278 .
( we write @xmath279 to indicate the transition matrix as a function of the payoff values of the game . )
@xmath191 extends to a map on @xmath280 defined by @xmath281 .
since @xmath191 preserves @xmath39 , @xmath34 , and the strategy mutations ( since they are homogeneous ) , it follows that the induced map @xmath282 is an automorphism of the markov chain on @xmath280 defined by @xmath275 .
if @xmath283 is a stationary distribution for the chain @xmath275 , then , for each @xmath284 and @xmath285 , @xmath286 it then follows from the uniqueness of @xmath287 that there exists @xmath288 such that @xmath289 .
if @xmath283 is such a stationary distribution , then , by lemma [ lem : stationarydistribution ] , @xmath290 for some other stationary distribution , @xmath291 , of the chain on @xmath280 .
this equation implies that @xmath292 for each @xmath285 .
therefore , @xmath293 for each @xmath85 and @xmath285 .
consequently , since @xmath294 , @xmath295 as a result , we have @xmath296
so eq . ( [ eq : partialsequation ] ) holds , which completes the proof .
we now perform explicit calculations using eqs .
( [ fixationprobabilityformula ] ) and ( [ absorptiontimeformula ] ) to show that the isothermal theorem extends to neither absorption times nor frequency - dependent games .
we also calculate the structure coefficients for the death - birth process on the graph of fig .
[ fig : transitivesymmetric](a ) to show that corollary [ cor : spatialaverage ] does not necessarily hold for graphs that are vertex - transitive but not symmetric .
consider the moran process on the frucht graph ( fig .
[ fig : fruchtgraph ] ) , and suppose that the mutant type has fitness @xmath0 relative to the wild type . by the isothermal theorem of @xcite , the fixation probability of a fixed number of mutants is independent of the configuration of those mutants on the graph . for @xmath297 ,
the absorption times ( of configurations of a single mutant in a wild - type population ) are listed in table [ table : absorption ] .
the fixation probability of a single mutant is @xmath298 for every vertex .
thus , unlike fixation probabilities , absorption times depend on the initial location of the mutant .
( some of the absorption times are similar in this case , but no two are the same . )
.the absorption times of the @xmath37 initial configurations of a single mutant in a wild - type population for the moran process on the frucht graph .
the fitness of the mutant relative to the wild type is @xmath297.[table : absorption ] [ cols="^,^,^",options="header " , ] | in evolutionary game theory , an important measure of a mutant trait ( strategy ) is its ability to invade and take over an otherwise - monomorphic population .
typically , one quantifies the success of a mutant strategy via the probability that a randomly occurring mutant will fixate in the population . however ,
in a structured population , this fixation probability may depend on where the mutant arises .
moreover , the fixation probability is just one quantity by which one can measure the success of a mutant ; fixation _ time _ , for instance , is another .
we define a notion of homogeneity for evolutionary games that captures what it means for two single - mutant states , i.e. two configurations of a single mutant in an otherwise - monomorphic population , to be evolutionarily equivalent " in the sense that _ all _ measures of evolutionary success are the same for both configurations . using asymmetric games , we argue that the term homogeneous " should apply to the evolutionary process as a whole rather than to just the population structure . for evolutionary matrix games in graph - structured populations , we give precise conditions under which the resulting process is homogeneous .
finally , we show that asymmetric matrix games can be reduced to symmetric games if the population structure possesses a sufficient degree of symmetry . |
early type galaxies ( etgs ) , despite having masses and luminosities that span several orders of magnitude , obey a number of tight phenomenological laws . these , collectively known as `` scaling relations '' , include the color - magnitude diagram ( cmd : * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , the color-@xmath3 and @xmath4 @xcite relations and the fundamental plane @xcite . with their remarkably small scatter ,
they impose strong constraints on the structure and evolution of etgs .
morphologically , etgs are either classified as ellipticals ( es ) or lenticular ( s0 ) galaxies . in the late 1980s , based on new and more accurate spectroscopy , es were divided into two groups : pressure supported and rotation supported @xcite .
it is thus particularly interesting to investigate how such dynamically distinct systems formed and evolved while still obeying the very same scaling relations . environment certainly plays a major role in galaxy evolution , as witnessed by the morphology - density relation @xcite : systems in denser surroundings are more likely to be etgs .
the advent of integral field spectroscopy ( ifs ) has brought a wealth of information to the field .
the ` sauron ` survey discovered the existence of two kinematically distinct classes of etgs , slow and fast rotators ( sr and fr , * ? ? ?
* ; * ? ? ?
the former are systems with little to no rotation , often exhibiting kinematically decoupled cores and misalignment between kinematics and photometry .
the latter are flattened systems , compatible with rotational symmetry , where ordered , large scale rotation is important for the gravitational equilibrium .
while overlapping with the existing dichotomy among etgs , the new classification crucially crosses the boundary between es and s0s , in that frs populate both morphological classes .
indeed atlas@xmath0 ( the volume limited follow - up survey to ` sauron ` , * ? ? ? * ; * ? ? ? * ) , found that many morphological es are frs .
they suggest a new classification paradigm based on kinematics rather than morphology @xcite .
+ atlas@xmath0 also presented the kinematic morphology - density relation ( kt-@xmath5 ) , linking the fraction of srs ( ) with the local number density of galaxies .
is insensitive to environment density over 5 orders of magnitude , with a sharp increase observed only in the inner core of the virgo cluster .
@xcite conclude by asking what would be measured in the denser environments beyond the local universe : does the fraction of slow rotators increase further or does it stay constant ? addressing this question
would give further insight on the processes that drive galaxy formation and evolution , and is indeed the goal of this work .
we exploited the unique capabilites of the flames / giraffe multiplexed integral field spectrograph @xcite at the very large telescope ( vlt ) to investigate internal kinematics of galaxies in the densest environment , so to extend the density baseline of the kt-@xmath5 relation .
after describing the observations in section [ sec : observations ] , we present the data reduction and analysis in section [ sec : data ] .
the results are presented in section [ sec : results ] , followed by their discussion in section [ sec : discussion ] and a summary in section [ sec : summary ] .
abell 1689 is a massive galaxy cluster at redshift @xmath6 @xcite . its regular , concentric x - ray contours suggest it is a relaxed system @xcite .
an x - ray luminosity of @xmath7 makes it considerably more luminous than coma , which has @xmath8 @xcite and virgo @xmath9 @xcite .
assuming 7-yr _ wilkinson microwave anisotropy probe _ cosmology ( * ? ? ? *
@xmath10 ) its comoving distance is 741 mpc , giving 1@xmath11 per 3.0 kpc , so that giraffe deployable integral field units ( see below ) sample up to 1 for most galaxies .
giraffe permits the observer to target 15 objects simultaneously and we chose to target 30 galaxies as a compromise between sample size and integration time . our selection was based on a catalogue from @xcite , and in order to gain the maximum possible signal - to - noise ratio ( snr ) , we initially selected the 30 etgs with the highest surface brightness within ( including the brightest cluster galaxy ) .
this sample was then subject to two practical constraints .
we needed all of our targets to have high resolution hst imaging , which limited our choice to candidates in the innermost regions of the cluster .
physical constraints from the instrument ( see section [ sec : data : subsec : instrument ] ) ruled out some targets in the most crowded regions , forcing us to re - select from a reserve list .
this left us with 29 galaxies inside the hst field of view and one outside ( galaxy 20 ) .
we used f625w band imaging from the hst advanced camera for surveys , combined with _
g@xmath12 _ and _ r@xmath12 _ band gemini imaging .
see @xcite for a thorough description of these data .
we present new data taken with the flames / giraffe spectrometer at the vlt unit telescope 2 .
the l612w filter gives a resolution r of 11800 ( the minimum allowed on the instrument ) with a wavelength range of 5732 - 6515 @xmath13 ( 4858 - 5521 @xmath14 in the rest frame ) , which includes prominent absorption features of old stellar populations ( for comparison , ` sauron ` has a wavelength range of 4800 - 5380 @xmath13 ) .
the observations were carried out between 24 may 2009 and 29 may 2009 , as detailed in table [ tab : obs.table ] , which also contains the observing conditions .
the instrument provides 30 independent integral field units ( ifus ) , deployable anywhere on the focal plane .
these are arranged in two positioner plates , each hosting 15 ; each deployable ifu is positioned by a magnetic button , with an accuracy better than @xmath15 and contains an array of 20 square microlenses , each with a side of @xmath16 on the sky .
they are arranged in 4 rows of 6 ( with 4 `` dead '' corners ) for a total field of view of @xmath17 .
each lenslet is then connected to the spectrometer with a dedicated optical fibre . alongside the 15 ifus
, each positioner plate also houses 15 sky fibres .
these are fully deployable just like the former but carry only one lenslet .
.[tab : obs.table]a summary of the vlt / flames giraffe / ifu spectroscopy .
the seeing was measured on site using telescope guide stars . [
cols="<,^,^,^,^,^ " , ] since the magnetic buttons are larger ( @xmath18 ) than the ifu field of view , they can not be deployed closer than a minimum distance of @xmath19 thus constraining the sample selection : galaxies closer than @xmath19 on the sky must be allocated on different plates , if at all . as a result , some targets lying in the most crowded regions of the cluster were omitted .
we proceeded to divide the sample in two equal sets , with galaxies numbers 1 to 15 assigned to plate 1 and galaxies numbers 16 to 30 to plate 2 .
each plate was exposed 5 times for 2 hours , for a total of 10 hours exposure time per galaxy .
we remark that , as detailed in table [ tab : obs.table ] , the seeing was comparable to the size of the lenslets ( @xmath16 ) .
this reduces the correlation between adjacent spaxels .
ccccccccccccccc galaxy & halkola & ra & dec & @xmath20 & @xmath21 & q & @xmath22 & pa & p(sr ) & @xmath23 & @xmath24 + & & ( deg ) & ( deg ) & ( mag ) & ( arcsec ) & ( 1 - 3 ) & & & ( 0 - 1 ) & & ( mpc@xmath25 ) + ( 1 ) & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) & ( 6 ) & ( 7 ) & ( 8) & ( 9 ) & ( 10 ) & ( 11 ) & ( 12 ) + 1 & 17 & 197.94862 & -1.15575 & -23.28 & 0.65 @xmath26 0.29 & 2 & 0.286 & 20 & 0.02 & 0.412 @xmath26 0.092 & 2.381 + 2 & 3 & 197.94831 & -1.14250 & -23.16 & 1.27 @xmath26 0.57 & 1 & 0.583 & 163 & 0.00 & 0.585 @xmath26 0.053 & 2.408 + 3 & 12 & 197.95400 & -1.13998 & -23.57 & 0.71 @xmath26 0.32 & 2 & 0.363 & 47 & 0.00 & 0.611 @xmath26 0.087 & 2.279 + 4 & 34 & 197.96433 & -1.14525 & -22.58 & 1.15 @xmath26 0.51 & 2 & 0.032 & 235 & 0.00 & 0.455 @xmath26 0.105 & 2.333 + 5 & 22 & 197.96239 & -1.15608 & -24.97 & 2.84 @xmath26 1.24 & 2 & 0.219 & 113 & 0.00 & 0.252 @xmath26 0.022 & 2.928 + 6 & 47 & 197.96764 & -1.16323 & -25.41 & 2.39 @xmath26 1.04 & 2 & 0.330 & 266 & 0.91 & 0.116 @xmath26 0.022 & 3.083 + 7 & 65 & 197.97640 & -1.15380 & -22.99 & 0.85 @xmath26 0.38 & 2 & 0.108 & 311 & 0.02 & 0.352 @xmath26 0.100 & 2.339 + 8 & 31 & 197.95769 & -1.17172 & -24.86 & 2.74 @xmath26 1.20 & 3 & 0.090 & 196 & 0.40 & 0.080 @xmath26 0.026 & 3.468 + 9 & 38 & 197.95643 & -1.17477 & -24.28 & 7.51 @xmath26 3.28 & 3 & 0.077 & -41 & 0.00 & 0.428 @xmath26 0.030 & 3.748 + 10 & 43 & 197.95238 & -1.18489 & -23.87 & 1.69 @xmath26 0.74 & 2 & 0.049 & -157 & 0.37 & 0.158 @xmath26 0.046 & 2.658 + 11 & 37 & 197.94754 & -1.18383 & -22.75 & 0.74 @xmath26 0.33 & 2 & 0.217 & -41 & 0.00 & 0.510 @xmath26 0.096 & 2.489 + 12 & 32 & 197.95437 & -1.17140 & -26.18 & 13.92 @xmath26 6.08 & 3 & 0.149 & -62 & 0.95 & 0.037 @xmath26 0.024 & 3.607 + 13 & 14 & 197.94564 & -1.16432 & -23.64 & 1.58 @xmath26 0.69 & 2 & 0.107 & 44 & 0.06 & 0.214 @xmath26 0.049 & 2.696 + 14 & 6 & 197.94071 & -1.16265 & -23.48 & 0.62 @xmath26 0.28 & 2 & 0.361 & 88 & 0.03 & 0.444 @xmath26 0.093 & 2.582 + 15 & 2 & 197.93615 & -1.15561 & -22.81 & 0.93 @xmath26 0.42 & 2 & 0.029 & 83 & 0.00 & 0.356 @xmath26 0.095 & 2.130 + 16 & 7 & 197.94359 & -1.15721 & -23.63 & 1.81 @xmath26 0.79 & 2 & 0.138 & 0 & 0.00 & 0.257 @xmath26 0.043 & 2.779 + 17 & 20 & 197.95362 & -1.15161 & -23.04 & 0.46 @xmath26 0.21 & 2 & 0.558 & 138 & 0.00 & 0.505 @xmath26 0.092 & 2.285 + 18 & 41 & 197.96484 & -1.15373 & -23.55 & 0.95 @xmath26 0.42 & 2 & 0.174 & 198 & 0.00 & 0.410 @xmath26 0.094 & 2.830 + 19 & 35 & 197.96782 & -1.15578 & -25.13 & 2.27 @xmath26 0.99 & 2 & 0.246 & 180 & 0.76 & 0.136 @xmath26 0.026 & 3.321 + 20 & - & 197.98916 & -1.15270 & -24.38 & 1.55 @xmath26 0.19 & 2 & 0.040 & 98 & 0.00 & 0.171 @xmath26 0.064 & 2.061 + 21 & 69 & 197.97680 & -1.16486 & -22.56 & 0.52 @xmath26 0.23 & 2 & 0.278 & 278 & 0.00 & 0.439 @xmath26 0.107 & 2.504 + 22 & 75 & 197.97635 & -1.18002 & -23.47 & 0.86 @xmath26 0.38 & 2 & 0.301 & 193 & 0.00 & 0.626 @xmath26 0.089 & 2.622 + 23 & 61 & 197.96103 & -1.17819 & -23.19 & 0.57 @xmath26 0.25 & 2 & 0.055 & 292 & 0.00 & 0.416 @xmath26 0.092 & 2.822 + 24 & 70 & 197.96526 & -1.19081 & -23.66 & 1.19 @xmath26 0.52 & 2 & 0.245 & -124 & 0.09 & 0.290 @xmath26 0.055 & 2.411 + 25 & 60 & 197.96108 & -1.18797 & -23.64 & 0.85 @xmath26 0.37 & 2 & 0.261 & -41 & 0.00 & 0.509 @xmath26 0.092 & 2.469 + 26 & 29 & 197.95675 & -1.17549 & -24.91 & 4.64 @xmath26 2.02 & 3 & 0.116 & -26 & 0.45 & 0.120 @xmath26 0.027 & 3.446 + 27 & 42 & 197.95666 & -1.17153 & -24.94 & 3.36 @xmath26 1.47 & 3 & 0.154 & -53 & 0.96 & 0.107 @xmath26 0.029 & 3.660 + 28 & 8 & 197.93336 & -1.18330 & -23.54 & 0.83 @xmath26 0.37 & 2 & 0.617 & -35 & 0.00 & 0.588 @xmath26 0.094 & 2.968 + 29 & 28 & 197.95005 & -1.17060 & -23.63 & 0.69 @xmath26 0.31 & 2 & 0.081 & 96 & 0.00 & 0.445 @xmath26 0.099 & 3.060 + 30 & 4 & 197.93953 & -1.16139 & -23.35 & 0.93 @xmath26 0.41 & 2 & 0.265 & 70 & 0.03 & 0.370 @xmath26 0.092 & 2.594 +
we extracted the spectra using the standard eso pipeline , following the guidelines eso offers . each morning
the telescope produces a number of calibration frames , including bias , lamp flats and arc lamp frames . to extract the spectra from the raw images we used the closest calibration available .
the pipeline is organized into 9 `` recipes '' , distinct applications with a number of user configurable parameters : we used the default values unless otherwise stated . for each night
we created a master bias out of the 5 raw frames provided .
we used the method _ ksigma _ and the recipe _ gimasterbias _ , with the keywords _ ksigma.low _ and _ ksigma.high _ set to 3.0 to remove cosmic rays .
we then proceeded to `` fibre localization '' , ( tracing the spectra on the chip ) .
this is done using a set of 3 very high snr lamp flat frames , in the recipe _
gimasterflat_. at each spectral pixel on the frame the recipe determines the locations , in the cross - dispersion direction , of the light peaks corresponding to the centres of each fibre signal .
a curve is fitted to each profile , and is stored as the trace shape .
we used the standard unweighted summation to extract the spectra ( we set the keyword _
extraction.method _ to _ sum _ ) .
we set to _ profile+curve _ the keyword _
biasremoval.method_ , as advised by eso on the website , while the keywords _ fibres.spectra _ and _ fibres.nspectra _ were modified to take into account the occurence of both broken and unused fibres .
the manual indicates that , using _ sum _ , the contamination between neighboring spectra is less than 10% of the counts .
the recipe also determines the pixel - to - pixel variation corrections and the fibre - to - fibre transmission variations .
the wavelength calibration was done separately for each night using the recipe _ giwavecalibration_. the resulting wavelength solution has an accuracy of @xmath27 and a resolution fwhm of @xmath28 .
the science extraction was performed using _
giscience_. we set the parameters _
biasremoval.method _ to _ profile+curve _ and _ flat.apply _ to
_ true_. we used _
g@xmath12 _ and _ r@xmath12 _ band gemini imaging to create a catalogue of all galaxies in the observed region of the sky @xcite .
we applied cuts at _ r@xmath12 _ = 22 and in the related error ( @xmath29 mag ) .
the resulting catalogue has been used to compute the number density of galaxies ( section [ sec : environment ] ) , the cluster luminosity function ( lf ) ( section [ subsec : sample.selection.effects ] ) and the cluster cmd ( section [ sec : red.sequence.selection ] ) .
hst imaging was used to determine de vaucouleurs @xcite effective radii ( using the curve of growth method of * ? ? ? * ) and ellipticities @xmath30 , whenever this was possible . in practice
one galaxy ( number 20 in table [ tab : sample ] ) lies partially outside the acs field of view , and takes its photometric parameters from the r@xmath12 gemini image .
since a large fraction of our sample is found in very dense regions , the surface photometry is often contaminated by that of a neighbour .
consequently the values in table [ tab : sample ] include a quality flag q , as in @xcite .
following @xcite we adopted the method of moments to determine ellipticities : after identifying the image isophotes we compute , for each of them , the position angle of the major axis pa , the ellipticity @xmath30 and the surface area @xmath31 .
the ellipticity of the k - th isophote @xmath32 is defined by @xmath33 where @xmath34 is the flux associated with the i - th pixel , and the coordinates @xmath35 are drawn from the galaxy centre , with the x - axis along the photometric major axis .
the sum is conducted on the set of all pixels comprised in the k - th isophote .
we associate to each isophote an ellipse of area @xmath36 equal to the isophote area , ellipticity @xmath32 and position angle @xmath37 , and associate to it a radius defined by @xmath38 . `
sauron ` and atlas@xmath0 based their classification on values computed at 1 .
we therefore define as the value of @xmath32 computed within the isophote of associated radius .
the results are listed in table [ tab : sample ] .
we find them to be robust against changes in , except for galaxy number 9 ( table [ tab : sample ] ) , which exhibits peculiar photometry , having an abrubt change in both @xmath30 and position angle at a radius of @xmath39 .
stellar kinematics were extracted using ppxf , a penalized maximum likelihood algorithm developed by @xcite .
it fits the line - of - sight velocity distribution ( losvd ) in pixel space , by convolving a linear combination of stellar template spectra with an losvd expressed by the truncated gauss - hermite series @xcite : @xmath40 \label{eq : losvd.gaussherm.dec}\ ] ] where @xmath41 and the @xmath42 are hermite polynomials . in practice
however , our snr was mostly lower than that @xmath43 required to reliably measure the weights @xmath44 and @xmath45 so we fitted a gaussian function , obtaining just @xmath46 and @xmath3 in the above expression .
while atlas@xmath0 team used miles stellar template library @xcite , its fwhm resolution of 2.54 @xmath13@xcite was lower than that of our data ( see section [ sec : data.data.reduction ] ) , so we used the high resolution version ( r = 40000 ) of the elodie template library @xcite , with fwhm of 0.13 @xmath13 .
the two libraries span similar regions in the age - metallicity space ; miles reaches lower ages ( @xmath47 gyr vs @xmath48 gyr ) and includes some old , metal poor stars ( @xmath49 ) , but these are not relevant when fitting etgs , and the change of library is unlikely to introduce any significant bias when compared to atlas@xmath0measurements .
all elodie templates have a gap at @xmath50 , so we cut the galaxy spectra at 5300 @xmath13 . for each galaxy
we computed a weighted average ( with sigma clipping rejection ) of all the 20 spectra , and fed it to ppxf along with all the templates available in elodie .
this resulted in @xmath51 templates being selected for each galaxy , and we use this subset to fit the individual fibre spectra of the galaxy .
we used formal errors derived by ppxf ( we did not exploit the penalizing functionality of the algorithm ) .
these are typically of the order of 15 for v and 17 for @xmath3 , but they do not take into account the correlation introduced when log - rebinning . due to the high spectral resolution and low snr , we decided not to subtract the sky , but rather to fit it simultaneously with the stellar templates . like @xcite
, we provided ppxf with all the simultaneous sky spectra and let the maximum likelihood algorithm rescale them to best fit the data .
the results of the kinematic extraction can be seen on fig .
[ fig : kinematic.maps ] . there are four frames for each galaxy , from left to right : high resolution photometry ( from either hst or gemini ) ; low resolution giraffe spectrograph photometry ; the extracted velocity map and the extracted velocity dispersion map . above each galaxy
we give the i d number ; the celestial orientation is given by the black compass arrows ( n and e ) .
corner spaxels and spaxels corresponding to broken / unused optic fibres are depicted in black .
although the spatial resolution is low , rotation can be clearly seen in some galaxies , while no such features are seen on others .
we can not detect kinematically decoupled cores ( kdcs ) and double @xmath3 peaks ( 2-@xmath3 ) as in @xcite , because our spatial resolution is too coarse .
if we try to detect srs from the velocity maps by eye , we identify at most six : these are galaxies 4 , 8 , 12 , 20 , 26 and 27 .
the overall fraction of srs in the sample would then be 0.20 , in line with what was found in the virgo core @xcite .
however we are subject to contamination from face on discs appearing as srs , which increases .
we also highlight five more objects which , despite exhibiting large scale rotation , have misaligned kinematic axes , a feature more common in srs than in frs @xcite : these are galaxies 1 , 3 , 5 , 9 , 17 and 25 .
galaxies 3 and 17 have very high ellipticities , and are thus unlikely to be srs .
galaxy 5 has high velocity dispersion , and also contains an inner disc ( r = 1.5 kpc ) in the hst imaging .
@xcite introduced the estimator @xmath52 to measure the projected specific angular momentum of galaxies and @xcite further show how the combination of @xmath52 and ellipticity @xmath30 conveniently captures the kinematic boundary between srs and frs .
@xmath52 is defined as @xmath53 where @xmath34 , @xmath54 , @xmath55 and @xmath56 are the flux , distance from the galaxy centre , velocity and velocity dispersion of the i - th spaxel ; the sum is conducted over all spaxels inside some subset @xmath57 of the ifu footprint . @xcite
define as the value of @xmath52 computed inside the ellipse of area @xmath58 ( see section [ sec : data.data.analysis.photometry ] ) . in our study
however that ellipse may either not comprise enough spaxels to reliably measure , or be too large to fit inside the ifu footprint .
therefore we used existing ` sauron ` data to estimate how our particular observing setup affects the measured value of @xmath52 .
the original ` sauron ` sample covers a wide range of etgs types @xcite , and its data is publicly available .
we use it to simulate observations with flames / giraffe , in order to determine how distance and reduced spatial resolution affect measurements of @xmath52 . for each galaxy
we created a kinematic model using ` k`inemetry @xcite ; each model was then projected to the distance of abell 1689 and convolved with a seeing of 0.8@xmath11 , before being `` observed '' with flames / giraffe .
we created 10000 realizations of each model , adding gaussian errors of 15 and 17 for @xmath46 and @xmath3 respectively ( section [ sec : stellar.kinematics ] ) , and proceeded to measure @xmath52 for each of them . in their @xmath52 vs @xmath30 diagram , @xcite plot values computed on the same ` sauron ` spectrograph images , at the same spatial scale of 1 . for the small galaxies in our sample
however , covers just a few pixels whereas the large galaxies have larger than the field of view of the ifu . for this reason we can not follow the atlas@xmath0 prescription precisely .
we therefore introduced @xmath23 , defined as the value of @xmath52 computed using all the available spaxels in the ifu field of view and show through simulation of the sauron results that it is a satisfactory proxy for .
[ fig : delta.lambda.vs.re ] shows @xmath59 plotted against , where @xmath59 is defined as the difference between @xmath23 and the value of of @xcite .
we can use this information to determine the correction and the uncertainty that we need to apply to @xmath23 to obtain .
it is clear how our ability to recover the true value of improves with increasing . to make use of this information we separate the sample into three groups , based on ( the divisions are at values of 1.15@xmath11 and 1.70@xmath11 which naturally divide the abell 1689 sample and are shown as vertical dashed lines in fig . [
fig : delta.lambda.vs.re ] ) .
we find the following biases ( mean offset ) and systematic errors ( dispersion ) : for galaxies with @xmath60 @xmath61 ; for galaxies with @xmath62@xmath63 , @xmath64 , and for the remainder , @xmath65 . .
@xmath59 is the difference between @xmath23 ( computed from simulated observations of ` sauron ` data with giraffe ) and the value given in @xcite , plotted against .
slow / fast rotators are denoted by red / blue dots , and classified as in @xcite ( upper panel ) . in the lower panel we show the values of for our sample of galaxies in abell 1689 , where the symbol colour indicates srs ( red ) and frs ( blue ) , according to a classification done using the @xmath52-@xmath30 plot ( fig .
[ fig : lambda.vs.epsilon ] ) , as described in section [ sec : lambda.measurements ] . dashed vertical lines define regions over which we estimate biases and systematic errors ( see section [ sec : effect.pixel ] ) . ]
we corrected @xmath23 according to the biases measured , summing the systematic errors in quadrature to the random errors .
this correction takes into account both the different spatial scale between @xmath23 and and the different spatial resolution between @xmath23 and . in fig .
[ fig : lambda.vs.epsilon.sauron ] we plot simulated values of @xmath23 against published values of ( from * ? ? ?
despite the aforementioned differences , there is little ( 10% ) misclassification in our diagram , especially at high values of .
we can calculate the probability distribution for the number of srs ( galaxies below the line defined by @xmath66 and the green line in fig .
[ fig : lambda.vs.epsilon.sauron ] , * ? ? ?
this is most easily done with a monte carlo approach . for each galaxy
we assume gaussian errors in @xmath52 , truncated so that @xmath67 and sample 100000 times .
the resulting probability distribution is gaussian - like and we find 12.3 @xmath26 1.7 slow rotators , where the true value is 12 .
this justifies both our choice of @xmath23 to substitute for , and the use of @xmath30 computed at a different resolution and radius than @xmath23 .
vs for our simulated flames / giraffe observations of the ` sauron ` sample of galaxies .
the green lines , which separates srs ( below it ) from frs , has equation @xmath68 .
red and blue dots denote srs and frs respectively , according to the original ` sauron ` classification @xcite . while @xmath23 has been measured on the redshifted and resampled data , the values of @xmath30 on the x axis
are the original values published in @xcite . despite the latter being measured on much higher resolution than @xmath23 , and at a different radius ,
the impact on the classification is low .
misclassified galaxies correspond either to red dots above the green line and blue dots below it . ] in fig .
[ fig : lambda.vs.epsilon ] we show the @xmath23 vs plot for our abell 1689 data .
given the simulation in the previous section , the values of @xmath23 have been corrected by -0.06 , -0.01 and 0.01 for galaxies with @xmath69 , @xmath62@xmath63 and @xmath70 respectively .
the errors include both the formal random error ( from ppxf ) and the systematic error ( 0.09 , 0.04 and 0.02 for the three ranges of from the previous simulations ) .
given these errors we can calculate the probability distribution for the number of srs , as done previously for the simulated ` sauron ` data .
we find 4.5 @xmath26 1.0 slow rotators , corresponding to = 0.15 @xmath26 0.03 .
galaxy number 9 , which has peculiar photometry and an uncertain value of , has no effect on the result , because its value of @xmath23 is greater than @xmath71 0.25 ( the maximum allowed for any sr ) by more than 3@xmath3 .
@xcite warn about using only @xmath52 to assign a galaxy to either the slow or fast rotator class .
the discrepancy between the `` by eye '' classification and the classification here bolsters that warning .
however , when studying galaxies beyond the local universe , such a detailed analysis as was done by the atlas@xmath0 team is unfeasable . we are thus forced to rely on a statistical approach .
vs for all target galaxies in abell 1689 .
the green line has equation @xmath72 and separates fast rotators ( blue dots above it ) from slow rotators ( red dots below it ) .
error bars are dominated by the systematic error ( section [ sec : lambda.measurements ] ) .
notice that we corrected the measured value of @xmath23 by subtracting -0.06 , -0.01 and 0.01 , depending on for each galaxy ( see again section [ sec : lambda.measurements ] ) .
the solid magenta line represents the edge - on view of axisymmetric galaxies with @xmath73 , while the black dashed lines represent the trajectories of 6 of these galaxies ( with @xmath74 ) as their viewing angle goes from edge - on ( on the magenta line ) to face on ( towards the origin ) . for more information on how these models were constructed
see @xcite ] over the etgs population in atlas@xmath0 , including virgo cluster ( green circles , solid green line ) , as given in @xcite , and fraction of slow rotators in our sample of abell 1689 galaxies ( red circles , solid red line ) .
numbers at the top are the total number of galaxies in that bin , with the same color code .
the error bars for the abell 1689 points represent the uncertainty in the slow rotators classification , as estimated in section [ sec : lambda.measurements ] .
the green square is the value of @xmath75 that we measure resampling virgo using our sample luminosity function .
the error bars are smaller than the marker size .
the red square is the average fraction of slow rotators found in our sample .
the lower , smaller circles and dashed lines are the fractions computed with respect to the total cluster population , for atlas@xmath0 ( green ) and abell 1689 ( red ) ; this is an estimate based on spirals and blue ellipticals counts . ] for each galaxy in the sample we computed the local environment density following @xcite .
we defined @xmath76 as the number density inside the circular area centred around the target galaxy and encompassing three other galaxies .
density estimates were done using only valid targets in the catalogue of section [ subsec : sample.selection.effects ] .
we applied a constant field correction of 0.49 gal arcmin@xmath25 , measured averaging data from one hundred 1 arcmin@xmath77 fields from sdss dr7 @xcite , without correcting for cluster / groups contamination . for comparison ,
the minimum value found in our sample is 3.83 gal arcmin@xmath25 . in fig .
[ fig : frac.vs.density ] we show versus @xmath76 ; we compare it to the results of the atlas@xmath0 survey , and in particular to @xcite .
the virgo core corresponds to the densest bin in atlas@xmath0 , with @xmath78 , double that typically found in less dense environments ( @xmath79 ) .
we probe environments with values of @xmath80 between @xmath81 and @xmath82 : the minimum is comparable to the core of virgo , and the maximum is 1.7 dex higher . in this sense
our work starts exactly where atlas@xmath0 finished .
we find a sharp increase in with projected density , ranging from @xmath83 in the least dense environment to @xmath84 in the densest environment .
errors due to misclassification , albeit large , show that the densest bin in abell 1689 has a higher fraction of srs than virgo core ( fig .
[ fig : frac.vs.density ] ) .
the intermediate bin has a value of compatible , within the errors , with both the field - group environments and overall virgo cluster value but is however lower than the virgo core . in the least dense bin is lower than atlas@xmath0 field and group values . however , considering the whole abell 1689 sample , we find for an average value of @xmath85 that the sr fraction is @xmath2 ( red square in fig .
[ fig : frac.vs.density ] ) , which is the same as the overall sr fraction in the virgo cluster , when sampled in the same way ( green square ) . furthermore , both values are similar to the field and group samples in atlas@xmath0 , suggesting little to no difference in when it is averaged over the whole cluster .
in order to assess the robustness of our result , it is important to study the relation between the sample of 30 galaxies and its parent population .
our sample selection , limited by both observing and instrument constraints , biases our study in different ways . in this section
we discuss how these effects change .
having in mind the atlas@xmath0 study of the virgo cluster as a point of comparison , we determined the abell 1689 k - band lf .
we took the r@xmath12-band catalogue ( section [ sec : data.data.analysis.photometry ] ) and , following @xcite , applied a k - correction to the gmos @xmath86-band magnitudes .
the results have been converted to k@xmath87 band ( and vega system ) , using @xcite models , where we assumed an age of 10.4 gyr @xcite and passive evolution .
we finally applied a cut at @xmath88 , thus matching atlas@xmath0 parent sample selection .
the result is shown in fig .
[ fig : lum.func ] ( blue circles ) , where we compare it with the cluster rs ( as determined by * ? ? ? * red diamonds ) and our sample ( yellow squares ) .
the virgo etg lf is also plotted ( green triangles ) . knowing the k - band magnitudes of our sample
, we can show as a function of magnitude . in fig .
[ fig : lum.func.frac ] , the value of observed in our sample ( red ) is compared to the fraction of srs over the etgs population of virgo ( green ) .
the two are , within the errors , remarkably similar ; however we do not reach magnitudes beyond @xmath89 to probe the faint srs . band and r@xmath12 band imaging .
r@xmath12 band magnitudes have been converted to @xmath90 band as described in section [ sec : discussion ] .
red diamonds trace the cluster rs , as determined by @xcite .
green triangles represent virgo etgs lf , from atlas@xmath0 survey ( data available at www-astro.physics.ox.ac.uk/atlas3d/ ) ] ) . ]
our sample falls entirely on the red sequence ( rs ) , a property that was not sought after .
we know that the rs does not necessarily trace the morphological etgs population , as it can include red spirals and omit blue ellipticals .
how many etgs lying off the rs have we left out of our sample ? a rich , relaxed cluster like abell 1689 comprises a very small fraction of spirals , particularly in the core .
in fact the ratio between the rs lf and the cluster lf goes from 1 at the bright end to 0.70 at = -22.5 @xmath91 .
this means that including 10 `` blue '' galaxies in the faintest magnitude bins would remove the rs bias , leaving us with a color `` fair '' sample .
these faint galaxies are more likely to be frs ( fig .
[ fig : lum.func.frac ] ) , so the bias introduced by selecting galaxies on the rs leads us to overestimate .
in fact , if we assume that these ten galaxies are all frs , and that they are distributed spatially much like the observed targets , we can determine the kinematic morphology - density relation for an unbiased sample ( with respect to color ) which we show in fig .
[ fig : frac.vs.density ] ( smaller red dots and red dashed line ) , with the relation from atlas@xmath0 ( smaller green dots and green dashed line ) . the good overlap between etgs and the rs in abell 1689 causes to stay the same , whether the fraction is computed against the rs or overall galaxy population .
this is not true in a less relaxed , spiral rich cluster like virgo ( fig .
[ fig : frac.vs.density ] ) . we know that in virgo , varies as a function of ( fig . [
fig : lum.func.frac ] ) , and that the lfs of virgo and abell 1689 are different , in that virgo is relatively richer in brighter objects ( fig .
[ fig : lum.func ] ) .
since our sample is not fully representative of the abell 1689 population , what bias does this introduce in the measured value of ?
a rigorous answer to this question is impossible , because we do not know if the sr lf varies as a function of redshift and/or environment . in particular
, virgo is a small and dynamically young cluster , whereas abell 1689 is a massive , relaxed system .
however , using a simulation , we can estimate what sr fraction we would measure in virgo with the same selection effects present in our abell 1689 sample .
let us assume that the sr fraction as a function of magnitude is the same in abell 1689 and virgo ( reasonable given fig .
[ fig : lum.func.frac ] ) and that the galaxies were selected only on their magnitude and no other properties ( not true , as discussed in section [ sec : sample.selection ] , but reasonable given the substitutions required to comply with proximity constraints ) .
using the actual abell 1689 lf , we drew random subsamples from the atlas@xmath0 virgo etg population .
these samples yielded = @xmath92 , in agreement with the actual virgo value of @xmath93 . thus despite being biased towards brighter galaxies , we should measure the same ; this is because we sample down to , but not including , the faintest magnitude bin of atlas@xmath0 , where suddenly increases .
we remark that the distribution of projected ellipticities of our sample is different from that observed in virgo . since the former is richer in round objects , and since srs generally appear rounder , it follows that our sample could be biased towards higher values of .
it may well be that the clusters @xmath30 distributions are different , in which case a higher fraction of round objects may increase . another possible source of bias
is the intrinsic shape of abell 1689 : according to @xcite and @xcite , abell 1689 is elongated along the line of sight , so that its measured @xmath76 is higher than what we would observe from another point of view . if the cluster length along the line of sight direction were @xmath94 times longer than the diameter of the sky projection , then the value of @xmath76 observed would be @xmath95 times the unbiased value . since the maximum reasonable value of @xmath94 is @xmath96 , in fig .
[ fig : frac.vs.density ] @xmath97 is overestimated by at most @xmath98 , which does not significantly affect our results .
finally we remark that the corrections to @xmath23 that we derived in section [ sec : effect.pixel ] increase ; had we applied no correction , we would have 3.8 @xmath26 1.0 srs , so an even lower value of .
abell 1689 has a higher average density than virgo , but the same value of . inside the cluster , in agreement with the findings of @xcite , rises with projected density . in the least dense region is smaller than the atlas@xmath0 field / group value .
given the low number of galaxies per bin , we can not rigorously claim that this is representative .
however , a similar `` depletion '' is observed in the outskirts of virgo cluster @xcite .
one explanation could be that massive srs are driven by dynamical friction towards the centre of the cluster .
if these were originally distributed in the cluster like other galaxies , dynamical friction would reduce their orbital velocity and radius .
since this process is more effective on more massive galaxies , it would concentrate srs ( more massive on average ) with respect to other galaxies .
we demonstrated the use of flames / giraffe in ifu mode to perform a survey of 30 galaxies in abell 1689 at @xmath99 .
the data has sufficient quality and spatial resolution to classify the majority of targets as either srs or frs . in summary : 1 .
we find , in agreement with atlas@xmath0 results , that srs populate the high luminosity end of the lf ; the sr lfs measured from the virgo atlas@xmath0 sample and our abell 1689 sample are identical down to @xmath100 mag . 2 .
the fraction of slow rotators in our sample is = @xmath2 .
if we apply the same selection criteria to all virgo galaxies in atlas@xmath0 , we find the same fraction ( assuming that the distribution of srs with magnitude is the same in both clusters ) .
this indicates that is not affected by the average number density of the cluster . both abell 1689 and virgo average are in line with the atlas@xmath0 value for field and group environments .
3 . the fraction of srs increases towards the denser , central region of the cluster .
this is in agreement with what was found in virgo , where srs concentrate in the cluster core .
this could be a consequence of dynamical friction , as srs dominate the high mass end of the galaxy population .
it is important to expand this study , both to further study abell 1689 down to lower luminosities and increase the number of observed clusters , to quantify the scatter in .
we thank michele cappellari and john magorrian for their helpful comments and suggestions .
we are particularly grateful to the referee eric emsellem , whose comments significantly improved this paper .
we wish to express our thanks to jonathan smoker , who was instrument scientist for flames when we took the observations .
fde acknowledges support from the physics department , university of oxford and travel support from merton college , oxford .
rld acknowledges travel and computer grants from christ church college , oxford , and support from the royal society in the form of a wolfson merit award 502011.k502/jd .
rld also acknowledges the support of the eso visitor programme which funded a 3-month stay in 2010 .
edb was supported by the grants cpda089220/08 and 60a02 - 5934/09 of padua university , and asi - inaf i/009/10/0 of italian space agency , and by a grant of accademia dei lincei and royal society .
edb acknowledges the sub - department of astrophysics , department of physics , university of oxford and christ church college for the hospitality while this paper was in progress .
this paper is based on observations made with eso telescopes at the paranal observatory under programme i d 083.b-0681(a ) and the nasa / esa hubble space telescope ( proposal i d 9289 ) , obtained from the data archive at the space telescope science institute .
stsci is operated by the association of universities for research in astronomy , inc . under nasa contract
nas 5 - 26555 .
this work also used data obtained at the gemini observatory ( proposal ids gn-2001b - q-10 & gn-2003b - dd-3 ) , 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 science and technology facilities council ( united kingdom ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , ministrio da cincia , tecnologia e inovao ( brazil ) and ministerio de ciencia , tecnologia e innovacion productiva ( argentina ) . | we present flames / giraffe integral field spectroscopy of 30 galaxies in the massive cluster abell 1689 at z = 0.183 . conducting an analysis similar to that of atlas@xmath0 , we extend the baseline of the kinematic morphology - density relation by an order of magnitude in projected density and
show that it is possible to use existing instruments to identify slow and fast rotators beyond the local universe .
we find @xmath1 slow rotators with a distribution in magnitude similar to those in the virgo cluster .
the overall slow rotator fraction of our abell 1689 sample is @xmath2 , the same as in virgo using our selection criteria .
this suggests that the fraction of slow rotators in a cluster is not strongly dependent on its density .
however , within abell 1689 , we find that the fraction of slow rotators increases towards the centre , as was also found in the virgo cluster .
[ firstpage ] galaxies : kinematics and dynamics , galaxies : clusters : individual : abell 1689 |
ir spectrometry is a powerful method to characterize the hundreds of exoplanets discovered by indirect methods , and ultimately to search for life on telluric ones , as it allows to evidence atmospheric signatures of interest ( h@xmath3o , co@xmath3 , ch@xmath4 , ... ) boosted by the thermal emission of the planet .
the main challenge is to isolate the planet from its parent nearby star ; one of the very few possible methods is spaceborne nulling interferometry , allowing implementing very efficient coronagraphs at the diffraction limit . within this framework ,
several space projects such as darwin@xcite , tpf - i@xcite , fksi@xcite or pegase@xcite , have been proposed in the past years .
most of them are based on a free - flying concept to create hectometric pupils with an acceptable launch mass .
but the price to pay is to very precisely position and stabilize the spacecrafts relatively to each other and also with respect to an inertial direction , in order to achieve a deep and stable extinction of the star .
understanding and mastering all these requirements is a difficult challenge and a key issue towards the feasibility of these missions .
thus , we decided to experimentally study this question and focus on some possible simplifications of the concept .
since 2006 , persee ( pegase experiment for research and stabilization of extreme extinction ) laboratory test bench is under development by a consortium composed of centre national dtudes spatiales ( cnes ) , institut dastrophysique spatiale ( ias ) , observatoire de paris - meudon ( lesia ) , observatoire de la cte dazur ( oca ) , office national dtudes et de recherches arospatiales ( onera ) , and thales alenia space ( tas)@xcite .
it is mainly funded by cnes r&d .
persee couples an infrared wide band nulling interferometer , based on a modified mach - zehneder ( mmz ) and a geometrical achromatic phase shifter , a disturbance simulator and a cophasing system based on a lqg tracker for the optical path difference ( opd ) .
the first results of the bench were already presented in ref . ; in this paper , we focus on complete it with the latest improvements and results of the bench . after a short description of persee goals and setup ( focused on recent changes ) , we will present in sec .
[ sec : calibration ] its various calibration procedures .
recent results are reported in sec .
[ sec : lqg validation ] for the cophasing system and in sec .
[ sec : nulling ] for the nuller .
this section provides the reader with a very brief description of persee .
much more details can be obtained by the interested reader from refs . .
the goal of persee is not to reach the deepest possible nulling . starting from a state of the art nuller of the 2006 - 2007 period
, it is an experimental attempt to better master the system flowdown of the nulling requirements both at payload ( instrument main optical bench ) and platform levels ( satellites ) .
the balance of the constraints between those two levels is a key issue .
the more disturbances the payload can face , the simpler the platforms are , the lower the cost .
the general idea is hence to simplify as much as possible the global design and reduce the costs of a possible future space mission .
the detailed objectives have been described in ref . .
our main requirement is to reach a @xmath5 nulling ratio in the [ 1.65 - 2.45 ] band ( 37% spectral bandwidth ) with a @xmath6 stability over a 10 h time scale .
another important requirement is to be able to find and stabilize fringes which have an initial drift speed ( as seen from the interferometer core ) up to 150 .s^-1^ , as this can greatly simplify the relative metrology and control needs .
we want to study and maximize the rejection of external disturbances introduced at relevant degrees of freedom of the optical setup by a disturbance module simulating various environments coming from the platform level . in the original design ,
three sources were used : a laser diode at 830 nm and a sled at 1320 nm for the dual band metrology system , and a black body used in the band @xmath2 $ ] for the science path .
the three sources were coupled in single mode optical fibers , that were joined close to each other at the focus of the collimator , to simulate the star . that configuration was not efficient , because the snr in the camera was too low ,
the science spectrum was polluted by the sled flux , and the contrast was too low in the fringe sensor ( fs ) .
so we had to change the configuration at the focus of the collimator , to have only one fiber that simulates the star .
unfortunately , the black body was not powerful enough for the fs detectors to be used for metrology in addition to science .
the adopted solution was to use an infrared supercontinuum source , a polychromatic source powerful enough to emit both in metrology and science paths .
its pulse rate is at 20 mhz , much higher than the loop frequency of the fs ( 1 khz ) and the camera ( 100 hz ) .
it was not possible to connectorize the supercontinuum fiber , so we designed a simple injection bench with two off axis parabola , to inject the flux in a single mode pcf fiber .
the flux was even too high for the science camera , so we have to defocalize the pcf fiber to reduce the coupled light .
another modification of the source module , presented in that conference @xcite , will include a planet and a exozodical disk . on the science spectrometer , two flux
are measured in order to perform the calculation of the nulling rate on each spectral channel @xmath7 , by @xmath8 where @xmath9 is the flux of the destructive interference , directly obtained at the nulled output iii of the mmz .
@xmath10 is the flux of the same output , in the case of a constructive interference .
it can not be measured simultaneously with @xmath9 , so we firstly used the flux of the related constructive interferometric output ii of the mmz .
this solution has two disadvantages : firstly , the single - mode fibers of outputs ii and iii must be perfectly aligned , in order to have optimal injection in both outputs , and the two fibers should not be subject to differential displacements .
secondly , the output ii is a coherent flux , so it is sensitive to opd variations . especially in that case , the differential phase at the output ii is not zero but 21 , thus the output ii is not strictly constructive and does not measure exactly @xmath10 . for those reasons , we modified that reference channel , by positioning the reference fiber in front of the m@xmath11 collimator , selecting a part of the unused flux of the source , before the interferometric train .
the flux @xmath12 of that output is not equal to the flux @xmath10 , but they are proportional : @xmath13 . also , the dynamics of the camera ( @xmath14 ) is not sufficient for the direct measurement of the desired nulling rate , that is why we use an optical density of transmission @xmath15 when we calibrate the coefficient @xmath16 , and we remove it for the measurement of @xmath9 .
the goal is to reduce the contrast of the measurement to fit in the range of the camera . finally , eq . ( [ eq : null ] ) becomes @xmath17
persee bench has two control loops in piston and tip / tilt , which require a few calibrations , as well as the scientific path
. there are three types of calibrations . the first type ( sec .
[ sec - spec]-[sec - actu ] ) depends only on the configuration , and need to be applied just once .
the second type of calibrations ( sec .
[ sec - tt - ref]-[sec - cam ] ) depends slightly on the environment , especially thermal drift .
the last type of calibrations ( sec .
[ sec - tt - cor ] , [ sec - opd - ref ] ) is strongly affected by thermal drift . to calibrate the spectral bands
, we used the method of fourier transform spectroscopy ( fts ) .
the fts uses the properties of the fourier transform to determine the spectrum of a source with an interferometer . in most cases , the interferometer is a michelson with a mirror mounted on a translation , to scan the interference fringes . with a mmz ,
the situation is slightly different . for each wave number @xmath18 ,
the intensity as a function of @xmath19 is @xmath20 , \label{eq : fts1}\ ] ] with @xmath21 and @xmath22 .
@xmath23 is the spectrum that we want to determine .
then the intensity measured at the output of the interferometer is @xmath24 then the fourier transform of @xmath25 , @xmath26 , is @xmath27 with @xmath28 the dirac delta function , and @xmath29 the total flux .
we find the case of the michelson interferometer , when the contrast is maximum and the phase is zero : @xmath30 . in the case of the mmz , we do not measure the spectrum of the source , but the product of the spectrum and the contrast . [
cols="^ " , ] figure [ fig - null-100s ] presents the first long - term measurement of the null rate over 100 s. we can see on the left figure that the null rate has a higher value than the best result : @xmath31 , but shows a very good stability over 7 h : @xmath32 .
that measured stability is a decade better than the specification : @xmath33 . in fig .
[ fig - null-100s ] ( right ) , the null rate spectral profile looks like the one presented in fig .
[ fig - best - null ] : the null rate is better at the center of the spectrum than at the edges .
the contribution of opd and flux mismatch are also higher at shorter wavelengths , and the contribution of polarization should be higher at longer wavelengths .
the chromatic dispersion of the null rate is between @xmath34 and @xmath35 , so the amplitude is 7 times higher than on the best result .
rccc + contribution & specification & best null ( 100 s ) & null rate ( 7 h ) + @xmath36 & @xmath5 & @xmath37 & @xmath38 + @xmath39 & @xmath40 & @xmath41 & @xmath42 + @xmath43 & @xmath40 & @xmath44 & @xmath45 + @xmath46 & @xmath47 & @xmath48 & @xmath49 + @xmath50 & @xmath6 & @xmath51 & @xmath52 + @xmath53 & @xmath54 & @xmath55 & @xmath56 + @xmath57 & @xmath54 & @xmath58 & @xmath56 + @xmath59 & @xmath60 & @xmath61 & @xmath62 + @xmath63 & @xmath6 & - & @xmath64 + table [ tab - null-100s ] summarizes the null rate analysis developped in this section . in the near futur , we should be able to achieve an average null rate of @xmath6 over 10 h , with a stability better than @xmath65 , i.e. a decade under the original specifications of the bench .
as we also demonstrate the capability of correcting injected disturbances , we should be able to achieve that result even with typical disturbances on the bench .
persee showed a very high quality and performance well beyond the initial specifications . the large majority of goals has already been demonstrated .
we have already presented some important lessons in ref . , about thermal stability and control loop feasibility .
however , we can complete these remarks by other conclusions .
firstly , we elaborated a calibration scheme that can be applied to a space mission , although it must take into account the linear - parabolic drift.that procedure is fully automatic , and allows also to make an accurate analysis of the instrument .
secondly , we showed that a lqg controller can manage the disturbances of the opd , to achieve a very deep null rate .
an experimental sub - nanometer control has been demonstrated , even in the case of strong disturbances ( more than 100 nm rms reduced to 1 nm rms ) .
that result was very important for persee , because we work in the near infrared , which is more sensitive to opd variations than thermal infrared used in projects like darwin .
finally , we showed that thanks to a a very efficient design , we obtained null rates more than 10 times better than the initial conservative goal of @xmath5 , thus close to the requirement of a nuller for the observation of telluric planets .
we achieved the deepest null ever achieved by a near - infrared nulling interferometer : @xmath0 on a 37% bandwidth , between 1.65 and 2.45 .
persee bench shows also a very good stability over 100 s , and also over 7 hours .
we still have to confirm this deep null rate over 10 hours , with the same type of stability , and with injected disturbances .
the last steps will be the correction of linear - parabolic drifts , together with the use of a new source module , which includes a planet and an exozodiacal disk , presented in ref . during this conference . | stabilizing a nulling interferometer at a nanometric level is the key issue to obtain deep null depths .
the persee breadboard has been designed to study and optimize the operation of a cophased nulling bench in the most realistic disturbing environment of a space mission .
this presentation focuses on the current results of the persee bench . in terms of metrology
, we cophased at 0.33 nm rms for the piston and 80 mas rms for the tip / tilt ( 0.14% of the airy disk ) . a linear quadratic gaussian ( lqg ) control coupled with an unsupervised vibration identification allows us to maintain that level of correction , even with characteristic vibrations of nulling interferometry space missions .
these performances , with an accurate design and alignment of the bench , currently lead to a polychromatic unpolarised null depth of @xmath0 stabilized at @xmath1 on the @xmath2 $ ] spectral band ( 37% bandwidth ) . |
most of physicists believe that the existence of magnetic monopoles had been ruled out by experiments .
however , experiments only indicated that the flux of magnetic monopoles on the earth is too low to be observed . in the primordial universe ,
the electromagnetic interaction between magnetic monopoles and plasma is so strong such that magnetic monopoles might store up in the center of quasars and active galactic nuclei ( agn ) during the collapsing process of the original giant nebulae , including at the collapsed core of the galactic center ( gc ) . due to the rubakov - callan ( rc ) effect@xcite
, magnetic monopoles may catalyze nucleon decay and is invoked as the energy source of quasars and agn .
the model of supermassive object with magnetic monopoles ( smomm ) @xcite , presented by one ( qp ) of the authors , gave five predictions , some of which match up with the recent astronomical observations quantitatively .
it may be the signals for existence of magnetic monopoles in the supermassive objects . from the symmetry and beauty of the dirac equation ,
dirac @xcite suggested the existence of magnetic monopoles which is the most natural explanation of the quantization of electric charge .
it is a fundamental problem in physics whether magnetic monopoles exist or not .
the plan of this letter is as follows . in section 2
, we will sketch the model of smomm and its predictions . in sections
@xmath0 , we will show some predictions of our model in agreement quantitatively or basically with some new astrophysical observations : a strong magnetic field around the gc , the observation of the rate of the emitted positrons , and the frequency of the spectrum peak of the thermal radiation from the gc . in the last section
, we will discuss and try to explain two recent observations with the model of the stars with the magnetic monopoles : the highly super - luminous supernova asassn-15lh and the gravitational waves gw150914 .
the model of smomm @xcite suggested that an amount of magnetic monopoles stored up in the center of quasars and agns , including at the collapsed core of the gc during the collapsing process of the original giant nebulae in the primordial universe because the electromagnetic interaction between magnetic monopoles and plasma was so strong . due to the rc effect @xcite , nucleon decay may be catalyzed by magnetic monopoles , @xmath1
@xmath2 in this model , the rc effect is invoked as the energy source . in the case of gc
, the smomm is located at the center with the radius of about @xmath3 ( @xmath4 is its schwarzschild radius ) .
the total mass of the smomm is derived to be @xmath5 ( now taken as @xmath6 ) from the observed luminosity of sgr a@xmath7 ergs @xmath8 ) .
the gravitational effect around smomm in the gc is similar to the model of supermassive black holes . in the black hole model , accretion of matter results in the huge luminosity , but the energy source in the model of smomm is supplied by the rc effect .
the main predictions of the model of smomm are as follows @xcite .
the production rate of positrons emitted from the smomm in the model is @xmath9 .
high - energy gamma - ray radiation has energy higher than 0.5 mev
. the integrated energy of these radiations would be much greater than both the bolometric luminosity and the energy of positron annihilation line .
\3 . the radial magnetic field at the surface of the smomm
is estimated to be @xmath10 .
the strong radial magnetic fields of the high - speed rotating smomm transforms a strong electric field for a distant observer in the rest frame .
a variety of produced particles ( @xmath11 , @xmath12 , @xmath13 ) would be accelerated by the strong electric field to very high energy , say @xmath14ev or greater .
we predict that these could just be the observed ultra - high - energy cosmic rays which have an initial energy of several hundred mev produced from the smomm .
the surface temperature of the smomm is derived to be about 121 k and the corresponding spectrum peak of the thermal radiation is at @xmath15 in the sub - mm wavelength regime .
the recent observation @xcite in 2013 indicated that * there is a dynamically important magnetic field near the black hole*. in particular , at @xmath16 the lower limit of the outward radial magnetic field near the gc is @xmath17 \left[\displaystyle \frac { n_{0}}{26 cm^{-3}}\right]^{-1 } mg , \label{3}\ ] ] where @xmath18 is the number density of electrons there , and rm denotes the measurement of the faraday rotation near the gc .
the lower limit of the observed data is in agreement with the the prediction 3 in the model of smomm because the magnetic field strength decreases as the inverse square of the distance from the source and has @xmath19 at @xmath16 . up to now no other physical mechanism
can produce this strong radial magnetic field . as analyzed in @xcite ,
jet magnetic field and accretion disk luminosity are tightly correlated over seven orders of magnitude for sample of 76 radio - loud active galaxies " .
they pointed out that the black hole models may require significant changes " , and
models of the galactic center accretion disk may also need to be revised , as a dynamically important magnetic field has been reported @xcite within a distance of @xmath20 from the central black hole . "
new observation @xcite reported that the measured 511 kev line flux located at the gc at a distance of @xmath21 converts into an annihilation rate of @xmath22 .
the observed flux is compatible with previous measurements @xcite that have been obtained using telescopes with small or moderate fields - of - view , yet it is on the low side when compared to osse measurements . "
@xcite those observations are in agreement with the the prediction 1 in the model of smomm quantitatively .
a review paper @xcite pointed out that the radio flux density @xmath23 from the gc shows a flat - to - inverted spectrum .
i.e. , it raised slowly with frequency of the power peaking around @xmath24 hz in the sub - mm band .
the observed power peak is in agreement basically with the prediction 5 in the model of smomm .
the agreement of the predictions of our model of the smomm with three new astrophysical observations quantitatively or basically issues the signals for existence of magnetic monopoles .
we are looking forward to seeing more astrophysical observation which will meet the predictions of our model . at the beginning of this year
two important astrophysical observations were reported : the highly super - luminous supernova asassn-15lh and the gravitational waves gw150914 .
we believe that our model and its development can be suitable to meet the new observations .
the recent observation @xcite reported the discovery of asassn-15lh ( sn 2015l ) , which was interpreted as the most luminous supernova to be found . at redshift @xmath25 , asassn-15lh reached an absolute magnitude of @xmath26 and bolometric luminosity @xmath27 erg s@xmath28 , which is more than twice as luminous as any previously known supernova " . in the 4 months since first detection , asassn-15lh radiated @xmath29 ergs , challenging the magnetar model for its engine " . up to now , the supernova explosion mechanism has not been solved . in a model of a supermassive star with magnetic monopoles , the phenomena of asassn-15lh and the super - luminous supernova ( slsn )
may be explained naturally .
although a few magnetic monopoles may be stored in the core of stars during the process of star formation from a neutral hydrogen nebulae due to very weak interaction between magnetic monopoles and neutral hydrogens , the stars may capture some flight magnetic monopoles in the space @xcite .
the flux of the flight magnetic monopoles is @xmath30 where @xmath31 is the number density of the magnetic monopoles , @xmath32 is the average velocity of the flight magnetic monopoles in space , @xmath33 is the number density of the baryons in the interstellar space , and @xmath34 denotes the newton saturation value of @xmath35 , where @xmath36 is the total number of the flight magnetic monopoles and @xmath37 is the total number of nucleons .
@xmath38 is the value of @xmath39 in the interstellar space , and its upper limit is @xmath40 @xcite .
after formation of stars the total number of the magnetic monopoles captured by the stars from space is estimated to be where @xmath45 @xmath46 is the cross section of the rc effect @xcite , @xmath47 is the thermal velocity of the nucleons , and @xmath48 is the central number density of the baryons for the star .
the temperature would reach @xmath49k in the collapsed core of the supernova , i.e. , @xmath50 .
thus , as long as we assume that the initial mass of the progenitor of the slsn is greater than @xmath53 , then its radius before the supernova explosion was greater than @xmath54 , and the life of the super - massive star is only @xmath55 .
thus , the luminosity produced from the nucleon decay catalyzed by the magnetic monopoles in the collapsed core of the supernova , where the density of the core is much greater than the nuclear density , will be greater than @xmath56 ergs .
this luminosity is enough to make the slsn exploding , so that the slsn is explained naturally .
on september 14 , 2015 two detectors of the laser interferometer gravitational - wave observatory ( ligo ) simultaneously observed a transient gravitational - wave signal ( gw150914 ) @xcite .
this is the first direct detection of gravitational waves which was commonly explained as the first observation of a binary black hole system mergering to form a single black hole . on december 26 , 2015 , a second gravitational - wave event , gw151226 , was observed by the twin detectors of ligo@xcite . on february 16 , 2016
, @xcite reported that the fermi gamma - ray burst monitor ( gbm ) had revealed the presence of a weak transient source above 50 kev , 0.4 s after the gw event : `` this weak transient lasting 1 s does not appear connected with other previously known astrophysical , solar , terrestrial , or magnetospheric activity . its localization is ill - constrained but consistent with the direction of gw150914''@xcite .
they also note that the electromagnetic signal from a stellar mass black hole binary merger is not expected if the gbm transient is associated with gw150914 . based on their measurement of the fluence seen by gmb
, a luminosity of @xmath57 erg s@xmath28 is derived in hard x - ray emission between 1 kev and 10 mev . in another paper @xcite lyutikov analyzed the physical requirements in some detail that the possible observation of the electromagnetic signal contemporaneous with gw150914 imposes on the circum - merger environment , and found that the required physical parameters at the source exceed by many orders of magnitude what is expected in realistic astrophysical scenarios .
he concluded that fermi gmb signal contemporaneous with gw150914 is unrelated to the black hole merger .
the merger of two black holes can produce the gw only , not the gamma - ray burst simultaneously .
however , in our model of stars with magnetic monopoles , the merger of two supermassive neutron stars with the magnetic monopoles , ( they can not collapse to black holes through the rc effect ) , whose mass density reaches to @xmath58 at the center core , not only produces the gw , but also produces the gamma - ray burst whose luminosity may reach @xmath59 through the rc effect . after we have finished this paper ,
@xcite reported on june 15 , 2016 that they had made observations of the event gw151226 and candidate lvt151012 with both the fermi gamma - ray burst monitor and the large area telescope ( lat ) .
no electromagnetic counterparts were detected by either gbm or lat .
we are looking forward to the future detections of the gw which may accompany with the gamma - ray burst .
if yes , it will verify the existence of magnetic monopoles .
we would like to thank prof .
daniel wang , prof .
huang , prof .
p.f . chen and prof .
j. l. han for their help to inform us some new information of observations .
this work was supported in part by the national natural science foundation of china under grants 10773005 , 11565020 , and the counterpart foundation of sanya under grant 2016pt43 , the special foundation of science and technology cooperation for advanced academy and regional of sanya under grant 2016yd28 , and the natural science foundation of hainan province under grant 114012 .
abbott , b. p. , abbott , r. , abbott , t. d. , et al . , 2016 , physical review letters , 116 , 061102 abbott , b. p. , et al . , 2016 ,
lett . , 116 , 241103 c. callan , 1983 , nucl .
, 212 , 391 cheng , l. x. , leventhal , m. , smith , d. m. , et al .
, 1997 , apjl , 481 , 43 connaughton , v. , burns , e. , goldstein , a. , et al .
2016 , arxiv:1602.03920 p. a. m. dirac .
, 1958 , the principle of quantum mechanics , ( clarendon press , oxford , 1958 ) .
dong , s. , shappee , b. j. , prieto , j. l. , et al .
, 2016 , science , 351 , 257 r.p .
eatough , h. falcke , r. karuppusamy et al . , 2013 ,
nature , 591 , 391 falcke , h. , & markoff , s. b. , 2013 , classical and quantum gravity , 30 , 244003 kndlseder , j. , lonjou , v. , jean , p. , et al .
, 2003 , aap , 411 , 457 lazarides , g. , shafi , q. , & walsh , t. f. , 1981 , physics letters b , 100 , 21 lyutikov , m. , 2016 , arxiv:1602.07352 milne , p. a. , kurfess , j. d. , kinzer , r. l. , leising , m. d. , & dixon , d. d. , 2000 , american institute of physics conference series , 510 , 21 milne , p. a. , kurfess , j. d. , kinzer , r. l. , & leising , m. d. , 2001 , gamma - ray astrophysics , 587 , 11 parker , e. n. , 1970 , apj , 160 , 383 peng , q. , lie , z. and wang , d. , 1985 , scientia sinica ( serries a ) , 28 , 970 peng , q. & chou , c. , 2001 , apj , 551 , 23 purcell , w. r. , cheng , l .- x . ,
dixon , d. d. , et al . , 1997 ,
apj , 491 , 725 racusin , j. l. , burns , e. , goldstein , a. , et al . , 2016 , arxiv:1606.04901 rubakov , v. 1981 , jetp lett .
33,644 share , g. h. , leising , m. d. , messina , d. c. , & purcell , w. r. , 1999 , apjl , 358 , 45 zamaninasab , m. , clausen - brown , e. , savolainen , t. , & tchekhovskoy , a. , 2014 , nature , 510 , 126 | we summarize some predictions from the model of supermassive object with magnetic monopoles which match up with recent astronomical observations quantitatively . they may be the signals for existence of magnetic monopoles in the supermassive objects , such as one at the galactic center . |
there has been much recent debate on the star formation histories ( sfhs ) of low surface brightness disc galaxies ( lsbgs ; galaxies with @xmath1 band central surface brightnesses fainter than 22.5 magarcsec@xmath2 ) .
the best studied lsbgs are blue in the optical and the near - infrared ( near - ir ) @xcite , indicating a young mean stellar age and/or low metallicity .
their measured region metallicities are low , at around or below @xmath3 solar abundance @xcite .
morphologically , the best studied lsbgs have discs , but little spiral structure @xcite . the current massive star formation rates ( sfrs ) in lsbgs are an order of magnitude lower than those of high surface brightness ( hsb ) galaxies @xcite .
observations show that lsbgs have high gas mass fractions , sometimes even approaching unity @xcite . as
yet , there have been no co detections of lsbgs , only upper limits on the co abundances which indicate that lsbgs have co / ratios significantly lower than those of hsb galaxies @xcite .
all of these observations are consistent with a scenario in which lsbgs are relatively unevolved , low mass surface density , low metallicity systems , with roughly constant or even increasing sfrs @xcite .
however , this scenario has difficulty accommodating giant lsbgs ( as , indeed , this scenario was not designed with giant lsbgs in mind ) : with scale lengths typically in excess of 510 @xmath4 kpc , these galaxies are similar to , but less extreme than , malin 1 .
quillen & pickering , in an as yet unpublished work , obtained near - ir @xmath5 band imaging of two giant lsbgs .
they concluded that the central optical near - ir colours of their galaxies were compatible with those seen in old stellar populations ( such as e / s0 galaxies ) , and that the ( more uncertain ) outer colours were consistent with somewhat younger stellar populations .
another difficulty for this scenario is posed by the recent discovery of a substantial population of red lsbgs @xcite .
the optical colours of these galaxies are similar to those of old stellar populations , but the red colours could be caused by age or metallicity effects ( note that dust is not expected to be an important effect in most face - on lsbgs ; section [ subsec : dust ] ) .
either way , the existence of old or metal - rich lsbgs is difficult to understand if all lsbgs are unevolved and gas - rich . this same age - metallicity degeneracy plagues the analysis of the colours of blue lsbgs .
padoan , jimenez & antonuccio - delogu question the apparent youth of blue lsbg stellar populations : they find that lsbg optical colours are consistent with those of old , very low metallicity stellar populations .
this uncertainty caused by the age / metallicity degeneracy is partially avoidable ; for stellar populations with ongoing star formation , it is possible to learn something of their sfh using a combination of optical and near - ir colours .
essentially , it is possible to compare the sfhs of galaxies in a relative sense , using a kind of ` birthrate parameter ' relating the amount of recent star formation to the cumulative amount of previous star formation ( in this work , we parameterise the sfhs using an exponential sfh with a timescale @xmath6 ) . as a guide to interpreting these optical and near - ir colours
, we use the latest multi - metallicity stellar population synthesis models of e.g. bruzual & charlot ( in preparation ) and kodama & arimoto . the limiting factor in applying this technique
is typically the availability of near - ir imaging ; especially so for lsbgs , where the near - ir central surface brightness can be up to a factor of 500 fainter than the sky surface brightness at these wavelengths . in this paper , we present optical and near - ir imaging for a diverse sample of 26 lsbgs in order to explore their sfhs . in our study
we include examples of i ) the well - studied blue - selected lsbgs taken from de blok et al . ,
de jong & van der kruit and the eso - lv catalogue , ii ) the more poorly studied , intriguing red - selected lsbgs from oneil et al . , and iii ) the lsbg giants , taken from sprayberry et al . .
earlier results of this programme were presented in bell et al . , where the stellar populations in a subset of five red and blue lsbgs were explored ; red lsbgs were found to be much older and metal - rich than blue lsbgs , indicating that the two classes of galaxy do not share a common origin . here , we explore the differences in sfh between red , blue and giant lsbgs
. we also attempt to understand which , if any , galaxy parameters ( e.g. mass , or density ) affect the sfhs of lsbgs , driving the differences between e.g. red and blue lsbgs .
the plan of the paper is as follows .
the observations and data reduction are described in section [ sec : obs ] . in section [ sec : phot ] we present the photometry for our sample , compare with existing published photometry , and discuss the morphologies of lsbgs in the optical and the near - ir . in section [ sec : sfh ] we present the optical near - ir colours for our sample of galaxies , discussing them in terms of differences in sfh . in section [ sec :
disc ] , we elaborate on this colour - based analysis and explore trends in sfh with physical parameters
. finally , we present our conclusions in section [ sec : conc ] .
our sample of 26 lsbgs was selected with a number of criteria in mind
. they must be detectable using reasonable exposure times on 4-m class telescopes in the near - ir , and must span a wide range of observed lsbg properties , such as physical size , surface brightness and colour .
in addition , we selected galaxies with as much existing optical and data as possible .
our sample is by no means complete , but is designed instead to span as wide a range as possible of observed lsbg parameters .
our northern sample was selected from a number of sources @xcite to have moderately low published @xmath1 band inclination corrected central surface brightness @xmath7 magarcsec@xmath2 .
furthermore , the northern sample was selected to have major axis radii to the 25 @xmath1 magarcsec@xmath2 isophote larger than @xmath8 16 arcsec .
galaxies with only moderately low surface brightness were chosen as the near - ir sky background at temperate sites is very high due to the large thermal flux from the telescope and sky . in order to explore the properties of galaxies with even lower surface brightnesses , we imaged a small sample of galaxies using the south pole 0.6-m telescope .
the @xmath0 band sky background at the south pole is suppressed by a factor of @xmath9 compared to temperate sites , allowing unprecedented sensitivity for faint extended @xmath0 band surface photometry ( see e.g. nguyen et al . 1996 ,
rauscher et al . 1998 ) .
our southern hemisphere sample is selected from the eso - uppsula catalogue @xcite to have larger sizes and lower surface brightnesses compared to the northern sample : @xmath10 magarcsec@xmath2 , 65 arcsec @xmath11 r@xmath12 arcsec , inclination less than 67@xmath13 and galactic latitude @xmath14 to avoid excessive foreground galactic extinction ( where r@xmath15 denotes the galaxy half light radius ) .
our sample is described in further detail in table [ tab : samplepar ] .
galaxy types and heliocentric velocities were typically taken from the nasa / ipac extragalactic database ( ned ) .
galaxy distances @xmath16 were determined from velocities centred on the local group @xcite assuming a a hubble constant @xmath17 of 65 kms@xmath18mpc@xmath18 . for galaxies within @xmath8 150 mpc , we further take into account the local bulk peculiar motions @xcite .
typical distance uncertainties , corresponding to the uncertainties in the peculiar motions , are typically @xmath8 7 mpc .
gas masses were calculated using @xmath19 @xcite , where @xmath20 is the line flux in jy and @xmath21 is the line width in kms@xmath18 .
fluxes were taken from ned , except for the fluxes from de blok et al . , sprayberry et al . and oneil , bothun & schombert .
note that in table [ tab : samplepar ] all the galaxies without masses have not yet been observed in c1 - 4 and c3 - 2 , which were not detected with arecibo in 5 @xmath22 5 minutes and 12 @xmath22 5 minutes respectively ( these limits roughly correspond to 3@xmath23 upper limits on the gas fraction of @xmath8 15 per cent ) . in table [
tab : samplepar ] we have also presented the foreground galactic extinction in @xmath1 band , as estimated by schlegel , finkbeiner and davis .
we have checked the infra red astronomical satellite ( iras ) point source and small scale structure catalogues @xcite for our sample : only the seyfert 1 lsbg giant 2327 - 0244 was detected ( at both 60@xmath24 m and 100@xmath24 m ) . [
cols="<,^,^,^,^,^,^,^,^",options="header " , ] + @xmath25 relatively unconstrained + @xmath26 these parameters may be affected by the strong bar the surface brightness profile @xmath27 ( in linear flux units ) was fit using one of four possible profiles : an exponential disc profile only , an exponential bulge plus disc profile , a de vaucouleurs @xmath28 law bulge plus exponential disc profile , and a de vaucouleurs @xmath28 law bulge only .
the exponential profile has an extrapolated central surface brightness @xmath29 ( in linear units ) , and an exponential disc scale length @xmath30 .
the de vaucouleurs profile has a surface brightness @xmath31 in linear units at the half - light radius @xmath32 .
these exponential or @xmath28 law model profiles were convolved with a gaussian with a fwhm equal to the seeing quality before fitting .
the sky level was not permitted to vary when performing these fits .
independent fitting for each passband was carried out using the method of levenberg and marquardt @xcite .
the reduced @xmath33 statistic was used to determine the quality of the fit ( using the measured rms in each ellipse as the error in the flux in that ellipse ) ; @xmath33 was minimised for each of the four types of fit , and the best values of @xmath33 for each fitting function compared .
one fitting function was chosen in preference to another if its @xmath33 was on average better , over most of the passbands .
most cases were clear cut , with reductions in @xmath33 by a factor of two or more compared to the other three fitting functions .
however , there were two main types of marginal case where the difference between minimum @xmath33 for two models was small ( i.e. less than @xmath8 10 per cent ) , and the best model varied between passbands . one case is where there is no evidence for a bulge component in the bluer passbands , but tentative evidence for an exponential bulge component in the red passbands . in this case , a disc only fit is chosen for simplicity . the other marginal case arises when a galaxy with a strong bulge is fit by an exponential or de vaucouleurs bulge equally well : in this case , the choice of bulge profile is fairly arbitrary , and the choice is made to minimise systematic deviations of the fit from the data .
these two types of case are both flagged in table [ tab : bulgedisc ] . in a number of cases ,
the disc parameters are relatively ill - constrained , as there is significant structure in the surface brightness profile in some of the bluer passbands , or in the @xmath34 profile .
this structure is typically due to the low signal - to - noise of the data ( especially in @xmath35 or @xmath34 bands ) , by real structure in the surface brightness profile caused by e.g. recent star formation ( affecting the @xmath35 and @xmath1 band profiles particularly ) or by asymmetries in the most irregular lsbgs in our sample . in these cases ,
we have chosen to fix some of the fit parameters to achieve a better convergence .
we usually choose to fix the disc scale length , as it is possible to use the disc scale lengths in other passbands , along with the mean variation in scale length with passband @xcite , to estimate the scale length in this colour . while clearly not optimal , this method allows us to use the same functional form for fitting the surface brightness profile for all the different passbands .
cases where the fit has been constrained have been flagged in table [ tab : bulgedisc ] .
we show a comparison between our measurements of total galaxy magnitude , disc central surface brightness and disc scale length and the corresponding literature measurements in all available passbands in fig.[fig : comp ] .
note that there is a mean offset in magnitude of @xmath360.28 mag between our magnitudes and those of sprayberry et al .
( stars in fig.[fig : comp ] ) .
this offset does not significantly affect the colours of the galaxies as it is constant over all passbands .
the source of this offset is unknown , however if it is corrected for their magnitudes agree with ours with a rms of @xmath8 0.1 mag .
therefore , in fig.[fig : comp ] and the subsequent comparison , we correct sprayberry et al.s magnitudes for this offset . for the complete sample , 68 per cent of the magnitude differences are smaller than 0.17 mag ( 68 per cent interval ) , and 50 per cent of the magnitude differences are smaller than 0.09 mag . for the central surface brightnesses and scale lengths ,
the 68 per cent intervals are 0.33 magarcsec@xmath2 and 18 per cent , respectively .
the distributions of all of these residuals are strongly non - gaussian , with a relatively narrow ` core ' of accurate comparisons with a more extended envelope of less accurate comparisons , with a few pathological cases .
it is important to understand the origin of the larger magnitude errors .
many of the less accurate magnitude comparisons are from very extended galaxies , which have a large portion of their flux outside the boundaries of the detector . hence their magnitudes are less accurate due to the more uncertain sky levels and larger total magnitude extrapolations
. these less accurate comparisons may be a concern however : do our quoted uncertainties reflect the true uncertainties ? to check this , we compared the ratio of the magnitude difference and the combined measurement errors : the median value of this ratio is @xmath8 0.8 . therefore , we conclude that our formal magnitude error bars are an accurate reflection of the true uncertainties . the central surface brightnesses and scale lengths compare less favourably , with many of the central surface brightnesses in particular disagreeing by more than their combined formal error bars .
however , the formal error bars only give the estimated errors for a given fitting method .
in this case , we have compared our measurements derived using an automated bulge / disc decomposition with measurements derived using a variety of fitting methods , including ` marking the disc ' fits @xcite , disc only fits to the whole surface brightness profile @xcite , bulge / disc decompositions similar to those presented here @xcite , and sophisticated two - dimensional modelling of the luminosity distribution @xcite .
therefore , the formal errors derived using each method are unlikely to reflect the true uncertainties introduced by the use of different fitting methods .
an additional source of scatter is the presence of a bulge component : when included , a bulge component can take some of the light from the disc component , and increase the uncertainty in the disc parameters accordingly .
we conclude that the formal errors for the central surface brightness and scale length are unlikely to represent the true range of uncertainties introduced by using different fitting methods : more representative uncertainties are given by the 68 per cent intervals and are @xmath8 0.3 mag and @xmath8 20 per cent respectively : these uncertainties are comparable to those found by e.g. de jong and de blok et al . .
we have also compared our adopted ellipticities and position angles with those in our sample s source papers .
our ellipticities compare well with those in the literature : most galaxies have ellipticity differences of @xmath37 or smaller .
a few galaxies have ellipticity differences greater than 0.1 ; these differences are typically due to the influence of bars , or due to low signal - to - noise in both sets of images .
comparison of our adopted position angles with those in the literature shows agreement to @xmath8 7@xmath13 .
it is interesting to compare and comment on the morphologies and surface brightness profiles of lsbgs in the optical and the near - ir .
bergvall et al . found a tendency towards similar morphologies in the optical and near - ir for their sample of lsbgs .
we confirm this trend ( despite our typically poorer signal - to - noise in the near - ir ) : this suggests that lsbgs lack the dust content and significant amounts of recent star formation that make morphological classification so passband - dependent for galaxies with higher surface brightness @xcite . our sample , because of its explicit selection to cover as wide a range of lsbg parameters as possible , has a wide range of morphologies , and as is discussed in the next section , sfhs .
the red selected lsbg subsample contains a lenticular galaxy ( c1 - 4 ) , two early type spirals , and two later type spirals .
oneil et al . ,
because of their typically lower spatial resolution , did not find strong evidence for bulges in most of the red lsbgs .
however , with our higher resolution data ( usually the int 2.5-m @xmath38 band images ) , we find that four out of the five red selected lsbgs show evidence for an exponential bulge component .
the lsbg giants all have very strong bulges , with bulge to disc ratios of the order of unity in most passbands .
two lsbg giants are better fit with an exponential bulge component , and the rest are better fit ( in the @xmath33 sense ) by a @xmath28 law bulge profile .
the discs of lsbg giants usually have pronounced spiral structure ( see , e.g. sprayberry et al .
1995 for images ) .
the blue selected lsbgs come with a variety of morphologies , from relatively well - defined spiral morphologies with weak bulges , to galaxies with nearly exponential disc profiles , to galaxies with central ` troughs ' in their luminosity profiles , compared with expectations from an exponential disc fit .
two examples of these ` trough ' galaxies are eso - lv 1040220 and 1040440 .
these galaxies have galactic extinction and inclination corrected central surface brightnesses in @xmath1 of 24.1 and 23.4 magarcsec@xmath2 respectively .
thus , in many respects , they are similar to the blue lsbgs studied by bergvall et al . .
they found that at surface brightnesses lower than @xmath8 23 @xmath1 magarcsec@xmath2 , centrally - depressed surface brightness profiles are quite common . on the basis of this limited and incomplete sample , it is difficult to properly confirm their finding , but our analysis of the surface brightness profiles certainly tentatively supports their observation . in table
[ tab : bulgedisc ] , and to a certain extent fig.[fig : surfphot ] , it is apparent that the disc scale lengths typically decrease and the bulge to disc ratios increase with increasing wavelength .
these are clear signatures of colour gradients in our sample of lsbgs . in the next section ,
we investigate these colour gradients : we will argue that these are primarily due to stellar population gradients .
the main motivation for this programme of optical and near - ir imaging was to study the stellar populations of a diverse sample of lsbgs . in this section ,
we construct accurate colours as a function of radius for our sample and compare these colours with model stellar populations .
we first degrade all of the edited images of a given galaxy to the same angular resolution , and carry out surface photometry in three radial bins : @xmath39 , @xmath40 and @xmath41 , where @xmath42 is the @xmath0 band disc scale length .
galactic extinction corrections were taken from schlegel et al . .
k - corrections for the lsbg giants ( as they are at non - negligible redshifts ) were computed using non - evolving sbc spectra from king & ellis , and are typically 0.2 mag in @xmath1 , negligible in @xmath38 and @xmath43 mag in @xmath0 .
if required , @xmath34 magnitudes were converted @xmath0 magnitudes using the relation in wainscoat & cowie , assuming a typical @xmath44 colour of @xmath8 0.3 mag @xcite .
the error in this correction is unlikely to exceed 0.05 mag .
we show galactic extinction and k - corrected @xmath45 and @xmath46 colours for our sample of lsbgs in fig .
[ fig : colcol ] . measurements in the three radial bins per galaxy are connected by solid lines , measurements of the central bin are denoted by a solid circle and measurements of the colours at two @xmath0 band disc scale lengths are denoted by open circles .
note that we only show colours with uncertainties smaller than 0.3 mag .
average zero point uncertainties and uncertainties in the inner and outer points due to sky level errors are also shown . because @xmath38 band is used in both colour combinations , errors in @xmath38 band affect both colours .
accordingly , the average 1@xmath23 @xmath38 band error is also shown ( the short diagonal line ) .
a @xmath45 against @xmath46 colour - colour plot was chosen to maximise the number of galaxies on the diagram .
other colour combinations ( e.g. @xmath47 against @xmath46 , etc . ) limit the number of galaxies that can be plotted , but yield consistent positions on the stellar population grids . note that p1 - 7 is placed on the plot assuming a @xmath48 colour of @xmath49 ( derived from the degenerate @xmath47 against @xmath48 colour - colour diagram using a typical galactic extinction corrected colour for p1 - 7 of @xmath50 ) .
overplotted on the same diagram are stellar population models and model dust reddening vectors .
we use the gissel98 implementation of the stellar population models of bruzual & charlot ( in preparation ) . for fig.[fig :
colcol ] we adopt a salpeter initial mass fuction ( imf ) and an exponentially decreasing star formation rate characterised by an e - folding timescale @xmath6 and a single , fixed stellar metallicity @xmath51 ( where @xmath52 is the adopted solar metallicity ) . for these models , the galaxy age ( i.e. the time since star formation first started in the galaxy )
is fixed at 12 gyr .
note that the shape of the model grid is almost independent of galaxy age @xmath53 5 gyr .
the solid grey lines represent the colours of stellar populations with a fixed metallicity and a variety of star formation timescales .
the dashed grey lines represent the colours of stellar populations with different , fixed stellar metallicities and a given star formation timescale .
there is some uncertainty in the shape and placement of the model grid : charlot , worthey and bressan discussed the sources of error in stellar population synthesis models , and concluded that the uncertainty in model calibration for older stellar populations is @xmath8 0.1 mag in @xmath45 and @xmath8 0.2 mag in @xmath46 , which is roughly comparable to the calibration error bars in fig.[fig : colcol ] .
the colour uncertainties are larger for some stellar populations than for others : e.g.the optical near - ir colours for older , near - solar metallicity stellar populations are relatively secure , whilst the colours for younger , extreme metallicity stellar populations are much more uncertain .
other sources of systematic error include dust reddening and sfh uncertainties ( as our assumption of a smoothly varying sfr is almost certainly unrealistic ) . for these reasons
the positions of galaxies and colour trends between galaxies should be viewed relatively , e.g. that one class of galaxies is more metal - rich than another class of galaxies .
we also include vectors describing the effect of dust reddening using either a screen model ( arrow ) or the more realistic geometry of a triplex dust model ( * ? ? ?
* curved line ) .
for the triplex model a closed circle denotes the central reddening , and an open circle the reddening at 2.5 disc scale lengths . for both reddening vectors
we use the milky way extinction law ( and for the triplex model , the albedo ) from gordon , calzetti & witt .
our triplex model assumes equal scale length vertically and radially exponential dust and stellar distributions , and a central optical depth ( for viewing a background object ) in @xmath54 band of 2 .
these parameters are designed to be a reasonable upper limit to the effects of reddening in most reasonably bright spiral galaxies : kuchinski et al .
find from their study of 15 highly - inclined spiral galaxies that they are well - described by the above type of model , with central @xmath54 band optical depths of between 0.5 and 2 . for the purposes of calculating the triplex reddening vector
, we use the optical depth due to absorption only for two reasons .
firstly , one might navely expect that for face - on galaxies at least as many photons will get scattered into the line of sight as out of it .
secondly , de jong finds that an absorption - only triplex model is a reasonably accurate description of the results of his realistic monte - carlo simulation ( including the effects of both scattering and absorption ) of a face - on triplex geometry spiral galaxy . in the following interpretation of fig.[fig : colcol ] , we for the most part _ explicitly neglect the possible effects of dust reddening_. there are a number of arguments that suggest that the effects of dust reddening , while present , are less important than stellar population differences both within and between galaxies .
we consider these effects more carefully in section [ subsec : dust ] .
the majority of the lsbgs in our sample with colours in at least two radial bins have significant optical near - ir colour gradients . for the most part _
these colour gradients are consistent with the presence of a mean stellar age gradient _ , where the outer regions of galaxies are typically younger , on average , than the inner regions of galaxies .
this is consistent with the findings of de jong who concludes that age gradients are common in spiral galaxies of all types .
note that this is inconsistent with the conclusions of bergvall et al .
, who find little evidence for optical near - ir colour gradients in their sample of blue lsbgs .
however , their sample was explicitly selected to lack significant @xmath45 colour gradients , and so that we disagree with their conclusion is not surprising .
the fact that bergvall et al . managed to find galaxies without significant colour gradients is interesting in itself ;
while colour gradients are common amongst disc galaxies , _ they are by no means universally present_. this is an important point , as by studying the systematic differences in physical properties between galaxies with and without colour gradients , it may be possible to identify the physical mechanism by which an age gradient is generated in disc galaxies .
there are galaxies which have colour gradients which appear inconsistent with the presence of an age gradient alone .
the two early - type lsbgs c1 - 4 and c3 - 2 have small colour gradients which are more consistent with small metallicity gradients than with age gradients .
in addition , a few of the bluer late - type lsbgs have colour gradients that are rather steeper than expected on the basis of age gradients alone , most notably f583 - 1 and p1 - 7 , both of which appear to have colour gradients more consistent with a metallicity gradient .
eso 1040440 has an ` inverse ' colour gradient , in that the central regions are bluer than the outer regions : this is probably due to a combination of very irregular morphology and the effects of foreground stellar contamination ( which in this case makes the galaxy colours difficult to accurately estimate ) .
there are also conspicuous ` kinks ' in the colour profiles of f574 - 1 and 2327 - 0244 , and a possible kink in the colour profile of ugc 128 . in the case of f574 - 1 ,
it is likely that the central colours of the galaxy are heavily affected by dust reddening .
f574 - 1 is quite highly inclined ( @xmath55 , assuming an intrinsic disc axial ratio @xmath56 of 0.15 ; holmberg 1958 ) and shows morphological indications of substantial amounts of dust extinction in the int 2.5-m @xmath1 and @xmath54 band images
. this may also be the case for ugc 128 , although the inclination is smaller in this case and there are no clear - cut morphological indications of substantial dust reddening .
the lsbg giant 2327 - 0244 has a red nucleus in the near - ir , but a relatively blue nucleus in the optical .
however , 2327 - 0244 is is a seyfert 1 and has a central starburst @xcite , making our interpretation of the central colours in terms of exponentially decreasing sfr models invalid . in considering colour trends between lsbgs
, we will first consider the star formation histories of different classes of lsbg , and then look at the star formation histories in a more global context .
[ [ blue - selected - lsbgs ] ] blue selected lsbgs + + + + + + + + + + + + + + + + + + + the majority of blue selected lsbgs , e.g. f561 - 1 , ugc 334 and eso - lv 2490360 , are also blue in the near - ir .
their colours indicate that most blue selected lsbgs are younger than hsb late - types , with rather similar metallicities ( compare the positions of the main body of blue selected lsbgs to the average sd
sm galaxy from de jong s sample ) .
this relative youth , compared to the typically higher surface brightness sample of de jong suggests that the age of a galaxy may be more closely related to its surface brightness than the metallicity is .
we will come back to this point later in section [ subsec : correl ] .
four out of the fifteen blue selected lsbgs fail to fit this trend : eso - lv 1040220 , eso - lv 1040440 , eso - lv 1450250 and eso - lv 1870510 all fall substantially ( @xmath57 mag ) bluewards of the main body of blue lsbgs in @xmath46 colour , indicating a lower average metallicity .
while it should be noted that the south pole subsample all have larger zero point uncertainties than the northern hemisphere sample , we feel that it is unlikely that the zero point could be underestimated so substantially in such a large number of cases . while the models are tremendously uncertain at such young ages and low metallicities , these optical near - ir colours suggest metallicities @xmath58 solar .
this raises an interesting point : most low metallicity galaxies in the literature are relatively high density blue compact dwarf galaxies @xcite .
thus , these galaxies and e.g. the blue lsbgs from bergvall et al . offer a rare opportunity to study galaxy evolution at both low metallicities _ and _ densities , perhaps giving us quite a different view of how star formation and galaxy evolution work at low metallicity .
[ [ red - selected - lsbgs ] ] red selected lsbgs + + + + + + + + + + + + + + + + + + in bell et al .
, we found that two red selected lsbgs ( c1 - 4 and c3 - 2 ) were old and metal rich , indicating that they are more evolved than blue selected lsbgs . in this larger sample of five red selected lsbgs , we find that this is not always the case .
the galaxies n10 - 2 , c3 - 2 , and c1 - 4 ( with @xmath1 band central surface brightnesses @xmath8 22.5 magarcsec@xmath2 ) _ are _ comparitively old and metal rich , with central colours similar to old , near - solar metallicity stellar populations .
however , i1 - 2 and p1 - 7 ( with lower @xmath1 band central surface brightnesses @xmath8 23.1 magarcsec@xmath2 ) both have reasonably blue galactic extinction corrected optical near - ir colours .
i1 - 2 has a stellar population similar to de blok et al.s f579-v1 , which is part of our blue selected subsample : these galaxies lie in the overlap between the two samples .
p1 - 7 is relatively similar to other , brighter blue lsbgs such as ugc 128 and f574 - 1 : it was included in the red selected subsample only by virtue of its relatively high foreground extinction @xmath59 mag .
our limited data suggests that the red - selected lsbgs catalogued by oneil et al .
are a very heterogeneous group . unlike the blue or giant lsbgs ,
the red - selected lsbgs seem to have relatively few common traits .
the five red - selected lsbgs in this study seem to be a mix of two types of galaxy : i ) early - type spirals or lenticulars which are genuinely red but have surface brightnesses @xmath60 22.5 magarcsec@xmath2 at the upper range of the lsb class , and ii ) objects with low surface brightnesses but colours that are not genuinely red .
objects in class ii ) appear in the red - selected subsample due to large galactic foreground reddening and photometric errors . recently , oneil et al . reported the detection of a small number of red , gas - rich lsbgs .
these galaxies would clearly contradict our above interpretation of the red lsbgs , suggesting that at least some of the red lsbgs are a distinct ( though potentially quite rare ) class of galaxy .
it is interesting to note that gerritsen & de blok predict that around 20 per cent of lsbgs should have @xmath61 and relatively low surface brightnesses @xmath62 magarcsec@xmath2 : these galaxies , which represent the fraction of the lsbg population that lack recent star formation , would appear to be both red and gas rich ( blue lsbgs would be galaxies with exactly the same past sfh , but more recent star formation ) .
however , note that oneil et al.s result is subject to significant observational uncertainties : for example , the @xmath47 colour of p1 - 7 adopted by oneil et al .
is 0.9 , whereas the foreground galactic extinction - corrected colour of p1 - 7 in this study is found to be [email protected] .
p1 - 7 is on the verge of being classified as a red , gas rich lsbg with a @xmath47 colour of 0.9 , however it lies well within the envelope of blue , gas rich lsbgs with a galactic extinction corrected @xmath47 colour of 0.57 .
proper observational characterisation of these red , gas rich lsbg candidates may prove quite crucial in testing lsbg formation and evolution models .
[ [ giant - lsbgs ] ] giant lsbgs + + + + + + + + + + + lsbg giants have galactic extinction and k - corrected optical near - ir colours which are similar to the redder lsbgs from oneil et al . , indicating central stellar populations that are reasonably old with roughly solar metallicity . however these central colours are not consistent with an old , single burst stellar population , although this conclusion is somewhat dependent on the choice of k - correction and model uncertainties ( k - correction uncertainties are typically @xmath64 0.05 mag in each axis at 20000 kms@xmath18 for a change in galaxy type from sbc to sab , or sbc to scd ) .
the outer regions of lsbg giants are much younger , on average , whilst still retaining near - solar metallicities . comparing the @xmath47 and @xmath46 colours for the lsbg giants with some of the bluer hsb sa sc galaxies in de jong s sample shows that both sets of galaxies have similar stellar populations .
thus , lsbg giants do not have unique sfhs ( implying that there need be no difference in e.g.star formation mechanisms between the hsb sa sc galaxies and the lsbg giants ) , however it is interesting that lsbg giants can have both a substantial young stellar population and a high metallicity , whilst possessing such low stellar surface densities ( assuming reasonable stellar mass to light ratios ) . [ [ summary ] ] summary + + + + + + + overall , there is a clear age and metallicity sequence , with red lsbgs and lsbg giants in the high stellar metallicity and older average age corner of the plot , progressing to the lower stellar metallicity and younger average age region of the plot for the blue lsbgs . _
this suggests that lsbgs , just like hsbgs , come in a variety of morphological types and sfhs ; moreover , their morphologies and sfhs are linked_. we have constructed fig.[fig : colcol ] using a salpeter imf .
however , the imf is still reasonably uncertain , especially at the very low mass and high mass ends ( see e.g. elmegreen 1999 , scalo 1998 ) .
variation of the low mass end of the imf only significantly changes the ( relatively ill - constrained ) stellar mass to light ratio .
variation of the high mass end of the imf , as illustrated by using e.g. the scalo or miller & scalo imf , only significantly changes the high metallicity , young stellar populations corner of the colour - colour plane : the colours in these regions are subject to considerable modelling uncertainties at any rate . we conclude that our results are robust to reasonable imf uncertainties .
we have interpreted fig.[fig : colcol ] in terms of stellar population differences , however we have so far neglected the effects of dust reddening on the colours of our stellar populations . in fig.[fig : colcol ] we show both a screen @xmath65 and absorption - only triplex model @xmath66 dust reddening vector , both assuming milky way extinction and albedo curves .
smc extinction and albedo curves make little difference to the direction of the reddening vectors , but increases the length of the triplex model reddening vector slightly because of the lower albedo at all wavelengths , compared to the milky way dust properties . the screen model vector is shown for illustrative purposes only , as the dust will be distributed roughly similarly to the starlight in realistic spiral galaxies , as assumed in the triplex model .
the colour gradients in our sample of lsbgs are unlikely to be due to the effects of dust reddening for four reasons . 1 .
the optical depth that we choose for the triplex model is on the high end of the plausible dust optical depths derived by kuchinski et al . .
the triplex model assumes that all of the dust is distributed smoothly throughout the galaxy . in real galaxies
typically 1/3 of the dust is gathered into optically thick clumps , which , when viewed face - on , simply tend to ` drill holes ' in the light distribution but not produce any significant colour changes @xcite .
this dust would be more easily found in more edge - on galaxies , as in the edge - on orientation the probability of a clump along the line of sight is much larger than in a face - on disc .
therefore , kuchinski et al.s dust optical depths are likely to account for both the clumped and unclumped dust .
most lsbgs have lower metallicity than their higher surface brightness counterparts .
this implies that the dust to gas ratio should be lower in lsbgs than in hsbgs .
4 . only one galaxy in our sample , the starbursting seyfert 1 2327 - 0244 , has been detected by iras .
this is in stark contrast to the diameter limited sample of de jong & van der kruit , for which 78 per cent of the sample was detected by iras .
this suggests that dust is much less important in lsbgs than in their higher surface brightness counterparts : indeed , many of the non - detections in de jong & van der kruit s sample were lsbgs .
the above arguments suggest that the triplex dust reddening vector in fig.[fig : colcol ] is an upper limit to the real effects of dust reddening .
comparison of this upper limit on the dust reddening effects with the colour gradients in our sample suggests that most galaxies must have stellar population gradients to produce such pronounced colour gradients .
this conclusion is supported by the direct modelling of kuchinski et al . , who use their own best - fit galaxy models to estimate the colour gradients in face - on spiral galaxies :
they conclude that colour gradients as large as the ones we observe in lsbgs , or observed by de jong , are too large to arise solely from reddening by realistic amounts of dust .
the same argument applies to relating different populations of lsbg : the amount of reddening required to make an intrinsically blue lsbg appear red is too large to be compatible with the observational evidence cited above .
in section [ subsec : sfhres ] , we saw a pronounced age / metallicity sequence .
lsbgs can have a diverse range of stellar populations : red selected lsbgs can be quite old and metal rich , whilst most blue selected lsbgs are much younger , and more metal poor .
however , from fig.[fig : colcol ] alone , it is difficult to see _ why _ the star formation histories of low surface brightness galaxies are as diverse as they are .
for this reason , we have chosen to quantify the positions of galaxies on the colour - colour plane by assigning average ages and metallicities from the gissel98 stellar population models of bruzual & charlot ( in preparation ) .
a finely interpolated grid of stellar population models was generated , covering a range of star formation timescales and metallicities .
the metallicity ranges from @xmath67 ( 1/100 solar ) to @xmath68 ( 2.5 times solar ) . in fig.[fig : colcol ] , we can see that some galaxies have optical near - ir colours too blue to be described adequately by a constant star formation rate : for this reason , we must include exponentially _ increasing _ sfhs in our model grid .
we include star formation timescales @xmath6 ranging from 0 ( an average age of 12 gyr ) to @xmath69 ( constant star formation with an average age of 6 gyr ) to @xmath70 gyr ( an average age of 1 gyr ) .
we can see from the range of models included above that @xmath6 is an inappropriate parameter for describing the sfhs : we parameterise the sfh using the average age of the stellar population . for each galaxy , the galactic extinction and k - corrected @xmath45 and @xmath46 colours between @xmath40 ( when available ; 0052 - 0119 and 0237 - 0159 have only central colours and are omitted from further consideration ) are compared with the colours of the finely interpolated stellar populations grid .
the quantity @xmath71 is minimised , where @xmath72 is the model colour , @xmath73 is the observed colour and @xmath74 is the error in the observed colour .
error bars are obtained for the average ages and metallicities by fitting the colours plus or minus their 1@xmath23 sky level and zero point errors added in quadrature .
this procedure has a number of limitations .
firstly , we neglect the effects of dust reddening .
we considered the possible effects of dust earlier in section [ subsec : dust ] , and concluded that the effects of dust are fairly minimal for this sample .
our use of the colours between @xmath40 should act to further lessen the importance of dust reddening on our results . secondly , because we use simple sfhs assuming a single metallicity , and because of stellar population model uncertainties , the ages and metallicities we determine are unlikely to be accurate in absolute terms .
however , the idea of this simple analysis is to identify relative trends in sfh and metallicity , and their possible causes .
the ages and metallicities we derive here will serve this purpose well : the optical near - ir colours provide a robust way to order galaxies in terms of the relative importance of @xmath53 5 gyr old stars compared to younger @xmath64 2 gyr old stars , which we parameterise using the average age of the stellar population . to better understand the relative uncertainties , we have also carried out the analysis using the stellar population models of kodama & arimoto : the results obtained with their models are indistinguishable ( in a relative sense ) to the results presented in figs .
[ fig : correl ] and [ fig : correl2 ] . an extended version of this analysis using all the available colour combinations for a much larger sample of galaxies will be presented in bell & de jong . in order to understand what physical mechanisms might be driving the sfhs of lsbgs
, we plot the best - fit average ages and metallicities against physical parameters , such as @xmath0 central surface brightnesses , @xmath0 absolute magnitudes , @xmath0 disc scale lengths and gas fractions .
these correlations are presented in fig.[fig : correl ] .
note that we do not attempt to provide a best fit or attach any kind of statistical significance to these relationships .
we feel that this is over - interpreting this relatively limited dataset at this stage .
galactic extinction and k - corrected @xmath0 absolute magnitudes are calculated using distances from table [ tab : samplepar ] .
these absolute magnitudes are combined with the masses ( multiplied by 1.33 to account for the helium mass fraction ) from table [ tab : samplepar ] to derive the gas fractions , using a @xmath0 band @xmath75 of 0.6 @xmath76 ( c.f .
verheijen 1998 ; chapter 6 ) and assuming a solar @xmath0 band absolute magnitude of 3.41 @xcite .
note that these gas fractions do not account for any molecular or hot gas component .
the same distances are used to derive the @xmath0 band disc scale lengths in kpc .
the @xmath0 band surface brightnesses have been corrected for galactic extinction , @xmath0 corrections , @xmath77 surface brightness dimming and inclination .
inclination corrections @xmath78 have been determined using an intrinsic edge - on disc axial ratio @xmath79 assuming that the disc is transparent ; these corrections are given by : @xmath80 where @xmath81 is the adopted disc ellipticity from table [ tab : epa ] @xcite .
before discussing fig.[fig : correl ] , it is useful to note that the average ages of eso - lv 1040220 , 1040440 , 1450250 and 1870510 are all potentially quite uncertain .
this is due to their low metallicities : they have optical near - ir colours suggesting metallicities @xmath64 1/100 solar metallicity ( which falls off the model grid ) . at these metallicities , the model predictions ( especially for young ages ) are all rather uncertain .
therefore , we feel that their ages are less constrained ( even in a relative sense ) than their error bars otherwise suggest .
these points are denoted by a naked set of error bars in figs.[fig : correl ] and [ fig : correl2 ] . in the discussion that follows
, we typically ignore the average ages for these four data points . in fig.[fig : correl ] , panel a , we show the relationship between @xmath0 band central surface brightness and @xmath0 band absolute magnitude in our dataset . in panels
b and c , we show the relationship between gas fraction and @xmath0 band surface brightness and magnitude respectively .
the main purpose of these panels is to simply show that in our sample of lsbgs , surface brightness , magnitude and gas fraction are quite strongly inter - related , and it is expected that correlations of any given quantity with surface brightness will be reflected by correlations of that same quantity with magnitude and gas fraction .
thus , on the basis of this dataset alone , it will be difficult to unambiguously differentiate between trends in the stellar populations of lsbgs driven by magnitude , trends driven by surface brightness , or trends driven by the gas fraction . note that because of selection effects ( primarily that we select only galaxies with low @xmath1 band surface brightness ) , panels a , b and c are unlikely to be representative of any universal correlation between surface brightness , magnitude and gas fraction .
fig.[fig : correl ] , panel d shows the correlation between the best - fit average age and metallicity . ignoring the uncertain ages of the ultra - low metallicity lsbgs
, there is a correlation , albeit a noisy one , between age and metallicity , in the sense that older galaxies tend to be more metal rich .
this correlation quantifies the global trend between age and metallicity observed earlier in fig.[fig : colcol ] .
this correlation is expected : galaxies appear old through a lack of recent star formation , implying a lack of gas ( see also panel h , the age
gas fraction correlation ) .
most of our sample lsbgs are reasonably massive , meaning that their gas is unlikely to have been ejected by e.g. supernova driven winds .
therefore , their low gas fraction is likely to be due to consumption by star formation : this results in reasonably high stellar metallicities ( see e.g. pagel 1998 ) .
note that some low surface brightness dwarf spheroidals may run counter to this argument , having low metallicities and reasonably old average ages ( e.g. mateo 1998 ) : to explain this one would have to invoke removal of the gas from these low mass galaxies by e.g.supernova-driven winds @xcite or , for cluster dwarf spheroidals , ram - pressure stripping @xcite . in panels
e , f , g and h , we show the trends in age with @xmath0 central surface brightness , @xmath0 absolute magnitude , @xmath0 disc scale length and gas fraction respectively .
age does not appear to correlate with disc scale length .
neglecting the uncertain ages of the four low metallicity eso - lv galaxies , there appear to be trends in age with surface brightness , magnitude and gas fraction , in the sense that lower @xmath0 band surface brightness , lower @xmath0 luminosity and higher gas fraction lsbgs all have younger average ages .
furthermore , the scatter between age and surface brightness appears smaller than the scatter in the age
magnitude and age gas fraction correlations .
this suggests , tentatively , that the age
surface brightness correlation is the main correlation in this dataset , and the other correlations are the result of the magnitude
surface brightness and gas fraction
surface brightness correlations .
we return to this issue later in section [ subsubsec : mod ] . in panels
i , j , k and l of fig.[fig : correl ] , we show the trends in lsbg metallicity with @xmath0 band surface brightness , magnitude , disc scale length and gas fraction .
note that the metallicities of the four low metallicity eso - lv galaxies are reasonably well - constrained , therefore it is fair to include them in the discussion of these trends .
metallicity , unlike age , seems to be correlated with @xmath0 band disc scale length .
this seems primarily because of the lsbg giants : they are bright and have large scale lengths , so their high metallicities will produce correlations between metallicity and magnitude and metallicity and scale length .
it is likely that a metallicity magnitude relation is more fundamental , however it is impossible to rule out scale length dependence in the metallicity on the basis of this dataset .
metallicity , like age , seems to correlate well with @xmath0 band surface brightness , magnitude and gas fraction in the sense that lsbgs with lower metallicities have lower surface brightnesses , lower luminosities and higher gas fractions than lsbgs with higher metallicities .
both age and metallicity are also linked ( panel d ) , so the similarity in trends with physical parameters between age and metallicity is not surprising . again , as was the case for the age , it is more or less impossible to tell , on the basis of this dataset alone , which of the magnitude metallicity or surface brightness
metallicity correlations are more fundamental . in order to investigate how the _ baryonic _ mass and surface density affect the sfhs of lsbgs , we have modified the @xmath0 surface brightnesses and absolute magnitudes by turning the neutral gas fraction of the galaxy into stars using the correction to the magnitude @xmath82 .
this correction assumes that the gas will turn into stars with the same @xmath0 band mass to light ratio as was assumed for the stellar population , which in this case is 0.6 @xmath76 .
this correction makes a number of assumptions . 1 .
the @xmath0 band mass - to - light ratio is expected to be relatively robust to the presence of young stellar populations , however our assumption of a constant @xmath0 band mass to light ratio is still a crude assumption .
note however that the relative trends in fig.[fig : correl2 ] are quite robust to changes in stellar mass - to - light ratio : as the stellar mass - to - light ratio increases , the modified magnitudes creep closer to their unmodified values asymptotically .
the presence of molecular gas is not accounted for in this correction .
assuming a typical galactic co to h@xmath83 conversion ratio , the non - detections of lsbgs by schombert et al . and de blok & van der hulst imply h@xmath83 to ratios smaller than 0.25 , and exceptionally as low as 0.04 .
this result is complicated by the expected metallicity dependence in the co to h@xmath83 conversion ratio @xcite , however , molecular hydrogen is unlikely to dominate for all plausible values of the conversion factor .
the correction to the surface brightness implicitly assumes that the gas will turn into stars with the same spatial distribution as the present - day stellar component .
this is a poor assumption as the gas distribution is usually much more extended than the stellar distribution ; most lsbg gas distributions are centrally peaked however , and so the correction is unlikely to be completely wrong .
all of the above arguments suggest that the modified magnitudes and surface brightnesses are unlikely to give an accurate measure of the true masses and densities of lsbgs .
these quantities do however better reflect the total mass and density of the galaxy than the @xmath0 band magnitudes and surface brightnesses alone , and so are useful as an indication of the kind of trends we might see if we could work out the baryonic masses and densities of lsbgs .
the median correction is -0.9 mag , and 68 per cent of the corrections lie within 0.7 mag of this value .
the trends in age and metallicity with modified @xmath0 band central surface brightness and absolute magnitude are shown in fig.[fig : correl2 ] .
three red and giant lsbgs with known redshifts but unknown ( or unmeasurable ) fluxes have been placed on fig.[fig : correl2 ] also : error bars denote their ( unmodified ) magnitudes or surface brightnesses , and solid triangles denote the direction that these galaxies move in with increasing gas fraction , with the vertex of the triangle indicating the offset that is produced by converting a gas fraction of 0.2 into stars . for reference , in panels c and f , we also show the relationship between @xmath1 band magnitude and surface brightness and the modified @xmath0 band magnitude and surface brightness .
surprisingly , we find that @xmath1 band magnitude and surface brightness are reasonable reflections of the modified magnitudes and surface brightnesses of lsbgs ( remembering the uncertain assumptions that went into constructing the modified magnitudes and surface brightnesses ) .
both the @xmath1 band magnitudes and surface brightnesses have a mean offset from the modified @xmath0 band values consistent with a + 3.9 mag offset , with an rms of around 0.5 mag ( solid and dotted lines in panels c and f ) .
this relationship may be caused by the sensitivity of the optical mass to light ratios to recent star formation : @xmath1 band mass to light ratios decrease for younger populations in a such a way as to make the increase in brightness for younger populations offset almost exactly ( to within a factor of 60 per cent or so , from the 0.5 mag scatter ) by their larger gas fractions , making the @xmath1 band a reasonable total mass indicator in lsbgs ( assuming that the @xmath0 band stellar mass to light ratio is fairly constant ) . in panels a and b of fig.[fig : correl2 ]
, we see that the lsbg ages and metallicities correlate with their modified @xmath0 band magnitudes ( baryonic masses ) . in stark contrast , in panels d and e , we see that the ages and metallicities correlate poorly with the modified @xmath0 band surface brightnesses ( baryonic surface densities ) .
this result is puzzling , especially in the light of the tight correlation between average age and surface brightness in panel e of fig.[fig : correl ] .
there are two possible interpretations of this huge scatter in the age
modified surface brightness plane .
is it possible that there is little or no correlation between the sfh and surface density of a galaxy , as panels d and e suggest ? in this case , we must explain the strong correlations between age and @xmath0 band surface brightness and metallicity and @xmath0 band surface brightness in panels e and i of fig.[fig : correl ] .
this may be possible to explain via selection effects : because we selected our galaxies to have a relatively narrow range of @xmath1 band surface brightness , and older , more metal rich galaxies are redder , we would naturally expect to see an artificial correlation between @xmath0 band surface brightness ( which would be higher because of the redder colours ) and age ( which correlates with these red colours ) .
note , however , that this interpretation has real difficulty in explaining the tightness and dynamic range of the age@xmath0 band surface brightness correlation .
conversely , is it possible that there is a correlation between baryonic surface density and age , and that for some reason it is masked in this dataset ?
there is quite a narrow range ( only @xmath8 2 magarcsec@xmath2 ) of modified surface brightness : the close correspondence between @xmath1 band and modified @xmath0 band surface brightnesses ( panel f of fig.[fig : correl2 ] ) means that because we selected our galaxies to have a relatively narrow range of @xmath1 band surface brightness , we have implicitly selected for a narrow range in baryonic central surface density . also , our modifications to the surface brightnesses were large for gas - rich galaxies with low @xmath0 band surface brightnesses and much smaller for gas - poor galaxies with higher @xmath0 band surface brightnesses , steepening the relationship between sfh and modified surface brightness considerably .
thus , we may simply lack an adequate surface density range to be able to distinguish a steep trend in age with surface density , especially in the presence of significant uncertainties in the ages and surface densities . note that this is the interpretation that we prefer : the correlation between age and @xmath0 band surface brightness is quite tight , which is quite suggestive of some kind of correlation between density and sfh .
note that the age
mass relation would survive in this case relatively unscathed : the dynamic range for the age
mass relation is in excess of 5 magnitudes , which makes it quite robust to changes @xmath8 1 mag in the relative positioning of the data points along the trend .
all the above correlations are consistent with the proposition that _ the age of a lsbg stellar population is primarily correlated with its surface density and that its metallicity is correlated with both surface density and mass _ , albeit with considerable scatter .
panels e , f , i and j of fig.[fig : correl ] all strongly support this scheme : in particular , lsbg giants have a range of @xmath0 band surface brightnesses ( and ages ; panel f ) but have bright @xmath0 band absolute magnitudes ( and high metallicities ; panel j ) .
the main obstacle for such a scenario is the lack of correlation between modified @xmath0 band surface brightnesses and age and metallicity , which _ may _ be due to our explicit selection of galaxies with only a relatively narrow range of @xmath1 band central surface brightness .
an important question to ask is if this scenario is physically plausible .
a relationship between surface brightness / density and average age would be easy to understand : either a density dependent ( e.g. schmidt 1959 ) star formation law or a star formation timescale which is proportional to the local dynamical timescale would result in such a correlation ( note that a magnitude age correlation might suggest a global dynamical timescale dependence in the star formation law ; e.g. kennicutt 1998 ) .
such a scenario would also very naturally account for the existence of age gradients in lsbgs : the outer regions of lsbgs are less dense , and would form stars less quickly than their inner regions , resulting in an age gradient .
such a scheme is qualitatively consistent with those proposed by e.g. de blok et al . , jimenez et al . or gerritsen & de blok .
in these schemes , it is primarily the low surface density that slows down the evolution of lsbgs ( either explicitly through a density - dependent star formation law or implicitly , through an inability to build up a high gas metallicity ) .
metallicity s mass dependence is likely to stem from the relative importance of feedback : the efficiency with which a galaxy ejects its metals in a supernova - driven wind is primarily driven by its mass @xcite .
while the above scenario is consistent with the data , it is unlikely to be unique : because of the correlation between surface brightness and magnitude , a purely mass - dependent age and metallicity may be plausible .
note , however that a mass - dependent age has some disadvantages : the ages of lsbg giants support a surface brightness dependent age , and age gradients would have to be explained using a different mechanism .
we have performed deep imaging of a sample of 26 lsbgs in the optical and near - ir in order to study their stellar populations . by comparing their optical near - ir colours with the latest stellar population models , it is possible to constrain the young to old star ratio ( parameterised by the average age of the galaxy ) and the galaxy metallicity .
we have found the following . *
optical near - ir colour gradients are common in lsbgs .
most colour gradients are consistent with a mean stellar age gradient , with the outer regions of galaxies appearing younger than the inner region of galaxies .
we argue against the effects of dust reddening as the only cause of lsbg colour gradients . as common as colour gradients are , they are not present in all lsbgs @xcite ; this is important , because it may provide a chance to directly observe what drives the age gradient in lsbgs . *
we find that lsbgs have a wide range in morphologies and stellar populations , ranging from old , near solar metallicity populations for the very reddest lsbgs , to younger , high metallicity populations in the lsbg giants , to young and metal poor populations in the blue gas - rich lsbgs .
* by comparing the observed optical near - ir colours between 0.5 and 1.5 disc scale lengths with stellar population models , we have determined best - fit average ages and metallicities that are robust in a relative sense , so that trends in age and metallicity will be quite secure .
* when the highly uncertain ages of the lowest metallicity galaxies are excluded , there are strong trends between both age and metallicity and @xmath0 band surface brightness , absolute magnitude and gas fraction .
lsbgs with low @xmath0 band surface brightnesses , low @xmath0 luminosities and high gas fractions are all fairly unevolved , young , low metallicity systems .
in contrast , lsbgs with higher @xmath0 band surface brightnesses , higher @xmath0 band luminosities and smaller gas fractions appear much more evolved , with older average ages and higher ( near solar ) metallicities .
* we have constructed crude estimates of the total mass and total densities of the atomic hydrogen , helium and stars in lsbgs by turning the gas into stars with a constant @xmath0 band mass to light ratio of 0.6 @xmath76 .
surprisingly , we find that the @xmath1 band absolute magnitudes and surface brightnesses are reasonable predictors of the mass and density estimators , to within a factor of @xmath8 60 per cent or so .
we find that the ages and metallicities of lsbgs correlate well with our mass estimator , but correlate poorly with our surface density estimator .
we argue that the poor correlation between our surface density estimator and sfh is the result of the narrow dynamic range in surface density probed by our sample , compounded by errors in sfh and density determination . *
our results are consistent with a scenario in which the age of a lsbg is correlated primarily with its surface density , and the metallicity of a lsbg is correlated with both its surface density and mass ( albeit with much scatter ) .
this kind of correlation would be observed if the star formation law depended either explicitly on gas surface density or on the local dynamical timescale , and if the efficiency with which a galaxy retained its newly - synthesised metal content was a function of its mass .
we would like to thank erwin de blok , karen oneil , stacy mcgaugh and david sprayberry for providing surface photometry and images of galaxies in their sample , and for helpful discussions . in particular , we would like to thank karen oneil for providing information about her lsbg sample before their publication , and enzo branchini for determining the peculiar motions of the galaxy sample .
we would also like to thank the referee for useful comments on the manuscript .
efb would like to thank the isle of man education department for their generous support .
support for rsdj was provided by nasa through hubble fellowship grant # hf-01106.01 - 98a from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
the northern hemisphere near - ir observations were obtained using the apache point observatory 3.5-m telescope , which is owned and operated by the astrophysical research consortium .
the united states national science foundation supported the near - ir observations made at the south pole through a cooperative agreement with the center for astrophysical research in antarctica , grant no .
nsf opp-8920223 .
some of the observations described in this paper were made during service time at the isaac newton telescope and at the united kingdom infrared telescope .
this project made use of starlink computing facilities in durham .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
moshir m. , kopan g. , conrow t. , mccallon h. , hacking p. , gregorich d. , rohrbach g. , melnyk m. , rice w. , fullmer l. , et al .
, 1990 , ` infrared astronomical satellite catalogs the faint source catalog ' , version 2.0 . | we have performed deep imaging of a diverse sample of 26 low surface brightness galaxies ( lsbgs ) in the optical and the near - infrared . using stellar population synthesis models , we find that it is possible to place constraints on the ratio of young to old stars ( which we parameterise in terms of the average age of the galaxy ) , as well as the metallicity of the galaxy , using optical and near - infrared colours .
lsbgs have a wide range of morphologies and stellar populations , ranging from older , high metallicity earlier types to much younger and lower metallicity late type galaxies . despite this wide range of star formation histories
, we find that colour gradients are common in lsbgs .
these are most naturally interpreted as gradients in mean stellar age , with the outer regions of lsbgs having younger ages than their inner regions . in an attempt to understand what drives the differences in lsbg stellar populations , we compare lsbg average ages and metallicities with their physical parameters .
strong correlations are seen between a lsbg s star formation history and its @xmath0 band surface brightness , @xmath0 band absolute magnitude and gas fraction .
these correlations are consistent with a scenario in which the star formation history of a lsbg primarily correlates with its surface density and its metallicity correlates both with its mass and surface density .
galaxies : spiral galaxies : stellar content galaxies : evolution galaxies : general galaxies : fundamental parameters galaxies : photometry |
single and dilepton probes in heavy ion collisions are of particular interest since such probes , once produced , are largely unaffected by the surrounding qcd medium .
they carry valuable information on the particle from which they originate and allow one to assess the properties of the medium formed in the early instants of the collision .
the following contributions to the dilepton invariant mass spectrum are discussed here , together with what one might learn from their measurement about the properties of the medium formed in the collision : * low mass dileptons originating from vector meson leptonic decay ( @xmath2 , @xmath3 and @xmath4 ) provide insight on the properties of these mesons in the high temperature expanding fireball produced immediately after the collision , where chiral symmetry may be ( at least partially ) restored @xcite ; * a significant fraction of the virtual and direct photons produced at low @xmath5 ( @xmath6 gev / c ) in heavy ion collisions originates from the thermal black - body radiation of the created fireball @xcite .
measuring these photons therefore allows one to quantify the temperature of the fireball ; * open heavy flavors , because of their high mass , allow one to study in - medium energy loss mechanisms in addition to what can be learned from light quarks @xcite ; * heavy quarkonia are of interest because of additional mechanisms that are predicted to occur in the presence of a qgp and that would affect the production of these bound states @xcite .
fig . [ low_mass_vector_mesons ] ( left ) shows the correlated dimuons invariant mass distribution at the @xmath2 vacuum mass , measured by the na60 experiment in semi - central in+in collisions @xcite .
the @xmath2 mass peak differs significantly from the expected vacuum @xmath2 and can be reasonably well described on the low mass side by the model presented in @xcite .
this model includes a detailed description of the baryonic matter created in the collision below the formation temperature of a qgp , @xmath7 .
interactions with this baryonic matter are responsible for a broadening of the @xmath2 ( but no modification of its mass ) when approaching chiral symmetry restoration near @xmath8 . [ cols="^,^ " , ]
in short : * low mass vector mesons exhibit strong shape modifications with respect to their vacuum properties , that can be well described at sps but not at rhic possibly because some contributions to the dilepton spectrum have not been properly accounted for ; * virtual photons can be used in addition to direct photon measurements to assess the medium temperature averaged over its expansion time and derive its initial temperature ; * a significant suppression of @xmath9 quarks is necessary to describe the observed heavy flavor @xmath10 in a way that is consistent with the b / b+d ratio measured in @xmath11 collisions ; * @xmath12 production in heavy - ion collisions is a puzzle .
the situation is more complex than the original picture , due to our poor knowledge of its production mechanism in @xmath11 collisions and to the existence of many cold nuclear matter effects which significantly modify this production even in the absence of a qgp .
efforts are being made to better understand the above so that one can quantify the _ hot _ , abnormal effects at both sps and rhic .
notably , it appears that the suppression measured at sps in in - in collisions can be entirely described in terms of such cold nuclear matter effects . 00 pisarski r d 1982 , _ phys .
lett . _ * 110*b 155 brown g e and rho m 2002 , _ phys .
rep . _ * 363 * 85 rapp r and wambach j 2000 , _ adv .
* 25 * 1 stankus p 2005 , _ ann .
nucl . part .
_ * 55 * 517 turbide s
_ 2004 , _ phys .
_ c*69 * 014903 baier r , schiff d and zakharov b g 2000 , _ annu .
nucl . part .
_ * 50 * 37 gyulassy m _
nucl - th/0302077 matsui t and satz h 1986,_phys .
_ b * 178 * 416 andronic a _ et al .
_ 2003 , _ phys .
36 thews r l 2007 , _ nucl .
_ a*783 * 301 arnaldi r _ et al . _
( na60 collaboration ) 2006 , _ phys .
lett . _ * 100 * 022302 rapp r 2002 arxiv : nucl - th/0204003 adare a _ et al . _
( phenix collaboration ) 2007 , arxiv : 0706.3034v1 [ nucl - ex ] turbide s , rapp r and gale 2004 , _ phys .
_ c * 69 * 014903 arnold p , moore g d , yaffe l g 2001 , _ jhep _ 0112 , 9 akiba y ( phenix collaboration ) , this proceedings adare a _ et al .
_ ( phenix collaboration ) , arxiv : 0804.4168v1 [ nucl - ex ] denterria d and peressounko d 2006 , _ eur .
c*46 * 451 aggarwal m m _ et al . _
( wa98 collaboration ) 2000 , _ phys .
lett . _ * 85 * 3595 turbide s , rapp r , gale c 2004 , _ phys .
_ c * 69 * 014903 abelev b i _ et al . _
( star collaboration ) 2007 _ phys .
lett . _ * 98 * 192301 adare a _ et al . _ ( phenix collaboration ) 2007 , _ phys .
lett _ * 98 * 172301 dion a ( phenix collaboration ) , this proceedings adare a _ et al . _
( phenix collaboration ) 2009 , _ phys lett .
_ b * 670 * 313 engelmore t ( phenix collaboration ) , this proceedings abelev b i _ et al . _
( star collaboration ) 2005 , _ phys .
lett . _ * 94 * 062301 abelev b i _ et al . _ ( star collaboration ) , arxiv:0805.0364 dokshitzer y l and kharzeev d e 2001 , _ phys . lett .
_ b * 519 * 199 .
armesto n _
_ 2006 , _ phys .
_ b * 637 * 262 van hees h , greco v and rapp r 2006 , _ phys .
_ c * 73 * 034913 moore g d , teaney d 2005 , _ phys .
_ c * 71 * 064904 dunlop j c , this proceedings cacciari m _
_ 2005 , _ phys .
lett _ * 95 * 122001 ; private communication da silva c l ( phenix collaboration ) , this proceedings haberzetti h and lansberg j p 2008 , _ phys . rev .
lett _ * 100 * 032006 collins j c and soper d e 1977 , _ phys . rev .
* 2219 faccioli p , lourenco c , seixas j , woehri h k 2009 , arxiv:0902.4462v1 [ hep - ph ] adare a _ et al . _
( phenix collaboration ) 2008 , _ phys .
_ c * 77 * 024912 eskola k j , paukkunen h , salgado c a 2009 , arxiv:0902.4154v1 [ hep - ph ] scomparin e ( na60 collaboration ) , this proceedings lourenco c , vogt r , woehri h k 2009 , arxiv:0901.3054 [ hep - ph ] arleo f and tram v n 2008 , _ eur .
c*55 * 449 liu h ( star collaboration ) , this proceedings linden levy l a , this proceedings | this contribution summarizes the main experimental results presented at the 2009 quark matter conference concerning single and dilepton production in proton and heavy ion collisions at high energy .
the dilepton invariant mass spectrum has been measured over a range that extends from the @xmath0 mass to the @xmath1 mass , and for various collision energies at sps , fermilab , hera and rhic .
this paper focuses on the various contributions ( photons , low mass vector mesons , open and hidden heavy flavors ) to this spectrum and discuss their implications on our understanding of the matter formed in heavy ion collisions . |
an immense number of recent experiments on high @xmath13 superconductors aims at a determination of the form of the bcs gap function in momentum space.@xcite josephson junction experiments,@xcite measurements of the london penetration depth,@xcite and of the cu nmr relaxation rate @xcite are consistent with a @xmath11-gap . particularly interesting
are the arpes data which provide rather precise information about the detailed angular dependence of the amplitude of the gap function .
the experiments on the bi2212 compound show that the order parameter is maximal along the @xmath14 direction @xcite and that its amplitude in the @xmath15 direction seems to attain a nonzero value at a new critical temperature below @xmath13 .
@xcite the @xmath16 symmetry of the gap function is generally considered to be a sign of a pairing interaction of electronic origin , implying the absence of the standard phononic mechanism for superconductivity .
the idea of a superconducting state induced by fluctuations of purely electronic origin in systems of electrons with coulomb repulsion is originally due to kohn and luttinger @xcite for the case of the three dimensional electron gas .
similar effects exists in a two - dimensional electron gas , and generally they depend strongly on the form of the fermi surface .
perturbative calculations of the four point vertex for a weakly filled band in the hubbard model show that the model is instable against @xmath9 superconductivity @xcite .
quantum monte carlo calculations on the same model , in the vicinity of metal - insulator transition , show that the attractive pairing interaction of @xmath11 symmetry is dominant,@xcite , in agreement with the earlier arguments that antiferromagnetic fluctuations are the mediator of pairing interactions @xcite .
direct evidence for @xmath11 superconducting ordering has however not yet been found in quantum monte carlo studies .
antiferromagnetic fluctuations become stronger and stronger as one approaches half - filling . for weakly interacting electrons ,
these fluctuations are associated with the @xmath17-divergence of the electron
hole ( e - h ) loop diagram , caused by the nesting property of the fermi surface and the van hove singularities . on the other hand the bcs fluctuations , characterized by the electron - electron ( e - e ) loop , normally only linear in logarithms , crosses over to a @xmath17 form in the vicinity of half - filling .
the perturbative treatment of an interacting system of electrons should thus be based on the summation of all iterations of these two types of loops .
the renormalization group is one way to do this .
applied to interactions between electrons placed at the van hove points it gives an antiferromagnetic instability at half - filling and superconductivity of @xmath11 symmetry if the deviation of the chemical potential @xmath18 from its value at half - filling becomes of the order of critical temperature of the antiferromagnetic state.@xcite the equivalent parquet approach has been used for half - filling and also finds the antiferromagnetic state.@xcite a direct calculation of the zero temperature free energy @xcite up to the second order in the bare interaction @xmath19 confirms that at half - filling the antiferromagnetic order is stable , but finds no finite superconducting order parameter at any filling . in the present analysis
we search to know whether the hubbard model with repulsive on - site interaction can lead to superconductivity , and if it does , to what form of the gap function .
in particular , we are interested in the dependence of the results on the density of electrons .
it is hoped that the results can help us to clarify the origins of the existence of a highly anisotropic bcs gap function in the cuprates .
the renormalization group technique for fermionic systems in two and three dimensions has recently been developed , but with very drastic limitations .
the wilsonian mode - elimination technique was applied by shankar @xcite only to systems with an either isotropic or open , perfectly nested fermi surface .
weinberg @xcite has written the flow equations for a general case of an anisotropic fermi surface , but taking into account only the electron - electron channel of the flow . moreover , a common tendency is to do the renormalization group procedure only for a thin ring of degrees of freedom around the fermi surface and to linearize the spectrum in the radial direction , taking as a starting model the low - energy effective action .
this makes it difficult or impossible to make statements about the phase diagram of lattice models like those relevant for the description of the cuprates , due to the absence of a proper description of high - energy degrees of freedom , the elimination of which may considerably affect the effective low energy action .
we formulate the renormalization group starting from the hubbard model , with the whole brillouin zone involved in renormalization .
we do no tree level scaling before the energy shell @xmath20 of available states around the fermi surface contains no van hove singularity .
once @xmath0 has become the smallest energy scale , performing a tree level analysis and deriving the renormalization group equations , we show that the effective coupling function contains some relevant contributions with the origins in the regime with higher @xmath0 , i.e. from the electronic degrees of freedom not included in the low energy effective action .
we must add that our renormalization group , since perturbative in interaction , can provide uniquely the information whether the fermi liquid is a fixed point or not .
if the coupling flows to strong coupling , we can say in which direction it flows , for example in the @xmath16type superconducting direction , but we can not say whether another fixed point with finite superconducting order parameter with @xmath16 symmetry exists or not . this kind of problem is well known e.g. from the renormalization group in quasi - one - dimensional compounds , where the most divergent flow in some direction is always associated with the corresponding long range order ( lro ) because already infinitesimal interchain coupling suffices for its stabilization.@xcite similarly , the dimensionality reason for the non - existence of lro in two dimensions at finite temperature can be ignored as soon as small hopping in the third direction exists .
this , however , does nt mean that lro and fermi liquid fixed points are the only possible , on this question the one - loop renormalization group simply can not give an answer .
the problem is formulated in section 2 in terms of the effective action and the corresponding renormalization group flow equation for the coupling function . in section 3 , after calculation of the crossover energy to the purely electron - electron ( e - e ) part of the flow , we derive the renormalization group equation for the pairing function @xmath5 . in section 4 .
we diagonalize the pairing @xmath5 for the case of a non - divergent electron - hole ( e - h ) channel and determine the most attractive eigenfunction and the resulting critical temperature as functions of the chemical potential @xmath18 . in section 5 we give our picture of the gap viewed by arpes experiments .
the conclusions are given in section 6 .
the hubbard model for a two - dimensional system of electrons on a square lattice is given by @xmath21 where @xmath22 and the momenta are within the first brillouin zone .
for this model we introduce now the action in terms of fermion coherent states represented by grassmann variables@xcite @xmath23 and @xmath24 , where @xmath25 .
we will write it in the form @xmath26 @xmath27 where @xmath28 constrains all four momenta to run within the energy shell @xmath29 around the fermi surface . the energy and momentum
are conserved so that @xmath30 .
note that the size of the cutoff is equal to the bandwidth , i.e. the whole brillouin zone is available for integration .
thus , the @xmath31 functions have no meaning yet : they become important when the cutoff , reduced by the renormalization group , become lower than a tilt from the fermi level to the band boundary .
note that for a non - half filled band the effective phase space is not particle - hole symmetric .
the renormalization group transformation that we will use , known as the field theory approach , is defined as the mapping @xmath32 where @xmath33 and @xmath1 depend on @xmath34 in a way so that the physical properties of @xmath35 and @xmath36 are the same for energies lower than @xmath37 .
this requirement is fulfilled if all one - particle irreducible vertices are invariant under reduction of the cutoff from @xmath38 to @xmath39 .
the renormalization group can be thought of as a set of successive , infinitesimally small steps @xmath40 .
this allows us to formalize the renormalization group requirement in a set of equations @xmath41 , where @xmath42 .
up to second order in @xmath1 , it suffices to consider only @xmath43 and @xmath44 , because all higher vertices are of higher order in @xmath1 . the solutions of the equations @xmath45 and
@xmath45 give us the renormalization group flow for @xmath46 and @xmath47 .
conservation of spin allows us to write the interaction part of the action as a sum of the singlet @xmath48 and triplet @xmath49 parts : @xmath50 where @xmath51 and @xmath52 are the variables of annihilation of the singlet and triplet states @xmath53 @xmath54 the singlet state is symmetric and the triplet antisymmetric under exchange of the momenta of two particles .
correspondingly , the coupling function @xmath55 can be taken to be symmetric and @xmath56 to be antisymmetric under the momentum exchange operation @xmath57 , defined as @xmath58 @xmath59 being a function of four energy
momenta which conserves energy and momentum . if @xmath59 possesses time - reversal symmetry @xmath60 which certainly is a property of the vertex , then it is equivalent whether @xmath57 exchanges @xmath61 and @xmath62 or @xmath63 and @xmath64 , i.e. @xmath65 .
formally , @xmath66 and @xmath67 are given by @xmath68 on the other hand , the interaction can also be written as a sum of one term with equal @xmath69 and one with opposite @xmath70 spin quantum numbers , with corresponding coupling functions named @xmath71 and @xmath72 , respectively . from two equal - spin electrons
one can build only a triplet state , which make us conclude that @xmath73 while @xmath74 containing the singlet and the triplet interactions .
another way to write the interaction is in terms of a charge and a spin part @xmath75 where @xmath76 and @xmath77 are @xmath78 the charge and spin coupling functions are @xmath79 in first order of perturbation theory , these functions determine the renormalization of the charge - charge and the spin - spin correlation functions .
we proceed now with the derivation of renormalization group equations . for simplicity
, we will ignore the renormalization flow of @xmath46 , which follows from conservation of @xmath43 , renormalizing the form of the fermi surface , the effective mass , etc .
this approximation is justified in the case of the circular fermi surface @xcite . in the anisotropic case ,
the diagrams for @xmath43 have a dependence on the direction of * k*. moreover , even a small renormalization of the fermi energy can give important changes of the form of the fermi surface if one is close to half filling , because of van hove singularities .
for filling not too close to one - half we can expect that the essential of the physics is given by just the renormalization of the coupling @xmath1 using the bare dispersion relation @xmath80 , which we will call @xmath46 from now on .
the feynman diagrams for @xmath43 and @xmath81 are given in fig .
the first loop in the expression for @xmath44 is of the electron - electron ( e - e ) and all others of the electron - hole ( e - h ) type . making use of the relations ( [ upar ] ) , ( [ uper ] ) , and ( [ as ] )
, we get the expression for @xmath44 in terms of @xmath1 and @xmath82 .
if we write the integration measure of the loop diagrams in the form @xmath83 @xmath51 being the curves of constant energy @xmath84 , then @xmath85 corresponds to the integration of the two energy shells of width @xmath86 with @xmath87 the functionals @xmath88 and @xmath89 are the partial derivatives with respect to @xmath34 of the e - e and e - h loops and both are bilinear forms in @xmath90 and @xmath91 .
they read @xmath92 and @xmath93 with @xmath94 @xmath95 @xmath96 @xmath97 the index @xmath98 symbolizes two energy shells at @xmath99 and @xmath100 ; @xmath101 stands for @xmath102 ; @xmath103 ; @xmath104 ; @xmath105 ; @xmath106 ; @xmath107 , where @xmath108 is the momentum running along the path @xmath109 .
@xmath110 and @xmath111 are non - analytic functions of the momenta , given by @xmath112 their origin is in the derivatives over @xmath0 of the products like @xmath113 when @xmath114 .
a particularity of the renormalization group approach treating e - e and e - h fluctuations in more than one dimension is the absence of self - similarity of the problem .
in fact , there is an intrinsic energy scale which is a function of the band filling .
it is associated with charge and spin fluctuations coming from the e - h term .
we will proceed by estimating the characteristic energy scales which appear in @xmath115 and @xmath116 , when all four particles are at the fermi surface , with zero energy .
if we are exactly at half - filling , it is known that in the limit @xmath117 both e - e and e - h loops scale like @xmath118 , which corresponds to the square - logarithmic divergence in both channels .
this gives an explicit @xmath119 dependence in the @xmath120-functionals .
let us suppose now that the filling is slightly lower than one - half , i.e. that @xmath121 is small and negative .
we expect two regimes .
one is for @xmath122 , where the flow is still unaffected by the small changes of the fermi surface due to nonzero @xmath121 and remains proportional to @xmath34 . in the second regime ,
where @xmath123 , the e - e flow is just a constant ( i.e. only a @xmath124-divergence ) , while the e - h flow decays exponentially due to disappearance of nesting . even far from half - filling it
is possible to define a crossover @xmath125 , beyond which the flow in the e - h channel disappears exponentially .
we can summarize saying that for any filling , @xmath125 is a crossover from a regime where both e e and e h loops contribute to a regime where @xmath116 starts to behave like @xmath126 . here
@xmath127 is positive for all values of the momentum transfer @xmath128 . to estimate the dependence of @xmath125 on the filling @xmath129 , we consider the static limit of the partially integrated e - h loop @xmath130 with the momentum transfer * q * equal to @xmath131 in the direction @xmath132 .
note that the energy integration is performed over @xmath133 .
the derivative of @xmath134 with respect to @xmath34 gives the explicit @xmath34-dependence in the @xmath135 functional .
2 shows @xmath136 and @xmath137 for two different values of @xmath18 .
it is reasonable to define @xmath125 as the point where @xmath137 starts to decrease . in the exponential regime
the function @xmath137 decays like @xmath138 ( i.e. @xmath139 , valid for any orientation of @xmath140 ) , while the regime @xmath141 remembers the @xmath142 divergence of @xmath143 at half filling .
if we consider @xmath136 for some large momentum transfer different from @xmath131 ( giving intersection rather than touching of the initial- and final - state fermi surface ) we get a shape like @xmath144 ( i.e. @xmath145 ) . the dependence @xmath146 is shown in fig.3 .
near @xmath147 there is a divergence of the form @xmath148 because of nesting , while the increasing @xmath125 as the filling goes to zero mirrors the fact that , for low density , the fermi energy appears as the new scale instead of the band - width being used .
the insert shows the function @xmath149 .
once in the exponential regime , @xmath116 can be neglected after it becomes smaller than @xmath150 of its value at @xmath151 .
putting @xmath152 , this defines the crossover @xmath153 corresponding to the crossover temperature @xmath154 .
suppose that we now integrate the flow equation ( [ flowu ] ) from @xmath155 to @xmath156 .
once @xmath34 has reached @xmath157 , only the term @xmath158 remains in the flow equations , and one has a partially renormalized @xmath159 as initial condition .
now we use the fact that @xmath160 is a small parameter , i.e. the inequality @xmath161 determines a thin ring of degrees of freedom , containing no van hove points , as one can conclude looking at fig.3 .
this allows us to rescale the momenta @xmath162 , where @xmath163 is the unit vector normal to fermi surface , dependent on direction of * k*. to clarify the reason for which a tree - level scaling is not allowed for energies higher than the deviation of the fermi level from van hove singularity , let us write the phase space integration measure in terms of energy @xmath164 and polar angle @xmath165 variables @xmath166 with @xmath167 , @xmath168 being the radial wave number and @xmath169 the group - velocity .
zeroth order ( tree level ) scaling tells us via power counting argument to consider @xmath170 as function of @xmath171 , neglecting any @xmath84 dependence about @xmath172 , which is possible if @xmath170 is an analytic function of @xmath84 at the whole shell @xmath173 which , consequently , should contain no singularity .
we can also rescale the frequencies if @xmath174 is an analytic function of @xmath175 in the interval @xmath176 about the fermi surface , which we assume to be the case@xcite . in the scope of this tree - level scaling , as it has already been shown by shankar @xcite , the slope of the electronic dispersion around the fermi surface is irrelevant and the two marginal interactions correspond to two different constraints on the four - momenta in @xmath1 .
since any @xmath177- and @xmath178-dependence in @xmath1 is irrelevant , both marginal interactions depend only on coordinates of the zero frequency particles placed at the fermi surface .
for the first , `` fermi liquid '' or forward interaction , the momenta satisfy the equation @xmath179 , where the meaning of momenta can be seen from the equation ( [ act ] ) .
this interaction is slightly @xmath180 renormalized by the high - energy - modes ( @xmath181 ) , and is not involved in further renormalization .
the second interaction is the pairing potential @xmath5 , where the momenta satisfy the condition @xmath182 .
the pairing @xmath5 depends only on angular coordinates of annihilated and created pairs .
keeping in mind the above remarks , we can write the action for the electrons in the ring @xmath176 around the fermi surface as @xmath183 @xmath184 where @xmath185 , closed loop integrations are over the the fermi surface and @xmath186 stands for the effective fermi liquid interaction .
@xmath187 is the energy - integrated number of pairs defined as @xmath188 note that the integration measure over small momentum @xmath189 goes to zero as @xmath190 .
a form similar to ( [ hlin ] ) has been used by weinberg@xcite , but taking @xmath191 phenomenologically and not as the partially renormalized pairing interaction which we get from @xmath192 putting incoming particles 1 and 2 to @xmath193 and @xmath194 and outgoing 3 and 4 to @xmath195 and @xmath196 . note that the loop integration over @xmath51 can be understood as the scalar product over `` vector components '' of a `` spin '' , where the number of components @xmath197 corresponds to @xmath198 @xcite .
the integrations have the weight factor @xmath199 , @xmath200 being the anisotropic fermi velocity , what suggests to introduce a new angular coordinate @xmath201 where @xmath202 is the density of states at the fermi level .
the function @xmath203 is shown in fig.4 for few different values of @xmath18 .
starting from the new hamiltonian ( [ hlin ] ) we can calculate now the function @xmath204 in @xmath205-space and obtain the flow - equation @xmath206 where the coordinate @xmath205 appears instead of @xmath207 .
for initial condition we take @xmath208 .
to make the differential equation ( [ flowv ] ) solvable one has to diagonalize the pairing potential @xmath209@xcite since it is invariant under all symmetry elements of the @xmath210 point group , its most general form in @xmath205-space can be written as @xmath211 where @xmath212 and @xmath213 .
the function @xmath214 is the @xmath215 basis state of the @xmath216representation of the point group @xmath210 .
it is proportional to the function @xmath217 , @xmath218 , @xmath219 , @xmath220 , and @xmath221 $ ] , for @xmath222 , and @xmath223 respectively . using ( [ decomp ] ) ,
the flow equation ( [ flowv ] ) becomes @xmath224 with the initial condition @xmath225 to solve exactly the equation ( [ flowdec ] ) one has to diagonalize five infinite dimensional matrices @xmath226 , thus decoupling completely the flow ( [ flowdec ] ) into a set of differential equations whose solution is @xmath227 here @xmath228 labels the eigenvalues within the representation @xmath216 .
if @xmath229 is negative , the denominator has a zero at @xmath230 and an instability occurs .
the renormalization group calculation of the pairing interaction @xmath231 is an extremely difficult problem because of the interplay between the cutoff @xmath0 and the geometry of the two - dimensional shells @xmath232 .
however for some range of @xmath233 and @xmath19 ( essentially small @xmath19 and @xmath234 not too close to half - filling ) the problem can be reduced to only one simple integration , i.e. @xmath235 with @xmath236 where @xmath237 is the partially integrated e - e loop , defined in a same way as @xmath136 in eq.([bubble ] ) .
@xmath136 comes just from the last diagram in the expression for @xmath44 in fig . 1 , because the other three e - h loops cancel exactly for any @xmath1 symmetric under exchange .
the approximation ( [ vco ] ) is allowed if @xmath238 is a small parameter . a good estimate of the magnitude of this parameter is given by the @xmath239-part of @xmath240 , giving the criterion for the validity of our approach in @xmath241-space .
the line @xmath242 is shown in fig .
5 . below the line
the approximation ( [ i0 ] ) is justified . by definition of @xmath157 , @xmath243 .
moreover , @xmath244 has no dependence on @xmath205 and @xmath245 , since it depends on external momenta only through @xmath246 , which we put to zero .
consequently , its only nonzero component is @xmath247 .
thus , for the calculation of all other components of @xmath248 we use just the bubble @xmath249 with the momentum transfer @xmath250 .
fig.6(a ) shows @xmath249 as a function of @xmath205 and @xmath245 for chemical potential @xmath251 .
the minimal eigenvalues of @xmath231 in all five channels , named @xmath252 , are shown in fig.6(b ) as functions of @xmath18 .
these curves indicate which kind of superconducting symmetry becomes critical at some given @xmath18 .
the eigenvalues for each channel are calculated taking only the first four harmonics for @xmath239 , @xmath253 , @xmath10 and @xmath8 , and the first six harmonics for the @xmath223 representation .
the corresponding eigenvectors determine the fourier spectrum of the gap function .
a very important result is that the relevant harmonic of the superconducting fluctuations in the @xmath10 channel occurs very close to just @xmath254 , being thus determined only by the structure of the fermi surface and not by the interaction .
7(a ) show the instable order parameters @xmath255 for @xmath256 , @xmath257 for @xmath258 and @xmath259 for @xmath260 as a function of the fermi surface angle @xmath193 .
the evolution of the function @xmath261 ( i.e. the first harmonic of @xmath10 ) with @xmath262 , given only by the dependence of @xmath205 on @xmath171 , is shown in fig .
the strength of the peaks near the van hove points increases and the magnitude in the area between the peaks decreases with @xmath262 .
the critical temperature is given by a cutoff for which the most attractive diagonal component of @xmath5 diverges , i.e. @xmath263,\ ] ] where @xmath264 .
. shows @xmath265 as a function of @xmath266 .
the critical temperature decreases extremely fast as we go away from the half - filling .
an increase of @xmath19 could save the situation , but in that case our perturbative method ceases to be sufficient ( see fig .
5 ) . since the cuprates are superconductors for fillings quite far from one electron per site ( @xmath267 )
, this result means that the small-@xmath1 hubbard model can not describe these systems quantitatively .
however , the model gives very precious informations about the form of the gap function in the @xmath10instable regime , which will not change considerably with increasing @xmath19 , as long as @xmath268 is the dominant attractive harmonic in @xmath231 .
once the renormalization flow has been integrated for @xmath181 , assuming that the interaction did not diverge earlier in the antiferromagnetic channel , the detailed angular dependence of the superconducting gap function can be easily found . in general , a superconducting state with the symmetry corresponding to the lowest of the eigenvalues @xmath269 will be formed .
however , when two of the @xmath269 are close to each other , a more complicated situation can occur : for definiteness , consider the region @xmath270 in fig.6(b ) , where the @xmath8 eigenvalue is the most attractive after the @xmath10 .
let us suppose that @xmath10 order of the simplest form @xmath271 has formed and that the temperature is close to @xmath13 . among the remaining symmetry channels ,
@xmath8 is the only one which can give a large gap function in the node - points of @xmath259 and zero in the points where @xmath259 is maximal .
consequently , we expect that the flow of the type ( [ solution ] ) with @xmath272 will not be strongly affected by the existing @xmath10 order .
considering to a first approximation the two flow equations ( @xmath273 and @xmath272 ) as independent , and taking only the first harmonics of the @xmath10 and @xmath8 representations , we can construct the relevant part of the pairing interaction which gives two phase transitions , one with @xmath10 and the other with @xmath8-symmetry : @xmath274 where all details of the fermi surface are contained in the dependence of @xmath205 on @xmath171 . from eq.([solution ] ) one finds that the ratio between two critical temperatures is given by @xmath275 $ ] . from fig.6(b ) note that the ratio @xmath276 is very sensitive to the variation of the chemical potential .
the gap function resulting from ( [ vapr ] ) has the form @xmath277 where @xmath259 , @xmath255 , @xmath278 and @xmath279 are real .
these parameters can be determined minimizing the mean - field expression for the free energy per site@xcite @xmath280 where @xmath281 and the theta function constrains the momentum summation to run only over the states within the energy shell @xmath176 about the fermi surface .
the minimization of @xmath282 with respect to @xmath283 gives @xmath284 i.e. the resulting gap function is of the type @xmath285 .
it is interesting to remark that the same kind of gap function has been obtained by laughlin using the anyon picture.@xcite the particularity of this gap function ( and of any gap consisting of two different symmetry terms with a phase difference of @xmath286 ) is that it breaks time reversal symmetry .
we can now try to understand recent arpes measurements by jian ma and coworkers @xcite on the bi2212 compound . from their experiment
it appears that two superconducting instabilities occur ; the first one is at @xmath287 and has probably the @xmath10 symmetry .
the second instability occurs at @xmath288 ; it introduces a nonzero gap at the points @xmath289 , i.e. halfway between the corners of the half - filled fermi .
the function measured by arpes is @xmath290 and in the picture discussed above has no zeros and minima on the diagonals of the brillouin zone if both @xmath259 and @xmath255 are finite , @xmath291 and @xmath292 .
this is in agreement with the experiments because the gap in the diagonal direction is just equal to @xmath255 , introduced at @xmath293 .
the minimum of @xmath290 on the diagonals is in agreement with other arpes experiments@xcite as well .
one should of course notice that in our model closeness of two different @xmath228 s only occurs in very narrow parts of the parameter space and therefore to a certain degree is accidental .
we have formulated the one - loop renormalization group for a two - dimensional system of interacting electrons on a square lattice , described by the hubbard model . for band filling different from one - half , the renormalization flow for @xmath34 superior to some crossover @xmath157 comes only from the contribution due to the electron - electron bubble - diagram , while the electron - hole contribution decays exponentially as @xmath294 , where @xmath157 depends on band filling , but is independent on the strength of the interaction .
we decompose the bcs pairing interaction @xmath5 for electrons in the vicinity of the fermi surface in fourier components of the five irreducible representations of the @xmath210 point group , defined at the fermi surface .
diagonalizing @xmath231 in each representation , we get five sets of decoupled bcs flow equations .
the minimal ( i.e. the most negative ) eigenvalue of @xmath231 determines the critical temperature and the eigenvector gives the form of the gap function . unlike the usual approach,@xcite
the characteristic of the procedure presented here to obtain the symmetry of the gap function is that only the axial coordinates at the fermi surface is relevant , while the radial dependence is `` scaled out '' .
moreover , the renormalization group treatment of the whole brillouin zone , and not only of the narrow belt about the fermi surface has allowed us to show that the origin of the attractive part of the pairing interaction in the hubbard model is in electron - hole fluctuations on rather high energy scales , up to the bandwidth .
we have calculated @xmath231 for the case where the flow due to the electron - hole channel can be treated perturbatively , i.e. when the filling is far enough from one - half .
the diagonalization of @xmath5 in terms of angular harmonics gave us the type of superconducting instability : for weak filling , the instability occurs in the @xmath8 ( @xmath9 ) singlet channel , while for filling close to one - half , the @xmath10 ( @xmath11 ) singlet instability strongly overwhelms all others , what are the results in agreement with previous work.@xcite particularly interesting is the fact that the order parameter @xmath295 can be very well approximated by the function @xmath261 , where the function @xmath203 depends only on the anisotropic fermi velocity and on the geometry of the fermi surface .
this gives for @xmath295 a function that has peaks in the directions of the van hove singularities .
the slope of the peaks increases as we approach half - filling .
this can be a justification to consider the interaction only between electrons in the close vicinity of the van hove points as relevant if we are very close to half - filling @xcite .
we believe that the form of @xmath295 does not depend considerably on the strength of interaction , and recent t - matrix calculations @xcite for realistic values of the interaction @xmath19 give a gap function in accord with our assumptions .
we find a superconducting instability at any electron concentration away from half filling .
the underlying physical mechanism , namely exchange of spin or charge density fluctuations , is the same as in previous approaches.@xcite we do however feel that our present results are on a more solid footing than the previous work because the present one loop renormalization group scheme does not make any a priori assumptions about important or unimportant diagrams and provides a more systematic way of handling the dynamics of the fluctuations being exchanged .
the only restrictions come from ( i ) the limitation to one loop order , necessitating weak coupling , and ( ii ) the requirement that the e - h diagram are a perturbation with respect to the e - e diagrams , implying that we can not be too close to half filling .
the region of validity of the approach is shown in fig.5 .
further , self
energy diagrams have been neglected , however , these are expected to produce important effects only at two loop order , and therefore are expected to be negligible in weak coupling . in our weak coupling model
the superconducting critical temperature is negligible ( but it exists , for any filling ! ) if we are not in the immediate vicinity of half - filling , which means that the hubbard model with small @xmath19 and small ( perturbative ) antiferromagnetic fluctuations does not suffice to describe the high critical temperature ( @xmath296 ) of the cuprates .
there exist two possible ways ( related to two restrictions of our calculations ) to increase the critical temperature .
the first is to simply increase @xmath19 and to remain far from the half - filling , keeping the e - h channel non - singular . to treat this case ,
an approach perturbative in @xmath19 is possibly only of limited use .
ideally , renormalization should be done exactly , and not using a simple one - loop ( or @xmath297-loop ) scheme ( which is actually just an `` intelligent '' version of the perturbative summation ) .
we can speculate and suppose that even in the case of strong coupling there exists the crossover @xmath298 , above which the flow is of the bcs type .
it is to expect that @xmath298 is not very different from @xmath157 that we have calculated .
this means that the cutoff @xmath299 for the effective bcs theory is @xmath300 ( see fig.3 . and eq.([co ] ) ) , where @xmath38 is the initial cutoff of the theory , equal to @xmath301 .
the second possibility to increase @xmath13 is to approach half - filling very closely , making fermi surface nesting important but remaining in a weak - coupling regime . for that case , a simplified one - loop renormalization group calculations
@xcite has shown that superconductivity wins over antiferromagnetism only if the e - h contribution to the flow decays before the divergence in the antiferromagnetic channel takes place .
thus , we can say that here too , the effective theory is of bcs type . the difference with the first scenario is that the effective cutoff is very small ( fig .
3 ) , and that the coupling constant is very strong , due to the strong flow in both e - e and e - h channels at all scales @xmath181 .
a very important feature of the nested case with a small @xmath19 is that it can be treated in terms of the one - loop renormalization group , renormalizing @xmath1 as a function only of three angular variables .
this is allowed because all important physics ( i.e. the majority of the e - e and e - h flow ) is contained in the vicinity of the fermi surface , making the effective phase space to be a rather narrow square ring @xmath302 ; @xmath303 where the marginally relevant interaction is a function only of the angular position of the particles on the ring .
finally , we have discussed the possibility of a superconducting state with a mixed symmetry . in the presence of @xmath10 order , the flow in the @xmath8 channel ( which is the second most attractive one for @xmath304 )
will be only weakly affected by a nonzero order parameter of @xmath10 symmetry .
this gives rise to two superconducting instabilities , with the critical temperatures @xmath13 for the @xmath10 and @xmath305 for the @xmath8 channel , and @xmath306 .
we have given the form of the pairing function for the effective bcs theory .
at @xmath307 the relative phase of two order parameters is @xmath292 .
the resulting form of the energy gap @xmath308 has no zeros and minima are in diagonal directions , providing a possible qualitative explanation of arpes experiments by shen@xcite and the decrease of the anisotropy with decreasing @xmath309.@xcite for recent short overview see : j. r. schrieffer , solid state commun . *
92 * , 129 ( 1994 ) .
d. a. wollman _
et al . _ ,
74 * , 797 ( 1995 ) .
j. r. kirtley _
et al . _ ,
nature * 373 * , 225 ( 1995 ) .
r. c. dynes , solid state commun .
, * 92 * , 53 ( 1994 ) .
w. n. hardy _ et al . _ ,
70 * , 3999 ( 1993 ) .
j. a. martindale _ et al .
_ , phys . rev . *
b47 * , 9155 ( 1993 ) .
z .- x . shen _
et al . _ ,
science * 267 * , 343 ( 1995 ) .
m. r. norman , m. randeria , h. ding , j. c. campuzano , and a. f. bellman , preprint ( submitted to sissa , cond - mat 9507021 ) .
jian ma _ et al .
_ , science * 267 * , 862 , ( 1995 ) .
w. kohn and j. m. luttinger , phys .
lett . * 15 * , 524 ( 1965 ) .
m. a. baranov , a. v. chubukov , and m. yu .
kagan , int .
phys * b6 * , 2471 ( 1992 ) .
n. bulut , d. j. scalapino , and s. r. white , phys . rev . *
b50 * , 9623 ( 1994 ) .
v. j. emery , synt . met .
* 13 * , 21 ( 1986 ) .
k. miyake , s. schmitt - rink , and c. m. varma , phys . rev . * b34 * , 6554 ( 1986 ) .
d. j. scalapino , e. loh , jr . and j. e. hirsch , phys . rev . *
b34 * , 8190 ( 1986 ) .
n. e. bickers , d. j. scalapino and r. t. scalettar , int .
b1 * , 687 ( 1987 ) .
n. e. bickers , d. j. scalapino and s. r. white , phys .
* 62 * , 961 ( 1989 ) .
n. e. bickers and d. j. scalapino , ann .
( n.y . ) * 193 * , 206 ( 1989 ) . n. e. bickers and s. r. white , phys .
rev . * b43 * , 8044 ( 1991 ) .
h. j. schulz , europhys . lett . * 4 * , 609 ( 1987 ) .
i. e. dzyaloshinskii , sov .
phys . jetp*66 * , 848 ( 1988 ) .
i. e. dzyaloshinskii and v. m. yakovenko , sov .
phys . jetp*67 * , 844 ( 1988 ) .
a. georges and j. s. yedidia , phys . rev .
* b43 * , 3475 ( 1991 ) .
r. shankar , rev .
* 66 * , 129 ( 1994 ) .
s. weinberg , nucl .
b413 * , 567 ( 1994 ) .
h. j. schulz , in `` low - dimensional conductors and superconductors '' , eds .
d. jrome and l. g. caron , ( plenum , new york , 1987 ) .
j. w. negele and h. orland , `` quantum many - particle physics '' , addison wesley ( 1988 ) .
the four point vertex is not an analytic function of momenta and frequency , because the e - h bubble has a singularity at @xmath310 .
this aspect is of importance for the fermi liquid properties ( for details see : n. dupuis and g. y. chitov , preprint ) , and since we are interested in logarithmic instabilities of the model , we can ignore it .
s. k. ma , modern theory of critical phenomena , benjamin ( 1976 ) .
g. kotliar , phys . rev . *
b37 * , 3664 ( 1987 ) .
r. b. laughlin , physica * c234 * , 280 ( 1984 ) .
g. e. volovik and l. p. gorkov , sov .
jetp * 61 * , 843 ( 1985 ) .
t. dahm and l. tewordt , phys .
lett . * 74 * , 793 ( 1995 ) .
( a ) the shape of the function @xmath317 in @xmath318 space at @xmath319 .
the nesting at half filling occurs for @xmath320 .
the split singular lines show the best incommensurate nesting vector .
( b ) the minimal eigenvalue of the pairing @xmath321 in every of 5 irreducible representations of @xmath210 point group .
( a ) the shape of three possible gap functions : @xmath322 for @xmath323 ( dot - dashed ) ; @xmath324 for @xmath325 ( dashed ) and @xmath326 for @xmath327 ( solid line ) .
( b ) a very good approximation for @xmath328 is just @xmath261 , shown here for the same choice of @xmath18 as in fig .
4 . the dashed line shows @xmath329 to comparison .
the scale @xmath330 as a function of the logarithm of the chemical potential , for a few values of coupling @xmath19 .
for very small @xmath18 and for @xmath331 the curves are out of the range of validity ( see fig.5 ) . | the problem of weakly correlated electrons on a square lattice is formulated in terms of one - loop renormalization group .
starting from the action for the entire brillouin zone ( and not with a low - energy effective action ) we reduce successively the cutoff @xmath0 about the fermi surface and follow the renormalization of the coupling @xmath1 as a function of three energy - momenta .
we calculate the intrinsic scale @xmath2 where the renormalization group flow crosses over from the regime ( @xmath3 ) where the electron - electron ( e - e ) and electron - hole ( e - h ) terms are equally important to the regime ( @xmath4 ) where only the e - e term plays a role . in the low energy regime
only the pairing interaction @xmath5 is marginally relevant , containing contributions from all renormalization group steps of the regime @xmath3 .
after diagonalization of @xmath6 , we identify its most attractive eigenvalue @xmath7 . at low filling
, @xmath7 corresponds to the @xmath8 representation ( @xmath9 symmetry ) , while near half filling the strongest attraction occurs in the @xmath10 representation ( @xmath11 symmetry ) . in the direction of the van hove singularities , the order parameter shows peaks with increasing strength as one approaches half filling . using the form of pairing and the structure of the renormalization group equations in the low energy regime ,
we give our interpretation of arpes experiments trying to determine the symmetry of the order parameter in the bi2212 high-@xmath12 compound . |
a substantial fraction ( @xmath3% ) of main - sequence stars show evidence for excess ir or sub - mm flux due to thermal emission from dust located at distances of 30 au or more from the stars ; i.e. , locations comparable to that of the kuiper belt in our own solar system .
this was first discovered using iras observations ( e.g. , aumann et al . 1984 ; 1985 ; gillett & aumann 1983 ) , and subsequently , observations with the james clarke maxwell telescope s submillimeter common - user bolometer array ( jcmt / scuba ) have directly imaged the dust distribution in a few of these systems ( holland et al .
1998 ) , including @xmath0 eri ( greaves et al 1998 ) .
@xmath0 eridani is a k2v main - sequence star at a distance of about 3.2 pc .
it is believed to be a relatively young system ( @xmath4 gyr ; song et al .
2000 ; soderblom & dappen 1989 ) , with a mass slightly less than our own sun . from radial velocity measurements , hatzes et al ( 2000 ) have reported evidence for a planet in this system with a semimajor axis of 3.4 au and @xmath5 .
the 850@xmath6 m observations of greaves et al .
show a ring - like structure around @xmath0 eri .
the maximum surface brightness of this ring is located at a radius of @xmath217 `` ( 55 au ) from the star , with some flux extending out as far as 36 '' ( 115 au ) .
the observed 850@xmath6 m flux shows the ring to be asymmetrical , with several bright clumps .
it has been suggested that structures of this kind can be caused by resonant interactions of dust with planets in or near the ring ( liou & zook 1999 ; ozernoy et al 2000 ; quillen & thorndike 2002 ) . in an attempt to detect an optical counterpart of this ring we undertook observations with the _ hubble space telescope s _ space telescope imaging spectrograph ( _ hst_/stis ) of @xmath0 eri , using this instrument s ccd camera , as part of hst go program 9037 ( mario livio pi ) .
a full description of the stis instrument can be found in kim quijano et al .
( 2003 ) .
while this camera can be used with a number of coronagraphic wedges , saturation of the detector by the wings of the stellar psf near the edges of the wedge would severely limit the exposure time achievable in a single image .
therefore , we instead used the 52``x52 '' clear ccd aperture and placed the star @xmath7 " off the edge of the detector .
the k0 iv star @xmath8 eri was also observed as a psf comparison star .
gillett ( 1986 ) reconsidered the iras observations of @xmath0 eri and concluded that the intrinsic fwhm of the source flux at 60@xmath6 m was less than 17 `` in the iras scan direction and less than 11 '' in the perpendicular direction .
if this suggestion is correct , it would imply that the bulk of the iras emission comes from a region inside of the ring detected by greaves et al .
however , both the sub - mm ring and the iras size limits suggested by gillett are significantly smaller than the nominal iras 60@xmath6 m resolution at of about 1 ( beichman et al . , 1985 ) , and therefore gillett s conclusions should be treated cautiously .
greaves et al .
observed the system with scuba at both 850 and 450@xmath6 m .
the 850 @xmath6 m observations used a beam of 15 " fwhm , and greaves et al published smoothed versions of the images produced by these observations . the s / n of the 450@xmath6 m observations is too low to give any useful spatial information ( no image was published ) , but it does supply a useful measure of the total flux .
schtz et al . (
2004 ) obtained 1200 @xmath9 m measurements using a 25 " beam , that are consistent with greave et al.s measurements . several previous sub - mm observations ( chini et al . 1990 ; 1991
; zuckerman & becklin 1993 ) at various wavelengths ( 800 1300 @xmath9 m ) had used single pointings with beam - sizes and background chopping too small to properly measure the structure detected by greaves et al .
( see also weintraub & stern 1994 ) .
while these observations may provide some useful constraints , they can not be used as direct measures of the flux and are not further considered here .
we summarize the iras , greaves et al . , and schtz et al .
data in table [ fluxtab ] . in this table
we adopt greaves et al s color corrections , as well as their corrections for the stellar contribution to the total flux , and present only the flux attributed to the dust alone . at 55 au ,
the iras and scuba data imply grain temperatures of about 30 k
close to the equilibrium black body temperature .
the very flat 450@xmath6 m to 850@xmath6 m flux ratio ( @xmath10 ) requires the presence of large grains ( @xmath11 m ) , which can emit efficiently at sub - mm wavelengths . at this distance ,
the time scale for the orbital decay of 100@xmath6 m grains due to the poynting - robertson effect is about @xmath12yr ; a time scale which is comparable to the inferred age of the system .
this would seem to argue against a substantial population of smaller grains in the outer parts of the system .
however , recent works on various debris disk systems ( e.g. , wyatt et al . 1999 ; li , lunine , & bendo 2003 ) have shown that this kind of interpretation is overly simplistic .
the sub - mm emissivity of grains drops rapidly enough with decreasing grain size that a substantial population of smaller grains has little effect on the sub - mm flux ratios , and the collision and fragmentation rate of the large grains required by the sub - mm data is still high enough to replenish the smaller grains faster than the poynting - robertson effect can remove them .
we will see that the nature and abundance of these smaller grains has dramatic effects on the optical detectability of the dust in the @xmath0 eri system .
the hst observations in this program were done on january 26 , 2002 using six adjacent single - orbit visits ( see table [ obstable ] ) . in each visit , after taking an acq exposure to determine the position of the targeted star , 15 offset exposures of 109 seconds each were taken using the unfiltered stis ccd in imaging mode . during these offset exposures
the bright star was located about 5 " off the detector ( near ccd pixel coordinates x=1127 , y=514 ) .
the first and last orbits were used to observe the psf of the comparison star , @xmath8 eri at two different orientations differing by 30@xmath13 .
the four intermediate orbits were used to observe the primary target at four different orientations separated by 10@xmath13 intervals . in each of these latter exposures , the position of the brightest sub - millimeter clump observed by greaves et al .
was imaged on the 1024x1024 pixel detector .
the clear aperture used with the stis ccd is designated as the `` 50ccd '' aperture in stis documentation .
it has a field of view of almost 52``@xmath1452 '' , and has a very broad bandpass , with significant throughput from about 2000 to 10200 .
prior to the installation of acs , stis 50ccd observations provided the most sensitive _ hst _ mode for deep imaging .
we will , for the most part , give our observed stis magnitudes in stmag units , where the magnitude is defined as @xmath15 , with @xmath16 in ergs@xmath17s@xmath17cm@xmath1/ . for stis 50ccd imaging , the conversion to a magnitude system that uses vega as the zero point is vegamag(50ccd)@xmath19stmag(50ccd)@xmath20 .
in addition , at the beginning of each orbit a very short observation was taken of each star using the stis ccd with the f25nd3 filter ( table [ nd3images ] ) .
the standard stis pipeline software was used to produce bias and dark subtracted and flat fielded images of individual subexposures ( flt files ) .
comparison with median filtered images was done to identify and correct hot pixels that were not handled properly in the standard analysis ; about 0.5% of pixels were corrected in this way .
each of the 15 subexposure flt files was rearranged into five separate files , each containing three adjacent subexposures , and these files were input into the stsdas stis routine ocrreject to produce five separate cosmic ray rejected ( crj ) files for each of the six visits - 30 crj files in total . in each of these crj files ,
the two diffraction spikes from the star ( @xmath0 eri or @xmath8 eri ) located 5 " off the edge of the detector are the brightest features visible .
these were used to register the images .
as we are primarily interested in the relative offsets between the images , any small systematic offset from the real location of the star relative to the detector is unimportant .
each of the 30 crj files was then shifted so as to put the intersection of the diffraction spikes at the mean location measured for all visits .
the shifts applied were up to @xmath21 pixels in x and @xmath22 pixels in y. because we wanted to distort the high s / n pattern of the psf as little as possible , we shifted the images using a seven point sinc interpolation function with the iraf imshift routine .
the shifted crj files for each visit were then combined by again using the stis ocrreject routine , producing a single aligned crj file for each visit .
we investigated whether there was any advantage in shifting each individual subexposure , rather than combining them first by groups of three before shifting and coadding , but this does not appear to significantly improve the final coadded image .
subtraction of the psf from the wings of the a bright star is very sensitive to small mismatches in the target and psf star s spectral energy distribution , as well as to small changes in telescope focus and breathing ( i.e. , changes in image quality caused by flexure of _ hst _ and stis optical elements ) .
we will use two separate techniques to subtract @xmath0 eri s psf .
roll deconvolution techniques use the target as its own psf star , by taking back - to - back observations at different orientations .
direct subtraction of the @xmath0 eri and @xmath8 eri observations will also be done .
a comparison of different psf subtraction techniques for stis coronagraphic observations was done by grady et al .
( 2003 ) , and much of their discussion will also be relevant here . the goal of the roll deconvolution is to use images taken at different orientations to separate the real sky image from the psf of the bright nearby star .
this technique eliminates any problem with mismatches in the shape of the psf due to differences in the spectral energy distributions , but it has the disadvantage that any circularly symmetric features or arc - like structures larger than the change in roll angle will be included with the psf rather than as part of the sky image .
unfortunately these are exactly the type of structures most likely in a circumstellar debris disk . to obtain a first approximation to the psf of @xmath0 eri , we combined the four aligned and coadded @xmath0 eri crj images , rejecting points that were high or low by more than three sigma from the median value for that pixel location by using the iraf imcombine routine with the `` ccdclip '' algorithm .
as the images were not yet rotated to align them on the sky , this clips out real objects as if they were cosmic rays .
this trial psf was then subtracted from each of the original images .
these subtracted images were each rotated about the position of @xmath0 eri to align the positions on the sky and then combined by taking a straight average of the values at each pixel , but with locations near the main diffraction spikes or the obstructed edges of the 50ccd aperture masked out of the average . we found that masking out a rather wide strip ( about 3.4 " in the diagonal direction ) around the main diffraction spikes gave the best results .
a number of faint objects are clearly visible in the field of view .
a mask was created for each unrotated image that identifies pixels that are affected by real objects on the sky.a final psf for @xmath0 eri was then made in the same way as the initial psf image , but with these sky objects masked out before taking the average .
the new psf was then subtracted from the shifted crj images , and the subtracted files were again rotated into alignment and averaged after masking out the diffraction spikes and aperture edges . to summarize , we masked out the sky objects when averaging the unrotated and unsubtracted images to create the psf , and then masked out the diffraction spikes and aperture edges when averaging the rotated and psf - subtracted images to create the image of the sky . in principle
this procedure could be iterated to refine the separation between the sky and psf images , but we will simply use the second version of the sky and psf images produced by this procedure .
the procedure used to produce the @xmath8 eri psf was similar to that used to produce the @xmath0 eri psf .
it was necessary to mask out a generous region around the brightest background star in each of the two @xmath8 eri images to avoid introducing obvious artifacts in the subtraction . before subtracting the @xmath8 eri psf from the observations of @xmath0 eri ,
it is necessary to know both @xmath23 , the relative normalization of the two psfs , as well as @xmath24 , the sky background level .
the psf subtracted images will then be calculated as @xmath25 . to calculate the relative normalization we need to consider the spectral energy distributions ( seds ) of the two stars .
we will approximate these seds by using the broad - band photometry ( table [ phottab ] ) taken from the lausanne online database ( mermilliod , mermilliod , & hauck 1997 ) .
longward of the @xmath26 band , the two stars have very similar spectral distributions , but @xmath0 eri is a bit bluer at shorter wavelengths .
we supplement this photometry with iue data for shorter wavelengths to provide a rough spectral energy distribution for each star ( table [ fluxtab2 ] ) , which is then used with synphot to predict stis ccd imaging magnitudes in the 50ccd and f25nd3 filters .
the predicted f25nd3 magnitudes are about 8% brighter than observed , but the predicted ratio of the two stars f25nd3 count rates matches the observed ratio to within 0.2% .
this gives us a fair measure of confidence that the predicted flux ratio for stis ccd imaging with the unfiltered 50ccd aperture will also be correct .
we have also constructed tiny tim ( krist 1993 ; 1995 ) models using the above spectral energy distributions as input .
tiny tim only calculates on - axis psfs for stis , and does not calculate the psf beyond 4.5 `` radius .
the ratio of the tiny tim psfs does show that , when normalized to the same total flux , the slightly bluer @xmath0 eri sed results in a psf at 4.5 '' that is about 0.4% lower than that of @xmath8 eri .
taking the predicted 50ccd ratio of 0.837 , and then assuming the same difference between the predicted and observed ratios as was found for the direct f25nd3 images ( @xmath27 ) , would give an expected ratio of 0.839 .
extrapolating the radial color difference found in the tiny tim models , we estimate an additional correction of about @xmath28 near 17 " , yields an estimated normalization ratio of 0.834 .
if this ratio is adopted , it is also necessary to assume a mean sky background of @xmath290.077 e@xmath30/pixel / s if the farthest parts of the two psfs are to match .
this is comparable to the expected sky brightness .
the results of this subtraction are shown in figure [ delsub834 ] .
there appears to be substantial excess light near @xmath0 eri , amounting to about 1% of the unsubtracted psf at the same radius , and a number of concentric rings are visible . to illustrate this diffuse structure we took the difference between the @xmath0 and @xmath8 eri psfs ( with the sky objects removed ) , and measure the mean surface brightness as a function of radius ( figures [ halo_cnts ] and [ halo_mags ] ) .
we can minimize the central halo around @xmath0 eri by increasing the normalization constant to about 0.842 ( see fig .
[ delsub842 ] ) , but the concentric rings remain , with a mean surface brightness of 25.5 to 25 stmag / arcsec@xmath31 .
if we assume the rings extend all the way around the star , this flatter normalization corresponds to a total integrated stmag of 17.4 for the rings .
this structure seems suspiciously symmetrical to be consistent with the observed clumpiness in the sub - mm observations , but the radial distribution of the flux , with a broad hump between 15 and 30 `` , is roughly consistent with the radial dependence of the 850@xmath6 m flux as shown in figure 2 of greaves et al .
the rings are broader than psf subtraction artifacts previously seen in stis coronagraphic imaging ( grady et al .
2003 ) , however few images expose the far psf wings this deeply .
one example of acs coronagraphic observations of arcturus did show that breathing induced changes between visits can lead to a ring in the subtracted image with a radius of 13 '' and a surface brightness 20.5 mag arcsec@xmath31 fainter than the star ( see figure 5.12 of pavlovsky et al .
it may be that the rings in our stis data are a similar artifact .
however , we will not completely ignore the possibility that the rings might be real , and will consider whether they could be plausibly modeled by dust in the @xmath0 eri system .
figure [ conmap ] shows the roll subtracted data from part of figure [ rollsub ] , but at two times the scale of that figure and after smoothing with a 5x5 boxcar filter to suppress the small scale noise . overlaid on this figure is the sub - mm contour map of greaves et al .
( 1998 ) , positioned assuming that the sub - mm emission shares the proper motion of @xmath0 eri .
the inner contour level of the brightest sub - mm peak is about 5 " in diameter .
there are a substantial number of faint objects detected in the region of the sub - mm emission , but the density of such objects is not appreciably greater than elsewhere in the image . there is no apparent correlation between these objects and the sub - mm flux , and most or all of these objects are probably background galaxies or stars . note that the orbital period for material in the dust ring is about 500 years , and the expected orbital motion is only about 0.2 " per year .
figure [ conmap98 ] shows the same data , but leaves the sub - mm contours at where they were actually observed in 1997 - 1998 ; i.e. , they are not corrected for the @xmath294 `` that @xmath0 eri moved during the intervening years . in this figure ,
the brightest sub - mm clump is just outside the 2 '' sub - mm pointing uncertainty of what appears to be a background galaxy with stmag=24.7 ( this brightness includes all the clumps in this extended object ) .
figure [ conmapdz ] compares the proper motion corrected sub - mm contours with a 5x5 boxcar smoothing of the image that results when subtracting the @xmath8 eri psf from the @xmath0 eri data .
one of the bright rings is roughly at the location of the sub - mm emission , but there is no correlation with the bright knots .
nearly sixty distinct objects are visible in the field of view , including about half a dozen within 20 `` of @xmath0 eri .
most of these objects appear to be slightly extended , and many have complex morphologies .
there is no apparent increase in their density at close distances to @xmath0 eri , and most are probably background galaxies .
two relatively bright point sources are visible at distances of 51.5 '' ( stmag=19.6 ) and near 54 " ( stmag=22.2 ) . a list of detected objects , their locations , and their magnitudes is given in the appendix ( tab .
[ objecttab ] ) . for comparison , we estimate , from the magnitudes and colors given by chabrier et al .
( 2000 ) for models of 1 gyr old brown dwarfs with dusty atmospheres , that a @xmath32 brown dwarf in the @xmath0 eri system would have a stis 50ccd stmag of about 19 , and a @xmath33 brown dwarf an stmag of about 27.5 .
we measured the pixel - to - pixel rms variation in apparently blank regions of the final composite roll - subtracted image as a function of radius from epsilon eri .
the point - to - point rms noise varies smoothly and can be fit as @xmath34 e-/sec , where r is the distance in arc - sec from epsilon eri .
the measured noise is close to that predicted by a simple noise model based on the read noise and the total number of counts in the psf as a function of radius . within 30 `` of the star ,
the noise in the psf subtracted image is dominated by the poisson noise of the star s psf wings ( although the psf has been subtracted , its poisson noise still affects the subtracted image ) . since surface brightness of the stellar psf increases steeply closer to the star , our limiting magnitudes become significantly brighter . at distances greater than 30 '' , the accumulated read noise dominates the total noise .
the measured and predicted noise are listed in table [ noisetab ] .
this noise will be somewhat higher in regions which were not observed at all four orients .
note that a total count rate of 1 e@xmath35/s corresponds to a 50ccd stmag of 26.405 , and that the magnitude of @xmath0 eri in these same units is 4.02 .
for extremely red objects , as little as 10% of the total point source counts will fall into the central pixel of 50ccd images . for an unambiguous point source detection
, this central pixel should be at least 5@xmath36 above the rms noise .
these assumptions lead to the detectability limits given in the final column of table [ noisetab ] . for extended objects ,
the entire flux should be considered , but a 5 sigma detection will still be required .
if noise in different pixels was uncorrelated , then the upper limit for detecting a fixed value of the surface brightness would decrease as the square root of the area .
for very large areas , however , any low - spatial frequency noise sources would limit the practically achievable faint limit .
we empirically tested the real faint limit for extended sources by masking out discrete sky objects and then comparing the measured background counts in a number of separate regions at various distances from @xmath0 eri in the roll subtracted image . taking the standard deviation of the mean fluxes in boxes of a given size at a given distance as the @xmath37sigma error in the background measurement together with the poisson noise from the potential source , we derive the @xmath38 detection limits given in table [ noisetabe ] . for boxes @xmath39``@xmath40 '' , we found about the variance expected from scaling the point - to - point fluctuations by @xmath41 , but for larger areas , the fluctuations were bigger than would be expected from poisson statistics .
for example , when averaging over a @xmath42 box at 20 " , the measured pixel - to - pixel variance would imply a @xmath38 detection limit of about 27.2 stmag / arcsec@xmath31 about 1 mag fainter than the directly measured limit .
this difference could be due in part to real sources on the sky , but the lack of a clear correlation of these fluctuations with the sub - mm map or distance from the star , requires us to treat such fluctuations as noise . note that at 20 `` from the star , the shift in position from a 10 degree roll change is about 3.5 '' .
circular or arc - like structures much larger than the roll separation will tend to be washed out in the roll - subtracted image .
unfortunately , long narrow arc - like structures oriented in the tangential direction are precisely the kind of structures that are most likely in circumstellar debris disks .
this limits the utility of the roll subtracted image for detecting circumstellar structures much larger than a few arc - seconds in extent .
collisional fragmentation in a sufficiently dense debris disk will lead to a collisional cascade that generates a wide spectrum of particle sizes .
wyatt et al . ( 1999 ) , discussed extensively the processes that influence the resulting size distribution .
theoretical arguments predict that , when collisions produce similar fragmentation at all size scales , the cascade leads to a distribution of particle sizes @xmath43 ( dohnanyi 1969 ; tanaka , inaba , & nakazawa 1996 ) , with @xmath44 .
such a distribution has most of the mass concentrated in the biggest objects , but with most of the surface area being dominated by particles near the lower end of the size distribution .
when the emissivity is constant for all sizes , then the observable characteristics will then be dominated by the smallest particles .
the distribution may , however , be considerably flattened if the orbital evolution of the smaller particles affects them rapidly enough .
the lower cutoff to the distribution is set either by the size at which the time scale for poynting - robertson driven orbital decay is smaller than the lifetime of these particles against collisional creation , or when the smallest grains can be blown out of the system by radiation pressure
. a small particle broken off a large object in a circular orbit will be on an unbound hyperbolic orbit when the ratio of radiation to gravitational forces @xmath45 .
we can calculate @xmath46 using the same mie calculations and stellar spectrum used to calculate the dust spectrum ( see [ dustcalc ] ) .
the time for poynting - robertson orbital decay is proportional to the ratio of gravitation to radiative forces on a dust grain .
the time scale for poynting - robertson drag to change a circular orbit from an radius @xmath47 to @xmath48 is ( burns , lamy , & soter 1979 ) @xmath49 where @xmath47 and @xmath48 are given in au . for numerical comparisons
we will define the time needed for the orbit to decay by 20% of its initial radius as the time scale for poynting - robertson drag ; this is 36% of the time to decay to zero radius . at a distance of 60 au from @xmath0 eri , and assuming the age of the system to be about 0.5 1 gyr , the poynting - robertson effect would be expected to remove any grains smaller than @xmath29100@xmath6 m size unless the disk is dense enough that the small particles are still being generated in collisions . for the smallest particles in a debris disk , wyatt et al .
( 1999 ) estimated that the collisional time scale is of order @xmath50 , where @xmath51 is the orbital period and @xmath52 is the effective face on optical depth ( i.e. , the geometric filling factor ) .
they also estimated that particles will only be destroyed in collisions with particles larger than a factor of @xmath53^{1/3}$ ] times their own size ( about 12% at 60 au from @xmath0 eri ) , so we will define @xmath54 to be a function of the particle size by integrating only over the appropriate size range , and the time scale in years for fragmenting collisions is then : @xmath55\,yr.\ ] ] for a dust model to be self consistent , this time scale should be shorter than @xmath56 for even the smallest particles in the model .
we will see below that this condition is easily satisfied for the debris disk around @xmath0 eri .
given an assumed grain composition and the associated wavelength dependent optical constants , we performed standard mie theory calculations for spherical particles covering a wide range of radii @xmath57 using the code of wiscombe ( 1979 ; 1980 ) .
this gives , among other quantities , the standard scattering and absorption coefficients @xmath58 and @xmath59 .
( the emission coefficient at each wavelength @xmath60 for the dust calculations we adopt the flux from a standard solar abundance kurucz model atmosphere with @xmath61 and @xmath62=4.75 , normalized to a total stellar luminosity of 0.35@xmath63 .
the dust is assumed to be optically thin , and the temperature of each dust grain can be determined by the equilibrium between absorbed stellar radiation and thermal emission .
once the temperature of the grain is determined , then we can calculate the total light from a dust grain at each wavelength as the sum of the scattered starlight and the thermal emission .
we will assume that the dust grain is small enough that the thermal emission is isotropic .
the scattered light , however , will be highly anisotropic .
the angular phase function @xmath64 for this scattering can be easily derived from the mie calculation , and we normalize this function so that @xmath65 for isotropic scattering . then , when viewed at a scattering angle @xmath66 , the apparent flux of the dust grain will be @xmath67 this result will be used below to calculate the expected optical spectrum for a given dust distribution . the observable dust particles in circumstellar debris disks are presumed to be collisionally produced fragments of larger objects resembling those in the solar system s kuiper belt .
these larger objects were probably formed as very porous agglomerations of interstellar grains early in the history of the system .
for the composition and optical properties of these grains we will adopt the model of interstellar and circumstellar dust developed by li & greenberg ( 1997 ; 1998 ) , which assumes that the dust grains are a mixture of silicates , organic refractories , and voids , with some fraction of the voids possibly filled in by water ice .
the bruggemann mixing rule ( krgel 2003 ) is used to calculate the effective optical constants for the resulting mixtures from the optical constants of organic refractories and amorphous silicates given by li & greenberg ( 1997 ) , and that of vacuum .
wyatt & dent ( 2002 ) were able to fit the sub - mm and iras data for the debris ring around the a3v star fomalhaut by assuming solid ( non - porous ) grains consisting of 1/3 silicate and 2/3 organic refractory material by volume , and a grain size distribution close to the theoretically expected @xmath68 that extended down to the radiation blowout limit for that star . while they could not completely exclude models with some degree of porosity in the grains , they argued that the collisional fragmentation should have resulted in significant compaction of the grains despite the high porosity expected in the primordial parent bodies .
in contrast , li & greenberg ( 1997 ; 1998 ) , and li & lunine ( 2003ab ) favor models for the debris disks around hd 141569a , @xmath69 pictoris , and hr 4796a that assume highly porous grains , with vacuum fractions , @xmath70 to 0.9 . in some cases
these require a dust size distribution close to @xmath71 , significantly flatter than the theoretically expected @xmath72 .
li , lunine , & bendo ( 2003 ) have recently fit such a model to the available data for @xmath0 eri .
they find that a model assuming highly porous particles , and a rather flat size distribution , can provide an excellent fit to both the iras and sub - mm data .
they assumed that the same dust - size distribution function applies at all distances from the star , and only varied the total number density of grains as a function of radius to match the distribution of the observed 850@xmath6 m flux .
it is not clear whether or not this is realistic .
moro - martn s & malhotra s ( 2002 ; 2003 ) dynamical studies of dust produced in our own solar system s kuiper belt found that the size distribution function is expected to change substantially as a function of radius .
however , the li et al .
( 2003 ) model does give an excellent fit to both the total sub - mm and iras fluxes observed in the @xmath0 eri system .
this model assumes @xmath73 , lower and upper size limits to the distribution , @xmath74 m , and @xmath75 cm , a porosity , @xmath76 .
the solid portion of the grains is assumed to consist of an organics / silicate mix in a 58:42 ratio by volume .
the radial distribution of dust density is modeled as a gaussian centered at 55 au , with a fwhm of 30 au .
we repeated li et al s calculations for the flux from such a dust distribution , but also included the contribution of the scattered light , assuming a scattering angle of 90 degrees ( i.e. , a face on disk , as indicated by the morphology of the 850 @xmath9 m flux ) .
it is instructive to examine the radiative forces and time scales for this model of the dust distribution .
figure [ bradplot ] shows @xmath77 , the ratio of radiative to gravitational forces , for dust grains of the modeled composition in the @xmath0 eri system ( see also figure 3 of sheret et al 2003 ) . at large grain sizes , the low density of porous grains substantially increases the effects of radiation pressure relative to that on solid grains of the same composition and size . at small sizes
, however , the porous grains no longer effectively scatter radiation , and @xmath77 drops below that of solid grains , never becoming large enough ( @xmath78 ) for such grains to be efficiently blown out of the system .
in reality , it is likely that the grains become less porous as they are fragmented to very small sizes . in any case , the time scale for fragmenting collisions expected for the model of li et al ( 2003 ) is much shorter than @xmath79 at all sizes ( fig .
[ tplot ] ) , so small grains should be abundant , but predicting the detailed distribution of grain size and porosities at the lower end of the distribution will be difficult .
attempts made to fit solid grains models ( @xmath80 ) have not resulted in as good a fit as the li et al .
( 2003 ) model .
however , sheret , dent , & wyatt ( 2003 ) found that a simple model of a thin ring at 60 au with solid silicate / organic grains , @xmath81 m , @xmath82 m , and @xmath83 , fits the ir and sub - mm data with a reduced @xmath84 of 2.8 .
near 55 au from @xmath0 eri the typical @xmath85 m surface brightness observed by greaves et al . is about 0.02 mjy / arcsec@xmath31 . after subtracting the mean emission of the ring , the 850@xmath6 m flux in the brightest clump totals to about 2.6 mjy .
if the angular extent of the clump is comparable to the scuba beam - size , this amounts to a surface brightness enhancement of an a additional 0.015 mjy / arcsec@xmath31 . in figure
[ fig_model1 ] , we show , for a dust model with the parameters of li et al .
( 2003 ) , the calculated total spectral energy distribution for the whole dust cloud , and also for just the dust at 55 au . normalizing this model to an 850@xmath6 m surface brightness at 55 au of 0.02@xmath86mjy / arcsec@xmath31 , yields a predicted stis 50ccd surface brightness of 27.9 stmag / arcsec@xmath31 ( equivalent to @xmath87 cnts / pixel / sec ) .
if we instead assume the parameters of the sheret et al .
( 2003 ) solid grain model , we predict an optical 50 ccd surface brigtness of 24.5 stmag rather _ brighter _ than the rings in the directly subtracted image
. in the roll subtracted image , any tangential feature large than the roll change between images will show up in the psf rather than in the sky image , and so circular rings would be invisible
. the bright clump might be detectable if it were not too diffuse or too spread out in the tangential direction .
if the dust parameters in the clump are the same as in the rest of the ring , we would predict a total 50ccd optical brightness of 22.6 stmag . spread over a scuba beam area
, this would give a surface brightness of about 28 stmag / arcsec@xmath31 well below our most optimitic detection limits of @xmath88 stmag / arcsec@xmath31 .
the clump would have had to be concentrated within an area of no more than 1/4 of the scuba beam size before we would have expected to see it .
this assumes that the dust distribution in the clump is the same as that of the ring as a whole .
if the clump was created by resonant interactions of dust with a neptune - like planet in the @xmath0 eri system , smaller particles , which are strongly perturbed by radiation pressure , may be less likely to collect in the same resonances .
if , for the enhancement in the clump , we change the lower limit of the size distribution to be @xmath89 m , and again normalize to 0.015 mjy / arcsec@xmath31 at @xmath85 m , then the predicted surface brightness of the clump drops to about 29.8 stmag / arcsec@xmath31 ( @xmath90cnts / pixel / s ) .
such a model also substantially reduces the clump s contrast against the rest of the ring in the thermal ir .
the rings seen in the directly subtracted image ( @xmath0 eri@xmath91 eri ) , with a surface brightness of about 25 stmag / arcsec@xmath31 , are much brighter than the predictions of li et al .
( 2003 ) , but slightly fainter than predicted by the model of sheret et al ( 2003 ) .
if the rings are psf subtraction artifacts , they are bright enough to obscure the expected signal from the dust .
if real , they would imply a much larger abundance of small grains that scatter efficiently in the optical than does the model of li et al .
the lack of obvious counterparts in the optical ring corresponding to the sub - mm clumpiness might be explained by the very different dynamical behavior of the small grains . unfortunately available
modeling of the hst / stis psf at distances of 20 " is inadequate to provide a clear answer regarding the reality of the ring - like features seen after the subtraction of the two stars psfs .
both the models of li et al .
( 2003 ) and sheret et al .
( 2003 ) assume that all excess ir and sub - mm flux in the @xmath0 eri system is due to a single dust distribution that is well traced by the 850@xmath6 m emission .
if a substantial fraction of the ir - excess is instead due to an inner zodiacal cloud of particles that contributes little at 850@xmath6 m , then constraints on possible models are considerably relaxed , although the iras measurements will still provide an upper limit to the allowed ir flux from the 55 au ring .
for example , if we change the model of li et al . ( 2003 ) by simply allowing the upper limit of the size distribution to extend to @xmath92 m instead of @xmath93 m , and adjust the overall normalization to again match the observed @xmath85 m flux , then both the predicted iras band and optical fluxes drop by a factor of about 4 . if we instead assume li et al s parameters , but with solid rather than porous grains , the predicted iras fluxes drop by a factor of two to three , but the optical surface brightness _ increases _ by a factor of four .
however , without better information on the spatial distribution of the 10 to @xmath94 m flux it is difficult to choose among the different possible models .
such information will eventually be provided by sirtf images of the @xmath0 eri system . however , there is still unique information about the distribution of small grains that would be provided by direct detection of the optical scattered light that can not be obtained from even the most detailed observations of the thermal dust emission .
the simple models discussed here clearly have some limitations .
real dust does not consist of the perfectly smooth spheres assumed in mie theory , but will have considerable surface roughness .
for example , lisse et al .
( 1998 ) in a study of cometary dust found that considering the expected fractal structure of the dust could increase the optical scattering at 90 degrees by as much as a factor of three . also , the porosity is unlikely to be the same for all grain sizes
. even if the parent bodies are highly porous , at sufficiently small sizes the dust may either be significantly compacted by collisions , or will have broken up into smaller but more solid component particles . however , currently there are insufficient observational data for the @xmath0 eri dust ring to constrain the additional free parameters needed by such models .
this situation will improve substantially when _ spitzer infrared telescope _
images of this system become available .
our deep optical observations of the @xmath0 eri sub - mm ring have not provided clear evidence for detection of an optical counterpart .
the upper limits measured are consistent with existing models of the dust in the sub - mm ring , and provide some constraints on the nature and amount of the smallest dust grains .
our optical limits should provide tighter constraints once _ spitzer infrared telescope _
images are available for this system .
we found approximately 59 objects between 12.5 `` and 58 '' from @xmath0 eri , with brightnesses between 19.8 and 28 magnitude ( stis/50ccd stmag ) .
if any of the more compact of these objects were associated with the @xmath0 eri system , they would correspond to brown dwarfs of @xmath29 0.03 to 0.05 @xmath95 .
however , it is much more likely that the majority of these objects are background stars and galaxies unrelated to the @xmath0 eri system .
a second epoch hst observation of comparable depth would immediately determine whether any of these objects shares @xmath0 eri s 0.98 " /year proper motion .
support for proposal go-09037 was provided by nasa through grants from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
table [ objecttab ] contains a list of detected objects .
some of the listed objects close into the star may well be noise or psf subtraction artifacts , while other faint but real objects may still be omitted from this list .
for each object the j2000 coordinates at epoch 2002.071 , the distance from @xmath0 eri , the approximate size of the object , and the total brightness in stis 50ccd stmag units are listed . for compact objects ,
the size given is the fwhm from a moffat fit , while for more extended objects the dimensions given are an approximate estimate .
one stis ccd pixel corresponds to about 0.05071 " on the sky .
macintosh et al . ( 2003 ) , performed a @xmath96 band adaptive optics search for close companions around @xmath0 eri , and found 10 candidates , although none are proper motion companions to @xmath0 eri .
four of these objects lie in our field of view .
objects # 4 and # 6 correspond to extended galaxies and are noted in the table below .
their objects # 5 and # 9 have no optical counterparts .
this , and other images of this field are aligned with the stis observation o6eo03020 , with the + y direction aligned 20 degrees east of north .
the compass drawn in the upper left corner of this figure is 5 " on a side .
[ rollsub ] ] /s ) as a function of radius from @xmath0 eri , after subtracting the @xmath8 eri psf with a normalization factor of 0.834 ( see fig .
[ delsub834 ] ) .
if we change this normalization factor to 0.842 , ( dashed line ) then the overall gradient is minimized , but the rings remain , as shown in fig .
[ delsub842 ] [ halo_cnts ] ] in the @xmath0 eri system for the porous grains assumed by li et al.(2003 ) are plotted as a function of grain size ( dashed line ) . for comparison ,
@xmath77 for solid grains of the same composition is also plotted ( solid line ) .
[ bradplot ] ] ) and porous ( @xmath76 ) grains .
this time scale is always much larger than the time - scale for the fragmentation rate of grains ( @xmath97 ) we calculate for the parameters of the li et al .
( 2003 ) dust model.[tplot ] ] eri , including the contribution from scattered optical light .
the solid line shows this model added to the stellar sed .
the ( + ) show the observed iras and sub - mm flux observations , and the ( diamonds ) show the flux values after subtracting the star light .
the dashed line , shows the sed of the surface brightness of this model at 55 au ( surface brightness values given on right axis ) , with the scales shifted to overlap the curves at 850@xmath6 m .
the dash - dot line shows the sed for the parameters of the thin ring solid grain model of sheret et al .
also marked are the approximate surface brightness levels corresponding to stis 50ccd brightnesses of 25 and 27 magnitudes / arcsec@xmath31 .
these levels roughly correspond to the brightness of the rings seen after the direct psf subtraction , and the upper limit to the surface brightness determined for the roll subtracted psf image respectively .
[ fig_model1 ] ] ccll 1200 & @xmath98 & schtz et al 2004 & + 850 & @xmath99 & greaves et al . &
+ 450 & @xmath100 & greaves et al . & + 100 & @xmath101 & _ iras _ & photospheric flux subtracted + 60 & @xmath102 & _ iras _ & photospheric flux subtracted + 25 & @xmath103 & _ iras _ & photospheric flux subtracted + cccccc @xmath8 eri off1 & o6eo01020 & 03 43 16.95 & @xmath10409 45 52.09 & 10.0544 & + @xmath0 eri off1 & o6eo02020 & 03 32 57.75 & @xmath10409 27 35.20 & 10.0546 & 735 , 446 + @xmath0 eri off2 & o6eo03020 & 03 32 57.66 & @xmath10409 27 40.42 & 20.0547 & 729 , 515 + @xmath0 eri off3 & o6eo04020 & 03 32 57.50 & @xmath10409 27 45.31 & 30.0548 & 734 , 584 + @xmath0 eri off4 & o6eo05020 & 03 32 57.29 & @xmath10409 27 49.74 & 40.0550 & 752 , 651 + @xmath8 eri off4 & o6eo06020 & 03 43 16.49 & @xmath10409 46 06.64 & 40.0548 &
+ lccc 2700 & @xmath112 & @xmath113 & 1.335 + 3000 & @xmath114 & @xmath115 & 1.036 + 3646.235 & @xmath116 & @xmath117 & 0.950 + 4433.491 & @xmath118 & @xmath119 & 0.865 + 5492.883 & @xmath120 & @xmath121 & 0.832 + 6526.661 & @xmath122 & @xmath123 & 0.831 + 7891.114 & @xmath124 & @xmath125 & 0.833 + 12347.43 & @xmath126 & @xmath127 & 0.801 + 22094.22 & @xmath128 & @xmath129 & 0.791 + cccc 7 & 0.064 & 0.052 & 25.0 + 8 & 0.055 & 0.043 & 25.3 + 9 & 0.041 & 0.037 & 25.5 + 10 & 0.033 & 0.032 & 25.7 + 12 & 0.023 & 0.026 & 26.1 + 15 & 0.017 & 0.021 & 26.3 + 20 & 0.013 & 0.017 & 26.6 + 30 + & 0.012 & 0.015 & 26.7 + cccccc 15 & @xmath130 & @xmath131 & @xmath132 & @xmath133 & + 20 & @xmath134 & @xmath135 & @xmath136 & @xmath137 & + 25 & @xmath138 & @xmath139 & @xmath140 & @xmath141 & + 30 & @xmath142 & @xmath143 & @xmath144 & @xmath145 & + 35 & @xmath146 & @xmath147 & @xmath148 & @xmath149 & + ccrcccccr 3:32:56.4185 & @xmath1049:27:36.314 & 12.563 & 3x6 & 25.2 & + 3:32:56.5551 & @xmath1049:27:34.156 & 13.517 & 2.3 & 26.7 & + 3:32:56.4763 & @xmath1049:27:21.908 & 14.150 & 3.8 & 25.4 & + 3:32:56.6711 & @xmath1049:27:28.703 & 14.602 & 7 & 25.9 & + 3:32:56.8063 & @xmath1049:27:33.636 & 16.962 & 12 & 25.7 & double object + 3:32:56.8713 & @xmath1049:27:28.768 & 17.549 & 2.0 & 26.7 & + 3:32:56.8330 & @xmath1049:27:22.815 & 18.370 & 2.3 & 26.1 & + 3:32:55.9059 & @xmath1049:27:49.906 & 20.223 & 26 & 24.2 & fuzzy patch 1.3 `` diam .
+ 3:32:56.5430 & @xmath1049:27:46.244 & 20.642 & 3x7 & 25.9 & faint line + 3:32:55.4660 & @xmath1049:27:50.758 & 21.065 & 3.4 & 25.1 & + 3:32:55.7102 & @xmath1049:27:51.351 & 21.406 & 8x16 & 23.0 & oval + 3:32:56.1177 & @xmath1049:27:51.611 & 22.582 & 5x9 & 24.4 & oval + 3:32:57.1903 & @xmath1049:27:25.101 & 22.740 & 2.9 & 26.4 & + 3:32:56.5049 & @xmath1049:27:09.433 & 23.767 & 3.0 & 25.6 & + 3:32:57.2179 & @xmath1049:27:19.614 & 24.862 & 3.2 & 25.8 & + 3:32:57.3883 & @xmath1049:27:34.610 & 25.586 & 3.0 & 25.1 & extended flux + 3:32:57.1796 & @xmath1049:27:45.611 & 27.068 & 3x10 & 26.9 & faint line + 3:32:57.4928 & @xmath1049:27:19.772 & 28.554 & 4x6 & 23.8 & oval galaxy + 3:32:56.7705 & @xmath1049:27:54.110 & 28.991 & 2.3 & 25.4 & + 3:32:57.6676 & @xmath1049:27:29.800 & 29.281 & 3.4 & 26.2 & + 3:32:57.2929 & @xmath1049:27:11.811 & 29.844 & 3 & 26.4 & on edge of fov + 3:32:57.8475 & @xmath1049:27:34.308 & 32.240 & 4.4 & 24.5 & + 3:32:57.2044 & @xmath1049:27:55.359 & 33.897 & 3.7 & 26.2 & + 3:32:58.0335 & @xmath1049:27:28.253 & 34.728 & 6.1 & 26.1 & extended structure mag 24.8 + 3:32:58.0057 & @xmath1049:27:19.861 & 35.718 & 3.1 & 26.6 & + 3:32:58.0150 & @xmath1049:27:41.904 & 36.439 & 7.6 & 25.2 & + 3:32:58.1936 & @xmath1049:27:30.231 & 37.056 & 5x9 & 25.6 & extended structure + 3:32:57.2856 & @xmath1049:27:58.997 & 37.445 & 4 & 26.1 & + 3:32:58.1967 & @xmath1049:27:36.068 & 37.606 & 4 & 26.1 & + 3:32:58.2965 & @xmath1049:27:37.911 & 39.394 & 4x10 & 24.1 & long oval + 3:32:58.3564 & @xmath1049:27:31.950 & 39.514 & 1.5 & 25.6 & extended structure + 3:32:58.1802 & @xmath1049:27:46.132 & 40.259 & 4.0 & 26.1 & + 3:32:58.3193 & @xmath1049:27:17.338 & 40.889 & 4 & 27.8 & + 3:32:58.0451 & @xmath1049:27:51.423 & 40.946 & 5 & 25.5 & extended structure + 3:32:58.1279 & @xmath1049:27:49.574 & 41.079 & 3.2 & 23.2 & double irr galaxy + 3:32:58.5861 & @xmath1049:27:24.382 & 43.210 & 2.0 & 27.4 & + 3:32:56.9571 & @xmath1049:28:08.992 & 43.308 & 3.7 & 25.3 & + 3:32:58.6919 & @xmath1049:27:24.300 & 44.772 & 3.2 & 27.1 & extended structure mag 25.8 + 3:32:58.4244 & @xmath1049:27:49.437 & 44.919 & 3.8 & 25.7 & + 3:32:58.7317 & @xmath1049:27:24.345 & 45.350 & 2.5 & 26.2 & + 3:32:58.7489 & @xmath1049:27:25.229 & 45.504 & 5 & 26.1 & + 3:32:58.6853 & @xmath1049:27:19.098 & 45.620 & 11 & 25.9 & noise ? + 3:32:58.3840 & @xmath1049:27:53.969 & 46.549 & 3 & 25.7 & + 3:32:57.8873 & @xmath1049:28:04.667 & 47.568 & 2.4 & 26.0 & + 3:32:58.9032 & @xmath1049:27:31.625 & 47.574 & 2 & 26.1 & + 3:32:57.8779 & @xmath1049:28:07.866 & 49.856 & 5 & 26.0 & + 3:32:59.0790 & @xmath1049:27:37.894 & 50.771 & 5.2 & 25.4 & + 3:32:59.0779 & @xmath1049:27:38.375 & 50.833 & 3 & 27.1 & + 3:32:58.9754 & @xmath1049:27:46.790 & 51.449 & 20 & 23.7 & fuzzy patch @xmath291 '' diam .
+ 3:32:58.0864 & @xmath1049:28:07.301 & 51.501 & 1.8 & 19.8 & bright star + 3:32:58.8881 & @xmath1049:27:52.372 & 52.366 & 3.3 & 26.6 & + 3:32:59.3340 & @xmath1049:27:29.536 & 53.912 & 1.8 & 22.2 & star + 3:32:58.5123 & @xmath1049:28:04.604 & 54.265 & 3.3 & 25.7 & + 3:32:59.4291 & @xmath1049:27:34.188 & 55.481 & 10x30 & 22.6 & oval + 3:32:59.3587 & @xmath1049:27:16.223 & 55.970 & 4 & 28.3 & + 3:32:59.4461 & @xmath1049:27:42.934 & 57.069 & 5x10 & 25.6 & + 3:32:59.5610 & @xmath1049:27:30.327 & 57.267 & 4 & 26.0 & + 3:32:59.5556 & @xmath1049:27:37.599 & 57.698 & 3.5 & 23.5 & oval outer isophote + 3:32:59.4106 & @xmath1049:27:49.428 & 58.392 & 4.2 & 25.1 & | the stis / ccd camera on the _ hubble space telescope ( hst ) _ was used to take deep optical images near the k2v main - sequence star @xmath0 eridani in an attempt to find an optical counterpart of the dust ring previously imaged by sub - mm observations .
upper limits for the optical brightness of the dust ring are determined and discussed in the context of the scattered starlight expected from plausible dust models .
we find that , even if the dust is smoothly distributed in symmetrical rings , the optical surface brightness of the dust , as measured with the _
hst_/stis ccd clear aperture at 55 au from the star , can not be brighter than about 25 stmag/``@xmath1 .
this upper limit excludes some solid grain models for the dust ring that can fit the ir and sub - mm data .
magnitudes and positions for @xmath259 discrete objects between 12.5 '' to 58 " from @xmath0 eri are reported . most if not all of these objects are likely to be background stars and galaxies .
1based on observations made with the nasa / esa hubble space telescope , obtained at the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
these observations are associated with proposal go-09037 .
2also the space telescope science institute , and the institute for astrophysics and computational science at the catholic university of america .
3also eureka scientific , 2452 delmer street , suite 100 , oakland , ca 94602 - 3017 4member of space telescope imaging spectrograph investigation definition team |
the most looked - for particles on earth at the present time are axions @xcite and weakly interacting massive particles(wimps ) @xcite .
recently , hints on these particles have been reported from outer space sources .
the 10 year old integral data @xcite gives the 511 kev line .
the white dwarf cooling has been suggested by the emission of very weakly interacting light particles(vwlp ) ( mass less than @xmath7 ev ) with a very small coupling strength to electron ( @xmath0 ) @xcite . at present , this white dwarf bound can be considered just as an upper bound , but in this paper we adopt a bold assumption that their best fit corresponds to the existence of a vwlp .
more recently , the remarkable observation of high energy positron excess in the satellite pamela data @xcite hints heavy ( 100 gev 10 tev ) cold dark matter(cdm ) particles @xcite .
since the 511 kev line can be explained by the astrophysical origin , or by the particles of mass @xmath8 @xcite , or by an excited state almost degenerate to the wimp @xcite , in this paper we attempt to understand the latter two of these observations from a grand unification ( gut ) viewpoint of the flipped - su(5 ) @xcite .
the pamela group @xcite reports the high energy positron excess , above 10 gev up to 60 gev , at the level @xmath9 , but has not reported any noticeable excess of anti - protons @xcite .
the charactersistic of the data between 1060 gev is a slightly rising positron flux , which is inconsistent to the lightest supersymmetric particle(lsp ) in the minimal supersymmetric standard model(mssm ) with supersymmetric(susy ) particles at @xmath10 @xcite .
thus , the mssm with @xmath10 lightest neutralino dm component @xmath11 ( ) has to be modified if susy has to be kept for the gauge hierarchy solution .
the minimal extension needs just three two component fermions @xmath12 and @xmath13 @xcite , with the coupling @xmath14 if one tries to introduce more fields around tev scale , certainly the rising positron flux can be explained , but it is very difficult @xcite to explain nonobservation of antiproton flux . this suggests a different treatment of electron from quarks .
indeed , the flipped - su(5 ) gut 1 treats charged leptons differently from quarks @xcite such that the charged lepton is an 1 singlet , @xmath15 the specific form of the coupling ( [ eecn ] ) is related to the problems of the yukawa couplings . in this regard ,
we note the five old but fundamental problems on the masses of the standard model fermions .
firstly , there is a hierarchy of the top quark mass @xmath16 gev at the electroweak scale of @xmath17 gev , and the rest of the standard model particle masses are much smaller than the top quark mass , @xmath18 .
second , there is a sizable cabibbo mixing , especially between the first and the second family quarks .
third , the quark mass ratio is reversed in the first family compared to those of the second and the third family quarks , i.e. @xmath19 while @xmath20 and @xmath21 . even though it is not proper to define an inverted mass pattern for leptons
, it looks like that the electron mass is also an inverted pattern because @xmath22 .
fourth , there is a big hierarchy of singlet neutrino mass of @xmath23 gev for the seesaw mechanism @xcite and the electroweak scale of @xmath17 gev . finally , we point out that the sm neutrino mass pattern shows a tri - bi maximal mixing pattern @xcite . except the tri - bi
maximal mixing , in this paper we try to understand the remaining four problems in a flipped - su(5 ) gut . to have a very light axion implied by the white dwarf data , we introduce a peccei - quinn ( pq ) symmetry u(1)@xmath4 so that the symmetry we consider is extended to 1 times u(1)@xmath4 . to have the inverted mass pattern of the quark mass ratio in the first family , we assign different @xmath24 charges for the first family members and the second and the third family members .
but , to have a large cabibbo mixing , we require q@xmath25 quarks or q@xmath26 quarks mix fully , restricted only by the strength of the corresponding yukawa couplings but without any restriction from the symmetry arguments . in the flipped - su(5 ) , since both the quark singlet @xmath27 and the quark doublet @xmath28 appear in @xmath29 , viz .
( [ fliprep ] ) , it is easy to mix q@xmath26 quarks fully without restrictions from the symmetry arguments .
so , we assign the same symmetry charges for three members @xmath30 .
then , q@xmath26 quarks can be mixed fully as desired .
but the charges of @xmath31 and @xmath32 can be different for different @xmath33 to allow the inverted mass pattern of the quark mass ratio for the first family .
for the singlet neutrino mass problem , we need a large vev , above the axion window of @xmath34 gev .
if there is no other gauge symmetry , the pq symmetry may be broken by this large vev .
thus , we introduce another u(1 ) gauge symmetry which will be called a horizontal gauge symmetry u(1)@xmath35 , distinguishing the first family and the second and third families .
finally , to keep @xmath36 pair at the electroweak scale and to have proton stability and two dark matter components @xcite , we introduce the @xmath37 symmetry and @xmath5 symmetry ( the @xmath37-parity ) .
thus , the symmetry we consider is @xmath38 in this model , the renormalizable couplings are gauge couplings and the top quark yukawa coupling .
all the other yukawa couplings are non - renormalizable .
the model we introduce does not have the @xmath39 term of ref .
@xcite , where the @xmath39 term is the axion and qcd anomaly coupling , and the pq charges of higgs doublets vanish .
therefore , the very light axion we obtain is not a proper kim - shifman - vainstein - zakharov ( ksvz ) axion @xcite , or the dine - fischler - srednicki - zhitnitskii ( dfsz ) axion @xcite , but can be called a ` variant very light axion ' .
the electron
axion coupling is known to be small for the ksvz axion , as calculated in ref .
@xcite . in our case ,
the axion arises from the @xmath40 term of ref .
@xcite , where the @xmath40 term is the light quark mass term , through a non - renormalizable interaction , and the electron
axion coupling turns out to be large . since this ` very light variant axion ' is required to have a large electron coupling , it will be called a very light _ electrophilic
_ axion . in sec .
[ sec : wd ] , we briefly review the white dwarf bound @xcite . in sec .
[ sec : flip ] , we present a flipped - su(5 ) model , satisfying the criteria presented above . in sec .
[ sec : phenomenology ] , we discuss the susy phenomenology for the model presented in sec .
[ sec : flip ] .
[ sec : conclusion ] is a conclusion .
white dwarfs can give us hints about their last stage evolution .
white dwarfs , with the mass range less than @xmath41 , are the fossil stars after consuming all their nuclear fuel . here
the electron degeneracy pressure is the one supporting the gravitational collapse , and the surface temperature ranges down from ten times the solar surface temperature .
the temperature difference between the core and the surface is not greater than 100 k , and hence if the vwlp is produced at the core then its mass can not be much greater than 1 ev . since the white dwarf has a relatively simple structure and its evolution mechanism
is just based on the cooling , its luminosity function has been rather accurately predicted .
thus , if there is a novel energy loss mechanism by a vwlp like axion , it can affect the resulting luminosity function of the white dwarf which can be used to constrain the properties of that particle or give a clue for its existence .
the most strong bound on vwlp comes from the existence of the sharp cutoff at the faint part in the luminosity function ( @xmath42 ) versus the bolometric magnitude ( @xmath43 ) plot .
the sharp cutoff is due to the finite age of our galaxy .
the luminosity function is given by @xcite @xmath44 where @xmath45 is the initial mass function , @xmath46 is the birth rate of the progenitor star , @xmath43 is the bolometric magnitude , @xmath47 is the mass of the progenitor star and @xmath48 is the characteristic time scale of the white dwarf cooling .
although the luminosity function depends on the birth rate which is not yet exactly known , the bright part of the curve is insensitive to it up to the overall normalization .
thus , one can exploit the bright part to constrain the hypothetical cooling mechanism due to the vwlp .
in this regard , recently the possible energy loss rate has been computed @xcite .
if a very light boson with mass less than 1 ev is produced inside white dwarfs , its coupling to electron is estimated as @xmath49 any boson with this coupling strength to electron can explain the white dwarf evolution _ a la _ ref .
@xcite , using ref .
@xcite . since axions are well - motivated light pseudoscalar bosons , ref .
@xcite tries to interpret the coupling @xmath50 in terms of the dfsz model @xcite . in
the interior of the white dwarf , the axion bremstralung coming from the axion - electron yukawa coupling , which is significant in the dfsz model but absent in the hadronic axion model , is dominant over the primakoff process .
the energy loss rate per mass with the crystalization effect is given by @xcite , @xmath51 where @xmath52 is the correction factor due to the crystalization effect and @xmath53 is the yukawa coupling between the electron and the axion .
in the dfsz model , it corresponds to @xmath54 with the domain wall number 6 . from this study , isern et al .
@xcite point out that the small deviations of the white dwarf luminosity function from the standard cooling model can constrain the axion mass @xmath55 mev , and even presented the best fit of the luminocity function with the axion emission for @xmath56 mev . in the dfsz model
, this white dwarf bound on the axion mass corresponds to the axion decay constant @xmath57 gev and the best fit corresponds to @xmath58 gev .
below , we try to construct a flipped - su(5 ) to accomodate the best fit of ref .
@xcite .
the electron axion coupling can be written as @xmath59 , where @xmath60 and @xmath61 is the pq charge of @xmath62 . since the axion mass is given by @xmath63 , the best fit of isern et al .
@xcite in the dfsz model interpreted as the electron axion coupling in any axion model , @xmath64 if the domain wall number is large and the pq charge of electron is 1 , then we need to have a smaller @xmath65 than @xmath66 gev since @xmath67 is required to be @xmath68 . generally , the domain wall number is counted as @xmath69 for any quark carrying the unit pq charge . therefore ,
if only one dirac quark carries the pq charge 1 so that each chirality contributes one unit of the pq charge , then the domain wall number is 1 . in this way
@xcite , the dfsz model gives @xmath70 since there are six quarks , and the electron
axion coupling is @xmath71 with @xmath72 gev .
there is an interesting case that only one weyl quark carries the pq charge 1 . in this case ,
our formula for the axion
electron coupling must use @xmath73.).[footdw ] ] in such a model eq .
( [ aewdcoupling ] ) implies @xmath74 gev .
below we concentrate on the latter possibility in a flipped su(5 ) model , obtaining eq .
( [ aewdcoupling ] ) for the very light _ electrophilic _ axion .
this moves @xmath65 to the supergravity favored scale @xcite even with the large electron axion coupling envisioned in ref .
so(10 ) gut groups can not realize our idea of ref .
. the so(10 ) gut with the su(5 ) gut assignment @xcite does not introduce our su(5 ) singlet @xmath6 and @xmath75 fields since the electromagnetic charge operator @xmath76 is fully embedded in su(5)@xmath77 .
the flipped - su(5 ) assignment of hypercharge in the so(10 ) subgroup @xcite , namely @xmath78 , allows the su(5 ) singlet ( in @xmath79 ) carrying the electromagnetic charge , and @xmath80 does not belong to @xmath81 or @xmath82 .
supersymmetric ( susy ) generalization of the flipped - su(5 ) chain from so(10 ) , including the vacuum expectation value of higgs field not belonging to the origin of the weight diagram to reduce the rank of the gauge group , has been discussed in @xcite .
a simple choice of such symmetry breaking with @xmath83 and @xmath84 from string construction was shown to be possible in the fermionic construction @xcite and in the orbifold compactification @xcite .
also , in the susy flipped - su(5 ) , the doublet - triplet splitting problem is solved elegantly as shown in the fermionic construction @xcite and in the orbifold compactification @xcite . here
, we do not consider the so(10 ) breaking chain to an 1 gauge group .
the 1 gauge group is considered as an ultraviolet completion gauge group , possibly arising from a compactification of string theory . as in ref .
@xcite , we use the right - handed chiral fields so that the electron singlet is represented as @xmath80 instead of @xmath85 and the quark doublet is represented as @xmath86 instead of @xmath87 .
this right - handed chiral field notation is simpler to discuss the interaction @xmath88 which we try to introduce for the pamela data .
we use the following right - handed notation for representations of the flipped - su(5 ) .
the electroweak hypercharge is @xmath89 where the sm singlet generator embedded in su(5 ) is @xmath90 the representations ( [ fliprep ] ) are consistent with this definition .
the matter field @xmath29 contains the quarks doublet @xmath91 , the quark singlet @xmath92 and the neutral lepton singlet @xmath93 .
@xmath82 contains the quark singlet @xmath94 and the anti - lepton doublet @xmath95 .
@xmath96 is the charged lepton of the unit charge 1 .
therefore , it is useful to remember that singlets @xmath27 , @xmath97 and @xmath80 are members of @xmath29 , @xmath82 and @xmath96 , respectively . the singlet neutral direction ( @xmath98 and @xmath99 directions ) of @xmath29 and @xmath100 can be assigned to gut scale vacuum expectation values(vevs ) , breaking su(5)@xmath101u(1)@xmath102 down to su(3)@xmath101 su(2)@xmath101u(1)@xmath103 .
for the spontaneous symmetry breaking of the flipped su(5 ) , thus , we need a vectorlike pair , @xmath83 and @xmath104 , which contain the sm singlet components , @xmath98 and @xmath105 .
the following superpotential , @xmath106 allow @xmath83 and @xmath84 to develop vevs along the direction of @xmath98 and @xmath105 . here
the superfield @xmath107 is a singlet of the flipped su(5 ) .
since the d flat direction is along @xmath108 , thus , at the susy minimum we have @xmath109 the phase modes of @xmath110 ( su(2 ) doublet ) and @xmath111 ( su(2 ) doublet ) in @xmath83 and @xmath84 correspond to the nambu - goldstone bosons and the real modes of @xmath110 and @xmath111 become heavy as the higgs boson becomes heavy in the higgs mechanism in the standard model .
the nambu - goldstone bosons are absorbed by the heavy gauge sectors when the flipped - su(5 ) is broken .
however , @xmath112 ( su(2 ) singlet ) and @xmath113 ( su(2 ) singlet ) in @xmath83 and @xmath84 need to be made somehow heavy . the quantum numbers of @xmath107 , @xmath83 , @xmath84 , and some other superfields needed later are displayed in table [ tab : vectorlike ] .
.vectorlike fields including the higgs fields .
@xmath107 and @xmath114 are the flipped - su(5 ) singlet fields .
[ cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ] here @xmath115 and @xmath116 are at the electroweak scale , @xmath117 is a free parameter , presumably around 100 gev , and @xmath118 is of order @xmath119 as discussed in eq .
( [ umass ] ) .
then , the physical right - handed up quark and the heavy quark are given by @xmath120 where @xmath121 and @xmath122 represent the @xmath123 quark in @xmath122 .
then , the coupling @xmath124 is suppressed by @xmath118 , and the antiproton excess is not expected in the flipped - su(5 ) model . the model presented in ref .
@xcite has three two component fermions . among these @xmath125
lepton @xmath6 can be produced at the lhc machine .
the most distinguishable feature of the model presented here is leptonic ( especially ` electronic ' ) property of the dark matter component .
@xmath126 and @xmath127 can be produced by proton
proton collision and eventually they decay to an electron positron pair and missing energy by the interaction ( [ eecn ] ) .
then , we compare it with the expected high energy @xmath128 spectrum from the lhc machine .
this signal distinguishes the model of ref .
@xcite from most other popular models .
we will calculate the high energy @xmath128 spectrum in the quark - antiquark ( @xmath129 ) center of momentum ( cm ) frame . here
, @xmath130 stands for @xmath131 etc .
then , the result is convoluted to the @xmath130 and @xmath132 structure functions in the proton . for the @xmath129 annihilation ,
the cm frame kinematics of the @xmath129 pair is useful , @xmath133 the @xmath129 parton contribution in the proton - proton collision is given by @xmath134 here @xmath135 and @xmath136 are parton distribution function(pdf ) of @xmath130 and @xmath137 with @xmath138 and @xmath139 , repectively .
we use mrst99 c++ code for pdf@xcite .
sum is for all quarks except for @xmath140 and is neglected from now on .
@xmath141 is relevent energy scale of pdf .
@xmath142 with @xmath14314 tev in the designed lhc energy .
@xmath144 is obtained by cross section at the parton level , and is given by @xmath145 [ ( \varepsilon_l^e)^2+(\varepsilon_r^e)^2],\ ] ] where @xmath146 for up - type quarks @xmath147 for down type quarks , @xmath148 are electromagnetic charge of up and electron , and @xmath149 , @xmath150 for up - type quarks , @xmath151 , @xmath152 for down type quarks,@xmath153 and @xmath154 are @xmath107 boson couplings of quarks , and electron , respectively .
@xmath155 is the weak mixing angle @xcite .
@xmath156 is given by @xmath157 where @xmath158 is fermi constant , @xmath159 and @xmath160 are @xmath107 boson mass and decay rate .
( [ cross : pp_ee ] ) can be rewritten as @xmath161 where @xmath162 which is cross section between @xmath163 and @xmath164 . here
@xmath165 .
similarly , @xmath166 is obtained by simple calculations , which is given by @xmath167 where @xmath168 .
@xmath169 is given by @xmath170 [ ( \varepsilon_l^e)^2+(\varepsilon_r^e)^2],\ ] ] where @xmath171 and @xmath172 .
the production cross section of @xmath173 is approximately a half of that of electron - positron pair at weak scale .
cross section is shown in fig .
[ fig : eeprod ] .
0.5 cm however , since @xmath6 decays inside the beam line , electron - positron excess appears at lower energy scale .
therefore we can detect electron - positron pairs in collider .
cross sectrion can be rewritten in elctron - positron transverse invariant mass @xmath174 where @xmath175 , @xmath176 and @xmath177 are transverse energy , transverse momentum and rapidity difference , respectively .
it is given by @xmath178 where @xmath179 \end{split}\ ] ] also , these events are accompanied by large missing energy and this physics is very similar to slepton pair production of odinary mssm .
this is shown in fig .
[ fig : eefromee ] .
the main difference of our calculation from the ordinary mssm result is the produced @xmath6 s totaly decay to electrons but not to the other leptons .
due to this feature , we can use transverse momentum variable , @xmath180 to determine the mass of @xmath6 @xcite . note that the phenomenology of @xmath181 production is very similar to that of the selectron
antiselectron production because in both cases the heavy particle decays to @xmath182 plus missing energy .
so , it is crucial to distinguish them .
but , the only way to distinguish them is to measure their spin at the lhc .
below , we concentrate on the total cross section but not the angular distribution .
in fact , there are many analyses for measuring the particle spin at the lhc machine @xcite . since our @xmath6
particle is the fermion , their polarization property looks the same as the kk fermion in the ued models , and it may be difficult to distinguish our fermion from the ued fermions .
but we can expect to distinguish @xmath6 from the selectron .
for the same @xmath6 and selectron mass , their kinematics are similar but angular distribution of electrons might be quite different in the two cases .
if both @xmath181 and the selectron antiselectron pair are produced , we can observe two @xmath180s which might be spectacular .
the recent pamela satellite data from the galactic source suggested the positron excess @xcite but no anti - proton excess @xcite . regarding these observations , two dark matter components , the neutralino @xmath11 and the singlet fermion @xmath183 ,
were suggested in ref .
the needed coupling has been given by eq .
( [ eecn ] ) . here , we studied the flipped - su(5 ) where charged su(2 ) singlet leptons such as @xmath184 and @xmath13 of eq .
( [ eecn ] ) can be its allowed representations , but a similar coupling for an anti - proton excess is not allowed because quarks must be embedded in su(5 ) non - singlet representations .
in addition , there exists an interesting suggestion of the possible existence of vwlp with the coupling strength ( [ aewdcoupling ] ) from the study of white dwarf cooling @xcite . in this paper
, we tried to interpret this vwlp as an _
axion with the axion decay constant @xmath74 gev . in the dfsz model
, the corresponding @xmath65 would be @xmath185 gev which would be barely consistent with the sn1987a bound .
this kind of _ electrophilic _ axion can arise in ` variant very light axion ' models where different families possess different pq quantum numbers @xmath24 .
we argued that the inverted mass pattern of the first family from that of the second and third families is the logic for assigning different pq quantum numbers for different family members .
allowing an effective domain wall number of @xmath69 [ see footnote [ footdw ] .
] , we constructed an _
the symmetries we introduced in ( [ symmetry ] ) is bigger than 1@xmath101u(1)@xmath4 , to allow a gut scale vev with an intermediate scale @xmath65 and r - parity . with these symmetries , we explained the mass hierarchy between the top and bottom quarks . in this model , the gauge couplings and the top quark yukawa coupling are the only renormalizable couplings .
this model needs a charged su(2 ) singlet lepton @xmath6 at the electroweak scale .
so , the existence of @xmath6 is absolutely needed for this suggestion of the positron excess to work . here
, the discovery potential of @xmath6 , at the lhc experiments by observing the electron and positron spectrum , is presented .
we calculated the @xmath128 spectrum at the center of momentum frame of the quark and anti - quark pair and compared it with the expected background .
we pointed out that the su(2 ) singlet @xmath6 with the coupling ( [ eecn ] ) is distinguishable from other possibilities we can imagine .
therefore , the discovery of @xmath6 at the lhc experiments will indirectly confirm the two dark matter component hypothesis of @xcite . or it can be ruled out from the lhc experiments by the absence of the @xmath128 excess above the background .
# 1#2#3phys .
* # 1 * ( # 3 ) # 2 # 1#2#3rev .
# 1 * ( # 3 ) # 2 # 1#2#3annu .
nucl . part .
* b#1 * ( # 3 ) # 2 # 1#2#3nucl .
* b#1 * ( # 3 ) # 2 # 1#2#3phys .
* b#1 * ( # 3 ) # 2 # 1#2#3phys .
* d#1 * ( # 3 ) # 2 # 1#2#3phys .
# 1 * ( # 3 ) # 2 # 1#2#3jhep * # 1 * ( # 3 ) # 2 # 1#2#3jcap * # 1 * ( # 3 ) # 2 # 1#2#3z .
* b#1 * ( # 3 ) # 2 # 1#2#3euro .
j. * b#1 * , # 2 ( # 3 ) # 1#2#3int .
j. mod .
. * b#1 * ( # 3 ) # 2 # 1#2#3mod .
* b#1 * ( # 3 ) # 2 # 1#2#3astrophys .
j. * b#1 * ( # 3 ) # 2 # 1#2#3nature * # 1 * ( # 3 ) # 2 # 1#2#3sov .
j. nucl .
phys . * # 1 * ( # 3 ) # 2 # 1#2#3astropart .
b#1 * ( # 3 ) # 2 for reviews , see , g. jungman , m. kamionkowski and k. griest , [ arxiv : hep - ph/9506380 ] ; l. bergstrm , rep .
* 63 * ( 2000 ) 793 [ arxiv : hep - ph/0002126 ] ; g. bertone , d. hooper and j. silk , [ arxiv : hep - ph/0404175 ] .
j. knodlseder _ et al .
_ , astron .
astrophys .
* 411 * , l457 ( 2003 ) [ arxiv : astro - ph/0309442 ] ; p.
jean _ et al . _ , astron .
astrophys .
* 407 * , l55 ( 2003 ) [ arxiv : astro - ph/0309484 ] ; j. knodlseder _ et al . _ ,
astrophys . * 441 * , 513 ( 2005 ) [ arxiv : astro - ph/0506026 ] ; g. weidenspointner _
et al._,astron .
astrophys .
* 450 * , 1013 ( 2006 ) [ arxiv : astro - ph/0601673 ] ; g. weidenspointner _
et al . _ , arxiv : astro - ph/0702621 ; p.
jean _ et al . _ ,
astrophys .
* 445 * ( 2006 ) 579 [ arxiv : astro - ph/0509298 ] . c. boehm , p. fayet and j. silk , [ arxiv : hep - ph/0311143 ] ; c. picciotto and m. pospelov , [ hep - ph/0402178 ] ; d. hooper and l. t. wang , [ arxiv : hep - ph/0402220 ] ; c. boehm and p. fayet , [ arxiv : hep - ph/0305261 ] ; j. f. beacom , n. f. bell and g. bertone , [ arxiv : astro - ph/0409403 ] ; j. m. frere , f. s. ling , l. lopez honorez , e. nezri , q. swillens and g. vertongen , phys . rev .
d * 75 * , 085017 ( 2007 ) [ arxiv : hep - ph/0610240 ] ; e. j. chun and h. b. kim , [ hep - ph/0607076 ] ; j .- h .
huh , j. e. kim , j .- c .
park , and s .- c .
park , [ arxiv:0711.3528 ] ; m. pospelov , a. ritz and m. b. voloshin , arxiv:0711.4866 [ hep - ph ] . | recently , there are two hints arising from physics beyond the standard model .
one is a possible energy loss mechanism due to emission of very weakly interacting light particles from white dwarf stars , with a coupling strength @xmath0 , and another is the high energy positrons observed by the pamela satellite experiment .
we construct a supersymmetric flipped - su(5 ) model , 1 with appropriate additional symmetries , [ u(1)@xmath1_{\rm gauge}\times$][u(1)@xmath2u(1)@xmath3_{\rm global}\times z_2 $ ] , such that these are explained by a very light _ electrophilic _
axion of mass 0.5 mev from the spontaneously broken u(1)@xmath4 and two component cold dark matters from @xmath5 parity .
we show that in the flipped - su(5 ) there exists a basic mechanism for allowing excess positrons through the charged su(5 ) singlet leptons , but not allowing anti - proton excess due to the absence of the su(5 ) singlet quarks .
we show the discovery potential of the charged su(5 ) singlet @xmath6 at the lhc experiments by observing the electron and positron spectrum . with these symmetries , we also comment on the mass hierarchy between the top and bottom quarks . |
throughout this paper we work on a 4-manifold @xmath0 equipped with prescribed lorentzian metric @xmath1 .
the accepted mathematical model for a neutrino field is the following linear partial differential equation on @xmath0 know as the _ weyl equation _ : @xmath2 the corresponding lagrangian is @xmath3 here @xmath4 , @xmath5 , are pauli matrices , @xmath6 is the unknown spinor field , and @xmath7 is the covariant derivative with respect to the levi - civita connection : @xmath8 where @xmath9 are christoffel symbols uniquely determined by the metric .
the purpose of our paper is to give an alternative representation of the weyl equation ( [ weyl s equation ] ) and the weyl lagrangian ( [ weyl s lagrangian ] ) . to this end , we follow @xcite in introducing instead of the spinor field a different unknown the so - called _
a coframe is a quartet of real covector fields @xmath10 , @xmath11 , satisfying the constraint @xmath12 where @xmath13 . in other words , the coframe is a field of orthonormal bases with orthonormality understood in the lorentzian sense .
we define an affine connection and corresponding covariant derivative @xmath14 from the conditions @xmath15 ) read @xmath16 giving a system of linear algebraic equations for the unknown connection coefficients @xmath17 .
the connection defined by the system of equations ( [ defining condition for connection ] ) is called the _ teleparallel _ or _ weitzenbck _ connection .
let @xmath18 be a nonvanishing real lightlike teleparallel covector field ( @xmath19 , @xmath20 ) .
such a covector field can be written down explicitly as @xmath21 where the @xmath22 are real constants ( components of the covector @xmath18 in the basis @xmath10 ) , not all zero , satisfying @xmath23 we define our lagrangian as @xmath24 where @xmath25 stands for the exterior derivative .
note that @xmath26 is the axial ( totally antisymmetric ) piece of torsion of the teleparallel connection .
let us emphasize that formula ( [ lagrangian as function of vartheta and c ] ) does not explicitly involve connections or covariant derivatives .
the lagrangian ( [ lagrangian as function of vartheta and c ] ) is a rank 4 covariant antisymmetric tensor so it can be viewed as a 4-form and integrated over the manifold @xmath0 to give the action @xmath27 .
variation with respect to the coframe @xmath10 subject to the constraint ( [ constraint for coframe ] ) produces an euler lagrange equation which we write symbolically as @xmath28 the explicit form of the field equation ( [ field equation 1 ] ) is given in griffiths and newing s paper @xcite .
let us define the spinor field @xmath6 as the solution of the system of equations @xmath29 where @xmath30 .
the above system determines the spinor field @xmath6 uniquely up to a complex constant factor of modulus @xmath31 .
this non - uniqueness is acceptable because we will be substituting @xmath6 into the weyl equation ( [ weyl s equation ] ) and weyl lagrangian ( [ weyl s lagrangian ] ) which are both @xmath32-invariant .
we will call @xmath6 the spinor field _ associated _ with the coframe @xmath10 .
the main result of our paper is the following [ main result ] for any coframe @xmath10 we have @xmath33 where @xmath6 is the associated spinor field .
the coframe satisfies the field equation ( [ field equation 1 ] ) if and only if the associated spinor field satisfies the weyl equation ( [ weyl s equation ] ) .
the proof @xcite of theorem [ main result ] is based on the observation that our lagrangian ( [ lagrangian as function of vartheta and c ] ) is invariant under the action of a certain class of local ( i.e. with variable coefficients ) transformations of the coframe , the class in question being the subgroup @xmath34 of the lorentz group @xcite .
this means that coframes come in equivalence classes ( cosets ) and the nature of these cosets is such that they can be identified with spinors .
our lagrangian ( [ lagrangian as function of vartheta and c ] ) has the unusual feature that it depends on a quartet of real parameters @xmath22 which can be chosen arbitrarily as long as they satisfy the condition ( [ constraint for constants ] ) .
this parameter dependence requires an explanation .
one possible explanation of the physical nature of the @xmath22 s is sketched out below .
consider the lagrangian @xmath35 ( `` axial torsion squared '' ) . putting @xmath36 and varying with respect to the coframe @xmath10 subject to the constraint ( [ constraint for coframe ] ) we get an euler lagrange equation which we write symbolically as @xmath37 it turns out that in the case of _ minkowski _ metric one can construct an explicit solution of ( [ field equation for quadratic lagrangian ] ) as follows .
let @xmath38 be a constant reference coframe and let @xmath39 be a constant real lightlike covector ; here `` constant '' means `` parallel with respect to the levi - civita connection induced by the minkowski metric '' .
of course , @xmath40 for some real constants @xmath22 satisfying the condition ( [ constraint for constants ] ) .
perform a rigid lorentz transformation @xmath41 so that @xmath42 for some @xmath43 , put @xmath44 where @xmath45 ( here @xmath46 are arbitrary local coordinates on the manifold @xmath0 and @xmath47 are minkowskian coordinates on @xmath0 associated with the constant coframe @xmath48 ) , and , finally , set @xmath49 .
straightforward calculations show that the coframe @xmath10 is indeed a solution of ( [ field equation for quadratic lagrangian ] ) .
we call this solution _ plane wave _ with _ momentum _ @xmath18 . note that for a plane wave @xmath50 and @xmath51 where the @xmath22 are the original constants from ( [ origin of the l_j s ] ) .
now consider a perturbation of a plane wave .
this perturbation can be the result of either a ) us looking for a wider class of solutions or b ) the metric ceasing to be minkowski .
application of a formal perturbation argument to the lagrangian ( [ axial torsion squared ] ) with @xmath52 as small parameter gives the lagrangian ( [ lagrangian as function of vartheta and c ] ) . | the main result of the paper is a new representation of the weyl lagrangian ( massless dirac lagrangian ) . as the dynamical variable we use the coframe , i.e. an orthonormal tetrad of covector fields .
we write down a simple lagrangian wedge product of axial torsion with a lightlike element of the coframe and show that variation of the resulting action with respect to the coframe produces the weyl equation .
the advantage of our approach is that it does not require the use of spinors , pauli matrices or covariant differentiation .
the only geometric concepts we use are those of a metric , differential form , wedge product and exterior derivative .
our result assigns a variational meaning to the tetrad representation of the weyl equation suggested by j. b. griffiths and r. a. newing . |
the tev source j2032 + 4130 ( aharonian et al . 2002 ) was the first unidentified very high energy ( vhe ) @xmath0-ray source , and also the first discovered extended tev source , likely to be galactic .
intensive observational campaigns at different wavelengths have been carried out on tev j2032 + 4130 .
butt et al .
( 2003 ) presented an analysis of the co , hi , and infrared emissions , together with first observations by _ chandra _ ( 5 ksec ) and a reanalysis of vla data .
these observations showed that the tev source region is positionally coincident with an outlying group of stars ( from the cygnus ob2 core ) , although they failed to identify a counterpart .
mukherjee et al . ( 2003 ) analyzed the same _ chandra _ data and provided optical follow - up observations of several of the brightest x - ray sources , confirming that most were either o stars or foreground late - type stars .
a deeper _ chandra _ observation ( 50 ksec , butt et al .
2006 ) , found hundreds of star - like sources and yet no diffuse x - ray counterpart emission .
a deep ( @xmath12 50 ksec ) _ xmm - newton _ exposure has also been obtained ( horns et al .
2007 ) . after the subtraction of the contribution of known sources from the data , an extended x - ray emission region with a fwhm size of @xmath12 12
arcmin was reported .
the centroid of the emission is co - located with the position of tev j2032 + 4130 and was proposed as the counterpart of the tev source .
the question whether the result reported by horns et al
. can be interpreted as a truly diffuse background , or it could be a result of unresolved x - ray sources , remains disputable .
paredes et al .
( 2007 ) and mart et al .
( 2007 ) have provided deep radio observations covering the tev j2032 + 4130 vicinity using the giant metrewave radio telescope and discovered a population of radio sources , some in coincidence with x - ray detections by butt et al .
( 2006 ) and with optical / ir counterparts .
at least three of these sources are non - thermal , and one has a hard x - ray energy spectrum .
they found extended non - thermal diffuse emission in the radio band apparently connecting with one or two radio sources .
it is yet to be determined if one or more of these sources is similar to some of the known @xmath0-ray binaries ( e.g. , aharonian et al .
2006 , albert et al .
2006a ) .
several theoretical explanations for the tev emission from j2032 + 4130 have been given . among them , those related with extragalactic counterparts , e.g. , a radiogalaxy ( butt et al .
2006 ) or a proton blazar ( mukherjee et al . 2003 ) , face the difficulty of explaining the extended appearance of the source .
gamma - ray production in hypothetical jet termination lobes of cyg x-3 was explored ( aharonian et al . 2002 ) , but the putative northern lobe of cyg x-3 ( now considered a mere thermal hii region , mart et al .
2006 ) is far from the location of the tev source .
a yet unknown pulsar wind nebula ( pwn ) was proposed by bednarek ( 2003 ) , although no clear pwn signal was observed .
a distant microquasar was proposed by paredes et al .
( 2007 ) , perhaps related with one of the x - ray / radio sources they discovered .
if such an association is accepted , the extension of the source could be explained by the diffusion of accelerated particles into a hypothetical nearby molecular enhancement ( see bosch - ramon et al .
torres et al .
( 2004 ) and domingo - santamara & torres ( 2006 ) studied the relationship between the tev emission and the known massive stars in the area , through the interaction of relativistic protons with wind ions .
the distribution of stars in the neighborhood favors this interpretation ( butt et al .
an explanation involving the excitation of giant dipole resonances of relativistic heavy nuclei in radiation dominated environments has also been suggested ( anchordoqui et al .
we start by making a brief summary of what has been claimed by other experiments observing at the highest energies .
the hegra iact using four years of data ( from 1999 to 2002 ) found a source to the north of cygnus x-3 , steady in flux over the years , extended , with radius [email protected]@xmath100.9@xmath11 arcmin , and exhibiting a hard energy spectrum with a photon index of @xmath9=-1.9@xmath2 [email protected]@xmath4 ( aharonian et al .
2005 ) . its integral flux above 1 tev amounts to @xmath125% of the crab nebula , assuming a gaussian profile for the intrinsic source morphology .
the center of the source position was determined quite accurately at @xmath13=20@xmath14 31@xmath15 57@xmath16 @xmath2 6@xmath17 @xmath2 13@xmath18 and @xmath19=41@xmath2029@xmath2156@xmath228@xmath23 1@xmath24 @xmath2 1@xmath25 .
the whipple collaboration reported an excess at the position of the hegra unidentified source ( 3.3@xmath1 ) in their archival data of 1989 and 1990 ( lang et al .
2004 ) , with a flux level of @xmath12 12% of the crab nebula for e@xmath26600 gev .
the detected flux is in conflict with the hegra flux level and steady nature of the source , assuming they all have the same origin .
this large difference between the detected flux levels , if physical , might suggest episodic emission ( with low duty cycle ) or variability over timescale measured in years .
nevertheless , the existence of @xmath0-ray variability is difficult to reconcile with the extended appearance of the source .
also the large difference might be in part due to unspecified systematic errors on the flux determination .
recently , the whipple collaboration reported new observations of this field done with their 10-m telescope for 65.5 hours during 2003 and 2005 ( konopelko et al .
their data is consistent with either a point - like or an extended source with less than 6@xmath27 angular size .
regarding the position , the hegra and the latest whipple data are barely in agreement : their centers of gravity are @xmath129@xmath27 apart , and only agree when adding up the spatial uncertainties in both data sets in opposite directions .
konopelko et al .
do not provide a energy spectrum for this source , but give a 8% crab - level flux ( although with no energy threshold specified ) under the assumption of a steep ( crab - like ) energy spectrum .
the cygnus region shows an excess in the milagro data ( abdo et al .
the flux at 20 tev in a 3x3 square degree region centered at the hegra position is ( [email protected]@xmath102.7@xmath4 ) @xmath510@xmath28 tev@xmath8 @xmath7 s@xmath8 assuming a differential energy spectrum e@xmath29 .
this flux is three times the hegra flux extrapolated at 20 tev .
the tibet air shower detector recently reported evidence for an excess also in their vhe @xmath0-ray candidate set from this region ( amenomori et al .
2006 ) .
in this rich observational and theoretical context we report here on magic telescope observations of tev j2032 + 4130 .
the magic single dish imaging air cherenkov telescope ( see e.g. , cortina et al .
2005 for a detailed description ) is located on the canary island of la palma .
its angular ( energy ) resolution is approximately 0.09@xmath30 ( 20% ) , and the trigger ( analysis ) threshold is 55 ( 60 ) gev at zenith in dark conditions ( see albert et al .
one of the unique characteristics of magic is its capability of observing under moderate moon light illumination ( albert et al . 2007b ) albeit with a slightly elevated threshold .
the field of view of tev j2032 + 4130 was observed with magic for more than 100 hours distributed in 2005 , 2006 and 2007 , see table [ tab1 ] . during the first period in summer 2005 ,
the observation was carried out in on / off mode , that is , the source was observed on - axis while observations from an empty , nearby field of view were used to estimate the background . in summer 2006 and 2007 , the data were taken in wobble mode , using five positions around the hegra position instead of the usual two symmetrical position in order to monitor a wider field of view .
quality cuts based on the trigger and after - cleaning rates were applied in order to remove bad weather runs and data spoiled by car or satellite light flashes . after these quality cuts the total observation time is 93.7 h. the energy range for which we report these results is significantly above the aforementioned trigger and analysis threshold energies due to the fact that the observations were scheduled during moonlight and at relatively high zenith angles ( up to 44@xmath31 ) .
the data analysis was carried out using the standard magic analysis and reconstruction software ( bretz & wagner 2003 ) .
it follows the general stream explained in albert et al .
( 2006b , c , d ) . after calibration and two levels of image cleaning tail cuts ( for image core and boundary pixels ,
see fegan 1997 ) , the camera images are parameterized by the so - called image parameters ( hillas 1985 ) .
the random forest method was applied for the @xmath0/hadron separation ( albert et al .
using this method a parameter , dubbed hadronness ( h ) , can be calculated for every event and which is a measure of the probability that the event is not @xmath0 like .
the @xmath0 like sample is selected for images with a h below a specified value , which is optimized using a sample of crab nebula data processed with the same analysis stream .
an independent sample of monte carlo @xmath0-showers was used to determine the cut efficiency . since part of our observations
was recorded during partial moon - shine , we have corrected the efficiency loss due to the increase of ambient light following the procedure outlined in albert et al .
( 2007b ) .
the @xmath32-distribution was calculated , being @xmath33 the angular distance between the source direction and the reconstructed arrival direction of the showers .
the reconstruction of individual @xmath0-ray arrival directions makes use of the disp method ( domingo - santamaria et al .
the expected number of background events is calculated using five regions symmetrically placed for each wobble position with respect to the center of the camera and refered to as anti - sources .
figure [ fig1 ] shows the distribution of the @xmath32 parameter for the excess observed from the direction of the source , for a size cut of 800 photoelectrons ( pe ) .
this relatively high size cut was selected in order to optimize the sensitivity for a source with such a hard energy spectrum observed during moonlight .
therefore , the total number of @xmath0-like excesses after hillas cuts and applying a cut in @xmath340.05 , is @xmath35 , for which a total significance of @xmath36 is obtained .
table [ tab2 ] shows the number of excesses above background for the different observing periods .
the excess is fitted to a gaussian function folded with the telescope psf , as obtained from monte carlo simulations and validated with crab nebula observations .
the source is extended with respect to the magic psf .
its intrinsic size assuming a gaussian profile is @xmath37 = [email protected]@xmath380.6@xmath4 arcmin .
the exact shape of the source , even if similar to the kev diffuse emission reported by horns et al .
2007 , can not be completely trusted due to limited statistics and telescope pointing systematics .
figure [ fig2 ] shows the gaussian - smoothed ( @xmath1=4 ) map ( [email protected]@xmath40 ) of @xmath0-ray ( background subtracted ) around tev j2032 + 4130 for energies
e @xmath26 500 gev .
the position of a few previously observed @xmath0-ray source candidates are also shown , namely cyg x-3 , the egret source 3eg j2033 + 4118 ( with its confidence contour at 95@xmath41 ) , the wolf rayet star wr 146 , and the whipple and hegra experimental positions .
the regions around cyg x-3 , wr 146 and 3eg j2033 + 4118 have been further investigated by us and no detection is obtained for a steady emission . the upper limit fluxes ( rolke et al . 2005 ) for 95@xmath41 confidence level , above 500 gev for a point - like source at these positions are given in table [ tab3 ] . to determine the best position of the magic detection the excess map was fitted to a 2d bell - shaped function .
the result is shown in the skymap with a black cross as well as by a circle indicating its size .
the best - fit coordinates are ra@xmath42=20@xmath14 32@xmath15 20@xmath43 @xmath2 11@xmath44 @xmath4511@xmath46 and dec@xmath42=41@xmath20 30@xmath21 36@xmath220 @xmath2 1@xmath47 @xmath2 1@xmath48 ( for more details on the systematic uncertainties in the source position determination , see bretz et al .
the position found is compatible within errors with the one determined by hegra , and barely compatible with the claims by whipple mentioned above ( in konopelko et al .
2007 ) . the tev j2032 + 4130 energy spectrum was obtained using the tikhonov unfolding technique ( tikhonov @xmath49 arsenin 1979 ) .
it can be fitted ( @xmath50 ) by a power law function .
the differential flux ( tev@xmath8@xmath7s@xmath8 ) is : @xmath51 the errors quoted are only statistical .
the systematic error is estimated to be 35@xmath41 in the flux level and 0.2 in the photon index ( see albert et al .
the differential energy spectrum is shown in figure [ fig3 ] . the hegra tev j2032 + 4130 and magic crab nebula measured spectra ( in albert et al .
2007a ) are shown in blue solid line and black dotted line , respectively .
the magic energy spectrum is compatible both in flux level and photon index with the one measured by hegra .
crab nebula data from the same periods and zenith angle distributions were studied with the same analysis chain to check for any systematic deviation due to the long observation period .
no indication of time variability was observed : the source integral flux is constant within errors , at 3@xmath41 of the crab nebula flux .
the relative systematic uncertainty in the ratio of both fluxes was estimated to be less than 10@xmath41 .
this uncertainty comes mainly from the slightly different atmospheric transmission conditions and differences in the detector parameters during data taking of the source and the crab nebula . for illustrative purposes ,
the dotted lines in figure [ fig3 ] represent one - zone hadronic and leptonic models of the high energy emission , both consistent with observations at lower energies in the region . under the hadronic scenario , the @xmath52
are obtained from a proton parent population described by a power law ( index @xmath53 ) with exponential cutoff at 100 tev .
the cutoff value was adopted to be consistent with the upper limit at the highest energies coming from the hegra spectrum .
the inverse compton spectrum is obtained from an electron population with equal index and a 40 tev exponential cutoff scattering off the cmb photons . as in aharonian et al .
2005 , we do not consider here the conditions under which particles are accelerated or how they lose energy .
our leptonic fits ( see also the quoted paper for an sed representation ) can only cope with the data if we are actually looking at a compton peak around the energy range of detection , which is not fully discarded within errors .
both models are compatible with the high energy emission . confirming the reality of the diffuse emission detected at lower energies is crucial to distinguish between these and more complex models .
magic observations confirm the location of tev j2032 + 4130 found by hegra .
the magic observation shows an extended source with a significance of 5.6@xmath1 . we find a steady flux with no significant variability within the three year span of the observations ( with the flux being at a similar level of the hegra data of the period 2002 - 2005 ) .
we also present the source energy spectrum obtained with the lowest energy threshold to date , which , within errors , is compatible with a single power law .
we thank the iac for the excellent working conditions at the orm .
the support of the german bmbf and mpg , the italian infn , the spanish cicyt , the eth research grant th 34/04 3 , and the polish mnii grant 1p03d01028 is gratefully acknowledged .
albert j. et al .
( magic collab . ) 2007b , `` very high energy gamma - ray observations during moonlight and twilight with the magic telescope '' submitted to astroparticle physics , astro - ph/0702475 .
albert j. et al .
( magic collab . ) 2007c , `` implementation of the random forest method for the imaging atmospheric cherenkov telescope magic '' submitted to astroparticle physics , arxiv:0709.3719 .
amenomori m. et al .
2006 , science , 314 , 439 .
-parameter for events coming from the direction of tev j2032 + 4130 ( size@xmath26800 pe ) , the background distribution subtracted ( black points ) . a convolved radial gaussian fit f = a@xmath5exp(-0.5@xmath32/(@xmath54 ) ) is indicated by the solid black line with @[email protected] arcmin .
the @xmath56 was measured from mc simulation and validated with crab nebula observations to be @[email protected] arcmin ( dashed black line ) . ]
= 4 ) map of @xmath0-ray excess events ( background - subtracted ) for energies above 500 gev .
the magic position is shown with a black cross .
the surrounding black circle corresponds to the measured 1@xmath1 width .
the last position reported by whipple is marked with a white cross while the hegra position is shown with a blue cross in the center of the field of view .
the error bars , in all cases , correspond to the linear sum of the statistical and systematic errors .
the green crosses correspond to the positions of cyg x-3 , wr 146 and the egret source 3eg j2033 + 4118 .
the ellipse around the egret source marks the 95@xmath41 confidence contour . ]
error in the fitted energy spectrum .
the flux observed by whipple in 2005 and in the milagro scan are marked with colored squares ( blue and pink , respectively ) .
the grey dotted line represents the crab nebula energy spectrum measured by magic .
the blue line shows the hegra energy spectrum .
theoretical one - zone model predictions are depicted with dashed lines . ] | we observed the first known very high energy ( vhe ) @xmath0-ray emitting unidentified source , tev j2032 + 4130 , for 94 hours with the magic telescope .
the source was detected with a significance of 5.6@xmath1 .
the flux , position , and angular extension are compatible with the previous ones measured by the hegra telescope system five years ago
. the integral flux amounts to ( [email protected]@xmath30.35@xmath4)@xmath510@xmath6 ph @xmath7 s@xmath8 above 1 tev .
the source energy spectrum , obtained with the lowest energy threshold to date , is compatible with a single power law with a hard photon index of @[email protected]@xmath100.2@xmath11 . |
some aspects of modeling lensed quasars are much as they were just after the discovery of the first double quasar 0957 + 561 . in one of the earliest lens - modeling papers , @xcite are concerned with some now - very - familiar issues : the effect on image positions of both the main lensing galaxy and other galaxies , the time delays predicted by the models , the non - uniqueness of the models despite the adequacy of the data , and the desirability of supplementary data about the lens , such as velocity dispersions or x - rays .
but other aspects of the subject these days would have been unimaginable in 1981 .
the first double quasar has been joined by dozens of others : the castles survey compilation @xcite currently lists 76 secure multiple - image galaxy - lens systems , with image positions measured to the mas level , even more precisely if there are compact radio sources . and
the time delay for 0957 + 561 , for which young et al.s preliminary estimate was about 5 years , is now measured as @xmath6 @xcite , along with time - delay measurements for eight other systems .
these excellent data demand new , more automatic , and more portable software tools for modeling the lenses .
thus motivated , we have developed a new code , _
pixelens , _ which we present in this paper .
_ pixelens _ works by reconstructing a pixelated mass map for the lens , an idea we first implemented in @xcite .
most lens modeling codes work by fitting a parametric functional form for example , @xcite fitted king models ; _ gravlens _ by @xcite is a modern example , offering the user a large choice of parametric models .
there are other possibilities too : @xcite reconstruct lenses non - parametrically as we do , but use multipole expansions rather than pixels .
_ pixelens _ generates large ensembles of models rather than one or a few mass maps , as a way of addressing the non - uniqueness problem .
we introduced this strategy in @xcite for pixelated models .
@xcite use a somewhat similar strategy , but with parametrized models .
but _ pixelens _ also brings two completely new features , one astrophysical , one computational : * it can model several lenses simultaneously , enforcing consistency of @xmath0 across different time - delay lenses . as a result
, lenses can be used to constrain other lenses in an interesting and surprising way ; * the code is highly portable and can run without change as a standalone program , or as a java applet inside a web browser . in the online version of this paper , it is available as an alternative version of figure [ gui ] .
the best way to explain what we have implemented and what problems remain is through an example .
so in the following section we will work through a simultaneous reconstruction of three lenses : the time delay quads 1115 + 080 and 1608 + 656 , and the ten - image system 1933 + 503 .
this will be the main part of the paper .
we turn to some computational issues after that .
in our first worked example , we consider the three lenses currently best - constrained by observations . 1 .
1115 + 080 is a quad where the lens is an elliptical galaxy supplemented by external shear from other group galaxies .
it was the second lens to be discovered @xcite .
@xcite , together with an improved time - series analysis by @xcite provide time - delays between two pairs of images .
2 . 1608 + 656 is a quad with the lens being apparently two interacting galaxies .
it was discovered @xcite in the class survey .
there are time delays between all pairs of images @xcite .
1933 + 503 is really two quads and a double .
the lens appears to be an elliptical , but the source is a radio source with a core and opposing lobes .
the core and one of the lobes are imaged as quads while the other lobe is imaged as a double .
it was also discovered in the class survey @xcite .
there are no measured time - delays , but there are prospects for time delays from the core quad .
we will predict time delays for the core quad .
the castles web page provides further information .
we can also make some preliminary inferences about the lenses by examining the morphology , as explained in @xcite : we can work out the time - ordering of each of the image systems , and for 1115 + 080 we can recognize an external shear and its approximate axis .
let us now proceed to apply _
pixelens_. figure [ gui ] shows the user interface , which we will explain in stages below .
data is entered by typing into the ` data area ' .
( there is also an ` input paste ' feature with a number of example inputs , which can be pasted and then edited by selecting the ` edit ' option of the menu . ) for 1115 + 080 , we enter the following .
object 1115 symm pixrad 12 maprad 2 redshifts 0.311 1.722 shear -45 quad 0.3550 1.3220 -0.9090 -0.7140 13.3 -1.0930 -0.2600 0 0.7170 -0.6270 11.7 the first line
just labels the object .
the second line specifies that the mass maps be inversion - symmetric with a radius of @xmath7 and 12 pixels ( thus setting the pixel size ) .
the fourth line gives the approximate shear axis ( ` shear 135 ' would be equivalent ) , in degrees relative to east . the program will consider positions angles within @xmath8 of what is given .
the last four lines give the image positions ( in arcseconds and relative to the lens center ) and the time delays ( in days and relative to the previous image ) .
the images must be ordered by arrival - time .
the ` 0 ' in the second - last line means that the time - delay is @xmath9 but unknown .
otherwise , uncertainties in image positions and time delays are assumed negligible .
note how the input format assumes that some preliminary analysis , of the qualitative variety explained in @xcite has already been done . for 1608 + 656
we enter the following .
object 1608 pixrad 9 maprad 2 redshifts 0.630 1.394 quad -1.300 -0.800 -0.560 1.160 31 -1.310 0.700 5 0.570 -0.080 40 the format is just as before , except that we have left out symm , because the lens here appears to be asymmetric . for 1933 + 503
we enter the following .
object 1933 symm pixrad 10 redshifts 0.76 2.63 shear 45 quad -0.40 0.53 0.43 -0.26 0 0.40 0.19
0 -0.19 -0.33 0 quad 0.54 -0.44 -0.15 0.44 0 -0.03 0.45 0 -0.38 -0.12 0 double -0.47 0.60 0.13 -0.30 0 here we have left out the mass - map radius , leaving the program to choose it .
we have also given image data on two quads and a double rather than just one quad , and indicated unknown time delays with many 0s .
line breaks in the input have no significance .
figure [ pix ] summarizes the image data and the pixelation .
( for symmetric mass maps , only half the pixels are shown here , the other half are specified by inversion symmetry . ) _ pixelens _ produces such plots before starting the real computations , and they provide a convenient goof test on the input .
now we type models 100 in the ` data area ' and click on the ` run ' button , and away we go . at this point
, we recall some lensing theory . for a source at @xmath10 , the arrival - time at @xmath11 is @xmath12 where @xmath13 is the dimensionless density ( also called convergence ) .
if we measure angles in arcsec , @xmath14 will have units of @xmath15 . to turn it into days we have to multiply by @xmath16 , where @xmath17 is a time scale and @xmath18
is @xmath19 in gyr ( see appendix a ) .
in addition to the indicated redshift dependence , @xmath20 has a weak cosmology dependence . by default _
pixelens _ assumes @xmath21 , which gives @xmath22 days , @xmath23 days , and @xmath24 days . in our models ,
the mass is pixelated .
therefore we can simplify equation ( [ eq - arriv ] ) to @xmath25 where @xmath26 is the density on the @xmath27-th pixel and @xmath28 is the integral from ( [ eq - arriv ] ) evaluated over that pixel . the explicit form of @xmath28 is a messy but straightforward mixture of logs and arctans and is given in @xcite .
the important thing about equation ( [ eq - arrivpix ] ) is that it is linear in the unknowns , @xmath10 and @xmath26 .
we also allow a constant external shear @xmath29 to be added to equation ( [ eq - arrivpix ] ) ; again , the shear is linear in the unknowns @xmath30 .
we now implement the following observational constraints . 1 .
an image observed at @xmath31 implies @xmath32 in a multiple - image system all the images have the same unknown @xmath10 , so each multiple - image system introduces @xmath33 constraints .
thus a double provides 2 image constraints , a quad 6 , while 1933 + 503 provides 14 .
2 . a time - delay measurement between images at @xmath31 and
@xmath34 implies @xmath35 here @xmath18 is another unknown and , fortunately for us , @xmath36 appears linearly . if a time - delay is unmeasured , but we are sure about the ordering , we replace ( [ eq - obsdel ] ) by an inequality constaint @xmath37 we have 3 equality constraints from time delays in 1608 + 656 , 2 in 1115 + 080 , and none in 1933 + 503 .
we require the images to have the expected parities , and for the image elongation when projected along the radial direction to be between @xmath38 and 10 .
although very modest , these requirements do suppress a tendency for spurious images to be created in the vicinity of a nearly - merging pairs of genuine images .
constraints of this type are also linear @xcite .
we remarked after equation ( [ eq - obsdel ] ) that @xmath18 is an unknown variable in the constraint equations .
now , if we write the constraint equations for several lenses , then each lens will introduce a separate unknown variable for @xmath18 , say @xmath39 , @xmath40 and so on . but
@xmath18 is universal , so we have additional constraint equations of the type @xmath41 though @xmath18 is still unknown , it now couples the constraint equations for different lenses .
this coupling is the central astrophysical contribution of this paper . in particular , it means that if one lens rules out a certain range of @xmath18 , _ pixelens _ will not consider models of any lens that require @xmath18 in that range .
we also impose secondary constraints based on our general knowledge of galaxy mass profiles . 1 .
all the @xmath42 , 2 .
if the galaxy does not appear very asymmetric , the mass profile is required to have inversion symmetry about the lens center .
3 . the density gradient anywhere must point within @xmath43 of the lens center . the precise formulation of this constraint is given in @xcite .
4 . the @xmath26 of any pixel must be @xmath44 , except for the central pixel , which is allowed to be arbitrarily dense .
the radial mass profile must be steeper than @xmath45 .
the stellar dynamical evidence indicates that the total density distribution in the central regions of ellipticals is well approximated by isothermal , @xmath46 @xcite , while the study of the dynamics of the gas in the center of our galaxy show that total density scales as @xmath47 @xcite .
guided by these studies we choose @xmath45 as a conservative minimal steepness of the 2d density profiles in galaxies . in bayesian terminology
, these secondary constraints constitute a prior .
the observational and prior constraints confine allowed lenses to a polyhedron in the space of @xmath48 .
_ pixelens _ searches this polyhedron by a monte - carlo technique , briefly explained in figure [ walk ] .
the key idea is to random - walk between models such that probability of moving from model @xmath49 to model @xmath50 equals the probability of moving from model @xmath50 to model @xmath49 .
this is a simple example of a metropolis algorithm , or a markov - chain monte - carlo method .
generalizations to non - uniform priors are also possible see , for example , @xcite . in our example , each model that _ pixelens _ generates is a three - lens model : there are mass maps , external - shear models where required , and inferred source positions for each of 1115 + 080 , 1608 + 656 , and 1933 + 503 . in any one model the lenses
are coupled by a common value of @xmath18 , but different models may have different values of @xmath18 . as soon as the first model is ready , the user can examine it .
but the users will not want to examine every model in an ensemble of 100 .
so _ pixelens _ presents two kinds of results from the ensemble : details of the average model of the ensemble - so - far and various statistics derived from the ensemble .
( by ensemble - average model we mean a model with the densities in each pixel , the source positions , and any external shear averaged over the ensemble .
the constraint equations being linear in the quantities being averaged , they are still satisfied by ensemble averages . ) let us consider the two kinds of results in turn .
figure [ mass ] shows mass maps for the three lenses from the ensemble - average model . for 1115 + 080 and 1608 + 656 ,
the mass maps are very similar to those in @xcite .
the modeling method and input data in the earlier paper were very similar except for the coupling ( [ hcoup])but the code was completely different .
we remark that the elongation of the mass profiles in 1115 + 080 is towards the other group galaxies , the elongations in 1608 + 656 and 1933 + 503 are what we might expect from the light in those lenses .
figure [ poten ] shows the lens potential in 1115 + 080 .
( the default is to pick the contours that pass through the images , but the user can try different contour spacings interactively , by typing in the appropriate values in the ` plot parameters ' boxes . ) notice how smooth the potential is : nothing of the pixelation so evident in the mass maps is visible here ; it has been washed out by the double integral in equation ( [ eq - arriv ] ) .
figure [ poten ] also illustrates a convenient way of gauging the importance of external shear in a model . _
pixelens _ allows the user to plot the lens potential with the external shear exaggerated .
( such exaggeration is purely for visualization , it has no effect on the models . )
the user can adjust an exaggeration parameter : the default is 1 , while 0 means omit the external shear .
figure [ poten ] shows that a 3-fold exaggeration makes a drastic difference to the potential , indicating that the external shear in this lens is large but not dominant .
figure [ arrivb ] show saddle - point contours in the arrival - time surfaces for the two quads and a double of 1933 + 503 .
saddle - point contours are the default , but as with the potential , the user can try different contour spacings interactively . examining
the arrival - time contours is a very good way of checking that the model is sensible : small errors in the input or the modeling code nearly always lead to noticeable spurious features in the arrival - time surface , such as contorted contours or extra images .
arrival - time contours also have another , somewhat surprising , use .
figure [ arrivc ] shows the result if the user specifies a contour - spacing of 0.004 .
( the units here are @xmath51 .
we can convert to physical time units , as explained after equation [ eq - arriv ] , and the result is about @xmath52 , but let us continue with @xmath51 . )
a plot of the arrival - time surface with closely - spaced contours tends to resemble an einstein ring .
in fact , it approximates see @xcite for details the image of a source with a conical light profile of radius @xmath53 now , the line thickness in all the plots in this paper is @xmath54 of the plot width . for figure [ arrivc ] , that amounts to @xmath55 .
hence figure [ arrivc ] predicts the image due to a conical profile of radius @xmath56 centered on the qso .
in fact it resembles the ring imaged by @xcite , though there are differences of detail .
thus , we can make quantitative inferences about the host galaxy simply by inspecting images and doing some mental arithmetic . out of the full ensemble of models , it is possible to extract many kinds of statistics . _ pixelens _ implements four kinds , all involving @xmath19 .
figure [ nh ] shows a histogram of @xmath18 values from the ensemble , and has a straightforward interpretation as the posterior distribution of @xmath19 .
this figure is analogous to figure 17 in @xcite , but better . in the earlier work we had separate model ensembles for 1115 + 080 and 1608 + 656 and multiplied their @xmath57-histograms bin - by - bin .
that procedure had the problems of ( i ) a reduced number of models in the common region , and ( ii ) increased shot noise from bin - by - bin multiplication . in _
pixelens , _ because of the constraint ( [ hcoup ] ) , there is only one histogram of @xmath18 , which eliminates both problems . this histogram ( or rather , the unbinned values it represents ) gives @xmath58 the median @xmath0 is somewhat higher than in @xcite , but this is expected because that paper used an einstein - de - sitter cosmology , which shortens @xmath20 by @xmath59 for 1115 + 080 and by @xmath60 for 1608 + 656 .
the uncertainties are similar .
figure [ del ] shows the predicted time - delays for the core - quad in 1933 + 503 .
figure [ kalf ] shows the radial index @xmath61 from a radial fit of @xmath62 against @xmath36 .
most lens systems when modeled independently show a correlation between @xmath61 and @xmath36 ( or @xmath57 ) , in the sense that steeper density profiles result in higher values of @xmath57 ( e.g. , figure 13a and 16a of williams & saha 2000 ) . for the three coupled lenses shown here , 1115
+ 080 shows the correlation weakly and the others hardly at all .
figure [ kann ] shows the mean @xmath13 in the annulus between the innermost and outermost images ( say @xmath63 ) against @xmath36 .
again , when modeled independently , most lens systems show an anti - correlation between @xmath63 and @xmath36 , whereas two of the three panels of figure [ kann ] show no trends at all .
this can be expected , because coupling different time - delay systems weakens correlations : coupling reduces the range of @xmath18 explored , and over a reduced range of @xmath18 any correlation involving @xmath18 will appear weaker . for an extreme case ,
suppose we artificially fix the value of @xmath18 , which we can do by adding ( say ) g 14 at the end of the input
. then figures [ kalf ] and [ kann ] will consist of vertical clusters of point at @xmath64 .
however , the above argument does not fully account for the lack of trends in the middle and lower panels of figure [ kann ] ; other causes are partially responsible .
@xcite derives a linear relation between time delay , @xmath63 , and @xmath57 , as a leading order approximation from lensing theory .
if one of these three ( sets of ) variables is fixed , say if the time delays are fixed by observations , then a well defined trend would be expected to relate the other two ( @xmath63 and @xmath57 ) .
this is demonstrated by 1115 + 080 .
1933 + 503 , on the other hand , is not expected to exhibit such a trend because no time delays exist for this system .
finally , 1608 + 656 shows no trend probably because it is a distinctly asymmetric lens . to finish this worked example
, we remark on some artifacts evident in pixelated models .
the most obvious spurious feature is the pixel - scale substructure . in principle
, one could replace mass pixels with basis functions , thereby preserving the linearity of the constraint equations and leaving the main algorithms unchanged , but making the mass maps smooth .
that would be a useful development for microlensing applications , where local density and shear are crucial .
but for studying image positions and time delays , basis functions are unlikely to make any difference , because the pixel scale structure is completely suppressed in the lens potential and arrival time see figure [ poten ] .
pixelation has a serious side - effect , however , if the pixels are made too large . we have found from experimenting with different pixel sizes that if the pixels are too large and the mass - distribution is poorly resolved , fewer steep mass - models are allowed . as a result ,
the prior is shifted to favor lower values of @xmath36 , and hence the hubble - time histograms also shift towards larger @xmath19 .
the effect is most noticeable in the @xmath61-@xmath36 correlation .
the pixel - sizes used for this paper are chosen small enough not to cause this particular problem . a second artefact concerns the central image .
now the central maximum itself is obvious from arrival time contour maps .
what is further often evident is that the central maximum is slightly offset from the lens center and not as steep as expected from the practical unobservability .
this happens because our mass maps lack a central density cusp , instead distributing the mass over the central pixel . in principle
, we could incorporate central density cusps into _ pixelens . _ but it would make no difference to the observed images because , as is well known , a circular mass disk has the same lensing effect outside itself as a point mass .
a third ( and rare ) artefact is spurious extra images .
the density - gradient constraint , we have found , is generally enough to suppress extra images , but it is possible that a few rogue models with spurious extra images are present in the ensembles . in principle we could reject those models by examining every single arrival - time surface by eye , but we have not implemented that .
in our second example we will be more brief . we reconstruct four time - delay doubles . * 1520
+ 530 was discovered by @xcite and @xcite measured its time - delay . *
1600 + 434 was discovered by @xcite and @xcite measured its time delay . * 1830 - 211
was discovered by @xcite and @xcite measured its time delay . *
2149 - 275 was discovered by @xcite and @xcite measured its time delay .
we have already explained the input format , and without further ado here are the data we input .
object 1520 symm pixrad 8 redshifts 0.71 1.855 shear 90 double 1.141 0.395 -0.288 -0.257 130 object 1600 symm pixrad 8 redshifts 0.42 1.59 double 0.610 0.814 -0.110 -0.369 51 object 1830 symm pixrad 8 redshifts 0.89 2.51 double -0.50 0.46 0.15 -0.26 26 object 2149 symm pixrad 8 redshifts 0.490 2.03 double 0.736 -1.161 -0.173 0.284 103 we can omit the number of models since 100 is the default . in the case of pks 1830 - 211
, the position of the lensing galaxy at @xmath65 remains uncertain .
various scenarios have been proposed in the literature @xcite ; we use the astrometry from @xcite .
since readers will have already seen the various kinds of results possible , we now give only a small selection .
figure [ mass2 ] shows the ensemble - average mass maps .
figure [ mass2 ] shows the common histogram of hubble times , from which we obtain @xmath66 figure [ kann2 ] shows @xmath63 against @xmath36 for the four doubles .
all four show a correlation , 1600 + 434 being the best .
these correlations are somewhat worse compared to what they would have looked like if the lens systems were modeled independently ( as the reader can readily verify using _ pixelens _ ) : as explained in section [ stats ] , coupling lens systems weakens the correlations .
the first decision we made about _ pixelens _ was to write it in java .
the overriding consideration was portability : we wanted to code to run without modification on a high - end workstation , a laptop computer used during a talk , or inside a web browser . while developing the software we also found other advantages of java .
for example , the graphics libraries are part of the standard , making guis very easy to write . also , the language is remarkably clean and easy to debug .
but these advantages come at a price , and we found there were four issues we had to confront .
first , java still lacks compilers optimized for numerical work .
java numerical code runs slower than c++ by a factor of 3 or more , and also takes up more memory .
this is an annoying inconvenience , but it is not fatal .
( given moore s law , it is like running c++ on a two - year - old computer . )
second , there is very little support for scientific programming . in particular , in all the elaborate java graphics libraries , there is no simple package to do a plot of @xmath67 against @xmath68 ! in response , we just wrote whatever utility software we needed .
for example , we wrote java code to generate postscript code to draw the big and small ticks in an @xmath69 plot .
but this utility code is now available for other projects and to other researchers .
third , disk input / output is not allowed from inside a web browser .
this is an essential security precaution clearly , untrusted code running over the internet must not be allowed disk access .
fortunately , the language allows an elegant workaround .
_ pixelens _ checks whether it is running inside a web browser or as a standalone program ( as an applet or an application in the jargon ) and activates disk input / output only in the latter case .
so users can try the code out over the internet inside a web browser ; if they are interested enough to want disk output , they can download and run like a normal program .
fourth , and most importantly , there is no batch mode inside a web browser . if the program takes more than a few seconds to produce results , which user will want to wait in front of their web browser ?
this seems to us the fundamental reason why numerically - intensive java programs were previously unknown .
one can write such programs and run them in batch mode in the normal way , but if they are useless inside a web browser the single largest advantage of java is lost .
but there is a way out .
that way is to provide the user with intermediate results , and the means to post - process them in non - trivial ways , while the main computation is still running . with _
pixelens , _ the user can start examining the results as soon as the first model of the ensemble is complete , even though only 1% of the total computation may have run .
we found that providing the user with intermediate results and sufficient flexibility to make them interesting influenced almost every aspect of our program design .
_ pixelens _ is internally documented in a literate programming idiom @xcite and we will not enter into minutae here , but figure [ schem ] gives a general idea of how the code is organized .
imagine each of the labelled ovals as consisting of many variables and several functions that can operate on those variables .
functions from one oval may know something about the variables and functions of another oval , but only the bare minimum needed to work together .
( text i / o ) is the simplest of the ovals .
it accepts the data and other input like ` run ' and ` pause ' , and outputs diagnostics and error messages , but does not interpret the input itself .
( router ) interprets enough of the input to work out how many lenses there are , allocates the required number of ( lens ) ovals two in figure [ schem]and passes the data to them .
the ( lens ) ovals are the only parts that know about lensing theory and pixelation .
each of them turns the data into constraint equations .
( router ) then takes the constraint equations from each ( lens ) , packs them into one large set of equations , and gives them to ( simplex ) .
( simplex ) generates solutions using linear programming and monte - carlo , and it does nearly all the numerical work involved . each time a new solution is ready , ( simplex ) passes it to ( router ) , which unpacks it and distributes among the ( lens ) ovals .
each ( lens ) oval then post - processes the solution , preparing grids of @xmath13 and @xmath70 and so on .
the post - processed material is passed via ( router ) to ( graphics i / o ) .
( graphics i / o ) can interact with the user and do some non - trivial computation on its own : it can make contour plots , scatter plots , and histograms to the user s specification , all without needing to know any lensing theory .
moreover ( graphics i / o ) can work without interrupting ( simplex ) or the other ovals , even for error - recovery . in effect
, model generation happens in the background while the user examines intermediate results in the foreground .
this paper continues our work on modeling lens quasars using pixelated mass maps .
we elaborate on ideas introduced in earlier papers @xcite : ( a ) formulating lens reconstruction as an undetermined linear inverse problem , ( b ) searching through model space to infer @xmath0 ( or alternatively , predict time delays ) together with systematic uncertainties , ( c ) predicting the main features of any einstein ring , and if observed , estimating the size of the host galaxy . but more importantly , we introduce and implement two new ideas : ( i ) reconstructing several time - delay lenses simultaneously while requiring them to agree about @xmath0 , ( ii ) making the software so portable that it can be run inside a web browser while reading this paper .
the results in this paper are from two worked examples of seven lenses in all . 1 .
first we model the time - delay quads 1115 + 080 and 1608 + 656 , and the ` dec ' 1933 + 503 . the two time - delay lenses help check against our earlier work using a different code . by coupling them with 1933 + 503
we can predict time - delays for the latter that are not conditional on @xmath0 . we also obtain @xmath71 2 .
then we model the time - delay doubles 1520 + 530 , 1600 + 434 , 1830 - 211 , and 2149 - 275 , and here the main result @xmath72 the reference cosmology has @xmath3 .
note that the two worked examples contain no lenses in common , so the close similarity of the two derived values of @xmath0 ( 1 and 2 above ) is very encouraging .
the well - known correlations(i ) steeper lenses give higher @xmath0 , and ( ii ) more mass in the annulus covering the images gives lower @xmath0are present for most of the lenses , and of these ( ii ) is better .
however , such simple correlations appear to get weaker as @xmath0 is better constrained , which happens as a consequence of coupling lenses .
comes to mind : hearing `` after all , a star is a pretty simple thing , '' redman retorted `` at a distance of 10 parsecs _ you d _ look pretty simple '' . ]
furthermore , lenses with no time delays , like 1933 + 503 , as well as lenses with very asymmetric mass distributions , like 1608 + 656 , are not expected to show the above correlations . at this stage
we remain somewhat cautious about the hubble - time estimates .
as we explained in subsection [ prior - sec ] , the distribution of models we obtain is conditional on the prior we use . in the past , when the uncertainties on the hubble time were very large , fine - tuning the prior was not so important .
but now , with the uncertainties shrinking down from better data and improved modeling , improving the prior is probably the next priority .
in lensing the hubble time @xmath19 plays a more fundamental role than the hubble constant @xmath0 . note that the crucial observable has the same dimensions as @xmath19 even though its value is many orders of magnitude different . to obtain convenient units for the computations ,
we define @xmath75 where the distances are computed using a fiducial value @xmath76 ( and in some chosen cosmological model ) .
multiplying further by @xmath77 converts @xmath20 to units of @xmath78 .
then , by using @xmath20 in appropriate places we can leave all positions in arcsec and all time delays in days .
we remark that in currently - favored cosmological models , @xmath19 is nearly equal to age of the universe . in a flat cosmology containing only matter and dark energy , the hubble time is related to the age by ( cf .
equation 5.63 in @xcite .
@xmath79 if @xmath80 then @xmath81 . for @xmath82 the hubble time would equal the age ; in other words , the current expansion rate would equal the historic mean . | we introduce and implement two novel ideas for modeling lensed quasars . the first idea is to require different lenses to agree about @xmath0 .
this means that some models for one lens can be ruled out by data on a different lens .
we explain using two worked examples .
one example models 1115 + 080 , 1608 + 656 ( time - delay quads ) and 1933 + 503 ( a prospective time - delay system ) all together , yielding time - delay predictions for the third lens and a 90%-confidence estimate of @xmath1 gyr ( @xmath2 ) assuming @xmath3 .
the other example models the time - delay doubles 1520 + 530 , 1600 + 434 , 1830 - 211 , and 2149 - 275 , which gives @xmath4 gyr ( @xmath5 ) .
our second idea is to write the whole modeling software as a highly interactive java applet , which can be used both for coarse - grained results inside a browser and for fine - grained results on a workstation .
several obstacles come up in trying to implement a numerically - intensive method thus , but we overcome them . |
the solar corona is often populated by peculiar objects , dense clouds of cold plasma inexplicably floating tens of thousands of kilometres above the photosphere .
such objects are routinely seen during solar eclipses , when they can be easily distinguished by their red glow , but they can also be unveiled with the help of filters , such as h@xmath1 , devised to observe the chromosphere ( fig .
[ fig1 ] ) . to put it in simple words , these objects ( usually called prominences or filaments ) are like chunks of chromospheric gas defying the downward pull of gravity and staying in a place higher than the one that apparently corresponds to their large density .
this is not the only enigma around solar prominences .
in contrast with the mk temperature of the surrounding corona , prominences remain at a comparatively cool 10,000 k , which prompts one to ask what prevents the mechanisms that heat the corona from also raising the temperature of prominences and consequently dispersing them . other pieces of the prominence puzzle concern their beginning and end : first , one may wonder not just how prominences form but also why they are born in an adverse environment .
second , despite their internal dynamics , prominences that have been stable and healthy for weeks suddenly die in an spectacular eruption .
hence , like all living beings these creatures are born , grow , ( do not reproduce ) and die , but unlike many living beings the processes shaping the lifetime of prominences are largely unknown .
nevertheless , the intervention of a decisive element is quite clear : the magnetic field , that is central to all the mentioned processes .
the reason of our limited insight into the nature of prominences probably has three causes @xcite : there is no such thing as a canonical prominence , but a wide range of parameters is observed in different objects ; no prominence has a uniform structure , but they are made of thin threads ( or fibrils ) and , in addition , different parameter values can be detected in different parts of a prominence ; and no structure is really isolated , so it is necessary to understand the physics of the prominence - corona interface , the effect of the coronal radiation field ( e.g. * ? ? ?
* ) and to trace the magnetic fields permeating the prominence to their origin at the sun s surface ( e.g. * ? ? ?
our knowledge about prominences has been well reviewed by @xcite , where ample information on the topic can be found .
one of the first studies about prominence oscillations was carried out by @xcite , who noted that in a sample of 68 non - active region prominences , 31% of the objects presented no significant velocity change along the line - of - sight , 28% showed apparently random line - of - sight velocity variations and 41% presented a definite oscillatory behaviour .
similar results were obtained for a set of 45 active region prominences .
there are several reasons that may lead to the absence of periodic variations in some prominences , e.g. the velocity amplitude or its projection along the line - of - sight are too small to stand above the instrumental noise level ; or the prominence material does not actually oscillate at the time the observations are performed ; or the light emitted or absorbed by various plasma elements along the line - of - sight and having different oscillatory properties results in a noisy signal .
after this leading work the subject of prominence oscillations remained dormant for about 15 years and revived in the 1980s , when a great amount of observations about prominence oscillations started being collected . according to the oscillatory amplitude ,
these events have been classified into two groups @xcite : * _ large amplitude oscillations _ , that arise when the whole prominence is shaken by a moreton wave @xcite impinging on it @xcite . as a consequence ,
the prominence gas undergoes a large displacement from its equilibrium position and the prominence as a whole vibrates with considerable velocity amplitude ( of the order of 20 km s@xmath0 or higher ) .
these phenomena are reviewed in a separate work of this volume ( tripathi et al .
* _ small amplitude oscillations _ , with velocity amplitudes much smaller than those of large amplitude oscillations have also been frequently observed .
the detected peak velocity ranges from the noise level ( down to 0.1 km s@xmath0 in some cases ) to 23 km s@xmath0 , although larger values have also been reported @xcite .
a distinguishing property of small amplitude oscillations , which gives them a distinction from large amplitude ones , is that they seldom influence the entire prominence .
thus , when spectrographic observations are carried out with a slit it is usually found that only a few consecutive points along the slit present time variations with a definite period , while all other points lack any kind of periodicity ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
@xcite made a comprehensive study of the two - dimensional spatial distribution of doppler velocity oscillations in a very large polar crown prominence and showed that , in their particular observation , oscillations are concentrated in a restricted area ( 54,000 @xmath2 40,000 km in size ) .
both propagating waves and standing oscillations are detected in this region and their wavelength and phase speed are determined .
using hinode sot observations of a quiescent prominence , @xcite also report on oscillations that do not affect the whole prominence body .
these oscillations have periods from 20 to 40 minutes and typically last for only one or two cycles .
they correspond to waves propagating vertically along threads with a projected phase speed of @xmath3 km s@xmath0 .
it has been mentioned that solar prominences are constituted by many thin , parallel magnetic threads filled with cold plasma , and as a consequence the dynamics of these components can form the basis of the dynamics of small amplitude oscillations .
early works @xcite already noted the possible link between small amplitude prominence oscillations and the fibril structure .
unfortunately , the spatial resolution of the data analysed by @xcite is not good enough to distinguish the prominence threads .
it was necessary , hence , to wait until the advent of telescopes with much better spatial resolution to have observations in which the prominence fine structure is well resolved @xcite . in the analysis of the doppler velocity in two threads belonging to the same filament , @xcite finds a clear sign of propagating waves and determines their period , wavelength and phase speed .
this study is followed by a much more profound one in which the two - dimensional motions and doppler shifts of 328 features ( or `` blobs '' ) of different threads are examined .
these features are observed to flow along the filament axis while oscillating at the same time . to simplify the examination of oscillations , @xcite computed average doppler signals for each fibril and found that groups of adjacent threads oscillate in phase with the same period .
this has two consequences : first , since the periodicity is outstanding in the averaged signal for each thread , the wavelength of oscillations is larger than the length of the thread .
second , fibrils have a tendency to vibrate bodily , in groups , rather than independently .
h@xmath1 observations conducted with the swedish 1-m solar telescope by @xcite lead to similar results concerning the collective dynamics of fibrils , although propagating doppler velocity signals with various periods and wavelengths in other threads of the same filament are also detected .
all these observations seem to indicate that prominence fibrils sometimes support collective oscillations and sometimes oscillate on their own .
this topic deserves a more detailed observational study and , given the simplicity of fibrils compared to the full filament structure , theoretical investigations can give rise to fruitful comparisons with observations . during the 1980s some observational works , e.g. those by @xcite , provided evidence about a temporal decrease of the oscillatory amplitude , which suggests that prominence oscillations are attenuated in time .
a very clear example of this phenomenon can also be found in figure 5a of @xcite . here the integrated line intensity of the he d@xmath4 line displays oscillations with a period around 25 min and with an amplitude that decreases in time .
these oscillations are present for a few cycles but , unfortunately , a precise value of the attenuation rate is not computed in this case ( figure 3 of * ? ? ?
* also presents similar results , although with a much better spatial and temporal resolution ) .
more detailed observational analyses of damped prominence oscillations can be found in @xcite , who studied the spatial and temporal features of doppler velocity oscillations in limb prominences . in the second of these works ,
already mentioned before , a sinusoidal function multiplied by a factor @xmath5 is fitted to the doppler series , which allowed these authors to obtain values of the damping time , @xmath6 , which are usually between 1 and 3 times the corresponding oscillatory period . in spite of the lack of similar studies
, the existing evidence suggests that small - amplitude oscillations in prominences are excited locally and are damped in a few periods by an unknown mechanism .
many previous observational and theoretical results on prominence oscillations have been reviewed by e.g. @xcite , so our purpose here is to extend these review works by summarising other papers not discussed by these authors .
thus , we describe the first seismological analysis of prominence oscillations ( section [ seismology ] ) and the theoretical studies on the attenuation of small amplitude oscillations ( sections [ leakage ] , [ nonideal ] and [ resabs ] ) .
the plausible theoretical explanations of this effect are reviewed here and the relevance of different mechanisms is assessed on the basis of the ratio of the damping time to the period ( @xmath7 ) , that must be in the range 110 to agree with the observational results .
works that invoke wave leakage from the prominence as the cause for the observed damping are described in section [ leakage ] , attenuation caused by non - ideal effects is discussed in section [ nonideal ] and wave damping by resonant absorption is presented in section [ resabs ] . finally , section [ conclusions ] contains the concluding remarks .
observations with the hinode solar optical telescope by @xcite display an active region prominence composed of many fine horizontal threads that oscillate independently in the vertical direction as they move along a path parallel to the photosphere .
six different threads are studied and in each of them several points are considered .
all points in each thread are found to oscillate in phase .
the relevant parameters for a seismological analysis of these events are flow velocities in the range 1546 km s@xmath0 , oscillatory periods in the range 135250 s and thread lengths in the range 170016,000 km .
a prominence thread is a cold plasma condensation that occupies a segment of a much longer magnetic tube . in the present case
it is not easy to directly measure the length of such magnetic tubes , although @xcite estimate that the wavelength of the oscillations is at least 250,000 km and so the minimum length of the magnetic tubes is 125,000 km . @xcite
provide another estimation of this quantity , which they argue must be at least 100,000 km .
these authors perform an independent seismological study of the six events by assuming that the thread is a dense plasma moving along a horizontal and straight magnetic tube which is tied to the dense photosphere at its ends .
additionally the low-@xmath8 and linear approximations are also imposed . for each thread
three different studies are done ; in increasing order of complexity they are : ( a ) thin tube approximation and no flow ; ( b ) thin tube approximation and flow ; and ( c ) full ideal mhd equations with flow .
while the thin tube approximation is well supported by the thinness of the threads compared to their length , ignoring the effect of the flow does not seem a good idea because the detected flow velocities are rather large .
nevertheless , as we describe below the results prove just the contrary . in the absence of flow ,
case ( a ) , transverse motions of the kind observed by @xcite can only be produced by the kink mode @xcite .
the dispersion relation of the kink mode in a magnetic tube containing prominence material can be deduced from equation ( 27 ) of @xcite .
@xcite conclude that , if no assumptions about the magnetic field , density and magnetic tube length ( @xmath9 ) are made , then this dispersion relation contains too many unknowns and can not be uniquely solved .
but if a value of @xmath9 is assumed ( with the restriction @xmath10 km from observations ) , then a dependence of the prominence versus the coronal alfvn speed can be obtained .
figure [ dr_hinode ] displays such a dependence for two of the six threads and several magnetic tube lengths .
this figure shows that , unless the ratio of prominence to coronal density is unrealistically low ( e.g. below 50 ) , the alfvn speed in the thread reaches a minimum value that is not modified by that density ratio .
this lower limit of @xmath11 is around 120 km s@xmath0 for thread # 6 and lies between 200 and 350 km s@xmath0 for the other five threads . to test the validity of this seismological result ,
@xcite consider case ( b ) , which involves solving a simple partial differential equation .
their numerical simulations of a thread that moves bodily along the magnetic tube yield in - phase oscillations of the whole thread , so a kink - like behaviour is recovered even in the presence of flow .
these results are in excellent agreement with the observations of @xcite . perhaps surprisingly , for all six threads
the period is only slightly smaller than that of case ( a ) , the difference being below 5% , i.e. smaller than the error bars of observations ( of the order of or larger than 6.8% ) .
hence , the period of these transverse oscillations is almost insensitive to the thread flowing motions .
finally , @xcite also investigate case ( c ) and recover the results of case ( b ) , which implies that the thin tube approximation is quite reasonable , as expected .
after these two studies the lower bounds for the alfvn speed derived from case ( a ) are firmly established and although this investigation does not lead to particular values of physical parameters , such alfvn speeds agree with the intense magnetic fields and large densities usually found in active region prominences .
wave leakage is a common phenomenon by which the energy imparted by a disturbance to a given structure ( e.g. a coronal loop or a solar prominence ) is not confined as standing oscillations of the structure , which emits waves carrying energy into the external medium .
this situation has been studied at length in relation with coronal loop oscillations ; see @xcite and references therein .
most of the works described in this section make no mention of wave leakage , but our interpretation is that the obtained attenuation of oscillations is simply a consequence of the emission of waves from the prominence into the corona . in the celebrated model of @xcite a prominence
is treated as a horizontal line current suspended in the corona .
such an approximation is justified by the fact that if the prominence plasma is supported against gravity by the lorentz force , then a current runs along the filament . in this work
the photosphere is substituted by a rigid , perfectly conducting plane with a surface current distribution caused by the coronal magnetic field generated by the filament current .
such induced currents in turn give rise to a coronal magnetic field whose effect on the prominence is to provide the lifting force necessary to counteract gravity . the full magnetic field arrangement in the corona
is determined by both the line current and the photospheric currents and can be of two types : either of normal polarity ( np ) or of inverse polarity ( ip ) ; see @xcite for more details .
this magnetic configuration was used by @xcite to study the oscillations of a filament .
it is worth to remark that in these works the prominence is treated as an infinitely thin and long line , so that it has no internal structure , although the interplay of the filament current with the surrounding magnetic arcade and photosphere is taken into account .
in addition , both np and ip structures were considered ; see figure [ fig_np_ip ] , in which the poloidal magnetic field is displayed ( the toroidal magnetic field component being zero ) .
the fundamental ingredient of these papers is that if a disturbance causes a displacement of the whole line current , that remains parallel to the photosphere during its motion , then the coronal magnetic field is reshaped and the photospheric surface current is modified .
this , in turn , influences the magnetic force acting on the filament current .
such a force can either strengthen the initial perturbation , so that the original equilibrium is unstable , or diminish it , so that the system is stable against the initial disturbance .
this simple description becomes more complicated when one takes into account that perturbations travel at a finite speed ( namely the alfvn velocity ) which causes time delays in the communication between the line current and the photosphere .
@xcite investigated the effect of these time delays on the filament dynamics . exponentially growing or decaying solutions
were found both for np and ip prominence models ; see figure 4 of @xcite and figure 5 of @xcite as examples .
an important conclusion that can be extracted here is that the attenuation is very efficient for many parameter values , and consequently the ratio of the damping time to the period is between 1 and 10 ( i.e. in agreement with observations ) .
this point is well illustrated by figure [ fig_leakage]a , in which the quality factor is depicted as a function of the alfvn speed .
different curves in this diagram correspond to different sets of parameters , with solid and dashed lines used to distinguish between ip and np configurations .
all the ip curves and one of the np curves in this figure take values below @xmath12 , equivalent to @xmath13 , which is an indication of the efficiency of the damping of oscillations . an ip magnetic configuration similar
to that used by @xcite was taken by @xcite , although in this study the prominence is not infinitely thin and instead is represented by a current - carrying cylinder .
@xcite carried out numerical simulations of the isothermal magnetohydrodynamic ( mhd ) equations and this implies that the equilibrium temperature is set to a constant value ( @xmath14 k ) everywhere .
the authors mention that , despite the large temperature difference between solar prominences and the corona , this isothermal assumption can only have a minor effect .
in addition , the photosphere is described as a perfectly conducting boundary , as in the papers examined above .
the inner part of the filament is disturbed by a perturbation with a velocity amplitude of 10 km s@xmath0 and at an angle of @xmath15 to the photosphere .
this causes the prominence to move like a rigid body in the corona , both vertically and horizontally , and to undergo exponentially damped oscillations ; see figure 7 in @xcite .
it is found that the horizontal and vertical motions are decoupled from one another ( and so can be investigated separately ) and that the period and damping time of horizontal oscillations are much larger than their respective counterparts for vertical oscillations .
again , strong damping can be achieved for some parameter values ; see figure [ fig_leakage]b .
it turns out that vertical oscillations are very efficiently attenuated for all the parameters considered in this work and that the same happens with horizontal oscillations for coronal densities above @xmath16 kg m@xmath17 .
these contrasting properties of damped horizontal and vertical oscillations contain some potential for performing seismology of prominences .
now we turn our attention to the interpretation of the results presented so far .
none of the works cited in the first part of this section , i.e. @xcite , mention wave leakage as the cause of the attenuation , but instead these papers connect this behaviour with the effect of the distant photosphere and the presence of time delays in the communication of disturbances between the filament and the photosphere . on the other hand , @xcite link the damping of oscillations to the emission of waves by the prominence : the damping of horizontal motions is attributed to the emission of slow waves , whereas fast waves are invoked as the cause of the damping of vertical motions .
given that the main difference between this work and the other ones lies essentially in the cross section of the filament , there is no reason to believe that the physics involved are much different regardless of the prominence being modeled either as a straight and infinitely thin current or as a current - carrying cylinder .
this issue must be examined in more detail to better understand the dynamics of solar prominences . a final remark about the exact nature of the wave leakage found by @xcite must be made . in this work ( see their figure 6a ) the plasma-@xmath8 in the prominence ranges from @xmath18 in its central part to @xmath19 at its boundary .
hence , waves emitted by the prominence into the corona propagate in a @xmath20 environment in which magnetic field lines are closed . under these conditions , slow modes propagate along magnetic field lines and are unable to transfer energy from the prominence into the corona and so wave leakage in the system studied by @xcite is only possible by fast waves . then , it is hard to understand how the prominence oscillations can be damped by the emission of slow waves in this particular model , in which the dense , cool plasma is only allowed to emit fast waves
. it must be mentioned , however , that the plasma-@xmath8 in the corona increases with the distance from the filament , which implies that the emitted fast waves can transform into slow waves when they traverse the @xmath21 region .
this effect has been explored by @xcite , but see also references therein for similar studies .
@xcite took into account some dissipative mechanisms and explored , through some order - of - magnitude calculations , their importance as possible damping agents . in this work
we are warned about the simplicity of the calculations and about the neglect of the solar corona , whose role on the damping of prominence oscillations is beyond this simple study .
dissipative mechanisms are separated into isotropic and anisotropic . amongst the first ones , viscosity and magnetic diffusion
are found to be rather inefficient in attenuating perturbations in only a few periods , as demanded by observations .
moreover , radiative losses ( here modeled by means of newton s cooling law , that gives an extremely simplified account of radiation in a solar prominence ) , may be of some importance , although their effect is difficult to establish unambiguously because of the presence of the unknown radiative relaxation time - scale in newton s cooling law .
therefore , radiation by the prominence plasma requires a more specific treatment to determine its significance .
regarding anisotropic mechanisms , in @xcite it is determined that , because of the very large density and very low temperature of prominences , viscosity can be considered isotropic and thus irrelevant as the cause of the observed damping of oscillations . in addition , thermal conduction is dominated by the presence of a magnetic field and is essentially parallel to the field direction .
this mechanism is very efficient as a damping agent for very short wavelengths .
since ballai s work has been extended in other papers discussed below , the meaning of `` very short wavelengths '' will become clear later .
although other damping mechanisms are not contemplated in @xcite , e.g. wave leakage , ion - neutral collisions or resonant absorption , this work is important since it allowed subsequent investigations to discard the irrelevant mechanisms and to concentrate only in the rest .
the efficiency of thermal conduction ( parallel to the magnetic field ) and radiation in transporting heat , and thus in causing the attenuation of oscillations , has been addressed in a number of papers ; see @xcite .
these works are concerned with the spatial and temporal damping of fast and slow waves propagating in a uniform medium .
radiative losses are simulated by equation ( 7 ) of @xcite , which contains two parameters that can be tuned to represent optically thin or thick radiation ; see @xcite .
different sources of plasma heating are also included , although this mechanism has a negligible consequence on the properties of the waves . to understand the results of these studies it is worth to consider the characteristic time - scales of thermal conduction and radiative losses ( @xmath22 and @xmath23 ) , that following e.g. @xcite can be defined as @xmath24 @xmath25 in these expressions @xmath26 , @xmath27 and @xmath28 are the plasma pressure , density and temperature
, @xmath9 is a characteristic length - scale ( the wavelength of oscillations , for example ) , @xmath29 the parallel thermal conductivity and @xmath30 and @xmath1 are the two parameters in hildner s radiative loss function .
we see that @xmath23 is independent of @xmath9 and so the rate at which the plasma looses energy through radiation does not vary with the wavelength or frequency of the waves . on the other hand , the conduction time - scale , @xmath22 , decreases linearly with @xmath9 and so conduction becomes more effective at transporting heat for short wavelengths , just as noticed by @xcite , for example . in the limit of very short wavelengths , conduction is so efficient that the isothermal regime is reached .
the critical wavenumber , @xmath31 , at which this transition occurs is given by @xcite as @xmath32 for slow and fast waves , respectively . here
@xmath33 is the number density , @xmath34 boltzmann s constant , @xmath35 and @xmath36 the sound and alfvn speeds and @xmath37 the angle between the wavenumber and the unperturbed magnetic field . to derive these formulas it is necessary to assume @xmath38 , together with @xmath39 for the slow mode and @xmath40 for the fast mode , where @xmath41 is the real part of the frequency and @xmath42 the component of the wavenumber parallel to the equilibrium magnetic field .
figure [ fig_nonadiab ] presents a plot of the characteristic conduction and radiative times together with the critical slow and fast mode wavenumbers marking the transition from the adiabatic to the isothermal regime .
now , let us consider a perturbation with tunable wavelength travelling in a magnetised , uniform medium in which thermal conduction and radiation can transport heat .
let us start with an extremely short wavelength , so that the wavenumber is larger than @xmath31 and the perturbation is isothermal ( hence , we are at the right of figure [ fig_nonadiab ] ) .
let us now increase the wavelength until the wavenumber becomes just smaller than @xmath31 .
now perturbations are no longer isothermal , but given that thermal conduction is still more efficient than radiation , since it works in a very short time - scale ( i.e. @xmath43 ) , the perturbation is dominated by conduction . next the wavelength is allowed to increase further and consequently the conduction time - scale increases proportionally to @xmath9 . under these conditions ,
conduction becomes less and less efficient and there is eventually a wavenumber ( or length - scale ) at which the conduction and radiation time - scales become equal ; it corresponds to the position in figure [ fig_nonadiab ] in which the solid and dashed lines cross .
this length - scale , @xmath44 , is obtained by imposing @xmath45 and with the help of equations ( [ tau_c ] ) and ( [ tau_r ] ) is @xmath46 finally , for wavelengths larger than @xmath44 radiative losses dominate the energy budget of the plasma
. these formulas are useful to explain some features of the damping time ( in the case of temporal damping of oscillations ) or the damping length ( in the case of spatial damping ) obtained by @xcite .
consider , for example , the damping time and the damping per period ( @xmath47 ) of the fast mode modified by radiative losses and thermal conduction in a uniform medium .
their variation with the wavenumber is displayed in figure [ fig_nonadiab2 ] , in which each of the three curves originates from a particular pair of values of @xmath30 and @xmath1 , the two parameters in the radiative loss function . in our discussion
we concentrate on the solid line , although similar conclusions can be deduced for the other two ( and also for the slow mode , which is not examined here ) .
all curves in figure [ fig_nonadiab2 ] display three changes of slope whose origin can be partly explained on the basis of the transitions between different regimes studied above . for the paramater values used in this work , we have @xmath48 m@xmath0 and @xmath49 m , that corresponds to the wavenumber @xmath50 m@xmath0 .
the right - most change of slope in figure [ fig_nonadiab2 ] is around @xmath51 and coincides roughly with the value of @xmath52 , so it is caused by the transition between the isothermal and the conduction - dominated regimes .
in addition , the intermediate change of slope is around @xmath53 and its agreement with @xmath54 confirms that it corresponds to the transition from the conduction - dominated to the radiation - dominated regimes .
therefore , the shift from one regime to another leaves its imprint in the variation of @xmath6 and @xmath7 with respect to the wavenumber .
we also note that , since the difference between the three curves of figure [ fig_nonadiab2 ] lies in the radiative loss function used , they overlap in the conduction dominated regime .
the main conclusions of @xcite regarding the damping of prominence oscillations are that the slow mode is strongly attenuated for wavenumbers within the range of observed values ( i.e. for @xmath55 in the range @xmath56 m@xmath0 ) , while the fast mode is not so strongly damped and its attenuation rate is too small compared with the values from observations . moreover , the considered approximations for optically thin or thick plasmas can give very different attenuation properties and all the considered heating mechanisms are of no relevance . the influence of the surrounding corona on the damping of magnetoacoustic oscillations via thermal conduction and radiative losses has been studied by @xcite .
in @xcite the filament is taken as a plasma slab embedded in the solar corona and with a uniform magnetic field parallel to the filament axis ; see figure [ fig_struct_along]a . both the prominence plasma and the coronal plasma are uniform and threaded by the same magnetic field . to understand the results of @xcite it is necessary to describe the main properties of the linear , adiabatic normal modes of such a configuration , that have been studied by @xcite . in the latter work
the solutions are classified in three types ( figure [ fig_modes_along ] ) : fast , internal slow and external slow modes and it is concluded that the perturbations associated to internal slow modes are essentially confined to the prominence slab , so that these waves are hardly influenced by the coronal environment . on the other hand ,
fast modes have tails that penetrate in the corona and as a result coronal conditions are important in determining the features of these waves .
moreover , the confinement of fast modes becomes poorer for small values of the wavenumber in the direction of the filament axis .
regarding external slow modes , they produce almost negligible perturbed amplitudes inside the cold slab and thus seem uninteresting in the context of prominence oscillations .
@xcite included the thermal conduction , radiative losses and plasma heating terms in the energy equation and studied the damping properties of the fast and internal and external slow modes .
they obtained values of @xmath7 in agreement with observations both for the internal slow and the fast modes ; this last result implies that the presence of the surrounding corona modifies the damping properties of fast waves . such as expected ,
the damping rate of the internal slow mode is not modified by the inclusion of the corona in the model and it only depends on the value of the wavenumber in the direction parallel to the magnetic field .
the attenuation of this mode can be recovered by inserting the parallel wavenumber , @xmath42 , in the expression @xmath57 and by substituting @xmath35 by a `` modified sound speed '' , @xmath58 , whose square is given by @xmath59 where @xmath60 and @xmath61 are determined from the partial derivatives of the heat - loss function with respect to density and temperature , respectively .
the damping time of the fast kink mode , however , is severely influenced by the addition of the coronal medium .
this can be clearly appreciated by comparing the damping time in a uniform medium ( figure [ fig_nonadiab2]a ) and in a structured medium with longitudinal magnetic field ( figure [ fig_fast_mode_along ] ) .
the reason for such a strong modification of the fast kink mode damping time is , firstly , that this mode couples with the external slow mode , whose properties are essentially dictated by the coronal plasma . and secondly , apart from this coupling the coronal medium has a direct influence because of the presence of coronal tails in the perturbed variables . in @xcite a close examination of the role of the corona in the damping of the fast mode
was done by removing one of the important non - adiabatic mechanisms ( thermal conduction or radiation ) either in the filament or in the corona and then computing the period and damping time .
the relevance of each mechanism is revealed by a discrepancy between the damping time computed with and without this mechanism ( figure [ fig_fast_mode_along ] ) .
the conclusion of this study is that coronal mechanisms govern the attenuation of the fast modes for wavenumbers @xmath62 m@xmath0 , whereas prominence mechanisms prevail for @xmath63 m@xmath0 . in each of these ranges radiation
is dominant in the low-@xmath42 range and thermal conduction is dominant in the high-@xmath42 range , such as found in section [ non - adiab - unif - medium ] for a uniform medium .
more specifically , coronal radiation controls the attenuation for @xmath64 m@xmath0 , coronal conduction does so for @xmath65 m@xmath0 @xmath66 m@xmath0 , while prominence radiation and conduction are the most important non - adiabatic mechanisms for @xmath67 m@xmath0 @xmath68 m@xmath0 and @xmath69 m@xmath0 , respectively .
hence , in the range of observed wavelengths both thermal conduction in the corona and radiation from the prominence gas are important to explain the reported damping times . in section [ non - adiab - unif - medium ] the characteristic time - scales of thermal conduction and radiation in a uniform medium were invoked to explain the transition from the radiation - dominated wavenumber range to the conduction - dominated wavenumber range .
the same idea can be applied to explain this transition in the present , structured configuration both for the prominence and for the corona . in the case of the prominence ,
the transition value is the same as before , namely @xmath70 m@xmath0 , whereas for the corona it is @xmath71 m@xmath0 .
these numbers coincide well with the transitions at which different mechanisms dominate in figure [ fig_fast_mode_along ] .
moreover , the wavenumber at which the transition between the ranges of dominance of coronal conduction and prominence radiation takes place can be derived from the equality of @xmath22 in the corona and @xmath23 in the prominence .
this calculation yields @xmath72 m@xmath0 , also in good agreement with the value from figure [ fig_fast_mode_along ] .
more details can be found in @xcite . in @xcite a prominence - corona configuration with an equilibrium magnetic field perpendicular to the prominence slab
is considered ; see figure [ fig_struct_along]b . such a structure can support fast and slow modes of internal or external character , depending on whether their properties are governed mainly by the prominence or the corona , respectively .
moreover , there is another type of fast and slow mode , called string by @xcite and hybrid by @xcite , that is simultaneously internal and external .
@xcite find that slow modes are efficiently damped by the considered non - adiabatic mechanisms ( thermal conduction and radiative losses ) , such as found by @xcite for a perpendicular orientation of the magnetic field .
it is also shown that prominence radiation dominates the attenuation of internal slow modes and that prominence radiation and coronal conduction together dominate that of the hybrid slow mode .
this result for the hybrid mode is coherent with the fact that its perturbations achieve large amplitudes both in the prominence and the corona , so one expects that the most relevant damping mechanisms of each medium govern together the attenuation of the hybrid mode . on the other hand ,
fast modes are very poorly damped in the present configuration , which implies that the assumed magnetic field orientation is decisive for these modes .
a remarkable feature of fast modes is that they can become unstable under certain circumstances .
thermal conduction allows energy transfer between the prominence slab and the coronal medium .
prominence radiation has an essential role in dissipating the extra heat injected from the corona and stabilises the oscillations .
nevertheless , thermal instabilities appear if radiative losses from the prominence plasma are omitted or significantly reduced ( e.g. caused by an increase of the optical thickness ) since the plasma can not dissipate the extra injected heat in an efficient way .
given that different parametrisations of the radiative loss function have been given in the literature @xcite and that they correspond to different optical thicknesses , the relevance of the optical thickness in the stability of fast waves is not firmly established yet .
material flows are common features in solar prominences @xcite .
@xcite investigated the attenuation of the slow and thermal waves by radiative losses and thermal conduction in a uniform medium with flow . in this work the fast mode is ignored because these non - adiabatic effects do not contribute in a significant way to its attenuation @xcite ; slow and thermal waves are investigated in the limit of field - aligned propagation .
@xcite find that the period and damping per period can show a strong dependence on @xmath73 , the flow speed , for certain values of this parameter ; figure [ fig_nonadiab_flow ] presents their results for the slow mode .
they also find that the greatest period and damping per period of slow waves is obtained for flow speeds close to the real part of the non - adiabatic sound speed , which in the presence of flow is deduced from equation ( [ nonadiab_sound_speed ] ) with @xmath74 substituted by @xmath75 .
regarding the thermal mode , which in the absence of flow does not propagate , it is found that it becomes propagating for @xmath76 .
therefore , when flows and time damped oscillations are observed in a prominence / filament , this wave should be also taken into account as potentially responsible of the observed damped oscillations
. these authors remark that it may not be possible to determine whether an observed period and damping time are associated to a slow wave or a thermal wave in a flowing material and that a possibility to determine the origin of such wave is to resort to the temperature perturbation , which should be rather large for the thermal wave and much smaller for the slow wave . instead of considering the attenuation of perturbations in a full - prominence configuration , as in section [ nonadiab_slab ] , @xcite
restrict themselves to a single thread .
their model consists of a homogeneous , infinite cylinder representing the thread embedded in an unbounded corona ; see figure [ fig_nonadiab_thread]a .
two additional ingredients of the model are a uniform magnetic field both inside and outside the thread and flow motions parallel to the magnetic field in both regions .
thermal conduction and radiative losses are taken into account as damping mechanisms , and the combined effect of these non - ideal physics and the steady flow on the attenuation of oscillations is assessed . in this work it is found that , in the absence of flow , slow modes are efficiently damped by non - adiabatic effects , while fast kink waves are in practice not attenuated , since their damping times are much larger than typical lifetimes of filament threads .
therefore , the damping by non - adiabatic mechanisms of kink oscillations is much less efficient in this cylindrical case than in the slab geometry ( figure [ fig_nonadiab_thread]b ) . in the presence of flow
it is well - known that the symmetry between waves propagating parallel or antiparallel to the flow is broken .
the thermal mode behaves as a wave that propagates parallel to the flow , and its motions are mainly polarized along the longitudinal direction . nevertheless , this oscillatory behaviour is not likely to be observed in practice due to its quick damping .
the damping time of both slow and thermal waves is not affected by the flow . on the contrary , for realistic values of the flow velocity , the larger the flow , the larger the attenuation of parallel fast kink waves , whereas the contrary occurs for antiparallel fast kink solutions . nevertheless , this effect is not enough to obtain realistic damping times in the case of fast kink modes .
chromospheric and prominence plasmas are not fully ionised and so several new effects are present in comparison to a fully ionised plasma .
for example , electric charges are frozen to magnetic field lines , but neutrals are not and thus the neutral and ionised fractions of the plasma behave differently .
collisions between neutrals , on one hand , and electrons and ions , on the other hand , arise and consequently a modified ohm s law is obtained . as a consequence ,
the main outcome of the interaction between neutrals and charged particles is the presence of enhanced joule heating and enhanced magnetic diffusion .
these ideas were presented by @xcite in connection with the heating of the solar chromosphere and corona and have also been invoked to explain the dynamics of chromospheric spicules ( e.g. * ? ? ?
* ; * ? ? ?
* ) and of flux tubes emerging from the solar interior @xcite .
the influence of partial ionisation of the prominence plasma on the properties of the fast and slow modes has been investigated by @xcite . in this work
the one - fluid mhd equations for a partially ionised plasma made of electrons , protons and neutral hydrogen are derived .
next , a uniform medium with a straight magnetic field is considered .
the linear regime is assumed and this results in the vanishing of the joule heating term in the energy equation , so that the corresponding physical effect is absent .
one may conclude that in more realistic configurations dissipation may be stronger than that found by @xcite .
the main result of this work is that the fast mode can be strongly damped ( in agreement with observations ) for small values of the ionisation fraction ( figure [ fig_pip ] ) , whereas the slow mode is hardly attenuated .
the mechanism responsible for the damping of the fast mode is cowling s diffusion , that vanishes in a fully ionised plasma .
an analytical approximation for the fast mode damping rate , @xmath77 is derived by @xcite , with @xmath78 the relative density of neutrals , @xmath27 the gas density , @xmath79 the magnetic field strength and @xmath55 the modulus of the wavenumber .
such a simple expression can be of practical use in the interpretation of observations .
it indicates that the damping is stronger , i.e. @xmath7 is smaller , for almost neutral plasmas with low density , strong magnetic field and for short wavelengths .
a natural extension of @xcite , concerned with the attenuation of fast and slow waves by thermal conduction and radiation , and @xcite , concerned with the attenuation of the same waves by ion - neutral effects , is to consider together a partially ionised plasma with thermal conduction and radiative losses .
this task was performed by @xcite , who derived the equations governing the dynamics of such medium .
a particularly interesting issue is the different conduction properties of electrons ( mainly along magnetic field lines ) and neutrals ( isotropic ) , which implies that the contribution of the neutral species to the thermal conduction must be taken into account separately following @xcite .
@xcite find that the period of the magnetoacoustic waves remains basically the same as in the ideal case and that their model gives values of the ratio of the damping time to the period similar to the ones obtained in observations for the three wave modes ( fast , slow and alfvn ) .
in addition , the inclusion of non - adiabatic terms in the partially ionised set of equations increases the damping of fast and slow waves in the interval of observed wavelegths as compared with the results obtained for a non - adiabatic fully ionised plasma @xcite and an adiabatic partially ionised plasma @xcite . for slow waves ,
the minima of the ratio of @xmath7 , corresponding to attenuation maxima , are displaced towards longer wavelengths as compared to when only thermal conduction and radiative losses are included in the model .
an increase of the neutral portion in the plasma produces a displacement of these ranges of maximum damping to longer wavelengths ( figure [ fig_pip_nonadiab]a ) .
radiative losses are dominant for long wavelengths ( just as found by * ? ? ?
* ) , while the rest of the wavenumber interval is dominated by thermal conduction and ion - neutral collisions .
regarding fast waves , radiation is the dominant damping mechanism for long wavelengths , while in the rest of the considered wavenumber interval the attenuation is dominated by the effect of ion - neutral collisions .
an important result is that fast waves only exist for wavenumbers smaller than a critical value that depends on the ionisation fraction ( see figure [ fig_pip_nonadiab]b ) . in spite of this , the critical wavenumber is large in comparison with the typical wavenumbers of waves in prominences .
finally , for typical prominence temperature values , the contribution of electrons to thermal conduction is negligible in front of the contribution of neutrals .
the plasma density varies by about two orders of magnitude between a filament thread and the surrounding corona . in such a highly inhomogeneous configuration
fast kink modes can be efficiently damped by transferring their energy to alfvn modes through resonant absorption .
@xcite examine the relevance of this mechanism in the attenuation of small amplitude prominence oscillations by modeling the magnetic and plasma configuration of an individual fibril as an infinitely long , straight , cylindrically symmetric flux tube of mean radius @xmath80 surrounded by a coaxial layer of thickness @xmath81 ( figure [ fig_thread_model ] ) . the @xmath82 approximation is used .
the coronal and thread densities are denoted by @xmath83 and @xmath84 , respectively , and their ratio as @xmath85 .
this parameter is much larger than unity and probably of the order of 100 or larger . in this simple model , the magnetic tube representing the thread
only contains cool material , i.e. the presence of hot material inside the tube is ignored . taking into account that the observed propagating waves in threads have a wavelength , @xmath86 , much larger than the tube radius , @xcite start by using the long - wavelength approximation to make some analytical progress .
next , they use the so - called thin boundary approximation ( @xmath87 ) to derive the following expression for the ratio of the damping time to the period @xmath88 where @xmath89 is a numerical factor that depends on the particular variation of the density in the nonuniform layer : for a linear variation @xmath90 @xcite , while for a sinusoidal variation @xmath91 @xcite .
a simple substitution in this formula of a density ratio @xmath92 and @xmath93 , which corresponds to a transitional layer whose thickness is 10% the thread radius , yields @xmath94 , in good agreement with observations .
then , this analytical approximation seems to indicate that the process of resonant absorption is an efficient damping mechanism for fast mhd oscillations propagating in these structures .
next , @xcite solve numerically the full set of linear , resistive , small - amplitude mhd wave equations for the transverse kink oscillations . in this manner ,
the approximations that lead to equation ( [ eq_res_abs ] ) are avoided .
these numerical calculations confirm the analytical estimate that , for the typical large filament to coronal density contrast , the mechanism produces rapid damping on timescales of the order of a few oscillatory periods only .
they also reveal that for thin layers ( @xmath95 ) the inaccuracy of the long - wavelength approximation produces differences of up to @xmath96% for the combination of short wavelength with high - contrast thread . for thick layers , differences of the order of 20%
are obtained ( in agreement with * ? ? ?
moreover , the authors conclude ( as shown by figure [ fig_res_abs ] ) that the obtained damping rates are only slightly dependent on the wavelength of perturbations and that the damping rate becomes independent of the density contrast for large values of this parameter .
this last result has two seismological consequences .
first , the observational determination of the density contrast is less critical than it is in the low - contrast regime , the one corresponding to coronal loops .
second , according to seismic inversion results that combine the theoretical and observed periods and damping times @xcite , high - density thread models would be compatible with relatively short transverse inhomogeneity length scales . in a subsequent work , @xcite note that in a @xmath97 plasma with prominence conditions embedded in the much hotter corona the fast kink mode can also undergo a resonant coupling to slow continuum modes .
hence , it is argued , the combined effect of the two resonant couplings may lead to an enhanced attenuation rate of the kink mode .
the structure considered by @xcite is a thread identical to that of figure [ fig_thread_model ] , although the @xmath82 assumption is removed from their calculations and , consequently , the plasma pressure is finite . such as in @xcite , in this work
a combination of analytical expressions and numerical analysis is used .
the analytical formulas not only allow to predict the positions of the slow and alfvn resonances , but they also provide the contribution of each of the two mechanisms to the damping time , @xmath98 and @xmath99 , whose ratio is @xmath100 with @xmath101 and @xmath102 the azimuthal and longitudinal wavenumbers and @xmath35 and @xmath36 the sound and alfvn speeds . using @xmath103 and a typical value for the wavelength ( @xmath104 ) results in @xmath105 , which means that the resonant coupling of the kink mode with the alfvn continuum produces an attenuation rate that is @xmath106 times smaller than that coming from the coupling with the slow continuum .
these results are also verified from the numerical computations ( see figure [ fig_slow_cont ] ) , which confirm that resonant absorption with the slow continuum is irrelevant in the damping of the observed transverse thread oscillations .
although the study of small - amplitude waves and oscillations in prominences constitutes a discipline that might complement the direct determination of prominence parameters by providing independent values based in the comparison between observations and theory , so far prominence seismology is more a promise than a reality because of the large gap between observation and theory .
the analysis presented in section [ seismology ] is a good example of a seismological application to prominences , but , unfortunately , it is the only one of its kind .
such a gap arises because of the little restrictions imposed by observational works ( which sometimes reduce to reporting the period of the detected oscillations ) and the simplicity of theoretical studies ( which neglect the intricate nature of prominences and substitute it by a very simplified physical model ) .
nevertheless , these two sides are coming together as the complexity of works increases .
in this paper the possible role of several physical mechanisms in the attenuation of prominence oscillations has been reviewed .
it has been shown that both the fast and slow waves can undergo strong damping under a variety of physical conditions and with the intervention of different effects .
nevertheless , the list of the presumably relevant mechanisms has not been totally explored and other effects should be investigated .
in addition , the geometry of the models is too simplified , which implies that the results derived so far only give a rough guide to the damping features in actual prominences .
the efficiency of wave leakage and coronal conduction in draining oscillatory energy from the prominence and transferring it to the corona calls for the study of complex configurations in which the structure of the full prominence - corona system is treated more realistically .
most of the works are based on the linear approximation , whose validity seems quite robust in view of the small amplitude of prominence oscillations in many cases .
nevertheless , apart from the so - called small - amplitude prominence oscillations , that in general affect small regions of a prominence and have doppler velocity peaks typically below 12 km s@xmath0 , there is a different kind of phenomenon , the large - amplitude oscillations ( tripathi et al . 2009 , this volume ) .
they are characterised by a much higher amplitude ( with oscillatory velocities up to 90 km s@xmath0 ) and disturb the whole filament ; see @xcite .
these prominence vibrations also damp very rapidly , in a few periods , so it is in order to ascertain whether the mechanisms at work in the damping of small - amplitude oscillations are also the main ones for the large - amplitude counterparts or whether non - linear effects can also cause the attenuation of the latter .
the author acknowledges the international space science institute for the financial support received to attend a meeting of the team `` coronal waves and oscillations '' in bern .
financial support received from the spanish micinn and feder funds , under the grant mec aya2006 - 07637 , and from the conselleria deconomia , hisenda i innovaci of the government of the balearic islands , under grant pctib-2005-gc3 - 03 , is also acknowledged . | quiescent prominences are thin slabs of cold , dense plasma embedded in the much hotter and rarer solar corona .
although their global shape is rather irregular , they are often characterised by an internal structure consisting of a large number of thin , parallel threads piled together .
prominences often display periodic disturbances mostly observed in the doppler displacement of spectral lines and with an amplitude typically of the order of or smaller than 23 km s@xmath0 , a value which seems to be much smaller than the characteristic speeds of the prominence plasma ( namely the alfvn and sound velocities ) .
two particular features of these small amplitude prominence oscillations is that they seem to damp in a few periods and that they seem not to affect the whole prominence structure .
in addition , in high spatial resolution observations , in which threads can be discerned , small amplitude oscillations appear to be clearly associated to these fine structure constituents .
prominence seismology tries to bring together the results from these observations ( e.g. periods , wavelengths , damping times ) and their theoretical modelling ( by means of the magnetohydrodynamic theory ) to gain insight into physical properties of prominences that can not be derived from direct observation . in this paper
we discuss works that have not been described in previous reviews , namely the first seismological application to solar prominences and theoretical advances on the attenuation of prominence oscillations . |
it is recognized that the corotation resonance where the rotation velocities of galactic matter and density waves , responsible for spiral arms coincide , plays an important role in the galactic evolution and the value of the corotation radius is one of the fundamental galactic parameters . in literature
one can find several methods of this quantity deriving in our galaxy , for instance , by means of analysis of arms geometry , stellar or open clusters kinematics , hi and hii emissions ( e.g. , lin et al .
1969 ; marochnik et al .
1972 ; mishurov & zenina 1999 ; lpine et al .
2001 ; fernandez et al . 2001 ; dias & lpine 2005 , etc . ) . in the present paper ,
we propose a new approach to evaluation of the corotation radius in our galaxy .
the method is based on the statistical analysis of oxygen radial distribution in the galactic disc ( notice that martin & roy 1995 and scarano et al .
2010 also mentioned the importance of the corotation effects in analysis of abundance gradient in external galaxies ) .
oxygen was used since it is mainly produced by sne ii which are strongly concentrated in spiral arms .
hence it is the most pure indicator of spiral arms influence on the formation of radial abundance pattern in the galactic disc . as an observed material ,
we use the data on oxygen distribution derived by andrievsky et al .
( 2002 a , b , c ) and luck et al .
( 2003 , 2006 ) over cepheids .
being bright and very young objects with precise distances , these stars give reliable information about abundances of heavy elements , close to the one in interstellar medium , in the significant part of the galactic disc .
the main finding of the above papers is that the radial distribution of metallicity in the galactic disc is _ bimodal _
, i.e. there is a rather steep gradient in the inner part of the disc at @xmath4 kpc and a plateau - like distribution for @xmath5 kpc and up to about 10 kpc ( the solar galactocentric distance @xmath6 kpc ) .
hence , there is a bending in the slope of the distribution at @xmath7 kpc .
the above fine structure in the radial abundance distribution is very important .
it indicates that in the galactic disc the distribution formation is caused by some non - trivial process .
mishurov et al .
( 2002 ) , acharova et al .
( 2005 ; 2010 ) developed a theory of spiral arms influence on the radial distribution of oxygen in the galactic disc .
they show that the bending in the slope of oxygen distribution is associated with the corotation resonance
. however , in our previous papers we fitted the theory to the observations `` by eyes '' .
now we propose a statistical method for the deriving of the corotation resonance location by means of analysis of oxygen distribution along the galactic radius .
simultaneously the so - called `` constant for the rate of oxygen synthesis '' is estimated .
let @xmath8^{ob}(r)$ ] be an observed distribution of any @xmath9 element along the galactic radius @xmath10 ( as usual @xmath8 = log(n_x / n_h)_s - log(n_x / n_h)_{\odot}$ ] , where @xmath11 is the number of @xmath9 element or hydrogen atoms in the object , the first item refers to a star located at the distance @xmath10 , the second one - to the sun ) . on the other hand ,
let us assume that @xmath8^{th}(r)$ ] is a theoretical distribution of the corresponding element which depends on some vector @xmath12 , the coordinates of the vector being the sought - for free parameters of our theory .
to fit the observations we minimize the variance @xmath13 over @xmath12 : @xmath14^{ob}_i - [ x / h]^{th}_i)^2,\ ] ] here the summation is taken over all @xmath15 points of galactocentric radius where the abundances of the element were measured , @xmath16 is the number of observational data , @xmath17 is the number of the sought - for free parameters .
unfortunately , @xmath8^{th}(r)$ ] can not be represented by any analytical or , at least , approximating formula .
the theoretical abundance is derived as a numerical solution of a system of equations , the solution being dependent on a particular set of the free parameters . in our approach
, there are 2 target free parameters ( @xmath18 ) .
to find them the following method is used : we solve numerically the equations of oxygen synthesis varying the coordinates of @xmath12 - vector within some regions and compute a net of @xmath19 .
after that we construct the surface @xmath20 as a function of the two free parameters and find the minimum ( @xmath21 ) which determines the best set of the sought - for parameters . to estimate the free parameters errors the confidence region
was constructed : according to draper & smith ( 1981 ) the confidence contour is determined by the intersection of the surface @xmath19 with the horizontal plane at level : @xmath22,\ ] ] where @xmath23 is fisher s @xmath23-statistics and the 95% level of confidence was adopted .
the equations which describe the evolution of oxygen synthesis in the galactic disc are as follows ( details see in acharova et al . 2005 and 2010 ) : @xmath24 @xmath25 @xmath26 where @xmath27 are surface densities correspondingly for interstellar gas and stars , @xmath28 is the density of oxygen in interstellar medium , @xmath29 is the fraction of oxygen ( hence @xmath30 = log ( z / z_{\odot}$ ] ) ) , @xmath31 is the star formation rate ( sfr ; @xmath32 , @xmath33 is a normalizing coefficient ) , @xmath34 is salpeter s initial mass function with the exponent of - 2.35 ( stellar masses @xmath35 are in solar units ) , @xmath36 is the rate of oxygen synthesis , @xmath37 is the infall rate of intergalactic gas on to the galactic disc , @xmath38 is the radial velocity of gas within the galactic disc averaged over the galactic azimuth , @xmath39 is time ( in gyr ) , @xmath40 is the life - time of a star of mass @xmath35 on the main sequence , @xmath41 , @xmath42 , @xmath43 is the mass of stellar remnants ( white dwarfs , neutron stars , black holes ) .
the infall rate of intergalactic gas on to the galactic disc is described as @xmath44 with the radial scale @xmath45 kpc ( marcon - uchida et al .
2010 ) which is an intermediate value between @xmath46 and @xmath47 kpc usually used in galactic nuclear synthesis modeling ( e.g. , naab & ostriker 2006 ; fu et al . 2009 ; schnrich & binney2009 ) .
constants @xmath48 and @xmath49 are defined below .
we made experiments with various abundances of the infall gas from @xmath50 to @xmath51 . in accordance with
lacey & fall ( 1985 ) our final abundances weakly depend on the exact value of @xmath52 if @xmath53 .
below we demonstrate the results for @xmath50 which is slightly less than the mean content of heavy elements in halo stars ( @xmath54 , prantzos 2008 ) .
the enrichment rate of interstellar medium by oxygen is represented as follows : @xmath55 , where @xmath56 is the mass ( in solar units ) of ejected oxygen per one snii explosion ( tsujimoto et al .
1995 ) , @xmath57 is the rate of sneii events @xmath58 the factor @xmath59 describes the influence of spiral arms on radial oxygen pattern formation and is represented by the expression : @xmath60 , where @xmath61 is the angular rotation velocity of galactic disc , @xmath62 is the angular rotation velocity of spiral density waves ( recall that whereas the galactic matter rotates differentially , i.e. @xmath61 is a function of @xmath10 , galactic density waves rotate as a rigid body , @xmath63 , the corotation radius @xmath0 is determined from equation @xmath64 ) , @xmath65 is the constant of the rate of oxygen synthesis , @xmath66 is a cut - off factor : @xmath67 if @xmath10 is contained between the inner and outer lindblad resonances and @xmath68 otherwise . in our computations ,
we use the rotation curve based on clemens ( 1985 ) data adjusted for the scale @xmath69 kpc . the last term in parentheses in equation ( 5 )
describes the turbulent radial diffusion of heavy elements . to derive the diffusion coefficient @xmath70 we use the simple gas kinetic approximation ( mishurov et al . 2002 ) .
the above system of equations may be divided into 2 groups : _
i ) _ equations ( 3,4 ) which describe galactic disc formation , and _
ii ) _ the `` chemical '' equation ( 5 ) of oxygen synthesis .
the target 2 free parameters , @xmath62 and @xmath65 , enter only the equation ( 5 ) .
hence , the first group of equations can be solved independently of the last equation .
therefore , the algorithm of solving the above equations is as follows . at first step ,
independently of equation ( 5 ) we solve equations ( 3,4 ) by means of lelevier method ( potter 1973 ) and find the time evolution of the radial profiles of gaseous and stellar densities , @xmath71 and @xmath72 ( simultaneously sfr , @xmath73 is derived ) . to perform this part of modelling
, we have to specify a value of the time - scale of gas infall on to the galactic disc , @xmath48 , and adopt an expression for @xmath74 .
after that , the constants @xmath33 and @xmath49 , entering these group of equations , are obtained by fitting the normalising conditions that at present epoch ( @xmath75 gyr - the age of galactic disc ) and solar position stellar and gaseous densities have to be equal to : @xmath76 m@xmath77 and @xmath78 m@xmath77 ( haywood et al . 1997 ) .
taking into account that the part of stellar remnants is of the order of 10% of @xmath79 ( our experiments confirm this supposition ) we have that at present time and solar galactocentric distance summary density has to be about 54 m@xmath80 pc@xmath81 - a value close to marcon - uchida et al .
however , below we also discuss other normalising conditions .
at the second step , we solve the chemical equation ( 5 ) for a set of @xmath62 and @xmath65 ( as boundary conditions we demand that the solution has to be finite at the galactic centre and the galactic disc edge which we dispose at @xmath82 kpc ) . repeated this procedure we find the minimum of @xmath83 , hence estimate the best values for @xmath62 and @xmath65 , then compute their errors as it was described in sec .
2 . at last
we derive the location of the corotation resonance @xmath0 .
in what follows we examine 2 limiting cases for @xmath48 : @xmath84 gyrs _ rapid disc formation _ , and @xmath85 gyrs _ slow disc formation_. for the radial gas flow 3 models were studied .. ] \1 ) no radial flow of gas : @xmath86 ( _ models m20 and m70 _ correspondingly for the above two rates of disc formation ) .
\2 ) gas radial inflow : _ case a ) _ @xmath87 km s@xmath88 , if @xmath89 kpc and @xmath90 km s@xmath88 , if @xmath91 kpc ( _ models m2a and m7a _ ) ; _ case b ) _ @xmath92 km s@xmath88 ( _ models m2b and m7b _ ) ; \3 ) gas radial outflow : @xmath93 km s@xmath88 ( _ models m2o and m7o _ ) . in fig.1 evolution of gaseous radial profiles with time is shown .
in significant part of galactic radii current gaseous density distributions happen to be close to the observed ones and profiles derived by portinari & chiosi ( 2000 ) , fu et al .
( 2009 ) , naab & ostriker ( 2006 ) and schnrich & binney ( 2009 ) . comparing the distributions
we can not prefer a certain model only on the basis of density profile ( models m2o and m7o have holes in the galactic centre but they do not describe oxygen radial distribution well ) . ; _ dashed - dotted line _ - for @xmath94 ; _ dashed line _ - for @xmath95 ; _ solid line _ for @xmath96 gyr .
] so , we need some additional arguments which enable us to make a choice among various scenarios of galactic disc formation .
such information comes from observations : according to sancisi & fraternali ( 2008 ) and bregman ( 2009 ) current rate of total ( integrated over the galactic disc ) gaseous mass falling on to the galactic disc is of the order of 0.1 - 0.2 m@xmath80 yr@xmath88 whereas the total star formation rate is @xmath97 m@xmath80 yr@xmath88 ( robitaille & whitney 2010 ) although one can meet a higher level of sfr @xmath98 m@xmath80 yr@xmath88 ( hartmann 2007 ) .
the above values may be considered as global normalisations which are complementary to the particular ones at solar position . in table 1
we give the computed total rate of forming stars , @xmath99 , and infall gas mass , @xmath100 at the present epoch . from the table
one can see that only models of rapid galactic disc formation ( @xmath3 gyr ) demonstrate low current infall rate , @xmath101 m@xmath80 yr@xmath88 , and sfr at level @xmath102 m@xmath80 yr@xmath88 , the both quantities fit the above observed values .
scenarios of slow disc formation ( @xmath103 gyr ) lead to greater current sfr , @xmath104 m@xmath80 yr@xmath88 , but they require the infall rate of the order of magnitude higher than that observed by sancisi & fraternali ( 2008 ) and bregman ( 2009 ) . obviously , one should be prudent in such estimates since future observations could discover such a huge amount of baryonic matter falling on to the galactic disc .
but at the moment we do not have such data , at least for our galaxy .
.constants @xmath49 ( m@xmath80 pc@xmath81 gyr@xmath88 ) and @xmath33 ( m@xmath105 pc gyr@xmath88 ) ; integrated rates of gas infall , @xmath106 , and star formation rate , @xmath107 ( m@xmath80 yr@xmath88 ) and minimal @xmath108 derived in various models + [ cols="<,^,^,^,^,^,^,^,^",options="header " , ] let us look at the problem from another angle .
there is a belief that star formation is mainly fuelled by the gas infall on to the galactic disc .
hence , the total rates of gas infall and star formation should be close to each other in order to the interstellar gas not to be consumed due to star formation .
but as we wrote above , at present epoch in our galaxy the total star formation rate is at least 5 times higher than the one of gas infall . to explain this problem , in fig .
2 we show evolution of the ratio of gas infall rate on to the galactic disc to star formation rate , @xmath109 , with time as a function of @xmath10 for 4 models . from the figure
it is seen that , at each moment of time there is a region in the galactic disc where @xmath110 .
this means that the interstellar gas is not fully consumed in the galaxy ( from fig .
1 one can also see that the gaseous density is not equal to zero at any galactocentric radius ) .
besides , as it was expected , in scenarios with gas radial inflow the region , where @xmath110 , is shifted closer to the galactic centre .
indeed , gaseous density decreases with galactocentric radius , star formation rate also decreases with @xmath10 .
hence , due to interstellar gas radial inflow the region with depressed star formation rate moves closer to the galactic centre and this results in the ratio @xmath109 increase .
in addition gas radial inflow replenishes the inner part of the galactic disc with building material for star formation fuelling .
so , as we can see , at each epoch of galactic evolution the reserve of gas forms into the galactic disc and it fuels star formation process . . here
a reserve of gas is formed for the future star formation . ]
how should we change the normalisations of the current gaseous and stellar densities at solar distance in order to obtain the global star formation rate at level , say , @xmath111 m@xmath80 yr@xmath88 ?
our experiments give the following results . in scenario of rapid galactic disc formation ( @xmath112 gyr ) and
no gas radial inflow , gaseous and stellar densities happen to be unrealistically large : @xmath113 m@xmath80 pc@xmath81 ; @xmath114 m@xmath80 pc@xmath81 albeit the rate of gas infall @xmath115 m@xmath80 yr@xmath88 ( in case of gas radial inflow like in models `` b '' @xmath11660 m@xmath80 pc@xmath81 ; @xmath117
m@xmath80 yr@xmath88 however @xmath118 keeps at very high level @xmath119 m@xmath80 pc@xmath81 ) . in scenario of slow galactic disc formation ( @xmath120 gyr ) and no gas radial inflow
we have : @xmath121 m@xmath80 pc@xmath81 ; @xmath122 m@xmath80 pc@xmath81 but the rate of gas infall @xmath123 m@xmath80 yr@xmath88 is of the order of magnitude higher than the observed value ( if we take into account gas radial inflow like in models `` b '' we derive : @xmath124 m@xmath80 pc@xmath81 ; @xmath125 m@xmath80 pc@xmath81 but @xmath126 m@xmath80 yr@xmath88 again is too high ) .
perhaps such situation can be implemented in some galaxies but it is not supported by observations for our galaxy .
there is one additional argument in favour of rapid disc formation scenario : according to allende prieto ( 2010 ) and zasov & silchenko ( 2010 ) galactic discs were formed rapidly , at least last 8 gyr their sizes did not change significantly .
despite the conclusion made in the previous section below we give the results for all the above models .
so , at the next step we solve the chemical equation ( 5 ) , compute the variance @xmath13 and find its minimum which determines the best values of @xmath62 and @xmath65 ( to catch the minimum of the variance we varied @xmath62 and @xmath65 in a wide region : @xmath127 km s@xmath88 kpc@xmath88 ; @xmath128 gyr . ) .
as an observed material we use the data of oxygen distribution along the galactic radius , derived over cepheids , which were averaged within bins of 0.5 kpc width ( see acharova et al . 2005 ; 2010 ) . the corresponding distribution is drawn in fig . 3 at step 0.25 kpc in the region from 5.25 up to 10 kpc ( in total , the number of observed points is @xmath129 ) .
the bars in the figure are the rms scatters of the mean abundance within the bin . and @xmath65 , _ dashed lines _ for ( @xmath130 ) and ( @xmath131 ) .
] to illustrate our method , in fig .
4 , we show an example of the surface @xmath132 for model m20 . in case of m20 . for better visualisation
we draw the figure `` bottom up '' .
the best estimates for the parameters are : @xmath133 km s@xmath88 kpc@xmath88 ; @xmath134 gyr . ] to estimate the errors of the sought - for free parameters , we compute the section of the above surface by the horizontal plane at @xmath135 ( see equation ( 2 ) , in our case the value of fisher s statistics happens to be @xmath136 , draper & smith 1981 ) . the section is demonstrated in fig . 5 by the shaded figure for model m20 .
its ellipse - like border delineates the confidence region .
the semi - axes of the ellipse give the errors .
the results of variance computations are shown in table 1 .
consider them in more detail . from fig . 4 by the horizontal plane @xmath137
this region determines the confidence one at probability level 95 % ( see sec .
its border gives the estimate of the parameters errors : @xmath138 ; @xmath139 km s@xmath88 kpc@xmath88 . ] first of all , by means of fisher s statistics analysis the significance of gas radial inflow inclusion into consideration can be estimated . to do that we have to compare models `` a'',``b '' and `` o '' with the ones without gas radial inflow separately for the above two scenarios of galactic disc formation .
the result is as follows .
_ i _ ) in scenario of slow disc formation the best model is m7b .
other models have significantly higher values of @xmath108 and they are to be rejected . _ ii _ ) in scenario of rapid disc formation two models , m20 and m2b , have the same values of @xmath140 , other ones should be rejected .
however , we can not decide between m20 and m2b : the null hypothesis ( @xmath86 ) can not be rejected . in other words , in this scenario inclusion of the gas inflow is insignificant .
so , below we discuss two models - m20 and m7b .
let us compare them and test the null hypothesis that the corresponding variances do not differ significantly against the alternative one that they do .
for this , according to martin ( 1971 ) we should check if the value @xmath141 lies in the range : @xmath142 ( @xmath143 is the significance level and we adopt that the degrees of freedom are approximately the same in the both alternatives ) . from table 1
we derive @xmath144 , whereas for @xmath145 @xmath146 ( draper & smith 1981 ) .
hence , the above inequality is satisfied and we conclude : the difference between models m20 and m7b is undetected statistically ( on the basis of available observational data of oxygen radial distribution and a rather simple theoretical approach ) . at the moment ,
the only way to discriminate between them is as it was discussed in sec .
4 . finally
, our estimates for the target parameters are as follows : \1 ) _ m20 _ : @xmath147 km s@xmath88 kpc@xmath88 ( the corotation radius @xmath148 km s@xmath88 kpc@xmath88 ) ; @xmath149 gyr .
\2 ) _ m7b _ : @xmath150 km s@xmath88 kpc@xmath88 ( the corotation radius @xmath151 km s@xmath88 kpc@xmath88 ) ; @xmath152 gyr . in fig .
3 the best theoretical oxygen distributions for the above two models and the distributions for the upper and lower values of @xmath62 and @xmath65 are shown . from the figure
it is seen that our theory explains the bimodal radial distribution of oxygen in the disc of our galaxy rather well .
in the present paper , a new method for evaluation of the corotation radius in the galaxy is developed .
our approach is based on a statistical analysis of the bimodal structure of oxygen radial distribution in the galactic disc determined over cepheids . by means of treatment of observational data
we derive that the corotation resonance happens to be situated at @xmath153 kpc depending on the rate of intergalactic gas infall on to the galactic disc ( the statistical error is @xmath1540.3 kpc ) .
the above value for the corotation radius is close to the solar galactocentric distance @xmath69 kpc .
simultaneously the constant for the rate of oxygen synthesis in the galactic disc @xmath155 gyrs was obtained .
we also argue in favour of a short time - scale formation of the galactic disc , namely : @xmath3 gyrs .
this scenario enables to solve the problem of the lack of intergalactic gas infall , i.e. the very low present integrated rate of gas infall on to the disc , @xmath156
m@xmath80 year@xmath88 observed by sancisi & fraternali ( 2008 ) and bregman ( 2009 ) , whereas the integrated star formation rate is expected to be @xmath157 m@xmath80 year@xmath88 ( robitaille & whitney 2010 ) . higher level of the current global star formation rate , @xmath111 m@xmath80 year@xmath88 , needs too high rate of gas infall on to the galactic disc or unrealistically high gaseous and stellar current densities at the solar galactocentric distance .
authors thank to referee for vary important comments , to profs . yu.shchekinov and a.zasov for helpful discussions . the work was supported in part by grants
no . 02.740.0247 and p685 of federal agency for science and innovations .
iaa thanks to the russian funds for basic research , grant no .
11 - 02 - 90702 . | we propose a new method for determination of the rotation velocity of the galactic spiral density waves , correspondingly , the corotation radius , @xmath0 , in our galaxy by means of statistical analysis of radial oxygen distribution in the galactic disc derived over cepheids . the corotation resonance happens to be located at @xmath1 kpc , depending on the rate of gas infall on to the galactic disc , the statistical error being @xmath2 kpc .
simultaneously , the constant for the rate of oxygen synthesis in the galactic disc was determined .
we also argue in favour of a very short time - scale formation of the galactic disc , namely : @xmath3 gyr .
this scenario enables to solve the problem of the lack of intergalactic gas infall .
[ firstpage ] galaxy : fundamental parameters galaxy : abundances galaxies : spiral galaxies : star formation . |
grb 970828 is one of the first grbs with an observed afterglow @xcite and a determined redshift of @xmath8=0.9578 from the identification of its host galaxy @xcite .
this source is still presenting today , after 15 years from its observations , an extremely rich problematic in the identification of its astrophysical nature .
the x - ray afterglow was discovered by the asca satellite 1.17 days after the grb trigger @xcite and the data showed the presence of a feature around @xmath9 5 kev , maybe associated to a radiative recombination edge of a h- or he - like ionized iron @xcite .
particularly interesting was also a variable and large intrinsic absorption column which was found from the asca : the absence of an optical afterglow @xcite , the large intrinsic absorption column detected in the x - ray data @xcite and the contemporary detection in radio - wavelengths of the grb afterglow , imply a very large value for the circum - burst medium ( cbm ) , this variable absorption might be an indication of a strong inhomogeneous cbm distribution .
@xcite described the presence of an evolving thermal component in the first 40 s of the emission of grb 970828 , as observed by the batse detector on board the compton gamma ray telescope @xcite .
this thermal emission was associated to the photospheric grb emission of a relativistic expanding fireball @xcite , and from the observed properties of the blackbody spectrum it was inferred the bulk lorentz gamma factor of the expanding plasma , @xmath10 .
this first thermal component is followed by a second emission episode , which starts @xmath9 50 s after the grb trigger .
this second episode was neglected in the analysis of @xcite , where they explicitly stated to `` neglect here late - time episodes of engine activity that occur after @xmath9 25 and @xmath9 60 s in this burst '' .
we have recently proposed that some grbs , namely grb 090618 @xcite and grb 101023 @xcite , can be described by two sharply different emission episodes . a first episode that is characterized by a thermal component in the early phases of its emission , which originates from an expanding source , not related to the grb emission .
we have shown that this component is characterized by an expansion of the blackbody emitter , whose origin is associated to the latest phase of the life of a massive star . in the late phase of its evolution ,
this massive star is well - described by a bare core that has lost its hydrogen envelope and most or entirely the helium one , during the previous stages of its evolution .
this core is then observed to expand at non - relativistic velocities . at the end of the first episode ,
the collapse to a black hole occurs and the actual grb emission happens . since this first episode occurs prior to the grb emission and to the formation of a black hole , we have referred to this as the proto - black - hole ( pbh - @xmath11 from the ancient greek that means anteriority in space and in time ) . in this work
we describe first the pbh component that we find in the first episode of grb 970828 light curve .
we then proceed to the description of the second episode , which in this work represents the authentic grb emission .
the grb emission is well explained in the context of the fireshell scenario , see e.g. @xcite for a complete review of the model .
we find for the grb an @xmath12-plasma total energy of @xmath1 erg , a baryon load @xmath2 and a lorentz gamma factor at transparency of @xmath3 . from the numerical simulations we confirm the presence in the cbm of large average particle density , of the order of @xmath4 part/@xmath5 , and dense clouds with @xmath7 .
we show that this can explain the `` dark '' nature of this grb and the observed large and variable column absorption in the asca x - ray afterglow .
these results are in very good agreement with the analysis presented in @xcite and represents a further confirmation of the presence of a pbh in the early emission of grb 970828 .
grb 970828 was discovered with the all - sky monitor ( asm ) on board the rossi x - ray timing explorer ( rxte ) on 1997 august 28th @xcite . within 3.6 hr
the rxte / pca scanned the region of the sky around the error box of the asm burst and detected a weak x - ray source @xcite .
grb 970828 was also observed by the burst and transient source experiment ( batse ) and the grb experiment on ulysses @xcite .
the batse - lad light curve is characterized by two main emission phenomena , see fig.[fig:4b.0 ] : the first lasts about 40 s and is well described by two main pulses , the second one is more irregular , being composed by several sharp pulses , lasting other 40 s. the x - ray afterglow observations started at about 0.1 days after the trigger time and continued up to 7 - 10 days .
it is well described by a double power - law behavior , but a possible resurgence of soft x - ray emission , together with marginal evidence for the existence of fe lines , was reported by beppo - sax and asca @xcite . the optical observations , which started about 4 hr after the burst , did not report any possible optical afterglow for grb 970828 up to @xmath13=23.8 @xcite . however , the observations at radio wavelengths of the burst position , 3.5 hrs after the initial burst , succeeded in identifying a source at a good significance level of 4.5 @xmath14 @xcite inside the rosat error circle ( 10 ) .
the following deep searches for a possible optical counterpart of this radio source led to the identification of an interacting system of faint galaxies , successively recognized as the host galaxy of grb 970828 .
the spectroscopic observations of the brightest of this system of galaxies led to the identification of their redshift , being @xmath8=0.9578 .
the lack of an optical transient associated with the afterglow of grb 970828 can be explained as due to the presence of strong absorption , due to dusty clouds in the burst site environment , whose presence does not affect the x - ray and the radio observations of the grb afterglow . to analyze in detail this grb ,
we have considered the observations of the batse - lad detector , which observed grb 970828 in the 25 - 1900 kev energy range , and then we have reduced the data by using the rmfit software package . for the spectral analysis we have considered the high energy resolution burst ( herb ) data , which consist of 128 separate high energy resolution spectra stored during the burst emission .
the light curve was obtained by using the medium energy resolution ( mer ) data , which consist of 4.096 16-channel spectra summed from triggered detectors .
[ tab:4b.1 ] [ cols="<,^,^,^,^,^,^,^",options="header " , ] 10@xmath15 g and 10@xmath16 cm in radii . the black line corresponds to the average value for the particle density.,title="fig:",width=340,height=226 ] + from the observed fluence and the values obtained from the spectral fits , we have computed the energy emitted in the p - grb by using the schaefer formula of the k - correction @xcite and the redshift of the source : it is @xmath17 erg .
this value of the p - grb energy is only 9 @xmath18 10@xmath19 the isotropic energy of the second episode , which we can assume as equal to the energy @xmath20 of the @xmath12-plasma of the grb emission . in this way , with the knowledge of the temperature at transparency and of the p - grb energy , we can derive the value of the baryon load from the numerical solutions of the fireshell equations of motion .
these solutions for four different values of the total e@xmath21e@xmath22-plasma energy are shown in the fig .
4 of @xcite .
we find that the baryon load is @xmath2 , which corresponds to a lorentz gamma factor at transparency @xmath3 .
the grb emission was simulated with very good approximation by using a density mask characterized by an irregular behavior : all the spikes correspond to spherical clouds with a large particle density @xmath23 part/@xmath5 , and with radius of the order of @xmath6 cm , see fig .
[ fig:4b.5 ] . considering all the clouds found in our analysis ,
the average density of the cbm medium is @xmath24 particles/@xmath5 .
the corresponding masses of the blobs are of the order of @xmath25 g , in agreement with the clumps found in grb 090618 .
the results of the fireshell simulation are shown in table [ tab:4b.3 ] while the simulated light curve and spectrum are shown in fig .
[ fig:4b.4 ] . + + ( blue line ) and @xmath26band ( red line ) light curve of sn 1998bw transposed at the same redshift of grb 970828 .
the purple and cyan line represents the limit given in the deep images by @xcite and @xcite respectively.,width=226,height=340 ] the analysis of grb 090618 @xcite and grb 101023 @xcite represents an authentic `` rosetta stone '' for the understanding of the grb - sn phenomenon .
the possible presence of a supernova emission , observed ten days after the burst in grb 090618 @xcite but not spectroscopically confirmed , is at the same time very intriguing in light of the still unknown grb - sn connection @xcite .
we have transposed the data of the `` bump '' rc - band light curve observed in the optical afterglow of grb 090618 , associated to the presence of a possible underlying supernova @xcite , to the redshift of grb 970828 .
this simple operation concerns only the transformation of the flux observed , under the assumption that the sn has the same intrinsic luminosity : the variation of the magnitude only depends on the square of the difference between the luminosity distances .
moreover , we have also transposed the @xmath27 and @xmath26band light curves of sn 1998bw @xcite , which is the prototype of a supernova associated to a grb . from the @xmath28 correction transformation formula ,
the @xmath29band light curve , transposed at @xmath8=0.9578 , corresponds approximately to the observed @xmath26band light curve , so in principle we should consider the @xmath30 nm transposed light curve as the actual one observed with the @xmath31 nm optical filter .
these transposed light curves are shown in fig .
[ fig:4b.6 ] .
we conclude that the possible supernova emission could have been seen between 20 and 40 days after the grb trigger , neglecting any possible intrinsic extinction .
the optical observations were made up to 7 days from the grb trigger , reaching a limit of @xmath32 @xcite and subsequent deeper images after @xmath33 days @xcite , so there are no observations in this time interval
. however , the @xmath13-band extinction value should be large since the observed column density from the x - ray observations of the grb afterglow is large as well @xcite : the computed light curve for the possible sn of grb 970828 should be lowered by more than 1 magnitude , leading to a sn bump below the @xmath13 = 25.2 limit , see fig .
[ fig:4b.6 ] .
the presence of very dense clouds of matter near the burst site might be the cause of these absorptions .
indeed we find the presence of clouds in our simulation at the average distances of @xmath34 cm from the grb progenitor , with average density of @xmath4 part/@xmath5 and typical dimensions of @xmath6 cm , see fig .
[ fig:4b.5 ] . the possibility to observe the energy distribution from a grb in a very wide energy range
, thanks to the newly space missions dedicated , has allowed to definitely confirm the presence of two separate emission episodes in grbs associated to sne .
future missions , as the proposed loft , will allow to observe the thermal decay from these objects down to @xmath35 kev .
the large area detector , designed for the loft mission , could be also able to observe in timing mode the emission from the newly born neutron star ( ns ) from supernovae ib / c , catching a possible characteristic signature of the neo - ns @xcite , shedding more light on the sn - grb enigma .
a general conclusion is therefore appropriate : it is mandatory to proceed to a detailed spectral analysis in parallel to an identification of the physical and astrophysical nature of the source .
this has been made explicitly clear in the analysis of grb 090618 @xcite .
the crucial role of grb 090618 has been to evidence the existence of a new family of grbs , associated to sne , composed of two distinct emission episodes .
this has been made possible thanks to different reasons : 1 ) the large number of detectors , fermi , swift bat and xrt , agile , suzaku - wam , konus - wind and coronas photon rt-2 , that allowed to study that grb in a very wide electromagnetic energy range , from 0.3 kev to 40 mev ; 2 ) the very large intensity emitted from grb 090618 , @xmath36 erg ; 3 ) the low redshift of the grb , @xmath8 = 0.54 .
although both episodes radiate in the hard x - ray energy range , the first episode originates from a non - relativistic expansion process , having an expansion velocity of @xmath9 1000 km / s , see fig .
[ fig:4b.2 ] , and occurring prior to the formation of a black hole ; the second one is the one occurring at the formation of the black hole and is the authentic grb emission .
of course , the understanding of these two components can not be done only on a phenomenological basis , but it requires an understanding of the underlying physical and astrophysical processes .
we verify in this work that indeed grb 970828 is a member of this family .
the new understanding leads to a wealth of information on the evolution of the thermal component and of the radius of the blackbody emitter , given by eq .
( 1 ) , in the first emission episode , see figs .
[ fig:4b.1 ] , [ fig:4b.2 ] .
similarly , for the second episode we obtained the cbm structure , see fig .
[ fig:4b.5 ] , the details of the simulation of the lightcurve and the spectrum of the real grb emission , see fig .
[ fig:4b.4 ] .
we have also shown in table [ tab:4b.2 ] the final results of the grb simulation , the total energy of the @xmath12 plasma , the baryon load @xmath37 , the temperature of the p - grb @xmath38 and the lorentz gamma factor at transparency @xmath39 , as well the average value of the cbm density @xmath40 and the density ratio of the clouds @xmath41 .
finally , from the analogy with the optical bump observed in grb 090618 , see fig .
[ fig:4b.6 ] , associated to a possible sn emission @xcite , we have given reasons why a possible sn associated to grb 970828 would not have been observable due to the large interstellar local absorption , in agreement with the large column density observed in the asca x - ray data @xcite and with the large value we have inferred for the cbm density distribution , @xmath42 particles/@xmath5 .
we thank j. rafelski for useful discussions about the writing of this manuscript .
we also thank z. cano for providing the data of the optical bump observed in the afterglow of grb 090618 .
hd , gp and avp acknowledge the support for the fellowship awarded for the erasmus mundus irap phd program . , r. , bernardini , m. g. , bianco , c. l. , et al .
2005 , in american institute of physics conference series , vol .
782 , xith brazilian school of cosmology and gravitation , ed .
m. novello & s. e. perez bergliaffa , 42 | following the recent theoretical interpretation of grb 090618 and grb 101023 , we here interpret grb 970828 in terms of a double episode emission : the first episode , observed in the first 40 s of the emission , is interpreted as the proto - black - hole emission ; the second episode , observed after t@xmath0 + 50 s , as a canonical gamma ray burst .
the transition between the two episodes marks the black hole formation .
the characteristics of the real grb , in the second episode , are an energy of @xmath1 erg , a baryon load of @xmath2 and a bulk lorentz factor at transparency of @xmath3 .
the clear analogy with grb 090618 would require also in grb 970828 the presence of a possible supernova .
we also infer that the grb exploded in an environment with a large average particle density @xmath4 part/@xmath5 and dense clouds characterized by typical dimensions of @xmath6 cm and @xmath7 .
such an environment is in line with the observed large column density absorption , which might have darkened both the supernova emission and the grb optical afterglow . |
vortices are simplest to understand for an interaction energy that has a single manifold , @xmath85 , of ground states and no hierarchy . the order parameter must move into @xmath85 far away from any spin texture .
vortices are classified by the topology of the circuit traced out in @xmath85 by the order parameter on a large circle containing a vortex or set of vortices .
although there are many ways to add wiggles to a given circuit , only the topological structure is important , leading to a discrete set of possible vortex charges .
[ fig : iron ] shows how a circuit may be tangled with holes in the space @xmath85 .
as a vortex evolves , such a circuit can evolve only into other circuits that are tangled in the same way ; thus the vortex charge is conserved .
this generalizes circulation - conservation in a single - component condensate : the circulation quantum number @xmath1 for a set of vortices is also the number of times that the wave function at infinity winds around the circle that minimizes the mexican hat potential@xcite .
the tangling in a multicomponent condensate can be described by a group of generalized winding numbers around the ground state manifold @xmath85 , @xmath86 , the fundamental group " ( see fig . [
fig : iron ] ) .
the order parameter at infinity of a vortex winds around a loop @xmath17 in the order parameter space @xmath13 .
realistic order parameter spaces are usually more symmetrical . ] besides the conservation of topological charge , two further properties of vortices follow from the geometry of the internal ground - state space .
first of all , the _ net winding _ or topological charge " of a set of vortices can be found by multiplying their charges together , using the definition of multiplication for the fundamental group .
( in a general space , the fundamental group has a multiplication operation defined by splicing circuits together . )
this rule is the generalization of adding the @xmath1 s of the individual vortices in the scalar order - parameter case .
( the fundamental group is noncommutative so one has to multiply the charges together in the right order , see appendix [ app : catstring ] . ) secondly , the _ energy _ of a vortex can be estimated as a function of the winding behavior at infinity . for a vortex
that _ minimizes _ this energy , the order parameter travels along a geodesic in @xmath85 .
since the interaction energy @xmath87 is constant ( and equal to its minimum @xmath69 ) at infinity , the energy of a vortex is determined by the elastic energy , that is , the cost of variations in @xmath40 as a function of the azimuthal angle @xmath71 .
the closed loop traced out by @xmath88 will relax to make this energy - cost small .
when it shrinks as much as is possible without leaving @xmath85 , it becomes a _
geodesic _ in @xmath85 .
the energy of the vortex is related to the length @xmath89 of this geodesic by @xmath90 where @xmath75 and @xmath91 are the radii of the condensate and the vortex core .
( note that replacing @xmath89 by @xmath0 gives the standard expression for a vortex in a scalar condensate . )
a small magnetic field introduces hierarchies into the order parameter space , leading to spin textures that wind around one manifold at an intermediate length scale and around a smaller manifold far away . classifying vortex topologies .
a ) the combined topology of a set of vortices can be found be seeing how the order parameter changes around a closed loop parameterized by @xmath71 .
the loop is drawn with a small gap between @xmath92 and @xmath93 to indicate that @xmath94 and @xmath95 vary continuously only between @xmath96 and @xmath97 .
the topological charge is defined as the jump in the values of @xmath98 and @xmath99 across the gap .
b ) vortices that can not exist by themselves in the presence of a magnetic field can still sometimes appear in clusters .
the charges @xmath100,@xmath101 in one of the clusters can involve rotations around arbitrary axes , but the combined charge @xmath102 has to use the direction of the magnetic field for its rotational axis to avoid a large energy - cost . ] let us consider how nested vortices hold themselves together in a continuous
spin texture .
" we will try to give a general argument showing how topology and the energy - hierarchy of the order parameter spaces implies that any texture is made up of a set of composite vortices which in turn are made up of clusters of nearly point - like vortices .
the order parameter subspaces in order of increasing energies are the two - dimensional @xmath11 , the four - dimensional @xmath13 and the @xmath103 dimensional @xmath15 ( for spin @xmath104 ) , with corresponding energy scales @xmath105 ( see eq .
( [ eq : effective ] ) ) and @xmath106 ( see eq .
( [ eq : hamiltonian ] ) ; as long as the density of the condensate does not vary , @xmath47 rather than @xmath107 sets the energy scale ) .
the condensate can move out of the ground state into one of the higher - energy subspaces if forced to by topology , spending less space in the manifolds with the greater energies .
( this is like saving eggnog with all that fat!for one week out of the year.)for a particular subspace @xmath108 in this sequence , in which @xmath109 , the size @xmath110 of the regions where @xmath40 is in @xmath108 is typically @xmath111 , where @xmath112 is the elasticity of the condensate .
this relation can be guessed at without understanding anything about the field configuration in such a region ; just assume equipartition between kinetic and interaction energies , so that @xmath113 is equal to @xmath114 , a typical scale for the kinetic energy density , @xmath115 . consequently , the composite vortex cores in fig .
[ fig : cherimoya ] , where @xmath40 is in @xmath19 , have diameters @xmath116 where the energy scale is taken from eq .
( [ eq : effective ] ) and the component cores , where @xmath40 varies through @xmath20 , have typical diameters @xmath117 . the more complete version of the equipartition argument , given in sec .
[ sec : molecules ] , estimates the total energy of the vortex as follows : the kinetic energy outside the composite core is about @xmath118 , according to eq .
( [ eq : chocolatepudding ] ) , the magnetic field energy inside the core is about @xmath119 . minimizing the sum , which has the form @xmath120 , gives @xmath121 .
the analogous argument also applies to vortices in a single - component condensate , giving the core size @xmath122 . because the composite cores in a cyclic condensate have some substructure ( the component vortices ) , the actual coefficient @xmath123 of the logarithm for the cyclic condensate will turn out to be less than @xmath124 once the kinetic energy of the vortices inside the core is included ( see section [ sec : enindex ] ) .
if @xmath6 is very small , the component cores are much smaller than the composite vortices , and may be regarded as points .
the wave function can therefore be approximated by a field that always stays in @xmath13 except at singularities " corresponding to the component cores : @xmath125 where @xmath126 is the spinor defined in eq .
( [ eq : order3 ] ) . over distances
much greater than @xmath22 even the composite cores seem to dwindle to points , suggesting the following approximation : @xmath127 these expressions generalize@xcite the phase - only approximation " for ordinary superfluids@xcite ; for example , eq .
( [ eq : manifold ] ) parameterizes elements of the space @xmath13 in terms of the symmetries of the @xmath14 hamiltonian , rotations @xmath128 , and rephasings , @xmath129 , which are applied to a representative state .
vortices are points which look like singularities at the level of resolution exposed by one of these approximations , although what one identifies as vortex singularity depends on which level of resolution one uses !
composite vortices result from the order parameter being forced out of the minimum - energy space @xmath11 by topology : if the order parameter goes around a hole in @xmath11 on some circle in the condensate , then inside the circle , the order parameter has to leave @xmath11 . if eq .
( [ eq : alphax ] ) continued to hold all the way to the center of the circle , then there would be a singularity since the angular variables @xmath107 and @xmath99 would run rapidly through multiples of @xmath97 .
so we can think of the circle as the core of an extended vortex in the field in eq .
( [ eq : alphax ] ) .
the singularity is actually filled in by a field of tetrahedra with higher - energy orientations , described by eq .
( [ eq : manifold ] ) , just as the wave function of a vortex in a single - component condensate avoids singularities by vanishing in the core . in the filled - in region of freely - oriented tetrahedra , there are also some elementary point "- vortices , actually spread out over the distance @xmath91 .
these result when the order parameter gets tangled around holes in @xmath13 , and then has to move outside of _ this _ space .
these cores are smaller than the composite core because of the large energy scale associated with @xmath13 .
the fields of these vortices can be described qualitatively by setting @xmath14 since kinetic energy surpasses the anisotropy energy very close to the cores .
the cores of the component vortices are filled to almost the same density @xmath34 as the rest of the condensate ( the spin - independent repulsion @xmath107 favors uniformity since @xmath130 ) .
vortices with this property are often referred to as
coreless , " but we use the word core " in a more general way , so that every vortex can have one . instead of defining core as a region where the _ density _ vanishes , our definition identifies a core as a region where the wave function departs from any particular form .
( in the core of a
coreless " vortex in a ferromagnetic condensate , the order parameter becomes polar rather than ferromagnetic . )
the core of the component tetrahedral vortices is the region where the order parameter leaves @xmath13 .
using the absolute ground state space @xmath11 leads to a different definition of vortex cores of field - aligned vortices : the core is where eq .
( [ eq : alphax ] ) breaks down .
this is why we call the region surrounding the tetrahedral vortices a composite core . "
the ordinary definition of a core arises when we focus on another space , @xmath131 , the sphere defined by @xmath132 .
( this space minimizes the largest of the interaction terms , @xmath133 . ) from the perspective of @xmath131 , a core would be a region where the wave function vanishes to avoid being discontinuous ; however , in a spinor condensate with @xmath130 , the wave function never has to vanish because a closed loop can not get snagged in the surface of the simply - connected sphere @xmath131 .
now we have to classify both the topologies of the point vortices ( or tetrahedral charges " ) and the composite vortices ( or aligned charges " ) , using the method described e.g. in @xcite .
for the point vortices we must parameterize the symmetry group that generates the composite core spin textures by a _ simply connected _
group @xmath134 .
we will take @xmath135 where multiplication is defined by @xmath136 .
we parameterize @xmath134 as follows , @xmath137 allowing us to regard @xmath134 as a ( redundant ) description of the symmetries of the hamiltonian . using this representation , one can define the net vortex charge inside any closed loop @xmath138(see fig . [
fig : plumpudding]a ) ; one simply expresses the wave function along the loop in the form @xmath139 where @xmath71 parameterizes the loop ( @xmath140 ) . in order for eq .
( [ eq : rays ] ) to be continuous when the circuit is closed , @xmath141 , or @xmath142 it follows that the classifying group element " or net topological charge " inside the curve @xmath143 is a symmetry describing the net rotation and rephasing around the closed loop .
( let us also use the briefer notation @xmath144 for the net rotation . ) since the tetrahedron has only @xmath145 symmetries ( see sec .
[ sec : namingofcats ] ) , this makes for a tractable classification of the topologies .
these topologies form the group of tetrahedral charges , multiplied together according to @xmath136 . when @xmath6 is very large then eq . ( [ eq : alphax ] ) has to apply basically everywhere and when @xmath6 is small then eq . ( [ eq : alphax ] )
has to apply outside of bound clusters of vortices . in order to classify the topology of vortices in the former case or of vortex clusters in the latter ,
we need a simply connected group that parameterizes the @xmath16 symmetry group @xmath146 ; we take @xmath147 and use the mapping @xmath148 .
then we write @xmath149 and define @xmath150 where the subscript
@xmath151 indicates that the magnetic field favors the 3-fold symmetric orientation illustrated in fig .
[ fig : bottler]a ( since we are assuming @xmath26 ) , and the rotation through @xmath152 is understood to be around the field axis . the possible values for @xmath102 can be referred to as field - aligned charges " since they describe the topologies of vortex - fields in which the tetrahedra keep the orientation favored by the magnetic field .
now let us consider the form of the fields near a tetrahedral vortex core .
near vortex @xmath153 , the spin texture will be rotationally symmetric , and given by @xmath154 where @xmath71 is now the azimuthal angle @xmath71 centered at this vortex . because the tetrahedra near this vortex may be tilted , we use @xmath155 , a _
generic _ member of the cyclic order parameter space .
the vortex is azimuthally symmetric since the parameters @xmath156 are constants .
the rotation axis @xmath157 is a local symmetry axis for the possibly _ tilted _ tetrahedra . to reduce the possible symmetries to a finite set
, we should relate the spinor @xmath158 to the spinor @xmath126 corresponding to the tetrahedron as oriented in fig . [
fig : bottler]a . if we write @xmath159 for an appropriate rotation @xmath75 and phase @xmath160 , the vortex in eq .
( [ eq : lathes ] ) can be written @xmath161 where @xmath162 and we have used the transformation rule for angular momentum : @xmath163 ( the right - hand side uses
the @xmath164 matrix @xmath165 associated with the rotation @xmath75 . )
continuity implies that the rotation axis @xmath166 is one of the finitely many symmetry axes illustrated in fig .
[ fig : bottler]a ; furthermore , according to the above scheme , the group element that classifies this vortex is @xmath167 .
there are only a discrete set of possible charges when we use @xmath166 , the axis relative to the body axes of the tetrahedron rather than @xmath157 , the axis relative to the lab coordinates .
( [ eq : lathes2 ] ) expresses the vortex as a product of a constant matrix @xmath168 ( the phase @xmath160 is unimportant ) and a standardized vortex configuration . the transformation @xmath168 rotates the standardized configuration in spin space , changing both the rotation axis ( from @xmath166 to the local axis @xmath157 ) and the orientation of the tetrahedra .
in the distant surroundings of a cluster of vortices , the spin texture will again have a rotationally symmetric form .
the group element describing the change in the order parameter as one tours the loop @xmath138 enclosing the entire cluster is given by @xmath169 this algebraic law has a few consequences .
first , it leads to a conservation law that constrains vortex alchemy : the net charge @xmath170 has to be conserved as the vortices inside the loop combine and metamorphose , since the topology on the loop can not change suddenly .
( there is only a discrete set of vortex charges because charges are defined using the body coordinates . )
second , eq .
( [ eq : berlin ] ) restricts the types of vortices which can form a cluster when @xmath16 .
( see fig .
[ fig : plumpudding]b . ) outside the cluster , the tetrahedra must be aligned with the field ( see eq .
( [ eq : alphax ] ) , so the charge is described by a field - aligned order parameter @xmath171 .
( [ eq : berlin ] ) requires that @xmath172 and @xmath173 .
the group elements must multiply to form a rotation about the @xmath23-axis to avoid the large zeeman cost outside the clusters , or composite cores " of the vortex molecules .
let us now review the example in the previous section : the vortex molecule was made out of two vortices of type @xmath174 ; since @xmath175 , the spin texture can align with the magnetic field outside the pair of vortices and @xmath176 .
this uses the fact that , in @xmath7 , all @xmath97 rotations are equal to @xmath177 , where @xmath178 is the @xmath179 identity matrix .
let us first assign names to the _ tetrahedral _ charges corresponding to closed loops in @xmath13 . at @xmath14 or within vortex clusters where @xmath6 can be neglected ,
the topological charges are described by a pair @xmath180 ( defined in eq .
( [ eq : wrongdivision ] ) ) where @xmath181 and @xmath182 .
note that for the phase of the order parameter to be continuous , as eq .
( [ eq : netcloses ] ) requires , the allowed values of @xmath99 and @xmath98 must be correlated@xcite , so only a discrete sequence of phases may accompany a given rotational symmetry .
also note that in @xmath7 , rotation angles are defined modulo @xmath183 rather than @xmath97 .
the twelve symmetries of the tetrahedron according to the ordinary method of counting become @xmath145 because , e.g. , a clockwise @xmath184 rotation around an axis is distinguished from the counterclockwise @xmath84 rotation .
this is more than a technical point : the vortex where the order parameter rotates through @xmath185 about an order three axis can not deform continuously into one where the order parameter rotates through @xmath84 , and it has more energy as well .
on the other hand , the @xmath186 vortex " can relax continuously to a state free of vortices . the necessity of using @xmath7 instead of @xmath164 is the biggest surprise to come out of the topological theory . since @xmath98 must be a symmetry of the tetrahedron corresponding to @xmath24 , illustrated in fig . [
fig : bottler]a , we can describe @xmath98 by indicating its rotation axis using the labels from the figure .
we refer to the minimal rotation around a given axis using just the label of the axis ; hence @xmath187 refer to the rotations through @xmath84 counterclockwise as viewed from the tips of the corresponding arrows , and @xmath29 refer to counterclockwise rotations through @xmath64 , about @xmath29 .
rotations through larger angles can be written as powers of these rotations .
therefore @xmath188 is a @xmath184 rotation ; also @xmath189 since @xmath76 rotations around any axis correspond to @xmath177 in @xmath7 .
we find it convenient to describe each rotation as a rotation through an angle @xmath107 around some axis , where @xmath190 .
( positive and negative @xmath107 s correspond to counterclockwise and clockwise rotations respectively . )
an arbitrary rotation angle can be replaced by an angle in this interval using the fact that @xmath191 rotations in @xmath192 are equivalent to the identity .
for example , the @xmath193 counterclockwise rotation @xmath194 is the same as the @xmath185 clockwise rotation @xmath195 because @xmath196 , a rotation through two full turns , corresponds to the identity of @xmath192 . on the other hand ,
the @xmath184 counterclockwise rotation @xmath188 is not equivalent to the clockwise @xmath84 rotation around the same axis , since the corresponding @xmath192 matrices differ by a minus sign .
now we can list all the pairs of rotations and phases which are allowed by the continuity condition , eq .
( [ eq : netcloses ] ) .
the possible vortices according to ref .
@xcite are @xmath197 and @xmath198 where @xmath1 and @xmath81 are integers , as well as the corresponding vortices with @xmath75 replaced by @xmath199 , @xmath102 , or @xmath200 and @xmath65 replaced by @xmath52 or @xmath201 .
one can work out explicit expressions for the @xmath192 elements corresponding to given rotations .
for example , let us find the @xmath192 element corresponding to @xmath65 ; since the rotation angle is @xmath64 , @xmath202 where @xmath203 is the @xmath65-axis . note that the tetrahedron has its vertices at @xmath204,@xmath205 , @xmath206 , @xmath207 .
the @xmath65 axis bisects the segment connecting the last pair of points ; the midpoint of these two points is @xmath208 the unit vector @xmath203 is obtained by normalizing this vector , so @xmath209 hence @xmath210 the net charge , eq .
( [ eq : berlin ] ) , of a set of vortices results from multiplying the matrices @xmath98 for the vortices of the set .
the result can be identified as one of the rotations @xmath211,@xmath212,@xmath213 , etc .
this procedure completely determines the @xmath192 product element , whereas the geometric method of applying the appropriate sequence of rotations to a tetrahedron does not determine the _ sign _ of the @xmath192 matrix .
we will describe a vortex reaction with the following notation , @xmath214 each factor describes the charge of an individual vortex-_atom _
, rather than a cluster of atoms , although @xmath215 can be used to describe the net charge of a set of vortices as well , as in fig .
[ fig : plumpudding]a .
this reaction describes onne point - vortex breaks up into three vortices .
the @xmath216 is just a separator between the different reaction products , reminding one to check conservation of charge by multiplying both sides of the reaction out .
another useful cyclic spinor is @xmath217 _ this _ spinor corresponds to the tetrahedron in fig .
[ fig : bottler]b .
its vertices are at the points of the form @xmath218 if we restrict the choices of signs so that there are always @xmath96 or @xmath28 minus signs .
the fact that the @xmath65 , @xmath52 , and @xmath201 axes of this tetrahedron correspond to the @xmath219 and @xmath220 coordinate vectors makes the spinor @xmath25 especially convenient for determining the consequences of the quadratic zeeman term in the next section .
the orientation of the tetrahedron in fig .
[ fig : bottler]b is also very convenient for working out the group of charges , since the expressions for the symmetry axes are so simple .
( e.g. , @xmath221 , since vertex @xmath199 is in the @xmath222 octant .
thus @xmath223 . )
at nonzero magnetic field , only rotations around the @xmath23-axis are symmetries . when @xmath26 , the ground state space is @xmath78 , consisting of rotations and rephasings of @xmath24 , as in eq .
( [ eq : alphax ] ) .
vortices are described as in eq .
( [ eq : tiedup ] ) by an ordered pair @xmath224 describing the rotation and rephasing angles of the vortex .
the subscript @xmath151 is used to indicate that the @xmath23-axis is an order three symmetry of the @xmath26 ground state tetrahedron .
the continuity of the phase limits the vortex types to the form @xmath225 when @xmath27 , minimizing eq .
( [ eq : effective ] ) implies that the magnetic field axis is an order two symmetry , and the ground state space is @xmath62 , the rotations and rephasings of @xmath25 .
now @xmath226 specifies the vortex types .
the possibilities are @xmath227
this section considers the effect of the magnetic field , which binds vortices , and the kinetic energy , which keeps the bound vortices from merging altogether .
these are both included in the full energy function @xmath228 where @xmath229 is given by eq .
( [ eq : hamiltonian ] ) .
when @xmath230 , the zeeman effect introduces an extra phase boundary dividing the cyclic phase into two phases , with a phase transition at @xmath63 where the tetrahedron changes its orientation relative to the magnetic field .
( this result applies for moderate magnetic fields ; very low magnetic fields cause different transitions@xcite . )
let us determine which orientations of the tetrahedron will be preferred by a magnetic field along the @xmath23 axis .
the preferred orientation can be calculated for small @xmath6 from an effective potential which is a function of the orientation of the tetrahedron . as long as @xmath231
the tetrahedron will be only slightly deformed .
it will move into a space @xmath232 displaced by a distance on the order of @xmath233 from the space @xmath13 of arbitrarily oriented perfect tetrahedra .
the spinors in the distorted space are given by @xmath234 where @xmath168 is the spin two rotation matrix corresponding to the rotation @xmath75 of space , multiplied by a phase .
the distortion @xmath235 depends on the orientation @xmath75 .
( [ eq : manifold ] ) , which omits the deformation , is a harmless shorthand description , emphasizing the orientation of the tetrahedron . in this section ,
we use @xmath25 as the standard spinor orientation instead of @xmath24 to simplify calculating the energy ; conveniently , the body axes @xmath29 of the corresponding tetrahedron for the _ former _ state are aligned with the coordinate axes @xmath236 .
the body axes of the rotated state @xmath237 ( which make the angles @xmath238 with the @xmath23-axis ) are thus @xmath239,@xmath240 , @xmath241
. therefore the @xmath23 component of the spin , in terms of the components of the spin along the body axes , is @xmath242 at first order the quadratic zeeman effect does not have any dependence on the orientation of the tetrahedron because a tetrahedral spinor is pseudoisotropic , " i.e. , @xmath243 . the first order energy is thus given by @xmath244 which does not prefer any orientation of the tetrahedron . in the last step we used @xmath245 which follows from the fact that @xmath220 , a unit vector , has body - coordinates @xmath70 . to find the second order energy due to the quadratic zeeman effect , which will break the tie
, we have to find the deformed state and its energy . the deformation @xmath235 is determined by minimizing the total interaction and zeeman energy in eq .
( [ eq : hamiltonian ] ) for each given orientation @xmath75 .
now if the deformation is not restricted somehow , the deformation " which minimizes the energy will be very large , involving the tetrahedron rotating all the way to the absolute ground state.we therefore allow deformations only of the form @xmath246 where @xmath247 are real numbers .
these terms correspond to the excitation modes found by @xcite .
this correction only perturbs @xmath40 in @xmath248 of the @xmath249 directions in the hilbert space .
the other @xmath250 directions are accounted for by the rotation @xmath75 and the phase @xmath160 which would be goldstone modes when @xmath14 .
( the energy remains @xmath160-independent even when @xmath251 . ) the particular six stiff deformations in eq .
( [ eq : modes ] ) are chosen because they are orthogonal to infinitesimal rotations and rephasings of the tetrahedral state .
we have to find the deformations that minimize @xmath229 , eq .
( [ eq : hamiltonian ] ) , for each rotation @xmath75 . to evaluate @xmath229 , note that @xmath168 cancels from all the terms in the energy except for the zeeman term , where one can use eq .
( [ eq : qmechspecial ] ) .
the resulting expression for the energy density reads @xmath252 where @xmath253 + ( e+if)\chi_{2t}$ ] is the perturbed wave function without the rotation
. the effective potential eq .
( [ eq : effective ] ) is obtained by minimizing @xmath229 over @xmath254 while keeping @xmath75 fixed .
further details are in appendix [ sec : feelingofdealing ] .
( working with @xmath255 is equivalent to fixing the orientation of the tetrahedron and rotating the magnetic field . )
the effective potential suggests an analogue of the magnetic and charge healing length in condensates without magnetic fields ( the tetrahedron tipping length")when the tetrahedra are rotated out of the appropriate ground state orientation at the edge of a condensate , the energy in eq .
( [ eq : effective ] ) returns the order parameter to @xmath11 within the distance @xmath256 .
now the ground states can be found as a function of @xmath44 .
when @xmath26 , a short calculation shows that eq .
( [ eq : effective ] ) has its minimum at @xmath257 ; i.e. , when the magnetic field is along the line connecting a vertex of the tetrahedron to the opposite face or vertex .
hence the order parameter space @xmath11 is as given in eq .
( [ eq : alphax ] ) .
when @xmath27 , the effective potential is minimized by an orientation in which the field points parallel to the line joining a pair of opposite edges ( see fig .
[ fig : bottler ] ) , the fluid is ferromagnetic i.e . , it becomes spontaneously magnetized either parallel or antiparallel to the external magnetic field with an @xmath258 that can be calculated from the expressions in appendix [ sec : feelingofdealing ] .
this magnetization spontaneously breaks a symmetry since the quadratic zeeman effect is invariant under @xmath259 for zero net magnetization , domains will form that have magnetization in both directions . even in the presence of a small cubic zeeman term
, one can show that there is a discontinuous phase transition ( as the thermodynamical variable conjugate to magnetization is varied ) although the symmetry between @xmath260 is not exact . ] .
the dependence of the ground - state orientation on @xmath44 can be understood intuitively using the geometrical representation of ref .
@xcite . in a tetrahedral state , @xmath261 and @xmath262 , minimizing the interaction energies in eq .
( [ eq : hamiltonian ] ) . when a magnetic field is applied , the base of the tetrahedron illustrated in fig .
[ fig : bottler]a gets pushed toward the vertex at @xmath263 .
the tetrahedron in fig .
[ fig : bottler]b , on the other hand , has its upper and lower edges pushed together , toward the @xmath264-plane .
both these deformations move the spin roots ( see @xcite ) away from the north and south poles , which increases the probability that @xmath265 in the corresponding spinors , as favored by the quadratic zeeman effect .
now these two types of deformed tetrahedra have different spin - dependent energies . in the first case ,
the magnetization @xmath258 of the spinor becomes nonzero , increasing the @xmath47 term of eq .
( [ eq : hamiltonian ] ) . in the second case , the magnetization of the spinor is still zero , by symmetry , but one can check that the spinor is no longer orthogonal to its time - reversal , so @xmath99 is nonzero with increasing magnetic field @xmath52 into @xmath266 with @xmath267 because the quadratic zeeman field favors large values of @xmath268 .
the ground state @xmath25 distorts into @xmath269 with @xmath270 and @xmath271 .
the deformed state @xmath272 has a nonzero @xmath258 and a vanishing @xmath99 and @xmath273 is the other way around .
incidentally , when @xmath52 becomes large enough , the ground state will become more symmetrical , with @xmath274 and @xmath275 ( compare with ref .
@xcite ) . ] .
therefore the orientation of the ground state tetrahedron is determined by whether the @xmath276 term or the @xmath277 has a larger coefficient in the hamiltonian . if the coefficient of the magnetization term is very large , then the state without any magnetization is preferred .
the detailed calculation shows that the relevant comparison is between @xmath278 and @xmath279 .
the main source of anisotropy is different at sufficiently low magnetic fields@xcite ; the cubic zeeman effect , proportional to @xmath280 , then dominates over the effective potential in eq .
( [ eq : effective ] ) , proportional to @xmath281 .
nevertheless the the order @xmath281 effect we have calculated can dominate over the @xmath280 effect , even when the magnetic field is small enough for the perturbation theory just described to apply .
this is possible because the denominator @xmath282 in eq .
( [ eq : effective ] ) is small compared to the hyperfine energy splitting @xmath283 . for spin 2 atoms the effect of the magnetic field
is given by @xmath284 where @xmath283 is the hyperfine coupling .
it follows that the quadratic and cubic zeeman effects are @xmath285 and @xmath286 .
the analysis we have given applies when the magnetic field is weak enough that the wave function is not drastically distorted , eq .
( [ eq : smallfield ] ) , but strong enough for the second order effect of the quadratic zeeman term to dominate over the cubic zeeman term .
these conditions combine into @xmath287 for @xmath45rb at density @xmath288 , @xmath289 nk and @xmath290 mk while @xmath291k / gauss ; hence the anisotropy potential used here is actually valid for a wide range of magnetic fields , between @xmath292 mg and @xmath293 g. now that we have estimated the energy due to misalignments with the quadratic zeeman field , let us determine the kinetic energy of vortices in terms of the rotations and rephasings , eq .
( [ eq : wrongdivision ] )
. this energy will be the source of the repulsion that keeps the vortices apart within the vortex molecules .
the kinetic energy is defined by the gradient terms in the hamiltonian @xmath15 ; the analogue for liquid crystals is the elastic energy favoring alignment of the order parameter . in order for the kinetic energy of a vortex to be minimal
, the field on a circle around it should trace out a geodesic , as mentioned above . for the cyclic phase ,
geodesics take the form given by eq .
( [ eq : lathes ] ) . to see that vortices are geodesics as functions of @xmath71 , suppose the field of a vortex is given far away by the radius independent expression @xmath294 for an appropriate spinor function @xmath295 .
the main contribution to the energy of this vortex is from the kinetic energy , which can be estimated by integrating from the core radius @xmath296 .
this is the same as the spin healing length ( see @xcite for the definition ) .
the kinetic energy is @xmath297 now the curve parameterized by @xmath295 adjusts itself so as to minimize the last integral , while maintaining the topology of the circuit traced out by @xmath295 in the order parameter space .
one can show that an integral of this form is minimized when @xmath295 traces out a closed _
geodesic _ in the ground state space .
the length of a closed curve is defined by @xmath298 , so it is not surprising that the geodesic of charge @xmath17 , which minimizes this expression , also minimizes eq .
( [ eq : kespinor1 ] ) . furthermore ,
if the geodesic has length @xmath299 , then @xmath300 .
a geodesic in the cyclic order parameter space ( which has the local geometry of a perfect sphere in four dimensions ) can be described by a rotation at a fixed rate around a single axis as in eq .
( [ eq : lathes ] ) .
( for a shape less isotropic than a tetrahedron , the order parameter would rotate around a wobbling axis according to the rigid - rotation equations . ) substituting the symmetrical @xmath295 into eq .
( [ eq : kespinor1 ] ) gives@xcite @xmath301 \label{eq : kespinor}\end{aligned}\ ] ] where the general expression for any spin and phase has been given in terms of the quantum fluctuation matrix @xmath302 and the magnetization per particle @xmath303 .
the spin 2 tetrahedron state appears to be isotropic as long as one does not go beyond second order correlators , as seen from the following calculations : @xmath304 and hence @xmath305 .
[ eq : kespinor ] implies that the energy is proportional to @xmath306 , a generalization of the pythagorean theorem showing how to combine the amount of rephasing and rotation to get the total geodesic length .
a spin rotation costs twice as much energy as a rephasing by the same angle . in order to study vortex stability and coulomb forces ,
let us define the energy index , " @xmath307 which is a fraction for each tetrahedral charge from sec .
[ sec : namingofcats ] .
the energy of a vortex is given as @xmath308 , a multiple of the energy of an ordinary phase vortex .
the force between a pair of vortices can be expressed very simply in terms of @xmath309 .
the force follows from an estimate of the energy of a cluster of vortices . using ideas from ref .
@xcite , we think of the cluster as forming the core " of a bigger vortex , as illustrated in fig . [
fig : maizy ] .
( we are not necessarily assuming that the vortices are bound together . )
draw a circle of radius @xmath274 just around the group of vortices .
the kinetic energy can then be found as the sum of the energies outside and inside of @xmath274 ; for the case illustrated in the figure this energy is approximately @xmath310 + [ ( i_e(1)+i_e(2)+i_e(3))\ln\frac{l_x}{a_c } ] ) , \label{eq : witches}\ ] ] where @xmath311 is the diameter of the group being combined together and @xmath75 is the radius of whole system .
the first term describes the energy outside of @xmath274 .
sufficiently far outside of @xmath274 , the field should have the form of a rotationally symmetric vortex .
the rotation and rephasing of this vortex , measured along a circle outside @xmath274 , is obtained by multiplying the group elements which describe all the individual vortices according to the rule for vortex unification eq .
( [ eq : berlin ] ) . hence the energy outside @xmath274
is given by an expression like in eq .
( [ eq : kespinor ] ) , except that the integral must start at the radius @xmath311 of the circle , so @xmath91 must be replaced by @xmath311 .
the energy inside @xmath274 is approximated by adding the energies of the three vortices in it , which are calculated like in eq .
( [ eq : kespinor ] ) but now with @xmath75 replaced by @xmath311 .
this approximation makes a small error ( compared to @xmath312 ) by ignoring the region where the vortices fields overlap .
this error has a special scaling form if @xmath14 and if there are no other vortices in the condensate . in this case
, the kinetic energy gives a complete description of the vortices outside their cores , and has a symmetry under rescaling .
therefore ( as in @xcite ) , one can show that the difference between eq .
( [ eq : witches ] ) and the actual energy has the form @xmath313 where @xmath314 refer to the sidelengths of the triangle formed by the vortices .
this correction can be ignored relative to the logarithmically divergent terms kept in eq .
( [ eq : witches ] ) , _ as long as the side - lengths all have the same order of magnitude _ , since in this case @xmath315 has no singularity .
a group of vortices is combined into the core of a composite vortex @xmath274 .
the energy outside the blacked - out core " is calculated from the resultant winding number of the vortices inside , and the energy inside the core " is calculated by looking inside the core to see which vortices are actually there . ] the general expression for the energy of a set of vortices with charges @xmath316 , after rearranging the formula to emphasize the dependence on the diameter of the set @xmath79 , is : @xmath317 \ln\frac{l}{a_c}\nonumber\\ & + \frac{\hbar^2\pi n_0}{m}i_e(\prod_i\gamma_i)\ln\frac{r}{a_c } \label{eq : nerffusion}\end{aligned}\ ] ] as long as the vortex cores are well - separated , the error in this estimate depends only on the ratios of the distances between the vortices , as in the three - vortex case , eq .
( [ eq : chandrasekhar ] ) .
the energy function has a similar form to the two - dimensional coulomb interaction between vortices in an ordinary single - component condensate , and shows that the vortices repel or attract each other according as the index of the combined vortex is greater or less than the sum of the separate vortices indices .
an interesting difference is that the interaction is _ not _ a sum of interactions of all pairs ( consider three @xmath318 vortices , for example , or almost any other set of three vortices ) .
when there are just two vortices , the force between them can be calculated by differentiating eq .
( [ eq : nerffusion ] ) , giving @xmath319 $ ] .
this is the _ exact _ ( if @xmath320 and @xmath14 ) expression for the force between the two vortices because the error ( see eq .
( [ eq : chandrasekhar ] ) ) reduces for two vortices to a constant . improving on eq .
( [ eq : nerffusion ] ) depends on finding the spin texture in the overlap regions , and these equations are nonlinear when not all the vortices use the same symmetry axes .
in the next section , we will apply eq .
( [ eq : nerffusion ] ) more broadly .
when there is a small cluster of vortices of size @xmath79 , but there are some other vortices besides , the energy of just the vortices _ in the cluster _ can be estimated by replacing @xmath75 in eq .
( [ eq : nerffusion ] ) by the distance @xmath321 to the nearest vortex not in the cluster .
likewise , for @xmath16 the energy of a cluster with a net charge that is field - aligned can be estimated by replacing @xmath75 by @xmath22 . as long as @xmath322 , the detailed expression for the energy has the form given by eq .
( [ eq : chandrasekhar ] ) , but when @xmath79 approaches @xmath22 , the anisotropy energy starts to compete with the kinetic energy and to distort the spin texture around the vortices .
in this section we will first discuss stability of isolated tetrahedral vortices and then determine when these vortices can combine to form molecules ( and what the spin texture around a molecule looks like ) .
some of these molecules are only metastable and can each break up in several ways .
bound states of vortices will be formed out of stable tetrahedral vortices .
these are the vortices based on @xmath84 symmetries of a tetrahedron , accompanied by a phase shift of @xmath323 or @xmath324 , and vortices based on @xmath64 symmetries without a phase shift .
vortices with larger rotations or phase shifts will not occur as components of molecules .
the net charge of the bound state must be field - aligned , with a schematic relation @xmath325 between the field - aligned charge and the component charges .
a composite vortex is stable if coulomb repulsions can prevent the collapse of any subset of its components ; since the coulomb interaction is not a sum of pairwise interactions , it is not enough that every pair of vortices repel each other . finally any bound state
whose charge @xmath326 has big enough rotation or rephasing angles can in principle decay into molecules with less energy , but this can only occur if thermal energy overcomes the coulomb repulsion between the component vortices .
the following expands on this general picture and points out some interesting sidelights .
only stable vortices will be found in the core - region of composite vortices .
since a weak anisotropy term can be neglected near component vortices , we can enumerate the stable vortex types assuming that @xmath14 . absolute stability of a vortex with charge @xmath17 implies that if @xmath17 is forced to break up into the fragments @xmath316 , then the energy of the fragments grows as they separate from one another .
the energy of the fragments can be found by applying eq .
( [ eq : nerffusion ] ) .
note that @xmath327 by conservation of charge . if @xmath17 has phase and rotation angles @xmath99 and @xmath107 then @xmath328\ln\frac{l}{a}+cnst .
\label{eq : nerffusionexpanded}\ ] ] thus , if @xmath329 , the energy decreases as the vortices move apart , so the fragments will move apart by themselves the rest of the way . on the other hand ,
the vortex is absolutely stable if @xmath330 any tetrahedral vortex not satisfying this absolute stability criterion " can break up into a lower - energy state , and we will assume that this break - up happens spontaneously for these _ component _ vortices .
for example , the vortex @xmath331 should break up into two @xmath332 s , halving the energy .
( in fact , these two singly quantized vortex can break up even further . )
generalizing this example , any vortex with charge @xmath215 whose circulation @xmath99 is bigger than @xmath97 can break up into @xmath333 and @xmath334 since @xmath335 .
this leaves only finitely many vortices that have the possibility of being stable : all the ones with phase winding numbers not more than @xmath97 .
some of the vortices with @xmath336 are also unstable ; trial and error finds decay processes such as : 1 .
@xmath337 or @xmath338 .
2 . @xmath339 3 .
@xmath340 can break up into @xmath341 and also @xmath342 4 .
@xmath343 might break up into @xmath344 .
@xmath345 could break up into @xmath346 .
all of these decays lower the energy index .
there are also two vortices which can break up without the energy indices changing : 1 .
@xmath334 has the same charge as @xmath347 .
@xmath348 has the same charge as @xmath349 .
the energy may either increase or decrease after one of these processes , the logarithmic term which dominates the energy ( see eq .
( [ eq : nerffusion ] ) ) does not change , so the remainder term needs to be calculated to determine whether these break - ups raise or lower the energy .
a point vortex whose charge is @xmath334 is probably unstable if @xmath130 because a pure phase vortex has to have an empty core ( with energy density of order @xmath350 ) , while the two fragments it breaks up into just have non - cyclic cores ( energy of order @xmath106 ) . and the fragments interact with a short - range repulsion , @xmath351 perhaps ; when the energy index of the fragments actually decreases the force is @xmath352 . whether @xmath348 is stable
or not we do not know , and the answer may depend on @xmath44 . ]
we can now enumerate the vortices which are stable , noting that some vortices ( such as @xmath353 ) can break up similarly to the ones just listed on account of symmetry .
the only vortex types that have not been eliminated are the vortex with one - third circulation , @xmath354 ( energy index @xmath355 ) and the currentless vortex @xmath356 ( energy index @xmath357 ) as well as possibly @xmath348 ( energy index @xmath358 ) , and also the inverses and conjugates of these . in this section
we will describe a qualitative wave function for the example in section [ sec : doublestar ] to illustrate how the field of a vortex molecule deforms in response to the anisotropy energy as one leaves the region containing the two vortices .
we will then give the binding criteria which describe how to use group theory to check which sets of the stable tetrahedral vortices ( from the previous section ) can combine together to form a vortex molecule at a nonzero magnetic field . a vortex with a charge that is not compatible with the magnetic field
has an energy that grows proportionally to the area of the condensate .
( [ eq : compatibility ] ) shows that the vortices in a cluster can avoid this energy cost if they have a net charge corresponding to a rotation around the @xmath23 axis .
more specifically , a set of tetrahedral vortices , with topologies @xmath316 , has to have a net topology of the form @xmath359 the molecule " can then have the rotational and phase windings @xmath360 at infinity , for any @xmath361 .
one example of a molecule for the @xmath78 phase is the composite vortex discussed at the beginning , @xmath362 , which partner up to make a compound vortex with @xmath176 in the field - aligned condensate . writing a wave function that describes this example even qualitatively is a little more complicated than it sounded in sec .
[ sec : doublestar ] . a possible wave function
is illustrated in fig .
[ fig : suntiles ] .
we can build this wave function up in stages .
the most obvious attempt at writing a wave function fails to eliminate the quadratic energy cost : @xmath363 where @xmath364 is the polar angle measured with respect to the location @xmath365 of the @xmath366 vortex .
the two @xmath64 rotations about @xmath65 combine into a @xmath97 rotation on a circle surrounding both vortices , but not about the magnetic field axis . _
this _ wave function takes the form @xmath367 at infinity where @xmath368 and @xmath369 approach @xmath71 .
the tetrahedra rotate around a tilted axis , so they do not stay aligned appropriately with the magnetic field except on the @xmath73-axis and the @xmath370-axis ( where the tetrahedra are reversed , but this is still a ground - state ) .
the zeeman energy of @xmath371 is still infinite .
this wave function agrees pretty well with the one illustrated in the figure _ within _ the region surrounded by the large circle .
the vortex cores are cordoned off by the small circles , where the @xmath64 symmetry axes are marked by dots .
note that the tetrahedra three rows inside the large circle behave like the tetrahedra at infinity described by eq .
( [ eq : bigsubscript ] ) ; they rotate through @xmath76 about the order - two axis bisecting their right edge , producing all sorts of arbitrary orientations .
another attempt , which uses the @xmath75 axis instead of the @xmath65 axis to eliminate the tilting of the tetrahedra , @xmath372 is a complete fiasco , since this function is not continuous along the line connecting the two cores .
( the @xmath75 axis has the wrong symmetry , so as @xmath373 circles around @xmath374 , @xmath375 changes from @xmath126 to @xmath376 . ) luckily , since a @xmath97 rotation around one axis can be deformed to any other ( @xmath97 rotations all correspond to @xmath377 ) one can produce a spin texture which has a finite zeeman energy and is continuous .
a hybrid of eqs .
( [ eq : bigsubscript ] ) and ( [ eq : biggersubscript ] ) which achieves this , illustrated in fig .
[ fig : suntiles ] , is : @xmath378 where @xmath379 must satisfy @xmath380 for example , we could define @xmath381 where @xmath321 is a variational parameter .
( probably @xmath382 is optimal . )
this field has two vortices at @xmath383 .
the texture varies continuously , except at the at the very centers of the vortices .
( in the figure , the cores are inside the small circles .
the tetrahedra s orientations change rapidly around the circles centers , but each tetrahedron inside has a similar orientation to the tetrahedron next to it outside the circle . ) around the cores , the tetrahedra rotate around a tilted @xmath384 axis which is aligned with the symmetry axis of the tilted tetrahedra around the vortex ; the tilting is carried out by the rotations described by the first two factors in eq .
( [ eq : tailoring ] ) , just as discussed in section [ sec : doublestar ] .
near vortex 1 , the wave function is approximately @xmath385 and near vortex 2 , @xmath386 ( the local symmetry axis @xmath384 is the rotation of the standard axis @xmath387 , through a half - turn about @xmath388 , according to eq .
( [ eq : lathes ] ) . ) though the two tetrahedra are rotated into misaligned positions inside the large circle , the excursions from @xmath78 to @xmath13 are brief and therefore cost only a finite amount of energy .
indeed , the first two factors fix the field up at infinity by applying a continuously varying rotation to the overall texture .
these factors change the axis from @xmath65 to @xmath75 at large @xmath77 as one can see by replacing @xmath389 by @xmath71 and using eq .
( [ eq:0 ] ) : @xmath390 the @xmath76 axis changes from the @xmath65 axis to the @xmath75 axis as one crosses through the transition region indicated by the large circle in fig .
[ fig : suntiles ] .
as you follow a radius outward past the circle , the tetrahedra are tipped by different amounts .
( the tetrahedra on the negative @xmath73-axis have to be tipped the most , though their original orientation is compatible with the magnetic field ! the face which is on top is changed .
this reorientation is required to make the amount of tipping continuous : the tipping angle increases more and more as @xmath71 goes from @xmath96 to @xmath391 . )
the first two factors manage to fix the orientation at infinity without introducing discontinuities like in eq .
( [ eq : biggersubscript ] ) .
although they seem to have a vortex - like discontinuity at the origin , their discontinuous parts cancel on account of eq .
( [ eq:1/3 ] ) .
( hence the tetrahedra near the origin in fig .
[ fig : suntiles ] all have roughly the same orientation . ) in short , though @xmath392 , the cancellation of the zeeman energy at infinity is not automatic !
the topological classification just implies that the field in eq .
( [ eq : bigsubscript ] ) _ can _ be deformed as @xmath18 so that the zeeman energy is small . the kinetic energy of this composite vortex ,
according to eq .
( [ eq : nerffusion ] ) , is @xmath393\ln\frac{l}{a_c}\nonumber\\ & + \frac{\pi\hbar^2n_0}{m}i_e((-id,0))\ln\frac{r}{a_c}. \label{eq : keamolecule}\end{aligned}\ ] ] since @xmath394 and @xmath395 , the coefficient of the first term is negative , so the vortices are driven apart in order to decrease the kinetic energy .
this effect competes with the anisotropy energy , which tries to bring the vortices together .
the net energy is therefore @xmath396 this is a more complete version of eq .
( [ eq : barrier ] ) . @xmath315 summarizes the dependence of the anisotropy energy eq .
( [ eq : effective ] ) on @xmath321 and @xmath79 , but is too complicated for us to figure out ! the competition between the attractive zeeman - term and the repulsive kinetic - energy term determines the size of the vortex . assuming @xmath397 and minimizing over @xmath79 gives @xmath398 this composite core - size is also the scale for the decay of tipping of tetrahedra due to the competition of anisotropy and kinetic energy ( see sec .
[ sec : cubeshape ] ) , just as the core size of a spin vortex in a spin 1 condensate is equal to the magnetic healing length .
vortex molecules will have a size of the same order of magnitude ( even if there are more than @xmath28 subvortices ) except when the component vortices have a short - range repulsion , like the two examples at the end of section [ sec : list ] .
( such components form smaller molecules ; vortices never form larger molecules . if a pair of tetrahedral vortices is stretched beyond @xmath22 , a cord " of tipped - tetrahedra forms between them , a simple version of the string imagined to hold the quarks together in a rapidly - rotating baryon . ) to find the total energy of the molecule , substitute @xmath79 into eq .
( [ eq : barrier2 ] ) and simplify using @xmath399 : @xmath400 we have dropped the contribution from the quadratic zeeman term since it adds something that is independent of @xmath6 . this formula can be compared to the energy one expects for a _ simple _ vortex with the same charge @xmath401 : @xmath402 which includes a core energy @xmath403 and allows for a physical definition of the core size . if we take @xmath404 to be the size of the molecule , @xmath405 , then the core energy has to be @xmath406 the core energy becomes large as @xmath21 because the kinetic energy of the component vortices diverges as the logarithm of the core size @xmath22 .
it makes sense to regard this composite vortex as a molecule because coulomb repulsion keeps the vortices in the core separate .
the previous discussion can be generalized by listing a set of binding criteria ; these ensure that a set of tetrahedral vortices will form a stable or metastable composite vortex : 1 .
each component vortex is one of the stable @xmath14 vortices from section [ sec : list ] .
the kinetic energy is not decreased when any subset of the component vortices coalesces into a single vortex .
3 . there is no way for the component vortices to form submolecules that can break apart .
this would occur if the components could be rearranged and then partitioned into @xmath77 sets @xmath407 such that each subset forms a molecule _ that is compatible with the magnetic field _
( i.e. , @xmath408 ) and such that _ the sum of the energy indices of these submolecules is less than the energy index of the original molecule_. the vortex molecule @xmath409 clearly satisfies all these conditions : condition 1 is satisfied because @xmath318 is one of the stable vortices found in section [ sec : list ] .
condition 2 is satisfied because the vortices repel each other .
condition 3 is easy to check for a diatomic molecule like this one , since it can only break up into individual atoms " ; neither of the fragments @xmath318 is compatible with the magnetic field
. not all of the vortex molecules satisfying the three conditions above are _ absolutely _ stable .
the analogue of the absolute stability condition , eq .
( [ eq : absolutestability ] ) , also selects a finite set of _ aligned _ vortex types when @xmath251 , this time from among the group elements listed in eq .
( [ eq : type3 ] ) .
the absolutely stable charges are@xmath410 for any _ other _ vortex topology @xmath326 , one can find vortex topologies @xmath411 such that @xmath412 and @xmath413 _ point _ vortices with such a topology @xmath102 would likely break apart spontaneously . there is actually another pair of charges that could be absolutely stable , but the energy index estimate is not accurate enough to decide the issue : @xmath414 these vortices can break up into pairs without changing the net energy indices , reprising the ambiguous behavior of the two vortices at the end of sec .
[ sec : list ] .
( see section [ subsec : bound ] for an answer . )
illustration of the absolute stability criterion . a composite vortex ( a ) and a possible set of composite vortices
( b ) it can break up into . even if @xmath415 and @xmath416 so that the energy would decrease , the break - up might not occur spontaneously .
if the component vortices in ( b ) are different from the component vortices in ( a ) , then the vortices making up @xmath417 , @xmath418 , and @xmath419 would have to be produced in a
chemical " reaction from the components of @xmath102 . ]
there are some _
composite _ vortices of other charges besides the four listed in eq .
( [ eq : stable3 ] ) which are long - lived .
the absolute stability criterion misses this possibility because it ignores the details of the vortex cores , drawing all its conclusions from the topology of the vortices far away : suppose the initial vortex @xmath102 is a cluster of vortices with topologies @xmath420 ( see fig .
[ fig : tc ] ) .
the decay products discussed in the previous paragraph , @xmath411 , may be the combined topology of other vortex clusters . the energy of the vortices after the reaction ( fig .
[ fig : tc]b ) is smaller if @xmath421 = e_{fin } ; \label{eq : degrad2}\end{aligned}\ ] ] the terms proportional to @xmath422 stand for the kinetic energies outside the composite cores , and the terms proportional to @xmath423 stand for the energies within the composite cores ; the latter contributions do not matter once the composite vortices are far apart ( @xmath424 ) .
therefore , the energy - index relation , eq .
( [ eq : degradation ] ) , implies that the energy decreases when @xmath425 .
however , at zero temperature , a vortex molecule satisfying the binding criteria _ can not _ break up , even if the total energy would end up smaller . although the metamorphosis of fig .
[ fig : tc]a into fig .
[ fig : tc]b lowers the energy , the process will not occur spontaneously . according to condition 3 , the vortices in fig .
[ fig : tc]b are different than those in fig .
[ fig : tc]a . and according to conditions 1 and 2 , there are no spontaneous chemical reactions that can occur to make the components in fig . [ fig : tc]b out of those in fig .
[ fig : tc]a . thus , at zero temperature , composite vortices besides the ones with charges listed in eq .
( [ eq : stable3 ] ) , whose energy indices seem to be too big , can still be stable . at _ nonzero _ temperature , such a molecule will only be metastable because it can decay after reactions inside its core produce the vortex types that appear in fig .
[ fig : tc]b .
these reactions are prevented by energy barriers , so the decay will occur only after a long time .
the vortex molecule made up of @xmath409 is an example of a metastable vortex ; its charge @xmath401 did not appear in eq .
( [ eq : stable3 ] ) because it is the same as the net charge of the three _ field - aligned _ point vortices @xmath426 .
the energy index of the molecule is @xmath427 ( see eq .
( [ eq : index ] ) ) while the energy index of the three point vortices is smaller , @xmath428 . nevertheless ,
since these three vortices are not present in the core of the original vortex the decay can not occur spontaneously .
a vortex with a composite core can break up by a combination of vortex - fusion and vortex fragmentation .
some subsets of the original component vortices fuse and the fused vortices each break up into some other vortices .
these regroup into clusters each of which has a charge compatible with the magnetic field .
then each cluster goes its own way .
( the fusion and fragmentation steps might sometimes happen more than once . )
conditions 1 and 2 ensure that at least one of these steps will be opposed by the coulomb potential , but the _ total _ energy will decrease if the energy index decreases
. the vortex molecule will be long - lived because its components do not know that the hard effort of fusing will allow them to change into vortices which can separate .
trial and error yields a couple of ways in which the @xmath409 bound state can break up .
one possibility begins with the two component vortices coalescing , @xmath429 the other begins when one of the components breaks up , @xmath430*(a,0)\nonumber\\ & \rightarrow ( r^{-1},-\frac{2\pi}{3})*[(q,\frac{2\pi}{3})*(a,0 ) ] .
\label{eq : fission}\end{aligned}\ ] ] in the first process , the two vortices come together , increasing the kinetic energy in accordance with condition 2 ( as shown by calculating the @xmath309 s and substituting into eq .
( [ eq : nerffusion ] ) ) .
the resulting vortex breaks up into three vortices which can separate from each other because they are compatible with the magnetic field .
the increase in energy during the first stage is given by @xmath431 where @xmath432 is the energy of the intermediate vortex .
thermal fluctuations have a chance of driving the @xmath318 s together , in spite of the energy increase @xmath433 . in the second process ,
( [ eq : fission ] ) , the @xmath318 vortex first splits up into two vortices .
the first of these , @xmath434 is compatible with the zeeman field and can leave .
the remaining two vortices form a new molecule which can not break up because @xmath102 is a rotation around the wrong order @xmath151 axis . for _ this _
process condition 1 requires that the energy increases during the initial fragmentation . to check this , note that the energy of the intermediate state is @xmath435\ln\frac{l}{a_c}\nonumber\\ & + i_e(-id,0)\ln\frac{r}{l}\}\end{aligned}\ ] ] and the energy barrier is @xmath436\ln\frac{l}{a}\nonumber\\ & = \frac{\pi n_0\hbar^2 m}{6}\ln\frac{l}{a}. \label{eq : fissionbar}\end{aligned}\ ] ] this energy barrier is lower than @xmath433 , so eq . ( [ eq : fission ] ) is a more common break - up route .
( in a finite condensate , thermally excited break - ups can be observed only if a vortex molecule is somehow prevented from wandering to the boundary of the condensate and annihilating before it can decay . )
these two examples illustrate the meaning of the stability conditions .
conditions 1 and 2 ensure that fragmentation and fusion processes can not happen spontaneously .
the third condition simply points out that vortex clusters like @xmath437 will not be stable because the components can sort themselves into field - aligned groups and break up without any thermal assistance
. the second condition can be difficult to check for a composite vortex with three or more sub - vortices .
one must consider subsets of every size and check that they can not lower their energy by collapsing all at once into one vortex . just knowing that any two vortices of the subset _ repel _ each other
does not guarantee that the set of vortices do not _ collectively attract _ each other ! an example is the set of three vortices @xmath438 .
any two of these vortices would combine to form a vortex with a @xmath76 rotation , @xmath439 , whose index @xmath28 is higher than the sum , @xmath440 , of the indices of the collapsing pair . on the other hand
all three vortices could form a vortex @xmath441 with energy index @xmath442 , so the three vortices can collapse simultaneously to lower the energy .
an even more counterintuitive complication is that , because of the noncommutative behavior of the combination rules , more complicated fusion processes can occur .
a vortex can change its type by circling around one vortex so that it can fuse with another vortex .
( see appendix [ app : catstring ] . ) for a bound state of many vortices there will be many possibilities for how the vortices meander around each other before some of them fuse . to test condition 3 , one also has to enumerate all possible wanderings .
a mathematical statement of our results is that there will be ( at least ) one solution to the time - independent gross - pitaevskii equation for any set of vortex topologies that satisfy the three conditions . if @xmath6 is small , the energy at the top of the barrier is greater , by a logarithmically large amount , than the energy of an initial variational state like the approximate wave function eq .
( [ eq : tailoring ] ) . a solution to the gross - pitaevskii equation should result if one starts from this qualitative texture and lets it relax to a local minimum of the energy .
there is not enough energy for the wave function to get over the energy barrier , so the wave function should get stuck in a local minimum .
( the energy of the intermediate state is not known precisely because of the rough estimates we have made of the zeeman energy and the kinetic energy , but these errors are small compared to the height of the barrier . )
[ cols="^,^,^,^,^",options="header " , ] now we can construct some other , more interesting , examples . we will use the algebra of the group of vortex charges to find molecules whose net charge is interesting in different ways , and we will use the energy index to test whether they are stable . the parameter @xmath44 will be less than 4 for some of these examples . if @xmath27 the ground state orientation of the tetrahedron will be as in fig .
[ fig : bottler]b so the @xmath23-axis is an order 2 axis .
the aligned topologies have the forms @xmath443 , as described in section [ sec : namingofcats ] . of these ,
the only absolutely stable topologies are @xmath444 , @xmath445 ( and possibly @xmath446 ) .
first let us find a vortex molecule whose phase winds by @xmath183 .
in single component condensates , such vortices are usually unstable ; one has been observed to break up , maybe into an entwined pair of @xmath97 vortices@xcite .
if phase and spin textures were completely independent of one another , doubly quantized vortices would not be any more stable in the cyclic condensates ; but fractional circulations are bound " to certain spin textures ( see sec . [
sec : namingofcats ] ) .
_ if _ we assume the vortex @xmath348 is stable ( at the end of sec .
[ sec : list ] we could not decide ) , then a doubly - quantized vortex can occur in a cyclic condensate when @xmath26 .
it consists of the three parts @xmath447 the phase changes by @xmath183 while the orientation of the tetrahedron does not change at infinity as we can check using the coordinate system from fig .
[ fig : bottler]b .
the three group elements are @xmath448 and their product is the identity .
let us discuss the conditions for binding .
we have not checked condition 1 ; it is not easy to check because @xmath449 has the same charge and energy index as @xmath450 ; an accurate solution for the spin texture around this pair of vortices is needed .
besides , @xmath451 might be stable for some ranges of @xmath44 values , but not others .
let us therefore hope that condition 1 is satisfied .
condition 3 is clear . to check condition 2 ,
let us first consider whether one of the _ pairs _ of vortices in the trio can coalesce .
using conservation of topological charge helps to avoid enumerating all the ways the vortices can braid around each other . if the first two vortices have coalesced into a vortex @xmath452 ( after some permutation ) and the third vortex , by winding around the other two vortices as they collapsed , has changed to @xmath453 , then @xmath454 by conservation of charge . hence @xmath455 .
also , braiding one vortex between other vortices can only _ conjugate _ its group element .
therefore , @xmath456 , like @xmath75 , is a counterclockwise rotation through @xmath84 . since @xmath455 ,
the rotation part of the coalesced vortex @xmath452 also is a @xmath84 turn and thus the energy index of this coalesced vortex is @xmath457 , which is _ greater _ than the sum of the energy indices of the two vortices which formed it . therefore the two vortices can not coalesce spontaneously .
( this argument can be generalized to any trio of vortices @xmath458 each of which commutes with the net charge @xmath17 .
fusing two of the vortices gives the same result ( up to conjugacy ) no matter how the vortices are mixed around first ; so braiding can not make a repulsive interaction between two vortices into an attractive one . ) finally , the three vortices can not coalesce simultaneously because @xmath459 .
returning to the original assumption , @xmath26 , where the ground state orientation is illustrated by fig .
[ fig : bottler]a , we can show that the second charge in eq .
( [ eq : dicey ] ) _ does _ correspond to a completely stable vortex molecule .
in fact , consider @xmath460 this molecule , one of the decay products in eq .
( [ eq : fission ] ) , has the topology @xmath461 or ( using the notation appropriate for the field - aligned tetrahedra outside the composite core ) , @xmath462 .
the energy of this molecule is approximately @xmath463\nonumber\\ & = \frac{n_0\pi\hbar^2}{m } ( \ln\frac{r}{a_c}-\frac{1}{6}\ln\frac{l_q}{a_c}).\end{aligned}\ ] ] where @xmath22 is the size of the composite core , given by eq .
( [ eq : collie ] ) .
this molecule answers a question from section [ sec : metastable ] .
are there stable vortices with charge @xmath462 ?
the two vortices @xmath464 , have the same net topology as a vortex of charge @xmath462 , and they have the _ same _ net energy index .
now we can check that the composite vortex @xmath465 is a stable realization for the charge @xmath462 .
its energy is lower _ by a finite amount _ than the energy @xmath466 of the pair of vortices .
this finite binding energy is @xmath467 . to take another point of view
, the minimum - energy spin texture with the topology @xmath462 imposed far away has an asymmetric structure : it has two singularities " with topologies @xmath468 and @xmath318 at a distance of order @xmath22 .
by contrast , when the topology imposed at a boundary corresponds to unstable vortices , the ground state has singularities whose spacing is on the order of the size of the system @xmath75 .
e.g. , in a scalar one might try to impose @xmath469 .
the spacing of the vortices in the energy minimizing wave function grows with @xmath75 , reflecting the fact that these vortices would repel each other to infinity in an infinite condensate .
the final example shows that point vortices are not necessary to hold a core together there is a
composite " vortices without any components ! in other words we can construct a vortex for which the order parameter stays in @xmath13 . there is still a composite core " where the tetrahedra leave @xmath11 and are no longer aligned with the field axis .
the trick is that the amount of rotation in a texture around a vortex is defined only modulo @xmath183 , in the absence of a magnetic field ( because of the @xmath7 charge classification ) .
when @xmath52 is turned on , the spin part of the order parameter space @xmath11 has the same topology as a circle , so each additional winding by @xmath97 changes the topological charge .
( a texture which rotates by @xmath183 about the field axis can relax only by using axes perpendicular to the magnetic field . )
thus a @xmath183-rotation - vortex is stable in a magnetic field , but since it has zero _ tetrahedral _ charge , it does not have to have point vortices inside of it .
( another way to say this : eq .
( [ eq : screened ] ) does not uniquely determine the aligned topology , because @xmath470 .
hence a @xmath471-vortex can be made from @xmath248 @xmath75 vortices ( i.e. , some @xmath472 s and @xmath348 s ) or out of no vortices at all ! ) such vortices occur for both the @xmath26 and @xmath27 cases .
a variational wave function can be constructed using the formula that shows how a @xmath183 rotation - vortex can relax in the absence of a magnetic field : @xmath473 where @xmath158 is an arbitrary cyclic spinor . at each moment of time
@xmath474 , the expression describes an @xmath77-independent texture as a function of @xmath71 . when @xmath475 , there is a vortex which is a full rotation through @xmath476 . by the time @xmath477
, this vortex has completely dissipated .
when @xmath16 a @xmath183 vortex can not relax in this way because the tetrahedra rotate away from the orientation preferred by the magnetic field before returning to the preferred orientation at the end .
but eq . ( [ eq : rubiksolution ] )
has a reincarnation as the description of a @xmath478 vortex .
we replace the time coordinate by a function of the radius to give a spin texture that winds through @xmath183 at infinity but does not have any singularities at @xmath96 : @xmath479 if @xmath480 or @xmath126 , then this wave function , at large @xmath77 s , has the winding number @xmath478 . at small @xmath77 s ,
the wave function is @xmath71-independent , giving a continuous and coreless " wave - function .
( the exact solution not only has a more complicated @xmath77-dependence , but also a less - symmetrical @xmath71-dependence . )
the region @xmath481 is the composite core of this vortex in the sense that @xmath482 rather than @xmath11 .
the optimal size @xmath79 of this region is again @xmath22 , as balancing the kinetic and zeeman energies shows .
this vortex can not disappear because the classification of vortices at _ nonzero _ @xmath6 implies that @xmath186 is conserved .
furthermore , though it does not satisfy the _
absolute _ stability criterion , since two @xmath483 s have a smaller energy , it is obviously metastable there are no vortices in the core to break apart
! the vortex can only break up if thermal energy causes a vortex - antivortex pair to nucleate in the core .
suppose a pair involving rotations through @xmath97 in _ opposite _ directions appears .
these vortices initially attract each other but if the thermal fluctuations pull them to opposite sides of the core the nonlinear coupling with the background field switches this force from attractive to repulsive , and the vortices can separate the rest of the way by themselves .
let us discuss the conditions under which the zeeman - effect bound states might be observed and the methods one can use for observing them .
first of all , we must assume that @xmath484 in order to justify neglecting @xmath6 near the tetrahedral vortices and to justify the perturbation theory of sec .
[ sec : cubeshape ] .
this is not just a technical assumption : above a certain magnetic field the component vortices probably merge . to estimate the maximum magnetic field note that @xmath6 is related to the hyperfine splitting @xmath283 via @xmath485 for rubidium and sodium atoms@xcite , and similar relations hold for other atoms .
also note that the spin independent interaction is @xmath486 where @xmath487 is an _ average _ of the scattering lengths corresponding to different net spins and that the spin dependent interaction is @xmath488 where @xmath489 @xcite depends on the _ differences _ between the scattering lengths@xcite .
the condition for our analysis to be applicable , @xmath490 , therefore implies @xmath491 about @xmath492 g for a condensate of rubidium atoms with density @xmath493cc . in order to observe vortex bound states
, one might start with a condensate prepared with a spin order other than the ground state and then watch it evolve as in ref .
thermal ( and less importantly quantum ) noise will produce perturbations that grow exponentially , producing complicated patterns .
if the magnetic field is small enough , vortex bound states might be found after some time . to test whether these vortex bound states behave in the way we have been describing
, one would have to identify the topological charges of the vortices .
one could then check that vortex sets whose net charge is compatible with the magnetic field have a size on the order of the theoretical value , @xmath22 .
one may have to use statistical correlations if too many vortices stay around .
( one could also take a more deliberate approach , choosing vortex types and imprinting them as in @xcite or @xcite .
one can then observe the subsequent evolution of the vortices to see whether they bind together . ) in fact , identifying the vortices that appear in a spinor condensate is difficult ; vortex cores in a spinor condensate are not empty like the vortices in an ordinary condensate ; they have nearly the same density as the rest of the condensate@xcite .
one thus has to measure something about the spins to observe the vortices .
two possibilities have already been invented .
one can either measure the magnetization field as in @xcite or use stern - gerlach separation to measure the density of the different spin species . measuring the magnetization as a function of position is less informative for a _
nonmagnetic _ phase like the cyclic phase than for the ferromagnetic phase studied in ref .
the magnetization outside the core of a vortex , where the spinor state is approximately a rotation of the unmagnetized cyclic state will be close to zero ( see eq .
( [ eq : sphere?1 ] ) ) , but inside the core , where the order parameter leaves the ground - state space @xmath13 , the magnetization can be nonzero .
measuring the magnetic moment in the core of a vortex helps to determine the topological charge of the vortex .
( the magnetization will not provide any _ direct _ evidence of the rotating orientation of the tetrahedral order parameter , though . )
any vortex one might have to identify involves a rotation about an arbitrary axis @xmath384 as in eq .
( [ eq : lathes ] ) , or eq . ( [ eq : lathes2 ] ) which is more convenient for understanding what a vortex will look like .
the latter description starts with a vortex whose rotation axis is special say it is parallel to @xmath220 , and applies some overall rotation to it .
for example a vortex of type @xmath472 can be obtained from a vortex whose axis is @xmath494 .
( [ eq : lathes2 ] ) implies that the vortex is described by @xmath495 where @xmath496 and @xmath497 are appropriate functions approaching @xmath49 at infinity and @xmath75 is a rotation that moves @xmath498 to @xmath499 .
( the phase , @xmath160 , does not matter . ) if @xmath500 , then @xmath75 is the identity , so @xmath501 and @xmath502n_0 .
\label{eq : magnetization}\end{aligned}\ ] ] the magnetization is parallel to the symmetry axis and is given by @xmath503\bm{\hat{n}}$ ] .
applying an arbitrary reorientation @xmath75 changes the magnetization axis and the symmetry axis in the same way , so the general result is @xmath504\bm{\hat{n}'}.\label{eq : coremag}\ ] ] far from the core , the magnetization vanishes . inside the core
, the magnetization can be found by noting that the top component of the vortex eq .
( [ eq : bump ] ) must vanish at @xmath505 in order to be continuous : @xmath506 since @xmath107 is much larger than @xmath47 and @xmath507 , the density of atoms will be almost uniform across the whole vortex and hence @xmath508 .
( [ eq : dimple ] ) therefore implies @xmath509 hence the magnetization , eq .
( [ eq : coremag ] ) , is approximately @xmath510 in the core ; the atoms have a single unit of hypefine spin in the direction of the vector from the center to the _ fixed _ vertex of the rotating tetrahedra near vortex the core , helps determine the orientation of the tetrahedra near each vortex core , in spite of eq .
( [ eq : sphere ? ] ) .
if there are several vortices nearby one can try to guess how the tetrahedron fields around them fit together . ] .
the inverse vortex , @xmath434 , has the same magnetization ( it does not change sign ) . on the other hand ,
similar arguments show that @xmath348 will have a magnetization approximately equal to @xmath511 at the core center because it is the @xmath512 component of the spinor in the analogue of eq .
( [ eq : bump ] ) that has the phase winding for this case .
the third stable vortex type , @xmath318 , will not have any magnetization in its center and would be hard to see using this method .
measuring the magnetization reveals vortices of order three but does not distinguish between vortices and antivortices and does not even indicate the presence of an order two vortex .
( one _ can _ observe the composite vortex described in section [ subsec : double ] . )
another deficiency is that the cores are only about @xmath513 m across , so the vortices might be hard to observe directly by this method .
however , one could first allow the condensate to expand in the transverse direction so that the atomic interactions decrease .
the vortex cores would expand ; as in experimental observations of vortices in single - component condensates , the depleted region in @xmath315 or @xmath98 ( whichever corresponds to the component of the transformed spinor with the phase winding ) would fly apart and the magnetized core would become much larger .
a magnetized ring would form at the edge of the core where the atoms of one magnetization accumulate more than the atoms of the other .
densities in the five spinor components around an order @xmath151 vortex , with a randomly oriented local axis @xmath384 . on the left is the pattern one might observe experimentally ; the lighter regions correspond to regions with fewer atoms . on the right
are plotted the percentage of atoms for each value of @xmath514 at some fixed distance from the vortex core .
the phases and amplitudes of these oscillations should help to determine the direction of the local axis . ]
the stern - gerlach method gives more information about the vortices .
though the density is depleted at the center , the field _ around _ a vortex in a single - component condensate is not observable , unless one reconstructs the phase variation of the condensate , perhaps using the technique described in ref .
but in a spinor condensate , the spin vortices produce observable patterns in the condensates stern - gerlach images .
these images capture separately the density of atoms in each of the five components of the spinor as functions of position . in these density profiles
each vortex ( aside from pure phase vortices with @xmath515 ) will be ornamented by radiating density ripples as illustrated in fig .
[ fig : shed ] . for example , according to eq .
( [ eq : bump ] ) , the density of atoms with @xmath516 is given by @xmath517 where @xmath518 are constants outside the vortex cores , since @xmath496 and @xmath497 approach 1 . while a vortex in a condensate of a single type of atom does not show any density modulation ( unless the condensate interferes with a second condensate , see e.g. ref .
@xcite ) , angular density ripples do result for a spinor vortex as a result of the interference between the @xmath315 and @xmath98 components of the spinor produced by of the unitary transformation changing the quantization axis from the vortex s rotation axis @xmath384 to the magnetic field direction .
if the @xmath384 axis happens to line up exactly with the axis of the stern - gerlach field , then there are no radial interference fringes , " but only the axes of _ point _ vortices with aligned charges will tend to line up with @xmath519 .
this is illustrated by the qualitative wave function in section [ sec : lavoisier ] .
( see eqs .
( [ eq : leaningtower ] ) , ( [ eq : ofpisa ] ) . ) both the order three and order two vortices will be visible based on the images of the five spin components .
one can determine the types of the vortices and their axes @xmath384 ( which are encoded in @xmath520 ) from the average magnitude of the densities @xmath521 together with the amplitudes @xmath522 and offsets @xmath523 of the density modulations .
( an order @xmath28 vortex will have @xmath524 and @xmath525 fourier modes in addition to the terms given in eq .
( [ eq : joseph ] ) . )
a possible difficulty with this method arises because , once the five spin components are separated in space , the density oscillations in each of them are no longer stable .
the ensuing dynamics in the clouds could mix the atoms up .
distinguishing among vortices with the same rotation but different phase winding numbers @xmath99 is not possible with this method without resolving the cores .
for example , the vortices @xmath472 and @xmath348 have the same density patterns , since they differ only by an overall phase @xmath526 .
one would also hope to check some predictions about the size and charges of the bound states .
one can select clusters of vortices in an image of the condensate ( if there are not too many vortices ) and use the methods just discussed to identify the vortex charges and check that each cluster satisfies eq .
( [ eq : screened ] ) .
additionally , a sign that the vortex clusters are actually _ bound _ states is that the bound state size depends in the right way on the magnetic field .
now atoms whose ground state is cyclic may be difficult to find ( @xmath45rb is likely to be polar@xcite , though it may be possible to adjust the interaction parameters by applying light fields . ) .
the general considerations of this article also apply to spin 3 condensates ( see ref .
@xcite ) , as well as to spin 1 condensates and pseudospin @xmath357 condensates as already studied by @xcite .
so we will estimate the bound state sizes for the more generic phases as well as for the cyclic phase .
the cyclic phase is unique because the misalignment energy is a second order effect ( see sec .
[ sec : cubeshape ] ) . for other phases where the misalignment energy is a first order effect of the zeeman energy , the effective potential would be on the order of @xmath527 .
hence @xmath528 in contrast , for the cyclic phase , eq .
( [ eq : collie ] ) can be expressed in terms of the scattering lengths as @xmath529 since @xmath530 , the size of the molecules in the cyclic phase is proportional to @xmath531 and the size of molecules in phases with @xmath532 is proportional to @xmath533 .
the size of the condensate must be large enough hold an entire vortex molecule .
substituting @xmath534 from eq .
( [ eq : minimumfield ] ) into eqs .
( [ eq : stbernard ] ) and ( [ eq : babycollie ] ) one finds that @xmath535 ( for either phase ) .
this size is the magnetic healing length of the condensates ( and the size of a vortex core ) and is on the order of @xmath513 m .
the condensate should be narrow in one direction ( so that the behavior of the order parameter is two - dimensional ) but at least several times wider than the magnetic healing length in the other two directions ; in order to measure the field dependence of the molecule sizes , one should be able to decrease the magnetic field by some factor below @xmath534 without the molecules leaving the condensate .
as a side - comment , vortex molecules probably undergo transitions at fields close to @xmath534 ( see fig .
[ fig : whiskers ] ) since the component vortices overlap at larger fields .
absolutely stable vortex molecules , like example 3 in table [ table : ledger ] , will be compressed so that the cores coincide and the vortex becomes rotationally symmetric at a finite field .
once the components cores overlap a little bit , being slightly offset might not lead to significant savings in kinetic energy . on the other hand , when the vortices in a _
metastable _ molecule are squeezed together , they form an unstable tetrahedral vortex .
metastability occurs only when the coulomb " force keeps the component vortices apart .
the evolution of vortex molecules as the magnetic field is increased .
the three curves illustrate how the sizes of the molecules from table [ table : ledger ] , for @xmath26 , might change as the strength of the magnetic field is increased .
the sizes decrease as @xmath531 . at a certain magnetic field
, an absolutely stable vortex will become rotationally symmetric and the component cores will coincide ( molecule 3 ) .
metastable vortices will become unstable when a certain magnetic field is reached , indicated by the x s terminating the curves corresponding to molecules @xmath49 and @xmath250 . ]
we have shown that vortex molecules can be understood reasonably well based only on simple group theory and rough energy estimates .
some of these vortex molecules are actually metastable , and we can study their possible break - up channels , " reminiscent of some of the decay processes in nuclear physics .
( in practice , the molecules will probably escape through the surface of the condensate before any kind of ultracold fusion " can happen ! ) more accurate calculations of the vortex fields and energies could address other interesting questions . as is well - known , unlike the coulomb interaction between vortices in an ordinary scalar superfluid , the interaction energy can not be written as a sum of two - vortex interaction terms , as indicated by our estimate in terms of the energy index .
a more accurate understanding of the kinetic energy landscape might show that the vortices in a molecule can have several spatial arrangements in the core .
another problem that requires more detailed calculations of vortex fields is determining whether a vortex with charge @xmath348 is stable : the energy index estimate shows that it can break up into vortices with only a finite change in energy , but whether the energy _ increases _ or _ decreases _
is not clear yet .
_ acknowledgments _ we wish to thank r. barnett , m. greiner , h. y. kee , j. moore , k. sengstock , d. stamper - kurn , and w. p. wong for useful conversations .
we also acknowledge financial support from cua , darpa , muri , afosr and nsf grant 0705472 .
( [ eq : rotatedfield ] ) is not as difficult to minimize as it appears , because of the special symmetry of @xmath536 .
we must substitute @xmath537 , where @xmath538 into the energy , eq .
( [ eq : rotatedfield ] ) . the perturbation @xmath539 is the deformation of the tetrahedron , measured relative to its body axes .
let us figure out how many powers of the coefficients @xmath254 to keep at each stage of calculating @xmath87 .
it helps to complete the square in eq .
( [ eq : rotatedfield ] ) to get @xmath540 where we note that the chemical potential for the cyclic state is @xmath541 and define @xmath542 .
we also use @xmath543 to stand for @xmath544 .
we need to find the minimum of this energy only to quadratic order in @xmath6 . at the end
we will find that @xmath545 are each _ linear _ in @xmath6 .
since each of the squared quantities in @xmath87 vanishes when @xmath546 , just the linear contributions from @xmath545 give the potential to quadratic order in @xmath6 .
the quadratic zeeman term , since it is multiplied by @xmath6 , also is not needed beyond linear order in @xmath247 .
next find which matrix elements of @xmath536 need to be calculated to evaluate all these contributions to the energy .
only the cross terms between the unperturbed part @xmath25 and the perturbation give linear functions of @xmath247 .
for example , one cross - term contained in the quadratic zeeman contribution is @xmath547 .\end{aligned}\ ] ] expanding this gives a sum of matrix elements such as @xmath548 and @xmath549 .
we need only the matrix elements of products of at most three @xmath104 s .
many of these ( e.g. , @xmath550 when @xmath551 are not all different , and @xmath552 when @xmath153 and @xmath361 are different ) are equal to zero because of the @xmath64 symmetries of @xmath536 around the coordinate axes . the numerical values of the few remaining ones can be worked out quickly . using these matrix elements to calculate all the terms in eq .
( [ eq : rotatedfield2 ] ) produces an expression @xmath553 . along the way ,
one notices that each of the variables @xmath545 contributes to only one term in the @xmath14 potential ( the first line of eq .
( [ eq : rotatedfield2 ] ) ) .
the variables @xmath554 determine the magnetization , @xmath555 determines the density perturbation and @xmath556 and @xmath315 determine the singlet - amplitude @xmath99 .
e.g. , @xmath557=2\sqrt{n_0}e\nonumber.\end{aligned}\ ] ] finally , minimize the potential .
it can be written as a sum of independent quadratic functions of @xmath558 and @xmath315 : @xmath559 .
\label{eq : boardmembers}\end{aligned}\ ] ] ( note that the second term , @xmath560 , is the first - order contribution of the zeeman energy , which is independent of orientation . ) minimizing each quadratic ( which gives @xmath561 ) and combining the results together with the help of eq .
( [ eq : pythagoras ] ) gives @xmath562 all the constant terms can be dropped to give eq .
( [ eq : effective ] ) . note that the magnetization varies with the orientation of the tetrahedron ( as can be checked by substituting the optimal values for @xmath554 into the magnetization , eq .
( [ eq : nah ] ) ) .
in particular , the @xmath26 ground state with @xmath563 has a small magnetization , @xmath564 ; since this has been calculated from @xmath255 , it is the magnetization _ relative to the body axes _ tetrahedron . comparing this to the magnetic field relative to the body axes , @xmath565 ,
shows that the state will become magnetized either parallel or antiparallel to the magnetic field .
the point of the effective potential is that it allows us to eliminate the @xmath248 most rigid degrees of freedom corresponding to @xmath558 and @xmath315 ; then it is easier to understand vortex textures by concentrating on the rephasing and rotation angles as a function of position .
the wavefunction in eq .
( [ eq : manifold ] ) can be parameterized in terms of euler angles for the rotation , e.g. , @xmath566 .
the first angle , @xmath567 , does not come into the angles @xmath568 that describe the orientation of the magnetic field ( @xmath569 ) relative to the tetrahedron .
( the tetrahedron can be rotated around the @xmath23-axis without changing these angles . )
one can check that @xmath570 and @xmath571 . now , working out the kinetic energy and combining it with the effective potential gives the phase - and - rotation - only " energy functional @xmath572\nonumber\\ & + ( c-4)\frac{3q^2}{4c\beta}[\cos^4\tau+\sin^4\tau(\sin^4\rho+\cos^4\rho ) ] .
\label{eq : anotherlandslide}\end{aligned}\ ] ] this can be solved ( in principle ) to give the textures around sets of vortices and the relative positions of the vortices in equilibrium .
each vortex type implies a certain type of discontinuity in the four angles as the core is encircled .
this expression does not seem too easy to use , but at least it shows just the two effects we have been balancing against one another ( kinetic and anisotropy energy ) .
the size of a vortex molecule can be estimated by assuming that the two terms are comparable , @xmath573 .
if distances are rescaled by @xmath22 , we then find that the energy function has a form that depends only on the _ sign _ of @xmath574 .
therefore , in a molecule with three vortices , the angles of the triangle they form will be independent of all the parameters , including @xmath44 , even though it is dimensionless .
( [ eq : anotherlandslide ] ) is derived from eq .
( [ eq : howtousebatteries ] ) by determining how the tetrahedra are distorted by the quadratic zeeman effect .
but the kinetic energy also causes distortions of the wave function from the perfect tetrahedral forms , and it seems possible that these distortions could lead to kinetic effects in the anisotropy term and anisotropy effects in the kinetic energy term .
however , at the lowest order , treating the two terms independently seems correct .
a simple argument for this ( neglecting the kinetic energy when finding the anisotropy potential ) is that the distortion due to the zeeman term is linear in @xmath6 while the distortion due to the kinetic energy is quadratic in @xmath6 . to see this ,
think of an ordinary scalar vortex , where the density varies as @xmath575 far from the core@xcite .
the amount of distortion " is @xmath576 . in the cyclic state ,
distortion ( i.e. , perturbations to the spinor components that take it out of @xmath13 ) implies changes in the magnetization as well as the density .
but we may assume that these distortions are still of order @xmath577 .
the majority of the pulp " in a molecule s core consists of points whose distance is of order @xmath22 from the actual vortex cores ( the seeds " ) , so the amount of distortion can be found by substituting @xmath578 from eq . ( [ eq : betterlate ] ) .
using the relation between @xmath91 and @xmath47 shows that the fractional distortion in the
pulp " regions is of order @xmath579 , to be compared with the deformations of order @xmath580 that result from minimizing eq .
( [ eq : boardmembers ] ) .
to give a complete description of charge conservation when the charges are described by the noncommuting rotations of a tetrahedron , one needs to give a rule for how to multiply the charges of a set of vortices together to get the net charge .
a convention we used is to multiply the topological charges together in order of increasing x - coordinates .
it seems that this definition has an awkward consequence : does the net vortex charge jump suddenly when two of the vortices are reordered , because of the noncommutativity of the group of charges ? in fig .
[ fig : marigolds ] vortices @xmath49 and @xmath28 are interchanged between frames a ) and c ) , which suggests that the net charge changes from @xmath581 to @xmath582 . but
this deduction is incorrect , and the net vortex charge _
is _ actually conserved .
the resolution of the paradox has to do with the fact that the charge of a vortex can only be determined up to conjugacy , unless one introduces a systematic convention .
for example , the charge of a @xmath84-rotation vortex is ambiguous
the rotation could be either @xmath199 , @xmath102 , @xmath75 , or @xmath200 , and there is no way to distinguish between these because the four vertices of the tetrahedra are indistinguishable .
( abstractly speaking , the four rotations are conjugate elements of the group . ) in order to identify the charge of each vortex , we have to choose a routine for labelling the vertices of the tetrahedra nearby . here is a convention that is consistent with the rule for ordering the vortex charges .
take a point @xmath583 far below all the vortices in the system and connect it with lines to points just below the vortices ( see fig .
[ fig : marigolds]a ) .
now identify the _ base _ tetrahedron at @xmath583 with the standard tetrahedron in fig .
[ fig : bottler]a , making a choice from among the twelve possible ways .
the labelling at @xmath583 can be communicated to the tetrahedron at the end - point of each line , by copying the labelling from @xmath583 to a nearby tetrahedron on the line , and then continuing to copy the labelling until the end of the line is reached .
now the charges of the vortices can be identified by using the labelling of the nearby tetrahedron to assign a letter to the rotation axis .
now that we have a consistent convention for assigning vortex charges , we can show that the net charge of a set of vortices does not change when two of them are interchanged .
the trick is that the charges of the _ individual _ vortices do change in such a way that the product charge does not change ! between fig .
[ fig : marigolds]a and fig .
[ fig : marigolds]c , vortex @xmath28 is moved over vortex @xmath49 . because vortex @xmath28 s tether gets tangled up with vortex @xmath49 when vortex @xmath49 passes below it , its charge gets redefined , as @xmath584 .
the other two vortices charges do not change .
the net charge , obtained by multiplying the vortex charges from left to right , is @xmath585 thus the net charge does not change . on the other hand ,
there is a sudden jump in the charge of vortex @xmath28 , but this does not mean that the fields of tetrahedra are changing suddenly ; the vortex has just been reclassified , with a vortex charge that is conjugate to the original charge . here is an interesting consequence of the noncommuting charges : the force between a pair of vortices changes from repulsive to attractive if a third vortex wanders between them . in fig .
[ fig : halfmagic ] one vortex ( the @xmath199 one ) catalyzes a reaction without touching the other two vortices involved .
( assume the phases are @xmath96 for the two @xmath52 vortices and
@xmath323 for the @xmath199 vortex . ) to figure out what happens , keep track of when a vortex s connection to the reference point ( below the figure ) is interrupted by another vortex .
the charge of the vortex passing underneath is not changed , and the charge of the vortex on top changes to keep the total charge the same .
this information is sufficient for working out all the charges : when @xmath199 passes below the @xmath52 on the right , the latter vortex changes to a @xmath586 , so that the net charge is still the same , even though @xmath199 has moved .
( this can be used to work out the charges : the net charge of the two vortices which have switched has to be the same , so @xmath587
so @xmath586 . )
next , when @xmath199 passes _ above _ the @xmath52 on the left , the _
former _ vortex changes to a @xmath588 vortex .
now the @xmath7 matrices for @xmath52 and @xmath199 are @xmath589 and @xmath590 ( using the axes associated with @xmath536 ) . multiplying out the charges shows that @xmath591 .
the force between the original pair of vortices ( say the @xmath199 is far away at the beginning and end ) is @xmath592 and the force between the vortices they turn into is @xmath593 .
( @xmath79 is the distance between the vortices . )
the vortices repel each other at first , but after @xmath199 s intervention , they attract each other , as one sees by checking that @xmath594 , and that @xmath595 is a @xmath97 rotation with energy index @xmath28 while the other reacting charges are all @xmath391 rotations , with index @xmath357 .
we have been assuming that @xmath14 , but a similar reaction could also happen in the core of a composite vortex when @xmath16 ; that is why one has to check all the possible ways for vortices to wander between one another before coalescing or before dividing into groups . | condensates of atoms with spins can have vortices of several types ; these are related to the symmetry group of the atoms ground state .
we discuss how , when a condensate is placed in a small magnetic field that breaks the spin symmetry , these vortices may form bound states . using symmetry classification of vortex - charge and rough estimates for vortex interactions , one can show that some configurations that are stable at zero temperature can decay at finite temperatures by crossing over energy barriers .
our focus is cyclic spin 2 condensates , which have tetrahedral symmetry . in an image of a nematic liquid crystal by polarized light
, one can identify topological defects of various topological charges ( see ref .
@xcite ) .
bose condensates ( see the books@xcite ) are starting to provide another context for studying topological defects : in the texture formed by the phase _ and _ spin of a spinor condensate@xcite .
( see also @xcite for reviews of the theory and experimental techniques being applied to spinor condensates . )
a topological defect in the _ phase _ _ of a superfluid _ is a quantized vortex .
the discontinuity in the phase as the defect is encircled must be @xmath0 for an integer @xmath1 , and the circulation of the vortex is then @xmath2 . in a single - component superfluid , multiply quantized vortices ( @xmath3 ) are usually not stable .
the widely known explanation is that the energy of a vortex is proportional to @xmath4 .
thus a doubly quantized vortex ( @xmath5 ) can lower its energy by splitting into two singly quantized vortices .
similar arguments can be formulated for multicomponent condensates , but we will find that some vortices in these condensates can be very long - lived in spite of having large energies .
our predictions are about condensates of atoms with spin in which the rotational symmetry has been weakly disrupted by a small parameter @xmath6 , such as the interaction between the spins and the magnetic field .
( we focus on the _ cyclic _ condensates of spin 2 atoms , see ref .
@xcite . )
long - lived multiply quantized vortices are particular examples of composite vortices , which are made up of several vortices bound together .
ground states with complicated symmetries have many types of vortices@xcite .
symmetry violating fields provide a force that can bind some groups types vortices together so that they form a composite core " for a larger vortex .
the composite vortex might have two quanta of circulation ( or have some other extra - large topological charge ) .
such a vortex would be surprising if observed under a low resolution , because the smaller vortices that make it up would be hidden in its core ; the behavior of the order parameter at infinity is determined entirely by the _ net _
topological charge .
some earlier theories about vortices in multicomponent condensates also describe vortices with asymmetrical cores which can be regarded as collections of closely spaced vortices .
the parameter @xmath6 makes this interpretation even more meaningful : when @xmath6 is very small , we find that the component vortices move far apart , so that their cores do not overlap , while still remaining bound . ref .
@xcite describes vortices that can occur in a condensate of two atomic states when there is an rf - field producing coherent transitions between the states .
the binding of these vortices also comes from an asymmetry , but the asymmetry comes from the dependence of the interaction strength on the internal states of the atoms , rather than from an external magnetic field . ref .
@xcite studies vortices of spinor atoms in a magnetic field , like us , but focuses on spin 1 , and describes composite vortices as well .
since the scenario involves rotation as well as a magnetic field , the vortices would be held close to the _ axis _ of the condensate by rotational confinement without the magnetic field . with just a magnetic field , we show that the vortices are attracted to _ one another_. vortices and bound states are not hard to picture , by taking advantage of the fact that the state of a spinor atom can be represented by a geometrical figure .
the appropriate shape depends on the type of condensate . for a ferromagnetic condensate ,
a stake pointing in the direction of the magnetization could represent the local state of the condensate .
other condensates can be represented by more complicated shapes .
now imagine a plane filled with identical shapes ( tetrahedra , for the cyclic phase ) , with orientations varying continuously as a function of position .
this shape field ( or spin texture " ) together with a phase field would represent a nonuniform state of a condensate .
if the shapes rotate around a fixed symmetry axis as some point is encircled then the spin texture has a topological defect at this point .
such configurations generalize vortices , because some of them are accompanied by persistent spin or charge currents . for each symmetry of the tetrahedron
, there will be a vortex when the hamiltonian is @xmath7 symmetric .
( each discrete subgroup of @xmath7 describes the vortices of some phase for atoms of some spin@xcite ; to find vortices for more even more complicated groups like @xmath8 or @xmath9 , one might want to study gases of spin @xmath10 atoms@xcite . )
the vortex spectrum is decimated when a magnetic field is applied , since the field favors a particular orientation of the order parameter e.g .
, tetrahedra may want to have an order - three axis aligned with the magnetic field .
the spectrum of vortices is then reduced from the full spectrum of tetrahedral " vortices ( based on arbitrary symmetries of the tetrahedron ) to field - aligned " vortices , where the tetrahedra must rotate around this order three axis , so as not to lose their alignment with the magnetic field .
we may introduce the ground state space @xmath11 ( of tetrahedra aligned with the magnetic field ) , where the interaction energy is @xmath12 , and the space of ground states of the @xmath7 invariant part of the hamiltonian , @xmath13 ( of tetrahedra with arbitrary orientations ) .
a set of vortices can be assigned a topological charge based on the loop traced out in one of these spaces by the values of the order parameter on a circle around the set .
specifically , the topological charge describes the symmetry transformation that brings the order parameter back to its initial value as the set of vortices is circumambulated .
( the relation between topology and symmetry is presented in @xcite . ) from the general point of view , the reduction of the charge types from the tetrahedral to the field - aligned ones when @xmath6 is turned on results because the smaller ground state space at nonzero @xmath6 , @xmath11 , has fewer closed loops also stabilizes vortices that would be unstable at @xmath14removing a portion of @xmath13 can leave a new hole for a vortex - circuit to wrap around . ] .
if @xmath6 is very small , the excess energy of an order parameter in @xmath13 rather than @xmath11 is small .
there is then a hierarchy of spaces @xmath11 , @xmath13 and @xmath15 ( the whole hilbert space , corresponding to arbitrarily _ distorted _ tetrahedra ) with increasing energy scales . when @xmath16 , a vortex has to have a field - aligned charge _ at infinity _ because the tetrahedra must eventually move into @xmath11 to avoid too big of an energy cost .
but if @xmath6 is small , the hierarchy of the order parameter space is reflected in the fact that a vortex with a field - aligned charge can have a _
composite _ core that may contain tetrahedral vortices , as in fig .
[ fig : cherimoya ] ; the core is sort of like the pulp and a composite vortex reflects the hierarchy of the order parameter space . in the white , grey , and black regions the order parameter moves from @xmath11 to @xmath13 to @xmath15 which has the highest energy of all .
the charges of the subvortices are tetrahedral charges , represented by @xmath17 and the charges of the composite vortex is a field - aligned " charge . ]
seeds of a fruit . inside
the fruit is a texture of arbitrarily oriented tetrahedra , almost as if @xmath6 were equal to zero .
the seeds are then the cores of tetrahedral vortices and the pulp is qualitatively the same as the texture that would surround these vortices in the absence of the magnetic field .
the net charge of the vortices has to be field - aligned so that the loop of order parameters that surrounds the whole core can move into @xmath11 as @xmath18 .
this picture becomes more accurate as @xmath6 becomes small , since the vortex cores within the composite core are far apart in that case the size of a region where the order parameter is in @xmath19 or @xmath20 is inversely proportional to the energy scale for each space . as @xmath21 ,
the binding of the tetrahedral charges becomes weaker and weaker , until they become free from each other ; each vortex therefore is described by independent degrees of freedom when @xmath14 .
the optimal size @xmath22 of a composite vortex results from competition between two forces a confinement force from the anisotropy term and the familiar logarithmic interaction of vortices .
the symmetry breaking energy favors minimizing the area over which the order parameter leaves @xmath11 , pushing the component vortices toward one another . on the other hand
, it can not compress them to a point since the coulomb - like repulsion one expects of vortices keeps them apart .
some of these vortex molecules will turn out to be metastable .
ref .
@xcite mentions an interesting clue to such a phenomenon ; namely there are multiple steady state wave functions describing a condensate with a given magnetization and rotational frequency ; these local minima of the energy function can maybe be analyzed using the group theoretic metastability conditions we discuss in section [ sec : metastable ] . for a spinor condensate ,
wave functions for states besides the ground state are experimentally important , since the experiments of ref .
@xcite , as well as the liquid crystal experiments of ref .
@xcite , reveal complicated textures produced by chance ; an initial fluctuation around a uniform excited state becomes unstable and evolves into an intricate _ nonequilibrium _ texture .
so it is useful to analyze spin textures which are only _ local _ minima of the gross - pitaevskii energy functional ( like the metastable vortex molecules considered here ) as well as unstable equilibria ( which take a long time to fluctuate out of their initial configuration ) .
examples of unstable equilibrium are described for single - component condensates by @xcite .
the process by which textures form out of uniform initial states has been discussed in theoretical articles , including @xcite ( on the statistics of the spin fluctuations and vortices that are produced from this random evolution ) , @xcite ( on the spectrum of instabilities ) and @xcite on the dynamics of spinor condensates .
the experiments described in @xcite show that the patterns that evolve in rubidium condensates are probably affected by dipole - dipole interactions , though we are not considering these .
dipole - dipole interactions lead to antiferromagnetic phases@xcite , which maybe can be described as _
ground state _ configurations of vortices .
besides just hoping for unusual types of vortices to form , one can make a vortex lattice by rotating a condensate .
the effects of rotating on a spinor condensate have been investigated theoretically in @xcite as well as the review @xcite .
experiments can also make a single vortex of a prescribed type@xcite .
excited by these possibilities , physicists have come up with several types of vortices and topological defects they would be interested in seeing : skyrmions@xcite , monopoles@xcite , textures whose order parameter - field lines make linked loops@xcite , as well as the noncommutative vortices of the cyclic phase that we will be expanding on here@xcite .
geometrical representation of the tetrahedral superfluid phases of spin 2 atoms .
two orientations of a tetrahedron , corresponding to the spin roots of the ground state spinor , are illustrated and the symmetry axes are labelled .
( the orientation of the coordinate axes , shown on the side , is the same in both cases . )
a ) an orientation with the vertical @xmath23-axis along an order three axis , corresponding to the ground state @xmath24 ( see eq .
( [ eq : order3 ] ) ) .
b ) an orientation with the @xmath23-axis along an order two axis , corresponding to the ground state @xmath25 ( see eq .
( [ eq : chi2 ] ) ) . for a magnetic field along the @xmath23 axis ,
orientation a ) is preferred when @xmath26 and orientation b ) is preferred when @xmath27 .
note that the three order @xmath28 axes , labelled @xmath29 can be used as a set of three orthogonal body axes . ]
the key to our discussion of vortices will be a geometrical representation of the order parameter , allowing us to visualize a texture of the cyclic phase as a field of tetrahedra with different orientations .
( see fig .
[ fig : bottler ] . ) without this representation , a spin texture would be given by a spinor field @xmath30 ; the fact that this spinor lies in the ground state manifold @xmath13 would have to be described by a set of polynomial relations between the five components . a more revealing way to represent
a spinor is to draw a geometrical figure consisting of spin - roots " ( as in @xcite ) , and in the cyclic phase , these spin roots form the vertices of a tetrahedron .
( a similar construction can be used to classify vortices in condensates of spin 3 atoms , see ref .
@xcite . ) even without using the spin - root interpretation , one can justify using tetrahedra to represent order parameters in the cyclic phase because they are a concrete way of representing the symmetry of this phase .
ref .
@xcite worked out the symmetry group of a state in the cyclic phase by finding all the pairs @xmath31 such that @xmath32 where @xmath33 the spinor @xmath24 is a representative cyclic state , if @xmath34 is the density of the condensate .
these symmetries are the same as the symmetries of a tetrahedron oriented as in fig .
[ fig : bottler]a ; hence we may represent the state @xmath24 by this tetrahedron .
any other ground state should be represented by a tetrahedron oriented so that its symmetry axes and symmetry axes of the spinor coincide .
the appropriate orientation of the tetrahedron for a given ground state can be determined in an automatic way by calculating the spin roots .
the hamiltonian for spin 2 atoms in a magnetic field , @xmath35 $ ] has a simple expression@xcite in terms of the density @xmath36 , magnetization @xmath37 , and singlet - pair amplitude @xmath38 ( @xmath39 stands for the time reversal of @xmath40 . ) @xmath41 the first three terms describe the rotationally symmetric interactions of pairs of atoms .
the first one describes repulsion between a pair of atoms and the next two terms describe additional , smaller interactions , that depend on the spin states of the two colliding atoms .
these terms determine the spinor ground state in the absence of a magnetic field@xcite .
( the spin - dependent interaction strengths @xmath42 can be expressed in terms of the scattering lengths . )
the properties of spin 2 atoms , which are described by this hamiltonian , and of spin 1 atoms have been investigated experimentally in refs .
@xcite ; @xcite reviews more experimental phenomena . ref .
@xcite found values for @xmath43 and @xmath44 for @xmath45rb that are consistent with theoretical predictions , although even the sign of @xmath44 is not known for sure , because @xmath44 is small .
the final term contains the chemical potential @xmath46 .
if @xmath47 and @xmath44 are positive , the ground state of a condensate of spin 2 atoms is cyclic .
the deformation of a cyclic state due to a magnetic field is the simplest if we assume that @xmath48 and that @xmath44 is on the order of @xmath49 .
we therefore assume @xmath50 , though @xmath51 for rubidium .
the magnetic field influences the atoms through the fourth term of the energy , breaking the rotational symmetry of the hamiltonian ; this term describes the quadratic zeeman shift due to a magnetic field @xmath52 along the @xmath23-axis and @xmath53 .
( see @xcite for the explanation of why the quadratic zeeman term @xmath6 but not the linear term is relevant if the condensate s initial magnetization is zero .
a nonzero magnetization is described by a lagrange multiplier term @xmath54 , which looks like a linear zeeman coupling .
we assume @xmath55 ; a nonzero @xmath56 should have similar consequences as a nonzero @xmath6 , since both break the rotational symmetry . ) introducing a small @xmath6 reduces the ground state space from @xmath13 to @xmath11 .
this can be explained ( see sec .
[ sec : cubeshape ] ) by finding the modulation of the energy of an arbitrary tetrahedral state as a function of the orientation of the corresponding tetrahedron : @xmath57 where @xmath58 and @xmath59 are the angles between the @xmath23-axis and three body - axes @xmath29 fixed to the tetrahedron ( see fig .
[ fig : bottler]a ) .
the orientations that minimize this energy are the true ground states at nonzero @xmath6 . if @xmath26 ( assumed until sec .
[ sec : sampler ] ) , then the tetrahedron in its ground state is oriented with the @xmath23-axis perpendicular to a face as in fig .
[ fig : bottler]a , with a ground state energy of @xmath60
. thus the absolute ground state space @xmath61 contains all the wave functions that are arbitrary rotations about the @xmath23-axis ( combined with rephasings ) of @xmath24 .
when @xmath27 , the ground states @xmath62 are rotations about @xmath23 of the tetrahedron illustrated in fig .
[ fig : bottler]b .
in particular , when @xmath16 , there is a phase transition at @xmath63 , though there is nothing special about @xmath63 in _ zero _ magnetic field . in this paper , we will mostly assume that @xmath26 because our initial example occurs under this condition .
consider first a single vortex associated with the rotation through @xmath64 of the ground state tetrahedron ( see fig .
[ fig : bottler]a ) about the @xmath65 axis , @xmath66 .
such a vortex is described at large distances by @xmath67 where @xmath68 are polar coordinates centered on the core of the vortex .
this vortex has an excess energy ( relative to the ground state energy of the condensate ) which diverges with the condensate size . in fact , its zeeman energy density ( with @xmath69 subtracted ) is given according to eq .
( [ eq : effective ] ) by s for this vortex may be calculated by noting that , relative to the tetrahedron , the @xmath23-axis of space rotates @xmath64 about @xmath65 so that @xmath70 , its coordinates relative to the body axes of the tetrahedron are easy to work out as functions of @xmath71 . ]
@xmath72 the tetrahedron is pointing in the wrong direction except along the positive and negative @xmath73-axis ; and the integrated energy is proportional to the area : @xmath74 where @xmath75 is the condensate s radius
. such a vortex can not be the only vortex in an infinite condensate !
the vortex described by eq .
( [ eq : bisect ] ) can form a partnership with another vortex of the same type , producing a molecule that can exist without costing too much energy .
this is because the net charge of the two vortices is compatible with the magnetic field .
the combination of two @xmath64 rotations of the tetrahedral order parameter about the @xmath65 axis is a @xmath76 rotation about the @xmath65-axis , but since any rotation axis is a @xmath76 symmetry of the tetrahedron , the rotation axis on circles of radius @xmath77 enclosing the vortices can be tilted continuously relative to the tetrahedron as @xmath77 increases until it becomes the @xmath75 axis instead .
then the tetrahedra align with the magnetic field far from the vortices and stay in @xmath78 .
each vortex screens the part of the other vortex s charge that produces the large zeeman cost , as illustrated in fig .
[ fig : murky ] .
illustration of the energy costs in a vortex molecule .
the energy in the region occupied by the two vortices is dominated by the quadratic zeeman cost and the energy outside it is dominated by kinetic energy .
increasing the molecule size decreases the kinetic energy and increases the zeeman energy .
the equilibrium size @xmath79 is determined by minimizing the sum . ] the zeeman energy in the region around the vortices , where the tetrahedra are still tilted , may be estimated by replacing the total condensate size @xmath75 in eq .
( [ eq : misalign ] ) by the diameter @xmath79 of this region .
the zeeman energy tries to pull the vortices toward one another but the elastic energy cost of rapid changes in the order parameter ( the gradient term in the hamiltonian ) opposes this tendency : the tetrahedra rotate twice as fast around circles beyond @xmath79 , where the two vortices act in concert .
therefore the elastic energy increases as @xmath79 becomes smaller ; this leads to the coulomb repulsion between the vortices , @xmath80 , where @xmath81 is the mass of the atoms in the condensate .
the energy of the vortex molecule is therefore @xmath82 where @xmath83 is a numerical constant .
the equilibrium size can be determined by minimizing over @xmath79 .
the quadratic zeeman force binds this molecule together while the repulsion keeps the vortices from merging .
if the vortices _ were _ to coalesce , then they could react to form a set of vortices that are not bound by the zeeman energy ; one possible set of decay products is three vortices each involving a rotation through @xmath84 about the @xmath75 axis ( which have the same net @xmath76 rotation as the original pair of vortices ) .
the vortex molecule of the two @xmath65 rotations is _ metastable _ because thermal fluctuations may overcome their coulomb repulsion and push them together , leading to such a fission process .
the rest of this paper elaborates : it gives a qualitatively correct expression for the spin texture surrounding the molecule just described and determines how this molecule _ actually _ decays . in order to do this
, we will give some more general results : first section [ sec : dali ] , summarizes the non - commutative group theory of combining vortex charges and a classification of the tetrahedral vortices ( the vortices that occur by themselves when @xmath14 and in clusters when @xmath16 ) ; then section [ sec : carnot ] estimates the elastic energies and zeeman energies of such clusters as functions of these vortex charges ; finally section [ sec : lavoisier ] gives criteria determining which types of tetrahedral vortices form bound states or metastable states ( see section [ sec : metastable ] ) .
the last two sections illustrate the criteria with a few additional surprising examples ( see section [ sec : sampler ] ) and give some basic ideas about how to observe metastable vortices ( section [ sec : fieldwork ] ) . |
for @xmath166 , @xmath167 , and @xmath168 the functions @xmath48 and @xmath47 are given by @xmath169 \
, , \end{aligned}\ ] ] @xmath170 \ , , \end{aligned}\ ] ] @xmath171 \ , , \end{aligned}\ ] ] @xmath172 \ , , \end{aligned}\ ] ] and @xmath173 and @xmath174 . | we study the spin polarization and its associated spin - hall current due to electric - dipole - induced spin resonance in disordered two - dimensional electron systems .
we show that the disorder induced damping of the resonant spin polarization can be strongly reduced by an optimal field configuration that exploits the interference between rashba and dresselhaus spin - orbit interaction .
this leads to a striking enhancement of the spin susceptibility while the spin - hall current vanishes at the same time .
we give an interpretation of the spin current in geometrical terms which are associated with the trajectories the polarization describes in spin space .
the ability to coherently control the spin of charge carriers in semiconductor nanostructures is the main focus of spintronics@xcite .
band - structure and confinement effects in these systems lead to a strong spin - orbit interaction ( soi ) offering the possibility to efficiently access the charge carrier spin via the control of its orbital motion@xcite .
a versatile and efficient scheme of spin control is electric dipole induced spin resonance ( edsr)@xcite where electric radio frequency ( rf ) fields give rise to internal fields coupling to the spin .
choosing an adequate configuration of the electric rf fields and a static magnetic field defining a quantization axis for the spin , arbitrary spin rotations can be realized .
this is analogous to standard paramagnetic spin resonance techniques , has the advantage , however , that it can be integrated in gated nanostructures thereby avoiding magnetic rf coils . in a two - dimensional electron gas ( 2deg ) with pure rashba
soi the amount of spin polarization which can be achieved by edsr is severely limited by disorder@xcite .
similar limitations are found for pure dresselhaus soi .
however , if both dresselhaus and rashba soi are present interference between the two soi mechanisms can occur and qualitatively new behavior emerges , such as anisotropy in spin relaxation@xcite and transport@xcite . for spin relaxation
this anisotropy is most pronounced if both sois have equal strength . in this case , the spin along the @xmath0 $ ] direction [ see fig .
[ fig : vectorfields ] ] is conserved@xcite , and the associated spin relaxation rates vanish , whereas they become maximal along the perpendicular direction @xmath1 $ ] . for the driven system considered here
we show that similar interference effects occur and that not only the internal rf field but also the edsr linewidth becomes dependent on the direction of the magnetic field . in a microscopic approach we show then that due to this dependence an optimal configuration exists where the linewidth and the internal field simultaneously become minimal and maximal , resp . , and that , as a remarkable consequence , the spin susceptibility gets dramatically enhanced . in other words , this optimal configuration allows one to obtain a high spin polarization with relatively small electric fields and thus making the power consumption for spin polarization minimal . due to spin - orbit interaction angular momentum can be transferred between spin and orbital degrees of freedom .
this fact leads , in particular , to a dynamical coupling between spin and spin current described by the heisenberg equation of motion@xcite . exploiting this coupling
we show that the spin current can be interpreted in geometrical terms : the spin dynamics generated by the rf fields describes an elliptical trajectory .
the spin - hall conductivity can then be expressed entirely in terms of the semi - minor and semi - major axis and the tilt angle of this ellipse .
since the spin dynamics ( trajectories ) is experimentally accessible , for instance with optical methods@xcite , this opens up the possibility for a direct measurement of the spin - hall current .
finally , we find that for the optimal configuration the spin current vanishes , in stark contrast to the spin polarization which , as mentioned , becomes maximal .
we consider a non - interacting 2deg consisting of electrons with mass @xmath2 and charge @xmath3 which are subject to a random impurity potential @xmath4 .
we take into account linear soi @xmath5 of the rashba - and dresselhaus type where @xmath6 are the pauli matrices and @xmath7 is the canonical momentum . taking the coordinate axes along the @xmath8 , [ 010 ] , [ 001]$ ] crystallographic directions , the internal magnetic field @xmath9 is then given by ( cf .
fig.[fig : vectorfields ] ) @xmath10 where @xmath11 and @xmath12 is the strength of the rashba and dresselhaus soi , respectively .
additionally , the external static magnetic field is given by @xmath13 with @xmath14 , and the external electric rf field by @xmath15 , where @xmath16 are the angles enclosed with the @xmath17 $ ] direction .
the system is described by the hamiltonian @xmath18 where @xmath19 is the vector potential associated with @xmath20 and @xmath21 with @xmath22 the electron g - factor and @xmath23 the bohr magneton .
_ spin polarization .
_ we turn now to the calculation of the spin polarization ( magnetization@xmath24 ) per unit area , @xmath25 , evaluated in linear response to an applied electric field @xmath26/2 $ ] and in the presence of both rashba and dresselhaus soi . due to the interference between these two soi mechanisms
we need to carefully extend earlier calculations@xcite , which were restricted to rashba ( dresselhaus ) soi only , to this new situation . we will then be able to identify a configuration that allows one to obtain a maximum degree of spin polarization due to this interference .
working in the linear response regime , @xmath27 is obtained from a kubo formula averaged over the random distribution of impurities in the 2deg .
we evaluate this average with standard diagrammatic techniques assuming the impurities to be short - ranged , isotropic and uniformly distributed . in this case , the impurity average @xmath28 is @xmath29 - correlated and proportional to the momentum relaxation time @xmath30 .
we further take the fermi energy @xmath31 to be the largest energy scale in the problem . then , to leading order in @xmath32 with @xmath33 the mean free path , the averaged spin polarization is given by the diffuson diagram , giving rise @xcite to a correction @xmath34 of the spin vertex ( cf . ref .
@xcite ) in the kubo formula .
thus , the spin susceptibility defined by @xmath35 is given by @xmath36 \omega_{kj } \ , , \end{aligned}\ ] ] where @xmath37 is the two - dimensional density of states and @xmath38 .
we evaluate the vertex correction @xmath39 of eq.([eq : spin - pol - rd ] ) for the case of a magnetic field @xmath40 that is large compared to the internal fields induced by soi .
this regime is characterized by @xmath41 and @xmath42 .
the components @xmath43 with @xmath44 of the vertex correction are then found to be given by @xmath45 where @xmath46 . here , the functions @xmath47 and @xmath48 are second order in @xmath49 and @xmath50 , and depend on the frequency @xmath51 , the larmor frequency @xmath52 , and the angle @xmath53 and @xmath48 are too lengthy to be written down here . in the following , only @xmath54 , @xmath55 and the linear combination @xmath56 are relevant for the linewidth and the spin - hall conductivity , resp . , which are explicitly given below in eqs.([eq : damping2 ] ) and ( [ eq : spin - current - explicit ] ) . ] . in the edsr system ,
pauli paramagnetism gives rise to a constant equilibrium polarization @xmath57 along @xmath58 which is independent of the electric field .
the polarization dynamically generated by @xmath59 , however , depends on the amplitude of the oscillating internal field perpendicular to @xmath58 .
it is thus instructive to consider the longitudinal ( along @xmath60 ) and the transverse ( along @xmath61 and @xmath62 ) polarization components given by @xmath63 and @xmath64 , @xmath65 , resp . as a result
, we find the polarization @xmath66 in terms of the transformed susceptibility @xmath67 .
to lowest order in @xmath68 only the transverse components ( @xmath69 ) are finite .
they are given by @xmath70 \notag \\ & \times w_{i } \left ( \frac{1}{\omega_l - \omega + \delta \omega -i\gamma } + \frac{1}{\omega_l + \omega -\delta \omega + i\gamma } \right ) , \end{aligned}\ ] ] where @xmath71 for the in - plane ( @xmath72 ) and @xmath73 for the out - of - plane component ( @xmath74 ) , and @xmath75 is proportional to the drude conductivity ) we can identify the component of the internal rf field @xmath76 ( which effectively drives the spin dynamics ) in terms of the electrically induced momentum drift @xmath77 .
we find that @xmath78 is given by the projection of the internal rf field ( induced by @xmath79 ) on the transverse direction @xmath80 .
note that due to disorder scattering the fourier transform @xmath81 and @xmath82 are phase - lagged with respect to @xmath83 . ] .
close to resonance the scattering from disorder leads to a renormalization of the magnetic field dependence .
the resonance is shifted by a term @xmath84 \frac { \omega_l \tau}{1 + ( \omega_{l } \tau)^2 } , \notag\end{aligned}\ ] ] corresponding to an effective g - factor which depends both on the amplitude and the orientation of the magnetic field . the linewidth @xmath85 of the resonance peak is given by @xmath86}\right ] .\notag\end{aligned}\ ] ] note that in @xmath85 the rashba and dresselhaus soi do not simply add up but can interfere with each other , enabling a strong enhancement of the susceptibility as we will see next . in fig .
[ fig : vectorfields ] we plot the spin susceptibility at resonance , @xmath87 /\gamma$ ] for the case @xmath88 measured in @xcite .
the angle @xmath89 has been tuned to maximize @xmath90 which displays a pronounced dependence on the magnetic field direction . in eq .
( [ eq : damping2 ] ) we note that @xmath85 scales with the mean square fluctuations of the internal magnetic fields @xmath91 and @xmath92 , where @xmath93 denotes a uniform average over all ( in - plane ) directions @xmath94 .
comparison with a simple model@xcite of spin relaxation ( bloch equation ) shows that the first term in eq .
( [ eq : damping2 ] ) comes from pure dephasing , i.e. from disorder induced fluctuations of the internal fields along @xmath58 , while the second term is due to fluctuations along @xmath80 . choosing a configuration with @xmath95 and tuning the soi strengths to @xmath96 the first term vanishes while the second is subject to narrowing due to the magnetic field .
the width becomes @xmath97 $ ] where @xmath98 is the dyakonov - perel spin relaxation rate for rashba soi .
increasing the frequency such that ( at resonance ) @xmath99 will lead to an increase of the inverse width @xmath100 and , hence , of the susceptibility at resonance , given by @xmath101 for comparison , we find the ratio to the resonance susceptibility @xmath102 in the pure rashba case as @xmath103 $ ] growing quadratically with @xmath104 .
thus , the spin polarization can be substantially enhanced by tuning the sois to equal strengths and by increasing the magnetic field .
finally , the range of validity for the linear response regime can be estimated as follows . assuming full polarization ( @xmath105 ) and parameters for a gaas 2deg @xcite with spin - orbit splitting @xmath106 , fermi wavelength @xmath107 and @xmath108
, we find from eq .
( [ eq : pol - res ] ) that the linear response is valid for electric fields with amplitudes up to @xmath109 . _ polarization and spin current .
_ we consider the spin current defined by @xmath110 .
using the heisenberg equation of motion the spin current components @xmath111 and @xmath112 along @xmath80 and @xmath113 can be expressed in terms of the polarization at frequency @xmath51 as @xmath114 ( i \omega s'^1 + \omega_l s'^3 ) - i \omega \beta \cos ( 2 \theta ) s'^2\\ ( \alpha + \beta \sin ( 2 \theta ) ) i \omega s'^2 - \beta \cos ( 2 \theta ) ( i \omega s'^1 + \omega_l s'^3 ) \notag \end{array } \right ) \ , .\end{aligned}\ ] ] we consider the configuration @xmath115 such that the soi induced internal rf field is perpendicular to @xmath58 and the longitudinal component @xmath116 is not altered in linear response in @xmath117 . note that in this case eq.([eq : spin - current - rd ] ) simplifies such that @xmath118 .
this relation differs from the naive model of an average spin - orbit field equating the internal field @xmath119 with its average @xmath120 .
contrary to eq.([eq : spin - current - rd ] ) , we then find @xmath121 where @xmath122 is a phenomenological transverse relaxation rate .
discrepancies to the model of an averaged spin - orbit field occur similarly for other effects such as the generation of an out - of plane polarization@xcite and zitterbewegung@xcite .
we proceed by evaluating the spin - hall current @xmath111 in terms of the vertex correction eq.([eq : vertex - correction ] ) which was obtained in the diagrammatic approach and is valid up to second order in @xmath68 .
the linear combination @xmath123 cancels in lowest order ( cf .
eq.([eq : sus3trans ] ) ) such that @xmath111 is given by the second order terms @xmath124 . from eq.([eq : spin - pol - rd ] ) and eq.([eq : spin - current - rd ] ) we find the spin - hall conductivity , defined as @xmath125 , to be given by @xmath126 remarkably , for high frequencies @xmath127 and @xmath128 eq.([eq : spin - current - explicit ] ) reaches the universal limit @xmath129 ( independent of the disorder details ) .
this limit depends only on the ratio @xmath130 of the strengths of the sois , but not on their absolute values , and agrees with the clean limit found in@xcite for @xmath131 . indeed , for the condition @xmath132 ( @xmath133 )
, many cycles of the electric rf field pass through between subsequent scattering events such that the system effectively behaves as ballistic .
this regime can be exploited to achieve high spin polarizations as described above .
moreover , the singularity in eq.([eq : spin - current - rd ] ) for @xmath96 is removed in eq.([eq : spin - current - explicit ] ) up to the accuracy @xmath134 considered here and we find that @xmath135 vanishes in the configuration where @xmath136 is maximal , i.e. for @xmath96 and @xmath137 .
we turn now to a geometrical interpretation of the spin hall current relating it to the trajectories @xmath138 followed by the tip of the polarization vector . for an applied electric field
@xmath139/2 $ ] with frequency @xmath140 this trajectory is given by the polarization ( as a function of time ) @xmath141 with the matrix @xmath142 containing the fourier components @xmath143 of the susceptibility evaluated at @xmath144 .
eq .
( [ eq : spin - polarization ] ) constitutes a quadratic form for the trajectory given by @xmath145 with real , positive eigenvalues @xmath146 ( say @xmath147 ) of the defining matrix @xmath148 .
thus , @xmath149 is of elliptic shape with semi - major and semi - minor axis @xmath150 and @xmath151 , resp .
we can further determine the angle @xmath29 enclosed by the semi - major axis of @xmath149 and the @xmath152 direction since the matrix @xmath153 is diagonalized by a rotation @xmath29 around @xmath154 .
the polarization of eq .
( [ eq : spin - polarization ] ) can thus be written as @xmath155{cc } \cos \delta & -\sin \delta \\
\sin \delta & \cos \delta \\ \end{array } \right ) \left(\begin{array}[c]{c } a \cos(\omega_0 t + \varphi ) \\
b \sin(\omega_0
t + \varphi ) \end{array } \right)\,.\end{aligned}\ ] ] here , @xmath156 is a phase shift between the electric field and the polarization . from eqs .
( [ eq : spin - polarization ] ) and ( [ eq : ellipse ] ) , we can relate the real and imaginary part of the susceptibilities @xmath157 and @xmath158 to the parameters @xmath159 , @xmath156 , and @xmath29 .
in particular , we obtain the spin hall current ( eq.([eq : spin - current - rd ] ) ) at resonance ( @xmath160 ) as @xmath161 eq . ( [ eq : spin - current - ab - res ] ) provides a remarkable interpretation of the spin hall current in terms of the geometric properties of the orbit @xmath149 .
the component @xmath111 is given by a complex phase depending on the rotation angle @xmath29 and the difference between the semi - minor and semi - major axis @xmath162 . in the linear response regime , the spin hall current characterizes the deviation from a circular orbit with @xmath163 to an elliptic shape ( with @xmath164 ) .
therefore , @xmath111 becomes accessible in terms of simple geometric properties of @xmath165 in experiments capable of resolving individual polarization components . in conclusion , we predict a substantially enhanced spin polarization due to interference effects of rashba and dresselhaus soi .
the spin hall current associated with this polarization can be interpreted in terms of the trajectory in spin space and vanishes if the polarization is maximal .
we thank o. chalaev , d. bulaev , j. lehmann , and h .- a .
engel for helpful discussions .
we acknowledge financial support from the swiss nf , nccr nanoscience basel , and the onr . |
field theories defined on a nc geometry are highly fashionable , in particular because they arise from a low energy limit of string theory @xcite
. a nc space may be defined by the non commutativity of some of its coordinates @xmath2 = i\theta_{\mu\nu } \,\overset{\text{2d}}{=}\ , i \theta \epsilon_{\mu \nu}\,.\ ] ] for a review of nc field theories , see ref .
the extension of actions of commutative field theories to their nc counterparts can be realized by replacing all products between fields by the _ star product _ @xmath3 in the nc @xmath0 model this replacement leads to the action @xmath4\,,\nonumber\ ] ] where only the interaction term requires the star product . in perturbation
theory the one loop contribution to the 1 pi two point function splits into two parts coming from the planar and the non planar graphs .
the planar terms are proportional to their ( uv divergent ) commutative counterparts @xcite . in the case of the non planar graphs the momentum cut off
@xmath5 is replaced by an effective cut off @xmath6 @xcite , @xmath7 where @xmath8 denotes the incoming momentum .
the cut off @xmath5 may be safely send to infinity , leading to a uv finite contribution .
however , the uv divergences reappear as ir divergences in the limit @xmath9 .
this mixing of uv and ir effects still causes serious problems in a perturbative treatment of nc field theories beyond one loop .
we studied the mixing of divergences non perturbatively in the 3d model , with a commutative time direction .
, see refs .
@xcite . ] to avoid the ( cpu ) time consuming lattice version of the star product , we mapped the system on a dimensionally reduced model @xcite . here the scalar fields @xmath10 defined on a @xmath11 lattice are mapped on @xmath12 hermitian matrices @xmath13 .
their action reads @xmath14&\hspace{-.5cm}=\vspace{-.1cm}n\textrm{tr}\sum_{t=1}^{t } \biggl [ \frac{1}{2}\sum_\mu\left(\gamma_\mu\hat{\phi}(t)\ , \gamma_\mu^\dagger\!-\hat{\phi}(t)\right)^2\\ & \hspace{-.5cm}+\frac{1}{2}\left(\hat{\phi}(t\!+\!1)-\hat{\phi}(t)\right)^2\!\!+ \!\frac{m^2}{2}\hat{\phi}^2(t)+\frac{\lambda}{4}\hat{\phi}^4(t ) \biggl]\nonumber\,,\end{aligned}\ ] ] where the twist
eaters @xmath15 are defined by @xmath16 this implies @xmath17 , where @xmath18 is the lattice spacing . for this action we studied the phase diagram and the dispersion relation .
based on the momentum dependent order parameter @xmath19 , with is the spatial fourier transform of @xmath10 . ]
@xmath20 we studied the phase diagram in the @xmath21@xmath22 plane .
our results for various values of @xmath23 are shown in fig .
[ fig : phase ] . model .
the connected symbols show the separation line between disordered and ordered regime , and the vertical lines mark the transition region between uniformly ordered and striped phase . _ _ ] we identify a clear separation line ( connected symbols ) between the disordered phase and the ordered regime .
the ordered regime splits into a uniformly ordered phase and a striped phase , where the transition region is marked by two vertical lines for each value of @xmath24 . to illustrate the striped phase we present in fig .
[ fig : snapshot ] snapshots of single configurations , which represent the ground state in this phase in the @xmath25@xmath26 plane at fixed time @xmath27 at a certain time @xmath27 , for @xmath28.__,title="fig : " ] at a certain time @xmath27 , for @xmath28.__,title="fig : " ] for @xmath29 at @xmath30 and @xmath31 .
the dotted areas indicate @xmath32 and in the blank areas @xmath33 is negative . here
we show configurations with two diagonal stripes resp .
four stripes parallel to the @xmath25 axis . at smaller values of the coupling @xmath21 or smaller system size
@xmath24 we also find two stripes parallel to one of the axis @xcite .
these results agree qualitatively with the conjecture by gubser and sondhi , who predicted the occurrence of a striped phase @xcite . to complete the agreement the striped phase has to survive the continuum limit , where the number of stripes should diverge , such that the width of the stripes remains finite .
the star product breaks explicitly the lorentz symmetry , which leads to a deformation of the standard dispersion relation .
the one loop result for this relation reads @xcite @xmath34 where @xmath6 is defined in eq .
( [ eq : eff - cutoff ] ) .
the deformation causes a shift in the energy minimum from @xmath35 to non
zero momenta .
we investigated numerically the energy
momentum relation in the disordered phase .
the energy @xmath36 can be computed from the correlator @xmath37 where the physical momenta are given by @xmath38 .
this correlator behaves like a cosh @xmath39 and the study of its decay allows to extract the energy . in fig .
[ fig : disp1 ] ( on top ) the system is close to the uniformly ordered phase transition . here the square of the energy is linear in @xmath40 as in a lorentz invariant theory .
close to the striped phase ( fig .
[ fig : disp1 ] below ) the situation is changed .
we see a clear deviation from lorentz symmetry .
the minimum of the energy is now at the lowest non zero ( lattice ) momentum and thus there will be two stripes parallel to the axes in the non uniform phase . in fig .
[ fig : disp2 ] the results at very large coupling @xmath21 , far outside the phase diagram in fig .
[ fig : phase ] , are shown .
now the energy minimum is shifted to larger momenta , leading to the more complicated patterns in the striped phase as in fig .
[ fig : snapshot ] . in figs .
[ fig : disp1 ] and [ fig : disp2 ] the solid lines are fits to the one loop result for the energy in eq .
( [ eq : deformation ] ) . .
close to the uniform phase at @xmath41 ( on top ) and close to the striped phase at @xmath42 ( below).__,title="fig : " ] + .
close to the uniform phase at @xmath41 ( on top ) and close to the striped phase at @xmath42 ( below).__,title="fig : " ]
we studied numerically the effects of uv / ir mixing in the 3d nc @xmath0 model . for the phase diagram we found that the ordered regime is split into an ising type phase for small coupling @xmath21 and a striped phase for larger coupling .
the patterns in the striped phase become more complex when @xmath21 or the system size @xmath24 is increased .
these results are in qualitative agreement with the conjecture of gubser and sondhi , if this type of stripes survives the large @xmath24 limit . the energy momentum relation behaves as predicted from one loop perturbation theory .
this is a remarkable result , since due to the uv / ir mixing there could be strong effects from higher order calculations .
our results imply that such effects do not change the results qualitatively
. however , for final conclusions one has to perform the continuum limit @xcite . | the uv / ir mixing in the @xmath0 model on a non - commutative ( nc ) space leads to new predictions in perturbation theory , including hartree
fock type approximations . among them
there is a changed phase diagram and an unusual behavior of the correlation functions . in particular
this mixing leads to a deformation of the dispersion relation .
we present numerical results for these effects in @xmath1 with two nc coordinates . |
many practical optimization problems must be addressed within the framework of large - scale structured convex optimization and need to be solved in a parallel and distributed manner .
such problems may appear in many fields of science and engineering : e.g. graph theory , networks , transportation , distributed model predictive control , distributed estimation and multistage stochastic optimization , see e.g. @xcite and the references quoted therein .
solving large - scale optimization problems is still a challenge in many applications @xcite due to the limitations of computational devices and computer systems .
recently , thanks to the development of parallel and distributed computer systems , many large - scale problems have been solved by using the framework of decomposition .
however , methods and algorithms for solving this type of problems , which can be performed in a parallel or distributed manner , are still limited @xcite . in this paper
we develop a new optimization algorithm to solve the following structured convex optimization problem with a separable objective function and coupling linear constraints : @xmath1 where , for every @xmath2 , @xmath3 is convex ( not necessarily strictly convex ) and possibly _ nonsmooth _ functions , @xmath4 is nonempty , closed and convex sets , @xmath5 and @xmath6 , and @xmath7 . here ,
@xmath8 is referred to as a local convex constraint and the final constraint is called _ coupling linear constraint_. in the literature , several approaches based on decomposition techniques have been proposed to solve problem . in order to observe the differences between those methods and our approach in this paper , we briefly classify some of these that we found most related . the first class of algorithms is based on lagrangian relaxation and subgradient methods of multipliers @xcite .
it has been observed that subgradient methods are usually slow and numerically sensitive to the choice of step sizes in practice @xcite .
moreover , the convergence rate of these methods is in general @xmath9 , where @xmath10 is the iteration counter .
the second approach relies on augmented lagrangian functions , see e.g. @xcite .
many variants were proposed and tried to process the inseparability of the crossproduct terms in the augmented lagrangian function in different ways . besides this approach , the authors in @xcite considered the dual decomposition based on fenchel s duality theory . another research direction is based on alternating direction methods which were studied , for example , in @xcite .
alternatively , proximal point - type methods were extended to the decomposition framework , see , e.g. @xcite .
other researchers employed interior point methods in the framework of decomposition such as @xcite . furthermore , the mean value cross decomposition in @xcite , the partial inverse method in @xcite and the accelerated gradient method of multipliers in @xcite were also proposed to solve problem .
we note that decomposition and splitting methods are very well developed in convex optimization , especially in generalized equations and variational inequalities , see e.g. @xcite .
recently , we have proposed a new decomposition method to solve problem in @xcite based on two primal steps and one dual step .
it is proved that the convergence rate of the algorithm is @xmath11 which is much better than the subgradient - type methods of multipliers @xcite but its computational complexity per iteration is higher that of these classical methods . moreover
, the algorithm uses an automatic strategy to update the parameters which improves the numerical efficiency in practice . in this paper
, we propose a new inexact decomposition algorithm for solving which employs smoothing techniques @xcite and excessive gap condition @xcite .
0.1 cm * contribution .
* the contribution of the paper is as follows : 1 . we propose a new decomposition algorithm based on inexact dual gradients .
this algorithm requires only one primal step and two dual steps at each iteration and allows one to solve the subproblem of each component inexactly and in parallel .
moreover , all the algorithmic parameters are updated automatically without using any tuning strategy .
2 . we prove the convergence of the proposed algorithm and show that the convergence rate is @xmath12 , where @xmath10 is the iteration counter .
due to the automatic update of the algorithmic parameters and the low computational complexity per iteration , the proposed algorithm performs better than some related existing decomposition algorithms from the literature in terms of computational time . 3 .
an extension to a switching strategy is also presented .
this algorithm updates simultaneously two smoothness parameters at each iteration and makes use of the inexactness of the gradients of the smoothed dual function .
let us emphasize the following points of the contribution .
the first algorithm proposed in this paper consists of two dual steps and one primal step per iteration .
this requires solving the primal subproblems in parallel only _ once _ but needs one more dual step . because the dual step corresponds only to a simple matrix - vector multiplication , the computational cost of the proposed algorithm
is significantly reduced compared to some existing decomposition methods in the literature . moreover , since solving the primal subproblems exactly is only _ conceptual _ ( except existing a closed form solution ) , we propose an inexact algorithm which allows one to solve these problems up to a given accuracy .
the accuracies of solving the primal subproblems are adaptively chosen such that the convergence of the whole algorithm is preserved .
the parameters in the algorithm are updated automatically based on an analysis of the iteration scheme .
this is different from augmented lagrangian approaches @xcite where we need to find an appropriate way to tune the penalty parameter in each practical situation . in the switching variant ,
apart from the inexactness , this algorithm allows one to update simultaneously both smoothness parameters at each iteration .
the advantage of this algorithm compared to the first one is that it takes into account the convergence behavior of the primal and dual steps which accelerates the convergence of the algorithm in some practical situations . since both algorithms are primal - dual methods , we not only obtain an approximate solution of the dual problem but also an approximate solution of the original problem without any auxiliary computation
. 0.1 cm * paper outline . *
the rest of this paper is organized as follows . in the next section ,
we briefly describe the lagrangian dual decomposition method for separable convex optimization .
section [ sec : smoothing ] mainly presents the smoothing technique via prox - functions as well as the inexact excessive gap condition .
section [ sec : decomp_alg ] builds a new inexact algorithm called _ inexact decomposition algorithm with two dual steps _ ( algorithm [ alg : a1 ] ) .
the convergence rate of this algorithm is established .
section [ sec : inexact_switching_alg ] presents an inexact switching variant of algorithm [ alg : a1 ] proposed in section [ sec : decomp_alg ] .
numerical examples are presented in section [ sec : num_results ] to examine the performance of the proposed algorithms and a numerical comparison is made . in order to make the paper more compact , we move some the technical proofs to appendix [ sec : proofs ]
. 0.1 cm * notation .
* throughout the paper , we shall consider the euclidean space @xmath13 endowed with an inner product @xmath14 for @xmath15 and the norm @xmath16 .
for a given matrix @xmath17 , the spectral norm @xmath18 is used in the paper .
the notation @xmath19 represents a column vector in @xmath13 , where @xmath20 is a subvector in @xmath21 and @xmath22 .
we denote by @xmath23 and @xmath24 the sets of nonnegative and positive real numbers , respectively .
we also use @xmath25 for the subdifferential of a convex function @xmath26 . for a given convex set @xmath27 in @xmath13
, we denote @xmath28 the relative interior of @xmath27 , see , e.g. @xcite .
in this section , we briefly describe the lagrangian dual decomposition technique in convex optimization , see , e.g. @xcite .
let @xmath29 be a vector and @xmath30 $ ] be a matrix formed from @xmath31 components @xmath20 and @xmath32 , respectively .
let @xmath33 and @xmath34 . the lagrange function associated with the coupling constraint @xmath35 is defined by @xmath36,\ ] ] where @xmath37 is the lagrange multiplier associated with @xmath35 .
the dual problem of is written as @xmath38 where @xmath39 is the dual function defined by @xmath40 note that the dual function @xmath41 can be computed in _
parallel _ for each component @xmath20 as @xmath42 we denote @xmath43 a solution of the minimization problem in .
consequently , @xmath44 is a solution of .
it is well - known that @xmath41 is concave and the dual problem is convex but nondifferentiable in general . throughout the paper , we assume that the following assumptions hold @xcite .
[ as : a1 ] the solution set @xmath45 of is nonempty and either @xmath27 is polyhedral or the slater constraint qualification condition for problem holds , i.e. @xmath46 for each @xmath47 , @xmath48 is proper , lower semicontinuous and convex in @xmath21 . if @xmath27 is convex and bounded then @xmath45 is also convex and bounded .
note that the objective function @xmath49 is not necessarily smooth .
for example , @xmath50 , which is nonsmooth and separable , can be handled in our framework . under assumption ,
the solution set @xmath51 of the dual problem is nonempty and bounded .
moreover , _ strong duality condition _ holds , i.e. for all @xmath52 we have @xmath53 .
if strong duality holds then we can refer to @xmath54 or @xmath55 as the _ primal - dual optimal value_.
since the dual function @xmath41 is in general nonsmooth , one can apply smoothing techniques to approximate @xmath41 up to a desired accuracy . in this section
, we propose to use a smoothing technique via proximity functions proposed in @xcite .
0.1 cm * 3.1 .
proximity functions .
* let @xmath56 be a nonempty , closed and convex set in @xmath13 .
we consider a nonnegative , continuous and strongly convex function @xmath57 with a convexity parameter @xmath58 . as usual , we call @xmath59 a _ proximity function _ ( prox - function ) associated with the convex set @xmath56
. let @xmath60 since @xmath59 is strongly convex , there exists a unique point @xmath61 such that @xmath62 . the point @xmath63 is called the _ proximity center _ of @xmath56 w.r.t .
@xmath59 . moreover ,
if @xmath56 is bounded then @xmath64 .
without loss of generality , we can assume that @xmath65 . otherwise , we can shift this function as @xmath66 , where @xmath67 .
[ re : quad_prox ] we note that the simplest prox - function is the quadratic form @xmath68 , where @xmath69 , @xmath58 and @xmath61 are given .
if the set @xmath56 has a specific structure then one can choose an appropriate prox - function that captures better the structure of @xmath56 than the quadratic prox - function .
for example , if @xmath56 is a standard simplex , one can choose the entropy prox - function as mentioned in @xcite .
if @xmath56 has no specific structure , then we can use the quadratic prox - function given above .
consequently , the convex problem generated using quadratic prox - functions reduces in some cases to a simple optimization problem , so that its solution can be computed numerically very efficient . 0.1 cm * 3.2
. smoothed approximations . * in order to build smoothed approximations of the objective function @xmath49 and the dual function @xmath41 in the framework of the primal - dual smoothing technique proposed in @xcite , we make the following assumption . [ as : a2 ] each feasible set @xmath70 admits a prox - function @xmath71 with a convexity parameter @xmath72 and the proximity center @xmath73 .
further , we assume @xmath74 if @xmath70 is bounded for @xmath75 , then assumption [ as : a2 ] is satisfied .
if @xmath70 is unbounded , then we can assume that our sample points generated by the proposed algorithms are bounded .
in this case , we can restrict the feasible set of problem on @xmath76 , where @xmath56 is a given compact set which contains the sample points and the desired solutions of .
we denote by @xmath77 since @xmath48 is not necessarily strictly convex , the function @xmath78 defined by may not be differentiable .
we consider the following function @xmath79 for @xmath22 and @xmath80 is a given _
smoothness parameter_. we denote @xmath81 the unique solution of , i.e. @xmath82 and @xmath83 .
we call each minimization problem in a _ primal subproblem_. note that we can use different smoothness parameters @xmath84 in for each @xmath85 .
first , we recall the following properties of @xmath86 , see @xcite . [
le : dual_smoothed_estimates ] for any @xmath80 , the function @xmath86 defined by is concave and differentiable .
the gradient of @xmath87 is given by @xmath88 which is lipschitz continuous with a lipschitz constant @xmath89 moreover , we have the following estimates : @xmath90 and @xmath91 next , we consider the variation of the function @xmath92 w.r.t . the parameter @xmath93 .
[ le : g_beta_1 ] let us fix @xmath94 .
the function @xmath95 defined by is well - defined , nondecreasing , concave and differentiable in @xmath24 .
moreover , the following inequality holds : @xmath96 where @xmath97 is defined by .
since @xmath98 and @xmath99 , it is sufficient to prove the inequality for @xmath100 , with @xmath22 . let us fix @xmath94 and @xmath85 .
we define @xmath101 a function of two joint variables @xmath20 and @xmath93 . since @xmath102 is strongly convex w.r.t . @xmath20 and linear w.r.t .
@xmath93 , @xmath103 is well - defined and concave w.r.t .
moreover , it is differentiable w.r.t . @xmath93 and @xmath104 , where @xmath105 is defined in .
hence , @xmath100 is nonincreasing . by using the concavity of @xmath100 we have @xmath106 by summing up the last inequality from @xmath107 to @xmath31 and then using we obtain .
@xmath108 finally , we consider a smooth approximation to @xmath49 .
let @xmath109 be a prox - function defined in @xmath110 with a convexity parameter @xmath111 .
it is obvious that the proximity center of @xmath112 is @xmath113 .
we define the following function on @xmath27 : @xmath114 where @xmath115 is the second smoothness parameter .
we denote by @xmath116 the solution of . from
, we see that @xmath117 and @xmath116 can be computed explicitly as @xmath118 it clear that @xmath119 for all @xmath120 .
now , we define the function @xmath121 as @xmath122 then , @xmath121 is exactly a quadratic penalty function of .
the following lemma shows that @xmath123 is an approximation of @xmath49 .
[ le : primal_smoothed_estimates ] the function @xmath124 defined by satisfies the following estimate : @xmath125 where @xmath126 .
moreover , the function @xmath26 defined by satisfies @xmath127 by the definition of @xmath124 , we have @xmath128 .
thus follows from this equality by applying some elementary inequalities .
the bounds follow directly from the definition of @xmath26 . @xmath108 0.1 cm * 3.3 .
inexact solutions of the primal subproblem . * regarding the primal subproblem ,
if the objective function @xmath48 has a specific form , e.g. univariate functions , then we can solve this problem analytically ( exactly ) to obtain a _ closed form _ solution .
a simple example of such function is @xmath129 .
however , in most practical problems , solving the primal subproblem exactly is only conceptual . in practice
, we only solve this problem up to a given accuracy .
in other words , for each @xmath130 , the solution @xmath105 in is approximated by @xmath131 in the sense of the following definition .
[ de : inexactness ] we say that the point @xmath132 approximates @xmath105 defined by up to a given accuracy @xmath133 if : * it is feasible to @xmath70 , i.e. @xmath134 ; * @xmath132 satisfies the condition : @xmath135 where @xmath136 . in practice , for a given accuracy @xmath137
, we can check whether the conditions of definition [ de : inexactness ] are satisfied by applying classical convex optimization algorithms , e.g. ( sub)gradient or interior - point algorithms @xcite .
since @xmath138 is strongly convex with a convexity parameter @xmath139 , we have @xmath140 where @xmath141 is defined as in definition [ de : inexactness ] .
consequently , we have : @xmath142 for @xmath22 .
let @xmath143 and @xmath144 the quantity @xmath145 can be referred to as an approximation of the gradient @xmath146 defined in lemma [ le : dual_smoothed_estimates ] .
if we denote by @xmath147 the vector of accuracy levels then we can easily estimate @xmath148 0.1 cm * 3.4 .
inexact excessive gap condition . * since problem is convex , under assumption strong duality holds .
the aim is to generate a primal - dual sequence @xmath149 such that for a sufficiently large @xmath10 the point @xmath150 is approximately feasible to , i.e. @xmath151 , and the primal - dual gap satisfies @xmath152 for given tolerances @xmath153 and @xmath154 .
the algorithm designed below will employ the approximate functions - to solve the primal - dual problems - .
first , we modify the excessive gap condition introduced by nesterov in @xcite to the inexact case in the following definition .
[ de : exessive_gap_cond ] a point @xmath155 satisfies the _ inexact excessive gap _ ( @xmath156-_excessive gap _ ) condition w.r.t . @xmath157 and @xmath115 and a given accuracy @xmath158 if @xmath159 if @xmath160 then reduces to the exact excessive gap condition considered in @xcite .
the following lemma provides an upper bound estimate for the primal - dual gap and the feasibility gap of problem . from the estimates
and we have @xmath164 then , by using , the last inequality implies the right - hand side of .
next , for a given @xmath162 we have @xmath165 .
thus we obtain the left - hand side of .
finally , the estimate follows from after a few simple calculations .
let us define @xmath166 the diameter of @xmath51 .
since @xmath51 is bounded , we have @xmath167 . the estimates and
can be simplified as @xmath168
in this section we first show that , for a given @xmath169 , there exists a point @xmath170 such that the condition is satisfied .
then , we propose a decomposition scheme to update successively a sequence @xmath171 that maintains the condition while it drives the sequences of smoothness parameters @xmath172 and @xmath173 to zero .
let us introduce the following quantities @xmath174 } & : = \left[\sum_{i=1}^m\sigma_i\varepsilon_i^2\right]^{1/2},\\ d_{\sigma } & : = \left[2\sum_{i=1}^m\frac{d_{x_i}}{\sigma_i}\right]^{1/2},\\ c_d & : = { \|a\|}^2d_{\sigma } + { \|a^t(ax^c - b)\|},\\ l_{\mathrm{a } } & : = m\max{\left\{\frac{{\|a_i\|}^2}{\sigma_i}~|~1\leq i\leq m\right\}}. \end{cases}\ ] ] from we see that the constant @xmath175 depends on the data of the problem ( i.e. @xmath176 , @xmath177 , @xmath178 , @xmath179 and @xmath63 ) . moreover , @xmath180 } = { \|\mathbf{\varepsilon}\|}$ ] .
if we choose the accuracy @xmath181 for all @xmath22 , then @xmath180 } = \sqrt{m}\hat{\varepsilon}$ ] and @xmath182 } = [ \sum_{i=1}^m\sigma_i]^{1/2}\hat{\varepsilon}$ ] . 0.1 cm * 4.1 .
finding a starting point .
* for a given a positive value @xmath80 , let @xmath183 be a point in @xmath184 computed as @xmath185 where @xmath186 is the origin and @xmath187 is defined by and @xmath188 is given by .
the following lemma shows that @xmath189 satisfies the @xmath190-excessive gap condition .
the proof of this lemma is given later in appendix [ sec : proofs ] .
[ le : init_point ] the point @xmath170 generated by satisfies the @xmath190-excessive gap condition w.r.t . @xmath93 and
@xmath191 provided that @xmath192 where @xmath193 } + \frac{1}{2}\varepsilon^2_{[\sigma]}\right ) \geq 0 $ ] .
note that if we use @xmath194 instead of @xmath187 into , i.e. the exact solution @xmath195 is used , then @xmath183 satisfies the @xmath196-excessive gap condition . 0.1 cm * 4.2 . the inexact main iteration with one primal step and two dual steps .
* let us assume that @xmath197 is a given point in @xmath184 that satisfies the @xmath156-excessive gap condition w.r.t .
@xmath93 , @xmath191 and @xmath156 .
the aim is to compute a new point @xmath198 such that the condition holds for the new values @xmath199 , @xmath200 and @xmath201 with @xmath202 , @xmath203 and @xmath204 . first , for a given @xmath37 and @xmath80 , we define the following mapping @xmath205 where @xmath206 is defined by and @xmath188 is the lipschitz constant . since this maximization problem is unconstrained and convex , we can show that the quantity @xmath207 can be computed explicitly as @xmath208.\ ] ] next , the main scheme to update @xmath198 is presented as @xmath209 here , the smoothness parameters @xmath93 and @xmath191 and the step size @xmath210 will be appropriately updated to obtain @xmath199 , @xmath200 and @xmath211 , respectively .
note that line 1 and line 3 in are simply matrix - vector multiplications , which can be computed distributively based on the structure of the coupling constraints and can be expressed as @xmath212 only line 2 in requires one to solve @xmath31 convex primal subproblems up to a given accuracy .
however , this can be done in _
parallel_. let us define @xmath213 then , by assumption , we can see that @xmath214 .
we consider an update rule for @xmath93 and @xmath191 as @xmath215 in order to show that @xmath198 satisfies the @xmath201-excessive gap condition , where @xmath201 will be defined later , we define the following function @xmath216}^2 + \left[\frac{\beta_1}{l_{\mathrm{a}}}c_d + ( 1-\tau)\tau\left ( \frac{c_d}{\beta_2 } + { \|a\|}{\|\bar{y}\|}\right)\right]\varepsilon_{[1]},\end{aligned}\ ] ] where @xmath182 } , c_d$ ] and @xmath217 are defined in . the next theorem provides a condition such that @xmath198 generated by satisfies the @xmath201-excessive gap condition . for clarity of the exposition we move the proof of this theorem to appendix [ sec : proofs ] .
[ th : alg_scheme ] suppose that assumptions and are satisfied
. let @xmath155 be a point satisfying the @xmath156-excessive gap condition w.r.t .
two values @xmath93 and @xmath191 .
then if the parameter @xmath218 is chosen such that @xmath219 and @xmath220 then the new point @xmath198 generated by is in @xmath184 and maintains the @xmath201-excessive gap condition w.r.t two new values @xmath199 and @xmath221 defined by , where @xmath222 with @xmath223 defined by . 0.1 cm * 4.3 . the step size update rule . *
next , we show how to update the step size @xmath210 .
indeed , from we have @xmath224 . by combining this inequality and we have @xmath225 . in order to ensure @xmath226 we require @xmath227 .
since @xmath228 and @xmath229 $ ] , we have @xmath230^{1/2 } - ( 1-\tilde{\alpha}\tau)\tau\right\ } < \tau.\ ] ] hence , if we choose @xmath231^{1/2 } - ( 1-\tilde{\alpha}\tau)\tau\right]$ ] then we obtain the tightest rule for updating @xmath218 . based on the above analysis
, we eventually define a sequence @xmath232 as follows : @xmath233^{1/2 } - ( 1-\tilde{\alpha}_k\tau_k)\tau_k\right\ } , ~ \forall k\geq 0,\ ] ] where @xmath234 is given and @xmath235 $ ] .
the following lemma provides the convergence rate of the sequence @xmath236 , whose proof can be found in appendix [ sec : proofs ] .
[ re : choice_of_tau ] the estimates of lemma [ le : choice_of_tau ] show that the sequence @xmath242 converges to zero with the convergence rate @xmath243 .
consequently , by , we see that the sequence @xmath244 also converges to zero with the convergence rate @xmath245 . from and
, we can derive an initial value @xmath246 . in order to choose the accuracy for solving the primal subproblem
, we need to analyze the formula .
let us consider a sequence @xmath247 computed by @xmath248 where @xmath249 is given in .
the sequence @xmath250 defined by @xmath251 where @xmath190 is chosen _ a priori _ , is nonincreasing if @xmath252 for all @xmath253 .
[ le : accuracy_level_choice ] if the accuracy @xmath254 at the iteration @xmath10 of algorithm [ alg : a1 ] below is chosen such that @xmath255 for @xmath22 , where @xmath256,\end{aligned}\ ] ] then the sequence @xmath250 generated by is nonincreasing . since @xmath257 for all @xmath22
, we have @xmath180}^k \leq \sqrt{m}\bar{\varepsilon}^k$ ] and @xmath258}^k)^2 \leq \left(\sum_{i=1}^m\sigma_i\right)(\bar{\varepsilon}^k)^2 \leq \left(\sum_{i=1}^m\sigma_i\right)\bar{\varepsilon}^k$ ] . by substituting these inequalities into of @xmath249 and then using and the notation @xmath259
, we have @xmath260 on the other hand , from we have @xmath261 for all @xmath262 .
thus , @xmath250 is nonincreasing if @xmath263 for all @xmath253 .
if we choose @xmath264 such that @xmath265 , i.e. @xmath266 , then @xmath252 . @xmath108 from lemma [ le : accuracy_level_choice ] it follows that if we choose @xmath267 sufficiently small , then the sequence @xmath268 generated by @xmath269 maintains the @xmath270-excessive gap condition with @xmath271 for all @xmath10 . now , by using lemmas [ le : excessive_gap ] and [ le : init_point ] , if we choose @xmath272 in lemma [ le : accuracy_level_choice ] such that @xmath273 , where @xmath274 and @xmath275 is a given accuracy level , then the condition holds with @xmath276 .
0.1 cm * 4.4 .
the algorithm and its convergence .
* finally , we present the algorithm in detail and estimate its worst - case complexity . for simplicity of discussion
, we fix the accuracy at one level @xmath264 for all the primal subproblems .
however , we can alternatively choose different accuracy for each subproblem by slightly modifying the theory presented in this paper .
_ inexact decomposition algorithm with two dual steps_[alg : a1 ] * initialization : * perform the following steps : * _ step 1 _ : provide an accuracy level @xmath275 for solving and a value @xmath277 .
set @xmath278 , @xmath279 and @xmath280 .
* _ step 2 _ : compute @xmath281 by . set @xmath282 and @xmath283 .
* _ step 3 _ : compute @xmath284 and @xmath285 from as @xmath286 and @xmath287 up to the accuracy @xmath272 .
* iteration : * for @xmath288 , perform the following steps : * _ step 1 _ : if a given stopping criterion is satisfied then terminate . *
_ step 2 _ : compute @xmath289 by . set @xmath290 and
update @xmath291 . *
_ step 3 _ : solve the primal subproblems in _ in parallel _ up to the accuracy @xmath264 . * _ step 4 _ : compute @xmath292 by . * _ step 5 _ : compute @xmath293 , where @xmath294 . *
_ step 6 _ : update @xmath295 and @xmath296 . *
_ step 7 _ : update @xmath297 as @xmath298^{1/2 } - ( 1-\tilde{\alpha}_k\tau_k)\tau_k\right\}$ ] .
the stopping criterion of algorithm [ alg : a1 ] at step 1 will be discussed in section [ sec : num_results ] . the maximum number of iterations @xmath299 provides a safeguard to prevent the algorithm from running to infinity .
the following theorem provides the worst - case complexity estimate for algorithm [ alg : a1 ] under assumptions and .
[ th : convergence ] suppose that assumptions and are satisfied .
let @xmath300 be a sequence generated by algorithm [ alg : a1 ] after @xmath301 iterations .
if the accuracy level @xmath302 in algorithm [ alg : a1 ] is chosen such that @xmath303 for some positive constant @xmath304 , then the following primal - dual gap holds @xmath305^{\alpha^{*}}},\ ] ] and the feasibility gap satisfies @xmath306 where @xmath307 and @xmath308 is defined by .
consequently , the sequence @xmath149 generated by algorithm [ alg : a1 ] converges to a solution @xmath309 of the primal and dual problems - as @xmath310 and @xmath311 . from lemma
[ le : excessive_gap ] , we have @xmath312 and @xmath313 .
moreover , @xmath314 due to the choice of @xmath190 and the update rule of @xmath315 at step 2 of algorithm [ alg : a1 ] . by combining these inequalities and and then using the definition of @xmath316 and @xmath317 we obtain and .
the last conclusion is a direct consequence of and .
@xmath108 the conclusions of theorem [ th : alg_scheme ] show that the initial accuracy of solving the primal subproblems needs to be chosen as @xmath11 .
then , we have @xmath318 and @xmath319 .
thus , if we choose the ratio @xmath320 such that @xmath321 then we obtain an asymptotic convergence rate @xmath11 for algorithm [ alg : a1 ] .
we note that the accuracy of solving has to be updated at each iteration @xmath10 in algorithm [ alg : a1 ] .
the new value is computed by @xmath322 at step 2 , which is the same @xmath323 order .
now , we consider a particular case , where we can get an @xmath324 worst - case complexity ( @xmath325 is a desired accuracy ) .
[ co : worst_case_complexity ] suppose that the smoothness parameter @xmath326 in algorithm [ alg : a1 ] is fixed at @xmath327 for all @xmath253 .
suppose further that the accuracy level @xmath302 in algorithm [ alg : a1 ] is chosen as @xmath328 and that the sequence @xmath242 is updated by @xmath329 starting from @xmath278 .
then , after @xmath330 iterations , one has @xmath331 where @xmath332 and @xmath333 . if we assume that @xmath326 is fixed in algorithm [ alg : a2 ] then , by the new update rule of @xmath242 we have @xmath334 due to and with @xmath335 .
since @xmath336 , if we choose @xmath337 then @xmath338 .
furthermore , by lemma [ le : excessive_gap ] we have @xmath339 and @xmath340 . by combining these estimates
, we obtain the conclusion .
@xmath108 ( _ distributed implementation_)[re : dist_opt ] in algorithm [ alg : a1 ] , only the parameter @xmath341 requires centralized information . instead of using @xmath341 , we can use its lower bound @xmath320 to compute @xmath297 and @xmath326 . in this case
, we can modify algorithm [ alg : a1 ] to obtain a distributed implementation .
the modification is at steps 5 , 6 and 7 , where we can parallelize these steps by using the same formulas for the all subsystems to compute the parameters @xmath326 , @xmath342 and @xmath297 .
we note that the points @xmath343 and @xmath344 in the scheme can be computed in parallel , while @xmath345 and @xmath346 can be computed distributively based on the structure of the coupling constraints of problem .
since the ratio @xmath347 defined in may be small , algorithm [ alg : a1 ] only provides a suboptimal approximation @xmath348 to the optimal solution @xmath309 such that @xmath349 in the worst - case .
for example , if we choose the prox - function @xmath350 , where @xmath351 , then worst - case complexity of algorithm [ alg : a1 ] is lower than subgradient methods , see of theorem [ th : convergence ] .
algorithm [ alg : a1 ] leads to a poor performance . in this section ,
we propose to combine the scheme @xmath352 defined by in this paper and an inexact decomposition scheme with two primal steps and one dual step to ensure that the parameter @xmath93 always decreases to zero .
apart from the inexactness , this variant allows one to update simultaneously both smoothness parameters at each iteration . 0.1 cm * 5.1 .
the inexact main iteration with two primal steps .
* let us consider the approximate function @xmath353 defined by .
we recall that @xmath49 is only assumed to be convex and possibly nonsmooth , while @xmath354 is convex and lipschitz continuously differentiable .
we define @xmath355 and the mapping @xmath356 where @xmath357 is the lipschitz constant of @xmath358 defined in lemma [ le : primal_smoothed_estimates ] . since @xmath359 is strongly convex , @xmath360 is well - defined .
[ re : bregman_dist ] note that we can replace the quadratic term @xmath361 in by any bregman distance as done in @xcite
. however , the convergence analysis based on this type of prox - functions is more complicated than the one given in this paper .
suppose that we can only solve the minimization problem up to a given accuracy @xmath362 to obtain an approximate solution @xmath363 in the sense of definition .
more precisely , @xmath364 and @xmath365 we denote @xmath366 and @xmath367 . in particular ,
if @xmath48 is differentiable and its gradient is lipschitz continuous with a lipschitz constant @xmath368 for some @xmath369 then one can replace the approximate mapping @xmath370 by the following one : @xmath371^t\!\!\!\!(x_i \!-\ !
\hat{x}_i ) \!+\ ! \frac{\hat{l}_i(\beta_2)}{2}{\|x_i - \hat{x}_i\|}^2\!\!\right\},\end{aligned}\ ] ] where @xmath372 , in the sense of definition [ de : inexactness ] .
note that the minimization problem defined in @xmath373 is a quadratic program with convex constraints .
now , we can present the decomposition scheme with two primal steps in the case of inexactness as follows .
suppose that @xmath155 satisfies w.r.t .
@xmath93 , @xmath191 and @xmath156 .
we update @xmath374 as @xmath375 where the step size @xmath210 will be appropriately updated and 1 . the parameters @xmath93 and @xmath191 are updated by @xmath376 and @xmath377 ; 2 .
@xmath378 is computed by ; 3 .
@xmath379 is an approximation of @xmath380 defined in and .
the following theorem states that the new point @xmath198 updated by @xmath381 maintains the @xmath201-excessive gap condition .
the proof of this theorem is postponed to appendix [ sec : proofs ] .
[ th : alg_scheme2 ] suppose that assumptions and are satisfied .
let @xmath197 be a point in @xmath184 and satisfy the @xmath156-excessive gap condition w.r.t .
two values @xmath93 and @xmath191
. then if the parameter @xmath218 is chosen such that @xmath219 and @xmath382 then the new points @xmath198 updated by maintains the @xmath201-excessive gap condition w.r.t .
two new values @xmath199 and @xmath221 , where @xmath383 } + \frac{1}{2}\sum_{i=1}^ml_i^{\psi}(\beta_2^{+})\varepsilon_i^2 $ ] , and @xmath182}$ ] and @xmath384 are defined in . finally , we note that the step size @xmath218 is updated by @xmath385 for @xmath253 starting from @xmath386 in the scheme , see @xcite for more details .
0.1 cm * 5.2 .
the algorithm and its convergence . *
first , we provide an update rule for @xmath156 in definition [ de : exessive_gap_cond ] . with @xmath182}$ ] and @xmath384 defined in , let us consider the function @xmath387 } + \frac{1}{2}\sum_{i=1}^ml_i^{\psi}(\beta_2^{+})\varepsilon_i^2,\ ] ] and a sequence @xmath388 generated by @xmath389 , where @xmath190 is given and @xmath390 is chosen appropriately .
the aim is to choose @xmath267 such that @xmath391 and @xmath388 is nonincreasing . by letting @xmath392 then , if we choose @xmath393 such that @xmath394 then we have @xmath395 . by combining both schemes and , we obtain a new variant of algorithm [ alg : a1 ] with a switching strategy as described as follows .
_ inexact decomposition algorithm with switching primal - dual steps_[alg : a2 ] * initialization : * perform as in algorithm [ alg : a1 ] with @xmath386 .
* iteration : * for @xmath396 perform the following steps : * _ step 1 _ : if a given stopping criterion is satisfied then terminate .
* _ step 2 _ : if @xmath10 is _ even _ then perform the scheme with two primal steps : * * 2.1 . compute @xmath397 by . set @xmath398 and update @xmath399 . * * 2.2 .
update @xmath296 . * * 2.3 .
compute @xmath400 up to the accuracy @xmath264 . * * 2.4 .
update @xmath401 . * * 2.5 .
update the step - size parameter @xmath297 as @xmath402 . * _ step 3 _ : otherwise , ( i.e. @xmath10 is _ odd _ ) perform the scheme with two dual steps : * * 3.1 .
compute @xmath289 by .
set @xmath290 and update @xmath403 . *
compute @xmath292 up to the accuracy @xmath264 . *
compute @xmath404 , where @xmath405 . * * 3.4 .
update @xmath406 and @xmath296 . * * 3.5 .
update @xmath297 as @xmath407^{1/2 } - ( 1-\tilde{\alpha}_k\tau_k)\tau_k\right\}$ ] .
note that the first line and third line of the scheme @xmath381 can be parallelized .
they require one to solve @xmath31 convex subproblems of the form and , respectively _ in parallel_. if the function @xmath48 is differentiable and its gradient is lipschitz continuous for some @xmath85 , then we can use the approximate gradient mapping @xmath373 instead of @xmath370 and the corresponding minimization subproblem in the third line reduces to a quadratic program with convex constraints .
the stopping criterion at step 1 will be given in section [ sec : num_results ] .
similar to the proof of lemma [ le : choice_of_tau ] we can show that the sequence @xmath232 generated by step 2.5 or step 3.5 of algorithm [ alg : a2 ] satisfies estimates .
consequently , the estimate for @xmath342 in is still valid , while the parameter @xmath326 satisfies @xmath408 . finally , we summarize the convergence results of algorithm [ alg : a2 ] in the following theorem .
[ th : convergence2 ] suppose that assumptions and are satisfied .
let @xmath300 be a sequence generated by algorithm [ alg : a2 ] after @xmath301 iterations .
if the accuracy level @xmath302 in algorithm [ alg : a2 ] is chosen such that @xmath409 for some positive constant @xmath304 , then the following primal - dual gap holds @xmath410 and the feasibility gap satisfies @xmath411 where @xmath412 and @xmath308 is defined as in .
consequently , the sequence @xmath149 generated by algorithm [ alg : a2 ] converges to a solution @xmath413 of the primal and dual problems - as @xmath310 and @xmath311 .
the proof of this theorem is similar to theorem [ th : convergence ] and thus we omit the details here .
we can see from the right hand side of in theorem [ th : convergence2 ] that this term is better than the one in theorem [ th : convergence ] .
consequently , the worst case complexity of algorithm [ alg : a2 ] is better than the one of algorithm [ alg : a1 ] .
however , as a compensation , at each even iteration , the scheme @xmath381 is performed .
it requires an additional cost to compute @xmath414 at the third line of @xmath381 . as an exception ,
if the primal subproblem can be solved in a _ closed form _ then the cost - per - iteration of algorithm [ alg : a2 ] is almost the same as in algorithm [ alg : a1 ] .
[ re : primal_update_variant ] note that we can only use the inexact decomposition scheme with two primal steps @xmath381 in to build an inexact variant of ( * ? ? ?
* algorithm 1 ) .
moreover , since the role of the schemes @xmath381 and @xmath352 is symmetric , we can switch them in algorithm [ alg : a2 ] .
in this section we compare algorithms [ alg : a1 ] and [ alg : a2 ] derived in this paper with the two algorithms developed in ( * ? ? ?
* algorithms 1 and 2 ) which we named ` 2pdecompalg ` and ` pddecompalg ` , the proximal center - based decomposition algorithm in @xcite , an exact variant of the proximal based decomposition algorithm in @xcite and three parallel variants of the alternating direction method of multipliers ( with three different strategies to update the penalty parameter ) .
we note that these variants are the modifications of the algorithm in @xcite , and they can be applied to solve problem with more than two objective components ( i.e. @xmath415 ) .
we named these algorithms by ` pcbdm ` , ` epbdm ` , ` admm - v1 ` , ` admm - v2 ` and ` admm - v3 ` , respectively . for more simulations and comparisons
we refer to the extended technical report @xcite .
the algorithms have been implemented in c++ running on a @xmath416 cores intel xeon @xmath417ghz workstation with @xmath418 gb of ram . in order to solve the general convex programming subproblems
, we either used a commercial software called ` cplex ` or an open - source software package ` ipopt ` @xcite .
all the algorithms have been parallelized by using ` openmp ` . in the four numerical examples below , since the feasible set @xmath70 has no specific structure , we chose the quadratic prox - function @xmath419 in the four first algorithms , i.e. algorithms [ alg : a1 ] and [ alg : a2 ] , ` 2pdecompalg ` and ` pddecompalg ` , where @xmath420 and @xmath421 are given , for @xmath22 , as mentioned in remark [ re : quad_prox ] . with this choice
we can solve the primal subproblem in the first example by using ` cplex ` .
we terminated these algorithms if @xmath422 and either the approximate primal - dual gap satisfied @xmath423 or the value of the objective function did not significantly change in @xmath424 successive iterations , i.e. : @xmath425 here @xmath426 is the approximate value of @xmath427 evaluated at @xmath428 . in `
admm - v1 ` and ` admm - v2 ` we used the update formula in ( * ? ? ?
* formula ( 21 ) ) to update the penalty parameter @xmath429 starting from @xmath430 and @xmath431 , respectively . in `
admm - v3 ` this penalty parameter was fixed at @xmath432 for all iterations . in ` pcbdm ` , we chose the same prox - function as in our algorithms and the parameter @xmath93 in the subproblems was fixed at @xmath433 .
we terminated all the remaining algorithms if the both conditions and were satisfied .
the maximum number of iterations @xmath434 was set to @xmath435 in all algorithms .
we warm - started the ` cplex ` and ` ipopt ` solvers at the iteration @xmath10 at the point given by the previous iteration @xmath436 for @xmath437 .
the accuracy levels @xmath267 in ` cplex ` and ` ipopt ` and @xmath315 were updated as in algorithms [ alg : a1 ] and [ alg : a2 ] starting from @xmath438 and then set to @xmath439 . in other algorithms , this accuracy level was fixed at @xmath440 .
we concluded that `` the algorithm is failed '' if either the maximum number of iterations ` maxiter ` was reached or the primal subproblems could not be solved by ` ipopt ` or ` cplex ` due to numerical issues .
we benchmarked all algorithms with performance profiles @xcite .
recall that a performance profile is built based on a set @xmath441 of @xmath442 algorithms ( solvers ) and a collection @xmath443 of @xmath444 problems .
suppose that we build a profile based on computational time .
we denote by @xmath445 .
we compare the performance of algorithm @xmath446 on problem @xmath447 with the best performance of any algorithm on this problem ; that is we compute the performance ratio @xmath448 .
now , let @xmath449 for @xmath450 .
the function @xmath451 $ ] is the probability for solver @xmath446 that a performance ratio is within a factor @xmath452 of the best possible ratio .
we use the term `` performance profile '' for the distribution function @xmath453 of a performance metric . in the following numerical examples
, we plotted the performance profiles in @xmath454-scale , i.e. @xmath455 . 0.1 cm * 6.1 . basic pursuit problem . *
the basic pursuit problem is one of the fundamental problems in signal processing and compressive sensing .
mathematically , this problem can be formulated as follows : @xmath456 where @xmath17 and @xmath457 are given .
since @xmath458 , the primal subproblem formed from in the algorithms can be expressed as @xmath459 this problem can be solved in a closed form without any subiteration .
we implemented algorithms [ alg : a1 ] and [ alg : a2 ] to solve this problem in order to compare the effect of the parameter @xmath320 on the performance of the algorithm .
the data of this problem is generated as follows .
matrix @xmath176 is generated randomly such that it is orthogonal .
vector @xmath460 , where @xmath461 is a @xmath10-sparse random vector ( @xmath462 ) .
we tested algorithms [ alg : a1 ] and [ alg : a2 ] with @xmath424 problems and the results reported by these algorithms are presented in table [ tb : basic_lasso ] with @xmath463 and @xmath464 .
1054 ) + + ` algorithm 1 ` & 8254/1.3858 & 5090/0.8769 & 10144/2.2876 & 29773/7.8386 & 35615/10.9315 + ` algorithm 2 ` & 4836/0.9025 & 4744/0.8808 & 8060/1.9809 & 13220/3.7619 & 12348/3.7967 + + ` algorithm 1 ` & 7115/1.1851 & 6644/1.1632 & 8749/1.9632 & 14927/4.0424 & 16128/4.9169 + ` algorithm 2 ` & 15016/2.6689 & 14284/2.6121 & 17140/4.0286 & 34048/9.6910 & 36668/11.1801 + as we can see from this table that algorithm [ alg : a2 ] performs better than algorithm [ alg : a1 ] in terms of number of iterations as well as computational time for the case @xmath463 . in the case
@xmath464 , algorithm [ alg : a1 ] performs better than algorithm [ alg : a2 ] .
this example claims the theoretical results .
0.1 cm * 6.2 .
nonsmooth separable convex optimization .
* let us consider the following simple nonsmooth convex optimization problem : @xmath465 where @xmath466 are given @xmath467 .
let us assume that @xmath468 $ ] is a given interval in @xmath469 .
then , this problem can be formulated in the form of [ eq : primal_cvx_prob ] with @xmath470 . since the lagrange function @xmath471 $ ] is nonsmooth , where @xmath472 is a lagrange multiplier , we choose @xmath473 such that the primal subproblem can be written as @xmath474}\!\!\!\!\big\ { i{|x_i \!-\ ! x_i^a| } \!+\ !
y\big(x_i \!-\ !
\frac{b}{n}\big ) \!+\ ! \frac{\beta_1}{2}{|x_i \!-\
! x_i^c|}^2 \!+\ ! 0.75d_{x_i}\big\},\ ] ] where @xmath80 .
now , we assume that we can choose the interval @xmath475 $ ] sufficiently large such that the constraint @xmath476 $ ] is inactive .
then , the solution of problem can be computed explicitly as @xmath477 , where the _ soft - thresholding - type operator _ @xmath478 is defined as follows : @xmath479 .
\end{cases}\ ] ] in this example , we tested five algorithms : algorithm [ alg : a1 ] , algorithm [ alg : a2 ] , ( * ? ? ?
* algorithm 1 ) , ( * ? ? ?
* algorithm 2 ) and ` pcbdm ` for @xmath480 problems with the size varying from @xmath481 to @xmath482 .
note that if we reformulate as a linear programming problem ( lp ) by introducing slack variables , then the resulting lp problem has @xmath483 variables and @xmath484 inequality constraints .
the data of these tests were created as follows .
the value @xmath485 was set to @xmath486 , @xmath487 , where @xmath488 .
the maximum number of iterations ` maxiter ` was increased to @xmath489 instead of @xmath490 .
the performance of the five algorithms is reported in table [ tb : exam1_results ] .
.performance comparison of five algorithms for solving [ cols="^,^ , > , > , > , > , > , > , > , > , > , > " , ] here , ` iters ` is the number of iterations and ` time ` is the cpu time in seconds .
as we can see from table [ tb : exam1_results ] , algorithm [ alg : a1 ] is the best in terms of number of iterations and computational time .
algorithm [ alg : a2 ] works better than ` pddecompalg ` .
the first four algorithms have consistently outperformed ` pcbdm ` in terms of number of iterations as well as computational time in this example .
0.1 cm * 6.3 .
separable convex quadratic programming .
* let us consider a separable convex quadratic program of the form : @xmath491 here @xmath492 is a symmetric positive semidefinite matrix , @xmath493 , @xmath494 for @xmath495 and @xmath457 . in this example
, we compared the above algorithms by building their performance profiles in terms of number of iterations and the total computational time . 0.1 cm _ problem generation . _
the input data of the test was generated as follows .
matrix @xmath496 , where @xmath497 is an @xmath498 random matrix in @xmath499 $ ] with @xmath500 .
matrix @xmath32 was generated randomly in @xmath501 $ ] .
vector @xmath502 , where @xmath503 is a given feasible point in @xmath504 and vector @xmath505 .
the density of both matrices @xmath32 and @xmath497 is @xmath506 .
note that the problems generated as above are always feasible .
moreover , they are not strongly convex .
the tested collection consisted of @xmath507 problems with different sizes and the sizes were generated randomly as follows : * _ class 1 : _ @xmath508 problems with @xmath509 , @xmath510 , @xmath511 and @xmath512 .
* _ class 2 : _ @xmath508 problems with @xmath513 , @xmath514 , @xmath515 and @xmath516 . * _ class 3 : _ @xmath480 problems with @xmath517 , @xmath518 , @xmath519 and @xmath520 . 0.1 cm _ scenarios .
_ we considered two different scenarios : _ scenario i : _ in this scenario , we aimed at comparing algorithms [ alg : a1 ] and [ alg : a2 ] , ` 2pdecompalg ` , ` pddecompalg ` , ` admm - v1 ` and ` epbdm ` , where we generated the values of @xmath521 relatively small .
more precisely , we chose @xmath522 = [ -0.1 , 0.1]$ ] , @xmath501 = [ -1 , 1]$ ] and @xmath523 . _
scenario ii : _ the second scenario aimed at testing the affect of the matrix @xmath176 and the update rule of the penalty parameter to the performance of ` admm ` .
we chose @xmath522 = [ -1 , 1]$ ] , @xmath501 = [ -5 , 5]$ ] and @xmath524 .
0.1 cm _ results .
_ in the first scenario , the size of the problems satisfied @xmath525 , @xmath526 and @xmath527 .
the performance profiles of the six algorithms are plotted in figure [ fig : perf_qp_exam02 ] with respect to the number of iterations and computational time .
scale for scenario i by using ` ipopt ` : left - number of iterations , right - computational time.,width=464,height=143 ] -0.2 cm from these performance profiles , we can observe that the algorithm [ alg : a1 ] , algorithm [ alg : a2 ] , + ` 2pdecompalg ` and ` pddecompalg ` converged for all problems . `
admm - v1 ` was successful in solving @xmath528 ( @xmath529 ) problems while ` epbdm ` could only solve @xmath530 @xmath531 problems .
it shows that algorithm [ alg : a1 ] is the best one in terms of number of iterations . it could solve up to @xmath532 ( @xmath533 ) problems with the best performance . `
admm - v1 ` solved @xmath534 ( @xmath535 ) problems with the best performance , while this ratio was only @xmath536 ( @xmath537 ) and @xmath538 ( @xmath539 ) in ` pddecompalg ` and algorithm [ alg : a2 ] , respectively .
if we compare the computational time then algorithm [ alg : a1 ] is the best one .
it could solve up to @xmath540 ( @xmath541 ) problems with the best performance . `
admm - v1 ` solved @xmath542 ( @xmath543 ) problems with the best performance .
since the performance of algorithms [ alg : a1 ] and [ alg : a2 ] , ` 2pdecompalg ` , ` pddecompalg ` and ` admm ` are relatively comparable , we tested algorithms [ alg : a1 ] and [ alg : a2 ] , ` 2pdecompalg ` , ` pddecompalg ` , ` admm - v1 ` , ` admm - v2 ` and ` admm - v3 ` on a collection of @xmath507 problems in the second scenario .
the performance profiles of these algorithms are shown in figure [ fig : perf_qp_exam01 ] .
-0.2 cm scale for scenario ii by using ` cplex ` with simplex method : left - number of iterations , right - computational time.,width=464,height=143 ] -0.5 cm from these performance profiles we can observe the following : * the six first algorithms were successful in solving all problems , while ` admm - v3 ` could only solve @xmath544 ( @xmath545 ) problems .
* algorithm [ alg : a1 ] and ` admm - v1 ` is the best one in terms of number of iterations .
it both solved @xmath546 ( @xmath547 ) problems with the best performance .
this ratio is @xmath548 @xmath549 in ` admm - v2 ` .
* algorithm [ alg : a1 ] is the best one in terms of computational time
. it could solve @xmath550 @xmath551 the problems with the best performance , while this quantity is @xmath552 in ` admm - v2 ` .
0.1 cm * 6.4 .
nonlinear smooth separable convex programming .
* we consider the following nonlinear , smooth and separable convex programming problem : @xmath553 here , @xmath554 is a positive semidefinite and @xmath555 is given vector , @xmath22 . 0.1 cm _ problem generation . _ in this example , we generated a collection of @xmath507 test problems as follows .
matrix @xmath554 is diagonal and was generated randomly in @xmath522 $ ] .
matrix @xmath32 was generated randomly in @xmath501 $ ] with the density @xmath556 .
vectors @xmath557 and @xmath558 were generated randomly in @xmath559 $ ] and @xmath560 $ ] , respectively , such that @xmath561 and @xmath562 .
vector @xmath563 for a given @xmath564 in @xmath565 $ ] .
the size of the problems was generated randomly based on the following rules : * _ class 1 : _ @xmath480 problems with @xmath566 , @xmath567 , @xmath515 and @xmath568 .
* _ class 2 : _ @xmath480 problems with @xmath569 , @xmath570 , @xmath571 and @xmath512 . * _ class 3 : _ @xmath480 problems with @xmath572 , @xmath573 , @xmath574 and @xmath516 . * _ class 4 : _ @xmath480 problems with @xmath575 , @xmath518 , @xmath574 and @xmath520 .
* _ class 5 : _ @xmath480 problems with @xmath576 , @xmath518 , @xmath574 and @xmath577 .
0.1 cm _ scenarios .
_ we also considered two different scenarios as in the previous example : _ scenario i : _ similar to the previous example , with this scenario , we aimed at comparing algorithms [ alg : a1 ] and [ alg : a2 ] , ` 2pdecompalg ` , ` pddecompalg ` , ` admm - v1 ` , ` pcbdm ` and ` epbdm ` . in this scenario
, we chose : @xmath522 \equiv [ -0.01 , 0.01]$ ] , @xmath559 \equiv [ 0 , 100]$ ] , @xmath501 \equiv [ -1 , 1]$ ] and @xmath578 . _
scenario ii : _ in this scenario , we only tested two first variants of ` admm ` and compared them with the four first algorithms . here , we chose @xmath522 \equiv [ 0.0 , 0.0]$ ] ( i.e. without quadratic term ) , @xmath559 \equiv [ 0 , 100]$ ] , @xmath501 \equiv [ -1 , 1]$ ] and @xmath579 .
0.1 cm _ results . _ for _ scenario i _ , we see that the size of the problems is in @xmath580 , @xmath581 and @xmath582 .
the performance profiles of the algorithms are plotted in figure [ fig : perf_scp2 ] .
-0.2 cm scale by using ` ipopt ` : left - number of iterations , right - computational time.,width=464,height=143 ] -0.3 cm the results on this collection shows that algorithm [ alg : a1 ] is the best one in terms of number of iterations
. it could solve up to @xmath583 ( @xmath584 ) problems with the best performance , while ` admm - v1 ` solved @xmath534 ( @xmath585 ) problems with the best performance .
algorithm [ alg : a1 ] is also the best one in terms of computational time .
it could solve @xmath586 ( @xmath587 ) problems with the best performance . `
pcbdm ` was very slow compared to the rest in this scenario
. for _ scenario ii _ , the size of the problems was varying in @xmath588 , @xmath589 and @xmath590 .
the performance profiles of the tested algorithms are plotted in figure [ fig : perf_scp1 ] .
-0.0 cm scale for ` scenario i ` by using ` ipopt ` : left - number of iterations , right - computational time.,width=464,height=143 ] -0.3 cm we can see from these performance profiles that algorithm [ alg : a1 ] is the best one in terms of number of iterations .
it could solve up to @xmath591 ( @xmath592 ) problems with the best performance , while this number were @xmath593 ( @xmath594 ) and @xmath595 ( @xmath596 ) problems in ` 2pdecompalg ` and ` admm - v1 ` , respectively .
algorithm [ alg : a1 ] was also the best one in terms of computational time .
it solved all problems with the best performance . `
admm - v2 ` was slow compared to the rest in this scenario . from the above two numerical tests
, we can observe that algorithm [ alg : a1 ] performs well compared to the rest in terms of computational time due to its low cost per iteration .
admm encounters some difficulty regarding the choice of the penalty parameter as well as the effect of matrix @xmath176 .
theoretically , pcbdm has the same worst - case complexity bound as algorithms [ alg : a1 ] and .
however , its performance is quite poor .
this happens due to the choice of the lipschitz constant @xmath217 of the gradient of the dual function and the evaluation of the quantity @xmath597 .
we have proposed a new decomposition algorithm based on the dual decomposition and excessive gap techniques . the new algorithm requires to perform only one primal step which can be parallelized efficiently , and two dual steps .
consequently , the computational complexity of this algorithm is very similar to other dual based decomposition algorithms from the literature , but with a better theoretical rate of convergence .
moreover , the algorithm automatically updates both smoothness parameters at each iteration .
we notice that the dual steps are only matrix - vector multiplications , which can be done efficiently with a low computational cost in practice .
furthermore , we allow one to solve the primal convex subproblem of each component up to a given accuracy , which is always the case in any practical implementation . an inexact switching variant of algorithm [ alg : a1 ] has also been presented .
apart from the inexactness , this variant allows one to simultaneously update both smoothness parameters instead of switching them .
moreover , it improves the disadvantage of algorithm [ alg : a1 ] when the constant @xmath320 in theorem [ th : alg_scheme ] is relatively small , though it did not outperform algorithm [ alg : a1 ] in the numerical tests .
the worst - case complexity of both new algorithms is at most @xmath324 for a given tolerance @xmath598 .
preliminary numerical tests show that both algorithms outperforms other related existing algorithms from the literature
. 0.2 cm * acknowledgments .
* this research was supported by research council kul : coe ef/05/006 optimization in engineering(optec ) , goa ambiorics , iof - scores4chem , several phd / postdoc & fellow grants ; the flemish government via fwo : phd / postdoc grants , projects g.0452.04 , g.0499.04 , g.0211.05 , g.0226.06 , g.0321.06 , g.0302.07 , g.0320.08 ( convex mpc ) , g.0558.08 ( robust mhe ) , g.0557.08 , g.0588.09 , research communities ( iccos , anmmm , mldm ) and via iwt : phd grants , mcknow - e , eureka - flite+eu : ernsi ; fp7-hd - mpc ( collaborative project strep - grantnr .
223854 ) , contract research : aminal , highwind , and helmholtz gemeinschaft : vicerp ; austria : accm , and the belgian federal science policy office : iuap p6/04 ( dysco , dynamical systems , control and optimization , 2007 - 2011 ) , european union fp7-embocon under grant agreement no 248940 ; cncs - uefiscdi ( project te231 , no .
19/11.08.2010 ) ; ancs ( project pn ii , no .
80eu/2010 ) ; sectoral operational programme human resources development 2007 - 2013 of the romanian ministry of labor , family and social protection through the financial agreement posdru/89/1.5/s/62557 .
in this appendix we provide the full proof of theorem [ th : alg_scheme ] , theorem [ th : alg_scheme2 ] , lemma [ le : init_point ] and lemma [ le : choice_of_tau ] .
0.2 cm _ a.1 .
the proof of theorem [ th : alg_scheme ] .
_ let us denote by @xmath599 , @xmath600 and @xmath601 . from the definition of @xmath26 , the second line of and
, we have @xmath602^ty \!-\ !
\frac{\beta_2^{+}}{2}{\|y\|}^2\right\ } \nonumber\\ & \overset{\tiny{\phi-\mathrm{convex}+\eqref{eq : beta_update}}}{\leq } \max_{y\in\mathbb{r}^m}\big\ { \ ! ( 1 \!-\ ! \tau)\big[\phi(\bar{x } ) \!+\ ! ( a\bar{x } \!-\
! b)^ty \!-\
! \frac{\beta_2}{2}{\|y\|}^2\big]_{[1 ] } + \tau\left[\phi ( \tilde{x}^1 ) \!+\ !
( a\tilde{x}^1 \!-\ ! b)^ty \right]_{[2 ] } \big\}.\end{aligned}\ ] ] now , we estimate two terms in the last line of .
first we note that @xmath603 for any vectors @xmath604 and @xmath37 and @xmath605 .
moreover , since @xmath606 is the solution of the strongly concave maximization with a concavity parameter @xmath191 , we can estimate @xmath607_{[1 ] } & & : = \phi(\bar{x } ) + ( a\bar{x } - b)^ty - \frac{\beta_2}{2}{\|y\|}^2 = \phi(\bar{x } ) + \frac{1}{2\beta_2}{\|a\bar{x } - b\|}^2 - \frac{\beta_2}{2}{\|y - \bar{y}^2\|}^2 \nonumber\\ & & = \phi(\bar{x } ) + \psi(\bar{x};\beta_2 ) - \frac{\beta_2}{2}{\|y - \bar{y}^2\|}^2 \overset{\tiny{\eqref{eq : smoothed_phi}}}{= } f(\bar{x } ; \beta_2 ) - \frac{\beta_2}{2}{\|y - \bar{y}^2\|}^2 \\ & & \overset{\tiny{\eqref{eq : excessive_gap_inexact}}}{\leq } g(\bar{y } ; \beta_1 ) - \frac{\beta_2}{2}{\|y - \bar{y}^2\|}^2 + \delta \overset{\tiny{g(\cdot;\beta_1)-\mathrm{concave}}}{\leq } g(\hat{y } ; \beta_1 ) + \nabla_yg(\hat{y } ; \beta_1)^t(\bar{y } - \hat{y } ) - \frac{\beta_2}{2}{\|y - \bar{y}^2\|}^2 + \delta \nonumber\\ & & \overset{\eqref{eq : app_deriv_g}}{\leq } { \!\ ! } g(\hat{y } ; \beta_1 ) \!+\ ! \widetilde{\nabla}_yg(\hat{y } ; \beta_1)^t(\bar{y } \!-\ ! \hat{y})\!-\ ! \frac{\beta_2}{2}{\|y \!-\ !
\bar{y}^2\|}^2 \!+\ ! ( \bar{y } \!-\ !
\hat{y})^ta(x^1 \!-\ !
\tilde{x}^1 ) \!+\ ! \delta . \nonumber\end{aligned}\ ] ] alternatively , by using , the second term @xmath608_{[2]}$ ] can be estimated as @xmath609_{[2 ] } & & : = \phi(\tilde{x}^1 ) + ( a\tilde{x}^1-b)^ty \nonumber\\ & & = \phi(\tilde{x}^1 ) + ( a\tilde{x}^1 - b)^t\hat{y } + \beta_1p_x(\tilde{x}^1 ) + ( a\tilde{x}^1 - b)^t(y - \hat{y } ) - \beta_1p_x(\tilde{x}^1 ) \nonumber\\ [ -1.6ex]\\[-1.6ex ] & & \overset{\tiny\eqref{eq : inexactness_xi}}{\leq}{\!\!\ ! }
\phi(x^1 ) \!+\ ! ( ax^1 \!-\ ! b)^t\hat{y } \!+\ !
\beta_1p_x(x^1 ) \!+\ ! ( a\tilde{x}^1 \!-\ ! b)^t\!(y \!-\ ! \hat{y } ) \!-\ !
\beta_1p_x(\tilde{x}^1 ) \!+\ ! \frac{\beta_1}{2}\varepsilon_{[\sigma]}^2 \nonumber\\ & & \overset{\tiny{\eqref{eq : smoothed_gi}+\mathrm{lemma}~\ref{le : dual_smoothed_estimates}}}{= } g(\hat{y } ; \beta_1 ) \!+\ ! \widetilde{\nabla}_yg(\hat{y } ; \beta_1)^t(y \!-\ !
\hat{y } ) \!-\ !
\beta_1p_x(\tilde{x}^1 ) \!+\ ! \frac{\beta_1}{2}\varepsilon_{[\sigma]}^2 . \nonumber\end{aligned}\ ] ] next , we consider the point @xmath610 with @xmath219 . on the one hand , it is easy to see that if @xmath94 then @xmath611 .
moreover , we have @xmath612 on the other hand , it follows from that @xmath613 by substituting and into and then using and , we conclude that @xmath614_{[1 ] } + \tau [ \cdot]_{[2 ] } \right\ } \nonumber\\ & & \overset{\tiny{\eqref{eq : th31_proof2}+\eqref{eq : th31_proof3}}}{\leq } \max_{y\in\mathbb{r}^m}\big\{(1-\tau)g(\hat{y};\beta_1 ) + \tau g(\hat{y};\beta_1 ) + \widetilde{\nabla}_yg(\hat{y};\beta_1)^t\left[(1-\tau)(\bar{y}-\hat{y } ) + \tau(y-\hat{y})\right ] \nonumber\\ & & - \frac{(1-\tau)\beta_2}{2}{\|y-\bar{y}^2\|}^2\big\ } - \tau\beta_1p_x(\tilde{x}^1 ) + 0.5\tau\beta_1\varepsilon_{[\sigma]}^2 + ( 1-\tau)\delta + ( 1-\tau)(\bar{y}-\hat{y})^ta(x^1-\tilde{x}^1 ) \nonumber\\ & & { \!\!\!\!\!\!\!\!\!}\overset{\tiny{\eqref{eq : th31_proof4}}}{=}{\!\!\!\ ! } \left[\max_{u\in\mathbb{r}^m}\left\ { g(\hat{y};\beta_1 ) + \widetilde{\nabla}_yg(\hat{y};\beta_1)^t(u - \hat{y } ) - \frac{(1-\tau)\beta_2}{2\tau^2}{\|u-\hat{y}\|}^2\right\ } \right]_{[3]}\nonumber\\ & & + \left[0.5\tau\beta_1\varepsilon_{[\sigma]}^2 + ( 1-\tau)\delta + ( 1-\tau)(\bar{y}-\hat{y})^ta(x^1-\tilde{x}^1 ) - \tau\beta_1p_x(\tilde{x}^1)\right]_{[4]}.\end{aligned}\ ] ] let us consider the first term @xmath608_{[3]}$ ] of . we see that @xmath615_{[3 ] } & & = \max_{u\in\mathbb{r}^m}\left\ { g(\hat{y};\beta_1 ) + \widetilde{\nabla}_yg(\hat{y};\beta_1)^t(u - \hat{y } ) - \frac{(1-\tau)\beta_2}{2\tau^2}{\|u-\hat{y}\|}^2\right\ } \nonumber\\ & & \overset{\tiny{\eqref{eq : th31_proof5}}}{\leq}{\!\!\!\!\ ! } \max_{u\in\mathbb{r}^m}\left\ { g(\hat{y};\beta_1 ) + \widetilde{\nabla}_yg(\hat{y};\beta_1)^t(u - \hat{y } ) - \frac{l^g(\beta_1)}{2}{\|u-\hat{y}\|}^2\right\}\nonumber\\ & & { \!\!\!\!\!\!\!\!\!\!\!\!}\overset{\tiny{\eqref{eq : gradient_mapping_ex}+\eqref{eq : alg_scheme}(\mathrm{line}~3)}}{=}{\!\!\!\!\!\!\!\!\!\ ! } g(\hat{y } ; \beta_1 ) + \widetilde{\nabla}_yg(\hat{y};\beta_1)^t(\bar{y}^{+ } - \hat{y } ) - \frac{l^g(\beta_1)}{2}{\|\bar{y}^{+ } - \hat{y}\|}^2\nonumber\\ & & \overset{\tiny{\eqref{eq : app_deriv_g_est}}}{= } { \!\!\!\ ! }
g(\hat{y};\beta_1 ) \!+\ ! \nabla_yg(\hat{y};\beta_1)^t\!\ ! ( \bar{y}^{+ } \!\!-\ !
\hat{y } ) \!-\ !
\frac{l^g(\beta_1)}{2}{\|\bar{y}^{+ } \!\!-\
! \hat{y}\|}^2 \!+\ ! ( \bar{y}^{+ } \!-\
! \hat{y})^t\!\!a(\tilde{x}^1 \!-\ !
x^1 ) \\ & & \overset{\tiny\eqref{eq : g_lipschitz_bound}}{\leq } g(\bar{y}^{+ } ; \beta_1 ) + ( \bar{y}^{+ } - \hat{y})^ta(x^1-\tilde{x}^1)\nonumber\\ & & \overset{\tiny\eqref{eq : g_beta1_lipschitz_bound}}{\leq } g(\bar{y}^{+ } ; \beta_1^{+ } ) + ( \beta_1-\beta_1^{+})p_x(x^{*}(\bar{y}^{+};\beta_1^{+ } ) ) + ( \bar{y}^{+ } - \hat{y})^ta(x^1-\tilde{x}^1)\nonumber\\ & & \overset{\tiny\eqref{eq : beta_update}+\eqref{eq : d_x_p_x}}{\leq } g(\bar{y}^{+ } ; \beta_1^{+ } ) + \left[\tilde{\alpha}\tau\beta_1d_x + ( \bar{y}^{+ } - \hat{y})^ta(x^1-\tilde{x}^1)\right]_{[5]}\nonumber.\end{aligned}\ ] ] in order to estimate the term @xmath608_{[4 ] } + [ \cdot]_{[5]}$ ] , we can observe that @xmath616 which leads to @xmath617 & \leq l_{\mathrm{a}}^{-1}\beta_1{\|a\|}^2{\|\tilde{x}^1-x^c\| } + l_{\mathrm{a}}^{-1}\beta_1{\|a^t(ax^c - b)\| } + \frac{(1-\tau)\tau}{\beta_2}{\|a\|}^2{\|\bar{x}-x^c\| } \nonumber\\ & + \beta_2^{-1}(1-\tau)\tau{\|a^t(ax^c - b)\| } + ( 1-\tau)\tau{\|a\|}{\|\bar{y}\| } .\end{aligned}\ ] ] note that similar to , we have @xmath618 and @xmath619 . by substituting these estimates into and using the definitions of @xmath608_{[4]}$ ] and @xmath608_{[5]}$ ] we have @xmath620_{[4 ] } + [ \cdot]_{[5 ] } \leq ( 1-\tau)\delta + \frac{\tau\beta_1}{2}\varepsilon_{[\sigma]}^2 + \tau\beta_1(\tilde{\alpha } d_x - p_x(\tilde{x}^1 ) ) + \left[\frac{\beta_1}{l_{\mathrm{a}}}c_d + ( 1-\tau)\tau\left(\frac{c_d}{\beta_2 } + { \|a\|}{\|\bar{y}\|}\right)\right]\varepsilon_{[1]}.\ ] ] by combining , and and noting that @xmath621
, we obtain @xmath622 which is indeed inequality w.r.t .
@xmath199 , @xmath200 and @xmath201 .
@xmath108 0.2 cm _ a.2 .
the proof of theorem [ th : alg_scheme2 ] .
_ let us denote by @xmath623 , @xmath624 , @xmath625 , @xmath626 and @xmath627 . from the definition of @xmath86 , the second line of and @xmath628 we have @xmath629_{[1 ] }
+ \tau \left [ \phi(x ) + ( y^2_{+})^t(ax - b)\right]_{[2]}\big\}.\end{aligned}\ ] ] first , we estimate the term @xmath608_{[1]}$ ] in .
since each component of the function in @xmath608_{[1]}$ ] is strongly convex w.r.t .
@xmath20 with a convexity parameter @xmath139 for @xmath22 , by using the optimality condition , one can show that @xmath630_{[1 ] } & & \overset{\tiny{\eqref{eq : smooth_dual_sol}}}{\geq } \min_{x\in x}\left\{\phi(x ) + \bar{y}^t(ax - b ) + \beta_1p_x(x)\right\ } + \frac{\beta_1}{2}{\|x - x^1\|}_{\sigma}^2 \overset{\tiny\eqref{eq : smoothed_gi}}{= } g(\bar{y } ; \beta_1 ) + \frac{\beta_1}{2}{\|x - x^1\|}_{\sigma}^2 \nonumber\\ [ -1.8ex]\\[-1.8ex ] & & \overset{\tiny{\eqref{eq : excessive_gap_inexact}}}{\geq } f(\bar{x } ; \beta_2 ) + \frac{\beta_1}{2}{\|x - x^1\|}_{\sigma}^2 - \delta \nonumber.\end{aligned}\ ] ] moreover , since @xmath631 , by substituting this relation into we obtain @xmath632_{[1 ] } & & \geq \phi(\bar{x } ) + \psi(\bar{x } ; \beta_2 ) + \frac{\beta_1}{2}{\|x - x^1\|}_{\sigma}^2 - \delta \nonumber\\ & & = \phi(\bar{x } ) + \psi(\bar{x } ; \beta_2^{+ } ) - \tau\psi(\bar{x};\beta_2^{+ } ) + \frac{\beta_1}{2}{\|x - x^1\|}_{\sigma}^2
- \delta \\ & & { \!\!\!\!\!}\overset{\tiny\mathrm{def.}~\psi}{\geq } \phi(\bar{x } ) \!+\ ! \psi(x^2 ; \beta_2^{+ } ) \!+\ ! \nabla_x{\psi}(x^2 ; \beta_2^{+})^t(\bar{x } \!-\ !
x^2 ) \!+\ ! \frac{\beta_1}{2}{\|x \!-\ ! x^1\|}_{\sigma}^2 \!-\ ! \delta + \frac{1}{2\beta_2^{+}}{\|a(\bar{x}-x^2)\|}^2 - \tau\psi(\bar{x};\beta_2^{+ } ) .
\nonumber\end{aligned}\ ] ] here , the last inequality follows from the fact that @xmath633 .
next , we consider the term @xmath608_{[2]}$ ] of .
we note that @xmath634 .
hence , @xmath635_{[2 ] } & & = \phi(x ) + ( y^2_{+})^ta(x - x^2 ) + ( ax^2 - b)^ty^2_{+ } \nonumber\\ & & \overset{\tiny\mathrm{lemma}~\ref{le : primal_smoothed_estimates}}{= } \phi(x ) + \nabla_x{\psi}(x^2 ; \beta_2^{+})^t(x - x^2 ) + \frac{1}{\beta_2^{+}}{\|ax^2-b\|}^2\\ & & = \phi(x ) + \psi(x^2;\beta_2^{+ } ) + \nabla_x{\psi}(x^2;\beta_2^{+})^t(x - x^2 ) + \psi(x^2 ; \beta_2^{+}).\nonumber\end{aligned}\ ] ] from the definitions of @xmath636 , @xmath384 and @xmath182}$ ] we have @xmath637 , @xmath638 and @xmath639}$ ] .
moreover , @xmath640 . by using these estimates
, we can derive @xmath641 ^ 2 \nonumber\\ & & = { \|x-\tilde{x}^1\|}_{\sigma}^2 - 2{\|x-\tilde{x}^1\|}_{\sigma}{\|x^1-\tilde{x}^1\|}_{\sigma } + { \|x^1-\tilde{x}^1\|}_{\sigma}^2 \nonumber\\ [ -1.5ex]\\[-1.5ex ] & & \geq { \|x-\tilde{x}^1\|}_{\sigma}^2 - 2{\|x^1-\tilde{x}^1\|}_{\sigma}\left[{\|x - x^c\|}_{\sigma } + { \|\tilde{x}^1 - x^c\|}_{\sigma}\right]\nonumber\\ & & \geq { \|x-\tilde{x}^1\|}_{\sigma}^2 - 4d_{\sigma}\varepsilon_{[\sigma]}.\nonumber\end{aligned}\ ] ] furthermore , the condition can be expressed as @xmath642 by substituting , and into and then using and note that @xmath643 , we obtain @xmath644_{[1 ] } + \tau [ \cdot]_{[2]}\right\}\nonumber\\ & & { \!\!\!\!}\overset{\tiny{\eqref{eq : thm41_proof2}+\eqref{eq : thm41_proof3}}}{\geq}{\!\!\!\ ! } \min_{x\in x}\big\ { ( 1 \!-\ ! \tau)\phi(\bar{x } ) \!+\ !
\tau\phi(x ) \!+\ ! \nabla{\psi}(x^2;\beta_2^{+})^t\left[(1 \!-\ !
\tau)(\bar{x } \!-\ !
x^2 ) \!+\ ! \tau(x \!-\ !
x^2)\right ] + \frac{(1-\tau)\beta_1}{2}{\|x - x^1\|}^2_{\sigma}\big\ } \nonumber\\ & & - ( 1-\tau)\delta + \left[\tau\psi(x^2;\beta_2^{+ } ) - ( 1-\tau)\tau\psi(\bar{x};\beta_2^{+ } ) + \frac{(1-\tau)}{2\beta_2^{+}}{\|a(\bar{x}-x^2)\|}^2\right]_{[3]}.\end{aligned}\ ] ] we further estimate as follows @xmath645_{[3 ] } - ( 1-\tau)\delta - 2(1-\tau)\beta_1d_{\sigma}\varepsilon_{[\sigma ] } \nonumber\\ & \overset{\tiny\eqref{eq : thm41_proof3b}}{\geq } { \!\!\!\!\!\!\ ! } \min_{u : = ( 1 \!-\ !
\tau)\bar{x } \!+\ ! \tau x \in x}{\!\!\!}\big\{\phi(u ) \!+\ ! \psi(x^2;\beta_2^{+ } ) \!+\ ! \nabla{\psi}(x^2;\beta_2^{+})(u \!-\ !
x^2 ) \!+\ ! \frac{(1 \!-\ ! \tau)\beta_1}{2\tau^2}{\|u \!-\ !
x^2\|}^2_{\sigma}\big\ } \nonumber\\ & - 2(1-\tau)\beta_1d_{\sigma}\varepsilon_{[\sigma ] } - ( 1-\tau)\delta + [ \cdot]_{[3]}\nonumber\\ & { \!\!\!\!\!\!\!\!\!}\overset{\tiny\eqref{eq : alg_scheme_cond_p_tmp}}{\geq}\!\!\ ! \min_{u \in x}\!\big\{\ !
\phi(u ) \!+\ ! \psi(\hat{x};\beta_2^{+ } ) \!+\ ! \nabla{\psi}(x^2;\beta_2^{+})(u \!-\ !
x^2 ) \!+\ ! \frac{l^{\psi}(\beta_2^{+})}{2}{\|u \!-\ ! x^2\|}^2_{\sigma}\big\ } \!-\ ! 2(1 \!-\
! \tau)\beta_1d_{\sigma}\varepsilon_{[\sigma ] } \!-\ !
( 1 \!-\ ! \tau)\delta \!+\ !
[ \cdot]_{[3]}\nonumber\\ & = q(\bar{x}^{*+},y^2_{+};\beta_2^{+ } ) - 2(1-\tau)\beta_1d_{\sigma}\varepsilon_{[\sigma ] } - ( 1-\tau)\delta + [ \cdot]_{[3]}\nonumber\\ & \overset{\tiny\eqref{eq : pi_approx}}{\geq } q(\bar{x}^{+ } , y^2_{+};\beta_2^{+ } ) - 2(1-\tau)\beta_1d_{\sigma}\varepsilon_{[\sigma ] } - ( 1-\tau)\delta + [ \cdot]_{[3 ] } - 0.5\varepsilon_{a}^2\nonumber\\ & \overset{\tiny{\eqref{eq : psi_lipschitz_bound}}}{\geq } f(\bar{x}^{+ } ; \beta_2^{+ } ) - 2(1-\tau)\beta_1d_{\sigma}\varepsilon_{[\sigma ] } - ( 1-\tau)\delta + [ \cdot]_{[3 ] } - 0.5\varepsilon_{a}^2,\end{aligned}\ ] ] where @xmath646^{1/2}$ ] .
to complete the proof , we estimate @xmath608_{[3]}$ ] as follows @xmath647_{[3 ] } & & = \tau\psi(x^2;\beta_2^{+ } ) - \tau(1-\tau)\psi(\bar{x};\beta_2^{+ } ) + \frac{(1-\tau)}{2\beta_2^{+}}{\|a(\bar{x}-x^2\|}^2 \nonumber\\ & & = \frac{1}{2\beta_2^{+}}\left[\tau{\|ax^2 - b\|}^2 - \tau(1-\tau){\|a\bar{x}-b\|}^2 + ( 1-\tau){\|a(\bar{x}-x^2)\|}^2\right]\\ & & = \frac{1}{2\beta_2^{+}}{\|(ax^2-b ) + ( 1-\tau)(a\bar{x}-b)\|}^2 \geq 0.\nonumber\end{aligned}\ ] ] by substituting into and using the definition of @xmath201 in we obtain @xmath648 where @xmath383 } + 0.5\sum_{i=1}^ml_i^{\psi}(\beta_2^{+})\varepsilon_i^2 = ( 1-\tau)\delta + \xi(\tau , \beta_1 , \beta_2 , \varepsilon)$ ] .
this is indeed with the inexactness @xmath201 .
@xmath108 0.2 cm _ a.3 .
the proof of lemma [ le : init_point ] .
_ for simplicity of notation , we denote by @xmath649 and @xmath650 , @xmath651 , where @xmath652 is defined in definition [ de : inexactness ] . by using the inexactness in inequality and @xmath653 , we have @xmath654}$ ] which is rewritten as @xmath655}^2.\end{aligned}\ ] ] since @xmath87 is concave , by using the estimate and @xmath656 we have @xmath657}^2\\ & = \phi(\tilde{x}^ { * } ) + ( a\tilde{x}^ { * } - b)^t\bar{y}^0 - \frac{l^g(\beta_1)}{2}{\|\bar{y}^0\|}^2 + ( \bar{y}^0)^ta(x^ { * } -\tilde{x}^ { * } ) + \beta_1p_x(\tilde{x}^ { * } ) - \frac{\beta_1}{2}\varepsilon_{[\sigma]}^2 : = t_0 \nonumber.\end{aligned}\ ] ] since @xmath658}$ ] , @xmath659 and @xmath285 is the solution of , we estimate the last term @xmath660 of as @xmath661}^2 \nonumber\\ ~&&~ \overset{\tiny\eqref{eq : init_point_cond}+\eqref{eq : accuracy_constants}}{\geq } \phi(\tilde{x}^ { * } ) + \max_{y\in\mathbb{r}^m}\left\{(a\tilde{x}^ { * } - b)^ty - \frac{\beta_2}{2}{\|y\|}^2\right\ } -{\|a^t\bar{y}^0\|}\varepsilon_{[1 ] } - \frac{\beta_1}{2}\varepsilon_{[\sigma]}^2 \\ ~&&~ \overset{\tiny\eqref{eq : psi_approx}}{\geq } f(\bar{x}^0 ; \beta_2 ) - \left[{\|a^t\bar{y}^0\|}\varepsilon_{[1 ] } + \frac{\beta_1}{2}\varepsilon_{[\sigma]}^2\right].\nonumber\end{aligned}\ ] ] now , we see that @xmath662 . thus , @xmath663 for all @xmath22 . by using the definition of @xmath384 in ,
the last inequalities imply @xmath664 finally , we note that @xmath665 due to .
this relation leads to @xmath666 \overset{\tiny\eqref{eq : lm31_proof4b}}{\leq } l_{\mathrm{a}}^{-1}\beta_1\left[{\|a\|}^2d_{\sigma } + { \|a^t(ax^c - b)\|}\right ] \\ & & \overset{\tiny\eqref{eq : accuracy_constants}}{= } l_{\mathrm{a}}^{-1}\beta_1c_d.\nonumber\end{aligned}\ ] ] by substituting and into and then using the definition of @xmath190 we obtain the conclusion of the lemma .
@xmath108 0.2 cm _ a.4 .
the proof of lemma [ le : choice_of_tau ] .
_ let us consider the function @xmath667 , where @xmath668 $ ] and @xmath669 .
after few simple calculations , we can estimate @xmath670 for all @xmath671 .
these estimates lead to @xmath672 from the update rule we can show that the sequence @xmath232 satisfies @xmath673 . if we define @xmath674 , then @xmath675 .
therefore , one can estimate @xmath676 for @xmath677 .
note that @xmath678 by assumption and by induction we can show that @xmath679 for @xmath253 and @xmath680 .
however , since @xmath681 , this leads to : @xmath682 which is indeed . here , @xmath683 and @xmath684^{-1}$ ] .
next , we prove the bound on @xmath326 . since @xmath688
, we have @xmath689 . by using the following elementary inequalities @xmath690 for all @xmath691
$ ] , we obtain @xmath692 , where @xmath693 and @xmath694 . from , on the one hand , we have @xmath695 which leads to @xmath696 for some constant @xmath697 . on the other hand
, we have @xmath698 converging to some constant @xmath699 . combining all estimates together
we eventually get @xmath700 for some positive constant @xmath241 .
m. kojima , n. megiddo , s. mizuno , and et al .
horizontal and vertical decomposition in interior point methods for linear programs . technical report . , information sciences , tokyo institute of technology , tokyo , 1993 .
q. tran - dinh , c. savorgnan , and m. diehl . combining lagrangian decomposition and excessive gap smoothing technique for solving large - scale separable convex optimization problems .
, ( in press):137 , 2012 . | in this paper we propose a new inexact dual decomposition algorithm for solving separable convex optimization problems .
this algorithm is a combination of three techniques : dual lagrangian decomposition , smoothing and excessive gap .
the algorithm requires only one primal step and two dual steps at each iteration and allows one to solve the subproblem of each component inexactly and in parallel .
moreover , the algorithmic parameters are updated automatically without any tuning strategy as in augmented lagrangian approaches .
we analyze the convergence of the algorithm and estimate its @xmath0 worst - case complexity .
numerical examples are implemented to verify the theoretical results .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore |
an important property of quantum - classical correspondence appears in the statistical property of energy levels of bounded quantum systems in the semiclassical limit .
universal behaviors are found in the statistics of _ unfolded _ energy levels at a given interval@xcite , which are the sequence of numbers uniquely determined by the energy levels using the mean level density obtained from the thomas - fermi rule@xcite .
it is widely known that , for quantum systems whose classical counterparts are integrable ( those systems will be referred to as classically integrable systems ) , the distribution of nearest - neighbor level spacing is characterized by the poisson ( exponential ) distribution@xcite , while for quantum systems whose classical counterparts are strongly chaotic , the quantal level statistics are well characterized by the random matrix theory which gives level - spacing distribution obeying the wigner distribution@xcite .
level statistics for the integrable quantum systems has been theoretically studied by berry - tabor@xcite , sinai@xcite , molchanov@xcite , minami@xcite , bleher@xcite , connors and keating@xcite , and marklof@xcite , and have been the subject of many numerical investigations .
still its mechanism is not well understood , the appearance of the poisson distributions is now widely admitted as a universal phenomenon in generic integrable quantum systems .
as suggested , e.g. , by hannay ( see the discussion of @xcite ) , one possible explanation would be as follows : for an integrable system of @xmath5 degrees - of - freedom , almost every orbit is generically confined in each inherent torus , and the whole region in the phase space is densely covered by invariant tori as suggested by the liouville - arnold theorem@xcite .
in other words , the phase space of the integrable system consists of infinitely many tori which have infinitesimal volumes in liouville measure .
then , the energy level sequence of the whole system is a superposition of sub - sequences which are contributed from those regions .
therefore , if the mean level spacing of each independent subset is large , one would expect the poisson distribution as a result of the law of small numbers@xcite .
this scenario suggested by hannay is based on the theory proposed by berry and robnik@xcite .
the berry - robnik theory relates the statistics of the energy level distribution to the phase - space geometry by assuming that the sequence of the energy spectrum is given by the superposition of statistically independent subspectra , which are contributed respectively from eigenfunctions localized onto the invariant regions in phase space .
formation of such independent subspectra is a consequence of the condensation of energy eigenfunctions on disjoint regions in the classical phase space and of the lack of mutual overlap between their eigenfunctions , and , thus , can only be expected in the semi - classical limit where the planck constant tends to zero , @xmath6 .
this mechanism is sometimes referred to as _ the principle of uniform semi - classical condensation of eigenstates_@xcite .
this principle states that the wigner function of a semiclassical eigenstate is connected on a region in phase space explored by a typical trajectory of the classical dynamical system . in integrable systems ,
the phase space is folded into invariant tori , and the wigner functions of the eigenstates tend to delta functions on these tori in the semiclassical limit@xcite . on the other hand , in a strongly chaotic system , almost all trajectories cover the energy shell uniformly , and hence the wigner functions of eigenstates are expected to become a delta function on the energy shell as suggested by the quantum ergodicity theorem@xcite . because of the suppression of the tunneling , each quantum eigenstate is folded into independent subsets in the semiclassical limit , and is expected to form independent spectral components .
indeed , formation of such independent components are checked numerically in a deep semi - classical regime@xcite . in the berry - robnik approach@xcite
, the overall level spacing distribution is derived along a line of mathematical framework by mehta@xcite , as follows : consider a system whose classical phase space is decomposed into @xmath7-disjoint regions .
the liouville measures of these regions are denoted by @xmath8 which satisfy @xmath9 .
let @xmath10 be the gap distribution function which stands for the probability that an interval @xmath11 contains no level .
@xmath10 is expressed by the level spacing distribution @xmath12 as follows ; @xmath13 when the entire sequence of energy levels is a product of the statistically independent superposition of @xmath7 sub - sequences , @xmath14 is decomposed into those of sub - sequences , @xmath15 , @xmath16 in terms of the normalized level spacing distribution @xmath17 of a sub - sequence , @xmath15 is given by @xmath18 and @xmath17 is assumed to satisfy@xcite @xmath19 equations ( [ eq:1 - 2 ] ) and ( [ eq:1 - 4 ] ) relate the level statistics in the semiclassical limit with the phase - space geometry . in most general cases ,
the level spacing distribution might be singular .
in such a case , it is convenient to use its cumulative distribution function @xmath20 ; @xmath21 in addition to equations ( [ eq:1 - 2 ] ) and ( [ eq:1 - 4 ] ) , we assume two conditions for the statistical weights : * assumption ( i ) : the statistical weights of independent regions uniformly vanishes in the limit of infinitely many regions ;
@xmath22 * assumption ( ii ) : the weighted mean of the cumulative distribution of energy spacing , @xmath23 converges in @xmath24 to @xmath25 @xmath26 the limit is uniform on each closed interval : @xmath27 . in the berry - robnik theory ,
the statistical weights of individual components are related to the phase volumes of the corresponding invariant regions .
this relation is satisfactory if the thomas - fermi rule for the individual phase space regions still holds , and thus supports eq.([eq:1 - 4])@xcite .
here we do not specify the validity of this problem and deal with the statistical weights as parameters . under assumptions ( i ) and ( ii ) , eqs.([eq:1 - 2 ] ) and ( [ eq:1 - 4 ] ) lead to the overall level spacing distribution
whose gap distribution function is given by the following formula in the limit of @xmath24 , @xmath28 } \label{eq:1 - 10},\ ] ] where the convergence is in the sense of the weak limit .
when the level spacing distributions of individual components are sparse enough , one may expect @xmath29 and the level spacing distribution of the whole energy sequence reduces to the poisson distribution , @xmath30 in general , one may expect @xmath31 which corresponds to a certain accumulation of the levels of individual components . in the following sections , the above statement is proved and the limiting level spacing distributions are classified into three classes .
one of them is the poisson distribution as discussed in the original work by berry and robnik@xcite .
the others are not poissonian .
we give examples of non - poissonian limiting level spacing distributions in section [ seciii ] . in the concluding section
, we discuss some relations between our results and other related works .
in this section , starting from eqs.([eq:1 - 2 ] ) and ( [ eq:1 - 4 ] ) , and assumptions ( i ) and ( ii ) introduced in the previous section , we show that , in the limit of infinitely many components @xmath24 , the gap distribution @xmath14 converges to the distribution function ( [ eq:1 - 10 ] ) with @xmath32 .
the convergence is shown as follows . following the procedure by mehta(see appendix a.2 of ref.@xcite ) , we rewrite @xmath14 in terms of the cumulative level spacing distribution @xmath33 of independent components : @xmath34 . \label{001}\ ] ] equation ( 1.4 ) and integration by parts lead to @xmath35 where @xmath36 follows from the existance of average , and hence to @xmath37 since the convergence of @xmath38 for @xmath24 is uniform on each interval @xmath39 $ ] by assumption ( ii ) , @xmath14 has the following limit , @xmath40}\nonumber\\ & = & -\sum_{i=1}^n\left[\rho_i\int_0^{s}d\sigma ( 1-\mu_i(\sigma))+o(\rho_i^2)\right]\nonumber\\ & = & -\int_0^s d\sigma \left[1-\mu(\sigma;n)\right ] + \sum_i^n o(\rho_i^2)\label{eq : sum}\\ & & \longrightarrow - \int_0^s d\sigma \left[1-\bar{\mu}(\sigma)\right ] \quad\mbox{as}\quad n\to + \infty \label{lim:01},\end{aligned}\ ] ] where we have used @xmath41 , @xmath42 in @xmath43 , and the following property obtained from assumption ( i ) , @xmath44 with @xmath45 a positive constant .
therefore , we have the desired result : @xmath46}. \label{eq:2 - 7}\ ] ] in this section , we show that in @xmath47 limit , the level spacing distribution @xmath48 converges weakly to @xmath49 , @xmath50 according to the helly s theorem@xcite , the weak convergence of the level spacing distribution is defined by @xmath51 since each side of the above equation is rewritten as , @xmath52_0^s,\ \int_0^s p_{\bar{\mu}}(x)dx = \left[\frac{d}{dx}e_{\bar{\mu}}(x)\right]_0^s,\ ] ] the limit ( [ aaa ] ) is equivalent to @xmath53 the above equation is proved as follows : by using equation ( [ 001 ] ) , we rewrite @xmath54 in terms of the cumulative level spacing distribution function @xmath33 of spectral components , @xmath55 in the limit @xmath47 , one has @xmath56 as shown by equation ( [ eq:2 - 7 ] ) and , @xmath57 where we have used equation ( [ 002 ] ) , @xmath58 in @xmath43 , and the limit ( [ lim : o ] ) .
therefore , we have the desired result : @xmath59 we remark that , when the limiting distribution function @xmath60 is differentiable , the asymptotic level spacing distribution is described as follows : @xmath61
\exp{\left[-\int_0^s \left(1-\bar{\mu}(\sigma)\right)d\sigma \right ] } \label{density}.\ ] ] since @xmath33 is monotonically increasing and @xmath62 , @xmath0 has the same properties .
then , @xmath63 for any @xmath64 and one has @xmath65 according to the above limit , the level spacing distribution is classified into the following three cases in the sense of weak limit : * case 1 , @xmath2 : the limiting level spacing distribution is the poisson distribution .
note that this condition is equivalent to @xmath66 for @xmath67 because @xmath0 is monotonically increasing .
* case 2 , @xmath68 : for large @xmath1 values , the limiting level spacing distribution is well approximated by the poisson distribution , while , for small @xmath1 values , it may deviate from the poisson distribution .
* case 3 , @xmath4 : the limiting level spacing distribution deviates from the poisson distribution for @xmath69 , and decays as @xmath70 more slowly than does the poisson distribution .
this case will be referred to as a sub - poisson distribution .
one has case 1 if the individual cumulative distribution function @xmath33 are bounded by a finite positive function @xmath71 , @xmath72 for @xmath73 and @xmath74 .
indeed , one has , @xmath75 more specifically , for example , one has case 1 if the individual level spacing distributions are derived from scaled distribution functions @xmath76 as ) is described by rewriting @xmath77 in the following way , @xmath78 , so that @xmath79 is a dimensionless function .
for instance , @xmath80 in the case of poisson distribution @xmath81 , and @xmath82 in the case of the wigner distribution @xmath83}$ ] .
] , @xmath84 where @xmath76 satisfy @xmath85 and are uniformly bounded by a positive constant @xmath86 : @xmath87 ( @xmath88 and @xmath64 ) .
indeed , one then has @xmath89 this includes the case studied by berry and robnik@xcite , where the gap distribution is a product of superposition of a single regular component characterized by the poisson distribution and @xmath7 equivalent chaotic components characterized by the wigner distribution , and the latter is expressed by the product of the scaled distributions as : @xmath90 where the statistical weights are @xmath91 and the individual level spacing distributions @xmath76 corresponding to the gap distributions @xmath92 are given by the dimensionless wigner distribution : @xmath93}.\ ] ] indeed , one has the poisson distribution in @xmath47 limit : @xmath94}\longrightarrow e^{-s}.\end{aligned}\ ] ]
as an example of the deviation from the poisson distribution , we study rectangular billiard system whose energy levels are described by using positive integer numbers , @xmath95 and @xmath96 as follows , @xmath97 where @xmath98 is the aspect ratio of two sides of a billiard wall . for a given energy interval @xmath99 $ ] , each energy level is classified into components according to the eigenvalues , @xmath100 ; @xmath101 , \label{eq : cx}\ ] ] where @xmath102 , and @xmath103 $ ] stands for the maximum integer which does not exceed @xmath104 .
the relative weight of each component is given by @xmath105 as easily seen , @xmath106 satisfies the assumption ( i ) ; @xmath107 note that the limit of infinitely many components , @xmath47 , corresponds to the high energy limit , @xmath108(see eq.([eq : cx ] ) ) , which is equivalent to the semiclassical limit in physical systems . in this limit ,
the statistical weight of each sub - spectrium becomes sparse , since each element of @xmath109 , @xmath110 , tends to zero : @xmath111 .
figures 1(a ) and 1(b ) show numerical results of the level - spacing distribution @xmath12 for two values of @xmath98 , and figures 2(a ) and 2(b ) show the gap distribution function corresponding to figures 1(a ) and 1(b ) , respectively . in case
that @xmath98 is far from rational , @xmath12 and @xmath10 are well approximated by the poisson distribution while in case that @xmath98 is close to a rational expressed as @xmath112 , where @xmath113 and @xmath114 are coprime positive integers , they deviate from the poisson distribution@xcite . in order to compare the non - poisson distribution and the classification given in the previous section , we consider , @xmath115 when @xmath47 , @xmath116 approaches @xmath0 , and this function distinguishes the three cases as follows : in case 1 i.e. , where the level spacing obeys the poisson distribution , @xmath117 . in case 2 , @xmath118 as @xmath70 @xmath119 , and in case 3 , where the sub - poisson distribution is expected , @xmath120 as @xmath70 .
figure 3 shows @xmath116 for different values of @xmath7 . the dotted line @xmath121 exhibits the poisson distribution . from this , one can think that @xmath116 for @xmath122 well approximates @xmath123 .
figure 4 shows @xmath116 for the two values of @xmath98 corresponding to figures 2(a)2(b ) , respectively . in case that the numerical data is well characterized by the poisson distribution ( figures 1(a ) and 2(a ) )
, the corresponding function @xmath116 agrees with @xmath124 , while in case that deviates from the poisson distribution ( figures 1(b ) and 2(b ) ) , @xmath125 and @xmath126 for @xmath70 .
therefore , this result corresponds to the case 2 . in this model , we have not yet observed the clear evidence of case 3 .
such a case is expected when there is stronger accumulation of the energy levels of individual components .
in this section , we propose one possible extension of the berry - robnik distribution ( [ eq2 - 12 ] ) for the level statistics of the nearly - integrable system with two degree - of - freedom .
since the classical phase space of the nearly - integrable system consists of regular and chaotic regions and the liouville measures of the chaotic regions are larger than zero , @xmath127 , this system does not support the assumption ( i ) .
however , the regular regions consist of infinitely many subsets , and our approach shown in section [ secii ] is partially applicable to the spectral components corresponding to the regular regions .
following to the assumption proposed by berry and robnik@xcite , the gap distribution functions in the nearly integrable system , which are contributed from the individual chaotic regions , are characterized by the random matrix theory(rmt ) .
then one has , @xmath128 where @xmath129 , and @xmath130 denote the gap distribution functions corresponding to the subsets in the regular regions . as shown by eq.(2.6 ) , @xmath131 has the following limit , @xmath132\ \mbox{as}\ n_2\to+\infty,\ ] ] where @xmath133 and @xmath134 .
accordingly , in the partial limit of @xmath135 , the original proposal for the gap distribution by berry and robnik is replaced by @xmath136}\prod_{i=1}^{n_1 } e_i^{\mbox{\tiny rmt}}(s;\rho_i ) , \label{nf1}\ ] ] where @xmath137 denotes the total amount of the liouville measures of the regular region in mixed phase space .
the above distribution formula is classified into the following three cases ; case 1 , @xmath2 : berry - robnik distribution , case 2 , @xmath3 : berry - robnik distribution for large @xmath1 , but possibly not for small @xmath1 , and case 3 , @xmath4 : a distribution function obtained by the superposition of spectral components obeying the sub - poisson statistics and the random matrix theory . from this classification
, one can see that the new formula ( [ nf1 ] ) admits deviations from the berry - robnik distribution when @xmath138 .
we remark that @xmath139 in the limit @xmath140 converges weakly to the limiting level spacing distribution : @xmath141 , and when the limiting function @xmath60 is differentiable , the asymptotic level spacing distribution admits the following density : @xmath142 \label{nf2}\ ] ] the validity of the berry - robnik distribution has been checked for generic nearly - integrable systems by many numerical investigations@xcite . among them , prosen and robnik found numerically for several systems that there is a high energy region in which the berry - robnik distribution formula ( [ eq2 - 12 ] ) well approximates the level spacing distribution@xcite .
this energy region is sometimes referred to as _ the berry - robnik regime_@xcite . while they also found that the level spacing distribution in the low energy region deviates from the berry - robnik formula , and approximates the brody distribution .
this behavior was studied in terms of a fractional power dependence of the spacing distribution near the origin at @xmath143 , which could be attributed to the localization properties of eigenstates on chaotic components@xcite . from the above classification , case 1:@xmath2
sould be satisfied in the berry - robnik regime . while case 2 and case 3 might propose another possibilities .
when the spectral components corresponding to regular regions show strong accumulation , the gap distribution obeys the distribution formula ( [ nf1 ] ) with @xmath144 , and the level spacing distribution shows deviations from the berry - robnik distribution .
therefore , this result might propose another possibility of the berry - robnik approach .
in this paper , we investigated the gap distribution function of systems with infinitely many independent components and discussed the level - spacing statistics of classically integrable quantum systems . in the semiclassical limit , reflecting infinitely fine classical phase space structures , individual energy eigenfunctions are expected to be well localized in the phase space and contribute independently to the level statistics . keeping this expectation in mind , we considered a situation in which the system consists of infinitely many components and each of them gives an infinitesimal contribution . and
by applying the arguments of mehta , and berry and robnik , the limiting level spacing distribution was obtained whose gap distribution function is described by a single monotonically increasing function @xmath0 of the level spacing @xmath1 : @xmath145 } \label{eq : xs}\ ] ] the weak convergence limit of the level spacing distribution is classified into three cases ; case 1 : poissonian if @xmath2 , case 2 : poissonian for large @xmath1 , but possibly not for small @xmath1 if @xmath3 , and case 3 : sub - poissonian if @xmath4 . thus , even when the energy levels of individual components are statistically independent , non - poissonian level spacing distributions are possible . in most general cases ,
the integral in equation ( [ eq : xs ] ) converges in @xmath146 and then @xmath147 , the limiting gap distribution @xmath148 does not work accurately .
in such case , however , its differentiation ( [ rrr ] ) still work accurately in @xmath70 limit@xcite , and thus the above classification ( case 13 ) holds in general .
note that the singular level spacing distribution can be taken into account in terms of non - smooth cumulative distributions .
such a singularity is expected when there is strong accumulation of the energy levels of individual components . for a certain class of systems ,
such accumulation is observable .
one example is shown in section iii where the results show clear evidence of case 2 .
another example is the two - dimensional harmonic oscillator whose level spacing distribution is non - smooth for arbitrary system parameter@xcite .
the final example is studied by shnirelman@xcite , chirikov and shepelyansky@xcite , and frahm and shepelyansky@xcite for a certain type of system which contains a quasi - degeneracy result from inherent symmetry(time reversibility ) .
as is well known , the existence of quasi - degeneracy leads to the sharp shnirelman peak at small spacings . finally , in section [ extension ] , we proposed one possible extension of the berry - robnik distribution for classically nearly - integrable quantum systems .
this extension admitted deviations from the berry - robnik distribution when there is strong accumulation of the energy levels of spectral components .
such possibilities will be studied elsewhere .
+ + acknowledgments + the authors would like to thank professor m. robnik , professor y.aizawa , professor a.shudo , and dr g. veble for their helpful advices and discussions .
the authors also thank the yukawa institute for theoretical physics at kyoto university .
discussions during the yitp workshop yitp - w-02 - 13 on `` quantum chaos : present status of theory and experiment '' were useful to complete this work . this work is partly supported by a grant - in - aid for scientific research ( c ) from the japan society for the promotion of science .
berry and m. robnik , j. phys .
a*17*,2413(1984 ) .
i. c. percival , j. phys .
b * 6*,l229(1973 ) .
berry and m. tabor , proc .
r. soc .
a * 356*,375(1977 ) .
o. bohigas , m.j .
giannoni , and c. schmit , phys .
* 52*,1(1984 )
. o. bohigas , _ random matrices and chaotic dynamics
_ ( leshouches 1989 session lii , chaos and quantum physics , north - holland ) .
mehta,_random matrices _
( 2nd ed . , san diego , ca : academic press , 1991 ) ya .
g. sinai , physica a*163*,197(1990 ) .
s. a. molchanov , commun .
, * 78*,429(1981 ) : s. a. molchanov , math .
ussr izvestija,*12*,69(1978 ) .
n. minami , prog .
sup pl.*116*(1994 ) .
m. bleher , j .
phys.*61*,869(1990 ) : p. m. bleher , j .
phys.*63*,261(1991 ) .
r. d. connors and j. p. keating , j. phys .
a*30*,1817(1997 ) .
j. marklof , comm .
phys.*199 * , 169(1998 ) . v. i. arnold and a. avez,_ergodic problems of classilcal mechaniocs _ ,
( benjamin , n. y.,1978 ) : v. i. arnold , russ . math . surveys * 18:6 * 85(1963 ) .
w. feller , _ an introduction to probability theory and its applications _
( 2nd ed . , john wiley & sons , inc .
, new york , 1957 ) m. v. berry , j .
a*10*,2083(1977 ) .
m. robnik , nonlinear phenomena in complex systems * 1*,n1,1(1998 ) .
m. v. berry , phil .
r. soc .
a*287*,237(1977 ) .
a. i. shnirelman , usp .
nauk * 29*,181(1974 ) .
s. zelditch , duke .
j.*55*,919(1987 ) .
m. robnik and t. prosen , j. phys .
a*30*,8787(1997 ) . ya .
g. sinai , _ kurs teorii veroyatnostej _ published by mgu , theorem 13.1 m. robnik and g. veble , j. phys .
a*31*,4669(1998 ) .
t. prosen and m. robnik , j. phys .
a*27*,l459(1994 ) .
t. prosen and m. robnik , j. phys .
a*27*,8059(1994 ) .
b. li and m. robnik , j .
a*28*,4843(1995 ) .
t. prosen , physica d*91*,244(1996 ) .
h. makino , t. harayama , y. aizawa , phys .
e,*59*,4026(1999 ) .
h. makino , t. harayama and y. aizawa , prog .
phys.suppl.*139*,477(2000 ) .
h. makino , t. harayama and y. aizawa , phys .
e , * 63*,056203(2001 ) .
h. makino and s. tasaki , ( to appear ) a. i. shnirelman , usp . mat .
a)*30*,n4,265(1975 ) .
b. v. chirikov and d. l. shepelyansky , phys .
lett . * 74*,518(1995 ) . k. m. frahm and d. l. shepelyansky , phys
. rev .
lett . * 78*,1440(1997 ) .
fig.1 numerical results of the level spacing distribution @xmath12 for ( a ) @xmath149 , ( b ) @xmath150 .
we used energy levels @xmath151 with @xmath152 $ ] .
total numbers of levels are ( a ) 10000016 , ( b ) 10000046 .
the dotted curve in each figure shows the poisson distribution : @xmath153 .
[ fig_1 ] fig.3 @xmath116 for @xmath155 and for @xmath156 . in each case
, we fixed @xmath157 , and used @xmath158 energy levels with @xmath159 $ ] and @xmath160 energy levels with @xmath152 $ ] , respectively .
the dashed line ( @xmath121 ) exhibits the poisson distribution . | along the line of thoughts of berry and robnik@xcite , the limiting gap distribution function of classically integrable quantum systems is derived in the limit of infinitely many independent components .
the limiting gap distribution function is characterized by a single monotonically increasing function @xmath0 of the level spacing @xmath1 , and the corresponding level spacing distribution is classified into three cases : ( i ) poissonian if @xmath2 , ( ii ) poissonian for large @xmath1 , but possibly not for small @xmath1 if @xmath3 , and ( iii ) sub - poissonian if @xmath4 .
this implies that even when the energy - level distributions of individual components are statistically independent , non - poissonian level spacing distributions are possible . |
Subsets and Splits
Data Mining Abstracts
This query searches for and retrieves 2000 documents containing the phrase 'data mining' in the abstract, providing limited but specific document samples.