article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
coronal mass ejections ( cmes ) , are large expulsions of magnetic field and plasma from the solar corona .
the kinetic and gravitational potential energy contained in a cme is around @xmath0 ergs , making these events some of the most energetic in our solar system @xcite .
it is thought that cmes derive their energy from the magnetic field of the solar corona because this field is the only possible source for such a large reserve of energy ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the flux rope model is one possible pre - eruptive cme configuration .
a flux rope is a length of magnetic field that has been twisted along its axis , often held in place in the corona by an overlying arcade or ambient field .
it is thought that cool photospheric plasma can become be trapped in the center of a flux rope , creating a solar filament or prominence @xcite .
the flux rope configuration easily explains the clear three - part structure seen in many cmes , specifically those associated with prominence eruptions @xcite . because these structures are present in the low corona where magnetic field is strong and plasma density is low , they are magnetic - field - dominated .
recently , movies from _
hinode _ have shown interesting dynamics that are not described by the flux rope model which imply that at least some prominences are not low-@xmath1 @xcite , but the flux rope model remains useful .
the lower coronal environment is also frequently modeled as being force - free because flow speeds are low , and @xmath2 is the dominant force in the equation of motion .
gravity is also frequently ignored because it is a factor of @xmath3 weaker than the magnetic forces .
low-@xmath1 flux ropes are stable when the outward magnetic pressure force is balanced by an inward - directed tension force . in models , an exterior arcade field is often added to increase the tension force and keep the flux rope from simply expanding in length and width as twist is added .
the approximate energy per unit length along the axis , @xmath4 , stored in an unconfined gold - hoyle flux rope @xcite is given by @xmath5 where @xmath6 is the magnetic flux , @xmath7 is the radius of the tube , and @xmath8 is the twist parameter such that @xmath9 @xcite .
as the twist accumulates , @xmath8 increases , @xmath6 is constant and @xmath7 remains approximately constant so the total energy increases . when an unconfined flux rope anchored at both ends on the photosphere accumulates twist , its equilibrium state is expanded in length relative to the untwisted state so the twist per unit length does not necessarily increase with the total twist .
thus , flux ropes that are confined by an overlying arcade contain more energy because their length changes very little as @xmath8 increases .
the energy stored in the magnetic field is given by the volume integral of @xmath10 , up to conversion factors .
the minimum energy of a magnetic system with a given photospheric boundary occurs in the potential , or vacuum field , configuration .
as the field is stressed away from this configuration due to photospheric movements , the energy is increased above the evolving potential state by an amount commonly referred to as the magnetic ` free energy ' . if the field reverts to the potential configuration , this free energy is released and in the case of solar active regions , is available to drive a cme .
if the reconnection is localized and helicity is conserved , the lowest accessible energy state may not be potential , so the free energy is an upper boundary on the amount of energy that can be released .
there is a global cap on the amount of energy that the magnetic field can provide .
the magnetic virial theorem asserts that the total pressure force can not exceed the tension force in a stable plasma environment @xcite .
the related aly - sturrock conjecture states that the global magnetic energy of a force free field is at a maximum when the field is completely ` open ' .
this refers to magnetic flux that is anchored at the solar surface and extends radially outward a significant distance so that , near the sun , the field appears open @xcite .
many cme observations show prominences lifting off of the surface of the sun , expanding to several solar radii and leaving behind long radial field lines .
if this conjecture is correct , then the implication is that cmes which open large amounts of field must derive the bulk of their kinetic energy from sources other than the magnetic field because the field energy is actually greater in the post - cme configuration .
order of magnitude analysis has shown , however , than the magnetic field is the only source of energy that can potentially drive a @xmath11erg cme @xcite .
this poses a significant problem to ideal cme models . in an azimuthally
symmetric 2.5-d case , all of the field lines originally above a prominence - like feature would have to open to release the filament ( fig .
[ fig:2.5d vs 3d]a ) .
thus , in the 2.5-d case , to have an eruption which results in a net decrease of magnetic energy , reconnection must be present .
there have been studies in which a 2.5-d field is shown to have magnetic energy exceeding the open field energy when mass loading is present , but it has not been demonstrated that these fields can erupt without reconnection @xcite . in three dimensions , the flux rope is anchored in the photosphere , and the surrounding field can move away in the direction parallel to the flux rope axis ( fig .
[ fig:2.5d vs 3d]b ) .
reconnection is not necessary in the fully three - dimensional case , as not all of the field must be open to have an eruption , only the flux rope opens , and thus the eruption is not relevant to the hypothesis of aly and sturrock because some of the field remains closed @xcite .
the full aly - sturrock conjecture has yet to be disproved .
however , it is not usually relevant to 3-dimensional models and analysis for the reasons stated above .
a more relevant question which has been asked , is whether a configuration with _ some _ open force free field can contain less energy than a configuration with the same boundary conditions that is fully closed @xcite .
this question has been addressed semi - analytically by @xcite and @xcite , who showed that a fully closed field can contain magnetic free energy above the partially open field threshold , but they did not demonstrate a release mechanism for this energy .
a demonstration of the ideal evolution of a field whereby free energy is stored then released , resulting in a partially open state , would once and for all eliminate the problem posed by the magnetic virial theorem and the aly - sturrock conjecture in a fully 3d case .
@xcite describe the possibility of driving cmes with metastable magnetic fields .
they analytically describe one system in particular as a known metastable state : the previously mentioned twisted flux rope under an overlying arcade .
this system is metastable because it is stable ( due to the confining arcade ) against small perturbations , but the energy of the erupted flux rope is lower than the contained flux rope .
if the rope is tightly wound , it can open and escape the arcade by herniating through it , leaving the deflated arcade near the footpoints .
the amount of twist needed is not unreasonable for the solar surface .
analytically , for a flux rope that is ten times longer than its radius , the rope need only exceed 1.5 total turns about its axis to be in this metastable state @xcite .
numerous simulations exist which model flux rope cme initiation of the metastable configuration described above ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
most of these simulations have found that it is possible to herniate through the arcade , but they do not agree exactly on how much twist is needed , or how unstable the resulting configuration is after the onset of writhe ( helical geometry in the central field line ) in the flux rope . typically , these codes agree that the the critical twist needed to erupt is around 1.5 turns , and that it is possible to get an eruptive event by twisting the footpoints of a flux rope .
cme initiation with these flux ropes has been modeled both with and without reconnection as the intended primary destabilizing factor .
theories that do not involve reconnection are referred to as ideal `` loss of equilibrium '' models ( e.g. * ? ? ?
the initial structure generally undergoes an ideal instability , such as the mhd kink instability , caused by a large amount of twist .
another possibility , if mass loading is critical in keeping the structure contained , is that mass displacement could upset the force balance and start the cme @xcite .
other theories , such as tether cutting @xcite or `` breakout '' @xcite , explicitly include reconnection to decrease the strength of the overlying arcade . in models such as tether cutting and breakout , there are generally two stages of reconnection . slow
sweet - parker style @xcite reconnection occurs early in the evolution and destabilizes the system , allowing it to expand .
this is often followed by fast petcheck style @xcite reconnection , which releases large amounts of energy in a short time and is believed to be the primary driver for fast , impulsive cmes .
essentially all existing numerical simulations of cme onset use eulerian methods , in which a 3-d grid of values is used . with magnetic fields ( and indeed all vector fields and flows ) , sharp gradients are not conserved because derivatives are represented as finite differences . for magnetic fields
, this means that the field will reconnect if gradients approach the size of the grid whether the modeler wishes it or not .
techniques such as adaptive mesh refinement can reduce the rate of numerical reconnection , but can not remove it altogether .
hence it is not possible to separate the effects of ideal mhd evolution and magnetic reconnection with an eulerian grid code .
this can be problematic insofar as the reconnection destabilizes a metastable system . by switching to a lagrangian ( field - aligned ) formulation ,
we eliminate all reconnection , allowing study of ideal mhd instabilities @xcite . with our model , we are able to analyze simplified systems where topology is locked in and reconnection is not present .
note that we do not hypothesize that reconnection is not present in the sun , only that to have a controlled numerical experiment , the effect of reconnection must be isolated , and we do this by eliminating it .
our method is unique in this way , and may offer insights that grid simulations can not .
in particular , we are able to demonstrate the existence of a true mhd instability that can release free energy into a cme , even with no triggering reconnection .
the code used in this work is called flux ( fieldline universal relaxer ) @xcite .
this quasi - lagrangian code represents a three - dimensional field as a collection of _ fluxons , _ or field lines with finite magnetic flux .
each fluxon is broken into piecewise linear segments called _ fluxels _ , which are joined at vertex points ( fig .
[ fig : fluxels ] ) .
to reconnect , a fluxon must be explicitly broken and connected to another fluxon . with no reconnection ,
the code preserves magnetic topology ; this is the case used in the current work .
flux is coordinate free , so in order to compare the simulations with the sun , we assume that our system is originally the size of an active region , which is a few tens of mm across ( 1 spatial unit = 25 mm ) .
flux is under development and the version we used for this work is not a full magnetohydrodynamic code .
it does not include the effects of mass or plasma ; thus it does not model dynamics .
we are neglecting short time - scale changes of the system in favor of concentrating on the large scale evolution .
flux computes nonlinear force - free equilibrium solutions from a prescribed initial topology and connectivity by balancing the components of the lorentz force , which are resolved as a magnetic pressure and a magnetic tension force:@xmath12 the tension force is computed from the geometry of the other fluxels on the same fluxon ( e.g. the angle between successive fluxels ) , and the pressure is computed based on the geometry of the nearest surrounding neighboring fluxels .
each vertex is moved in the direction of net calculated force until the ratio of the net force to the sum of the magnitudes of the forces on each vertex is below a threshold level .
once all vertices are below this threshold , e.g. @xmath13 , the field is deemed to be in equilibrium .
( a more detailed mathematical description is available in @xcite ) .
initial conditions in the code consist of a planar line - tied photosphere - like boundary with a set connectivity .
the footpoint of each fluxon can be moved independently to simulate photospheric motions .
after each footpoint movement , the field is allowed to relax to equilibrium before the next movement occurs . in this way , it is possible to create a quasi - static evolution of equilibrium states .
the simulation is bounded at the top by an open hemisphere .
fluxons that intersect this surface are free to move around on it .
closed loops that approach the surface are truncated and become two separate fluxons that then move independently .
open fluxons move to equalize magnetic pressure ( there is no curvature on the final fluxel ) , which has the effect that the magnetic field at the upper boundary is normal to the surface .
the simulated systems consist of a flux rope , an overlying arcade , and an outer ring of open field lines .
figure [ fig : sequence](a ) shows this set - up .
the fluxons are tied to a planar lower boundary and evolve with a prescribed surface motion .
the central flux rope is twisted incrementally in a solid body rotation pattern by four degrees each step and allowed to relax to equilibrium.the flux rope footpoints are set at 2 spatial units apart , or 50 mm .
( a)-(c ) initial simulation set - up and evolution of a system with 75 fluxons where the arcade is 25 fluxons shown in red and the flux rope is 25 fluxons in shown in a gradient from blue to green to show directionality .
the central fluxon is in black to highlight the flux rope axis .
the outer ring of 25 fluxons is shown in light red .
( a ) the equilibrium system with no twist , ( b ) with 1 turn , which has not kinked , ( c ) with 1.43 turns which has herniated and is open at the upper boundary .
( d ) shows a different system with 9 fluxons in the flux rope and 3 units of flux ( 27 fluxons ) in the arcade at 2.1 turns , which demonstrates kinking without herniation .
note that the systems with one unit of arcade flux do exhibit a slight kink while confined as evidenced by a small distortion in the central fluxon , but the systems with higher arcade flux exhibit much stronger kink before herniation.,title="fig:",width=264 ] ( a)-(c ) initial simulation set - up and evolution of a system with 75 fluxons where the arcade is 25 fluxons shown in red and the flux rope is 25 fluxons in shown in a gradient from blue to green to show directionality .
the central fluxon is in black to highlight the flux rope axis .
the outer ring of 25 fluxons is shown in light red .
( a ) the equilibrium system with no twist , ( b ) with 1 turn , which has not kinked , ( c ) with 1.43 turns which has herniated and is open at the upper boundary .
( d ) shows a different system with 9 fluxons in the flux rope and 3 units of flux ( 27 fluxons ) in the arcade at 2.1 turns , which demonstrates kinking without herniation .
note that the systems with one unit of arcade flux do exhibit a slight kink while confined as evidenced by a small distortion in the central fluxon , but the systems with higher arcade flux exhibit much stronger kink before herniation.,title="fig:",width=264 ] ( a)-(c ) initial simulation set - up and evolution of a system with 75 fluxons where the arcade is 25 fluxons shown in red and the flux rope is 25 fluxons in shown in a gradient from blue to green to show directionality .
the central fluxon is in black to highlight the flux rope axis .
the outer ring of 25 fluxons is shown in light red .
( a ) the equilibrium system with no twist , ( b ) with 1 turn , which has not kinked , ( c ) with 1.43 turns which has herniated and is open at the upper boundary .
( d ) shows a different system with 9 fluxons in the flux rope and 3 units of flux ( 27 fluxons ) in the arcade at 2.1 turns , which demonstrates kinking without herniation .
note that the systems with one unit of arcade flux do exhibit a slight kink while confined as evidenced by a small distortion in the central fluxon , but the systems with higher arcade flux exhibit much stronger kink before herniation.,title="fig:",width=264 ] ( a)-(c ) initial simulation set - up and evolution of a system with 75 fluxons where the arcade is 25 fluxons shown in red and the flux rope is 25 fluxons in shown in a gradient from blue to green to show directionality .
the central fluxon is in black to highlight the flux rope axis .
the outer ring of 25 fluxons is shown in light red . (
a ) the equilibrium system with no twist , ( b ) with 1 turn , which has not kinked , ( c ) with 1.43 turns which has herniated and is open at the upper boundary .
( d ) shows a different system with 9 fluxons in the flux rope and 3 units of flux ( 27 fluxons ) in the arcade at 2.1 turns , which demonstrates kinking without herniation .
note that the systems with one unit of arcade flux do exhibit a slight kink while confined as evidenced by a small distortion in the central fluxon , but the systems with higher arcade flux exhibit much stronger kink before herniation.,title="fig:",width=264 ] one difficulty in examining these results has been with the energy calculation .
our code calculates the energy of every fluxel based on the cross sectional area it occupies and its length . in regions where the fluxons are close to each other
, this method works extremely well , but it has more trouble for the outer - most fluxons in a system .
we call this the ` last - fluxon ' problem and it is discussed by @xcite .
the volume that the last fluxon occupies is infinite , so it can not be treated as small , violating the approximation used by the code .
once the system has herniated and expanded fully , the number of last - fluxons is much greater , exaggerating the difficulty in determining the post - eruption energy .
the open hemispherical surface at @xmath14 and an outer ring of fluxons were added to alleviate this problem . as a consequence of the surface
, the field opens once it encounters the hemisphere and expands to fill the volume .
the footpoints of the outer ring remain stationary throughout the simulation .
this outer ring is present to minimize the ` last fluxon problem ' because with the ring , none of the arcade or flux rope fluxons will be a last fluxon .
the ring is positioned far from the rope and the arcade , about @xmath15 away , so it does not effect the evolution .
the outer boundary at a radius of @xmath14 is much farther than the physical regime of applicability of flux .
the transition to solar wind occurs at @xmath16@xcite , so at most , our results have physical meaning up to this height .
we performed several simulations with varying numbers of fluxons in the flux rope and the arcade , each with the same basic set - up .
the flux rope consists of 9 , 16 , 25 , 36 or 49 fluxons arranged in a square on the photosphere .
the footprint of the flux rope is the same size in each case . for the rest of this paper
, a _ unit _ of magnetic flux refers to the amount of magnetic flux associated with the flux rope , or the number of fluxons in the flux rope . in these simulations , the flux rope and the ring consist of one unit of flux , and the arcade contains one , two , or three units .
in all cases , we find that the flux rope herniates through the arcade after a certain amount of twist has been applied , entraining a few arcade fluxons with it as it goes .
figure [ fig : sequence](a)-(c ) demonstrates a typical sequence of events for a case of a 25 fluxon flux rope and a one - unit magnetic flux arcade .
first , the flux rope twists about its central axis under the arcade . after about 1.4 turns have been applied to the flux rope ( for one unit of arcade flux ) , the rope herniates . in this case
, the flux rope does not significantly kink the central axis remains mostly un - twisted but in the case of a stronger arcade , the flux rope does kink before it herniates .
the stronger the arcade , the flatter the flux rope is , and the more twist is required to initiate herniation .
figure [ fig : sequence](d ) shows the three - unit arcade system after it has undergone writhe ; the black central fluxon is no longer straight .
the onset of kink does not trigger the herniation through the arcade .
the flux rope continues to twist and writhe until it begins to herniate . in every case , after the onset of herniation , the flux rope expands rapidly while the arcade deflates .
figure [ fig : height plot ] shows a plot of height of the flux rope vs. twist imparted , for various fluxon densities with one unit of arcade flux .
the expansion occurs extremely rapidly , within one equilibrium step , once the flux rope breaks through the arcade .
the flux rope remains well confined as it is twisted until it herniates rapidly through the arcade .
the data shown are for systems with one unit of flux in the arcade . ]
we label this rapid expansion as an `` eruption , '' because the size of the flux rope increases by a factor of 100 or more and breaks through the open surface in a single equilibrium time step , while releasing a non - trivial amount of energy .
thus , these large expansions are deemed eruptions and the flux rope fluxons are labeled as open . the amount of twist needed to herniate through a given arcade strength varies with the number of fluxons used to represent the field ( fig .
[ fig : height plot ] ; this may be indicative of grid effects that are setting an unstable twist level or seeding the instability . because of the discrete nature of fluxons , the system is not always symmetric , and this probably accounts for some uncertainly in the critical twist
there is not always a consistent trend with fluxon density , and so there may be other reasons behind this behavior . the energy of the final erupted state is less than the energy of the confined flux rope ( figure [ fig : energy ] ) . note that the presence of the open boundary does not skew these results .
the magnetic energy that escapes through the boundary would not be available to drive the cme in any case because it is present beyond the transition to the solar wind . in the latter stages of the expansion , the twist per unit length in the flux rope is small , and hence the free energy is low .
it is the initial expansion that drives the cme , not the later expansion .
compared to the physical case , we overestimate the final magnetic energy of the system because our boundary is much farther out than the transition .
a significant amount of energy is available to drive a cme , even without reconnection . a typical energy plot for the simulations run .
this one shows the energy evolution of a 25 fluxon flux rope with a one unit flux arcade .
as the flux rope is twisted under the arcade , the energy increases steadily .
once the system herniates , the energy loss is substantial .
note that the energy in the flux rope and arcade only ( black line ) drops to lower than the initial level .
this is because most of the energy in the system at that point is in the outer ring fluxons which are highly stressed toward the horizontal.,width=216 ] this simulation allows us to put a strong upper and lower bound on the amount of energy that is released with the eruption .
of the free energy injected , @xmath17 is lost after herniation .
the energy calculated in the ring field after herniation is an overestimate , and the energy in the flux rope and the arcade is an underestimate for the following reasons .
the ring field was added so that all of the last fluxons were that ring . as stated earlier , we do not trust the energy calculation for the ring field , especially considering that before herniation less than @xmath18 of the system s energy was in the ring field compared to @xmath19 after herniation .
also , on a spherical solar surface , the ring field would be farther away than the disk limb if the flux rope were at disk - center , and consequently would not be highly sheared away from radial after the eruption , so it would not store much more energy than it had initially .
because of the unreliability of the final energy in the ring field , we also looked at the energy in only the flux rope and the arcade . in this partial system ,
the final energy is less than the initial potential energy in part because some of the energy is carried through the open boundary and lost from the calculation , and in part because this limited system does not account for any background solar field that may be deformed by the cme . despite these effects
, we are able to conclude that a significant amount of energy is released and could be used to drive an impulsive cme .
the best way to resolve the energy would be to run a similar system as a full - sun simulation in spherical geometry so that there are no last - fluxons , and include an estimate of the energy outside the upper boundary with a force free field extrapolation
. this future work may be able to determine quantitatively how much magnetic energy available ; a figure which is highly dependent on the geometry of each event .
our simulations show that reconnection is not necessary to initiate a cme and that impulsive cmes may be possible without explosive reconnection .
this theory is not new ; it was originally published by @xcite , who describe the existence of metastable states , specifically a system similar to the one we have studied . since then
, other solar physicists have studied this system computationally ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* etc . ) .
the results from these studies show that a highly twisted flux rope can herniate through a confining magnetic arcade and reconnect into a plasmoid , causing an eruption . however , this is the first study of this system in the complete lack of reconnection . in general ,
our results agree with those of other research groups .
the fact that many of these simulations , including ours , agree that about 1.5 total turns is needed to herniate through an arcade implies that reconnection is not greatly important to the overall stability of the system .
if it were , then we would expect our ideal simulation to support significantly more twist and therefore release more energy than the dissipative simulations .
the exact value of the critical twist will depend on the configurations of the system : strength of the arcade , width of the flux rope , the twist profile within the flux rope , etc .
but even with these variables , we conclude that highly twisted flux ropes can not easily be confined by external field , even when the reconnection rate is extremely small . current research on the twist available photospheric fields indicates that there may be an excess of one full turn available in many active regions @xcite .
this implies that many pre - eruptive active regions may be on the brink of an ideal instability when they flare or erupt , regardless of the eventual trigger mechanism .
our results together with the results of @xcite and @xcite can finally put to rest the concerns that the aly - sturrock conjecture have created over the initiation of cmes .
although it has not been proven that a closed force free field can have more energy than a fully open one , that question is not relevant in the complex three dimensional system that is our sun .
a more appropriate question is whether a closed force free field can contain more energy than a configuration with some open field given the same lower boundary and connectivity .
@xcite began to answer that question with semi - analytic techniques and successfully showed that it was possible .
we have proven that it is possible to transition between closed and partially open field while still releasing free energy without reconnection and without the need for gravitational or other non - magnetic confinement . at the beginning of the century
there was , `` still no model which demonstrates that a partly open magnetic field can be achieved solely by a loss of ideal mhd equilibrium or stability . ''
happily , this statement is no longer true .
this work was funded by nasa s lws - tr&t program .
flux is open source software available from http://flux.boulder.swri.edu .
thanks to the pdl development team http://pdl.perl.org .
we also owe thanks to spiro antiochos , bernhard kliem and zoran mikic for valuable discussions of the fluxon technique .
t. e. berger , r. a. shine , g. l. slater , t. d. tarbell , a. m. title , t. j. okamoto , k. ichimoto , y. katsukawa , y. suematsu , s. tsuneta , b. w. lites , and t. shimizu . .
_ _ , 676:0 l89l92 , march 2008 .
doi : 10.1086/587171 .
r. c. canfield , c .- c .
cheng , k. p. dere , g. a. dulk , d. j. mclean , e. j. schmahl , r. d. robinson , jr . , and s. a. schoolman . .
in p. a. sturrock , editor , _ skylab solar workshop ii _ , pages 451469 , 1980 .
a. j. hundhausen , a. l. stanger , and s. a. serbicki . . in j. j. hunt , editor , _ solar dynamic phenomena and solar wind consequences , the third soho workshop
, volume 373 of _ esa special publication _ , pages 409+ , december 1994 .
j. l. kohl , g. noci , e. antonucci , g. tondello , m. c. e. huber , l. d. gardner , p. nicolosi , l. strachan , s. fineschi , j. c. raymond , m. romoli , d. spadaro , a. panasyuk , o. h. w. siegmund , c. benna , a. ciaravella , s. r. cranmer , s. giordano , m. karovska , r. martin , j. michels , a. modigliani , g. naletto , c. pernechele , g. poletto , and p. l. smith . .
_ _ , 175:0 613644 , october 1997 .
doi : 10.1023/a:1004903206467 . | we demonstrate that magnetic reconnection is not necessary to initiate fast cmes .
the aly - sturrock conjecture states that the magnetic energy of a given force free boundary field is maximized when the field is open .
this is problematic for cme initiation because it leaves little or no magnetic energy to drive the eruption , unless reconnection is present to allow some of the field to escape without opening .
thus , it has been thought that reconnection must be present to initiate cmes .
this theory has not been subject to rigorous numerical testing because conventional mhd numerical models contain numerical diffusion , which introduces uncontrolled numerical reconnection .
we use a quasi - lagrangian simulation technique to run the first controlled experiments of cme initiation in the complete lack of reconnection .
we find that a flux rope confined by an arcade , when twisted beyond a critical amount , can escape to an open state , allowing some of the surrounding arcade to shrink and releasing magnetic energy from the global field .
this mechanism includes a true ideal mhd instability .
we conclude that reconnection is not a necessary trigger for fast cme eruptions . |
assume that @xmath0 is a domain , namely a connected open set , in @xmath1 , with @xmath2 , and let @xmath3 .
the negative sobolev norm of the distributional gradient of a function @xmath4 can be defined as @xmath5 here , @xmath6 stands for the divergence operator , and @xmath7 , the hlder conjugate of @xmath8 . moreover , in , and in similar occurrences throughout the paper , we tacitly assume that the sumpremum is extended over all functions @xmath9 which do not vanish identically . observe that the notation @xmath10 is consistent with the fact that the quantity on the right - hand side of agrees with the norm of @xmath11 , when regarded as an element of the dual of @xmath12 , where @xmath13 denotes the sobolev space of @xmath1-valued functions in @xmath0 with zero traces .
definition goes back to neas @xcite , who showed that , if @xmath0 is regular enough a bounded lipschitz domain , say and @xmath14 , then the @xmath15 norm of a function is equivalent to the @xmath16 norm of its gradient .
namely , there exist positive constants @xmath17 and @xmath18 , such that @xmath19 for every @xmath20 , where @xmath21 the mean value of @xmath22 over @xmath0 , and @xmath23 denotes the lebesgue measure of @xmath0 .
this result is known as neas negative norm theorem . in the present paper
we are concerned with a version of this theorem when a negative norm is introduced with the lebesgue space @xmath15 replaced with a general orlicz space . loosely speaking
, orlicz spaces extend lebesgue spaces in that the role of the power @xmath24 in their definition is played by a more general convex function ( a precise definition is recalled in the next section ) .
clearly , a full analogue of can not hold for arbitrary orlicz spaces , since fails , for instance , in the borderline cases when either @xmath25 , or @xmath26 .
our main result in this connection asserts that , however , an inequality in the spirit of still holds if , on the leftmost side , an orlicz norm appears which , in general , has to be slightly weaker than that on the rightmost side .
a precise balance between the relevant norms for a conclusion of this kind to hold is the content of theorem [ thm:3.2 ] , section [ main ] .
a key step in our approach is an analysis , of possible independent interest , of the divergence equation in orlicz spaces , via boundedness properties of the ( gradient of the ) operator in these spaces .
@xmath27 is introduced with the lebesgue space @xmath15 is replaced with a more general orlicz space @xmath28 associated with a young function @xmath29 , and , accordingly , the sobolev space @xmath13 replaced with the orlicz - sobolev space @xmath30 , associated with the conjugate function @xmath31 to @xmath29 .
a full analogue of can not hold , since fails , for instance , in the borderline cases when @xmath25 or @xmath26 .
our main motivation for a discussion of negative orlicz norms are applications to a mathematical model for non - newtonian fluids . in the stationary case ,
the relevant model tells us that the velocity field @xmath32 and the pressure @xmath33 of a fluid solve the following system of partial differential equations : @xmath34 see , for instance , @xcite . here
, @xmath35 is a positive constant , whose physical meaning is the density of the fluid , the operation @xmath36 denotes tensor product , the function @xmath37 describes the given volume forces , and the stress deviator @xmath38 is related to @xmath9 via a constitutive law . of course , the physically relevant dimensions are @xmath39 and @xmath40 . in a common model for fluids with non - newtonian behavior ,
the dependence of @xmath41 on @xmath9 is through a nonlinear function of the symmetric part @xmath42 of its @xmath43-valued gradient , defined as @xmath44 , where @xmath45 stands for transpose .
lebesgue and usual sobolev spaces provide an appropriate functional framework for the study of existence , uniqueness and regularity of solutions to when this nonlinear function is of power type . on the other hand ,
if nonlinearities of non - polynomial type are allowed , such as in the eyring - prandtl model @xcite , the more flexible orlicz and orlicz - sobolev spaces have to be called into play .
in particular , in section [ fluids ] we show how our negative orlicz norm theorem applies in the description of a suitable space for the pressure @xmath46 in .
a related result of use in a finite elements method for a numerical analysis of is also established .
many problems in pdes are related to distributions defined on sobolev - spaces .
for instance on @xmath47 , with @xmath48 bounded and @xmath49 , one considers the dual space @xmath50 .
it is isometrically isomorphic to the closure of @xmath15 w.r.t .
the norm @xmath51 the norm introduced above is called negative norm of a lebesgue function .
of course it is naturally defined also for its distributional gradient by @xmath52 neas @xcite showed the equivalence of the @xmath53-norm of a function and the negative norm of its gradient , i.e. @xmath54 for every @xmath55 with @xmath56 .
inequality ( [ eq:1.1 ] ) is known as neas negative norm theorem .
the main application he had in mind was korn s inequality : @xmath57 for every @xmath58 .
korn s inequality , which involves the symmetric gradient @xmath59 of a function @xmath32 , has important consequences in fluid mechanics and in the mechanics of solids ( see @xcite and @xcite ) .
due to the formula @xmath60 it follows easily by ( [ eq:1.1 ] ) .
+ a further application of the negative norm theorem is the pressure reconstruction in fluid mechanics . in the stationary case one
has to find the velocity field @xmath32 and the pressure @xmath33 of a fluid motion as solution of the following set of pdes : @xmath61 here @xmath37 is a given system of volume forces and ( in case of a generalized newtonian fluids ) one has a relation between the stress deviator @xmath62 and @xmath42 ( see section 4 for more details ) .
it is common to split the problem in two parts .
first one has to find a sobolev function @xmath9 such that @xmath63 holds in the sense of distributions on @xmath64 .
] by de rahms theorem one can find a distribution @xmath46 such that @xmath65 if @xmath66 for some @xmath67 we obtain from ( [ eq:1.1 ] ) that @xmath46 can be represented by a @xmath68-function and @xmath69 a generalization of ( [ eq:1.1 ] ) is shown in @xcite .
let @xmath70 be a young function with @xmath71 and @xmath72-condition ( see section 2 for definitions ) then a version of ( [ eq:1.1 ] ) holds on the orlicz space @xmath73 .
it is straightforward to extend the concepts of negative norms and boundary values to the fashion of orlicz - spaces in this situation since it is reflexive and smooth functions are dense .
the results follow from an extrapolation argument and ( [ eq:1.1 ] ) .
however this is only possible for `` nice '' young functions and the limit cases are not included .
those are not only interesting in order to complete the mathematical theory but also for the applications in fluid mechanics .
we will explain in detail in section 4 why the cases @xmath74 and @xmath75 are important for the study of the pressure in the prandtl - eyring fluid model invented in @xcite .
+ the aim of this paper is the extension of the negative norm concept to general orlicz spaces and relation between a young function and a its distributional gradient . without the @xmath71 and @xmath72-condition we can not expect that it can be described by meanings of the same orlicz space
this is also the case for singular integral operators and korn s inequality ( see @xcite and @xcite ) .
in fact we will give conditions for young functions @xmath29 and @xmath76 such that @xmath77 holds ( see theorem [ thm:3.2 ] ) .
in our proof we use the connection between the solution to the divergence - equation given by the -operator @xmath78 and negative norms .
hence in theorem [ thm:3.1 ] we first give conditions for young functions such that @xmath79 this result is also interesting by itself since the @xmath80 does not satisfy the conditions supposed on singular integral operators in @xcite .
+ our main application is connected to the problem of finding the pressure for stationary navier - stokes equations . given a function @xmath81 such that @xmath82 we give conditions for a young function @xmath76 such that there is @xmath83 with @xmath84
we also investigate the same problem on a discrete level using finite elements .
precisely we give conditions for an inf - sup condition in orlicz spaces which allows the reconstruction of the pressure .
in the present section we collect some definitions and fundamental results from the theory of orlicz and orlicz - sobolev spaces .
we refer to @xcite for a comprehensive treatment of this topic .
a function @xmath85 $ ] is called a young function if it is convex , left - continuous , and neither identically equal to @xmath86 , nor to @xmath87 .
thus , with any such function , it is uniquely associated a ( nontrivial ) non - decreasing left - continuous function @xmath88 $ ] such that @xmath89 the young conjugate @xmath90 of @xmath29 is the young function defined by @xmath91 note the representation formula @xmath92 where @xmath93 denotes the ( generalized ) left - continuous inverse of @xmath94 .
one has that @xmath95 moreover , @xmath96 for any young function @xmath29 .
then @xmath97 to see this , recall that @xmath98 thus , @xmath99 if @xmath29 is any young function and @xmath100 , then @xmath101 as a consequence , if @xmath100 , then @xmath102 where @xmath103 denotes the ( generalized ) right - continuous inverse of @xmath29 .
a young function @xmath29 is said to satisfy the @xmath71-condition if there exists a positive constant @xmath104 such that @xmath105 we say that @xmath29 satisfies the @xmath72-condition if there exists a constant @xmath106 such that @xmath107 for @xmath108 . if [ resp .
] just holds for @xmath109 for some @xmath110 , then @xmath29 is said to satisfy the @xmath71-condition [ @xmath72-condition ] near infinity .
we shall also write @xmath111 [ @xmath112 to denote that @xmath29 satisfies the @xmath71-condition [ @xmath72-condition ] .
one has that @xmath113 [ near infinity ] if and only if @xmath114-condition [ near infinity ] .
a young function @xmath29 is said to dominate another young function @xmath76 [ near infinity ] if there exists a positive constant @xmath104 @xmath115\,.}\ ] ] the functions @xmath29 and @xmath76 are called equivalent [ near infinity ] if they dominate each other [ near infinity ] .
let @xmath0 be a measurable subset of @xmath1 , and let @xmath29 be a young function .
the luxemburg norm , associated with @xmath29 , is defined as @xmath116 for any measurable function @xmath117 .
the collection of all functions @xmath22 for which such norm is finite is called the orlicz space @xmath28 , and is a banach function space .
the subspace of @xmath28 of those functions @xmath22 such that @xmath118 will be denoted by @xmath119 .
a hlder type inequality in orlicz spaces takes the form @xmath120 for every @xmath121 . if @xmath29 dominates @xmath76 , then @xmath122 with embedding norm depending on the constant @xmath104 appearing in .
when @xmath123 , embedding also holds if @xmath29 dominates @xmath76 just near infinity , but , in this case , the embedding constant also depends on @xmath76 , @xmath124 and @xmath23 .
the decreasing rearrangement @xmath125 $ ] of measurable function @xmath117 is the ( unique ) non - increasing , right - continuous function which is equimeasurable with @xmath22 .
thus , @xmath126 the equimeasurability of @xmath22 and @xmath127 implies that @xmath128 for every @xmath129 .
the lebesgue spaces @xmath15 , corresponding to the choice @xmath130 , if @xmath131 , and @xmath132 for @xmath133 $ ] and @xmath134 for @xmath135 , if @xmath26 , are a basic example of orlicz spaces
. other customary instances of orlicz spaces are provided by the zygmund spaces @xmath136 , and by the exponential spaces @xmath137 .
if either @xmath138 and @xmath139 , or @xmath25 and @xmath140 , then @xmath141 is the orlicz space associated with a young function equivalent to @xmath142 near infinity .
given @xmath143 , @xmath144 denotes the orlicz space built upon a young function equivalent to @xmath145 near infinity .
the orlicz space @xmath146 of @xmath1-valued measurable functions on @xmath0 is defined as @xmath147 , and is equipped with the norm given by @xmath148 for @xmath149 .
the orlicz space @xmath150 of @xmath151 matrix - valued measurable functions on @xmath0 is defined analogously .
assume now that @xmath0 is an open set .
the orlicz - sobolev space @xmath152 is the set of all weakly differentiable functions in @xmath28 whose gradient also belongs to @xmath28 .
it is a banach space endowed with the norm @xmath153 we also define the subspace of @xmath152 of those functions which vanish on @xmath154 as @xmath155 in the case when @xmath130 for some @xmath156 , and @xmath157 is regular enough , such definition of @xmath73 can be shown to reproduce the usual space @xmath47 defined as the closure in @xmath158 of the space @xmath159 of smooth compactly supported functions in @xmath0 . in general , the set of smooth bounded functions is dense in @xmath160 only if @xmath29 satisfies the @xmath161-condition ( just near infinity when @xmath162 ) , and hence , for arbitrary @xmath29 , our definition of @xmath73 yields a space which can be larger than the closure of @xmath163 in @xmath73 even for smooth domains . on the other hand , if @xmath0 is a lipschitz domain , then @xmath164 where @xmath165 is defined as usual . recall that an open set @xmath0 is called a lipschitz domain if it is bounded and there exists a neighborhood @xmath166 of each point of @xmath157 such that @xmath167 is the subgraph of a lipschitz continuous function of @xmath168 variables .
an open set @xmath0 is said to have the cone property if there exists a finite cone @xmath169 such that each point of @xmath0 is the vertex of a finite cone contained in @xmath0 and congruent to @xmath169 . clearly , any lipschitz domain has the cone property , but the converse is not true in general .
a poincar type inequality in orlicz - sobolev spaces tells us that , if @xmath0 is a lipschitz domain , then there exists a constant @xmath104 , depending on @xmath170 and on the lipschitz constant of @xmath0 , such that @xmath171 for every @xmath172 .
inequality is established in ( * ? ? ?
* lemma 4.1 ) in the special case when @xmath0 is a ball .
its proof makes use of a rearrangement type inequality for the norm @xmath173 which holds , in fact , for sobolev functions @xmath22 on any lipschitz domain @xmath0 ( * ? ? ?
* lemma 4.1 and inequality ( 3.5 ) ) .
the same proof then applies to any lipschitz domain , and one can verify that the constant in the resulting poincar inequality has the form claimed in .
the orlicz - sobolev space @xmath174 of @xmath1-valued functions is defined as @xmath175 , and equipped with the norm @xmath176 .
the space @xmath177 is defined accordingly .
let @xmath29 be a young function , and let @xmath0 be a bounded domain in @xmath1 .
we define the negative orlicz - sobolev norm associated with @xmath29 of the distributional gradient of a function @xmath178 as @xmath179 the alternative notation @xmath180 will also occasionally be employed to denote the negative orlicz - sobolev norm @xmath181 associated with the orlicz space @xmath28 .
our orlicz - sobolev space version of the negative norm theorem involves pairs of young functions @xmath29 and @xmath76 which obey the following balance conditions : @xmath182 and @xmath183 for some positive constant @xmath184 .
let us mention that assumptions and also come into play in a version of the korn inequality for the symmetric gradient in orlicz spaces @xcite .
[ thm:3.2 ] let @xmath29 and @xmath76 be young functions fulfilling and .
assume that @xmath0 is a bounded domain with the cone property in @xmath1 , @xmath2 .
then there exist constants @xmath185 and @xmath18 such that @xmath186 for every @xmath178 . here , @xmath184 denotes the constant appearing in and .
[ 3.2bis ] inequality continues to hold even if conditions and are just fulfilled for @xmath187 for some @xmath188 , but with constants @xmath189 and @xmath190 depending also on @xmath29 , @xmath76 , @xmath191 and @xmath23 .
indeed , the young functions @xmath29 and @xmath76 can be replaced , if necessary , with young functions equivalent near infinity , and fulfilling and for every @xmath192 .
owing , such replacement leaves the quantities @xmath193 , @xmath194 and @xmath195 unchanged , up to multiplicative constants depending on @xmath29 , @xmath76 , @xmath191 and @xmath23 .
as recalled in section [ intro ] , the standard negative - norm theorem expressed by breaks down in the borderline cases when @xmath25 or @xmath26 .
this shows that , in general , equation can not hold with @xmath196 on the left - hand side .
in fact , condition , or fails , with @xmath196 , if , loosely speaking , the norm @xmath197 is close " to @xmath198 , or to @xmath199 , respectively .
note that , if either or holds , then @xmath29 dominates @xmath76 globally ( * ? ? ?
* proposition 3.5 ) . in a sense , assumptions and provide us with a quantitative information on if , and how much , the norm @xmath200 has to be weaker than @xmath197 for a version of the negative - norm theorem to be restored in orlicz - sobokev spaces .
however , if @xmath111 , then certainly holds with @xmath196 ( * ? ? ?
* theorem 1.2.1 ) .
hence , in this case , assumption can be dropped in theorem [ thm:3.2 ] . on the other hand ,
if @xmath201 , then @xmath202 and hence holds with @xmath196 , and assumption can be dropped in theorem [ thm:3.2 ] .
in particular , if @xmath203 , then both conditions and are fulfilled with @xmath196 .
hence , we have the following corollary which also follows from the results of @xcite . [ cor:3.2 ]
assume that @xmath0 is a bounded domain with the cone property in @xmath1 , @xmath2 .
let @xmath29 be a young function in @xmath204 .
then there exist a constants @xmath205 and @xmath18 such that @xmath206 for every @xmath178 .
[ exlog ] assume that @xmath207 is a young function equivalent to @xmath208 near infinity , where either @xmath138 and @xmath209 , or @xmath25 and @xmath210 .
hence , if @xmath162 , then @xmath211 assume that @xmath0 is a bounded domain with the cone property in @xmath1 .
if @xmath138 , then @xmath212 , and hence corollary [ cor:3.2 ] tells us that @xmath213 for every @xmath178 .
however , if @xmath25 , then @xmath214 , but @xmath215 .
an application of theorem [ thm:3.2 ] now yields @xmath216 for every @xmath178 .
in particular , @xmath217 for every @xmath178 .
[ exexp ] let @xmath143 , and let @xmath207 be a young function equivalent to @xmath218 near infinity .
then @xmath219 if @xmath162 .
one has that @xmath220 , but @xmath221 .
theorem [ thm:3.2 ] ensures that , if @xmath0 is a bounded domain with the cone property in @xmath1 , then @xmath222 for every @xmath178 .
moreover , @xmath223 for every @xmath178 .
[ ex1 ] if @xmath138 and @xmath224 , then @xmath225 for every @xmath226 . if @xmath227 , then @xmath228 for every @xmath229 . [ exrn ]
let @xmath29 and @xmath76 be young functions such that @xmath230 where either @xmath231 and @xmath232 , or @xmath233 and @xmath234 , and either @xmath235 and @xmath224 , or @xmath25 and @xmath236 , and @xmath237 near @xmath86 , and @xmath238 near infinity .
then @xmath239 for every compactly supported function @xmath240 .
[ ex2 ] assume that @xmath138 and @xmath224 .
then @xmath241 for every @xmath242 .
if @xmath227 , then @xmath243 for every @xmath244 .
[ ex3 ] assume that @xmath143 .
then @xmath245 for every @xmath246 .
our proof of theorem [ thm:3.2 ] relies upon an analysis of the divergence equation @xmath247 in orlicz spaces .
this is the objective of the next result . in what follows ,
we set @xmath248 and @xmath249 [ thm:3.1 ] assume that @xmath0 is a bounded domain with the cone property in @xmath1 , @xmath2 .
let @xmath29 and @xmath76 be young functions fulfilling and .
then there exists a bounded linear operator @xmath250 such that @xmath251 and @xmath252 for every @xmath253 .
in particular , there exists a constant @xmath254 such that @xmath255 and @xmath256 for every @xmath253 . here , @xmath184 denotes the constant appearing in and .
the proof of theorem [ thm:3.1 ] in turn makes use of a rearrangement estimate , which extends those of ( * ? ? ?
* theorem 16.12 ) and @xcite , for a class of singular integral operators of the form @xmath257 for an integrable function @xmath258 . here
, the kernel @xmath259 fulfils the following properties : ( i ) @xmath260 ( ii ) @xmath261 ( iii ) for every @xmath262 , there exists a constant @xmath189 such that @xmath263 where @xmath264 denotes the unit sphere , centered at @xmath86 , in @xmath265 , and @xmath266 stands for the @xmath267-dimensional hausdorff measure ; \(iv ) there exists a constant @xmath190 such that @xmath268 and , if @xmath269 , then @xmath270 let @xmath0 be a bounded open set in @xmath1 , and let @xmath271 be a kernel satisfying . if @xmath272 and @xmath273 in @xmath274 , then the singular integral operator @xmath275 given by is well defined for a.e .
@xmath276 , and there exists a constant @xmath277 such that @xmath278 , the boundedness of singular integral operators given by between orlicz spaces associated with young functions @xmath29 and @xmath76 fulfilling and can be established . [ singularcor ]
let @xmath0 , @xmath279 and @xmath275 be as in theorem [ singular ] .
assume that @xmath29 and @xmath76 are young functions satisfying and .
then there exists a constant @xmath280 such that @xmath281 and @xmath282 for every @xmath283 .
here , @xmath184 denotes the constant appearing in and .
* proof*. by ( * ? ? ?
* lemma 1 ) , we have that , if @xmath29 and @xmath76 are young functions satisfying , then there exists a constant @xmath284 such that @xmath285 for every @xmath286 .
moreover , if @xmath29 and @xmath76 fulfill , then there exists a constant @xmath284 such that @xmath287 for every @xmath286 . combining , and , and making use of property yield inequality .
as far as is concerned , observe that , inequalities and continue to hold , with the same constant @xmath184 , if @xmath29 and @xmath76 are replaced with @xmath288 and @xmath289 , where @xmath290 is any positive constant .
thus , inequality continues to hold , with the same constant @xmath104 , after this replacement , whatever @xmath279 is , namely @xmath291 for every @xmath292 .
now , given any such @xmath293 , choose @xmath294 .
the very definition of luxemburg norm tells us that @xmath295 .
hence , by , @xmath296 .
the definition of luxemburg norm again implies that @xmath297 , namely .
* proof of theorem [ singular]*. let @xmath298 be such that @xmath299 , the ball centered at @xmath86 , with radius @xmath300 . fix a smooth function @xmath301 such that @xmath302 in @xmath303 $ ] and @xmath304 in @xmath305 . define @xmath306 by properties of @xmath271 ,
one has that : @xmath307 @xmath308 for every @xmath262 , there exists a constant @xmath309 such that @xmath310 where @xmath189 is the constant appearing in ; there exists a constant @xmath311 such that @xmath312 and , if @xmath276 , @xmath313 and @xmath269 , then @xmath314 @xmath315 where @xmath190 is the constant appearing in . define @xmath316 inequality will follow if we prove that @xmath317 for some constant @xmath318 , and for every @xmath272 such that @xmath319 in @xmath320
. a proof of inequality can be accomplished along the same lines as that of theorem 1 of @xcite , which in turn relies upon similar techniques as in @xcite . for completeness
, we sketch such proof hereafter .
the key step in the derivation of consists in showing that , for every @xmath321 , there exists a constant @xmath322 such that @xmath323 for every @xmath272 such that @xmath319 in @xmath324 .
fix @xmath325 , and define @xmath326 then , there exists an open set @xmath327 such that @xmath328 . by whitney s covering theorem ,
there exist a family of disjoint cubes @xmath329 such that @xmath330 , @xmath331 , and @xmath332 the operator @xmath333 is of weak type ( 1,1 ) , namely , there exists a constant @xmath334 such that @xmath335 for @xmath272 , as proved in ( * ? ?
* proof of lemma 6.3 ) .
we shall now show that there exists a constant @xmath336 such that @xmath337 fix any @xmath338 , choose @xmath339 such that @xmath340 and denote by @xmath341 the cube , centered at @xmath342 , with @xmath343 .
define @xmath344 so that @xmath345 .
if we prove that there exist constants @xmath346 and @xmath347 such that @xmath348 and @xmath349 then follows with @xmath350 . consider first .
let @xmath347 be such that @xmath351 .
since @xmath352 for @xmath353 , an application of with this choice of @xmath354 tells us that @xmath355 , @xmath356 indeed , since @xmath357 , we have that @xmath358 , and hence implies . we may thus focus on .
fix @xmath359 , and set @xmath360 . observe that @xmath361 .
given any @xmath353 , define @xmath362 .
one has that @xmath363 observe that , if @xmath364 , then @xmath365 and hence @xmath366 .
thus , owing to , @xmath367 moreover , @xmath368 .
therefore , there exists a constant @xmath369 such that @xmath370 on the other hand , @xmath371 where the first inequality holds since @xmath372 in @xmath373 if @xmath374 , and trivially holds ( with equality ) if @xmath375 . since @xmath376 in the last integral in , and @xmath293 vanishes in @xmath324 , by @xmath377 hence , @xmath378 for some constant @xmath379 .
note that , in the first inequality , we have made use of the inclusion @xmath380 , which holds since @xmath381 , and @xmath382 . combining inequalities
yields . inequality is fully established . via summation in @xmath383 , we obtain from that @xmath384 coupling with the inequality @xmath385 whence follows , by the very definition of decreasing rearrangement .
starting from inequality leads to , via the same iteration argument as in the proof of ( * ? ? ?
* theorem 1 ) .
[ galdi ] let @xmath0 be a bounded domain with the cone property in @xmath1 , with @xmath2 .
then there exist @xmath386 and a finite family @xmath387 of domains which are starshaped with respect to balls , such that @xmath388 .
moreover , given @xmath389 , there exist @xmath390 , @xmath391 , such that @xmath392 in @xmath393 , @xmath394 and @xmath395 for some constant @xmath396
. * proof , sketched*. any bounded open set with the cone property can be decomposed into a finite union of lipschitz domains ( * ? ? ? * lemma 4.22 ) . on the other hand
, any lipschitz domain can be decomposed into a finite union of open sets which are starshaped with respect to balls ( * ? ? ?
* lemma 3.4 , chapter 3 ) .
this proves the existence of the domains @xmath387 as in the statement .
the same argument as in the proof of ( * ? ? ?
* lemma 3.2 , chapter 3 ) then enables one to construct the desired family of functions @xmath397 on @xmath0 , @xmath398 , according to the following iteration scheme .
we set @xmath399 , @xmath400 , and , for @xmath401 , @xmath402 and @xmath403 observe that , since @xmath0 is connected , we can always relabel the sets @xmath404 in such a way that @xmath405 for @xmath406 . finally , we define @xmath407 the family @xmath408 satisfies the required properties .
the only nontrivial one is . to verify the latter , fix @xmath409 , and observe that , by , the second inequality in , inequality , and inequality @xmath410 on the other hand , by and a chain similar to
, one has that @xmath411 from , and an iteration of , one infers that @xmath412 and follows .
* proof of theorem [ thm:3.1 ] * by lemma [ galdi ] , it suffices to prove the statement in the case when @xmath0 is a domain starshaped with respect to a ball @xmath76 , which , without loss of generality , can be assumed to be centered at the origin and with radius @xmath413 . in this case
, we shall show that the ( gradient of the ) bogovskii operator @xmath78 , defined at a function @xmath414 as @xmath415 where @xmath416 is any ( nonnegative ) function in @xmath417 with @xmath418 , agrees with a singular integral operator , whose kernel fulfills
, plus two operators enjoying stronger boundedness properties . to be more precise , we set @xmath419 and claim that @xmath420 , and @xmath421 where @xmath422 is the linear operator defined at @xmath293 as @xmath423 for @xmath424 . here
, @xmath425 is the kernel of a singular integral operator satisfying the same assumptions as the kernel @xmath279 in theorem [ singular ] , and the kernels @xmath426 satisfy @xmath427 to verify this assertion , recall that , if @xmath428 , then @xmath429 , and moreover equations and hold for every @xmath430 ( * ? ? ? * proof of lemma iii.3.1 ) .
consider next the general case when @xmath431 .
owing to , @xmath432 , since @xmath433 grows at least linearly near infinity , and hence @xmath207 dominates the function @xmath434 near infinity .
since the space @xmath435 is dense in @xmath436 , there exists a sequence of functions @xmath437 such that @xmath438 in @xmath439 .
one has that @xmath440 ( in fact , @xmath78 is also bounded into @xmath441 )
. furthermore , @xmath442 as a consequence of and of a special case of theorem [ singularcor ] , with @xmath443 and @xmath444 .
thus , @xmath445 in @xmath446 and @xmath447 in @xmath448 .
this implies that @xmath449 , and and hold . by theorem [ singularcor ] , the singular integral operator defined by the first addend on the right - hand side of is bounded from @xmath450 into @xmath451 . by inequality ,
the operator defined by the second addend on the right - hand side of has ( at least ) the same boundedness properties as a riesz potential operator with kernel @xmath452 . such an operator is bounded in @xmath448 and in @xmath453 , with norms depending only on @xmath23 and on @xmath170 .
an interpolation theorem by calderon ( * ? ? ? * theorem 2.12 , chap .
3 ) then ensures that it is also bounded from @xmath28 into @xmath450 , and hence , a fortiori , from @xmath28 into @xmath451 , with norm depending on @xmath170 and @xmath23 .
finally , the operator given by the last addend on the right - hand side of is pointwise bounded ( in absolute value ) by @xmath454 .
thus , it is bounded from @xmath28 into @xmath450 , and hence from @xmath28 into @xmath451 .
equations and are thus established .
inequality can be derived from via a scaling argument analogous to that which leads to from see the proof of theorem [ singularcor ] .
we need a last preliminary result in preparation for the proof of theorem [ thm:3.2 ] .
[ supapprox ] let @xmath0 be an open subset in @xmath265 such that @xmath455 , and let @xmath29 be a young function .
assume that @xmath456 .
then : @xmath457 and @xmath458 note that equation is well known under the assumption that @xmath459 near infinity , namely @xmath460 near infinity , since @xmath461 is dense in @xmath462 in this case .
equation also easily follows from this property when @xmath463 near infinity .
the novelty of proposition [ supapprox ] is in the arbitrariness of @xmath29 .
* proof of proposition [ supapprox]*. consider first .
it clearly suffices to show that @xmath464 and @xmath465 given any @xmath466 , define , for @xmath338 , the function @xmath467 as @xmath468 clearly , @xmath469 , and @xmath470 a.e . in @xmath0 as @xmath471
. hence , @xmath472 by the monotone convergence theorem for integrals , and , by the fatou property of the luxemburg norm , @xmath473 thus , since @xmath474 equation follows . as far as is concerned , consider an increasing sequence of compact sets @xmath475 such that @xmath476 , @xmath477 for @xmath338 , and @xmath478 .
moreover , let @xmath479 be a family of ( nonnegative ) smooth mollifiers in @xmath265 , such that @xmath480 and @xmath481 for @xmath338 .
given @xmath482 , define @xmath483 as @xmath484 and @xmath485 as @xmath486 classical properties of mollifiers ensure that @xmath487 thus , if @xmath488 , then @xmath489 by the dominated convergence theorem for integrals . moreover , @xmath490 indeed , by dominated convergence and the definition of luxemburg norm , @xmath491 in particular , for every @xmath359 , there exists @xmath492 such that @xmath493 hence , by the arbitrariness of @xmath494 and the definition of luxemburg norm , @xmath495 we also have that @xmath496 indeed , assume that fails
. then , there exists @xmath497 and a subsequence of @xmath498 , still denoted by @xmath498 , such that @xmath499 a contradiction .
equation follows from and .
coupling with yields .
the proof of is complete .
the proof of follows along the same lines , and , in particular , via the equations @xmath500 and @xmath501 on defining , for any @xmath502 , the sequence of functions @xmath503 as @xmath504 for @xmath338 , where @xmath505 is given by , one can prove equation via a slight variant of the argument employed for . here
, one has to use the fact that @xmath506 as @xmath507 .
similarly , equation can be established similarly to on replacing , for any given @xmath508 , the sequence @xmath509 defined by with the sequence @xmath510 defined as @xmath511 where @xmath512 is any function in @xmath513 such that @xmath514 .
note that now , for every @xmath359 , there exists @xmath515 such that @xmath516 , provided that @xmath517 .
* proof of theorem [ thm:3.2]*. let @xmath518.then @xmath519 note that the inequality in holds since , by the first inequality in , @xmath520 and the last equality in relies upon
. by theorem [ thm:3.1 ] , applied with @xmath29 and @xmath76 replaced with @xmath521 and @xmath31 , respectively , there exists a constant @xmath522 such that @xmath523 the first inequality in follows from and .
the second inequality is trivial , since @xmath524 for some constant @xmath525 .
in many customary mathematical models , the stationary flow of a homogeneous incompressible fluid in a bounded domain @xmath526 is described by a system with the structure . with a slight abuse of notation with respect to
, we also denote by @xmath527 the function , acting on the symmetric gradient @xmath42 of the velocity field @xmath9 , which yields the stress deviator of the fluid .
thus , we shall consider systems of the form @xmath528 in the simplest case of a newtonian fluid , the function @xmath41 is linear , and @xmath529 , the laplacian of @xmath9 .
however only fluids with an easy molecular structure , such as water , oil , and several gases are governed by this low .
more complex liquids are not , and are called non - newtonian fluids see e.g. @xcite .
the most common nonlinear model among rheologists is the power law model , corresponding to the choice @xmath530 here , @xmath531 and @xmath532 are constants , and @xmath533 is an exponent which need to be specified via physical experiments .
an extensive list of specific @xmath8-values for different fluids can be found in @xcite . in various instances of interest in applications ,
the term @xmath536 is negligible in , compared with the other terms appearing in the first equation .
this is the case , for example , if the modulus of the velocity @xmath9 is small .
another situation where the role of the term is immaterial is that of plastic or pseudo - plastic fluids .
indeed , accounts for the inner rotation in the fluid flow , and for such fluids the impact of this term is very limited .
dropping the term reduces to the simplified system @xmath537 a standard approach to ( [ eq:4.1 ] ) or ( [ eq:4.1 ] ) consists in two steps .
firstly , a velocity field @xmath9 is exhibited such that @xmath538 for every @xmath539 , where either @xmath540 or @xmath541 according to weather the convective term @xmath542 is included in the model or not . here
, : " stands for scalar product between matrices , and @xmath543 denotes the space of compactly supported , infinitely differentiable @xmath265-valued functions whose divergence vanishes in @xmath0 .
the function @xmath9 belongs to a proper sobolev type space depending on the constitutive law underlying the definition of the function @xmath41 .
secondly , the pressure @xmath46 is reconstructed .
a discussion of the first issue falls beyond the scopes of the present paper , and will not be addressed here .
let us just mention that the standard power type model has been investigated in the classical contributions @xcite , and in the recent papers @xcite .
stationary flows of fluids whose constitutive law satisfies ( [ eq:4.4 ] ) with a young - function @xmath544 are studied in @xcite .
an unconventional constitutive law , where @xmath41 has the form with @xmath545 near infinity , and hence @xmath546 near infinity , was introduced by eyring in @xcite , where it is assumed that @xmath547 .
similar results are due to prandtl ( see e.g. @xcite for an overview on this kind of models ) .
an analysis of the simplified system for the eyring - prandtl model is the object of @xcite , whereas the complete system ( [ eq:4.1 ] ) , in the case @xmath39 , is considered in @xcite . in the remaining part of this paper
, we focus , instead , on the second question , namely the reconstruction of the pressure @xmath46 in a correct orlicz space . in case of fluids governed by a general constitutive low of the form
, the function @xmath548 belongs to some orlicz space @xmath549 . if @xmath550 , then @xmath551 as well .
however , in general , one can only expect that @xmath46 belongs to some larger orlicz space @xmath552 .
the balance between the young functions @xmath29 and @xmath76 is determined by conditions and , as stated in the following result .
a very general constitutive law for non - newtonian fluids is @xmath553 with a young function @xmath29 .
the natural space for the symmetric gradient of the velocity is @xmath28 so that ( [ ineqa1 ] ) implies @xmath554 .
using theorem [ thm:4.1 ] we can introduce the pressure in the correct orlicz space @xmath552 .
+ let us have a look at the most interesting cases : * if @xmath29 satisfies a @xmath71- and the @xmath72-condition we can choose @xmath555 . + in @xcite , @xcite and @xcite the model ( [ eq:4.4 ] ) is considered where @xmath556 satisfies the @xmath71- and the @xmath72-condition . note that this includes power law fluids where @xmath557 or @xmath558 with @xmath533 .
the existence of a weak solution @xmath559 ( together with regularity properties ) to ( [ eq:3.1weak ] ) with @xmath41 given by ( [ eq:4.4 ] ) is shown .
now we have @xmath560 and can reconstruct the pressure in the same space by theorem [ thm:4.1 ] . * if @xmath561 we can choose @xmath562 for all @xmath563 .
+ in the case of slow flows of prandtl - eyring fluids we have @xmath564 and hence @xmath565 .
* if @xmath566 we can choose @xmath567 for all @xmath568 . + considering the complete equation of motion for prandtl - eyring fluids in 2d we have @xmath569 and hence @xmath570 .
this was already shown in @xcite ( end of section 3 ) using continuity properties of the -operator in @xmath571-spaces .
[ thm:4.1 ] let @xmath29 and @xmath76 be young functions fulfilling and .
let @xmath0 be a bounded domain with the cone property in @xmath1 , @xmath2 .
assume that @xmath572 and satisfies @xmath573 for every @xmath539 .
then there exists a unique function @xmath574 such that @xmath575 for every @xmath576 .
moreover , there exists a constant @xmath254 such that @xmath577 and @xmath578 here , @xmath184 denotes the constant appearing in and . in particular , theorem [ thm:4.1 ] reproduces , within a unified framework , various results appearing in the literature .
for instance , when the constitutive relation is in force , the function @xmath207 is just a power @xmath579 , where the exponent @xmath580 , and depends on @xmath8 , on @xmath581 , and on whether the system or is taken into account . in any case
, @xmath582 agrees with the lebesgue space @xmath583 , and theorem [ thm:4.1 ] recovers the fact that @xmath46 belongs to the same lebesgue space @xmath584 .
as far as the simplified system for the eyring - prandtl model is concerned , under appropriate assumptions on @xmath581 one has that @xmath585 . hence , via theorem [ thm:4.1 ] , we infer the existence of a pressure @xmath586 .
more generally , if @xmath587 for some @xmath143 , one has that @xmath588 . the complete system for the eyring - prandtl model , in the @xmath589-dimensional case , admits a weak solution @xmath9 such that @xmath590 and hence @xmath591 @xcite .
again , one can not expect that the pressure @xmath46 belongs to the same space .
in fact , theorem [ thm:4.1 ] yields the existence of a pressure @xmath592 , thus reproducing a result from @xcite . in general ,
if @xmath593 for some @xmath210 , then we obtain that @xmath594 .
* proof of theorem [ thm:4.1]*. by de rahms theorem , in the version of @xcite , there exists a distribution @xmath595 such that @xmath596 for every @xmath597 . replacing @xmath598 with @xmath599 in , where @xmath600 , yields @xmath601 for every @xmath602 .
we claim that the linear functional @xmath603 is bounded on @xmath604 equipped with the @xmath453 norm . indeed , by ( [ 1.1 ] ) , one has that @xmath605 . moreover , by a special case of theorem [ thm:3.1 ] , @xmath606 .
thus , since @xmath607 and @xmath608 are orlicz spaces built upon young functions which are conjugate of each other , @xmath609 for every @xmath602 , where @xmath610 and @xmath611 .
hence , the relevant functional can be continued to a bounded linear functional on @xmath612 , with the same norm .
now , as a consequence of riesz s representation theorem , there exists a radon measure @xmath595 such that @xmath613 for every @xmath614 .
fix any open set @xmath615 . by theorem [ thm:3.1 ] again , there exists a constant @xmath104 such that @xmath616 one can verify that the norm @xmath617 is absolutely continuous , in the sense that for every @xmath359 there exists @xmath618 such that @xmath619 if @xmath620 , and since any lebesgue measurable set can be approximated from outside by open sets , inequality implies that the measure @xmath621 is absolutely continuous with respect to the lebesgue measure .
hence , @xmath621 has a density with respect to the lebesgue measure .
so @xmath595 can be represented by a function @xmath622 fulfilling ( [ eq : thm4.1 ] ) holds .
the function @xmath46 is uniquely determined if we assume that @xmath623 . by this assumption ,
theorem [ thm:3.2 ] , and equation ( [ eq : thm4.1 ] ) we have that @xmath624 where @xmath522 .
this proves inequality .
inequality follows from , on replacing @xmath29 and @xmath76 with @xmath288 and @xmath289 , respectively , with @xmath625 , via an argument analogous to that of the proof of .
let us turn to another constitutive relation where the situation is not that clear anymore : the prandtl - eyring fluid model .
it was invented by eyring @xcite who supposes @xmath626 with physical constants @xmath627 .
similar results are due to prandtl ( compare @xcite for an overview ) . fuchs and seregin
@xcite study the slow flow of prandtl - eyring fluids ( cf . system ( [ eq:4.1 ] ) ) .
the natural function space for weak solutions is @xmath628 note that the function @xmath629 is equivalent to @xmath207 .
the existence of a function @xmath630 satisfying @xmath631 with @xmath41 given by ( [ eq:4.3 ] ) can be shown by the direct method in the calculus of variations . under appropriate assumptions on @xmath581 we have that @xmath632 belongs to the space @xmath633 .
now it is not obvious at all which is the correct space for the pressure .
since we are in a limiting situation we do not expect that it belongs to the same space than @xmath548 .
if @xmath634 near infinity for some @xmath563 , we can choose @xmath635 near infinity in theorem [ thm:4.1 ] .
hence , we obtain by theorem [ thm:4.1 ] the existence of a pressure with @xmath586 .
+ in @xcite the full equation is considered in case @xmath39 .
the existence of a weak solution @xmath636 to ( [ eq:4.1 ] ) is shown and also that @xmath637 .
this means we are now in the opposite limit case and have @xmath638 .
again we do not expect that it is possible to obtain the pressure in the same space . if @xmath566 we can choose @xmath567 for every @xmath568 . considering the complete equation of motion for prandtl - eyring fluids we get by theorem [ thm:4.1 ] the existence of a pressure with @xmath592 .
this was already shown in @xcite ( end of section 3 ) using continuity properties of the -operator in @xmath571-spaces .
+ a very general constitutive law for non - newtonian fluids is @xmath553 with a young function @xmath29 .
the natural space for the symmetric gradient of the velocity is @xmath28 so that ( [ ineqa1 ] ) implies @xmath554 . using theorem [ thm:4.1 ]
we can introduce the pressure in the correct orlicz space @xmath552 .
+ let us have a look at the most interesting cases : * if @xmath29 satisfies a @xmath71- and the @xmath72-condition we can choose @xmath555 .
+ in @xcite , @xcite and @xcite the model ( [ eq:4.4 ] ) is considered where @xmath556 satisfies the @xmath71- and the @xmath72-condition . note that this includes power law fluids where @xmath557 or @xmath558 with @xmath533 .
the existence of a weak solution @xmath559 ( together with regularity properties ) to ( [ eq:3.1weak ] ) with @xmath41 given by ( [ eq:4.4 ] ) is shown .
now we have @xmath560 and can reconstruct the pressure in the same space by theorem [ thm:4.1 ] . * if @xmath561 we can choose @xmath562 for every @xmath563 .
+ in the case of slow flows of prandtl - eyring fluids we have @xmath564 and hence @xmath565 .
* if @xmath566 we can choose @xmath567 for every @xmath568 .
+ considering the complete equation of motion for prandtl - eyring fluids in 2d we have @xmath569 and hence @xmath570 .
this was already shown in @xcite ( end of section 3 ) using continuity properties of the -operator in @xmath571-spaces .
let us turn to a further consequence of theorem [ thm:4.1 ] , which is related to a numerical analysis of problem , or of its simplified version .
we shall adopt the scheme of the finite element method for the @xmath8-stokes system exploited in @xcite in the special case when @xmath41 is given by .
in what follows , we assume that @xmath0 is a polyhedron .
the goal is to compute an approximate solution to problem via discretization . to this purpose
, one needs a triangulation @xmath639 of @xmath0 into simplices of diameter bounded by @xmath640 .
recall that a simplex in @xmath265 is the convex hull of @xmath641 points which do not lie on the same hyperplane .
we also need that such a triangulation is regular enough for the lipschitz constant of the functions , which locally represent the boundaries of the relevant simplices , to be uniformly bounded in @xmath642 .
we denote by @xmath643 the space of those functions in @xmath0 whose restriction to each simplex of @xmath639 is constant , and by @xmath644 its subspace of those functions from @xmath645 whose mean value over @xmath0 is zero . also , for @xmath646 , we denote by @xmath647 the space of first - order weakly differentiable functions in @xmath0 whose restriction to each simplex of @xmath639 is a polynomial of degree @xmath648 , and by @xmath649 its subspace of those functions from @xmath647 which vanish on @xmath154 .
finally @xmath650 stands for the space of all functions from @xmath647 whose mean value over @xmath0 is zero . clearly , given any young function @xmath29
, one has that @xmath651 and @xmath652 the spaces @xmath653 , @xmath654 and @xmath655 of @xmath265-valued functions are defined accordingly .
an important tool for the numerical analysis of problem is a projection operator @xmath656 which , for given @xmath657 and @xmath658 , is such that @xmath659 , and enjoys the following properties : + ( i ) @xmath656 preserves zero boundary values , i.e. @xmath660 ( ii ) @xmath656 is divergence preserving , namely , @xmath661 ( iii ) @xmath656 is continuous in the @xmath662-sense , i.e. there exists a constant @xmath396 such that @xmath663 for every @xmath664 , and every @xmath665 . here
@xmath666 denotes the diameter of @xmath667 , and @xmath668 the union of @xmath667 and all its direct neighbors in the triangulation @xmath639 .
the existence of such an operator for suitable couples of @xmath669 and @xmath290 is standard
see e.g. ( * ? ? ?
* appendix ) .
for instance , the choice @xmath670 and @xmath671 is admissible , whereas @xmath672 and @xmath671 is not .
this can again be accomplished in two steps .
first , on setting @xmath677 and @xmath678 one has to find @xmath679 such that @xmath680 the existence of a unique function @xmath681 satisfying follows easily from the theory of monotone operators . if @xmath41 has variational structure one may equivalently solve the corresponding strictly convex minimizing problem on the finite dimensional function space @xmath682 .
next , the pressure has be to reconstructed .
precisely , one has to find @xmath683 such that @xmath684 this is the discrete analogone of the problem from theorem [ thm:4.1 ] .
as far as existence is concerned , we only need to solve an algebraic linear system .
the following questions then arise : the answer to all this questions follows from the so - called inf - sup condition .
such a condition , in the standard case when @xmath41 has a power type growth as in , reads as @xmath688 for some positive constant @xmath104 independent of @xmath642 .
in fact , the uniqueness of @xmath685 does not even require @xmath104 to be independent of @xmath642 .
[ thm:4.2 ] let @xmath29 and @xmath76 be young functions fulfilling and .
assume that @xmath0 is a polyhedron in @xmath1 , @xmath2 , and that @xmath657 and @xmath658 are such that there exists an operator @xmath689 satisfying ( [ itm : a1 - 2])-([itm : a2 ] ) .
then , @xmath690 for some positive constant @xmath254 , where @xmath184 is the constant appearing in and * proof*. as a first step , we show that the operator @xmath656 is continuous in every orlicz space , in the sense that there exists a constant @xmath396 such that , for every young function @xmath29 , @xmath691 inequality has been established in ( * ? ? ?
4.5 ) and in ( * ? ? ?
3.2 ) under the additional assumption that @xmath29 fulfils a global @xmath71-condition .
a variant of those proofs shows that , in fact , this assumption can be dropped .
we outline the argument hereafter .
since @xmath692 belongs to a finite dimensional function space , it follows from ( [ itm : a2 ] ) that @xmath693 for some constants @xmath396 and @xmath694 , and for every @xmath695 .
hence , owing to jensen s inequality and the convexity of @xmath29 , @xmath696 for every @xmath697 .
now , given any @xmath698 , we deduce from the convexity of @xmath29 , and latexmath:[\[\begin{aligned } \dashint_{\mathcal s}a(h_\mathcal s|\nabla\bfu-\nabla \pi^h\bfu|)\,dx & \leq \frac12\dashint_{\mathcal s}a(2h_\mathcal s|\nabla(\bfu-\bfq ) |)\,dx+ \frac 12\dashint_{\mathcal s}a(2h_\mathcal s|\nabla \pi^h(\bfu-\bfq)|)\,dx\nonumber\\ & \leq \dashint_{m_{\mathcal s}}a(ch_\mathcal s|\nabla(\bfu-\bfq ) for some constant @xmath396 and for every @xmath700 .
finally we choose @xmath701 . by inequality and a scaling argument as in the proof of theorem [ thm:3.1 ]
one can show that there exists a constant @xmath702 such that @xmath703 for every @xmath697 .
inequalities and imply that @xmath704 for every @xmath697 .
the convexity of @xmath29 and inequality ( [ eq : appr ] ) yield @xmath705 for some constant @xmath396 , and for every @xmath700 .
this implies , on replacing @xmath706 with @xmath707 and summing up over all simplexes @xmath667 .
note , in this connection , that the number of neighbours of each @xmath667 only depends on @xmath170 .
we are now in a position to prove inequality . by theorem [ thm:3.2 ]
, there exists a constant @xmath708 such that @xmath709 for every @xmath710 . owing to and to properties and of the operator @xmath656 , @xmath711 for some constants @xmath522 and @xmath712 , and for every @xmath713 .
hence , follows . since the functions @xmath714 belong to finite dimensional spaces it would be possible to replace @xmath715 in theorem [ thm:4.2 ] by @xmath716 .
however , the constant then depends on the dimension of @xmath717 and hence on @xmath642 . in that form ,
the result would be of no use in the study of the convergence of the finite element method approximation .
* there exists a unique function @xmath683 such that @xmath721 for every @xmath722 .
* there exists a constant @xmath104 , independent of @xmath642 , such that @xmath723 * there exists a constant @xmath104 , independent of @xmath642 , such that , if @xmath724 and @xmath725 satisfy @xmath726 for every @xmath727,then @xmath728 * proof*. ( i ) consider a basis @xmath729 , @xmath730 of @xmath731 and a basis @xmath732 , @xmath733 of @xmath734 .
note that @xmath735 .
then the problem @xmath736 for every @xmath722 is equivalent to the algebraic linear system @xmath737 , where @xmath738 such system has a solution , provided that @xmath739 belongs to the image of the matrix @xmath740 , or , equivalently , if it is orthogonal to @xmath741 .
now , observe that a vector @xmath742 belongs to @xmath741 if and only if @xmath743 whence , @xmath744 .
thus , owing to ( [ cor:4.1.0 ] ) , @xmath745 for every @xmath746 .
hence , @xmath747 .
next , we show that @xmath740 is injective , whence the uniqueness of @xmath685 follows . to verify the injectivity of @xmath740 ,
assume that @xmath748 is such that @xmath749 .
hence , on setting @xmath750 , we have that @xmath751 for every @xmath409 .
let us apply theorem [ thm:4.2 ] with any pair of young functions @xmath29 and @xmath76 fulfilling and , for instance @xmath752 for @xmath753 (
this means that we are in fact applying with @xmath754 ) .
we then obtain that @xmath755 whence @xmath756 .
+ ( ii ) by theorem [ thm:4.2 ] and ( [ b.5 ] ) @xmath757 for some constants @xmath254 and @xmath712 .
this proves inequality .
+ ( iii ) the triangle inequality and the embedding @xmath758 ensure that , for every @xmath759 , @xmath760 for some constant @xmath284 . by theorem [ thm:4.2 ] , @xmath761 for some constant @xmath762 , and for every @xmath763 .
moreover , by , , and an approximation argument for functions in @xmath764 via functions in @xmath765 , we have that @xmath766 for every @xmath759 and @xmath767 .
hence , via hlder s inequality in orlicz spaces , @xmath768 for some constant @xmath525 .
altogether , we conclude that @xmath769 for some constant @xmath104 independent of @xmath642 , namely . this research was partly supported by leopoldina ( german national academy of science ) , by the research project
geometric properties of partial differential equations and related topics " 2008 of the italian miur ( ministry of education , university and research ) , and by gnampa of the italian indam ( istituto nazionale di alta matematica ) .
m. e. bogovski , solutions of some problems of vector analysis , associated with the operators @xmath771 and @xmath772 . in _ theory of cubature formulas and the application of functional analysis to problems of mathematical physics ( russian ) _ ,
pages 540 , * 149 * , akad .
nauk sssr sibirsk .
inst . mat . ,
novosibirsk , 1980 .
j. frehse , j. malek & m. steinhauer , an existence result for fluids with shear dependent viscosit
steady flows , _ nonlinear anal .
_ * 30 * ( 1997 ) 30413049 .
j. frehse , j. malk & m. steinhauer , on analysis of steady flows of fluids with shear - dependent viscosity based on the lipschitz truncation method , _ siam j. math .
* 34 * ( 2003 ) 10641083 .
o. a. ladyzhenskaya , _ the mathematical theory of viscous incompressible flow _ ,
gorden and breach , 1969 . o. a. ladyzhenskaya , on some new equations describing dynamics of incompressible fluids and on global solvability of boundary value problems to these equations , _ trudy steklov s math . institute _
* 102 * ( 1967 ) 85104 .
o. a. ladyzhenskaya , on some modifications of the navier - stokes equations for large gradients of velocity , _ zap .
nauchn . sem .
otdel . mat .
inst . steklov ( lomi ) _ * 7 * ( 1968 ) 126154 .
j. neas , sur les normes quivalentes dans @xmath778 et sur la coecivit des formes formellement positives .
sminaire des mathmatiques suprieures , in : _ equations aus drives partielles _ * 19 * , 102128 , les presses de luniversit de montral , montral ( 1965 ) | we prove a version of the negative norm theorem in orlicz - sobolev spaces .
a study of continuity properties of the -operator between orlicz spaces is a crucial step , of independent interest , in our approach .
applications to the problem of pressure reconstruction for non - newtonian fluids governed by constitutive laws , which are not necessarily of power type , are presented .
a key inequality for a numerical analysis of the underlying elliptic system is also derived . |
meteorites are a partial sample of asteroids that survive the passage through the earth s atmosphere .
the identification of the source regions of the different type of meteorites is essential to be able to link the mineralogical properties of meteorites to the parent asteroids and , consequently , to address the mineralogical evolution in the asteroid belt .
however , this is not an easy task and only some of these links could be established : for instance , the group of howardites , eucrites , and diogenites ( heds ) meteorites are thought to come from the vesta family of asteroids ( e.g. * ? ? ?
* ) ; more speculatively , l ordinary chondrites could come from the gefion family @xcite , while asteroids of the the flora family bear spectral similarities with the ll chondrites @xcite .
however , the parent bodies of most meteorite types , if still intact , are unknown .
the discovery and spectroscopic observation of the near - earth asteroid ( nea ) 2008 tc@xmath0 ( henceforth tc@xmath0 ) 20 hours before it impacted the earth s high atmosphere , and the subsequent recovery of meteorites ( called almahata sitta ) clearly coming from this body was a major result in this respect @xcite .
it allowed a direct link between an asteroid and meteorites to be established for the first time : the asteroid was classified as belonging to the spectroscopic f - class ( in the tholen classification , * ? ? ? * ) or b - class ( in the bus classification , * ? ? ? * ) on the basis of the flat shape of its reflectance spectrum in the region between 500 and 1000 nm .
moreover , among the 47 meteorites initially recovered , it was observed that the visible spectrum of the fragment @xmath3 matches the telescopic spectrum of tc@xmath0 obtained before the impact with the earth s atmosphere @xcite .
fragment @xmath3 is an achondrite polymict ureilite @xcite .
ureilites are in a group of achondritic meteorits that exhibit both primitive and evolved characteristics @xcite : in particular , they are characterized by olivine and pyroxene - rich clasts among carboneous material ( mainly graphite ) ; fine - grained graphite is also present , which lower the albedo of the meteorites ( about 10 - 12% ; * ? ? ?
ureilites were initially thought to derive from @xmath4-class asteroids ( see for instance @xcite ) .
however , because of its spectral properties , @xcite propose a link with @xmath5-class asteroids according to the bus classification or @xmath6-class in the tholen classification .
this is more plausible than a link with @xmath4-class asteroids , given the low albedos of ureilite meteorites , consistent with @xmath5 or @xmath6-class asteroids but not with the @xmath4-class .
it is worth to note that the f - class can be distinguished from the b - class from a much weaker uv drop - off in the spectra of the former compared to the latter and also because b - class asteroids show a moderately higher average albedo than f - class bodies .
however , in the bus classification these two classes are merged in a unique class ( the b - class ) .
this is because the @xcite spectral classification is based on spectra acquired with ccd spectrographs , which in general do not observe far enough in the uv to observe the above - mentioned drop - off feature .
we will refer in this work to b - class asteroids including both tholen b- and f - classes , and bus b - class .
interestingly , mineralogical studies of almahata sitta show that , among the @xmath7 fragments recovered @xcite , about @xmath8 are ureilites , while the remaining @xmath9 are enstatite chondrites , h and l ordinary chondrites .
more specifically , @xcite show that , among a subsample of 40 deeply studied meteorites from almahata sitta , 23 fragments are achondritic ureilites and 17 have chondritic litologies with 14 of them corresponding to enstatite chondrites , 2 to h ordinary chondrites and 1 to a new type of chondrite ( see @xcite for more details ) .
although small clasts of different types are quite common in brecciated meteorites ( see * ? ? ?
* for a review ) and carboneous material is found in some he d meteorites , it is the first time that meteorites of very different mineralogies ( i.e primitive fragments with achondrite polymict ureilites and evolved ones such as ordinary chondrites or enstatite chondrites ) are associated to the same fall .
this led to make the hypothesis that tc@xmath0 was an asteroid made of blocks of different mineralogies , held together very loosely ( given the explosion of the body at the anomalously high altitude ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
tracing back tc@xmath0 to its source region in the asteroid main belt would allow us to understand the origin of the almahata sitta meteorites and how tc@xmath0 was put together by loosely assembling meteorites of different mineralogies .
establishing this link would also be fundamental to shed light on the source region of ureilites , that albeit rare , are the forth major class of primitive meteorites recovered on earth after the cv , ci and co carbonaceous chondrites . in their attempt to find the source region of tc@xmath0 and almahata sitta , @xcite selected all @xmath5-class asteroids , according to the bus classification , and objects of the tholen f and b classes and searched for spectra similarities with tc@xmath0 and almahata sitta . as a result of their study
, they showed spectral similarities between tc@xmath0 and ungrouped asteroids as well as several dynamical asteroid groups ( or families ) as possible origins for the tc@xmath0 asteroid , namely polana ( @xmath10 au , @xmath11 ) , hoffmeister ( @xmath12 au , @xmath13 ) , pallas ( @xmath12 au , @xmath14 ) , themis ( @xmath15 au , @xmath16 ) , and theobalda ( @xmath17 au , @xmath18 ) .
later , from dynamical grounds , @xcite concluded that asteroids from the inner asteroid belt ( i.e. with orbital semimajor axis @xmath19au ) are the likely parent bodies of tc@xmath0 .
this reduces the choice to dispersed b - class asteroids in the inner main belt and the polana asteroid group . in section [ section_dynamics ] , we revisit this issue studying the possible dynamical source regions for tc@xmath0 .
we recall here that the polana group is part of a cluster of asteroids known as the nysa - polana family @xcite , which is located in the inner main belt , between the @xmath20 secular resonance ( at heliocentric distance @xmath21 2.1 au ) and the 3:1 mean motion resonance with jupiter ( at heliocentric distance of 2.5 au ) .
this family has a complex twofold
structure in orbital proper element space @xcite , suggesting that it is the outcome of at least two independent break - up events in the same orbital region . from the few spectral data available at the time
, @xcite argued that the nysa - polana family contains asteroids of three spectral classes .
the first class is that of b - class objects , like asteroid ( 142 ) polana itself
note that @xcite uses the f - class classification from the @xcite taxonomy ; the second class is the s - class , with the largest member being identified as the asteroid ( 878 ) mildred ; the third class is that of x - class objects , such as the asteroid ( 44 ) nysa . in this manuscript
, we revisit this result using a much wider dataset of spectro - photometric data provided by the moving objects catalog of the sloan digital survey ( sdss , * ? ? ?
* ) , which is analyzed here using a new classification algorithm ( * ? ? ?
* described in section [ section_method ] ) developed for the gaia space mission of the european space agency .
a detailed study of the mineralogy of the nysa - polana family is of great importance also for better understanding the origin of other neos . in partcular
, @xcite claimed that the asteroid ( 101955 ) 1999 rq@xmath22 , target of the sample return mission osiris - rex ( approved in the program new frontiers of nasa ) , was delivered to near - earth space via the @xmath20 secular resonance from the polana group .
moreover , the binary asteroid ( 175706 ) 1996 fg@xmath0 , primary target of the sample return mission marco polo - r , under study at the european space mission ( esa ) , might have formed within the polana group and delivered to the near - earth space via the overlapping jupiter 7:2 and mars 5:9 mean motion resonances rather than the @xmath20 ( see * ? ? ?
* ) . as a consequence , in sections [ section_polana ] and [ background ]
, we perform a spectroscopic analysis using the sdss data of the asteroids of the nysa - polana family as well as dispersed asteroids of the inner main belt ( called objects of the background ) in order to find spectral matches with tc@xmath0 and almahata sitta .
finally , in section [ section_formation ] , we investigate a possible formation scenario for the tc@xmath0 asteroid as a rubble pile of rocks of different mineralogical types , which is based on the peculiar low inclination of the nysa - polana family and dispersed asteroids .
the large majority of neas are fragments generated by the collisional disruption of larger asteroids of the main belt ; said fragments drift in orbital semimajor axis by the so - called yarkovsky effect until they reach regions of orbital instabilities ( mean motion resonances with jupiter and secular resonances ) which , by enhancing their orbital eccentricities , deliver them to the near - earth space ( see * ? ? ?
* for a review ) .
according to model by @xcite , there are five main intermediate sources of neas : * the @xmath23 secular resonance , which marks the inner edge of the asteroid belt and occurs when the precession frequency of the longitude of perihelion of an asteroid is equal to that of saturn ; * the 3:1 mean motion resonance located at @xmath24au , where the orbital period of an asteroid is 1/3 of that of jupiter ; * mars - crossing asteroids , defined as objects which are not neas ( i.e. their perihelion distance @xmath25 is larger than 1.3 au ) , but whose semimajor axis evolves in a random - walk fashion as a result of close and distant encounters with mars . * the outer belt population , whose eccentricities can increase up to planet - crossing values due to a network of high order orbital resonances with jupiter and three - body resonance of type asteroid - jupiter - saturn @xcite * dormant jupiter family comets .
the orbital elements of tc@xmath0 before impact are not known very precisely . here
are different estimates : @xmath26 au , @xmath27 , and @xmath28 ( neodys website : http://newton.dm.unipi.it/neodys ) ; @xmath29 au , @xmath30 , and @xmath31 @xcite . for each of these orbits we computed , using the original orbital evolution files of @xcite , which are the most likely intermediate sources among those enumerated above .
we found that tc@xmath0 has a probability of 63 - 66% to come from the @xmath23 source and 34 - 37% to come from the mars crosser population .
apparently , none of the simulations for the 3/1 resonance in @xcite produce objects within @xmath32 au in @xmath33 , @xmath32 in @xmath34 and @xmath35 in @xmath36 of tc@xmath0 .
this contrasts with the claim reported in @xcite , that tc@xmath0has a 20% probability to come from the 3/1 resonance in the @xcite model .
we suspect that there has been an error in the manipulation of the @xcite model in that work . of the mars crossers that can produce the tc@xmath0 orbit in @xcite simulations , none has semimajor axis larger than 2.5 au . from this , and remembering that the @xmath23 resonance lays at the inner edge of the belt , we conclude - in agreement with @xcite - that tc@xmath0 most likely comes from the inner asteroid belt , inside of 2.5 au .
as noted in @xcite , there are two populations of asteroids in the inner belt with spectra broadly consistent with that of tc@xmath0 : the polana members , with an orbital inclination of 23 degrees , and a population of dispersed b - class asteroids , with inclinations ranging up to 15 degrees .
these objects are too dispersed to belong to a relatively young collisional family , but their broad cluster in orbital space suggests that they might belong to an old collisional family , dynamically dispersed probably during the phase when the orbits of the giant planets changed substantially , about 4 gyr ago . to determine whether it is more likely that tc@xmath0 comes from the nysa - polana family or from the population of dispersed b - class asteroids , we turn , once again , to the original simulations in @xcite . in that work ,
the initial distribution of asteroids in the @xmath23 was uniform in inclination .
in particular , 50% of the initial conditions had initial inclination larger than 8 degrees . however
, 93% of the particles which reproduced the orbit of tc@xmath0 at some time during their evolution , have initial inclination @xmath2 . for the mars crossers , all those reproducing the orbit of tc@xmath0 have initial @xmath2 .
this suggests that the most likely source of tc@xmath0 is at low inclination , consistent with the nysa - polana family and with the dispersed b - class asteroids with @xmath2 .
this is not surprising , given that the inclination of tc@xmath0 is 2.32.5 degrees .
however , the dynamics rules out a tc@xmath0 origin from the dispersed b - class asteroids at higher inclination ( i.e. i@xmath37 ) .
finally , @xcite have shown that the polana group can easily deliver small enough fragments in the @xmath20 resonance .
in fact , the extrapolated yarkovsky - induced semimajor axis distribution of polana group members predicts that asteroids fainter than @xmath38 18.5 can reach the border of the @xmath20 resonance , which is at 2.15 au for a polana - like inclination . assuming a polana - like geometric visible albedo of 0.05
, @xmath39 18.5 translates into a diameter @xmath40 2 km , which is much larger then the size of tc@xmath0 .
for all these reasons we conclude that , from the dynamical viewpoint , the nysa - polana family and dispersed b - class asteroids with @xmath2 are the most likely sources of tc@xmath0 .
thus , we will focus on the nysa - polana family and the background at low inclination in the next sections .
in order to search for spectral groups in the nysa - polana asteroid family or in the inner main belt background at low inclination , we used the unsupervised classification algorithm that will be adopted for the classification of asteroid spectra from the multiband photometers @xcite on board the gaia space mission of the european space agency ( esa ) .
while gaia is mainly devoted to the observations of @xmath41 stars , it is expected that this mission will also observe more than 250,000 asteroids multiple times over 5 years ( mission lifetime ) @xcite .
spectral classification of asteroids will be performed by an unsupervised classification algorithm based on the works of @xcite and @xcite .
an unsupervised approach has the advantage that no _ a priori _ information is taken into account to build the spectral groups .
the algorithm is based on a method for partitioning a set @xmath42 of @xmath43 data points ( @xmath44 ) where @xmath45 in our case is the number of spectral bands into @xmath46 non - overlapping clusters ( or groups of data points ) such that ( a ) the inter - cluster variance is maximized and ( b ) the intra - cluster variance is minimized .
each spectrum is considered as a vertex of tree .
a tree is a graph that is connected ( i.e. every vertex is connected to at least another vertex ) and acyclic ( i.e. the graph has no loops ) .
the spanning tree ( mst , i.e. the tree passing through each vertex of the set ) with the minimal length is calculated using the so - called prim s algorithm ( see * ? ? ? * and references therein ) .
the length of each edge of the tree ( i.e. the distance between two spectra ) is determined using the kullback - leibler metric : let @xmath47 the feature vector corresponding to a reflectance spectrum ( @xmath45 being the number of wavelengths ) ; at a given wavelength @xmath48 , each spectrum is associated with a ( positive ) normalized quantity : @xmath49 , which can be interpreted as the probability distribution that a certain amount of information has been measured around the wavelength @xmath50 . to measure the similarity between two probability density functions , we compute the symmetrized kullback - leibler divergence : @xmath51 the identification of the cluster
is performed by first computing , at each step of the mst construction , the length of the newly connected edge and then by identifying valleys in the curve obtained by plotting the mst edge length as a function of the iteration of the construction ( see figure [ fig : mst ] ) .
the valleys in this curve identify the number and the position of high density region of points , i.e. the clusters ( see * ? ? ? * for a more thorough description of the algorithm ) . in order to increase the sample statistics compared to the relatively small number of asteroids observed by @xcite and identify meaningful spectral groups within the nysa - polana family , we used our spectral classification algorithm on the observations contained in the sdss moc4 ( * ? ? ? * http://www.sdss.org ) .
the moc4 contains magnitudes of 104449 main belt asteroids .
each asteroid was observed , in general , at multiple epochs over 5 spectral bands in the visible light , namely @xmath52 , @xmath53 , @xmath54 , @xmath55 , @xmath56 at the following central wavelengths ( @xmath50 ) of @xmath57 , @xmath58 , @xmath59 , @xmath60 , @xmath61 nm , respectively . to obtain the surface reflectance of the asteroids , we needed to remove the solar contribution ( @xmath62 , @xmath63 , @xmath64 , @xmath65 , @xmath66 ) to the observed magnitudes .
the solar contribution is calculated from transformation equations between the sdss ugriz magnitudes and the usual @xmath67 system .
we find : @xmath68 , @xmath69 , @xmath70 , @xmath71 , @xmath72 .
( see the sdss website for more details : http://www.sdss.org ) . in order to compare asteroid spectral reflectances and be able to classify them , we choosed to normalize each spectrum relatively to the @xmath54 band ( @xmath73 nm )
. finally , we obtained the following asteroid reflectance for the @xmath74 wavelength : @xmath75 @xmath76 the reflectance colors at @xmath74 and @xmath77 respectively .
the same computations were performed for the @xmath78 , @xmath79 , and @xmath80 wavelengths .
we searched in the moc4 database for all asteroids belonging to the nysa - polana family and for dispersed objects .
concerning the family members identification , we used the definition of dynamical families of @xcite . among the nysa - polana family ,
4134 objects have been observed at least once by sdss ( as a comparison , the number of spectra obtained in the visible light or near - infra red , for the nysa - polana family , in the smass , ecas and 52-color catalogs are : 15 , 13 , and 2 , respectively ) . for those asteroids with more than one observation , we calculated the weighted mean reflectance by averaging the reflectancies derived at each epoch weighing each point with its respective signal - to - noise ratio .
however , likely due to non - photometric conditions some of the observed objects have high uncertainties in the measured magnitudes . in the visible light
, the separation of asteroid spectroscopic group is based on the overall slope of their reflectance and in the presence ( or absence ) and strength of the 1 absorption feature . in order to avoid the superposition of spectral groups we selected asteroid with the less noisy observations .
namely , we rejected observations with a relative uncertainty @xmath81 on the in - band photometric flux derived from the moc4 magnitudes . in the end
, the taxonomic classification of the nysa - polana dynamical family was performed over @xmath82 objects and over @xmath83 objects for the background at low inclination . -class spectrum , group 2 with @xmath5-class spectrum and group 3 with @xmath84-subclasses .
we also find a good agreement with several almahata sitta fragments .
spectral differences with fragment @xmath85 ( h - chondrite ) appearing in panel ( a ) are likely due to space weathering . fragments @xmath86 and @xmath3 in panel ( b ) are ureilites.,title="fig : " ] + -class spectrum , group 2 with @xmath5-class spectrum and group 3 with @xmath84-subclasses . we also find a good agreement with several almahata sitta fragments .
spectral differences with fragment @xmath85 ( h - chondrite ) appearing in panel ( a ) are likely due to space weathering
. fragments @xmath86 and @xmath3 in panel ( b ) are ureilites.,title="fig : " ] + -class spectrum , group 2 with @xmath5-class spectrum and group 3 with @xmath84-subclasses .
we also find a good agreement with several almahata sitta fragments .
spectral differences with fragment @xmath85 ( h - chondrite ) appearing in panel ( a ) are likely due to space weathering .
fragments @xmath86 and @xmath3 in panel ( b ) are ureilites.,title="fig : " ] ) , b - class @xmath87 ) , and x - class ( @xmath88 ) objects of the nysa - polana family as a function of @xmath89 and @xmath90 ( calculated from sdss magnitudes ) .
ellipses represent the location of s , b , and x regions .
the color palette is the same of @xcite .
although polana ( b ) , nysa ( x ) , and herta ( x ) have not been observed by sdss , we have calculated their @xmath89 and @xmath90 values from the ecas catalog and located them in this plot . ]
figure [ fig : mst ] shows the mst length as a function of the number of iterations ( i.e. addition of spectra ) in the case of the sdss - moc4 spectrophotometric data of the 579 asteroids of the nysa - polana family .
the figure clearly shows three main valleys corresponding to three groups composed of 118 , 13 , and 378 asteroids , respectively .
the remaining objects ( @xmath91 ) are considered as unclassified asteroids .
this is due to their large value of the mst length which means no spectral similarity among the family is found for these @xmath91 objects .
figure [ fig : spectra ] shows asteroid spectra of each group and their average spectrum .
error bars are the 1-@xmath92 standard deviation .
we also compared the spectra of each spectral group of asteroids with the mean spectrum of different taxonomic classes from the bus classification ( http://smass.mit.edu/busdemeoclass.html ) . in each panel , we only plot the bus mean spectra which present the closer similarity with the mean spectra of each group : 1 .
for the first group ( top panel of figure [ fig : spectra ] , 378 objects ) , we find that all spectra resemble the @xmath4-class mean spectra of bus .
2 . in the middle panel of figure [ fig : spectra ] ,
the 118 spectra of the second group fits with @xmath5-class asteroids ( such as 142 polana , for instance ) .
the bottom panel shows the third more populous group containing 13 asteroids . comparing the spectra of this group and the mean spectrum of the @xmath84-class , @xmath93-class , @xmath94-class , and @xmath95-class
, we find that the third group of 13 objects corresponds to asteroids of the @xmath84-complex and its subclasses .
we note that groups ( 1 ) and ( 2 ) are consistent with the two main spectral classes found by @xcite .
moreover , the same authors had already identified the presence of asteroids with @xmath84-class spectral reflectances .
they also noted that two of the largest asteroids , ( 44 ) nysa and ( 135 ) herta have spectra consistent with an @xmath84-class .
as a consequence , we confirm their results using the sdss moc4 large sample of spectra . in figure
[ fig : astariz ] we plot the asteroids of the three spectroscopic groups of the nysa - polana family as a function of @xmath89 and @xmath90 ( with @xmath96 ) .
we show that the @xmath4-class group is well defined with @xmath97 $ ] and @xmath98 $ ] .
the @xmath5-class region stretches from @xmath99 to @xmath100 for @xmath89 and from @xmath101 to @xmath102 for @xmath90 and we find for the @xmath84-class region , @xmath103 $ ] and @xmath104 $ ] . the color palette used in fig .
[ fig : astariz ] is that of @xcite .
blue dots correspond to asteroids with neutral or slightly blue spectra ( b - class ) , asteroids with neutral to slightly red spectra ( x - class ) are displayed in the plot with purple colors , while s - class asteroids are red to yellow .
it is very interesting to note that the three asteroid spectroscopic groups that we found in the nysa - polana family are likely to be the analogs of the different meteorite mineralogies found in almahata sitta . to strengthen this argument
, we also compare in figure [ fig : spectra ] the mean spectrum of each asteroid spectroscopic class of the nysa - polana family ( s , b , x ; red curves ) with published spectra ( from * ? ? ?
* ) of some fragments of almahata sitta .
we note that at the time of writing , spectra of e - chondrites from almahata sitta are not yet publicly available .
so , comparisons between the classes of asteroids in the nysa - polana family and fragments of almahata sittawere performed for the s and b classes only .
however , comparison of spectra of other e - chondrites ( from the relab database from * ? ? ?
* ) with the average spectrum of the the nysa - polana x - class shows a good agreement . concerning the b class , we find that the mean spectrum of the nysa - polana b - class asteroids matches the spectra of the fragments @xmath86 and @xmath3 ( ureilites ) of almahata sitta .
it is important to note that this spectral match was obtained only considering the visible wavelengths between about 350 and 900 nm .
we remind that , from spectroscopic and albedo similarities , @xcite have proposed a link between the b - class ( or the f - class ) and ureilite meteorites .
an important caveat is that the link between b - class asteroids and the almahata sitta ureilites proposed by @xcite is based on the noisy spectrum of tc@xmath0 .
moreover , b - class asteroids are more commonly associated with carbonaceous chondrites by several studies ( see * ? ? ? * for a review about b - class objects , their meteorite analogues and their composition ) . however , a recent spectroscopic survey of b - class asteroids by @xcite shows that the ensemble of the reflectance spectra of the 45 b - class asteroids analyzed in their work have a continuous shape variation in the range between 500 and 2500 nm , from a monotonic negative ( blue ) slope to a positive ( red ) slope . @xcite apply a clustering technique to reduce the ensemble of the spectra to 6 optimized averaged spectra or `` groups '' .
interestingly the relab spectrum of the fragment # 4 of almahata sitta shows a good match with the group # 3 of @xcite in the region between 500 and 2500 nm ( de leon , j. private communication september , 2011 ) . concerning the s class , we show , in fig .
[ fig : spectra ] , that fragment @xmath85 ( an ordinary chondrite ) is rather close to the s - class mean spectrum , as expected , but not in very good agreement ( see e.g. the review of * ? ? ? * for spectroscopical links between ordinary chondrites and s - class asteroids ) .
one possible explanation for this spectral mismatch is the space weathering of s - class asteroids .
we discuss this possibility in the following subsection .
space weathering is a physical process caused by cosmic rays , collisions , ion bombardment , that alters physical and spectral properties of the surface of atmosphere - less planetary bodies .
more particularly , due to space weathering , s - class asteroids become darker ( their albedo is reduced ) , redder ( the reflectance increases with increasing wavelength ) , and the depth of absorption bands are reduced . as a consequence , the spectral slope changes due to this process @xcite . in @xcite , a strong relation was found between the spectral slope of s - class asteroids and their `` exposure '' to space weathering .
the exposure of an asteroid to space weathering corresponds to time integral the flux of solar ions that the body receives along its orbit .
the exposure to space weathering depends on the age of the asteroid and its average distance from the sun .
therefore , in order to know if space weathering affected s - class asteroids of the nysa - polana family , we calculate the exposure to space weathering of these bodies as shown in @xcite and @xcite . .
] we compute the slope @xmath105 of a sdss spectrophotometric data by fitting a linear equation normalized to @xmath106 at @xmath107 ( @xmath59 nm ) to the datapoints .
to calculate the exposure , we need to estimate the average collisional age of main belt asteroids ( @xmath108 ) , which can be constrained by the age of the belt ( @xmath109 gyr ) and by collisional times ( @xmath110 ) : @xmath111 + t_{lhb}\,e^{-\frac{t_{lhb}}{\tau_{coll}}}\ ] ] according to @xcite , the collisional lifetime can be estimated as a function of asteroid diameter ( using sdss h absolute magnitudes and assuming an albedo of @xmath112 for all s - class objects ) . in the end , we have : @xmath113 with @xmath33 and @xmath34 the asteroid proper semimajor axis and eccentricity , respectively . in figure
[ fig : spaceweathering ] , we plot both the spectral slope of nysa - polana s - class objects as a function of the exposure to space weathering and the best linear fit to this distribution .
the two - tailed probability for the linear correlation is lower than @xmath114 which means the linear correlation is significant . ] .
this shows that the spectral slope increases with the exposure .
semimajor axes and eccentricities do not vary much within the nysa - polana family . as a consequence ,
the exposure variation mainly comes from asteroid diameter , through the computation of @xmath108 .
hence , plotting the spectral slope as a function of the exposure , or @xmath108 , or asteroid diameters ( @xmath115 ) provide quite similar figures ( not shown here ) where the exposure increases with time or diameter but with the following abscissa ranges : @xmath116 gyr or @xmath117 km .
the spectral slope increase with exposure or asteroid age is then the proof that space weathering occurs for s - class asteroids of the nysa - polana family .
we also note that the spectral slope of the fragment @xmath85 of almahata sitta has the value of @xmath118 . as a consequence , the spectral mismatch between asteroids of the s - component of the nysa - polana family and the ordinary chondrites of almahata sitta as observed in figure [ fig : spectra ]
can be explained by space weathering .
as mentioned in sections [ section_intro ] and [ section_dynamics ] , the other possible origin source for tc@xmath0 , from a dynamical point of view , is the population of dispersed asteroids at low inclination . in the present section ,
we perform the same spectroscopic study as for the nysa - polana family ( section [ section_polana ] ) but using objects of the background located in the inner main belt ( @xmath119 au ) with a proper inclination lower than @xmath120 . selecting the best observations of the sdss moc4 catalog ( i.e. observations with a relative uncertainty @xmath81 on the in - band photometric flux , corresponding to 2828 objects ) , we find 4 large groups of asteroids corresponding to the following spectral classes : b , c , x , s , as well as small clusters of q , v , and a - class asteroids .
in addition , for the b , x , and s groups of the background , we obtain mean spectra very similar to those of the nysa - polana family ( not shown here ) .
we also note that space weathering is also found for s - class asteroids of the background ( see fig .
[ fig : spaceweathering ] ) . as a consequence , the three ingredients required to form tc@xmath0 ( s - class asteroid / h - chondrite ; b - class asteroid / ureilite ; x - class asteroid / enstatite chondrite ) are both found in the nysa - polana family and in the background at low inclination . in our attempt to definitively conclude on the origin region of tc@xmath0 , we have compared the distribution of b - class asteroids for the two source regions . in order to get a homogeneous selection of b - class asteroids both from the nysa - polana family and the background at low inclination , we have selected objects such as their spectra are similar to the b - class mean spectrum of the nysa - polana family , within the 1-@xmath92 error bar .
although this selection reduces the size of the original b - class groups , we limit the overlap of classes ( b and x classes can overlap ; see fig .
[ fig : spectra ] ) . from this selection ,
we plot in fig .
[ fig : distribution ] , the cumulative number of asteroids as a function of @xmath121 absolute magnitude as well as the distribution in eccentricity , inclination and @xmath121 magnitude as a function of semimajor axis .
we find more b - class asteroids from the background at low inclination ( dashed curve in the first plot of fig .
[ fig : distribution ] ) than from the polana group ( black curve ) . as a consequence
, the background at low inclination seems to dominate the polana group , even if , around magnitude 17 , the slope of the polana group curve looks a little steeper than the curve for the background . in the other plots in orbital elements
, we can see that the two regions overlap in some points and that b - class objects of the background are much more dispersed compared to polana asteroids . as in @xcite
, we also represent the absolute magnitude @xmath122 a a function of semimajor axis , both for the polana group and the background ( the plot of * ? ? ?
* was done for the whole nysa - polana family ) .
the plot appears to be v - shaped which is a feature known to be associated both with the size - dependent ejection velocity field and with the drift in the proper semimajor axis induced by the yarkovsky effect ( e.g. * ? ? ?
* ) . in the plot ,
the black curve shows the @xmath121-dependent semimajor axis distribution induced by the yarkovsky effect that best fits the boundaries of the observed distribution for the polana b - class asteroids .
objects below this curve are expected to be interlopers and may belong to the background .
the extrapolated yarkovsky - induced distribution predicts that the polana group should reach the outer edge of the @xmath20 resonance for objects with @xmath123 , which for a polana - like albedo of @xmath124 translates into a diameter @xmath125 km .
this means that objects smaller than @xmath126 km , such as tc@xmath0 , can easily escape the polana group and the inner main belt . concerning the background at low inclination
, different size limits can be computed due to a large variation of proper orbital inclinations . for an inclination of @xmath127 ,
the @xmath20 resonance boundary is found at @xmath128 au , which gives us an @xmath121 magnitude of @xmath129 and an asteroid diameter limit of @xmath130 km .
for the highest inclination ( @xmath131 ) , we obtain @xmath132 at @xmath133 au , which is equivalent to a b - class asteroid ( @xmath134 ) of 8 km . as a result
, the background at low inclination could have also delivered tc@xmath0 through the @xmath20 resonance . because the background at low inclination is not a dynamical family , its v - shaped structure in @xmath135 was not especially expected .
we then think that the background at low inclination could correspond to an old break - up of the nysa - polana family . as a consequence
, we can conclude that tc@xmath0 comes from the inner main belt , more particularly from the background at low inclination or the nysa - polana family , and that these two sources could be genetically linked .
meteorite strewn fields with fragments of different mineralogical types are very rare .
almahata sitta and kaidun @xcite are probably the only known specimen . in the scenario
that mixing occurs by a collisional process , the paucity of mineralogically - mixed meteorites suggests that the process that formed almahata sitta
i.e. mixing the material between projectile and target is very rare ; in most cases the projectile is pulverized and leaves negligible traces in the target @xcite .
it is unclear which conditions allow for mixing between projectile and target .
impact velocity is probably a key parameter .
the average impact velocity between asteroids in the main belt is 4.4 - 5.3 km / s @xcite . if projectile / target mixing were possible at these impact speeds , meteorites like almahata sitta would probably be frequent , which is not the case
thus , we think that unusually low impact velocities are needed for mixing .
this could prevent the target from pulverizing and could lead to macroscopic projectile fragments being implanted in the regolith of the impacted body or gravitationally bound to fragments of the target , if the latter is catastrophically disrupted by the impact .
the fact that we give evidence that tc@xmath0 comes from the nysa - polana family ( see section [ section_polana ] ) or the background at low inclination ( see section [ background ] ) , which are characterized by a mixing a taxonomic classes , also suggests that a specificity of the members of these families / regions may be unusually low collision velocities with projectiles that are also on low - inclination orbits .
for all these reasons we did a systematic search for projectiles that could hit nysa - polana family members or dispersed asteroids of the background at very low speeds .
we did this search using the algorithm for the calculation of the intrinsic collision probability @xmath136 between pairs of asteroids , described in @xcite .
this algorithm , given the semimajor axis @xmath33 , eccentricity @xmath34 and inclination @xmath36 of the two orbits ( we use proper values for each pair of selected asteroids ) assumes that the angles @xmath137 , @xmath138 , @xmath139 ( mean anomaly , argument of perihelion and longitude of the ascending node ) of the two objects have a uniform probability distribution over the range @xmath140 ; then it computes which fraction of these angles corresponds to the two objects being closer to each other than 1 km ; finally , this fraction is translated into an intrinsic collision probability per year ( @xmath136 ) , using the orbital periods of the two objects and assuming that they are not in resonance with each other . for our goals , we modified this algorithm in order to take into account only orbital intersections corresponding to relative speeds smaller than @xmath141 km / s .
admittedly , this velocity threshold is arbitrary . given the exceptional character of almahata sitta , we need a threshold much smaller than the typical impact velocity among random asteroids , that is to say a sub - sonic impact velocity , in order to preserve the target . because most of almahata sitta fragments recovered are mainly made of ureilites , which we consider as analogs of b - class asteroids ( as previously mentioned in section [ section_intro ] and according to * ? ? ?
* ) , we assume that the favorite scenario for the formation of tc@xmath0 involves a low velocity collision of asteroids of x and s classes ( projectiles ) with a b - class member ( target ) of the nysa - polana family . and @xmath90 , with @xmath142 , @xmath53 , @xmath54 , @xmath55 , @xmath56 corresponding to sdss magnitudes at the following central wavelengths : 477 nm , 623 nm , 763 nm and 913 nm , respectively .
bluer dots ( @xmath143 $ ] and @xmath144 $ ] ) are related to dark objects while yellow to red colors ( @xmath145 $ ] and @xmath146 $ ] ) gather brighter objects ( see figure [ fig : astariz ] for more details ) .
data come from the sdss moving objects catalog 4.,title="fig : " ] + and @xmath90 , with @xmath142 , @xmath53 , @xmath54 , @xmath55 , @xmath56 corresponding to sdss magnitudes at the following central wavelengths : 477 nm , 623 nm , 763 nm and 913 nm , respectively .
bluer dots ( @xmath143 $ ] and @xmath144 $ ] ) are related to dark objects while yellow to red colors ( @xmath145 $ ] and @xmath146 $ ] ) gather brighter objects ( see figure [ fig : astariz ] for more details ) .
data come from the sdss moving objects catalog 4.,title="fig : " ] as for projectiles we considered all asteroids known in the astorb file ( asteroid orbital elements database : ftp://ftp.lowell.edu/pub/elgb/astorb.html ) of the whole main belt and we computed the probability of collision of each asteroid with a b - class object near the center of the b - class group of the nysa - polana family .
figure [ fig : collision ] shows the region of orbital element space where impacts can occur ( impact probability @xmath1470 ) .
this figure shows that asteroids included in the region of low - velocity impacts have different sdss colors and thus different taxonomies . as a consequence ,
collisions involving b - class , s - class , and x - class asteroids are likely to be possible . in the end , depending on the location of the target within the polana group , we find that projectiles come from one of the following families : ( i ) nysa - polana , ( ii ) flora or ( iii ) massalia , located in the inner main belt , or ( iv ) hestia which is in the central main belt , very close to the 3:1 mean motion resonance with jupiter .
more particularly , for targets with a semimajor axis ( @xmath33 ) smaller or equal than @xmath148 au , the flora family and objects of the background , very close to the flora and nysa - polana family are the main source of projectiles ; for targets with 2.3 @xmath149 2.4 au and eccentricities larger than @xmath1500.15 , projectiles in general are likely to come from the nysa - polana family ; for targets with 2.3 @xmath149 2.4 au and eccentricities smaller than @xmath151 0.15 , projectiles are likely to come from the massalia family ; for targets with @xmath152 2.4 au projectiles come from the hestia family .
similar results are found when studying collisions involving b - class asteroids of the background at low inclination .
once identified the region of potential impactors , we proceded to estimate the likelihood and the frequency of impacts , which depend on the sizes , number density of asteroids , and average impact probability in the region .
the average impact probability in the region was derived by averaging the intrinsic impact probabilities ( one of the output of our algorithm ) of the asteroids in the region ; we found @xmath153 impacts / km@xmath154/year .
first of all , we estimated the number of impacts and the size of the impactors hitting tc@xmath0 during its collisional lifetime assuming a diameter of 4 m ( given by its h value of 30.9 @xcite and assuming a geometric visible albedo of 0.05 ) .
the impact probability per projectile for tc@xmath0 is given by @xmath155 with @xmath156 the radius of tc@xmath0 and @xmath110 its collisional lifetime estimated from @xcite ( @xmath157 myr ) .
as a consequence , we find that tc@xmath0 has an impact probability per projectile of the order of @xmath158 during its collisional lifetime . to obtain the number of impacts on tc@xmath0 during its collisional lifetime , we estimated the number of projectiles by calculating the size frequency distribution ( sfd ) of the population of the asteroids in the impact region and extrapolating the corresponding function to asteroid sizes below the detection limit of current discovery surveys . figure [ fig : sfd ] shows the sfd in terms of the cumulative number of asteroid as function of their absolute magnitude ( @xmath121 ) .
we used @xmath121 values from the astorb file . in order to extrapolate the sfd asteroid sizes below the current detection limits
, we calculated the equation of the straight line that best fits the base-10 logarithm of the known sfd and that is compatible with the size frequency distribution of the whole main belt ( see * ? ? ? * ) ( see figure [ fig : sfd ] ) .
the line has equation @xmath159 where @xmath160 is the @xmath161 10-base logarithm of the cumulative number of asteroids with @xmath121 smaller or equal than a threshold value .
the best - fit value for @xmath105 and @xmath162 are @xmath141 and -4.5 respectively , implying about @xmath163 asteroids with @xmath164 , @xmath165 objects with @xmath166 , @xmath41 with @xmath167 and @xmath168 with @xmath169 , if the linear extrapolation is assumed to be valid .
we can deduce that for @xmath169 , we obtain only @xmath170 impacts with tc@xmath0 . in the end , 1 impact at v @xmath171 0.5 km / s during tc@xmath0collisional lifetime would require @xmath172 objects ( i.e. @xmath173 objects ) .
using the sfd equation , this means we need a collision with an object of magnitude @xmath174 .
this correspond to the size of a pebble of 2.4 cm of diameter .
clearly the contamination brought by low speed impacts is very small ( the volume ratio between tc@xmath0 and the pebble is 1.1@xmath175 ) and certainly not at the level of 20 - 30% as implied by the number of chondrites found amongst the ureilites of the almahata sitta meteorites .
much more likely , tc@xmath0 was a fragment of a larger asteroid that was liberated by an impact near the @xmath20 resonance few millions years ago .
in fact , we note that ( 1 ) the dynamical time to deliver asteroids from the @xmath20 resonance to orbits similar to that of tc@xmath0 is of the order of 19 myr ( see * ? ? ?
* for 1999 @xmath176 coming from the polana region ) ; ( 2 ) the cosmic ray exposure age of almahata sitta meteorites is of [email protected] myr @xcite ; as a consequence , tc@xmath0 spent most of its life as a single asteroid traveling from the @xmath20 resonance to the orbit that brought it to impact the earth .
it makes sense to investigate what was the contamination brought by low velocity impacts on the parent body of tc@xmath0 .
the average impact probability implies that we have @xmath151 @xmath178@xmath179 impacts / km@xmath154 over the age of the solar system , which correspond to 1 impact of an asteroid of @xmath180 taking into account the sfd of the objects in the region .
the diameter of an asteroid with h=27 is between 10 and 25 m depending on the albedo . again
the level of contamination is very little , of the order of 1.@xmath181 .
when we take into account the necessity of mixing meteorites of at least three different mineralogies , we conclude that the formation of tc@xmath0 by means of low - velocity collisions is not a realistic scenario .
from our study , it seems that the nysa - polana family and the background at low inclination are good candidates for the origin of 2008 tc@xmath0 and almahata sitta .
first of all , as mentioned in section [ section_dynamics ] , the nysa - polana family is located close to the @xmath20 secular resonance which is the favorite route leading to primitive neos and more particularly to asteroid 2008 tc@xmath0 .
moreover , the proper inclination of the nysa - polana family is very similar to that of tc@xmath0 , which should have been maintained during the transfer of tc@xmath0to the neos region through the @xmath20 secular resonance .
we also know , from our algorithm of spectral classification ( section [ section_method ] ) , that the nysa - polana family gathers the 3 spectral classes ( s , b , and x ) , which are proposed analogs to almahata sittafragments ( section [ section_polana ] ) under the hypothesis that ureilites are linked to b - class asteroids .
more specifically , ( 1 ) the mean spectrum of b - class asteroids of the nysa - polana family is spectrally matched at least in the visible with available spectra of ureilitic fragments of almahata sitta , ( 2 ) considering space - weathering effects , the mean spectrum of s - class asteroids of the nysa - polana family is compatible with the spectrum of the h - chondrite fragment @xmath85 , ( 3 ) a good agreement is found between x - class asteroids and enstatite chondrites from other meteorite falls ( we remind that enstatite chondrites are part of almahata sitta ) .
of course , a comparison with enstatite chondrites from almahata sitta fragments would be very useful to get a definitive match between the nysa - polana family and almahata sitta . in section [ background ] , the same kind of work was performed for objects of the inner main belt coming from the background at low inclination ( @xmath2 ) .
we concluded that these dispersed asteroids could also be at the origin of tc@xmath0 . in section [ section_formation ]
, we tried to explain the formation of tc@xmath0 by low velocity impacts ( below @xmath141 km / s ) between different mineralogies in the neighborhood of and within the nysa - polana family . selecting tc@xmath0 as the target asteroid ( d=4 m ) , we find a probability per projectile about @xmath182 impacts during its collisional lifetime ( i.e. in 16.2 myr ) .
as a consequence , impacts at low velocity are extremely rare and there is little chance that tc3 was formed by low - velocity impacts in the current asteroid belt .
this implies that the heterogeneous composition of the parent body of tc@xmath0has to be inherited from a time when the asteroid belt was in a different dynamical state , most likely in the very early solar system .
one could think that an asteroid of ureilite composition was contaminated by impactors of different nature when the asteroid belt was still massive and dynamically cold , so that mutual collisions were frequent and at low velocity . however , this view is probably simplistic .
in fact , a body of ureilite composition needs to be formed in the interior of a large carbonaceous asteroid which underwent significant thermal alteration @xcite .
this asteroid needs to have undergone a collisional disruption to expose the ureilite material in space .
the same is true for the bodies of hn composition , with n larger than 3 @xcite .
but collisional disruptions require large relative velocities , in contrast with the view of a dynamically cold belt .
thus , the asteroid belt could not be overall dynamically cold when the parent body of tc@xmath0 formed .
these considerations suggest that , conversely to what is usually thought , accretion and collisional erosion had to co - exist for some time in the asteroid belt . for this to be possible
, presumably there had to be still a significant amount of gas in the system so that , although large asteroids could be on dynamically excited orbits , the orbit of their small fragments were rapidly re - circularized by gas - drag .
consequently , the mutual relative velocities of these fragments were small and a new phase of accretion was possible for them .
we remark that the heterogeneity of tc@xmath0 is not at the microscopic level ; each of the meteorites delivered to the ground are of a distinct class .
thus , tc@xmath0 seems to be an agglomeration of meteorite - sized ( i.e. few dm ) pebbles of different nature .
pebbles of this size are strongly coupled with the gas and are extremely sensitive to pressure gradients .
they play the key role in the new models of planetesimal formations , based on the concentration of dm - size pebbles in the eddies of a turbulent disk and on the process of streaming instability @xcite
. these models of planetesimal formation in a turbulent disk seem a priori to be particularly favorable to explain the coexistence of collisional erosion and accretion .
large planetesimals are formed by the concentration of a large number of pebbles , forming self - gravitating clumps .
once formed , the orbits of these large planetesimals are rapidly excited by the stochastic gravitational perturbations exerted by the turbulent disk @xcite . if the threshold for collisional break - up is achieved , the pebble - size fragments of these large bodies are re - injected into the game : by being concentrated into new eddies they can give origin to new large planetesimals and so forth .
admittedly , quantitative work is needed to support this scenario ; also , other more classical planetesimal formation mechanism in presence of gas drag @xcite might explain the coexistence of erosion and accretion as well . in this respect , it will be important to understand from the observational point of view if macroscopic heterogeneity as that of tc@xmath0 is the exception or the rule among asteroids .
tc@xmath0 is the first object of this kind that has ever been observed , but it is also the first fall of an asteroid on earth documented live and for which an extensive and exhaustive search for meteorites has been done .
so , it might not be as rare as one could be tempted to believe .
indeed , a second similar case has just been reported @xcite .
now that the possibility for macroscopic heterogeneity is recognized , careful investigations ( also conducted by remote sensing techniques ) may reveal additional interesting cases . understanding which fraction of the asteroids are of primary or secondary accretion will be a fundamental step to constrain the asteroid formation and evolution models .
we thank o. michel and p. bendjoya for providing us their method of classification as well as a. cellino , p. tanga , m. mller , h. campins , b. carry , and p. vernazza for helpful discussions .
programming tools made available to us by the gaia data processing analysis consortium ( dpac ) have been used within this work .
j. gayon - markt is also grateful to the centre national detudes spatiales ( cnes ) for financial support .
clark , b. e. , ziffer , j. , nesvorny , d. , campins , h. , rivkin , a. s. , hiroi , t. , barucci , m. a. , fulchignoni , m. , binzel , r. p. , fornasier , s. , demeo , f. , ockert - bell , m. e. , licandro , j. , moth - diniz , t. 2010 , journal of geophysical research ( planets ) , 115 , 6005 .
jenniskens , p. , shaddad , m. h. , numan , d. , elsir , s. , kudoda , a. m. , zolensky , m. e. , le , l. , robinson , g. a. , friedrich , j. m. , rumble , d. , steele , a. , chesley , s. r. , fitzsimmons , a. , duddy , s. , hsieh , h. h. , ramsay , g. , brown , p. g. , edwards , w. n. , tagliaferri , e. , boslough m. b. , spalding , r. e. , dantowitz , r. , kozubal , m. , pravec , p. , borovicka , j. , charvat , j. , vaubaillon , j. , kuiper , j. , albers , j. , bishop , j. l. , mancinelli , r. l. , sandford , s. a. , milan , s. n.n nuevo , m. & worden s. p. 2009 , nature , 458 , 485 jenniskens , p. , vaubaillon , j. , binzel , r. p. , demeo , f. e. , nesvorn , d. , bottke , w. f. , fitzsimmons , a. , hiroi , t. , marchis , f. , bishop , j. l. , vernazza , p. , zolensky , m. e. , herrin , j. s. , welten , k. c. , meier , m. m. m. , & shaddad m. h. 2010 , meteoritics and planetary science , 45 , 1590 shaddad m. h. , jenniskens , p. , numan , d. , kudoda , a. m. , elsir , s. , riyad i. f. , ali , a. e. , alameen , m. , alameen , n. m. , eid , o. , osman , a. t. , abubaker , m. i. , yousif , m. , chesley , s. r. , chodas , p. w. , albers j. , edwards , w. n. , brown , p. g. , kuiper , j. , & friedrich , j. m. 2010 , meteoritics and planetary science , 45 , 1557 | asteroid 2008 tc@xmath0 was a near earth asteroid that impacted the earth on 2008 october 7 .
meteorites were produced by the break - up of 2008 tc@xmath0 in the high atmosphere and at present , about @xmath1 meteorites - called almahata sitta - coming from 2008 tc@xmath0 have been recovered .
a mineralogical study of almahata sitta fragments shows that the asteroid 2008 tc@xmath0 was made of meteorites of different types ( ureilites , h , l , and e chondrites ) .
understanding the origin of this body and how it was put together remain a challenge . here
we perform a detailed spectroscopical and dynamical investigation to show that the most likely source region of 2008 tc@xmath0 is in the inner main belt at low inclination ( @xmath2 ) .
we show that asteroids with spectroscopic classes that can be associated with the different meteorite types of almahata sitta are present in the region of the main belt that includes the nysa - polana family and objects of the background at low inclination . searching for a possible scenario of formation for 2008 tc@xmath0
, we show that there is little chance that 2008 tc@xmath0 was formed by low velocity collisions between asteroids of different mineralogies , in the current asteroid belt .
it seems more likely that the heterogeneous composition of 2008 tc@xmath0 was a inherited from a time when the asteroid belt was in a different dynamical state , most likely in the very early solar system . because ureilites are fragments of a large , thermally metamorphosed asteroid , this suggests that the phases of collisional erosion ( the break - up of the ureilite parent - body ) and collisional accretion ( the formation of the parent body of 2008 tc@xmath0 ) overlapped for some time in the primordial asteroid belt .
techniques : spectroscopic catalogues meteorites , meteors , meteoroids minor planets , asteroids : individual : 2008 tc@xmath0 |
in this paper we introduce a numerical code designed to solve the einstein field equations for axisymmetric spacetimes . even though the predominant focus in numerical relativity in recent years has been to study situations of relevance to gravitational wave detection , and hence lacking symmetries , there are still numerous interesting problems , both physical and computational , that can be tackled with an axisymmetric code .
the advantages of restricting the full 3d problem to axisymmetry ( 2d ) are that the complexity and number of equations are reduced , as are the computational requirements compared to solving a similar problem in 3d .
prior numerical studies of axisymmetric spacetimes include head - on black hole collisions @xcite , collapse of rotating fluid stars @xcite , the evolution of collisionless particles applied to study the stability of star clusters @xcite and the validity of cosmic censorship @xcite , evolution of gravitational waves @xcite , black hole - matter - gravitational wave interactions @xcite , and the formation of black holes through gravitational wave collapse @xcite and corresponding critical behavior at the threshold of formation @xcite .
our goals for creating a new axisymmetric code are not only to explore a wider range of phenomena than those studied before , but also to provide a framework to add adaptive mesh refinement ( amr ) and black hole excision to allow more thorough and detailed investigations than prior works .
the outline of the rest of the paper is as follows . in section [ sec : formalism ]
we describe the @xmath0 decomposition @xcite of spacetime that we adopt to arrive at our system of equations
. the @xmath0 formalism is the familiar adm space + time decomposition ( in this case @xmath1 ) applied to a dimensionally reduced spacetime obtained by dividing out the axial killing vector , following a method devised by geroch @xcite . in section [ sec : coords_and_vars ] we specialize the equations to our chosen coordinate system , namely cylindrical coordinates with a conformally flat @xmath2metric . at this stage
we do not model spacetimes with angular momentum , and we include a massless scalar field for the matter source . in section [ sec : apph ]
we discuss how we search for apparent horizons during evolution . in section [ sec : implementation ] we describe our numerical implementation of the set of equations derived in section [ sec : coords_and_vars ] .
a variety of tests of our code are presented in section [ sec : test ] , which is followed by conclusions in section [ sec : conclusion ] . some details concerning our finite difference approximations ,
solution of elliptic equations via the multi - grid technique , and a spherically symmetric code used for testing purposes are given in appendices a and b. unless otherwise specfied , we use the units and conventions adopted by misner , thorne and wheeler @xcite .
the most common approach in numerical relativity is to perform the so - called @xmath3 , or adm , split of the spacetime that one would like to evolve . in this procedure , a timelike vector field
is chosen together with spatial hypersurfaces that foliate the spacetime .
if axisymmetry is assumed , it is usually incorporated once the adm decomposition has been done , and is reflected in the independence of the various metric and matter quantities on the ignorable , angular coordinate , @xmath4 .
our approach , which follows geroch @xcite and nakamura _ et al _
@xcite , reverses this procedure .
we assume axisymmetry from the outset and perform a reduction of the spacetime based on this assumption .
once we have projected out the symmetry , we perform an adm - like split ( now a @xmath1 split ) of the remaining @xmath5manifold . more specifically , we begin with a @xmath6dimensional spacetime metric manifold and latin letters to denote coordinates in the reduced @xmath5manifold . ]
@xmath7 on our manifold @xmath8 .
the axisymmetry is realized in the existence of a spacelike killing vector x^= ( ) ^ , with closed orbits .
we define the projection operator @xmath9 , allowing us to project tensors from the 4dimensional manifold @xmath10 with metric @xmath11 to a @xmath5dimensional manifold @xmath12 with metric @xmath13 , as g _
= _ - x_x _ , where @xmath14 is the norm of the killing vector s^2=x^x_. with the definition of the vector y_= x _ , the metric on the full , @xmath6dimensional spacetime can be written as _
= ( cc g_ab + s^2 y_a y_b & s^2 y_a + s^2 y_b & s^2 ) . projecting , or dividing out the symmetry amounts to expressing @xmath6dimensional quantities in terms of quantities on the @xmath5manifold .
for instance , the connection coefficients @xmath15 associated with the @xmath6dimensional metric are ^(4 ) ^ _ & = & ^(3 ) ^ _ + ^ _ + & = & ^(3 ) ^ _ + s^2 g^ + y^ , [ christof ] where @xmath16 are the connection coefficients constructed from the @xmath5metric @xmath9 , and we have defined the antisymmetric tensor z _ = _ y _ - _ y_. notice that @xmath17 is an intrinsically @xmath5dimensional object , in that @xmath18 . with some algebra , the ricci tensor on the @xmath6manifold , @xmath19 ,
can now be written as the ricci tensor on the reduced space , @xmath20 , together with additional terms involving fields coming from the dimensional reduction ^(4)r _
= ^(3)r _ + d _ ^ _ - d_^
_ + ^ _ _ ^- ^ _ ^ _ , [ ricci_4 ] where @xmath21 is the covariant derivative on the @xmath5 dimensional manifold . expressed in terms of @xmath14 , @xmath22 , @xmath23 and @xmath24 , the components of ( [ ricci_4 ] ) are ^(4)r _ & = & 14 s^4 z_bcz^bc - s d^a d_a s , [ eq : r_phiphi ] + ^(4)r_a & = & 12s d^c(s^3
z_ac ) + y_a , [ eq : r_rhia ] + ^(4)r_ab & = & ^(3)r_ab - 1s d_a d_bs - 12 s^2 z_acz_b^c - 1s d^cy_b ) + y_a y_b .
[ eq : r_ab ] taking the trace of ( [ ricci_4 ] ) , and using the definitions described above , gives the decomposition of the ricci scalar as ^(4)r = ^(3)r - 2s d^a d_as - 14 s^2 z_bcz^bc .
[ eq : ricci ] the @xmath6dimensional einstein equations , with a stress - energy tensor @xmath25 , are ^(4)r _ - ^(4)r _
= 8t_. [ eq : einstein ] using equations ( [ eq : r_phiphi]-[eq : ricci ] ) , we can write the einstein equations as d^ad_a s & = & - w_aw^a 2s^3 - 8s ( t _ - 12 s^2 t_^ ) , [ eq : s_eom ] + d_[a w_b ] & = & 8 s _ ab^c t_c , [ eq : w_eom ] + ^(3)r_ab & = & 1s d_a d_bs + 12 s^2 z_acz_b^c + 8 ( t _ g^_a g^_b - 12 g_ab t_^ ) , [ eq : r_eom ] where we have introduced the twist @xmath26 of the killing vector w_= s^42 _ y^z^ , and the four and three dimensional levi - civita symbols @xmath27 and @xmath28 , respectively .
the twist vector @xmath26 is intrinsically @xmath5dimensional , _
i.e. _ @xmath29 . furthermore ,
@xmath30 is divergence free d^a=0 .
[ eq : div_w_eom ] at this point , the first reduction is essentially done .
equation ( [ eq : r_eom ] ) can be viewed as the @xmath5dimensional einstein equations , coupled to the projection of the stress - energy tensor @xmath25 and to induced `` matter fields '' @xmath14 and @xmath31 ( or @xmath23 ) .
the equations of motion for @xmath14 and @xmath31 are given by ( [ eq : s_eom ] ) and ( [ eq : w_eom ] ) respectively ; additional equations of motion will need to be specified for whatever true matter fields one incorporates into the system .
this procedure so far is completely analogous to the kaluza - klein reduction from five to four dimensions , in which the @xmath32dimensional geometry becomes gravity in @xmath6 dimensions coupled to electromagnetism and a scalar field @xcite . here
, however , the reduced @xmath5manifold has no dynamics and we have the rather appealing picture of having divided out " the dynamics of the four dimensional gravitational system and reinterpreted the two degrees of freedom in the gravitational field as scalar ( @xmath14 ) and `` electromagnetic '' ( @xmath23 , or @xmath33 ) degrees of freedom .
dimensions has only a single degree of freedom as compared to the two degrees of freedom in @xmath6 dimensions . ]
we now perform the adm split of the remaining spacetime .
this is done by first foliating the @xmath5dimensional spacetime into a series of spacelike hypersurfaces with unit , timelike normal vector @xmath34 .
then , similar to the dimensional reduction above , we decompose quantities into components orthogonal and tangent to @xmath34 , using the projection tensor @xmath35 h_ab = g_ab + n_a n_b .
we now define the components of @xmath34 , and the induced @xmath2dimensional metric @xmath36,-dimensional tensor components . ] using the following decomposition of the @xmath5metric g_ab dx^a dx^b = -^2 dt^2 + h_ab ( dx^a + ^adt ) ( dx^b + ^bdt ) , where @xmath37 is the lapse function and @xmath38 the shift vector .
the gravitational equations now become the evolution equations for the components of the @xmath2metric @xmath36 and the @xmath2extrinsic curvature @xmath39 _
t h_ab & = & -2k_ab + _ a _ b + _ b _ a [ eq : hdot ] + _ t k_ab & = & ^c _ c k_ab + k_ac _ b ^c + k_bc _ a ^c + + & & - 2k_ack_b^c - _ a_b - ^(3)r_ab , [ eq : kdot ] [ eq : k_ab_dot ] the hamiltonian constraint equation ^(2)r - k^a_bk^b_a + k^2 = ^(3)r + 2 ^(3)r_abn^a n^b [ eq : hc ] and the @xmath40 and @xmath41 momentum constraint equations _ a k_b^a - _ ak = - ^(3)r_cd n^d h^c_b . [ eq : mc ] in the above , @xmath42 is the covariant derivative compatible with the @xmath2metric @xmath36 , @xmath43 , and @xmath44 and @xmath45 are the @xmath2dimensional ricci tensor and ricci scalar , respectively .
note that because gravity in @xmath5 dimensions has no propagating degrees of freedom , the constraint equations fix the @xmath5dimensional geometry completely .
thus , if desired , one can use the constraint equations ( [ eq : hc]-[eq : mc ] ) instead of the evolution equations ( [ eq : hdot]-[eq : kdot ] ) to solve for @xmath36 .
the freely specifiable degrees of freedom of the @xmath6manifold are encoded in @xmath14 and @xmath31 , which are evolved using ( [ eq : s_eom ] ) , ( [ eq : w_eom ] ) and ( [ eq : div_w_eom ] ) . note that ( [ eq : w_eom ] ) and ( [ eq : div_w_eom ] ) constitute four equations for the three components of @xmath33the purely spatial have four equations for the three components of @xmath33 .
the purely spatial part of ( [ eq : w_eom ] ) is , in @xmath3 language , the angular momentum constraint equation and only needs to be solved at the initial time in a free evolution of @xmath33 .
we note that the restricted class of axisymmetric spacetimes having no angular momentum ( rotation ) is characterized by the existence of a _ scalar twist _
, @xmath46 , such that @xmath47 . in the vacuum case
, @xmath46 generally represents odd parity gravitational waves , while @xmath14 encodes even parity , or brill waves@xcite .
we further note that this class includes the special case @xmath48 , which will be the focus of our discussion below .
in this section we describe a particular coordinate system and set of variables which , in the context of the formalism described in the previous section , provides us with the concrete system of partial differential equations that we solve numerically .
we also detail the outer boundary conditions we use , and the on - axis regularity conditions necessary to obtain smooth solutions to these equations .
we only consider spacetimes with zero angular momentum , and no odd - parity gravitational waves ; therefore @xmath49 .
we choose a conformally flat , cylindrical coordinate system for the @xmath2metric h_ab dx^a dx^b = ( , z , t)^4 ( d^2 + dz^2 ) .
[ hab_def ] this choice for @xmath36 exhausts the coordinate freedom we have to arbitrarily specify the two components of the shift vector@xmath50 and @xmath51 . in order to maintain the form ( [ hab_def ] ) during an evolution
, we use the momentum constraints , which are elliptic equations , to solve for @xmath52 and @xmath53 at each time step .
the hamiltonian constraint provides a third elliptic equation that we can use to solve for the conformal factor @xmath54 . for a slicing condition
, we use maximal slicing of @xmath55 hypersurfaces _ in the @xmath6dimensional manifold__i.e .
_ we impose @xmath56 , where @xmath57 is the trace of the extrinsic curvature tensor of @xmath55 slices of @xmath11 .
this condition ( specifically @xmath58 ) gives us an elliptic equation for the lapse . instead of directly evolving the norm of the killing vector , @xmath14
, we evolve the quantity @xmath59 , defined by s = ^2 e^| , and furthermore , we convert the resultant evolution equation for @xmath59 ( [ eq : s_eom ] ) to one that is first order in time by defining the quantity @xmath60 , which is `` conjugate '' variable to @xmath59 , via |&= & - 2k_^- k_z^z + & = & - n^a ( s)_,a + .
[ eq : bo_def ] part of the motivation behind using @xmath59 and @xmath60 as fundamental variables is to simplify the enforcement of on - axis regularity . in particular , regularity as @xmath61 implies that @xmath59 and @xmath60 must exhibit leading order behavior of the form @xmath62 and @xmath63 respectively , and experience has shown it to be easier to enforce such conditions , than to enforce the leading order behavior of @xmath14 ( or its time derivative ) near the axis , which is @xmath64 . as mentioned previously , the only matter source we currently incorporate is a massless scalar field @xmath65 , which satisfies the usual @xmath6dimensional wave equation @xmath66 .
we convert this equation to first - order - in - time form by defining a conjugate variable @xmath67 : ^2 n^a _ , a .
the stress - energy tensor for the scalar field is t _ = 2 _ , _ , - _ ^ , _ , . using all of the above definitions and restrictions within the formalism detailed in the previous section
, we end up with the following system of equations that we solve with our numerical code , described in the next section .
the maximal slicing condition results in the following elliptic equation for @xmath37 @xmath68 - \frac{\psi^4}{6\alpha } \bigl [ 2 \alpha \rho \bo + \beta^\rho{}_{,\rho } - \beta^z{}_{,z } \bigr]^2 = 16 \pi \alpha \pi^2 .
\label{eq : cons_alpha}\end{aligned}\ ] ] the hamiltonian constraint gives an elliptic equation for @xmath54 @xmath69 + { \psi^4
\over 6\alpha^2 } \bigl [ 2\alpha\rho\bar{\omega } + \beta^\rho{}_{,\rho } - \beta^z{}_{,z } \bigr]^2 \quad\qquad\qquad \nonumber\\ \qquad \ , = \ , - 16 \pi \left ( \pi^2 + \phi_{,\rho}{}^2 + \phi_{,z}{}^2 \right ) % - 6 \left ( \rho^2 \left(\rho \bs\right)_{,\rho } \right)_{,\rho^3 } - 6 \bigl ( \rho^2 \left(\rho \bs\right)_{,\rho } \bigr)_{,\rho^3 } \nonumber\\ % - 2 \left ( \left ( \rho \bs \right)_{,\rho } \right)^2 % - 2 \left ( \rho \bs \right)_{,zz } % - 2 \left ( \left ( \rho \bs \right)_{,z } \right)^2 . - 2 \bigl ( \left ( \rho \bs \right)_{,\rho } \bigr)^2 - 2 \bigl ( \rho \bs \bigr)_{,zz } - 2 \bigl ( \left ( \rho \bs \right)_{,z } \bigr)^2 .
\label{eq : cons_psi}\end{aligned}\ ] ] the @xmath40 and @xmath41 momentum constraints , which provide elliptic equations that we use to solve for @xmath52 and @xmath53 , are @xmath70 % - \frac{8}{3 } \alpha \bar{\omega } % -\frac{2 \alpha \rho } { 3 } \left [ % 6 \bo \frac{\psi_{,\rho } } { \psi } % + \bo_{,\rho } % + 3 \bo \left(\rho \bs \right)_{,\rho } % \right]=0 , \frac{2}{3}\beta^{\rho}{}_{,\rho\rho } + \beta^{\rho}{}_{,zz } + \frac{1}{3 } \beta^z{}_{,z\rho } -\frac{2 \alpha \rho } { 3 } \bigl [ 6 \bo \frac{\psi_{,\rho } } { \psi } + \bo_{,\rho } + 3 \bo \left(\rho \bs \right)_{,\rho } \bigr ] - \frac{8}{3 } \ , \alpha \bar{\omega } \qquad\qquad\qquad\qquad\qquad\qquad \nonumber\\ - \frac{2}{3 } \bigl [ \frac { \alpha_{,\rho } } { \alpha } - 6 \frac { \psi_{,\rho } } { \psi } \bigr ] \bigl ( \beta^\rho{}_{,\rho } - \beta^z{}_{,z } \bigr ) - \bigl [ \frac { \alpha_{,z } } { \alpha } - 6 \frac { \psi_{,z } } { \psi } - \left ( \rho\bs \right)_{,z } \bigr ] \bigl ( \beta^\rho{}_{,z } + \beta^z{}_{,\rho } \bigr ) \ ; = \ ; - 32\pi { \alpha \over \psi^2 } \pi_{,\rho } , \label{eq : cons_betarho}\end{aligned}\ ] ] and @xmath71 % \left ( \beta^\rho{}_{,z } + \beta^z{}_{,\rho } \right ) % \nonumber\\ % + \left [ 2 \left ( \rho \bs \right)_{,z } % - \frac{4}{3 } \left ( \frac { \alpha_{,z } } { \alpha } % - 6 \frac { \psi_{,z } } { \psi } \right ) % \right ] % \left ( \beta^z{}_{,z } - \beta^\rho{}_{,\rho } \right ) % - \frac{2 \alpha\rho}{3 } \left ( %
6 \bo \frac{\psi_{,z}}{\psi } % + \bo_{,z } %
\right ) % \nonumber\\ % + 32\pi { \alpha \over \psi^2 } \pi_{,z } % - 2\alpha(\bs_{,z})\rho^2\bo=0 .
\beta^z{}_{,\rho\rho } + \frac{4}{3}\beta^z{}_{,zz } - \frac{1}{3 } \beta^{\rho}{}_{,z\rho } - \frac{2 \alpha\rho}{3 } \bigl [ 6 \bo \frac{\psi_{,z}}{\psi } + \bo_{,z } + 3 \bo ( \rho \bs)_{,z } \bigr ] \qquad\qquad\qquad\qquad\qquad\qquad\qquad\quad\quad \nonumber\\ + \frac{4}{3 } \ , \bigl [ \frac { \alpha_{,z } } { \alpha } - 6 \frac { \psi_{,z } } { \psi } - \frac{3}{2 } \left ( \rho \bs \right)_{,z } \bigr ] \bigl ( \beta^\rho{}_{,\rho } - \beta^z{}_{,z } \bigr ) + \bigl [ \frac { 2 \alpha } { \psi^6 } \bigl ( \frac{\rho \psi^6}{\alpha } \bigr)_{,\rho^2 } \ , + \left ( \rho \bs \right)_{,\rho } \bigr ] \bigl ( \beta^\rho{}_{,z } + \beta^z{}_{,\rho } \bigr ) \ ; = \ ; - 32\pi { \alpha
\over \psi^2 } \pi_{,z } .
\label{eq : cons_betaz}\end{aligned}\ ] ] the definition of @xmath60 in eq .
( [ eq : bo_def ] ) gives an evolution equation for @xmath59 ( where the overdot denotes partial differentiation with respect to @xmath72 ) @xmath73_{,\rho}. - \bigl ( { \beta^{\rho } \over \rho } \bigr)_{,\rho}.\ ] ] the evolution equation for @xmath60 is @xmath74 + 64\pi{\alpha\over \psi^4 } \rho ( \phi_{,\rho^2})^2 .
\label{eq : evol_bo}\end{aligned}\ ] ] we also have an evolution equation for @xmath54 , which we optionally use instead of the hamiltonian constraint ( [ eq : cons_psi ] ) to update @xmath54 @xmath75 the definition of @xmath67 and the wave equation for @xmath76 give @xmath77 and @xmath78 + \frac{\alpha}{\psi^2 } \left [ \left ( \rho \bs \right)_{,\rho } \phi_\rho + \left ( \rho \bs \right)_{,z } \phi_z \right ] .
\label{eq : evol_pi}\end{aligned}\ ] ] to complete the specification of our system of equations , we need to supply boundary conditions . in our cylindrical coordinate system , where @xmath40 ranges from @xmath79 to @xmath80 and @xmath41 ranges from @xmath81 to @xmath82 , we have two distinct boundaries : the physical outer boundary at @xmath80 , @xmath83 , and @xmath84 ; and the axis , at @xmath79 .
historically , the axis presented a stability problem in axisymmetric codes .
we solve this problem by enforcing regularity on the axis , and , as described in section [ sec : implementation ] , adding numerical dissipation to evolved fields .
the regularity conditions can be obtained by inspection of the equations in the limit @xmath85 , or more formally , by transforming to cartesian coordinates and demanding that components of the metric and matter fields be regular and single valued throughout @xcite .
garfinkle and duncan @xcite have further proposed that in order to ensure smoothness on the axis , one should use quantities that have either even or odd power series expansions in @xmath40 as @xmath61 , but which do not vanish faster than @xmath86 .
it is interesting that the quantities which we found to work best also obey this requirement . as discussed earlier , the particular choice of @xmath59 and @xmath60 as fundamental variables was partly motivated by regularity concerns .
the results are _ , ( 0,z , t ) & = & 0 [ eq : bc_alpha ] + _ , ( 0,z , t ) & = & 0 + ^z_,(0,z , t ) & = & 0 + ^(0,z , t ) & = & 0 [ eq : bc_betarho ] + ( 0,z , t ) & = & 0 + ( 0,z , t ) & = & 0 + _ , ( 0,z , t ) & = & 0 + _ , ( 0,z , t ) & = & 0 at the outer boundary , we enforce asymptotic flatness by requiring _
r ( r , t ) & = & 1 + + o(r^-2 ) [ eq : obc_alpha ] + _ r ( r , t ) & = & 1 + + o(r^-2 ) + _ r ^z(r , t ) & = & + o(r^-2 ) + _ r ^(r , t ) & = & + o(r^-2 ) , [ eq : obc_betarho ] for undetermined functions @xmath87 , @xmath88 , @xmath89 , @xmath90 , and @xmath91 .
these latter relations are converted to mixed ( robin ) boundary conditions ( see appendix a for details ) and then are imposed at the outer boundaries of the computational domain : @xmath92 , @xmath93 , and @xmath94 .
we have also experimented with the use of dirichlet conditions on @xmath95 and @xmath53 at the outer boundaries ( specifically @xmath96 and @xmath97 there ) , and have found that these work about as well as the robin conditions . for the scalar field , we assume that near the outer boundary we can approximate the field as purely radially outgoing , and require ( r ) _ , t + ( r ) _ , r = 0 . for scalar field configurations far from spherical symmetry , this approximation suffers and reflections are relatively large .
however , in general , the reflections do not grow and are somewhat damped . for the other two evolved quantities , @xmath59 and @xmath60 , we use this same naive condition for lack of any better , more physically motivated conditions .
while this condition proves to be stable with damped reflections , a better condition is sought and this issue remains under investigation .
for initial conditions , we are free to set @xmath98 . once the free data is chosen , we then use the constraint and slicing equations to determine @xmath99 .
specifically , we define a general pulse shape g_x(,z ) = a_x [ g_x ] characterized by six parameters @xmath100 and then choose initial data of the form ( 0,,z ) & = & g_(,z ) ( 0,,z ) & = & g_(,z ) ( 0,,z ) & = & g_(,z ) ( 0,,z ) & = & 0 .
for @xmath101 , these pulses are gaussian , spherical shells centered at @xmath102 with radius @xmath103 and pulse width @xmath104 . for @xmath101 and @xmath105 ,
the pulses are spherical .
the factor of @xmath40 in the initial data for @xmath59 and @xmath60 ensures the correct behavior on axis for regularity .
for the evolutions presented here , we let @xmath106 so that the initial configuration represents a moment of time symmetry .
we note , however , that we are also able to generate and evolve approximately ingoing initial data .
in this section we describe the equation and technique we use to search for apparent horizons ( ahs ) within @xmath107 spatial slices of the spacetime ( see @xcite for descriptions of some of the methods available to find ahs in axisymmetry ) .
we restrict our search to isolated , simply connected ahs . in axisymmetry , such an ah can be described by a curve in the @xmath108 plane , starting and ending on the axis at @xmath79 .
we define the location of the ah as the level surface @xmath109 , where @xmath110 and @xmath111 the ah is the outermost , marginally trapped surface ; hence , we want to find an equation for @xmath112 such that the outward null expansion normal to the surface @xmath109 , is zero .
to this end , we first construct the unit spatial vector @xmath113 , normal to @xmath114 @xmath115 then , using @xmath113 and the @xmath107 hypersurface normal vector @xmath34 , we construct future - pointing outgoing ( @xmath116 ) and ingoing ( @xmath117 ) null vectors as @xmath118 the normalization of the null vectors is ( arbitrarily ) @xmath119 .
the outward null expansion @xmath120 is then the divergence of @xmath121 projected onto @xmath114 @xmath122 using the definition of the extrinsic curvature @xmath39 , and substituting ( [ null ] ) into ( [ exp1 ] ) , we arrive at the familiar form for the null expansion when written in terms of adm variables @xmath123 note that because the normalization of @xmath121 is arbitrary , so ( to some extent ) is that of @xmath120 .
the above normalization is chosen so that @xmath120 measures the fractional rate of change of area with time measured by an observer moving along @xmath34 . substituting ( [ fs2 ] ) and ( [ sadef ] ) into ( [ exp2 ] ) , and setting @xmath124 , we are left with an ordinary differential equation for @xmath112 .
this equation takes the following form , where a prime @xmath125 denotes differentiation with respect to @xmath126 : @xmath127 @xmath128 is a rather lengthy function of its arguments , non - linear in @xmath129 and @xmath130 ; for brevity we do not display it explicitly .
all of the metric functions and their gradients appearing in ( [ rpp_eqn ] ) are evaluated along a given curve of integration , and hence are implicitly functions of @xmath126 .
@xmath126 ranges from @xmath131 to @xmath132 , and regularity of the surface @xmath109 about the axis requires @xmath133 .
integration of ( [ rpp_eqn ] ) therefore proceeds by specifying @xmath129 at @xmath134 ( for instance ) , and then `` evolving '' @xmath129 until either @xmath135 , or @xmath129 diverges at some value of @xmath136 . if an ah exists , and assuming @xmath137 is inside the ah , then the ah can be found by searching for the ( locally ) unique ) augmented with the conditions @xmath133 .
we want the outermost of these surfaces . ]
initial value @xmath138 such that integration of ( [ rpp_eqn ] ) ends at @xmath139 , with @xmath140 and @xmath141 finite . for @xmath142
slightly larger than @xmath143 ( outside the ah ) , the integration will end at @xmath135 with @xmath144 , indicating an irregular point on the surface ; similarly , for @xmath142 slightly smaller than @xmath143 ( inside the ah ) the integration will end with @xmath145 . therefore ,
if we can find a reasonable bracket about the unknown @xmath143 , we can use a bisection search to find @xmath143 .
currently , we find a bracket to search by testing a set of initial points , equally spaced in @xmath41 at intervals of @xmath146
. this seems to work well in most situations , and the search is reasonably fast .
we use a second - order runge - kutta method to integrate equation ( [ rpp_eqn ] ) .
the metric functions appearing in @xmath128 are evaluated using bilinear interpolation along the curve .
in this section we describe the numerical code that we have written to solve the equations listed in section [ sec : coords_and_vars ] .
some details are deferred to appendix [ app : solver ] .
we use a uniform grid of size @xmath147 points in @xmath40 by @xmath148 points in @xmath41 , with equal spacing @xmath149 in the @xmath40 and @xmath41 directions .
the value of a function @xmath150 at time level @xmath151 and location @xmath152 within the grid , corresponding to coordinate @xmath153 , is denoted by @xmath154 . for the temporal discretization scale we use @xmath155 , where @xmath156 is the courant factor , which for the type of differencing we employ ,
should be less than one for stability ; typically we use @xmath157 . the hyperbolic equations ( [ eq : evol_bs]-[eq : evol_pi ] )
are discretized using a second - order accurate crank - nicholson type scheme , whereby we define two time levels , @xmath72 and @xmath158 , and obtain our finite difference stencils by expanding in taylor series about @xmath159 .
this gives the following second - order accurate approximation to the first derivative of @xmath150 with respect to time @xmath160 second - order accurate approximations to functions and spatial derivative operators at @xmath159 are obtained by averaging the corresponding quantity , @xmath161 , in time : @xmath162 thus , after discretization of the evolution equations using ( [ ref_cn_time ] ) and ( [ ref_cn_spat ] ) , function values are only referenced at times @xmath72 and @xmath158 , even though the stencils are centered at time @xmath163 .
specific forms for all the finite difference stencils that we use can be found in appendix [ app : solver ] .
we add kreiss - oliger dissipation @xcite to the evolution of equations of @xmath164 and @xmath60 ( in addition to @xmath54 during partially constrained evolution ) , as described in appendix [ app : solver ] . to demonstrate that this is _ essential _ for the stability of our numerical scheme , we compare in fig .
[ fig : diss ] the evolution of @xmath59 from simulations without and with dissipation , but otherwise identical . the elliptic equations ( [ eq : cons_alpha]-[eq : cons_betaz ] )
are solved using brandt s fas multigrid ( mg ) algorithm @xcite , described in some detail in appendix [ app : solver ] .
there are no explicit time derivatives of functions in these equations , and we discretize them at a single time level @xmath151 ( _ i.e. _ we do not apply the crank - nicholson averaging scheme for the elliptics ) .
we use either a fully constrained evolution , solving for @xmath165 and @xmath54 using the constraint equations and slicing condition , or a partially constrained evolution where instead of using the hamiltonian constraint to update @xmath54 , we use the evolution equation ( @xmath166 ) .
partially constrained evolution has proven to be useful due to the occasional failure of the mg solver in the strong - field regime ( _ i.e. _ close to black hole formation ) .
use of the evolution equation for @xmath54 ( rather than the hamiltonian constraint ) circumvents this problem in many instances ; however , in certain brill - wave dominated spacetimes , free evolution of @xmath54 is not sufficient to ensure convergence of the mg process .
we are currently working to make the mg solver more robust in these situations .
the code is written in a combination of rnpl ( rapid numerical prototyping language @xcite ) and fortran 77 .
the hyperbolic equations are implemented in rnpl , which employs a point - wise newton - gauss - seidel iterative relaxation scheme to solve these equations , while the mg solver is implemented in fortran ( see appendix [ app : solver ] for more mg details ) .
a pseudo - code description of the time - stepping algorithm used is as follows : .... a ) as an initial guess to the solution at time t+dt , copy variables from t to t+dt b ) repeat until ( residual norm < tolerance ) : 1 : perform 1 newton - gauss - seidel relaxation sweep of the evolution equations , solving for the unknowns at time t+dt 2 : perform 1 mg vcycle on the set of elliptic equations at time t+dt end repeat .... for the residual norm used to terminate the iteration we use the the infinity norm of the residuals of all updated unknowns .
in this section , we describe some of the tests we have performed to check that we are solving the correct set of equations .
the first test consists of checking the equations against those derived with a computer algebra system @xcite . by inputting the metric and
coordinate conditions , the computer derived equations can then be subtracted from our equations and simplified . by finding that the differences simplify to 0
, we can conclude that two sets of equations agree . for diagnostic purposes and as tests of the equations and of their discretization , we compute several quantities during the numerical evolution .
the first is the adm mass @xcite @xmath167 where the integral is evaluated on a flat 3-space , _ i.e. _ with metric @xmath168 .
the spatial 3-metric @xmath169 is that from our curved space solution , but has its indices raised and lowered with the flat metric .
integrating around the boundaries of our numerical grid , the normal vectors @xmath170 are @xmath171 and @xmath172 .
after some algebra , the adm mass becomes m_adm & = & _ z_max ^4 d & & - _
z_min ^4 d & & + _ _ max ^4 dz . a second set of quantities we calculate are the @xmath173-norms of the residuals of the evolution equations for the extrinsic curvature ( [ eq : k_ab_dot ] ) , which we denote @xmath174 .
because we do not directly evolve individual components of the extrinsic curvature , these residuals will not be zero ; however , they _
should _ converge to zero in the limits as the discretization scale @xmath175 , and the outer boundary positions @xmath176 .
note that we include these last conditions because it is only in the limit @xmath177 that our outer boundary conditions are fully consistent with asymptotic flatness .
the convergence properties of our code are measured by computing the convergence factor , @xmath178 , associated with a given variable , @xmath179 , obtained on grids with resolution @xmath180 , @xmath181 and @xmath182 via q_u = .
in particular , for the case of @xmath183 ( second order ) convergence , we expect @xmath184 as @xmath185 .
the first set of tests we present here are comparisons of @xmath76 and @xmath54 from the evolution of spherically symmetric initial data to the corresponding functions computed by a 1d spherically symmetric code , the details of which are presented in appendix [ app : spherical ] . in general , the results from the two codes are in good agreement . a sample comparison is illustrated in fig .
[ fig : spherical ] which shows the scalar field obtained with the 1d code as well as two radial slices of the corresponding solution calculated using the axisymmetric code .
note , however , that we do not expect _ exact _ agreement in the limit @xmath175 for a fixed outer boundary location , as the `` rectangular '' boundaries of the axisymmetric code are , in general , incompatible with precise spherical symmetry .
in the second series of tests , we examine evolutions of brill waves and non - spherical scalar pulses . figs . [
fig : brill_10][fig : oblate_20 ] show results from two typical initial data sets , each computed using two distinct outer boundary positions .
each figure plots ( a ) the adm mass @xmath186 , ( b ) the @xmath173-norm of the residual of the @xmath187 component of the evolution equation for the extrinsic curvature @xmath188 , and , ( c ) the convergence factor @xmath189 of @xmath54 , as functions of time ( the convergence factor @xmath161 for other functions exhibit similar behavior as @xmath189 , and so for brevity we do not show them ) . here
, one expects to see an improvement of the results namely trends toward mass conservation early on , a zero residual , and a convergence factor of 4in the limits @xmath190 and @xmath191 .
after energy has reached the outer boundary , and to a lesser extent before ( as is evident in the scalar field example in figs .
[ fig : oblate_10 ] and [ fig : oblate_20 ] ) , we fail to get consistency with the evolution equation ( [ eq : k_ab_dot ] ) as @xmath190 , for _ fixed _ @xmath192 .
this is a measure of the inaccuracy of our outer boundary conditions ( [ eq : obc_alpha]-[eq : obc_betarho ] ) ; though the trends suggest that we _ do _ achieve consistency in the limit @xmath191 .
= 8.5 cm = 8.5 cm = 8.5 cm = 8.5 cm finally , we show some results of a simulation of black hole formation from the collapse of a spherically symmetric distribution of scalar field energy . again , by looking at spherically symmetric collapse we can compare with the 1d code ( obtaining the same level of agreement as seen in the example in fig .
[ fig : spherical ] ) .
however , here we want to show the behavior of our coordinate system ( in particular the maximal slicing ) in the strong - field regime , which demonstrates the need to incorporate black hole excision techniques and/or adaptive mesh refinement ( amr ) before attempting any serious investigation of physics with this code .
[ fig : bh_plots ] shows plots of the adm mass estimate ( [ madm ] ) , an estimate of the black hole mass @xmath193 , where @xmath194 is the area of the apparent horizon , and the minimum value of the lapse as a function of time from the simulation .
[ fig : bh_psi ] shows the conformal factor @xmath54 at several times , in the central region of the grid .
maximal slicing is considered _ singularity avoiding _ , because as the singularity is approached in a collapse scenario , the lapse @xmath37 tends to 0 , as demonstrated in fig .
[ fig : bh_plots ] .
this effectively freezes the evolution inside the black hole , though it causes a severe distortion in the @xmath107 slices as one moves away from the black hole .
this particular coordinate pathology is evident in fig .
[ fig : bh_psi ] .
recall from the @xmath2metric ( [ hab_def ] ) that @xmath195 determines proper length scales in the @xmath40 and @xmath41 directions ; thus the rapid growth with time of @xmath54 shown in fig .
[ fig : bh_psi ] means that a given coordinate area represents increasing proper area .
furthermore , the increase in magnitude of @xmath54 in the strong - field regime ( which happens even when black holes do not form , and in non - spherical scalar field and brill wave evolution , though not to the same extent as shown in fig .
[ fig : bh_psi ] ) implies that our effective numerical resolution decreases in those regions , as some feature of the solution with a given characteristic size will span less of the coordinate grid .
thus , in the end , even though maximal slicing may prevent us from reaching a physical singularity , the `` grid - stretching '' effect is just as disastrous for the numerical code , preventing any long - term simulation of black hole spacetimes . for these reasons we will add black hole excision techniques and amr before exploring physics with this code ; our efforts in this regard are well underway , and will be described elsewhere .
we have described a @xmath0 gravitational evolution model which evolves axisymmetric configurations of gravitational radiation and/or a scalar field .
a thorough battery of tests confirms that the correct equations are being solved . in particular
, we have provided evidence that the code is second - order convergent , consistent and conserves mass in the limit where the outer boundary position goes to infinity .
the unigrid code described here is the first step towards our long - term goal of studying a range of interesting theoretical and astrophysical phenomena in axisymmetry .
these include gravitational collapse of various matter sources and gravitational waves , the corresponding critical phenomena at the threshold of black hole formation , head - on black hole collisions and accretion disks . to this end
, we need to include support for angular momentum and additional matter fields in the code , as well as to add additional computational and mathematical infrastructure adaptive mesh refinement , black hole excision and the capability of running in parallel on a network of machines .
all of these projects are under development , and results will be published as they become available .
another goal of this project is to provide a platform from which to develop computational technology for 3d work . in particular , we see development of amr in axisymmetry as a precursor to its incorporation in 3d calculations . likewise ,
accurate and stable treatment of boundary conditions presents a continual challenge in numerical relativity , and it is possible that we can develop an effective treatment of boundaries in axisymmetry that will generalize to the 3d case .
the authors would like to thank david garfinkle for helpful discussions .
mwc would like to acknowledge financial support from nserc and nsf phy9722068 .
fp would like to acknowledge nserc , the izaak walton killam fund and caltech s richard chase tolman fund for their financial support .
ewh and sll would like to acknowledge the support of nsf grant phy-9900644 .
sll acknowledges support from grant nsf phy-0139980 as well as the financial support of southampton college .
ewh also acknowledges the support of nsf grant phy-0139782 . the majority of the simulations described here were performed on ubc s * vn * cluster ( supported by cfi and bckdf ) , and the * maci * cluster at the university of calgary ( supported by cfi and asra )
in this appendix we briefly mention some aspects of our multigrid ( mg ) routine , and list the set of finite difference approximations that we use .
the constraint equations ( [ eq : cons_alpha]-[eq : cons_betaz ] ) are four elliptic equations which , for a fully constrained system , must be solved on every time slice ( _ i.e. _ spatial hypersurface ) . as such , it can be expected that the time taken by a given evolution will be dominated by the elliptic solver and hence we look for the fastest possible solver .
currently , multigrid methods are among the most efficient elliptic solvers available , and here we have implemented a standard full approximation storage ( fas ) multigrid method with v cycling ( see @xcite ) to solve the four nonlinear equations simultaneously .
( when using the evolution equation for @xmath54 in a partially constrained evolution , we use the same multigrid routine described here , except we only solve for the three quantities @xmath95 and @xmath53 during the v - cycle ; @xmath54 is then simply considered another `` source function '' . )
a key component of the mg solver is the relaxation routine that is designed to _ smooth _ the residuals associated with the discretized elliptic equations .
we use point - wise newton - gauss - seidel relaxation with red - black ordering ( see fig . [
fig : redblack ] ) , simultaneously updating all four quantities @xmath196 at each grid - point during a relaxation sweep .
in addition to its use for the standard pre - coarse - grid - correction ( pre - cgc ) and post - cgc smoothing sweeps , the relaxation routine is also used to _ solve _ problems on the coarsest grid .
we use half - weighted restriction to transfer fields from fine to coarse grids and ( bi)linear interpolation for coarse to fine transfers .
we generally use 3 pre - cgc and 3 post - cgc sweeps per @xmath197-cycle , and likewise normally use a single @xmath197-cycle per crank - nicholson iteration .
one complicating factor here is the treatment of the boundary conditions eqs .
( [ eq : bc_alpha]-[eq : bc_betarho ] ) and eqs .
( [ eq : obc_alpha]-[eq : obc_betarho ] ) . in accordance with general multi - grid practice
, we view the boundary conditions as logically and operationally distinct from the interior equation equation .
the outer boundary conditions can generally be expressed as rx [ ap : bc ] where @xmath198 .
taking the derivative of ( [ ap : bc ] ) with respect to @xmath199 , we arrive at the differential form that is applied on the outer boundaries of the computational domain ( @xmath200 ) : x - x _ , - z x_,z = 0 . on the @xmath41-axis ( @xmath79 ) , the conditions eqs .
( [ eq : bc_alpha]-[eq : bc_betarho ] ) take one of the following two forms : a_i , = 0 [ eq : qfit ] or a_i = 0 [ eq : zfit ] equations of the former form are discretized using an @xmath183 backwards difference approximation to the @xmath40 derivative . the interior and boundary differential equations are solved in tandem via the multigrid approach : 1 .
the residual is smoothed using some number of relaxation sweeps . for the interior , eqs .
( [ eq : cons_alpha]-[eq : cons_betarho ] ) are relaxed using red - black ordering as discussed above . after each call of this relaxation routine , a second routine that `` relaxes '' the boundary points
is called @xcite .
2 . for quantities restricted from a fine to a coarse grid ,
discrete forms of ( [ eq : qfit ] ) and ( [ eq : zfit ] ) are applied during the @xmath197-cycle . at the other boundaries ,
straightforward injection is used .
the key idea here is to ensure that the boundary relaxation process does not substantially impact the smoothness of the interior residuals , because it is only for smooth residuals that a coarsened version of a fine - grid problem can sensibly be posed .
finally , in table [ table : diff ] , we show all of the difference operators we use to convert the differential equations listed in [ sec : coords_and_vars ] to finite difference form , using the crank - nicholson scheme described in section [ sec : implementation ] .
in addition , as discussed in section [ sec : implementation ] , we use kreiss - oliger dissipation @xcite to maintain smoothness in the evolved fields .
specifically , we add the kreiss - oliger filter to discretized evolution equations _
t a^n = _ t f^n ( ) by replacing the crank - nicholson time difference operator @xmath201 with @xmath202 : ^_t a^n = _ t f^n ( ) empirically , we find that a value of @xmath203 generally keeps our fields acceptably smooth .
.finite difference operators and their correspondence to differential operators .
here , @xmath204 is an arbitrary grid function defined via @xmath205 , where @xmath180 and @xmath206 are the spatial and temporal grid spacings , respectively .
@xmath207 denotes either of the two spatial coordinates @xmath40 or @xmath41 , with the dependence of @xmath204 on the other suppressed .
the parameter @xmath208 represents a user - specifiable `` amount '' of kreiss - oliger dissipation .
[ cols="<,<,<",options="header " , ]
one simple test of the code compares the results for spherically symmetric initial data with the output of a code which explicitly assumes spherical symmetry . here
we present the equations for this 1d code .
the spacetime metric is : ds^2 = -(^2 + ^4 ^ 2)dt^2 + 2 ^ 4dt dr + ^4 ( dr^2 + r^2 d^2 ) , where @xmath37 , @xmath209 and @xmath54 , are functions of @xmath199 and @xmath72 , @xmath210 is the line element on the unit 2-sphere , and @xmath209 is the radial component of the shift vector ( i.e. @xmath211 ) . adopting maximal slicing to facilitate direct comparison to the axisymmetric code , we have k^i_j = diag(k^r_r(r , t),0,0 )
. then a sufficient set of equations for the coupled einstein - massless - scalar system is @xcite + + 2 + ( k^r_r)^2 ^5 & = & 0 [ eq : ss_psi ] + ( k^r_r ) + 3 k^r_r + & = & 0 [ eq : ss_krr ] + - & = & 0 [ eq : ss_alpha ] ( ) & = & [ eq : ss_beta ] + & = & + [ eq : ss_ph ] + & = & ( + ) [ eq : ss_phi ] + & = & + & & - .
[ eq : ss_pi ] here dot and prime denote derivatives with respect to @xmath72 and @xmath199 , respectively .
the evolution equations ( [ eq : ss_ph]-[eq : ss_pi ] ) are discretized using an @xmath183 crank - nicholson scheme .
( [ eq : ss_psi ] ) and ( [ eq : ss_krr ] ) are similarly discretized using @xmath183 finite difference approximations , then solved iteratively for @xmath54 and @xmath212 at each time step .
once @xmath76 , @xmath67 , @xmath54 and @xmath212 have been determined , @xmath37 and @xmath209 are found from @xmath183 finite - difference versions of eqs .
( [ eq : ss_alpha ] ) and ( [ eq : ss_beta ] ) .
the code is stable and second - order convergent . 1 l. l. smarr , ph.d .
dissertation , university of texas at austin , unpublished ( 1975 ) k. r. eppley , ph.d .
dissertation , princeton university , unpublished ( 1977 ) s. l. shapiro and s. a. teukolsky , `` collisions of relativistic clusters and the formation of black holes '' , phys . rev .
* d45 * , 2739 ( 1992 ) p. anninos , d. hobill , e. seidel , l. smarr and w. suen , `` the collision of two black holes , '' phys .
lett . * 71 * , 2851 ( 1993 ) [ gr - qc/9309016 ] .
j. baker , a. abrahams , p. anninos , s. brandt , r. price , j. pullin and e. seidel , `` the collision of boosted black holes '' , phys .
* d55 * , 829 ( 1997 ) p. anninos and s. brandt `` head on collision of two unequal mass black holes , '' phys .
lett . * 81 * , 508 ( 1998 ) [ gr - qc/9806031 ] .
t. nakamura , `` general relativistic collapse of axially symmetric stars leading to the formation of rotating black holes , '' _ prog . of theor .
physics _ * 65 * , 1876 - 1890 ( 1981 ) .
r. f. stark and t. piran , `` gravitational wave emission from rotating gravitational collapse , '' phys .
. lett . * 55 * , 891 ( 1985 ) .
t. nakamura , k. oohara and y. kojima , `` general relativistic collapse of axially symmetric stars '' , prog .
. suppl . *
90 * , 13 ( 1987 ) m. shibata , `` axisymmetric simulations of rotating stellar collapse in full general relativity criteria for prompt collapse to black holes '' prog .
phys . * 104 * , 325 ( 2000 ) .
a. m. abrahams , g. b. cook , s. l. shapiro and s. a. teukolsky , `` solving einstein s equations for rotating space - times : evolution of relativistic star clusters , '' phys .
d * 49 * , 5153 ( 1994 ) .
shapiro and s.a .
teukolsky , `` formation of naked singularities : the violation of cosmic censorship , '' phys .
* 66 * , 994 ( 1991 ) .
d. garfinkle and g. c. duncan , `` numerical evolution of brill waves , '' phys .
rev . * d 63 * , 044011 ( 2001 ) [ gr - qc/0006073 ] . m. alcubierre , s .
brandt , b .
brugmann , d .
holz , e .
seidel , r .
takahashi , j .
thornburg `` symmetry without symmetry : numerical simulation of axisymmetric systems using cartesian grids '' int .
j. mod .
d * 10 * , 273 ( 2001 ) .
s. brandt , j.a .
font , j.m .
ibanez , j. masso and e. seidel , `` numerical evolution of matter in dynamical axisymmetric black hole spacetimes '' , comput .
. commun . * 124 * 169 ( 2000 ) s. brandt and e. seidel , `` the evolution of distorted rotating black holes i : methods and tests '' , phys . rev .
* d52 * , 856 ( 1995 ) s. brandt and e. seidel , `` the evolution of distorted rotating black holes ii : dynamics and analysis '' , phys . rev . *
d52 * , 870 ( 1995 ) s. brandt and e. seidel , `` the evolution of distorted rotating black holes iii : initial data '' , phys . rev .
* d54 * , 1403 ( 1996 ) a. m. abrahams and c. r. evans , `` trapping a geon : black hole formation by an imploding gravitational wave .
, '' phys .
d. * 46 * , 4117 ( 1992 ) .
a. m. abrahams and c. r. evans , `` critical behavior and scaling in vacuum axisymmetric gravitational collapse , '' phys .
* 70 * , 2980 ( 1993 ) .
k. maeda , m .
sasaki , t .
nakamura , s .
miyama `` a new formalism of the einstein equations for relativistic rotating systems . ''
phys . * 63 * , 719 ( 1980 ) .
k. maeda , `` [ ( 2 + 1)+1]-dimensional representation of the einstein equations , '' in _ proceedings of third marcel grossmann meeting on general relativity _ , ed .
hu ning , ( science press , 1983 ) .
r. geroch , `` a method for generating solutions of einstein s equations , '' j. math .
* 12 * , 918 ( 1971 ) .
misner , k.s .
thorne and j.a .
wheeler , _ gravitation _ , new york , w.h .
freeman and company ( 1973 ) j.m .
bardeen and t. piran , `` general relativistic axisymmetric rotating systems : coordinates and equations '' , phys . rep .
* 96 * 205 ( 1983 ) .
t. appelquist , a. chodos , and p.g.o .
freund , `` modern kaluza - klein theories , '' addison - wesley , menlo park , 1987 ) .
j. thornburg , `` finding apparent horizons in numerical relativity '' , phys . rev .
* d54 * , 4899 ( 1996 ) m. alcubierre , s. brandt , b. brugmann , c. gundlach , j. masso , e. seidel and p. walker , `` test beds and applications for apparent horizon finders in numerical relativity '' , class .
* 17 * 2159 ( 2000 ) kreiss , h .- o . , and oliger , j. , `` methods for the approximate solution of time dependent problems '' , global atmospheric research program publication no . 10 , world meteorological organization , case postale no . 1
, ch-1211 geneva 20 , switzerland ( 1973 ) .
software available from http://laplace.physics.ubc.ca/members/matt/rnpl/index.html we use waterloo maple along with a tensor package written by one of the authors ( mwc ) .
a. brandt , `` multilevel adaptive solutions to boundary value problems , '' _ math . of computation
_ * 31 * , 333 - 390 ( 1977 )
. a. brandt , `` guide to multigrid development , '' in _ lecture notes in mathematics _ * 960 * , 220 - 312 , ( springer - verlag , new york , 1982 ) .
choptuik , `` a study of numerical techniques for radiative problems in general relativity , '' ph.d .
thesis , the university of british columbia , unpublished ( 1982 ) . a subtle point here
concerns the `` relaxation '' occurring on the boundaries .
the finite - difference equations on the boundary yield algebraic equations which determine the given fields there `` exactly . ''
the subtlety arises because these algebraic conditions couple neighboring boundary points and thus despite the fact that we solve these equations exactly , the residual ( as computed after the entire boundary is `` relaxed '' ) will not be identically zero .
we find this procedure suffices to keep residuals sufficiently smooth over both interior and boundary domains . | we present a new numerical code designed to solve the einstein field equations for axisymmetric spacetimes .
the long term goal of this project is to construct a code that will be capable of studying many problems of interest in axisymmetry , including gravitational collapse , critical phenomena , investigations of cosmic censorship , and head - on black hole collisions .
our objective here is to detail the ( 2 + 1)+1 formalism we use to arrive at the corresponding system of equations and the numerical methods we use to solve them .
we are able to obtain stable evolution , despite the singular nature of the coordinate system on the axis , by enforcing appropriate regularity conditions on all variables and by adding numerical dissipation to hyperbolic equations . |
transport phenomena and in particular directed transport are at the heart of many fundamental problems in physics , chemistry and biology @xcite . at the same time they are also of great relevance to technological applications based on a variety of transport - based devices such as rectifiers , pumps , particle separators , molecular switches , and electronic diodes and transistors . of special interest is the realization of novel classes of integrated photonic devices that allow one - directional flow of information - e.g .
optical isolators @xcite .
currently , such unidirectional elements rely mainly on the faraday effect , where external magnetic fields are used to break the space - time symmetry .
this in general requires materials with appreciable verdet constants - typically non compatible with light - emitting wafers @xcite . to anticipate these problems , alternative proposals for the creation of optical diodes and isolators
have been suggested in recent years .
some representative examples include the creation of optical diodes based on asymmetric nonlinear absorption @xcite , second harmonic generation in asymmetric waveguides @xcite , nonlinear photonic crystals @xcite , and photonic quasi - crystals and molecules @xcite . in this paper , we propose a new mechanism for unidirectional optical transport based on configurations involving nonlinear optical materials with parity ( @xmath0 ) and time ( @xmath1)-reflection . this is possible by judiciously interleaving gain and loss regions , in such a way that the ( complex ) refractive index @xmath3 profile satisfies the condition @xmath4 .
a first experimental realization of such ( linear ) arrangements has been recently reported in refs .
@xcite where a @xmath5 dual coupled structure was fabricated and the beam dynamics was investigated . here we show that the interplay of non - reciprocal dynamics arising from @xmath5-symmetry @xcite , and self - trapping phenomena associated with kerr nonlinearities @xcite , can mold the flow of light in a surprising way .
such novel directed dynamics can be exploited in the realization of a new generation of optical isolators or diodes . even though the validity of our arguments can be demonstrated for a variety of non - linear @xmath5-configurations , below
, we will highlight its basic principles , using the simplest possible arrangement , consisting of two @xmath2-coupled waveguide elements with kerr nonlinearity of strength @xmath6 .
each of the waveguides is single - moded one providing gain and the other an equal amount of loss ( see fig .
[ fig1 ] ) .
we have obtained the phase diagram in the @xmath7 plane for which our system acts as an optical diode , and we have identified the minimum propagation length needed , in order to achieve this unidirectional functionality .
detail numerical simulations support our theoretical predictions .
this paper is structured as follows . in sec .
ii an overview of the linear @xmath5-symmetric dimer is presented .
the nonlinear @xmath5-symmetric dimer will be introduced in sec .
iii , where the equations of motion are given in terms of stokes parameters . in subsection iii.a , we present both our theoretical and numerical results on the dynamics of the non - linear @xmath5-symmetric dimer . in subsection
iii.b we calculate the critical value of the non - linearity for which diode action is possible .
finally we will draw our conclusions in sec .
in this section we will briefly review the basic properties of the linear @xmath5-symmetric dimer @xcite . in integrated optics
this simple @xmath5 element can be realized in the form of a coupled system , with only one of the two parallel channels being optically pumped to provide gain @xmath8 for the guided light , whereas the neighbour arm experiences equal amount of loss ( see fig .
1 ) . under these conditions , and by using the coupled - mode approach ,
the optical - field dynamics in the two coupled waveguides are described by the following set of equations : @xmath9 where @xmath10 are modal electric field amplitudes in the amplifying ( eq . [ eq1a]a ) and lossy ( eq . [ eq1a]b ) waveguide channels , @xmath11 represents a dimensionless propagation distance- normalized in units of coupling lengths , and @xmath8 is a scaled gain / loss coefficient , also normalized to the coupling strength . the hamiltonian corresponding to the linear problem of eq .
( [ eq1a ] ) , is written as : @xmath12 and commutes with the combined @xmath5 operator .
a surprising result associated with this class of problems is the possibility that such a @xmath5 symmetric hamiltonian @xmath13 can have an entirely real energy spectrum , despite the fact that , in general , they are non - hermitian @xcite . for the specific example of the non - hermitian hamiltonian of eq .
( [ pth ] ) , a direct diagonalization gives the following set of eigenvalues : @xmath14 which are real as long as the gain / loss parameter @xmath8 is smaller than some critical value , @xmath15 ( _ exact _
@xmath5-symmetric phase ) .
as the gain / loss parameter @xmath8 increases above @xmath16 , the eigenvalues becomes complex ( _ broken _ @xmath5-symmetric phase ) .
the corresponding eigenvectors of the hamiltonian eq .
( [ pth ] ) are @xmath17 in the exact @xmath5-symmetric phase , both the @xmath13 and @xmath5 operators share the same set of eigenvectors . in this regime ,
the mode intensity is symmetric with respect to the mirror axis of the two waveguides .
as @xmath8 increases above @xmath18 the eigenfunctions of @xmath13 cease to be eigenfunctions of the @xmath5-operator , despite the fact that @xmath13 and the @xmath5-operator still commute .
this happens because the @xmath5-operator is anti - linear , and thus the eigenstates of @xmath13 may or may not be eigenstates of @xmath5 . in the broken @xmath5-symmetric phase ,
the spacial distribution of the modes is asymmetric , one of them living predominantly in the amplifying waveguide and the other in the lossy one . at the phase - transition point
@xmath19 the two eigenfunctions and their corresponding eigenvalues coalesce leading to an exceptional " point singularity @xcite .
the beam dynamics associated with eq .
( [ eq1a ] ) were investigated theoretically in refs .
@xcite while direct measurements were performed in @xcite .
these authors recognized that as the gain / loss parameter @xmath8 reaches @xmath18 , the total beam power starts growing exponentially , while for @xmath20 power oscillations are observed ( see figs .
the most dramatic effect in the beam evolution is the appearance of non - reciprocal wave propagation ( see figs .
specifically , the beam propagation pattern differs depending on whether the initial excitation is on the left or right waveguide .
this is contrasted with the @xmath21 case ( fig .
1a , b ) , where the beam propagation is insensitive to the initial condition .
( color online ) beam propagation in two coupled linear waveguides . for the parameters of the simulation ( we use normalized coupling units ) ,
the spontaneous @xmath5-breaking take place at @xmath22 . in all cases ,
left ( right ) panels correspond to an initial excitation at the left ( right ) channel .
( a , b ) a passive system corresponding to @xmath21 .
the propagation is reciprocal ; ( c , d ) @xmath23 corresponding to the exact @xmath5-phase .
a non - reciprocal beam propagation is evident .
although the dynamics is non - hermitian , the evolution is pseudo - unitary " and the total beam power remains bounded ; ( e , f ) for @xmath24 , beam power grows exponentially ( vertical scale is logarithmic ) in both waveguides , while the beam propagation is again non - reciprocal with respect to the mirror axis of the two waveguides .
waveguides are color - coded , indicating balanced gain ( red ) and loss ( green ) regions .
grey - color waveguides indicate a passive ( @xmath21 ) system . ]
we begin our analysis by providing the mathematical model that describes optical wave propagation in a kerr nonlinear @xmath5 symmetric coupled dual waveguide arrangement ( see fig . [ fig1 ] ) .
the two modal field amplitudes are governed by the evolution equations : @xmath25 where @xmath6 is the strength of the kerr - nonlinearity .
equations ( [ eq1 ] ) can be rewritten in terms of the ( real ) stokes parameters @xmath26 , where @xmath27 denote the pauli spin matrices @xcite . in this representation ,
the total field intensity is given by @xmath28 , @xmath29 is the intensity imbalance between the two waveguides , while @xmath30 and @xmath31 . in this representation eqs .
( [ eq1 ] ) take the form : @xmath32 where we have introduced the two real vectors @xmath33 and @xmath34 , and the @xmath35-dimensional stokes vector @xmath36 .
we note that the condition @xmath37 is always satisfied .
it is worth mentioning that eqs .
( [ eq5 ] ) , are identical to the equation of motion of a relativistic negatively -charged particle with zero mass , in a pseudo - electromagnetic field @xmath38 , where @xmath39 represents the energy and @xmath40dimensional momentum of the particle , while the propagation distance @xmath11 has the role of the time .
nonlinear @xmath2-symmetric optical coupled systems can be realistically synthesized on semiconductor wafers- known for their high kerr - like nonlinearities @xcite . as in ref .
@xcite , coupling lengths as low as @xmath41 can be obtained , in which case a gain / loss level below @xmath42 ( readily available in such materials ) will suffice to keep the arrangement in the @xmath2 phase .
in addition , critical switching ( @xmath43 ) can also occur at milliwatt power levels in multi - quantum well configurations .
( color online ) beam propagation in two coupled nonlinear waveguides with non - linearity strength @xmath6 and a complex @xmath5-symmetric refractive index profile .
waveguides are color - coded , indicating balanced gain ( red ) and loss ( green ) regions ( @xmath44 )
. left columns correspond to an initial excitation at the gain waveguide port , while right columns correspond to an initial excitation at the lossy waveguide : ( a , b ) the non - linearity @xmath45 is below the critical value @xmath46 while for ( c , d ) the non - linearity strength @xmath47 is above . ] for @xmath21 , eqs .
( [ eq5 ] ) admits two constants of motion : the total input power @xmath48 and the total energy @xmath49 .
these two constants allow for an analytic solution of the stokes vector @xmath50 in terms of elliptic functions @xcite .
depending on the initial preparation and strength of nonlinearity @xmath6 , we observe two distinct dynamical behaviors .
for example , if the initial beam of total input power @xmath51 , is prepared in one of the two waveguides ( i.e. @xmath52 ) , we observe either rabi oscillations , or self - trapping dynamics @xcite .
the former case corresponds to @xmath53 and results in beam oscillations between the two waveguides , while the latter case , occurs for @xmath54 and leads to localization of the field ( for all times ) at the waveguide that was initially placed . in both cases
, symmetric initial preparation will result to a dynamics which is reciprocal with respect to the axis of symmetry of the two coupled waveguides . for @xmath55 , neither the energy @xmath13 nor the beam power @xmath48 are anymore conserved quantities .
nevertheless , @xmath5-symmetry enforces two other constants of motion @xmath56 : @xmath57 thus indicating that the system of eqs .
( [ eq1 ] ) is fully integrable .
below we will consider the case where initially @xmath51 , @xmath52 , while @xmath58 . in this case , the constants of motion , as defined in eqs .
( [ eq6 ] ) , take the values @xmath59 and @xmath60 .
using @xmath61 and @xmath62 , in this particular case we can express the components of the stokes vector in terms of @xmath63 in the following way @xmath64 substituting the expression for @xmath65 from eq .
( [ eq7]c ) , to the first of the eqs .
( [ eq5 ] ) , we get that @xmath66 even though the problem is soluble by quadratures , the integral in eq .
( [ eq7b ] ) can not be evaluated further and thus a closed expression for @xmath63 is not possible .
it is therefore instructive at this point to gain insight on the properties of the dynamics of this @xmath5 non - linear coupler by numerically solving eqs .
( [ eq1],[eq5 ] ) .
the accuracy of the numerical integration was checked via the conservation laws eq .
( [ eq6 ] ) , which were satisfied up to @xmath67 .
( color online ) main figure : a semi - logarithmic plot of @xmath68 vs. @xmath8 . for the numerical evaluation of @xmath68
we have integrated eq .
( [ eq1 ] ) inset : the efficiency factor @xmath69 vs. nonlinearity strength @xmath6 , for a fixed gain / loss parameter @xmath44 and three different waveguide lengths @xmath70 and @xmath71 . for non - linearity strength
@xmath72 the isolator reaches its optimal efficiency .
] examples of the resulting beam dynamics for @xmath73 and two representative non - linearity strengths @xmath45 and @xmath47 are reported in fig .
[ fig1]a , b and [ fig1]c , d respectively . in contrast to the @xmath21 case @xcite , now the dynamics is non - reciprocal with respect to the axis of symmetry of the system .
while this is true for both values of non - linearity strength @xmath6 , it is much more pronounced for the case of figs .
[ fig1]c , d . in this latter case
, the output field always leaves the sample from the waveguide with gain ( red - colored ) irrespective of the preparation of the input beam . at the same time the output beam intensity at the lossy waveguide approaches zero for longer waveguides .
it is important to stress that in the case of the linear @xmath5-dimer ( see figs 1e - f ) the beam intensity at the lossy waveguide never goes to zero .
instead it increases exponentially , albeit with a smaller prefactor with respect to the one of the gain waveguide .
this novel unidirectional propagation of the @xmath5-symmetric non - linear dimer is the key mechanism for establishing optical isolators ( diodes ) .
it has to be contrasted with the corresponding cases shown in figs .
[ fig1]a , b where the output beam depends on the input state , i.e. an initial excitation at the gain waveguide results in an output field at the lossy guide and vise versa . to quantify the ability of our set - up to act as an optical non - reciprocal device , we have defined the efficiency factor @xmath69 of unidirectional propagation as @xmath74 where @xmath75 is the normalized transmission coefficient associated with the gain ( @xmath76 ) waveguide of length @xmath11 . in our definition we have always assumed that the initial input beam has total power @xmath77 , while the beam is launched either in the gain ( @xmath76 ) or in the loss ( @xmath78 ) waveguide .
the efficiency factor takes values from @xmath79 : a perfect diode corresponds to @xmath80 ( since the term inside the absolute value in eq .
( [ effic ] ) will be zero ) , while the opposite limit of @xmath81 indicates total revival of the field . in the inset of fig .
[ fig2 ] we report our numerical findings for the efficiency factor @xmath69 as a function of the non - linearity strength @xmath6 for three different waveguide lengths @xmath70 , and @xmath71 , and for a fixed value of the gain/ loss parameter @xmath44 .
it is clear that an optimal diode is achieved once the non - linearity strength @xmath6 is larger than a critical value @xmath68 .
( color online ) the numerically extracted first passage distance @xmath82 versus the gain / loss parameter @xmath8 .
the initial conditions are chosen to be @xmath51 and @xmath83 . an inverse power law is observed .
left inset : the proportionality coefficient @xmath84 is plotted versus the nonlinearity strength @xmath6 for @xmath85 .
the red line correspond to the best linear fit . ]
next we present a heuristic argument that aims to estimate the critical non - linearity strength @xmath68 ( as a function of @xmath8 ) , above which the @xmath5 symmetric non - linear dimer acts as an optical diode of high efficiency factor @xmath80 . to this end
we focus our analysis on the temporal behavior of the total power @xmath63 . in the case of ( rabi - like ) oscillations
@xmath63 is bounded between a minimum and a maximum value . instead , in the regime where the coupled system acts as an optical diode , @xmath63 is bounded only from below , while asymptotically it grows in an exponential fashion @xcite .
using the first of eq .
( [ eq5 ] ) together with eq .
( [ eq7]c ) , and requesting the extrema condition @xmath86 ( which is equivalent to @xmath87 ) together with the condition @xmath88 for the existence of a global maxima , we find that @xmath63 shows oscillatory behavior ( i.e. rabi - like oscillations ) if the non - linearity @xmath6 is smaller than @xmath68 , given by @xmath89 in the main panel of fig .
[ fig2 ] we compare eq .
( [ eq8 ] ) with the numerical values found for @xmath68 .
the latter has been evaluated via a direct integration of eq .
( [ eq1 ] ) for systems sizes up to @xmath90 .
the critical nonlinearity @xmath68 was evaluated up to a fourth digit accuracy as the nonlinearity strength for which the total power @xmath63 , is bounded . in all cases
the accuracy of the integration scheme has been guaranteed by requesting that the constants of motion eq .
( [ eq6 ] ) are conserved with accuracy up to @xmath91 . a nice agreement between the theoretical and numerical value of @xmath68
is evident for small values of the gain / loss parameter @xmath8 , while deviations from the theoretical prediction start to be visible as @xmath8 approaches the @xmath5 transition point ( i.e. @xmath92 ) of the linear system .
( color online ) dynamics of the rescaled stokes variables @xmath93 for @xmath94 and various gain / loss parameters @xmath8 : dashed blue line correspond to @xmath95 ; pink line ( @xmath96 ) correspond to @xmath97 ; solid yellow line correspond to @xmath98 ; and red line ( @xmath99 ) to @xmath100 .
the green line ( @xmath101 ) corresponds to the passive system @xmath21 with critical non - linearity @xmath102 , where the motion of the trajectory is on the separatrix .
the trajectory associated with @xmath95 ( see red arrow in fig .
[ fig3 ] ) is typical to the cases where @xmath82 diverge and correspond to the closest one to the separatrix of the passive @xmath21 system . ] finally , we investigate the minimal waveguide length @xmath103 which is required in order to have a high@xmath104 diode . from figs .
[ fig1]c , d we see that the beam evolution follow two distinct scenaria depending on the initial conditions : if the beam is launched initially at the gain waveguide , the propagation is mainly along this channel .
if on the other hand the beam excites the lossy waveguide , there is a minimum propagation distance @xmath103 which is required before the light intensity is concentrated in the gain waveguide .
we have found that @xmath103 is proportional to the `` first passage distance '' @xmath82 associated with the point that @xmath65 becomes zero for the first time . in fig .
[ fig3 ] we report the results of our simulations for @xmath105 for various @xmath106 values or input power levels
. an intriguing feature of @xmath82 is the existence of singularities ( peaks in the @xmath82 ) for some characteristic values of the gain / loss parameter @xmath8 . to understand the origin of these singularities , we have plotted the evolution of the stokes vector @xmath50 , by making use of the rescaled variables @xmath107 . in this representation ,
the magnitude @xmath108 remains constant , and thus we can visualize the evolution on the bloch sphere ( see fig .
[ fig4 ] ) .
it should be emphasized that the bloch trajectories can in general show self - intersections , as they are a projection from a higher dimensional phase space .
one must also distinguish between closed orbits and those approaching an asymptotic state , as this is in general connected to broken and unbroken @xmath5-symmetry @xcite .
we note that closely related bloch dynamics appear in different physical model systems like the ones reported in @xcite .
our analysis , indicated that the singularities in @xmath82 are associated with trajectories that during their evolution , they stay close to the separatrix associated to the critical value @xmath102 ( transition between rabi - oscillations and self - trapping ) of the passive system . leaving aside the issue of the singularities ,
we have found that for all @xmath6-values larger than @xmath68 , the first passage distance @xmath82 follows an inverse power law i.e. @xmath109 where the proportionality factor @xmath84 is @xmath6-dependent .
a best least square fit allow us to extract the various @xmath84 which is in this case @xmath110 ( see inset of fig .
[ fig3 ] ) .
in conclusion , we have proposed a new mechanism for directed transport in nonlinear optical coupled systems , that relies at the interplay between nonlinearity and @xmath5-reflection symmetries .
more specifically , we have observed that above a critical non - linearity strength , the beam evolution is unidirectional , i.e. the output beam remains in the gain channel , irrespective of initial conditions .
such behavior implies that these systems can be used to realize new classes of optical diodes and other unidirectional photonic elements
. of great interest will be to extend these notions to more involved arrangements like nonlinear @xmath5 lattices where nonlinear excitations are expected lead to even more intriguing phenomena .
we would like to acknowledge v. kovanis for useful comments and suggestions .
this work was supported by the dfg for760 , and a grant from the us - israel binational science foundation ( bsf ) , jerusalem , israel .
kittel , _ introduction to solid state physics _ , wiley ( 2005 ) ; r. d. astumian and p. hnggi , phys .
today * 55 * , 33 ( 2002 ) ; f. jlicher , a. ajdari , j. prost , rev . mod . phys . * 69 * , 1269 ( 1997 ) ; p. reimann , phys . rep . * 361 * , 57 ( 2002 ) . f. biancalana , j. appl . phys . * 104 * , 093113 ( 2008 ) ; h. zhou , k - f .
zhu , w. hu , q. guo , s. lan , x - s .
lin , a. v. gopal , j. appl . phys . *
99 * , 123111 ( 2006 ) .
a. guo , et al .
* 103 * , 093902 ( 2009 ) c. m. bender , s. boettcher , phys .
* 80 * , 5243 ( 1998 ) ; c. m. bender , d. c. brody , h. f. jones , phys . rev . lett . * 89 * , 270401 ( 2002 ) ; c. m. bender , d.c .
brody , h. f. jones , and b. k. meister , phys .
* 98 * , 040403 ( 2007 ) .
u. gnther and b. f. samsonov , ibid .
* 101 * , 230404 ( 2008 ) . c. m. bender , s. boettcher , p. n. meisinger , j. math . phys . * 40 * , 2201 ( 1999 ) ; c. m. bender , rep .
. phys . * 70 * , 947 ( 2007 ) ; c. m. bender , d. c. brody , and h. f. jones , phys .
lett . * 89 * 270401 ( 2002 ) . in this case , the system of eq .
( [ eq1 ] ) act as a set of two uncoupled waveguides i.e. in the long time limit one can assume that the coupling constant is essentially zero .
as a result the intensity at the gain waveguide ( which is approximately the same as the total intensity ) increases exponentially , while the one at the lossy waveguide decays exponentially .
we have checked the accuracy of this statement with direct numerical simulations . | we show that non - linear optical structures involving a balanced gain - loss profile , can act as unidirectional optical valves .
this is made possible by exploiting the interplay between the fundamental symmetries of parity ( @xmath0 ) and time ( @xmath1 ) , with optical nonlinear effects .
this novel unidirectional dynamics is specifically demonstrated for the case of an integrable @xmath2-symmetric nonlinear system . |
classical work related to determine the polarization of x - ray photons emitted near rbh considers the newtonian limit , which consists , for symmetry reasons , on horizontal polarization , parallel to the disk and vertical polarization , perpendicular to the disk @xcite .
however , the polarization of photons is determined by local processes in the source and by general relativistic effects within the source and during the propagation to the observer . in recent numerical simulations
@xcite , @xcite , @xcite , @xcite , @xcite the role of general relativistic effects is strongly emphasized and determines the angle of polarization to acquire values in the interval @xmath0 $ ] .
it was pointed out that the radiation deflected by the strong gravitational field near rbh and scattered by the accretion disk toward a distant observer , as returning radiation , suffers larger rotations of the polarization vector .
we chose here to characterize the polarization of the radiation emitted near the rbh by the density matrix , a less explored @xcite , @xcite , @xcite , @xcite , but equivalent approach@xcite , having in mind to take further advantages of the lately development of quantum information theory .
the density matrix completely determines the polarization states of photons @xcite , @xcite and directly points out to polarization as a degree of freedom of photons that can encode quantum information , in the form of one qubit .
accordingly , we consider horizontal and vertical polarization of photons as encoding the qubit state @xcite , @xcite and we infer polarization as a two - level quantum system that spans the two - dimensional hilbert space .
polarization is not the only degree of freedom the single - photons acquire near rbh .
it was advocated @xcite , @xcite that the twisting of spacetime , reflected in the dragging frame effect near rbh , imprints orbital angular momentum ( oam ) to radiation beams emitted by the accretion disk .
numerical simulations have shown a spectrum of oam that widens as the bh spins faster .
the oam flips to wider modes for photons emitted in the innermost regions of the disk that are deflected and scattered to a distant observer as returning radiation .
traditionally , in quantum information theory @xcite , @xcite the parity of oam modes of photons may encode quantum information , as a single qubit . in this context
, to fully characterize the states of photons emitted by accretion disk , a distant observer must measure both degrees of freedom acquired by photons traveling in the spacetime deformed by the presence of rbh polarization and oam . considering the quantum information encoded in the two degrees of freedom of photon , as two qubits , @xcite , @xcite , @xcite
, we determine here how the particular shape of spacetime near rbh manipulates the quantum information and implements elementary quantum gates and simple quantum circuits , like bell states .
we demonstrate the implementation of the polarization hadamard gate @xcite , @xcite not gate and cnot gate by general relativistic effects probed by photons near rbh .
implementation of cnot quantum gate is important since it provides evidence that the shape of spacetime entangles the two degrees of freedom of photons .
accordingly , photons in diagonal and anti - diagonal states could realize the all four bell states .
it is expected that the photons detected from rbh sources should carry a spectrum of quantum information , from photons in separable states to nonmaximally entangled states photons @xcite , @xcite and photons carrying bell states @xcite .
the future measurements of the bell states encoded in single - photons emitted near rbh are promising since the technology required by such a process is already in use .
the main source of information on the efforts to unveil some of the black holes hidden characteristics , such as the speed of spinning , is the polarization of radiation emitted by the accretion disk .
the detection of the polarization vector rotations of photons may be a strong evidence of general relativistic effects that manifest near rbh .
rotations of polarization angle of radiation emitted by accretion disks in the thermal state , have been shown in numerical simulations @xcite , @xcite , @xcite , @xcite , @xcite , assuming the simplest model of a geometrically thin , optically thick , steady - state accretion disk , aligned with the bh spin axis .
due to its simplicity , the stokes parameters formalism is preferred in numerical simulation models in order to characterize the polarization angle of radiation emitted by the accreting black holes .
considering that the polarization is induced by compton scattering , such that no circular polarization is present , the polarization angle is derived from the stoke parameters in the following manner : @xmath1 where @xmath2 , @xmath3 are the stokes parameters related to linear polarization of photons , @xmath4 is the polarization angle .
the polarization angle of the radiation emitted near rbh is sensitive to the black hole spin parameter ( @xmath5 ) and the angle of inclination ( @xmath6 ) of the accretion disk in relation to the distant observer .
the angle of inclination is not relevant here and we assigned it a constant value @xmath7 , inferring further only the spin parameter of the black hole .
numerical simulations @xcite , @xcite pointed out , in the context of the standard disk model , that the angle of polarization shows large variations with energy , corresponding to emission of different energy photons from different regions of the disk .
thus , it was shown that the radiation emitted near the outer region of the accretion disk , at low energies , is horizontally polarized , parallel to the plane of the disk , with the angle of polarization , @xmath8 .
accordingly , the radiation originating from the innermost region of the disk , probing the strong gravitational field near rbh is deflected back to the disk and scattered to the distant observer as returning radiation .
the scattered radiation is vertically polarized , with the polarization angle perpendicular , @xmath9 , to the plane of the disk . in the context of rapidly spinning black holes ( @xmath10 ) , photons emitted between the outer and innermost regions of the accretion disk , in the transition region , as direct radiation ,
suffer large rotations of the polarization vector toward positive values @xmath11 , peaking at energies around 1 kev with diagonal polarization , @xmath12 .
consequently , for the returning radiation , at high energies , it is prescribed a rotation of the polarization vector in the clockwise direction , giving a negative angle of polarization , @xmath13 , that may reaches the anti - diagonal polarization , @xmath14 .
slowly rotating black holes ( @xmath15 ) and non - rotating bh ( @xmath16 ) have shown reduced rotations of polarization angle and at energies above 10 kev . the polarization of radiation can be characterized either by the stokes parameters or by , a less considered approach @xcite , @xcite , @xcite the 2x2 density matrix .
the two approaches are equivalent as it was demonstrated in @xcite .
we consider here the density matrix representation of the polarization of radiation emitted near rbh , having in mind to speculate the close analogy with the quantum mechanics density matrix @xcite and take advantages of the latest development of quantum information science .
we construct the density matrix via the coherency matrix starting from stokes parameters of the photons emitted near rotating black holes .
the coherency polarization matrix unitary contains all physical information about the polarization state of radiation .
the polarization matrix , considering the absence of circular polarization is given by : @xmath17 straightforward calculations considering @xcite , @xcite , lead to : @xmath18 the polarization matrix equals the density matrix of the photons source @xcite , @xcite , since density matrix is the unit trace scaling of the polarization matrix and @xmath19 , such as for the rest of the present paper we will refer to @xmath20 as the density matrix of polarization of photons . the density matrix in eq.(3 ) is hermitian , and expresses the two - level orthogonal system of x - ray photons polarization , considering the polarization as the degree of freedom .
the states of the x - ray photons span the hilbert ( @xmath21 ) two - dimensional space of polarization @xmath22 .
the density matrix in eq.(3 ) corresponds to the general quantum state of x - ray photons emitted by the accretion disk near the rotating black hole : @xmath23 where @xmath24 and @xmath25 are horizontal and vertical polarization , respectively .
we recognize in the eq.(4 ) the general normalized representation of the qubit state @xcite , @xcite .
the polarization degree of freedom encodes quantum information , as a single qubit .
quantum information processing based on optical polarization is a flourishing field of research encompassing optical quantum computation and quantum communication .
the state of photons is completely specified by knowing the polarization angle as it results from eq.(4 ) .
thus , the state of photons emitted at the outer edge of the disk , with the polarization angle @xmath8 , is @xmath26 , while the photons coming from the inner edge , with @xmath27 , are in the state @xmath28 .
an important observation we have to infer here is that at higher energies , in the inner regions of the disk , as we stated earlier , the angle of polarization rotates clockwise to negative values , @xmath13 , such that the states : @xmath29 should also be considered .
the polarization of photons emitted by the bh s accretion disk is determined by local processes in the source ( compton scattering ) and by general relativistic effects within the source and during the propagation toward the observer . in the present approach
we are interested in changes of the polarization generated by the strong general relativistic effects near rbh .
accordingly , we consider the polarization states of the photons , horizontal of vertical , due to the local processes in the source as input states , while the radiation beams escaping the bh s influence towards the distant observer are considered as output states . although , initially polarized horizontally or vertically , the strong general relativistic effects near rbh alter the state of photons by changing the polarization angle .
the general relativistic effects allow polarization angles in any direction .
more energetic photons coming from nearer the bh suffer larger general relativistic rotations with the angle of polarization taking values in the interval , @xmath30 $ ] . on the outer regions of the accretion disk , characterized by low energies , the polarization vector rotates to positive angles , @xmath11 , for the direct radiation .
nevertheless , near the transition region where the effects of gravitational field start to manifest , the angle of polarization may reaches @xmath31 , for a large portions of photons at energies around 1 kev , emitted close to rapidly spinning accreting black holes . the photons escaping the disk , initially horizontally polarized , as the input state , rotate the polarization angle during the propagation toward the distant observer , under the general relativistic effects into the output state characterized by diagonal polarization .
accordingly , measured by a distant observer , the output state of photons yields the diagonal polarization : @xmath32 close to the rbh , in the innermost regions of accretion disk , at high energies , polarization vector rotates in the clockwise direction , giving a negative angle of polarization , @xmath13 .
it is expected that initially vertically polarized photons , with high energies above the thermal peak , to shift the polarization state , as returning radiation , to the anti - diagonal polarization output state : @xmath33 the general relativistic effects , transform the initial horizontally polarized state of photons emitted near the transition region of the disk as : @xmath34 while the initial vertically polarized state as : @xmath35 the transformations of the states of photon s polarization in eq.(8 ) and eq.(9 ) are well known and intensively studied lately in optics quantum information theory @xcite,@xcite .
these quantum transformations realize the elementary hadamard gate @xcite , with the qubit encoded in the polarization degree of freedom of photons .
the polarization is not the only degree of freedom the photons emitted by the accretion disk near rbh may acquire . performing numerical simulations of the radiation originated from the accretion disk of rotating black holes , tamburini et al .
@xcite observed nontrivial oam generation and asymmetric spectra in terms of the lg - modes .
thus , the radiation beams from the accretion disk acquire independent azimuthal phase term @xmath36 possessing an orbital angular momentum of @xmath37 per photon , where @xmath37 is the integer topological charge .
it was pointed out @xcite that the oam modes @xmath37 the frame dragging effect nearby spinning black holes imprint to x - ray radiation emitted by accretion disk are determined by the bh spin parameter ( @xmath5 ) and the inclination angle of the disk ( @xmath6 ) towards a remote observer .
we maintain here for the inclination angle the same value @xmath7 as discussed in the case of x - ray polarization and analyze further only the bh spin parameter influence on the radiation coming from the accretion disk .
the spectrum of oam modes acquired by direct radiation emitted near rbh that have two different values for the spin parameter , a relatively moderate spinning bh ( @xmath15 ) to the left and a near - extreme rotating bh ( @xmath38 ) to the right is depicted in fig.1 .
oam modes spectrum of x - ray photons emitted by the accretion disk of two different spin parameter black holes : moderate spinning bh ( left ) and fast spinning bh ( right).,width=325 ] we should infer from fig.1 that the moderate spinning bhs ( @xmath15 ) are characterized by a narrow spectrum of oam modes dominated by zero oam mode , with ( @xmath39 ) . the oam spectrum reduces to zero lg mode ( @xmath39 ) for schwarzschild bh ( @xmath16 ) .
on the other hand , rapidly spinning black holes ( @xmath38 ) have the innermost region of the accretion disk closer to the black hole event horizon .
x - ray photons emitted by the accretion disk are influenced stronger by the gravitational field because of the bh proximal vicinity and tends to flip toward wider oam modes .
as it can be noticed from fig.1 the near - extreme spinning black hole has a broader oam spectrum with a significant reduction of zero lg mode .
it is worth being emphasized here that tamburini et al .
numerical simulations have shown that the major contribution to a widen oam spectrum for the photons emitted by accretion disk , comes from the inner regions closer to the bh s horizon , as returning radiation .
accordingly , photons flip the oam modes parity more rapidly as the radiation originates closer to the rbh . the photons emitted from the outer regions
the accretion disk possess narrow oam modes ( @xmath40 ) , while photons emitted from the innermost regions , experiencing stronger gravitational field near the rbh flip to wider oam modes ( @xmath41 ) .
although , numerical simulations conclude that the initially zero oam mode ( @xmath39 ) of photons flips to wider oam modes is independent of the photon frequency , x - ray photons emitted at higher energies , closer to the bh , are likely to acquire wider oam modes .
oam modes of x - ray photons are intensively sudied lately @xcite , @xcite .
our interest in the oam modes of photons emitted by accretion disks is motivated here by the capacity of photons oam modes to form a degree of freedom that can encode quantum information of a single qubit .
traditionally , in quantum information theory @xcite , @xcite , the oam spans an infinite dimensional hilbert space , and can theoretically take values in the interval , @xmath42 $ ] .
we restrict here the oam to a two - dimensional hilbert space by considering the parity of oam modes .
the even / odd ( @xmath43 ) grouping of oam parity spans a two - dimensional hilbert space and may encode a single qubit @xcite . we can write the initial state of photons originating from the rbh s accretion disk , considering the oam parity as the degree of freedom to encode a qubit is @xmath44 ( @xmath39 ) , hence the oam even parity , @xmath45 , as : @xmath46 it is likely that photons emitted al larger radii of the disk , probing low gravitational field , to escape the rbh influence toward a distant observer in the state in eq.(10 ) .
the photons emitted by the innermost regions of the disk , probing stronger twist of spacetime near rbh , flip the initial zero oam to a wider mode , the even oam parity , @xmath45 transforms to odd oam parity , @xmath47 .
@xmath48 changes in the oam modes to wider values , hence the flipping of the initial even / odd oam parity to a final odd / even oam parity , acts similarly to unitary transformation that implements a not gate @xcite .
the local processes at the source ( compton scattering ) polarize photons horizontally or vertically depending on the region of the disk they are emitted , and hence , depending on the photons energy . during the propagation to the distant observer , influenced by the general relativistic effects near the black holes , photons suffer rotation of polarization angle and flip of the oam parity to wider values . to completely determine the state of x - ray photons emitted by the accretion disk , both of the two degrees of freedom , polarization and parity of oam , must be considered .
accordingly , when escaping the rbh influence toward the distant observer , the states of photons should be expressed as a bipartite quantum system , @xmath49 , that spans a 4-dimensional hilbert space : @xmath50 the state of photons is represented by a tensor product that aggregates the two distinct degrees of freedom of photons , polarization and parity of oam , in a composite system @xcite @xcite .
we could consider the state of photons emitted by the accretion disk and reaching a distant observer in the general form : @xmath51 the initial parity oam mode of the photons polarized after the local processes at the source is @xmath52 , with @xmath39 , hence , maintaining the even / odd parity notation , @xmath53 . considering @xmath54 and @xmath55 in the forms we stated earlier , the general state of photons in eq.(13 ) yields : @xmath56 which , by dropping the tensor product notation , can be simplified to : @xmath57 it is likely that photons emitted by the accretion disk closer to the black hole , probing more the strong gravitational field , to flip the oam modes to wider values . for the x - ray emitted by the innermost regions
there is a flip of the oam mode to odd values , escaping toward a remote observer in the status : @xmath58 we noted here the state of photons , @xmath59 to maintain unitarily the notation in quantum information science .
we emphasize here that if the photons are emitted at larger radii it is expected to escape toward a distant observer , as direct radiation in the state in eq.(15 ) , while the photons coming from regions closer to the black hole , in the state in eq.(16 ) .
the radiation emitted by the accreting rbh near the innermost regions likely returns to the disk and is scattered toward the observer as returning radiation .
the returning radiation , strongly vertically polarized , probing more the general relativistic effects ( twisting of spacetime ) flips the oam modes to wider values .
the state of the photon in eq.(15 ) taken as initial state of the returning radiation , will suffer changes in the oam modes that affects the vertical polarization component . as result
, the parity of the photons vertical component flips from even to odd oam values and yields the output state in the form : @xmath60 we apply similar reasoning to returning radiation initially in the state eq.(16 ) and considering the shift of vertical component oam parity from odd to even , we find the output state : @xmath61 these shifts of states are considered in the quantum information theory @xcite , @xcite , @xcite as unitary transformations that implements c - not gates .
the x - ray photons that returning radiation consist of are carrying the quantum information associated to nonmaximally entangled states of the two degrees of freedom , polarization and oam parity , of x - ray photons .
this result is important since it proves that the curvature and twist of spacetime near rbhs could literally entangle the two degrees of freedom of single - photons . the incidence for such special output quantum states of photons emitted by accretion disks it is expected to be high , when measured by a distant observer for near extreme rotating black holes .
the nonmaximal entangled states in eq.(17 ) and eq.(18 ) are not the only output states the photons emitted by accretion disks can acquire traveling the distorted spacetime near rbh toward the distant observer .
we should mention here the presence of two other nonmaximal entangled states that photons near rbh can encode , as returning radiation . returning radiation , vertically polarized ,
have the polarization vector rotated to negative angle ( @xmath13 ) under the general relativistic effects closer to rbh . as returning radiation
, the oam of the vertical component of polarization flips to wider values changing the party from the initial even / odd state to the odd / even output state .
the photons emitted near the thermal peak , at larger radii , likely to possess initially even oam party flip to odd oam parity the vertical component of the polarization leading to the nonmaximally entangled state of the form : @xmath62 finally , we prescribe for photons coming from the inner edge of the disk , with the initial odd oam parity a flip to even oam parity of the vertical component of polarization : @xmath61 all four nonmaximally entangled states are widely considered lately in the quantum information science @xcite , @xcite .
it was advocated by schnitman and krolik @xcite that the vast majority of x - ray photons above the peak energy , 1kev in the case of rapidly spinning black holes ( @xmath10 ) , in the transition zone and the innermost regions of the accretion disk , acquire either diagonal or anti - diagonal polarization .
the photons in diagonal polarization states can escape the black hole influence toward the observer , as direct radiation , with the @xmath44 initial oam mode , @xmath39 , hence even parity @xmath45 , in the output state : @xmath63 if photons are emitted at larger radii , or @xmath64 in the case photons are emitted in the innermost regions of the disk , having the initial even oam parity flipped to final odd oam parity .
photons emitted above the thermal peak are likely deflected by the strong gravitational field near black holes and scattered by the accretion disk to a remote observer as returning radiation . during the propagation from the disk through the regions outside the rbh s influence ,
probing more the general relativistic effects , the vertically polarized component of returning radiation flips the initial even oam mode ( @xmath45 ) to the odd oam mode,(@xmath47 ) .
returning radiation originating from the regions near the transition region leaves the rbh influence , in the output state : @xmath65 while the photons originating from the innermost regions of the disk in the output state : @xmath66 these particular states of photons emitted by the accreting black holes are two bell states , the most explored states in the quantum information science @xcite , @xcite .
bell states express the maximal entangled states of the two degrees of freedom , polarization and oam , that photon acquire in the vicinity of rbhs .
we infer here that the above bell states are similar to a c - not applied to the input states determined by the diagonal states of photons .
we conclude mentioning that the other two bell states originate from the anti - diagonal polarization state of returning radiation . accordingly , applying similar reasoning as in the case of c - not gate implementation , we infer that the photons in anti - diagonal polarization state are escaping the rbh influence in the state : @xmath67 if the returning radiation originated from near transition region , above the thermal peak , or in the state : @xmath68 if the photons are emitted closer to the inner edge of the accretion disk .
we exemplify the creation of all four maximally entangled states of photons under the strong influence of the gravitational field near rapidly spinning black holes .
originating from regions closer to the rbh , at higher energies , x - ray photons are the perfect candidates to detect and measure the quantum information encoded in their degrees of freedom acquired by probing the bh strong gravitational field .
however , the processing of quantum information encoded in the photons degrees of freedom in the gravitational field of rotating black holes should be observed to all spectrum of light .
although , radiation with low energies likely may not encode bell states , since it originates from very large radii , probing less the gravitational field of rbh , it certainly encodes nonmaximaly entangled states .
we emphasize that the most probable quantum state to be detected and measured in low energy radiation , like visible light , is : @xmath69 since it originates from the larger radii of the accretion disk .
the incidence for the other three states involving wider oam modes , negative angle of polarization and diagonal states of the photons polarization may not be present in the quantum measurements of visible light coming from accreting black holes .
detection and measurement of bell states encoded in the degrees of freedom of radiation should be focused on very energetic photons , with frequencies in the band of x - ray or gamma - ray , originating from the inner edge of the accretion disk , closer to the rbh .
probing more the general relativistic effects in the vicinity of the black hole , x - ray photons may encode wider oam values , may acquire larger rotations of polarization vector and negative angle of polarization .
recalling the oam spectrum in fig.1 and considering a near - extreme spinning black holes we should notice here that the most probable values for the even oam modes are @xmath70 and @xmath71 , and the odd oam modes @xmath72 . under these circumstances we emphasize that the bell states which could be measured as being encoded in x - rays are : @xmath73 and , @xmath74 also from fig.1 , we could infer that the incidence for @xmath75 states should be substantially higher for rapidly spinning black holes while the probability to measure the @xmath76 states should increase for slow rotating black holes .
x - ray photons emitted near bh s event horizon may experience more than one bending and scattering as the bh spins faster .
near - extreme spinning black holes may account for bell states having wider oam modes , like : @xmath77 measurement of bell states encoded by various degrees of freedom of photons is very common technique in quantum optics @xcite , @xcite .
setups meant to measure the bell states are common in every laboratory working in quantum information processing of qubits encoded in visible light .
however , the processing of quantum information encoded in the degrees of freedom of photons with very short wavelengths , like x - ray or gamma - ray , is hard to implement since it necessitate sophisticate equipment required by their high energy . still processing quantum information with x - ray photons
is a flourishing field of research lately .
the extreme robustness of the quantum information encoded by the two degrees of freedom of x - ray photons , polarization and oam , attracted a large interest @xcite , @xcite of researchers in this particular field of quantum information processing .
we conclude by expressing our reasonable expectations that the near future space missions around the world would contain apparatus capable to detect bell states in x - ray photons originating from spinning black holes across the universe .
we have shown in the present paper that strong general relativistic effects near rotating black holes may process quantum information encoded in the polarization and oam degrees of freedom of x - ray photons emitted by the accretion disk .
the strong gravitational field distorts the spacetime nearby spinning black holes , forcing the spacetime around them to curve and twist . extreme curvature and twist of spacetime
affect the photons emitted by the accretion disk .
propagating throughout the particular shape of spacetime near rbh , the photons polarization angle suffers large rotations and acquire a spectrum of wide oam modes . considering a density matrix approach constructed from the stokes parameter via the coherency matrix
, we have shown that these two degrees of freedom polarization and oam - the photons acquired near spinning black holes , encode two qubits , characterize the states of photons .
we emphasized here that the particular shape of spacetime near spinning black holes implements unitary transformations over the two degrees of freedom of photons which act similarly to the elementary quantum gates .
transformations suffered by the photons due to general relativistic effects nearby rbh are similar to quantum information processing in quantum optics .
we explained the setups of spacetime required to perform unitary transformations onto single - photons in order to implement quantum gates .
we exemplified these quantum processes by some elementary quantum gates implementation .
the elementary hadamard gate , not gate , and cnot gate were constructed with the help of photons , considering only the general relativistic effects nearby fast rotating black holes . | the curvature and twisting of spacetime rotate the angle of polarization and imprint orbital angular momentum to photons emitted by the accretion disk near rotating black holes . considering polarization and orbital angular momentum as two degrees of freedom of single - photons that can encode quantum information , we emphasize that the particular shape of spacetime around rotating black holes implements quantum gates and simple quantum circuits .
consequently , we demonstrate the implementation of some elementary quantum gates , like hadamard or c - not , and simple quantum circuits , like bell states , with photons in the presence of spinning black holes .
detection and measurement of quantum information encoded in photons emitted in the accretion disk around rotating black holes may be performed by actual quantum information technology .
pacs numbers : : 04.70.dy , 03.67.bg , 42.50.ex , 95.30.gv |
in these years there has been introduced and intensively investigated an operation called _ hairpin completion _ in formal language theory , inspired by intra molecular phenomena in molecular biology .
a hairpin structure is well - known as one of the most popular secondary structures for a single stranded dna ( or rna ) molecule to form , with the help of so - called _ watson - crick complementarity _ and _ annealing _ , under a certain biochemical condition in a solution .
this paper continues research directed by a series of works started in @xcite where the hairpin completion operation was introduced , followed by several other related papers ( @xcite ) , where both the hairpin completion and its inverse operation ( the hairpin reduction ) were investigated . inspired by threefold motivations
, we will introduce the notion of _ hairpin incompletion _ in this paper .
firstly , the hairpin incompletion is a natural extension of the notion of _ bounded hairpin completion _ introduced and studied in @xcite which is a restricted variant of the hairpin completion with the property that the length of the prefix ( suffix ) prolongation is constantly bounded .
thus , the bounded hairpin completion involves the lengthening of prefix ( suffix ) with a constant length of the strand at the end , which implies that the resulting strand always bears a specific property that its prefix and suffix _ always form complementary sub - strands _ of a certain constant length .
in contrast , our notion of hairpin incompletion can produce a resulting strand with more complexity , due to the nature of its prolongation , which will be formally explained later .
secondly , the hairpin incompletion is also regarded as a restricted variant of the notion of _ hairpin lengthening _ recently introduced in @xcite which is an extension of the ( original ) notion of the hairpin completion .
more specifically , the hairpin lengthening concerns the prolongation of a strand that allows to stop itself at any position in the process of completing a hairpin structure . from the practical and molecular implementation point of view , here we are interested in the case where the prolongation in the hairpin lengthening is bounded by a constant , which leads to our notion of the hairpin incompletion . in this respect , one may take the hairpin incompletion as the _ bounded _ variant of the hairpin lengthening .
thirdly , the hairpin incompletion can provide a purely formal framework that exactly models a bio - molecular technique called _ whiplash pcr _ that has nowadays been recognized as a promising experimental technique and has been proposed in an ingenious paper @xcite by hagiya et al .
they developed an experimental technique called polymerization stop and theoretically showed in terms of thermal cycling how dna molecules can solve the learning problem of @xmath0-formulas ( i.e. , boolean formulas with each variable appearing only once ) from given data .
suppose that a dna sequence is designed as given in ( a ) of figure 1 , where a sequence of transition ( program ) is delimited by a special sequence ( called _ stopper sequence _ ) and @xmath1 and its reversal complementarity @xmath2 may hybridize , leading to a hairpin structure ( b ) .
then , the head @xmath2 ( current state ) is extended by polymerization ( with a primer @xmath2 and a template @xmath3 ) up to @xmath4 , where the stopper sequence is specifically designed to act as the stopper . in this way
, this cycle can execute one process of state transition and be repeatedly performed .
following the work of @xcite , sakamoto et al .
has shown how some np - complete problems can be solved with whiplash pcr ( or whiplash machines ) ( @xcite ) . recently , komiya et al .
has demonstrated the applicability of whiplash pcr to the experimental validation of signal dependent operation ( @xcite ) . ;
( c ) polymerization extension of @xmath3 ; ( d ) simulation of one state transition . ]
the paper is organized as follows . after providing the definitions of the basic concepts used in the paper , we define the central notion of _ hairpin incompletion _ ( as an extension of the bounded hairpin completion and also as a bounded variant of the hairpin lengthening ) in section 2 .
we first show in section 3 that any family of languages with a certain closure properties is closed under the hairpin incompletion .
we then consider the case of applying the _ iterated _ hairpin incompletion operations , and show that every afl is closed under the iterated one - sided hairpin incompletion .
this result is further extended to the general case of the iterated hairpin incompletion , and it is shown that any family of languages including all linear languages and with a certain closure properties is also closed under the iterated hairpin incompletion , and as a corollary that the family of context - free languages is closed under the iterated hairpin incompletion , followed by a brief discussion with concluding remarks in section 4 .
this paper assumes that the reader is familiar with the basic notions of formal language theory @xcite .
in particular , for the notions of abstract family of languages , we refer to @xcite . for an alphabet @xmath5
, @xmath6 is the set of all finite - length strings of symbols from @xmath5 , and @xmath7 is the empty string . while @xmath8 denotes @xmath9 . for @xmath10
, @xmath11 is the length of @xmath12 . for @xmath13 , we define @xmath14 .
note that for a set @xmath15 , @xmath16 denotes the cardinality of @xmath15 . for @xmath13 ,
let @xmath17 and @xmath18 be the prefix and the suffix of a word @xmath12 of length @xmath19 , respectively . for @xmath13 , we define @xmath20 and @xmath21 . for @xmath22 ,
let @xmath23 be the set of infixes of @xmath12 of length @xmath19 .
if @xmath24 , then @xmath17 , @xmath18 and @xmath25 are all undefined .
( note that for @xmath26 , @xmath17 and @xmath18 are elements in @xmath25 . ) by @xmath27 ( @xmath28 ) we simply denote @xmath29 ( @xmath30 ) , i.e. , the concatenation of @xmath12 with a language @xmath31 .
the left derivative of a language @xmath31 with a word @xmath12 is defined by @xmath32 . for a word @xmath33 , @xmath34 is the palindrome of @xmath12 , that is , @xmath35 . a morphism @xmath36 such that @xmath37 for all @xmath38 is called a _ coding _ , and it is called a _ weak coding _ if @xmath39 for all @xmath38 . an _ involution _ over @xmath5 is a bijection @xmath40 : @xmath41 such that @xmath42 . in particular
, an involution @xmath40 over @xmath5 such that @xmath43 for all @xmath44 is called _ watson - crick involution _ ( in molecular computing theory ) in a metaphorical sense of dna complementarity . in this paper
, we fix an involution @xmath45 over @xmath5 such that @xmath46 for @xmath38 and extend it to @xmath6 in the usual way .
note that for all @xmath47 , it holds that @xmath48 . for the original definitions of the ( unbounded ) @xmath19-hairpin completion ,
the reader is referred to precedent papers ( for example , @xcite ) .
a variant of the notion called bounded @xmath19-hairpin completion and its modified operation were introduced and investigated in @xcite and @xcite , respectively , while a recent paper @xcite introduces and studies an extended version of the hairpin completion , called hairpin lengthening . in this paper , we are interested in a new variant of both the bounded @xmath19-hairpin completion and the hairpin lengthening which will be introduced as follows .
let @xmath49 .
for any @xmath10 , we define the _
@xmath50-bounded @xmath19-hairpin incompletion _ of @xmath12 , denoted by @xmath51 , as follows : @xmath52 where @xmath53 ( or @xmath54 ) is called _
@xmath50-bounded right ( or left ) @xmath19-hairpin incompletion_. moreover , @xmath50-bounded right ( or left ) @xmath19-hairpin incompletion is also called @xmath50-bounded _ one - sided _
@xmath19-hairpin incompletion .
( see figure [ hi1 ] , for pictorial illustration of the operations @xmath53 and @xmath54 . ) thus , from a mathematical viewpoint , we consider the hairpin incompletion operations whose prolongations take place at both ends in a hypothetical ( and ideal ) molecular biological setting
. * note . * for @xmath10 not satisfying the condition to apply the @xmath50-bounded @xmath19-hairpin incompletion , here we assume @xmath55 . + the iterated version of the @xmath50-bounded right @xmath19-hairpin incompletion is defined in a usual manner :
@xmath56 the `` left '' counterpart of the iterated version of this operation is defined in an obvious and similar manner and is denoted by @xmath57 .
further , the iterated version of the @xmath50-bounded @xmath19-hairpin incompletion operation is defined in a similar manner as follows : @xmath58 finally , the iterated version of the @xmath50-bounded ( right or left ) @xmath19-hairpin incompletion operation is naturally extended to languages as follows : @xmath59 \mbox{and \ } hi^{*}_{m , k } ( l ) & = \bigcup_{w \in l } hi^{*}_{m , k } ( w).\\ \end{array } \right.\ ] ] -bounded right @xmath19-hairpin incompletion operation ; ( b ) @xmath50-bounded left @xmath19-hairpin incompletion operation , where @xmath60 and @xmath61 . ]
note that the bounded hairpin _
in_completion in this paper is an extension of bounded hairpin completion in the sense that @xmath51 is exactly the same as @xmath62 in @xcite when the prefix ( suffix ) @xmath63 of @xmath12 is empty .
further , the hairpin lengthening @xmath64 in @xcite is corresponding to the union of all @xmath51 , where @xmath50 is arbitrary , in this paper .
as is expected from the definitions , non - iterated bounded hairpin incompletion operation behaves as the bounded hairpin completion operation does .
let @xmath65 be a class of languages and @xmath49 .
if @xmath65 is closed under @xmath66-mappings , @xmath65 is also closed under @xmath50-bounded @xmath19-hairpin incompletion . for any @xmath67 , consider a generalized sequential machine ( _ gsm _ ) @xmath68 which adds a suffix ( or prefix ) @xmath69 of length at most @xmath50 to the input word @xmath12 if @xmath12 is of the form @xmath70 ( or @xmath71 ) with @xmath72 , @xmath73 .
it is easily shown that this @xmath66 simulates @xmath50-bounded @xmath19-hairpin incompletion @xmath51 . since every trio - free morphisms , inverse morphisms and intersection with regular languages . ]
is closed under _
gsm _ mapping ( @xcite ) , the following is straightforwardly obtained .
every @xmath74 is closed under @xmath50-bounded @xmath19-hairpin incompletion for any @xmath67 .
this result extends the corresponding one ( proposition 1 ) in @xcite , while it is contrast to the results ( propositions 1 and 2 ) in @xcite . in this section
, we consider the closure property of iterated bounded one - sided hairpin incompletion .
especially , we show that every afl is closed under this operation . to this aim
, we start by preparing some notions required in the proof of the main result . a key idea of the proof is to construct a certain equivalence relation which is right invariant and of finite index .
first , we consider the iterated @xmath50-bounded right @xmath19-hairpin incompletion operation : @xmath75 . given @xmath49 and a word @xmath76 ,
we define : @xmath77 [ def - dmk ] we also define a binary relation @xmath78 as follows : for @xmath79 , @xmath80 intuitively , a pair @xmath81 in @xmath82 implies that it is a candidate of @xmath83 where @xmath1 and @xmath3 satisfy the conditions to apply @xmath50-bounded right @xmath19-hairpin incompletion to @xmath12 , producing a word in @xmath84 . from the definition
, it holds that @xmath85 is in @xmath82 with @xmath72 if and only if @xmath86 is in @xmath87 .
the binary relation @xmath78 is clearly an equivalence relation and of finite index , that is , the number of equivalence classes @xmath88 is finite .
moreover , the following claim holds .
the equivalence relation @xmath78 is right invariant , that is , for @xmath89 , @xmath90 implies that for any @xmath91 , @xmath92 .
[ cla - ri ] we prove it by induction on the length of @xmath93 . if @xmath94 , then the claim trivially holds .
assume that @xmath90 implies that @xmath92 with @xmath95 .
then , it suffices to show that for any @xmath38 , @xmath96 .
we observe that @xmath97 is constructed from _ only _ @xmath98 as follows : @xmath99 @xmath100 note that if @xmath101 , then @xmath102 for some @xmath103 , so that @xmath104 can be rewritten by @xmath105 .
therefore , @xmath106 is contained in @xmath107 . from the induction hypothesis , since @xmath108
, we can construct @xmath109 from _ only _ @xmath98 in the same way .
thus , it holds that @xmath96 .
we first show that the language obtained by applying iterated bounded right hairpin incompletion to a singleton is regular .
let s consider the equivalence classes : @xmath111 , [ w_2 ] , \dots , [ w_t ] \
, | \ , w_i \in v^{\ge 2k } , 1 \le i \le t \},\ ] ] where @xmath112 is the representative of @xmath113 $ ] . for @xmath76 ,
the regular grammar @xmath114 is constructed as follows : @xmath115 we need the following two claims .
let @xmath12 be in @xmath116 , and @xmath117 .
then , for @xmath118 , if a derivation of @xmath119 is of the form @xmath120 for some @xmath91 , then it holds @xmath121 .
[ cla - sf ] the proof is by induction on @xmath122 .
if @xmath123 , then @xmath124 and from the manner of constructing @xmath125 , it holds @xmath126 , thus , the claim holds .
assume that the claim holds for @xmath127 and consider a derivation of the form @xmath128 for some @xmath129 , @xmath130 .
from the assumption and the form of @xmath125 , it holds that @xmath121 and @xmath131 . by claim [ cla - ri ] , we obtain that @xmath132 . for @xmath118 and @xmath91 , there exists a derivation of @xmath119 of the form @xmath133 if and only if @xmath134 is in @xmath135 . [ cla - sf2 ] the proof is by induction on @xmath122 .
if @xmath123 , it obviously holds that @xmath136 if and only if @xmath12 is in @xmath137 .
assume that the claim holds for @xmath122 and consider the case for @xmath138 .
+ ( if part ) let @xmath139 .
then there exists @xmath140 such that @xmath141 with @xmath142 .
from the definition of @xmath143 , @xmath85 is in @xmath144 with @xmath72 . from the induction hypothesis and claim [ cla - sf ] ,
there exists a derivation : @xmath145 with @xmath121 .
since @xmath85 is in @xmath146 , there exists the derivation @xmath147 for some @xmath129 .
+ ( only if part ) if there exists the derivation @xmath148 @xmath149 for some @xmath129 , it holds that @xmath121 from claim [ cla - sf ] .
moreover , from the form of @xmath125 , there exists @xmath150 .
hence , @xmath151 is in @xmath152 . from the induction hypothesis ,
@xmath142 so that @xmath153 .
it follows from the claim that the language obtained by applying iterated bounded right hairpin incompletion to a singleton is regular . for any word @xmath10 and @xmath49 , a language @xmath154 is regular .
[ lem - single ] in the case of @xmath155 , from the definition , @xmath156 is regular .
for @xmath76 it follows from claim [ cla - sf2 ] that there exists a derivation of @xmath119 which derives a terminal string @xmath157 if and only if @xmath158 .
thus , we have that @xmath159 which is regular . in order to show more general results , we need to prove the claims regarding the language @xmath154 . for @xmath79 and @xmath118 , if @xmath90 then there exists a finite language @xmath160 such that @xmath161 and @xmath162 .
[ cla - fin ] the proof is by induction on @xmath122 .
if @xmath123 , it obviously holds that @xmath163 and @xmath164 , where @xmath165 .
we assume that the claim holds for up to @xmath122 .
let @xmath161 and @xmath162 for some finite language @xmath166 .
for any @xmath167 , it holds that @xmath92 from claim [ cla - ri ] .
hence , from the induction hypothesis , there exists a finite language @xmath168 such that @xmath169 therefore , it holds that @xmath170 where @xmath171 . for @xmath79 , if @xmath90 then there exists a regular language @xmath172 such that @xmath173 and @xmath174 .
[ cla - reg ] from claim [ cla - fin ] , if @xmath90 , then there exists a sequence of finite languages : @xmath175 where @xmath176 , with the property that for @xmath118 , @xmath177 and @xmath178 .
then it holds that @xmath179 let @xmath180 .
then , we obtain @xmath173 and @xmath174 . recall that @xmath181 and @xmath182 are regular from lemma [ lem - single ] . the class of regular languages is closed under left derivative , so that @xmath183 is also regular .
we are now in a position to show the main theorem in this section .
it is shown that iterated bounded one - sided hairpin incompletion can be simulated by several basic language operations , which leads to the following theorem .
let @xmath65 be a class of languages and @xmath49 .
if @xmath65 is closed under intersection with regular languages , concatenation with regular languages and finite union , then @xmath65 is also closed under iterated @xmath50-bounded right ( left ) @xmath19-hairpin incompletion .
let @xmath184 be the language over @xmath5 .
we can write @xmath185 where @xmath186 note that @xmath187 . since the number of the elements in @xmath188 is finite from the definition of @xmath78 , we can set @xmath189 , [ w_2 ] , \dots , [ w_s ] \ , | \ , w_i \in l_1 \text { for } 1 \le
i \le s \}$ ] for some @xmath190 . from the way of construction of @xmath191
, it holds that
for @xmath192 , @xmath193 = l_1 \cap ( \bigcap_{(xy , z ) \in c_{m , k}(w_i ) } v^ * xy v^ * z ) \cap ( \bigcap_{0 \le j \le m } v^ * \cdot suf_{j+k-1}(w_i ) ) , \ ] ] for @xmath194 , since all words in @xmath113 $ ] are equivalent , it follows from claim [ cla - reg ] that there exists regular language @xmath195 such that @xmath196 ) = [ w_i ] r_i$ ] .
moreover , it holds that @xmath197 r_i$ ] .
thus , @xmath198 can be constructed from @xmath31 by intersection with regular languages , concatenation with regular languages and finite union , which completes the proof . as a corollary
, we immediately obtain the following .
every @xmath199 is closed under iterated @xmath50-bounded right ( left ) @xmath19-hairpin incompletion for any @xmath67 .
it is known in @xcite that there exists _ no universal _ regular grammar @xmath200 with the property that for any regular grammar @xmath201 , there exists a coding @xmath202 of @xmath201 such that @xmath203 .
this can be strengthened in the form that no morphism @xmath204 can help to satisfy the equation @xmath205 . in this context
, the next lemma shows that the bounded hairpin incompletion operation can play a role of the universal - like grammar for all regular languages .
a language @xmath206 is regular if and only if there exists a word @xmath207 and a weak coding @xmath208 such that @xmath209 , where @xmath210 and @xmath211 .
[ renma2 ] ( if part ) this clearly holds , because the class of the regular languages is closed under iterated bounded right hairpin incompletion , intersection and weak codings .
( only if part ) for a regular grammar @xmath212 , we construct @xmath213,@xmath214 , @xmath215 and @xmath208 as follows : * @xmath216 \ , | \ , a \in v , x \in n \cup \ { \lambda \ } \ } \cup \ { \overline{[a , x ] } \ , | \ , a \in v , x \in n \ } \cup \ { \ # , \overline { \ # } \}$ ] , * @xmath217 \ , | \ , a \in v \}$ ] , * @xmath218 } \
, \overline{[b , x_i ] } ) \cdot [ \lambda , s]$ ] , * @xmath219 for @xmath220 \in \ { [ a , x ] \ , | \ , a \in v , x \in n \cup \ { \lambda \ } \}$ ] , @xmath221 otherwise .
note that for any @xmath118 and @xmath222 with @xmath223 , if @xmath224 , then @xmath3 is the symbol just right of @xmath225 .
then , from the way of construction of @xmath12 , it holds that there exists a derivation of @xmath201 , @xmath226 if and only if @xmath227 } \ , \overline{[b , x_i ] } ) [ \lambda , s ] [ a_1 , x_1 ] [ a_2 , x_2 ] \dots [ a_{n-1 } , x_{n-1 } ] [ a_n , \lambda]\ ] ] is in @xmath228 , which can be shown by induction on @xmath122 . by applying @xmath204
, we obtain @xmath229 .
we note that theorem 3 in @xcite proves the _ only if _ part of this lemma for the iterated ( unbounded ) hairpin lengthening .
thus , lemma [ renma2 ] complements the result for the case of bounded hairpin lengthening . in this section
, we consider the closure property of iterated bounded hairpin incompletion .
for the ( unbounded ) hairpin lengthening operation , the paper @xcite has proved that the family of context - free languages is closed under iterated hairpin lengthening in theorem 4 .
we will show that the result also holds for the case of iterated bounded hairpin lengthening , in a more general setting of afl - like formulation .
the proof is based on the similar idea to the previous section and claim [ cla - ri ] , [ cla - sf ] , [ cla - sf2 ] are corresponding to claim [ cla - rili ] , [ cla - sfcf ] , [ cla - sfcf2 ] ( below ) , respectively . in order to consider both - sided hairpin incompletion
, we modify the equivalence relation . for @xmath49 and the word @xmath76 , @xmath230 , @xmath231 and @xmath232
are defined by @xmath233 , where @xmath234 is the relation defined in definition [ def - dmk ] .
the binary relation @xmath235 is defined as @xmath236 if @xmath237 for @xmath79 . the binary relation @xmath235 is clearly an equivalence relation and of finite index .
note that @xmath238 and @xmath239 are symmetrically defined .
we show that the equivalence relation @xmath235 is right invariant and left invariant .
the equivalence relation @xmath235 is right invariant and left invariant , that is , for @xmath79 , if @xmath236 then for any @xmath240 , @xmath92 and @xmath241 holds .
[ cla - rili ] we firstly show that for @xmath91 , @xmath92 .
the proof is by induction on the length of @xmath93 . if @xmath242 , it clearly holds .
we assume that the claim holds for @xmath122 , i.e. , @xmath243 with @xmath244 .
let @xmath245 be a symbol in @xmath5 .
[ proof of @xmath96 ] it can be shown by the same way as claim [ cla - ri ] . [ proof of @xmath247 we construct @xmath248 from _ only _ @xmath249 as follows : @xmath250 @xmath251 note that for @xmath252 , @xmath253 with @xmath254 and some @xmath255 , @xmath104 can be represented as @xmath256 .
hence , @xmath257 is in @xmath258 .
since @xmath259 , we can construct @xmath260 from _ only _ @xmath249 in the same way .
therefore , it holds that @xmath261 . from @xmath96 and @xmath261 ,
we eventually get @xmath262 . for the left invariance of @xmath235
, we can show in the symmetrical manner . for the proof of theorem [ ibhi ] ( below ) regarding @xmath50-bounded @xmath19-hairpin incompletion
, we need to construct a linear grammar . for @xmath206 ,
let @xmath264 for some @xmath265 and @xmath266 , [ w_2 ] , \dots , [ w_s ] \}$ ] for some @xmath267 , where @xmath112 is the representative of @xmath113 $ ] .
a linear grammar @xmath268 is constructed as follows : @xmath269 we set @xmath270 and @xmath271 .
let @xmath272 and @xmath273 .
for @xmath118 , if a derivation of @xmath274 is of the form @xmath275 , then for any @xmath276 , it holds that @xmath277 , where for each @xmath278 , @xmath279 , @xmath280 , one of @xmath281 and @xmath282 is @xmath7 and the other is not @xmath7 .
[ cla - sfcf ] the proof is by induction on @xmath122 .
if @xmath123 , then @xmath124 and from the manner of constructing @xmath125 , for any @xmath276 , it holds that @xmath283 , thus the claim holds .
assume that the claim holds for @xmath127 and consider a derivation of the form @xmath284 for some @xmath285 , @xmath286 @xmath287 .
from the assumption and the form of @xmath125 , for any @xmath276 , it holds that @xmath288 @xmath289 and @xmath290 . by claim [ cla - rili ]
, we obtain that @xmath291 a word @xmath292 is generated by @xmath274 if and only if for any @xmath293 , @xmath294 is in @xmath295 , where for each @xmath278 , @xmath279 , @xmath280 , one of @xmath281 and @xmath282 is @xmath7 and the other is not @xmath7 .
[ cla - sfcf2 ] the proof is by induction on @xmath122 .
if @xmath123 , it obviously holds that @xmath296 if and only if for any @xmath297 , @xmath12 is in @xmath298 .
assume that the claim holds for @xmath122 and consider the case for @xmath138 .
+ ( if part ) let @xmath299 , where for each @xmath300 , @xmath279 , @xmath280 , one of @xmath281 and @xmath282 is @xmath7 and the other is not @xmath7 . from the definition of @xmath143 and @xmath301 ,
either @xmath302 is in @xmath303 or @xmath304 is in @xmath305 with @xmath72 . from the induction hypothesis and claim [ cla - sfcf ]
, there exists a derivation : @xmath306 with @xmath307 .
therefore , it holds that either @xmath308 or @xmath309 , from which there exists the derivation either + @xmath310\end{aligned}\ ] ] or + @xmath311\end{aligned}\ ] ] for some @xmath285 .
+ ( only if part ) consider the case where there exists a derivation @xmath312 for some @xmath285 .
then , it holds that for any @xmath276 , @xmath307 from claim [ cla - sfcf ] .
moreover , from the way of construction of @xmath125 , there exists @xmath313 .
hence , @xmath314 is in @xmath315 . from the induction hypothesis ,
@xmath316 so that @xmath317 . for the other case , there exists a derivation @xmath318 @xmath319 for some @xmath285
. then we can show in a similar way that for any @xmath276 , @xmath320 . in order to prove the next result
, we need a language operation called _ circular permutation _ @xmath321 which maps every word in the set of all its circular permutations and every language in the set of all circular permutations of its words .
the proof is due to an idea similar to the one in @xcite .
let @xmath65 be a class of languages which includes all linear languages and let @xmath49 .
if @xmath65 is closed under circular permutation , left derivative and substitution , then @xmath65 is also closed under iterated @xmath50-bounded @xmath19-hairpin incompletion .
[ ibhi ] recall the construction of the linear grammar @xmath274 .
let @xmath31 be in @xmath65 and @xmath322 be a substitution over @xmath323 defined by @xmath324 for @xmath325 and @xmath326 otherwise . from claim [ cla - sfcf2 ]
, it holds that @xmath327 where @xmath328 .
hence , it is easily seen that @xmath329 . since the family of context - free languages meets all of preconditions in theorem [ ibhi ] , the following corollary holds .
the family of context - free languages is closed under iterated @xmath50-bounded @xmath19-hairpin incompletion for any @xmath67 .
in many works on dna - based computing and the related areas , dna hairpin structures have numerous applications to develop novel computing mechanisms in molecular computing . among others ,
these molecules of hairpin formation called whiplash pcr have been successfully employed as the basic feature of new computational models to solve an instance of the 3-sat problem ( @xcite ) , to execute ( and simulate ) state transition systems ( @xcite ) , to explore the feasibility of parallel computing for solving dhpp ( @xcite ) , and so forth . on the other hand ,
different types of hairpin and hairpin - free languages are defined in @xcite and more recently in @xcite , where they are studied from a language theoretical point of view .
we have proposed a new variant of hairpin completion called hairpin incompletion , and investigated its closure properties of the language families .
the hairpin incompletion is in fact a bounded variant of the hairpin lengthening in @xcite where not only closure properties of language families but also the algorithmic aspects of the hairpin lengthening operations are investigated .
the hairpin incompletion is also an extended version of the bounded hairpin completion recently studied in @xcite that has been more recently followed up by slightly modified operations in @xcite where two open problems from @xcite have been solved .
we have shown that every afl is closed under the iterated one - sided hairpin incompletion , and therefore , the family of regular languages is closed under the operation .
further , it has been shown that the family of context - free languages is closed under the iterated hairpin incompletion .
these complement some of the corresponding results for ( unbounded ) hairpin lengthening operations in @xcite .
moreover , since the hairpin incompletion nicely models a bio - molecular technique ( whiplash pcr ) , the obtained results in this paper may provide new insight into the computational analysis of the experimental technique .
m. hagiya , m.arita , d.kiga , k.sakamoto , and s.yokoyama . towards parallel evaluation and learning of boolean @xmath0-formulas with molecules , _
dna based computers iii _
( rubin , h. and wood , d .. eds . ) , _ dimacs series in discrete mathematics _ , vol .
57 - 72 , 2000 .
m. ito , p. leupold , and v. mitrana . bounded hairpin completion , in : lata 09 : _ proceedings of the 3rd international conference on language and automata theory and applications _
, springer - verlag , berlin , heidelberg , pp.434 - 445 , 2009 .
k.komiya , k.sakamoto , a.kameda , m.yamamoto , a.ohuchi , d.kiga , s.yokoyama and m. hagiya , m. : dna polymerase programmed with a hairpin dna incorporates a multiple - instruction architecture into molecular computing , _ biosystems _ , vol . 83 , pp .
18 - 25 , 2006 .
k.komiya and j.a . rose .
experimental validation of signal dependent operation in whiplash pcr , _ dna computing .
14th international workshop on dna - based computers _
( goel , a. and simmel , f.c . , eds . ) , _ lecture notes in computer science _ , vol.5347 , pp.1 - 10 , 2009
. s. kopecki . on the iterated hairpin completion . in y. gao ,
h. lu , s. seki , and s. yu ( editors ) , _ 14th developments in language theory _ , lncs 6224 , springer , pp.438 - 439 , 2010 .
also , in arxiv:1010.3640v2 [ cs.fl ] . | hairpin completion and its variant called bounded hairpin completion are operations on formal languages , inspired by a hairpin formation in molecular biology .
another variant called hairpin lengthening has been recently introduced and studied on the closure properties and algorithmic problems concerning several families of languages . in this paper
, we introduce a new operation of this kind , called _ hairpin incompletion _ which is not only an extension of bounded hairpin completion , but also a restricted ( bounded ) variant of hairpin lengthening .
further , the hairpin incompletion operation provides a formal language theoretic framework that models a bio - molecular technique nowadays known as whiplash pcr .
we study the closure properties of language families under both the operation and its iterated version .
we show that a family of languages closed under intersection with regular sets , concatenation with regular sets , and finite union is closed under one - sided iterated hairpin incompletion , and that a family of languages containing all linear languages and closed under circular permutation , left derivative and substitution is also closed under iterated hairpin incompletion . |
perovskite type manganites such as r@xmath2a@xmath3mno@xmath4 ( r = rare earth ion and a = alkaline earth ion ) have been extensively investigated due their properties of colossal magnetoresistance ( cmr ) @xcite .
the cmr effects originates from a competition between a ferromagnetic metallic ( fmm ) state and an antiferromagnetism insulating ( afmi ) state .
the appearance of the fmm state is explained by the double exchange ( de ) mechanism @xcite , whereas the afmi state originates from the jahn - teller ( jt ) distortions @xcite , leading in the case of small a - site cations to the charge / orbital ordered ( co / oo ) phenomenon @xcite .
the efforts made for understanding the magnetotransport properties of manganite thin films @xcite have shown that the physical properties are often different from the bulk materials .
the main reason deals with the strains included the substrate - induced strains which modify the lattice parameters of the film and change the properties as shown recently for @xmath5 @xcite , @xmath6 @xcite and @xmath7 @xcite . in manganite
thin films , the strains are susceptible to modify the mn - o - mn angles and mn - o distances .
consequently , this affects the bandwidth and also the jt distortions , so that de and co / oo phenomena may be drastically influenced .
based on these considerations , the strain effects should decrease as the thickness increases and it should be possible to establish relationships between the structural distortions and the transport properties , by studying the properties of the films with various thicknesses . starting from @xmath7 films where a spectacular cmr effect was previously evidenced @xcite ,
compared to the bulk @xcite , we have carried out a systematic study of films with different thickness grown in situ using the pulsed laser deposition technique ( pld ) on both ( 100)-srtio@xmath8 and ( 100)-laalo@xmath4 substrates . in this paper ,
a clear correlation between the structural distortions and the transport properties ( insulator - to - metal @xmath9 transition under a magnetic field , co / oo @xmath10 transition ) .
we suggested that for small films thickness the elongation of the mn - o bonds prevails , reinforcing the jt distortion and consequently the co / oo state , whereas for large thickness the flattering of the mn - o - mn bond angles prevails , increasing the bandwidth favoring the insulator - to - metal transition to the detriment of the co / oo state .
films with different thicknesses were deposited on single crystal substrates of ( 100)-srtio@xmath4 ( sto ) and ( 100)-laalo@xmath4 ( lao ) from a dense targets of nd@xmath11ca@xmath11mno@xmath4 using the pld technique .
detailed of the process can be found elsewhere @xcite .
the structural study was done by x - ray diffraction ( xrd ) using a seifert xrd 3000p for the @xmath12 - 2@xmath12 scans and a phillips mrd xpert for the in - plane measurements ( cu k@xmath13 , @xmath14=1.5406 ) .
the in - plane lattice parameters were obtained from the ( 103)@xmath15 reflection ( where c refers to the ideal cubic perovskite cell , @xmath16=3.9 ) . an electron microscope jeol 2010 equipped with an energy dispersive spectroscopy ( eds ) analysis , was used for the electron diffraction ( ed ) study .
the eds has shown that the cationic composition of the film is homogeneous and remains very close to the nominal one nd@xmath17ca@xmath18mno@xmath3 within experimental error .
the resistivity ( @xmath19 ) of the films was measured with a ppms quantum design as a function of the magnetic field ( @xmath20 ) and the temperature ( @xmath21 ) using four - probe contacts . the thickness ( @xmath22 ) of the films
is measured using a dektak@xmath23st surface profiler .
besides the classical growth parameters such as the deposition temperature , the oxygen pressure , we found that the energy of the laser was also important in order to obtain a single phase .
the optimum energy value is @xmath24 which corresponds to @xmath25 on the target .
this is evidenced on fig.1 for a 2000 thick film of ncmo on sto .
a low ( @xmath26 ) or a high ( @xmath27 ) laser energy leads to a poor crystallization of the film with appearance of a secondary phase .
the structure of bulk ncmo is orthorhombic ( @xmath28 ) with @xmath29 , @xmath30 and @xmath31 @xcite .
some area of the film have been investigated by ed and we found that the films are single phase , [ 010]-oriented , i.e. with the [ 010]-axis perpendicular to the sto substrate plane , and [ 101]-oriented , i.e. with the [ 101]-axis perpendicular to the lao substrate plane in the space group @xmath28 .
this orientation is not surprising and resulting from the lattice mismatch between the film and the substrate as previously reported for pr@xmath11ca@xmath11mno@xmath1 thin films grown on ( 100)-srtio@xmath1 and ( 001)-laalo@xmath1 substrates @xcite .
details of the transmission electron microscopy study are currently undertaken and will be published elsewhere .
fig.2a shows the evolution of the lattice parameters of the ncmo film grown on sto versus thickness .
one observed that the in - plane lattice parameter @xmath32 decreases as the thickness of the film increases while at the same time , the out - of - plane lattice parameter @xmath33 increases .
the lattice parameters of films on lao were also determined ( fig.2b ) but , due to the twins of this substrates , we were not able to obtain the in - plane lattice parameters .
the out - of - plane lattice parameters are almost constant on lao when the thickness is changing , suggesting that the film is quasi - relaxed even for small thickness . figure 3a shows the resistivity dependence of the temperature for ncmo films on sto . without a magnetic field ,
all films present a semiconducting behavior with an anomaly around 275k , corresponding to the @xmath34 .
a similar effect is observed on the bulk , but at a lower temperature close to 250k @xcite .
in fact , this anomaly in the @xmath35 curves is more clear ( inset of fig .
3a ) , when the resistance is plotted in a logarithmic scale versus the inverse of the temperature . when applying a magnetic field of 7 t , while the thinner film ( @xmath36 ) remains insulating in the whole temperature range of 4 - 300k , the thicker films ( @xmath37 ) show an insulator - to - metal transition ( @xmath9 ) at 100k and 130k for the 1340and 2000 films respectively .
such an effect is typical of the metastability of the co / oo state where a magnetic field ( abbreviated @xmath38 ) can induce a metallic behavior at low temperature .
however , the magnetic field required is much higher in the bulk compound ( close to 20 t for the same temperature ) @xcite .
in fact , there is a dependence of the critical field with the temperature ( figure 3b ) .
the resistivity shows an important decrease on a logarithmic scale at a critical field ( @xmath38 ) indicating the field - induced melting of the co / oo state .
this field - induced insulator - to - metal transition , which accompanies the collapsing of the co / oo state takes place below @xmath34 .
this decrease can be viewed as a cmr of about four orders of magnitude at 75k . a clear hysteresis ( between the lower and the upper critical fields ) is seen at these temperatures as previously reported on several co / oo compounds @xcite but this hysteretic region is more pronounced when the temperature is decreasing ( see @xmath39 at 25k in fig.3b ) .
in addition , the temperature dependence of the large hysteresis region is a feature of a first order transition and has been extensively studied for the composition nd@xmath11sr@xmath11mno@xmath1 in ref .
note also the reentrant nature of the co at 25k due to the fact that a 7 t magnetic field is not enough to completely melt this state . on lao substrates ,
the @xmath35 curves are practically similar ( fig.4a ) to those on sto except that in this case , even the thinnest film ( 670 ) displays an insulator - to - metal transition when applying a magnetic field of 7 t .
moreover , the @xmath9 is higher than in the case of sto which is consistent with the quasi - relaxed ncmo film found on lao ( for a 1340film , @xmath40 on sto and @xmath41 on lao ) .
large hystereses indicating the first - order phase transition are also observed ( fig.4b ) and , in the same way , there is an increase of the @xmath34 when the thickness of the film decreases ( inset of fig .
4a ) .
this study shows that the thickness of ncmo films deposited on stoinfluences strongly the transport properties of the material . for the largest thickness ( @xmath42 )
, an insulator - to - metal transition is induced under 7 t in contrast to the bulk @xcite , although the cell parameters reach values close to those of the bulk ceramics @xcite.moreover , the @xmath9 under 7 t magnetic field decreased as the thickness decreases whereas at the same time , the @xmath34 increases so that for a small thickness ( @xmath43 ) the film remains insulator .
though it was suggested by several authors in bulk samples of the same composition that electronic phase separation can be at the origin of the drop in resistivity @xcite , no evidence of such phase separation was found in the thin films .
we have indeed observed a continuous variation of the lattice parameters with the thickness of the films .
thus , in order to explain this evolution , we have take into consideration the following features concerning the strains induced by the sto substrates \(i ) the tensile stain along the [ 101 ] direction ( in the plane of the substrate ) can increases both the mn - o - mn angles and mn - o distances in the equatorial plane ( see fig.5a ) .
the increase of the mn - o@xmath44-mn @xcite bond angle increases the bandwith @xcite , the angles tending toward @xmath45 , so that de is involved .
consequently this structural effect should destabilize the co / oo state and favor the @xmath9 .
the increase of the mn - o bond length induces the antagonist effect , favoring the distortion of the mno@xmath46 octahedra , i.e. the jahn - teller distortion , so that the co / oo state should be stabilized .
\(ii ) the compression of the structure along the [ 010 ] direction ( along the out - of - plane direction on sto ) , which appears simultaneously , results in a decrease of the mn - o@xmath47-mn @xcite angle along this direction .
consequently , this decreases the bandwidth and stabilizes the co / oo state .
\(iii ) in the @xmath28 structure ( such as ncmo ) , there are eight mn - o@xmath44-mn bonds against four mn - o@xmath47-mn bons .
thus , the bandwidth along the mn - o@xmath48 is consistently larger than the bandwidth along the mn - o@xmath47 direction which indiquates that the de along the [ 101 ] direction prevails over the [ 010 ] direction @xcite .
\(iv ) viewing these remarks , it is reasonable to assume that the substrate - induced strains modify in a first step the tilting of the octahedra , i.e. the mn - o - mn angles and only in a second step the mn - o bond lengths in the plane of the substrate . for the larger films thickness ( @xmath49 ) , the cell parameters are close to those of the bulk leading consequently to a tilting of the octahedra rather than a modification of the bond lengths . in other words ,
the variation of the mn - o - mn angles prevails over the variation of the mn - o distances .
thus , as @xmath50 increases the mn - o@xmath44-mn angles increases with respect to the bulk .
in contrast , as @xmath51 decreases the mn - o@xmath47-mn angle decreases , resulting in an increase of the bandwidth along the [ 101 ] direction and a decrease of the bandwidth along the [ 010 ] direction.but as pointed out above , the first effect prevails over the second one . as a consequence ,
the de favored with respect to the bulk , so that the co / oostate is less stable and an @xmath9 is obtained under a 7 t magnetic field contrary to the bulk .
as the thickness of the film decreases from 2000 , @xmath50 increases significantly so the mn - o@xmath44-mn bond angle reached 180@xmath52 rapidly and the mn - o@xmath44 bond length increases .
the increase of the mn - o@xmath48 distance induces a jt distortion of the octahedra , stabilizing the co / oo state at the detriment of the @xmath9 transition .
this explains why the @xmath34 transition increases as the thickness of the film decreases and the @xmath9 transition under 7 t decreases correlatively .
finally , for the smaller film thickness ( @xmath43 ) , the @xmath50 is very close to the sto parameter , and the bond angle mn - o@xmath44 effects prevails over the mn - o@xmath44-mn effect , so that the co / oo state is the most stable . note that the simultaneous decrease of the @xmath51 , i.e. decrease of mn - o@xmath47-mn also favors the stabilization of the co / oo state . in the case of ncmo films grown on lao ,
the substrate - induced strains lead in fact to rather similar effects .
indeed , one observes a compression along the [ 010]-axis in the plane of the substrate . as shown in fig.5b , this induces a decreases of the mn - o@xmath47-mn angles and of the mn - o@xmath47 bond lengths similar to what occurs for sto ; an increase of the mn - o@xmath44-mn out - of - plane bond angle is induced , but in this case the increase of the mn - o@xmath44 bond distance does not prevail over the mn - o@xmath44-mn angle ( not imposed by the substrate ) . for small thickness ( @xmath43 ) , the @xmath9transition is favored with respect to the co / oo state .
in conclusion , we grown high quality thin films of nd@xmath11ca@xmath11mno@xmath8 on srtio@xmath4 and laalo@xmath4 using the pld technique .
we have shown that we are able to induce an insulator - to - metal transition at a magnetic field much lower than the corresponding bulk compound by the modification of the structure through the utilization of the appropriate substrate .
moreover the substrate - induced strains increases the mn - o@xmath44-mn bond angles on both lao and sto substrates ( even if the orientation of the film is different ) favoring the double exchange and consequently destabilizing the charge / orbital state leading to a decrease of the critical magnetic field as compared to the bulk .
r. von helmolt , j. wecker , r. holzapfel , l. schultz , and k. samwer , phys .
71 , 2331 ( 1993 ) , mccormack , s. jin , t. tiefel , r. m. fleming , j.m .
philips , and r. ramesh , appl .
64 , 3045 ( 1994 ) .
p.m. woodward , t. vogt , d.e .
cox , a. arulraj , c.n.r .
rao , p. karena and a.k .
cheetham , chem .
10 , 3652 ( 1998 ) , a. arulraj , p.n .
santhosh , r.s .
gopalan , a. guha , a k raychaudhuri , n. kumar and c.n.r .
rao , j. phys .
: condens .
mater 10 , 8497 ( 1998 ) .
fig.2 : : ( a ) : evolution of the lattice parameters of ncmo on sto as a function of the thickness .
( b ) : evolution of the out - of - plane lattice parameters of ncmo on lao as a function of the thickness the values of the bulk are indicated as a dot line ( see text for details ) .
lines are only guide for the eyes .
fig.3 : ncmo films on sto with various thickness .
( a ) : @xmath35 under 0 t ( empty symbols ) and 7 t ( full symbols ) .
the inset depicts the evolution of the logarithmic resistivity versus the inverse of the temperature under 0 t ( empty symbols ) and 7 t ( full symbols ) .
note the anomaly associated to the @xmath55 ( marked by arrow ) .
( b ) : @xmath39 at various temperature for a 2000 film .
fig.4 : ncmo films on lao with various thickness .
( a ) : @xmath35 under 0 t ( empty symbols ) and 7 t ( full symbols ) . the inset depicts the evolution of the logarithmic resistivity versus the inverse of the temperature under 0 t ( empty symbols ) and 7 t ( full symbols ) .
note the anomaly associated to the @xmath55 ( marked by arrow ) .
( b ) : @xmath39 at various temperature for a 2000 film . | strained thin films of charge / orbital ordered ( co / oo ) @xmath0 ( ncmo ) with various thickness have grown on ( 100)-srtio@xmath1 and ( 100)-laalo@xmath1 substrates , by using the pulsed laser deposition ( pld ) technique .
the thickness of the films influences drastically the transport properties . as the thickness decreases , the co transition increases while at the same time the insulator - to - metal transition temperature decreases under application of a 7 t magnetic field .
clear relationships between the structural distortions and the transport properties are established .
they are explained on the basis of the elongation and the compression of the mn - o - mn and mn - o bond angles and distances of the _ pnma _ structure , which modify the bandwidth and the jahn - teller distortion in these materials . |
one of the differences between laboratory and space laser interferometer gravitational wave detectors is that , in the laboratory , the two arms of the interferometer that is used to detect changes in the spacetime geometry are maintained at precisely equal lengths .
therefore , when the signals from the two perpendicular arms are combined , the laser phase noise in the differenced signals cancels exactly . in space , a laser interferometer gravitational wave detector such as lisa @xcite will have free - flying spacecraft as the end masses , and precise equality of the arms is not possible .
other methods must then be used to eliminate laser phase noise from the system @xcite .
these methods involve a heterodyne measurement for each separate arm of the interferometer and data processing that combines data from both arms to generate a signal that is free of laser phase noise . in a previous paper ( @xcite ,
hereafter called paper i ) , the sensitivity curves for space detectors using these techniques were generated by explicitly calculating transfer functions for signal and noise , as modified by the data processing algorithms . while the algorithms have been shown @xcite , in principle , to eliminate the laser phase noise in the detectors regardless of the lengths of the two arms ,
the transfer functions have previously only been calculated for the case of equal arms @xcite . in this paper
we extend the calculation of the noise and signal transfer functions to the case of arbitrarily chosen armlengths .
one of the goals of paper i was to provide a uniform system for evaluating the sensitivity of various configurations of space gravitational detectors .
this paper extends that capability to configurations in which the armlengths are significantly different from each other .
for example , a proposal by bernard schutz at the 2000 lisa symposium in golm , germany @xcite , suggested a modification to the current lisa design in which a fourth spacecraft is inserted in the middle of one of the legs of the interferometer to produce two independent interferometers , each with one leg half the length of the other ( see fig . [ fig : detector ] ) .
the goal of such a design was to be able to cross - correlate the independent interferometers to search for the stochastic cosmic gravitational wave background .
using the analysis presented here , one will be able to determine the sensitivity of such an interferometer and judge the scientific value of the proposed modification .
as in paper i , the analysis begins with the response of a round - trip electromagnetic tracking signal to the passage of a gravitational wave , as derived by estabrook and wahlquist @xcite . a gravitational wave of amplitude @xmath0 will produce a doppler shift @xmath1 in the received frequency , relative to the outgoing signal with fundamental frequency @xmath2 .
the shift is given by @xmath3 , \label{dopplersignal}\end{aligned}\ ] ] where @xmath4 is the one - way light travel time between spacecraft , @xmath5 is the angle between the line connecting the spacecraft and the line of sight to the source , and @xmath6 is a principal polarization angle of the quadrupole gravitational wave .
it is desirable to work in frequency space , so @xmath0 is written in terms of its fourier transform @xmath7 .
if the doppler record is sampled for a time @xmath8 then @xmath0 is related to its fourier transform by @xmath9 where the @xmath10 normalization factor is used to keep the power spectrum roughly independent of time . using this definition of the fourier transform , the frequency shift of eq .
( [ dopplersignal ] ) can be written as @xmath11 e^{i \omega t } d\omega \ , \label{dopplersignal2}\end{aligned}\ ] ] where @xmath12 .
the quantity that is actually read out in a laser interferometer tracking system is phase , so eq . ( [ dopplersignal2 ] ) is integrated to find the phase in cycles @xmath13 in paper i , a strain - like variable @xmath14 was formed by dividing the @xmath15 in eq .
( [ phasesignal ] ) by @xmath16 and the analysis was done using this variable . since both arms had roughly the same length in paper i and carried nearly the same frequency , there was only a scale difference between using @xmath15 and using @xmath14 as the observable , and linear combinations of @xmath14 were the same as linear combinations of @xmath15 .
however , when the two armlengths are different , this is no longer the case , and one must be careful as to what is taken to be the observable for use in noise - cancelling data analysis . in the laser phase - noise - cancellation algorithms that will be presented in section
ii , it is _ relative phase _ and _ not strain _ that can be combined to create laser - noise - free signals . to understand how this arises ,
consider a case where laser signals in two arms are phase - locked to each other , with @xmath17 as the frequency of the master laser in the first arm and @xmath18 as the frequency in the second arm , with @xmath19 as the ratio of the two frequencies .
then a phase noise excursion @xmath20 in the first arm will produce a phase noise excursion @xmath21 in the second arm .
thus it will be linear combinations of @xmath22 that will allow the two noise terms to cancel .
therefore , in this paper , the gravitational wave observable in the @xmath23 arm is defined to be @xmath24 \frac{1}{\omega } e^{i \omega t } \ , \label{zsignal}\end{aligned}\ ] ] where eq .
( [ dopplersignal2 ] ) has been used to expand @xmath25 and where arbitrary constant phases have been set to zero in the integration .
it should be noted that @xmath26 is a different observable than the strain variable that was labelled @xmath26 in paper i. it should also be noted that @xmath26 , as it is now defined , has units of time , so eq . ( [ zsignal ] )
gives the _ time delay _ in seconds produced by the passage of a gravitational wave through the detector .
tinto and armstrong @xcite originally showed that the preferred signal for purposes of data analysis is not the traditional michelson combination ( difference of both arms ) , but rather a new combination @xmath27 , given in the time domain by assumes that the lasers in the end spacecraft are phase - locked to the signals they receive from the central spacecraft .
there is a form for @xmath27 that does not make this assumption and that can thereby be converted to interferometers centered on the other spacecraft in the constellation .
however , the sensitivity we derive using this form will be valid for the more general form as well , and will therefore apply to signals formed with any spacecraft as vertex . ]
@xmath28 where @xmath29 is the data stream from the @xmath30 interferometer arm , composed of the signal @xmath31 of interest ( given by eq.([zsignal ] ) ) and the combined noise spectra in each of the interferometer arms , @xmath32 .
the armlengths are taken to be unequal , with armlength @xmath33 in the @xmath30 arm .
this combination is devoid of laser phase noise for all values of the two armlengths @xmath34 and @xmath35 to determine the sensitivity using the @xmath27 variable , it is necessary to establish a relationship between the amplitude of a gravitational wave incident on the detector and the size of the @xmath27 signal put out by the instrument .
the noise in the detector will limit this sensitivity , and must also be included in the analysis .
the part of @xmath27 containing the gravitational wave signal is was called @xmath36 in the limit where @xmath37 . ] : @xmath38 the transfer function @xmath39 , which connects the spectral density of the instrument output , @xmath40 with the spectral density @xmath41 in frequency space , is defined via @xmath42 where the bar over the @xmath43 in eq .
( [ transferequation ] ) indicates an average over source polarization and direction .
the gravitational wave amplitude spectral density @xmath41 is defined by @xmath44 where @xmath7 is the fourier amplitude defined in eq .
( [ fouriertransform ] ) , so that the mean - square gravitational wave strain is given by @xmath45 similarly , the instrumental response @xmath40 is defined such that @xmath46 where the brackets indicate a time average . in the next section ,
the transfer function from the gravitational wave amplitude @xmath47 to the instrument signal @xmath48 is worked out .
let us take the ratio of the two armlengths in the interferometer to be an adjustable parameter , @xmath49 , taking on values between @xmath50 and @xmath51 , such that @xmath52 and @xmath53 .
the average power in the part of @xmath27 which contains the gravitational wave signal is given by @xmath54 where @xmath43 is defined by eq .
( [ lambdadef ] ) . using the definition of @xmath14 from eq .
( [ zsignal ] ) this can be expanded to yield @xmath55 \ , \label{power}\ ] ] where @xmath56 \ , \label{t1 } \\
t_{2}(u ) & = & \cos^{2}(2 \psi_{2 } ) \cdot 4 \sin^{2}(u ) \left [ \mu_{2}^{2}\left(1 + \cos^{2}(\beta u ) - 2 \cos(\beta u)\ \cos(\beta u \mu_{2 } ) \right ) \right .
\nonumber \\ & - & \left . 2 \mu_{2 } \sin(\beta u)\ sin(\beta u \mu_{2 } ) + \sin^{2}(\beta u ) \right ] \ , \label{t2 } \\
t_{3}(u ) & = & \cos ( 2 \psi_{1 } ) \
\cos ( 2 \psi_{2 } ) \cdot 4 \sin(u ) \sin(\beta u )
\ \eta(u ) \ , \label{t3}\end{aligned}\ ] ] with @xmath57 , @xmath58 , and where @xmath59 \left [ \cos ( \beta u ) - \cos ( \beta u \mu_{2 } ) \right ] \mu_{1 } \mu_{2 } \nonumber \\ + \left [ \sin ( u ) - \mu_{1 } \sin ( u \mu_{1 } ) \right ] \left [ \sin ( \beta u ) - \mu_{2 } \sin ( \beta u \mu_{2 } ) \right ] \label{eta}\end{aligned}\ ] ] has been defined for convenience
. the propagation angles @xmath60 and principal polarization angles @xmath61 are defined with respect to the @xmath30 arm using the geometric conventions of paper i. the expression for the power in the detector , as given by eq .
( [ power ] ) , is a complicated function of frequency and of the orientation between the propagation vector of the gravitational wave and the interferometer , and represents the antenna pattern for the detector .
it is customary to describe the average sensitivity of the instrument by considering the isotropic power , obtained by averaging the antenna pattern over all propagation vectors and all polarizations ) is linearly polarized .
as in paper i , the averaging procedure over all linearly polarized states produces the same response function as averaging over a more general elliptically polarized state with an appropriately weighted distribution ] .
using the definition of @xmath39 from eq .
( [ transferequation ] ) , with the average isotropic power computed using the geometric averaging procedure of paper i with eqs .
( [ t1 ] - [ eta ] ) , the gravitational wave transfer function is found to be @xmath62 \right .
\nonumber \\ & + & \left .
2 \sin^{2}(u ) \left [ \left(1 + \cos^{2 } ( \beta u ) \right ) \left(\frac{1}{3 } - \frac{2}{(\beta u)^{2 } } \right ) + \sin^{2 } ( \beta u ) + \frac{4}{(\beta u)^{3 } } \sin ( \beta u ) \ \cos ( \beta u ) \right ] \right .
\nonumber \\ & - & \left .
\frac{1}{\pi } \sin(u ) \sin(\beta u ) \int_{0}^{2\pi } d\epsilon \int_{-1}^{+1 } d\mu_{1 } \
\left(1 - 2 \sin^{2 } \alpha \right ) \eta(u , \theta_{1 } , \theta_{2 } ) \right\ } \ . \label{transferfunction}\end{aligned}\ ] ] the remaining integral can be evaluated using simple numerical techniques , after relating the angular variables as described in paper i , where : @xmath63 and @xmath64 here @xmath65 is the opening angle of the interferometer , and @xmath66 is the inclination of the gravitational wave propagation vector to the plane of the interferometer .
the complete gravitational wave transfer function is plotted in fig .
[ fig : gwtransfers ] for @xmath67 ( `` equal arm '' ) and fig . [ fig : gwtransfers2 ] for @xmath68 ( `` unequal arm '' ) examples . as may be seen in the figure , the low - frequency ( small @xmath69 ) response of the detector to a gravitational wave signal is four orders of magnitude lower for the @xmath70 detector than for the equal arm detector , implying that the ( amplitude ) signal will be two orders of magnitude lower the detected signal level is proportional to the length of the shortest arm . however , once the period of the gravitational wave falls inside the light - time of the longest arm , @xmath71 , the equal - arm detector ( @xmath67 ) response begins to fall off while the unequal - arm detector ( @xmath68 ) response is roughly flat up to a period corresponding to the light - time in the shortest arm .
the dropoff at low frequencies is a result of the fact that the variable @xmath27 is formed by subtracting each @xmath26 from itself , offset by the light - time in the opposite arm .
thus , in the low - frequency limit , the two copies of the signal strongly overlap and the signal is almost entirely subtracted away . for equal arms , the response of the detector is likewise subtracted to zero when an integer number of wavelengths fits in the arm length , as seen in the high - frequency portion of the @xmath72 curve . for the unequal - arm case , this does not occur , because the subtraction of two versions of the signal in each arm are done at different light - times in the two arms , so whatever period signal cancels in one arm will not cancel in the other .
however , as may be seen in the @xmath68 case , the response drops sharply to zero at @xmath73 ( equivalent to @xmath74 hz for lisa armlength of @xmath75 m ) , where exactly one wavelength fits into the short arm and exactly one hundred fit into the long arm . however , the response of the detector s @xmath27 signal is not the whole story .
the ability of a detector to detect a signal depends on both the signal in the detector and on the competing noise .
as we shall see in the next section , when the @xmath27 variable is formed , the noise in each arm is likewise subtracted away in most of the places where the signal is lost ( _ e.g. _ , at low frequency ) , so the ratio of signal to noise remains high .
the noise sources for lisa may be divided into categories in two different ways .
first , a noise source may be either one - way ( affecting only the incoming or the outgoing signal at a spacecraft , but not both ) or two - way , ( affecting both incoming and outgoing signals at the same time ) .
a one - way noise source will have a transfer function of @xmath76 , since there are @xmath76 spacecraft in each leg contributing equal amounts of such noise . as @xmath77 , where @xmath78 is the noise spectral density of a single type in a single spacecraft ] the transfer function for two - way noise sources , however , will be more complicated due to the internal correlation .
a single two - way noise fluctuation in the central spacecraft of the interferometer will affect the incoming signal immediately , and then , a round - trip light - time later , will affect the measured signal again in the same way . in the time domain , the effect in the @xmath30 arm of a fluctuation @xmath79 will be @xmath80 .
the transfer function for this time - delayed sum is @xmath81 .
if an end spacecraft has noise that affects both incoming and outgoing beams , it will affect them at almost the same time , with no delay , giving a transfer function contribution of 4 .
the noise transfer function for a single arm for a two - way noise source is therefore @xmath82 examples of one - way noise are thermal noise in the laser receiver electronics or a mechanical change in the optical pathlength in the outgoing laser signal before it gets to the main telescope optics .
examples of two - way noise are parasitic forces on the accelerometer proof mass or thermal changes in the optical pathlength in the main telescope .
a second way in which noise sources may be classified is by how they scale when there is a change in armlength in the interferometer .
the first type of noise in this classification scheme is what we call `` position noise '' , in which the size of the noise in radians of phase is independent of the length of the arm .
accelerometer noise and thermal noise in the laser electronics are examples of position noise .
the second type of noise is what we call `` strain noise '' , in which the size of the noise scales with armlength .
examples of strain noise include shot noise and pointing jitter ( if it is dominated by low power in the incoming beacon )
. position noises may be either one - way or two - way , but we can think of no two - way strain noise sources .
the transfer functions that connect the noise in the instrument to the @xmath83 variable depend on the type of noise .
we begin by considering the noise terms in eq .
( [ xsignal ] ) : @xmath84 \ . \label{noisepart}\ ] ] we then go to the frequency domain , squaring and time - averaging to obtain the power spectrum .
@xmath85 \ , \label{noisepower}\ ] ] where cross - terms ( _ e.g. _ , @xmath86 ) have been neglected under the assumption that noise in the two arms will be independent and uncorrelated .
note that @xmath87 is the power spectrum in the long arm ( length @xmath88 ) and @xmath89 is the power spectrum in the short arm ( length @xmath90 ) . since the noise in the detectors includes different types , with different transfer functions , it is not possible to write a single transfer function giving the response of the @xmath83 variable to noise , so let us consider the various noise categories one at a time .
we first consider position noise , for which @xmath91 .
then , using eq .
( [ noisepower ] ) , we find the transfer function for one - way position noise to be @xmath92 where , as we noted above , there is a factor of @xmath76 representing the noise from the two spacecraft in each arm .
two - way position noise must include the transfer function from eq .
( [ twoway ] ) , giving @xmath93\ .
\label{twowaytransfer}\ ] ] strain noise scales with armlength , and is hence smaller in the shorter arm , so that @xmath94 .
its transfer function is therefore @xmath95 where the factor of @xmath76 for the two spacecraft has again been included . when @xmath72 , the transfer functions for strain noise and one - way position noise ( eqs .
[ positiontransfer ] and [ shottransfer ] ) are identical and have zeros at @xmath96 , where @xmath97 is zero or a positive integer .
these are exactly the places where the @xmath72 transfer function for gravitational wave signal ( fig .
[ fig : gwtransfers ] ) has its zeros . when @xmath98 , the situation is more complicated .
both @xmath99 and @xmath100 share the @xmath101 term which will go to zero at @xmath102 and at multiples of @xmath103 .
the @xmath104 terms in @xmath99 and @xmath100 have their zeros at multiples of the lower frequency , @xmath105 . in @xmath99
, this term will be larger than @xmath101 term at low frequencies , since near @xmath102 , @xmath106 , while @xmath107 . in @xmath100
, these terms will be equal in the low - frequency limit , because of the factor @xmath108 that multiplies the @xmath104 term .
thus , in the low frequency limit , the strain noise transfer function will be @xmath109 times the the one - way position noise transfer function .
when @xmath110 , the transfer function for one - way position noise will have sharp drops at multiples of @xmath105 , down to the level of its @xmath101 term .
these behaviors are shown in fig .
[ fig : noisetransfer ] and fig .
[ fig : noisetransfer2 ] . the signal to
noise ratio is the ratio of the signal power in the detector to the noise power in the detector : @xmath111 where @xmath112 , @xmath113 , and @xmath114 are the spectra of strain noise and one - way and two - way position noise , respectively , and @xmath115 is the gravitational wave transfer function given by eq.([transferfunction ] ) . setting @xmath116 and solving for @xmath117 yields the instrument sensitivity curve as defined in paper i : @xmath118 where @xmath115 is the gravitational wave transfer function , given by eq .
( [ transferfunction ] ) .
figures [ fig : equalarmsense ] and [ fig : unequalarmsense ] show the sensitivity curves , computed using eq .
( [ hf ] ) , for @xmath67 and @xmath68 respectively .
the noise values used are taken to be the lisa target design values ( computed as described in paper i ) .
the shot noise and acceleration noise levels are set at the standard lisa values . in addition
, a flat one - way position noise spectrum is assumed at @xmath119 the lisa shot - noise value . also plotted in figures [ fig : equalarmsense ] and [ fig : unequalarmsense ] are sensitivity curves representing each of the three components of the total noise , taken one at a time .
as may be seen in fig .
[ fig : unequalarmsense ] , the low - frequency sensitivity for unequal arms , being set by the two - way position noise in the accelerometer , is degraded over the equal - arm case by the ratio of the two arms . in other words ,
the sensitivity at lowest frequencies is set by the sensitivity of the shortest arm . at middle and high frequencies , the situation is more complicated . if the dominant noise is strain noise , then the sensitivity is independent of @xmath49 in this frequency range . however ,
if the dominant noise is position noise , then the sensitivity curve at high frequencies will rise in proportion to @xmath49 , though its flat floor will extend to higher frequency , from the @xmath120 of the equal - arm case to @xmath121 when the armlength ratio is @xmath49 .
the implications of these results for mission design are obvious .
if the armlengths are not equal , the low - frequency sensitivity is degraded by a factor @xmath122 , the ratio of the armlengths .
if the high - frequency noise can be guaranteed to be strain noise , even in the shorter arm , then the high - frequency sensitivity is unaffected by the unequal arms .
if the noise at high frequency is dominated by position noise , then the high frequency sensitivity is degraded by the factor @xmath122 , but the sensitivity remains flat up to a frequency @xmath123 , where it turns over and joins the strain noise curve .
thus , as long as the position - noise sources can be kept well below the shot noise and other strain - noise contributions , a change in armlength ratio from strict equality will not degrade the high - frequency portion of the sensitivity curves .
as the length of one of the arms is shortened , small position noise sources will become important and eventually dominate .
let us consider the example of schutz s 4-spacecraft configuration ( fig .
[ fig : detector ] ) .
since this configuration will have @xmath124 , the low - frequency sensitivity curve will be a factor of 2 higher ( hence less sensitive ) .
the current error budget for lisa assumes that the high - frequency portion of the window is dominated by position noise approximately three times the shot noise .
if this remains the case , then the high - frequency section of the curve will likewise be a factor of 2 higher up to a frequency twice as high as the lisa sensitivity `` knee '' at @xmath125 , at which point it would turn up and join the current lisa high - frequency ramp .
the shot noise is determined by the power of the laser and by the size and efficiency of the optics , and there is nothing beyond brute - force improvements in these parameters that will lower the shot noise .
the contributions to position noise , on the other hand , are due to optics quality , the attitude control system , brownian noise in the electronics , thermal noise in the optical path length , _
etc_. these are more complex and are amenable to reduction by careful or innovative engineering design .
if these noise sources can be reduced to a fraction of the shot noise , not only will the lisa noise floor be reduced by a factor of 4 , but the schutz modification will have high - frequency performance that is undiminished by the reduction of the length of one arm . finally , we describe a totally unfeasible mission design that is nevertheless interesting for instructive purposes .
let us consider a two - spacecraft `` interferometer '' , where one of the spacecraft contains a fiber optic delay line , of length 5 km , that acts as the second arm of the interferometer .
if the distance between the two spacecraft is @xmath126 km , we will have @xmath127 .
the use of the @xmath27 variable will eliminate laser phase noise , exactly as it does in arms that are more nearly equal .
a rigidly - attached reflector at the far end of the fiber - optic line would eliminate accelerometer noise , but , of course , would replace it with thermal fluctuation in the optical path length in the fiber .
however , a concatenation of fibers with well - chosen thermal pathlength coefficients could produce a fiber tuned to have a coefficient very near zero .
this , combined with multilevel thermal isolation , could keep this noise source very small .
the key to the sensitivity of this configuration is the position noise .
if a way can be found to reduce position noise to less than @xmath128 of the lisa shot noise , then this two - spacecraft interferometer would have the same sensitivity as a conventional three - spacecraft interferometer .
s. l. l. acknowledges support for this work under lisa contract number po 1217163 , and the nasa epscor program through cooperative agreement ncc5 - 410 . the work of w. a. h. was supported in part by nsf grant no .
phy-0098787 and the nasa epscor program through cooperative agreement ncc5 - 579 .
r. w. h. was supported by nasa grant nags5 - 11469 and ncc5 - 579 . | unlike ground - based interferometric gravitational wave detectors , large space - based systems will not be rigid structures .
when the end - stations of the laser interferometer are freely flying spacecraft , the armlengths will change due to variations in the spacecraft positions along their orbital trajectories , so the precise equality of the arms that is required in a laboratory interferometer to cancel laser phase noise is not possible . however , using a method discovered by tinto and armstrong , a signal can be constructed in which laser phase noise exactly cancels out , even in an unequal arm interferometer .
we examine the case where the ratio of the armlengths is a variable parameter , and compute the averaged gravitational wave transfer function as a function of that parameter .
example sensitivity curve calculations are presented for the expected design parameters of the proposed lisa interferometer , comparing it to a similar instrument with one arm shortened by a factor of 100 , showing how the ratio of the armlengths will affect the overall sensitivity of the instrument . |
consider the following generative model for independent component analysis ( ica ) @xmath3 where the elements of the non - gaussian source vector @xmath4 are mutually independent with zero mean , @xmath5 is an unknown nonsingular mixing matrix , @xmath6 is an observable random vector ( signal ) , and @xmath7 is a shift parameter . let @xmath8 be the whitened data of @xmath9 , where @xmath10 .
an equivalent expression of model ( [ ica ] ) in @xmath11-scale is @xmath12 where @xmath13 is the mixing matrix in @xmath11-scale .
it is reported in literature that prewhitening the data can make the ica inference procedure more stable . in the rest of the discussion
, we will work with model ( [ ica.z ] ) in estimating the mixing matrix @xmath14 based on the prewhitened @xmath11 .
it is easy to transform back to the original @xmath9-scale via @xmath15 .
note that both @xmath14 and @xmath16 are unknown , and there exists the identifiability problem .
this can be seen from the fact that @xmath17 for any nonsingular diagonal matrix @xmath18 . to make @xmath19 identifiable , we assume the following conventional conditions for @xmath16 : @xmath20 where @xmath21 is the identity matrix .
it then implies that @xmath22 and @xmath23 which means that the mixing matrix @xmath14 in @xmath11-scale is orthogonal .
we will use notation @xmath24 to denote the space of orthogonal matrices in @xmath25 . note that
, if @xmath26 is a parameter of model ( [ ica.z ] ) , so is @xmath27 .
thus , to fix one direction , we consider @xmath28 , where @xmath29 consists of orthogonal matrices with determinant one .
this set @xmath30 is called the special orthogonal group .
the main purpose of ica is to estimate the orthogonal @xmath31 based on the whitened data @xmath32 , or equivalently , to look for a recovering matrix @xmath33 so that components in @xmath34 have the maximum degree of independence . in the latter case
, @xmath35 provides an estimate of @xmath19 .
we first briefly review some existing methods for ica .
one idea is to estimate @xmath35 via _ minimizing the mutual information_. let @xmath36 be the joint probability density function of @xmath37 , and @xmath38 be the marginal probability density function of @xmath39 . the mutual information , denoted by @xmath40 , among random variables @xmath41 , is defined to be @xmath42 where @xmath43 and @xmath44 are the shannon entropy . ideally , if @xmath35 is properly chosen so that @xmath45 has independent components , then @xmath46 and , hence , @xmath47 .
thus , via minimizing @xmath48 with respect to @xmath35 , it leads to an estimate of @xmath35 .
another method is to estimate @xmath35 via _ maximizing the negentropy _ , which is equivalent to minimizing mutual information as described below .
the negentropy of @xmath45 is defined to be @xmath49 where @xmath50 is a gaussian random vector having the same covariance matrix as @xmath45 ( hyvrinen and oja , 2000 ) .
it can be deduced that @xmath51 where the second equality holds since , by @xmath52 , @xmath53 . moreover , as @xmath54 with @xmath33 , we have @xmath55 , which does not depend on @xmath35 .
that is , the negentropy is invariant under orthogonal transformation .
thus , minimizing the mutual information @xmath48 is equivalent to maximizing the negentropy @xmath56 .
the negentropy @xmath57 , however , involves the unknown density @xmath38 . to avoid nonparametric estimation of @xmath38
, one can use the following approximation ( hyvrinen , 1998 ) via a non - quadratic contrast function @xmath58 , @xmath59 ^ 2,\label{ne_approx}\end{aligned}\ ] ] where @xmath60 is a random variable having the standard normal distribution . here
@xmath61 can be treated as a measure of non - gaussianity , and minimizing the sample analogue of @xmath62 to search @xmath35 corresponds to the fast - ica ( hyvrinen , 1999 ) . another widely used estimation criterion for @xmath35 is via _ maximizing the likelihood_. under model ( [ ica.z ] ) and by modeling @xmath63 with some known probability density function @xmath64 , the density function of @xmath11 takes the form @xmath65 since @xmath33 and hence @xmath66 .
the mle - ica then searches the optimum @xmath35 via @xmath67 where @xmath68 is the kullback - leibler divergence ( kl - divergence ) , and @xmath69 is the empirical distribution of @xmath32 .
possible choices of @xmath64 include @xmath70 for sub - gaussian models , and @xmath71 for super - gaussian models , where @xmath72 and @xmath73 are constants so that @xmath64 is a probability density function .
it can be seen from ( [ likelihood.z ] ) that , for any row permutation matrix @xmath74 , we have @xmath75 .
that is , we can estimate and identify @xmath14 only up to its row - permutation . as will become clear later that the above mentioned methods are all related to _ minimizing the kl - divergence _ , which is not robust in the presence of outliers .
outliers , however , frequently appear in real data analysis , and a robust ica inference procedure becomes necessary .
for the purpose of robustness , instead of the kl - divergence , mihoko and eguchi ( 2002 ) considers the _
minimum @xmath2-divergence _ estimation for @xmath35 ( @xmath2-ica ) .
the issues of consistency and robustness of @xmath2-ica are discussed therein . on the other hand , the @xmath1-divergence , which can be induced from @xmath2-divergence ,
is shown to be super robust ( fujisawa and eguchi , 2008 ) against data contamination .
it is our aim in this paper to propose a unified ica inference procedure by minimum divergence estimation .
moreover , due to the property of super robustness , we will focus on the case of @xmath1-divergence and propose a robust ica procedure , called @xmath1-ica .
hyvrinen , karhnen and oja ( 2001 ) have provided a sufficient condition to ensure the validity of mle - ica under the orthogonality constraint of @xmath35 , in the sense of being able to recover all independent components .
amari , chen , and cichocki ( 1997 ) studied necessary and sufficient conditions for consistency under a different constraint of @xmath35 , and this consistency result is further extended by mihoko and eguchi ( 2002 ) to the case of @xmath2-ica . in this work , we also derive necessary and sufficient conditions for the consistency of @xmath1-ica . in the limiting case @xmath76 ,
our necessary and sufficient condition for the consistency of mle - ica is weaker than the condition stated in hyvrinen , karhnen and oja ( 2001 ) . to the best of our knowledge
, this result is not explored in existing literature .
some notation is defined here for the convenience of reference . for any @xmath77 ,
let @xmath78 be the commutation matrix such that @xmath79 ; @xmath80 ( resp .
@xmath81 ) means @xmath18 is strictly positive ( resp .
negative ) definite ; and @xmath82 is the matrix exponential . note that @xmath83 for any nonsingular square matrix @xmath18 . for a lower triangular matrix @xmath18 with 0 diagonals
, @xmath84 stacks the nonzero elements of the columns of @xmath18 into a vector with length @xmath85 .
there exist matrices @xmath86 and @xmath87 such that @xmath88 and @xmath89 .
each column vector of @xmath90 is of the form @xmath91 , @xmath92 , where @xmath93 is a vector with a one in the @xmath94-th position and 0 elsewhere , and @xmath95 is the kronecker product . @xmath21 is the identity matrix and @xmath96 is the @xmath97-vector of ones .
the rest of this paper is organized as follows .
a unified framework for ica estimation by minimum divergence is introduced in section 2 .
a robust @xmath1-ica procedure is developed in section 3 , wherein the related statistical properties are studied . a geometrical implementation algorithm for @xmath1-ica
is further illustrated in section 4 . in section 5 ,
the issue of selecting @xmath1 value is discussed .
numerical studies are conducted in section 6 to demonstrate the robustness of @xmath1-ica .
the paper is ended with a conclusion in section 7 .
all the proofs are placed in appendix .
in this section we introduce a general framework for ica by means of a minimum @xmath0-divergence , which covers the existing methods reviewed in section 1 .
the aim of ica is to search a matrix @xmath33 so that the joint probability density function @xmath36 for @xmath54 is as close to marginal product @xmath99 as possible .
this aim then motivates estimating @xmath35 by minimizing a distance metric between @xmath36 and @xmath99 .
a general estimation scheme for @xmath35 can be formulated through the following minimization problem @xmath100 where @xmath101 is a divergence function .
different choices of @xmath102 will lead to different estimation criteria for ica . here
we will consider a general class of divergence functions , the @xmath0-divergence ( murata et al . , 2004 ; eguchi , 2009 ) , as described below .
the @xmath0-divergence is a very general class of divergence functions .
consider a strictly convex function @xmath103 defined on @xmath104 , or on some interval of @xmath104 where @xmath103 is well - defined .
let @xmath105 be the inverse function of @xmath106 .
consider @xmath107 which defines a mapping from @xmath108 to @xmath109 , where @xmath110 .
define the @xmath0-cross entropy by @xmath111 and the @xmath0-entropy by @xmath112 .
then the @xmath0-divergence can be written as @xmath113 in the subsequent subsections , we will introduce some special cases of @xmath0-divergence , which will lead to specific methods of ica . by taking the @xmath114 pair @xmath115
the corresponding @xmath0-divergence is equivalent to the kl - divergence @xmath116 . in this case
, it can be deduced that @xmath117 where @xmath48 is the mutual information defined in ( [ mi ] ) .
as described in section 1 that @xmath118 we conclude that the following criteria , minimum mutual information , maximum negentropy , and fast - ica , are all special cases of ( [ div_ica ] ) . on the other hand , observe that @xmath119 where @xmath120 is the joint probability density function of @xmath11 .
if we consider the model @xmath121 , and if we estimate @xmath120 by its empirical probability mass function @xmath69 , minimizing ( [ mutual_mle ] ) is equivalent to mle - ica in ( [ mle_ica_z ] ) . in summary , choosing the kl - divergence @xmath116 covers minimum mutual information , maximum negentropy , fast - ica , and mle - ica .
consider the convex set @xmath123 .
take the @xmath114 pair @xmath124 the resulting @xmath0-divergence defined on @xmath125 is calculated to be @xmath126 which is called @xmath2-divergence ( mihoko and eguchi , 2002 ) , or density power divergence ( basu et al . , 1998 ) .
note that @xmath127 if and only if @xmath128 for some @xmath129 . in the limiting case @xmath130 ,
it gives the kl - divergence .
if we replace @xmath131 in ( [ mle_ica_z ] ) by @xmath132 , it gives the @xmath2-ica of mihoko and eguchi ( 2002 ) .
the @xmath1-divergence can be obtained from @xmath2-divergence through a @xmath0-volume normalization , @xmath134 where @xmath135 is defined the same way as ( [ beta_div ] ) with the plug - in @xmath136 , and where @xmath137 is some normalizing constant .
here we adopt the following normalization , called the volume - mass - one normalization , @xmath138 it leads to @xmath139 .
then , @xmath140 it can be seen that @xmath1-divergence is scale invariant .
moreover , @xmath141 if and only if @xmath128 for some @xmath129 .
the @xmath1-divergence , indexed by a power parameter @xmath1 , is a generalization of kl - divergence . in the limiting case @xmath142
, it gives the kl - divergence .
it is well known that mle ( based on minimum kl - divergence ) is not robust to outliers . on the other hand ,
the minimum @xmath1-divergence estimation is shown to be super robust ( fujisawa and eguchi , 2008 ) against data contamination .
hence , we will adopt @xmath1-divergence to propose our robust @xmath1-ica procedure .
in particular , the main idea of @xmath1-ica is to replace @xmath131 in ( [ mle_ica_z ] ) by @xmath143 .
though the idea is straightforward , there are many issues need to be studied .
detailed inference procedure and statistical properties of @xmath1-ica are discussed in section 3 .
the ica is actually a two - stage process .
first , we need to whiten the data .
the whitened data are then used for the recovery of independent sources . since the main purpose of this study is to develop a robust ica inference procedure ,
the robustness for both data prewhitening and independent source recovery should be guaranteed . here
we will utilize the @xmath1-divergence to introduce a robust prewhitening method called @xmath1-prewhitening , followed by illustrating @xmath1-ica based on the prewhitened data . in practice , the @xmath1 value for @xmath1-divergence should also be determined . in the rest of discussion
, we will assume @xmath1 is given , and leave the discussion of its selection to section [ sec.select_gamma ] .
although prewhitening is always possible by a straightforward standardization of @xmath9 , there exists the issue of robustness of such a whitening procedure .
it is well known that empirical moment estimates of @xmath145 are very sensitive to outliers . in mollah ,
eguchi and minami ( 2007 ) , the authors proposed a robust @xmath2-prewhitening procedure .
in particular , let @xmath146 be the probability density function of @xmath97-variate normal distribution with mean @xmath147 and covariance @xmath148 , and let @xmath149 be the empirical distribution based on data @xmath150 . with a given @xmath2 , mollah et al .
( 2007 ) proposed the following minimum @xmath2-divergence estimators @xmath151 and then suggested to use @xmath152 for whitening the data .
interestingly , @xmath152 can also be derived from the minimum @xmath1-divergence as @xmath153 at the stationarity of ( [ gamma.prewiten.obj ] ) , the solutions @xmath154 will satisfy @xmath155 where @xmath156 the robustness property of @xmath152 can be found in mollah et al .
we call the prewhitening procedure @xmath157 the @xmath1-prewhitening .
the whitened data @xmath32 then enter the @xmath1-ica estimation procedure .
we are now in the position to develop our @xmath1-ica based on the @xmath1-prewhitened data @xmath32 . as discussed in section [ sec.gamma.div ] , the @xmath35 estimator is derived from @xmath158 where @xmath159 and @xmath64 is the working model for @xmath38 . since @xmath33 , @xmath160 which does not involve @xmath35 .
thus , @xmath161 can be equivalently obtained via @xmath162 finally , the mixing matrix @xmath163 is estimated by @xmath164 .
let @xmath165 and @xmath166^\top,\;\ ; { \rm where}\;\;\phi_j(y)=\frac{d}{dy}\ln f_j(y).\ ] ] we have the following proposition .
[ prop.stationarity ] at the stationarity , the maximizer @xmath167 defined in ( [ gamma_ica_z0 ] ) will satisfy @xmath168^\top-\phi(\widehat w^\top z_i)\left[\widehat w^\top z_i\right]^\top\right\}=0.\end{aligned}\ ] ] from proposition [ prop.stationarity ] , it can be easily seen the robustness nature of @xmath1-ica : the stationary equation is a weighted sum with the weight function @xmath169 .
when @xmath170 , an outlier with extreme value will contribute less to the stationary equation . in the limiting case of @xmath171 , which corresponds to mle - ica , the weight @xmath169 becomes uniform and , hence , is not robust .
a critical point to the likelihood - based ica method is to specify a working model @xmath64 for @xmath38 .
a sufficient condition to ensure the consistency of mle - ica can be found in hyvrinen , karhnen and oja ( 2001 ) .
here the ica consistency means recovery consistency .
that is , an ica procedure is said to be recovery consistent if it is able to recover all the independent components .
note that the consistency of mle - ica does not rely on the correct specification of working model @xmath64 , but only on the positivity of @xmath172 $ ] , @xmath173 .
this subsection aims to investigate the consistency of @xmath1-ica for @xmath174 $ ] , where @xmath175 is some constant .
we will deduce necessary and sufficient conditions such that @xmath1-ica is recovery consistent .
the main result is summarized below .
[ thm.consistency ] assume the ica model ( [ ica.z ] ) .
assume the existence of @xmath176 for @xmath177 $ ] such that 1 .
@xmath178=0\,$ ] for all @xmath179 and all @xmath173 .
then , for @xmath179 , the associated @xmath1-ica is recovery consistent if and only if @xmath180 where @xmath181 , @xmath182 , @xmath183 $ ] , @xmath184 , @xmath185 $ ] , and @xmath186 $ ] . condition ( a ) of theorem [ thm.consistency ] can be treated as a weighted version of @xmath187 .
it is satisfied when @xmath188 is symmetrically distributed about zero , and when the model probability density function @xmath64 is an even function .
we believe condition ( a ) is not restrictive and should be approximately valid in practice .
notice that @xmath189 .
thus , for the validity of ( [ hessian_tangent ] ) , we must require that @xmath190 , and the effect of @xmath191 can be exceeded by @xmath190 .
fortunately , due to the coefficient @xmath192 , when @xmath1 is small , the effect of @xmath193 will eventually outnumber the effect of @xmath194 , so that @xmath195 can be ensured .
in this situation , the negative definiteness of @xmath196 mainly relies on the structure of @xmath197 .
moreover , a direct calculation gives @xmath198 to be a diagonal matrix with diagonal elements @xmath199 .
we thus have the following corollary . [ cor.consistency ]
assume the ica model ( [ ica.z ] ) . assume the existence of a small enough @xmath176 for @xmath177 $ ] such that 1 .
@xmath178=0 $ ] for @xmath179 , @xmath173 .
@xmath200 + e[f^\gamma(s)\{\phi_k(s_k)s_k-\phi_k'(s_k)s_j^2\}]>0 $ ] for @xmath179 , for all pairs @xmath201 , @xmath202 .
then , for every @xmath179 , the associated @xmath1-ica can recover all independent components . to understand the meaning of condition ( b )
, we first consider an implication of corollary [ cor.consistency ] in the limiting case of @xmath171 , which corresponds to the mle - ica . in this case ,
condition ( a ) becomes @xmath187 , which is automatically true by the model assumption of @xmath16 .
moreover , since @xmath203 , condition ( b ) becomes @xmath204 + e[\phi_k(s_k)s_k-\phi_k'(s_k)]>0,\quad \mbox{for all pairs $ ( j , k)$ , $ j\neq k$}.\label{consistency.mle}\end{aligned}\ ] ] a sufficient condition to ensure the validity of ( [ consistency.mle ] ) is @xmath205>0,\quad\forall j , \label{consistency.mle2}\ ] ] which is the same condition given in theorem 9.1 of hyvrinen , karhnen and oja ( 2001 ) for the consistency of mle - ica .
we should note that ( [ consistency.mle ] ) is a weaker condition than ( [ consistency.mle2 ] ) .
in fact , from the proof of theorem [ thm.consistency ] , ( [ consistency.mle ] ) is also a necessary condition .
one implication of ( [ consistency.mle ] ) is that , we can have at most one @xmath64 to be wrongly specified or at most one gaussian component involved , and mle - ica is still able to recover all independent components .
this can also be intuitively understood that once we have determined @xmath206 directions in @xmath207 , the last direction is automatically determined . however
, this fact can not be observed from ( [ consistency.mle2 ] ) which requires all @xmath64 to be correctly specified .
we summarize the result for mle - ica below .
assume the ica model ( [ ica.z ] ) .
then , mle - ica is recovery consistent if and only if @xmath208 + e[\phi_k(s_k)s_k-\phi_k'(s_k)]>0 $ ] for all pairs @xmath201 , @xmath202 . turning to the case of @xmath1-ica , condition ( b ) of corollary [ cor.consistency
] can be treated as a weighted version of ( [ consistency.mle ] ) with the weight function @xmath169 .
however , one should notice that the validity of @xmath1-ica has nothing to do with that of mle - ica , since there is no direct relationship between condition ( b ) and its limiting case ( [ consistency.mle ] ) .
in particular , even if ( [ consistency.mle ] ) is violated ( i.e. , mle - ica fails ) , with a proper choice of @xmath1 , it is still possible that condition ( b ) holds and , hence , the recovery consistency of @xmath1-ica can be guaranteed . by theorem [ thm.consistency ] , a valid @xmath1-ica procedure must correspond to @xmath195 , or equivalently , the maximum eigenvalue of @xmath196 , denoted by @xmath209 , must be negative .
how should one pick a @xmath210-interval so that @xmath179 is legitimate in the sense that @xmath211 ? our suggestion for a rule of thumb is as follows .
let @xmath212 be the empirical estimator of @xmath196 based on the estimated source @xmath213 , where @xmath214 .
the plot of @xmath215 then provides a guidance in determining @xmath210 , over which @xmath216 should be far away below zero . with the @xmath210-interval , a further selection procedure , introduced in section [ sec.select_gamma ] ,
can be applied to select an optimal @xmath1 value from @xmath210 .
it is confirmed in our numerical study in section 6 that , the interval @xmath210 , where @xmath217 , is quite wide , and the suggested rule does provide adequate choice of @xmath210 .
it also implies that the choice of @xmath176 in corollary [ cor.consistency ] is not critical , as @xmath176 is allowed to vary in a wide range and not limited to very small number .
it is the condition ( b ) that plays the most important role to ensure the recovery consistency of @xmath1-ica . by using @xmath2-divergence ,
mihoko and eguchi ( 2002 ) proposed @xmath2-ica to recover independent components .
the objective function of @xmath2-ica being maximized is of the form @xmath218 where @xmath219 is a known constant .
if we restrict @xmath33 , then @xmath220 and maximizing ( [ div.beta ] ) is equivalent to maximizing @xmath221 , which has the same form with the population objective function of @xmath1-ica in ( [ gamma_ica_z0 ] ) .
we should emphasize that mihoko and eguchi ( 2002 ) considered the ica problem under the original @xmath9-scale , while the constraint @xmath33 is a consequence of prewhitening . without considering the constraint @xmath33 ,
the objective function of @xmath1-ica is deduced to be @xmath222 which is different from ( [ div.beta ] ) .
however , ( [ div.beta ] ) is similar to ( [ div.gamma ] ) when @xmath219 is small .
this fact also confirms the observation of mihoko and eguchi ( 2002 ) that setting @xmath223 does not affect the performance of @xmath2-ica . in summary ,
@xmath1-ica and @xmath2-ica based on the whitened data @xmath11 are equivalent . for data @xmath9 in original scale ,
however , @xmath1-ica maximizing ( [ div.gamma ] ) is different from @xmath2-ica maximizing ( [ div.beta ] ) , but they will have similar performance for small @xmath2 .
in this section , we introduce an algorithm for estimating @xmath35 constrained to the special orthogonal group @xmath30 , which is a lie group and is endowed with a manifold structure .
is a lie group if the group operations @xmath225 defined by @xmath226 and @xmath227 defined by @xmath228 are both @xmath229 mappings ( boothby , 1986 ) . ]
the lie group @xmath30 , which is a path - connected subgroup of @xmath230 , consists of all orthogonal matrices in @xmath25 with determinant one .
is that @xmath230 itself is not connected . in the case that the desired orthogonal matrix @xmath35 has determinant @xmath231 , our algorithm in fact searches for @xmath232 for some permutation matrix @xmath74 with @xmath233 . ]
recall @xmath234 being the objective function of @xmath1-ica maximization problem defined in ( [ gamma_ica_z0 ] ) .
a desirable algorithm is to generate an increasing sequence @xmath235 with @xmath236 , such that @xmath237 converges to a local maximizer @xmath238 of @xmath234 .
various approaches can be used to generate such a sequence @xmath237 in @xmath30 , for instance , geodesic flows and quasi - geodesic flows ( nishimori and akaho , 2005 ) . here
we focus on geodesic flows on @xmath30 . in particular , starting with the current @xmath239 , the update @xmath240 is selected from one geodesic path of @xmath239 along the steepest ascent direction such that @xmath241 . in fact , this approach has been applied to the general stiefel manifold ( nishimori and akaho , 2005 ) .
below we briefly review the idea and then introduce our implementation algorithm for @xmath1-ica .
we note that the proposed algorithm is also applicable to mle - ica by changing the corresponding objective function .
let @xmath242 denote the tangent space of @xmath30 at @xmath35 .
consider a smooth path @xmath243 on @xmath30 with @xmath244 .
differentiating @xmath245 yields the tangent space at @xmath35 @xmath246 clearly , @xmath247 is the set of all skew - symmetric matrices .
each geodesic path starting from @xmath248 has an intimate relation with the matrix exponential function .
in fact , @xmath249 if and only if @xmath250 is skew - symmetric ( see page 148 in boothby , 1986 ; proposition 9.2.5 . in marsden and ratiu , 1998 ) . moreover , for any @xmath251 , there exists ( not unique ) a skew - symmetric @xmath250 such that @xmath252 .
if the killing metric ( nishimori and akaho , 2005 ) @xmath253 is used , the geodesic path starting from @xmath248 in the direction @xmath250 is given by @xmath254 since the lie group is homogeneous , we can compute the gradient and geodesic at @xmath255 by pulling them back to the identity @xmath248 and then transform back to @xmath239 . in the implementation algorithm , to ensure all the iterations lying on the manifold @xmath30 , we update @xmath240 through @xmath256 where the skew - symmetric matrix @xmath257 and the step size @xmath258 are chosen properly to meet the ascending condition @xmath241 . since , from ( [ geodesic.ip ] ) , @xmath259 lies on the geodesic path of @xmath248 , then @xmath260 must lie on the geodesic path of @xmath239
. moreover , since @xmath261 by @xmath262 , the sequence in ( [ w.update ] ) satisfies @xmath263 for all @xmath264 .
the determination of the gradient direction @xmath257 and the step size @xmath258 is discussed below . to compute the gradient and geodesic at @xmath239 by pulling them back to @xmath248 , define @xmath265
we then determine @xmath260 from one geodesic at @xmath248 in the direction of the projected gradient of @xmath266 .
specifically , to ensure the ascending condition , we choose each skew - symmetric @xmath257 to be @xmath267 , the projected gradient of @xmath266 at @xmath248 , defined to be @xmath268^\top-\phi(w_k^\top z_i)\left[w_k^\top z_i\right]^\top\right\}.\end{aligned}\ ] ] this particular choice of @xmath257 ensures the existence of the step size @xmath258 for the ascending condition .
note that in the case of @xmath30 imposed with the killing metric , the projected gradient coincides with the natural gradient introduced by amari ( 1998 ) .
see also fact 5 in nishimori and akaho ( 2005 ) for further details . as to the selection of the step size @xmath258 at each iteration
@xmath264 with @xmath239 and @xmath269 , we propose to select @xmath258 such that @xmath270 is the first improved rotation " . in particular , we consider @xmath271 for some @xmath272 and @xmath273 , where @xmath274 is the nonnegative integer .
to proceed , we search @xmath274 such that @xmath275 and then update @xmath276 . in our implementation , @xmath277 and @xmath278 are used . for the convergence issue
, one can instead consider the armijo rule for @xmath258 ( given in equation ( [ armijo ] ) ) .
our experiments show that the above first improved rotation " rule works quite well .
lastly , in the implementation , to save the storage for @xmath239 , we rotate @xmath279 directly " instead of manipulating @xmath35 , where @xmath279 is the @xmath280 data matrix whose columns are @xmath281 , @xmath282 .
that is , we use the update @xmath283 . to retrieve the matrix @xmath35
, we simply do a matrix right division of the final @xmath279 and the initial @xmath279 .
the algorithm for @xmath1-ica based on gradient ascend on @xmath30 is summarized below .
+ ' '' '' 1 .
initialization : @xmath277 , @xmath278 , prewhitened data @xmath284 ( @xmath285 matrix ) .
2 . for each iteration @xmath286 ,
* compute the skew - symmetric matrix @xmath257 in ( [ v ] ) . * for @xmath287 , if @xmath288 , then break the loop .
* update @xmath289 by @xmath290 .
check the convergence criterion .
if the criterion is not met , go back to ( i ) .
3 . output @xmath291 . ' '' '' + finally , we would like to mention the convergence issue .
the statement is similar to proposition 1.2.1 of bertsekas ( 2003 ) .
[ thm.armijo ] let @xmath234 be continuously differentiable on @xmath30 , and @xmath292 be defined in ( [ f ] ) .
let @xmath293 be a sequence generated by @xmath294 , where @xmath257 is a projected gradient related ( see ( [ projected ] ) below ) and @xmath258 is a properly chosen step size by the armijo rule : reduce the step size @xmath295 , @xmath296 until the inequality holds for the first nonnegative @xmath274 , @xmath297 where @xmath298 is a fixed constant .
then , every limit point @xmath299 of @xmath293 is a stationary point , i.e. , @xmath300 for all @xmath301 , or equivalently , @xmath302 .
the statement that @xmath257 is a projected gradient related corresponds to the condition @xmath303 this condition is true when @xmath257 is the projected gradient @xmath304 itself or some natural gradient @xmath305 ( theorem 1 , amari , 1998 ) , where @xmath18 is a riemannian metric tensor , which is positive definite .
the estimation process of @xmath1-ica consists of two steps : @xmath1-prewhitening and the geometry - based estimation for @xmath35 , in which the values of @xmath1 are essential to have robust estimators .
hence , we carefully select the value of @xmath1 based on the adaptive selection procedures proposed by minami and eguchi ( 2003 ) and mollah et al .
we first introduce a general idea and then apply the idea to the selection of @xmath1 in both @xmath1-prewhitening and @xmath1-ica . define the measurement of generalization performance as @xmath306,\ ] ] where @xmath307 is the underlying true joint probability density function of the data , @xmath308 is the considered model for fitting , @xmath309 is the minimum @xmath1-divergence estimator of @xmath310 , and @xmath311 is the empirical estimate of @xmath307 .
the @xmath312 is called the anchor parameter and is fixed at @xmath313 throughout this paper .
this value is empirically shown to be insensitive to the resultant estimators ( minami and eguchi , 2003 ) .
let @xmath314 be the sample analogue of @xmath315 .
we propose to select the value of @xmath1 over a predefined set @xmath210 through @xmath316 for @xmath1-prewhitening , @xmath317 and @xmath318 with @xmath319 . for @xmath1-ica , @xmath320 and @xmath321 with @xmath322 .
the above selection criterion requires the estimation of @xmath315 . to avoid the problem of overfitting
, we apply a @xmath323-fold cross - validation .
let @xmath324 be the whole data , and let @xmath323 partitions of @xmath324 be @xmath325 , that is , @xmath326 if @xmath327 and @xmath328 . the whole selection procedure is summarized below
. + ' '' '' 1 . for @xmath329 , 1 .
for every @xmath179 , obtain @xmath330 , where @xmath331 is the empirical estimate of @xmath307 based on @xmath332 .
2 . compute the cross validation estimate @xmath333 , where @xmath334 is the empirical estimate of @xmath307 based on @xmath335 .
estimate @xmath315 by @xmath336 and obtain @xmath337 . ' '' ''
+ eventually , we have two optimal values of @xmath1 : @xmath338 for @xmath1-prewhitening and @xmath339 for estimation of the recovering matrix @xmath35 .
we conduct two numerical studies to demonstrate the robustness of the @xmath1-ica procedure . in the first study ,
the data is generated from independent sources with some known distributions . in the second study ,
we use transformations of lena images to form mixed image .
we independently generate the two sources @xmath188 , @xmath340 , from a non - gaussian distribution with sample size @xmath341 .
the observable @xmath9 is then given by @xmath342 , where @xmath343.\end{aligned}\ ] ] among the @xmath344 observations , we add to each of the last @xmath345 observations a random noise @xmath346 .
the data thus contains @xmath347 uncontaminated i.i.d .
observations from the ica model , @xmath342 , and @xmath345 contaminated i.i.d . observations from @xmath348 , where @xmath349 with @xmath350 and @xmath351 .
we consider two situations for the independent source @xmath352 : 1 .
uniform source : each @xmath188 , @xmath340 , is generated from uniform@xmath353 .
student-@xmath354 source : each @xmath188 , @xmath340 , is generated from @xmath354-distribution with 3 degrees of freedom .
for the case of uniform source , we use the sub - gaussian model @xmath355 with @xmath356 , which ensures the variance under @xmath64 is close to unity . as to the case of @xmath354 source , the super - gaussian model @xmath357 is considered , and we follow the suggested range of hyvrinen and oja ( 2000 ) and set @xmath358 .
we also implement mle - ica ( using the geometrical algorithm introduced in section 4 ) and fast - ica ( using the code available at ` http://www.cis.hut.fi/projects/ica/fastica/ ` ) based on the @xmath1-prewhitened data for fair comparisons . to evaluate the performance of each method ,
we modify from the performance index of parmar and unhelkar ( 2009 ) by a rescaling and by replacing the 2-norm with 1-norm and define the following performance index @xmath359 where @xmath360 is the @xmath361-th element of @xmath362 .
we will expect @xmath74 to be a permutation matrix , when the method performs well .
in that situation , the value of @xmath363 should be very close to 0 , and attains @xmath364 if @xmath74 is indeed a permutation matrix .
simulation results with 100 replications are reported in figure [ fig.sim ] .
for the case of no outliers ( @xmath365 ) , all three methods perform well except the performance index @xmath363 of @xmath1-ica increases as @xmath1 increases .
this is reasonable since , according to theorem [ thm.consistency ] , @xmath1-ica may fail to apply when @xmath1 is too large .
however , this influence is not severe as the the performance index @xmath363 is slightly increased only . as to the case of involving outliers ( @xmath366 ) ,
it can be seen that the proposed @xmath1-prewhitening followed by @xmath1-ica does possess the advantage of robustness for a wide range of @xmath1 values , while the other two methods are not able to recover the latent sources .
the performance of @xmath1-ica becomes worse when @xmath1 is small , since in the limiting case @xmath171 , @xmath1-ica reduces to the non - robust mle - ica .
we note that both @xmath1-prewhitening and @xmath1-ica are critical .
this can be seen from the poor performance of mle - ica and fast - ica , even they use the @xmath1-prewhitened data as the input .
indeed , @xmath1-prewhitening only ensures that we shift and rotate the data in a robust manner , while the outliers will still enter into the subsequent estimation process and , hence , a non - robust result is expected . in figure [ fig.sim.scatter ]
we report the scatter plots of the recovered sources @xmath367 from each method , of @xmath9 , and of @xmath368 for one simulation run ( @xmath366 ) .
these plots still convey the same message that @xmath1-ica is the winner among three methods , where the pattern of the reconstructed sources from @xmath1-ica is the most close to that of @xmath368 .
we use the lena picture to evaluate the performance of @xmath1-ica . in our experiment
, we use the lena image with @xmath369 pixels .
we construct four types of lena as the latent independent sources @xmath16 as shown in figure [ fig.lena ] .
we randomly generate the mixing matrix to be @xmath370 , where the elements of @xmath371 are independently generated from uniform@xmath372 .
the observed mixed pictures are also placed in figure [ fig.lena ] , wherein about @xmath373 of the pixels are added with random noise generated from @xmath374 for contamination .
the aim of this data analysis is to recover the original lena pictures based on the observed contaminated mixed pictures . in this analysis ,
the pixels are treated as the random sample , each with dimension 4 .
we randomly select @xmath375 pixels to estimate the demixing matrix , and then apply it to reconstruct the whole source pictures .
we conduct two scenarios to evaluate the robustness of each method : 1 . using the original image @xmath9 as the input ( see the second row of figure [ fig.lena ] ) .
2 . using the filtered image @xmath376 from @xmath9 as the input ( see the third row of figure [ fig.lena ] ) .
the filtering process in the second scenario can be treated as a pre - processing to alleviate the influence of additive gaussian noise . in both scenarios ,
the estimated demixing matrix is applied to the original images @xmath9 to recover @xmath16 .
note that with gaussian noise contamination , conventional prewhitening by empirical moment estimators is not robust and , hence , both fast - ica and mle - ica may fail to apply .
therefore , we prewhiten the data by @xmath1-prewhitening first and then apply @xmath1-ica , mle - ica , and fast - ica to the same whitened data for fair comparison .
the plot @xmath377 introduced in the end of section 3.3 is placed in figure [ fig.eig ] , which suggests that @xmath378 $ ] is a good candidate for possible @xmath1 values .
we then apply the cross - validation method developed in section 5 to determine the optimal @xmath379 .
the estimated values of @xmath314 are plotted in figure [ fig.cv ] , from which we select @xmath380 for @xmath1-prewhitening and @xmath381 for @xmath1-ica .
the recovered pictures are placed in figures [ fig.lena_gamma_recover]-[fig.lena_fast_recover ] , where for each figure the first row is for scenario-1 and the second row is for scenario-2 . it can be seen that @xmath1-ica is the best performer under both scenarios , while mle - ica and fast - ica can not recover the source images well when data is contaminated .
it also demonstrates the applicability of the proposed @xmath1-selection procedure .
we detect that mle - ica and fast - ica perform better when using filtered images @xmath376 , but can still not reconstruct images as good as @xmath1-ica does .
interestingly , @xmath1-ica has a reverse performance , where the best reconstructed images are from the original images instead of the filtered ones .
the filtering process , which aims to achieve robustness , replaces the original pixel value by the median of the pixel values over its neighborhood .
therefore , while filtering process will alleviate the influence of outlier , it is also possible to lose useful information at the same time .
for instance , a pixel without being contaminated will still be replaced by certain median value during the filtering process .
@xmath1-ica , however , works on the original data @xmath9 that possesses all the information available , and then weights each pixel according to its observed value to achieve robustness .
hence , a better performance for @xmath1-ica based on the original images is reasonably expected .
in this paper , we introduce a unified estimation framework by means of minimum @xmath0-divergence . for the reason of robustness consideration , we further focus on the specific choice of @xmath1-divergence , which gives the proposed @xmath1-ica inference procedure .
statistical properties are rigorously investigated . a geometrical algorithm based on gradient flows on orthogonal group
is introduced to implement our @xmath1-ica .
the performance of @xmath1-ica is evaluated through synthetic and real data examples .
there are still many important issues that are not covered by this work .
for example , we only consider full ica problem , i.e. , simultaneous extraction of all @xmath97 independent components , which is unpractical in the case of large @xmath97 .
it is of interest to extend our current @xmath1-ica to partial @xmath1-ica .
another issue of interest is also related to the large-@xmath97-small-@xmath344 scenario . in this work , data have to be prewhitened before entering the @xmath1-ica procedure .
prewhitening can be very unstable especially when @xmath97 is large . how to avoid such a difficulty is an interesting and challenging issue .
tensor data analysis is now becoming popular and attracts the attention of many researchers .
many statistical methods include ica have been extended to deal with such a data structure by means of multilinear algebra techniques .
an extension of @xmath1-ica to a multilinear setting to cover tensor data analysis is also of great interest for future study .
eguchi , s. ( 2009 ) .
information divergence geometry and the application to statistical machine learning . in _ information theory and statistical learning _ , f. emmert - streib and m. dehmer ( eds . ) , 309 - 332 .
springer , berlin .
parmar , s. d. and unhelkar , b. ( 2009 ) . performance analysis of ica algorithms against multiple - sources interference in biomedical systems .
_ international journal of recent trends in engineering _ , 2 , 19 - 21 . ,
and for the true sources @xmath16 .
in each plot , the red dots are observations without contamination , and the blue pluses are contaminated ones .
( a)-(e ) : uniform source ( scenario-1 ) , ( f)-(j ) : @xmath354 source ( scenario-2 ) .
, title="fig : " ] , and for the true sources @xmath16 . in each plot , the red dots are observations without contamination , and the blue pluses are contaminated ones .
( a)-(e ) : uniform source ( scenario-1 ) , ( f)-(j ) : @xmath354 source ( scenario-2 ) .
, title="fig : " ] with @xmath313 for ( a ) @xmath1-prewhitening and ( b ) @xmath1-ica for the lena data analysis .
the red dot indicates the place where the minimum value is attained.,title="fig:",height=226 ] with @xmath313 for ( a ) @xmath1-prewhitening and ( b ) @xmath1-ica for the lena data analysis . the red dot indicates the place where the minimum value is attained.,title="fig:",height=226 ] | independent component analysis ( ica ) has been shown to be useful in many applications . however , most ica methods are sensitive to data contamination and outliers . in this article
we introduce a general minimum @xmath0-divergence framework for ica , which covers some standard ica methods as special cases . within the @xmath0-family
we further focus on the @xmath1-divergence due to its desirable property of super robustness , which gives the proposed method @xmath1-ica .
statistical properties and technical conditions for the consistency of @xmath1-ica are rigorously studied . in the limiting case
, it leads to a necessary and sufficient condition for the consistency of mle - ica .
this necessary and sufficient condition is weaker than the condition known in the literature .
since the parameter of interest in ica is an orthogonal matrix , a geometrical algorithm based on gradient flows on special orthogonal group is introduced to implement @xmath1-ica .
furthermore , a data - driven selection for the @xmath1 value , which is critical to the achievement of @xmath1-ica , is developed .
the performance , especially the robustness , of @xmath1-ica in comparison with standard ica methods is demonstrated through experimental studies using simulated data and image data . + * key words and phrases * : @xmath2-divergence ; @xmath1-divergence ; geodesic ; independent component analysis ; minimum divergence ; robust statistics ; special orthogonal group . |
the minimal supersymmetric standard model ( mssm ) provides the most attractive and best studied extension of the standard model ( sm ) .
the mssm predicts the existence of not yet observed supersymmetric partners to the sm particles . among those
there are the sfermions @xmath0 , the scalar partners of the sm fermions , the neutralinos @xmath1 , the fermionic partners of the photon , @xmath2-boson and the neutral higgs bosons and last but not least the charginos @xmath3 , the fermionic partners of the @xmath4-bosons and the charged higgs bosons .
it is expected that some of these particles will be detected at the lhc or the tevatron . to determine their precise properties we will have to wait for a linear @xmath5 collider .
the future linear collider should allows measurements with high precision . after measuring the masses and the couplings of the supersymmetric particles
the main focus will be the extraction of the fundamental parameters of the mssm . although straightforward at tree - level , it is no longer simple when higher orders are included which are necessary in view of the precision of the linear collider .
the definition of the parameters is not unique beyond the leading order and depends on the renormalization scheme .
therefore , a well - defined theoretical framework has been recently proposed within the so - called spa ( susy parameter analysis ) project @xcite .
the `` spa convention '' provides a clear base for calculating masses , mixing angles , decay widths and production processes .
it also provides a clear definition of the fundamental parameters using the ( dimensional reduction ) renormalization scheme which allows one to extract them from future data . here
we present the one - loop results for the neutralino , chargino and sfermion ( of the 3rd generation ) production processes which were calculated in the spa convention mentioned above ( for details and other references see @xcite ) .
despite the fact that the parameters in the spa convention are defined in the renormalization scheme , we have used an on - shell renormalization scheme .
we show how one can obtain results full in accord with the spa convention using the on - shell renormalization scheme . for this purpose ,
the parameters are transformed into on - shell parameters which can then be used as input in the calculation .
this transformation is presented here in some detail . at the end ,
we show numerical results for the pair production of stops , sbottoms , staus , charginos and neutralinos in @xmath5 collisions .
the precision achieved by the linear collider implies the need for next - to - leading order predictions to be included . in principle one
is free to choose any consistent renormalization scheme for the next - to - leading order corrections .
we show here that by using any such scheme one can still make a full use of the spa conventions .
one just has to transform the parameters in a consistent way . from now on we focus on transforming the parameters from the to the on - shell renormalization scheme .
as an example , we take the sfermion mass matrix ( see fig .
[ figpara ] ) which is a non - diagonal matrix at the beginning .
there are two equivalent ways how to obtain the sfermion pole masses ( as one can see in fig .
[ figpara ] ) .
first , in the scheme one diagonalizes the spa input matrix to obtain the masses . to get the pole masses one has to subtract the finite parts of the on - shell counterterms @xmath6 .
( 178,51)(0,0 ) ( 10,5 ) ( 135,25 ) [ cols="^,^,^",options="header " , ] if one uses the on - shell renormalization scheme , one first has to transform the spa input matrix into the on - shell input matrix .
this is done by subtracting the finite parts of the on - shell counterterms which correspond to the parameters found in the mass matrix . however , there is a subtlety hidden . as some parameters are common to more than one mass matrix ( e.g. @xmath7
is found in the stop and in the sbottom mass matrix ) , there are more possibilities how to define these parameters .
this results in more possible counterterms for one parameter which one can use in the transformation .
if one chooses one particular definition for the parameter ( say we define @xmath7 in the sbottom mass matrix ) one has to take into account finite shifts @xmath8 for the same parameter appearing elsewhere ( i. e. we use @xmath9 in the stop sector ) .
there are two different approaches how to define these shifts which can be found in @xcite , where we follow the approach given in @xcite .
the advantage of using the on - shell input values is that the well - established procedure of on - shell renormalization can be applied .
we use the transformation mentioned above for the sps1a benchmark point which we need for the numerical analysis of the production processes . the table with the values of the parameters in different schemes
can also be found in fig .
[ figpara ] .
the plots in fig . [ figneuchar ] and [ figsfer ] depict the total cross - sections of the complete @xmath10 and leading higher order results for the neutralino , chargino and 3rd generation sfermion production processes ( as presented in detail in @xcite ) .
the numerics was done using the sps1a benchmark point and the tree - level results were done also in accord with the convention i. e. using couplings and on - shell masses . in the case of neutralino and chargino production extra attention
is paid to the problem of the separation of the qed and weak corrections .
the approach applied here is identical with the one specified in the spa project ( see @xcite ) , where the weak corrections are defined as @xmath11 with @xmath12 taking the cut - off dependent terms from final state radiation and inital / final state radiation interference into account ( the initial state radiation contribution is contained in the remaining terms ) .
using this definition one can separate the weak and the qed corrections in a gauge invariant and cut - independent way .
the qed corrections are further divided into two parts , the universal one @xmath13 fully given by a structure function approach and the remainder @xmath14 ( for details see @xcite ) .
the sfermion production cross - sections are presented in fig .
[ figsfer ] and the emphasis is put on showing the effect of the polarization of the @xmath5 beams .
as one can see the polarization has a significant effect on both the tree - level cross - sections and the one - loop corrections . | recent high precision calculations for production processes of the susy particles at the next linear collider are reviewed .
special attention is paid to the input parameter definition .
numerical results for the sps1a benchmark point as proposed in the spa project , are presented . |
electron photon scattering , @xmath11 studied in high energy @xmath12 collisions with tagged electron is an analogue of the inelastic lepton
nucleon scattering . here
the probe a virtual photon of four momentum @xmath13 ( @xmath14 ) , tests the target particle , the real photon of four momentum @xmath15 ( @xmath16 ) .
the corresponding spin averaged cross section , fig.1 , can be parametrized e.g. by the photon structure functions @xmath17 and @xmath18 .
the bjorken parameter @xmath2 is conventionally defined as @xmath19 .
thus the process ( 1 ) permits an insight into the inner structure of the real photon .
+ [ fig : fig1 ] at large @xmath1 the photon structure function is described by the perturbative qcd @xcite . however in the low @xmath1 region , @xmath20 1
gev@xmath21 , it is expected that the vector meson dominance ( vmd ) contribution @xcite is important . in this paper
we present a model of the photon structure function @xmath5 which includes both the vmd contribution and the qcd term , suitably extrapolated to the low @xmath1 region .
this approach is based on the extension of a similar representation of the nucleon structure function @xcite to the case of the photon .
possible parametrisations of the photon structure function which extend to the low @xmath1 region have also been discussed in ref .
the parametrisation proposed in ref.@xcite is based upon the quark parton model supplemented by the contribution from the hadronic structure of the photon .
the energy dependence of the latter has a regge form .
the @xmath1 dependence is parametrised in terms of the simple form - factors which if combined with the regge - type energy dependence generate at large @xmath1 the bjorken scaling behaviour of the corresponding part of the structure function . in ref .
@xcite the energy dependence of the cross - section is also parametrised in a regge - like form with the @xmath1 dependence specified by the suitable form factors which contain terms corresponding to the vmd contribution .
the parametrisation discussed in @xcite is based upon a model corresponding to the interaction of colour dipoles , i.e. the @xmath22 pairs which the photon(s ) fluctuate into . in our approach the contribution coming from the light vector mesons within the vector meson dominance model is similar to that used by other authors ( see eg .
@xcite ) although the details concerning estimate of the relevant total cross - sections are slightly different .
the novel feature of our model is the treatment of the contribution coming from high masses of the hadronic states which couple to the virtual photons . in our scheme
this contribution is directly related to the photon structure function in the large @xmath1 region .
the low and high mass hadronic states are separated at @xmath23 , a parameter whose value was taken identical with that for the @xmath24 .
our framework permits to describe the @xmath25 and @xmath26 total cross sections as functions of energy .
the energy dependence of @xmath27 is also described by other models , @xcite .
most of them incorporate the regge - like parametrisation of the total @xmath28 cross - sections ; some provide a detailed insight into the structure of final states and a decomposition of the @xmath29 total cross - section into terms corresponding to the appropriate subdivision of photon interactions and event classes @xcite .
possibility that part of the @xmath28 cross - section is driven by the production of minijets has been discussed in @xcite .
certain approaches analyse the behaviour of the cross - sections on the virtualities of both interacting photons @xcite .
the content of our paper is as follows : in the next section we recall the qcd description of the photon structure functions and in section 3 we briefly describe the vector meson dominance model in the process @xmath30 . in section 4
we present a parametrisation of the photon structure function , as well as the total cross section for the interaction of two real photons and of the virtual and real photon . in section 5
we compare our theoretical predictions with the experimental data on the @xmath0 and on the total cross sections @xmath3 and @xmath4 .
we also give predictions for @xmath3 in the very high energy range which can become accessible in future linear colliders .
finally in section 6 we give the summary of our results .
in the large @xmath1 , i.e. in the deep inelastic limit the virtual photon probes the quark ( antiquark ) structure of the ( real ) photon in analogy to the deep inelastic lepton hadron scattering , fig.2 . [
fig : fig2 ] the corresponding photon structure function @xmath18 may thus be related to quark and antiquark distributions @xmath31 , @xmath32 in the photon : f_2^(x , q^2 ) = x _ i e_i^2 [ q_i^(x , q^2 ) + |q_i^(x , q^2 ) ] , [ eq : a1 ] where @xmath33 denote the charges of quarks and antiquarks and the sum is over all active quark flavours . to be precise equation ( [ eq : a1 ] ) holds in leading logarithmic approximation of perturbative qcd .
it acquires higher order corrections in next - to - leading approximation and beyond @xcite .
a special feature of the quark structure of the photon with respect to the proton is a possibility of a direct @xmath34 production through the process @xmath35 , see fig.3 , leading to the parton model predictions for the @xmath36 and @xmath37 . a contribution of this process introduces an inhomogeneous term into the equation describing the qcd evolution of the quark ( and antiquark ) distributions in the photon .
[ fig : fig3 ] the process @xmath35 ( modified by the qcd evolution ) with the pointlike quark coupling both to real and virtual photons dominates in the large @xmath1 limit making the photon structure functions exactly calculable in this limit @xcite .
striking features of these functions are : * @xmath38 rise with increasing @xmath2 at large @xmath2 ; * @xmath38 show the scaling violation , @xmath39 . at low values of @xmath2
the dominant role in the photon structure functions is played by gluons .
the situation here is similar to that of the hadronic structure functions which exhibit a very strong increase with decreasing @xmath2 , see e.g. @xcite . those effects are still rather weak in the kinematical region of @xmath5 probed by present experiments but they will be very important in the regime accessible in the future linear @xmath40 ( @xmath41 , @xmath25 ) colliders @xcite .
besides the direct , point - like coupling to quarks , the target photon can fluctuate into vector mesons and other hadronic states which can also have their partonic structure .
the latter can not be calculated perturbatively and thus has to be parametrised phenomenologically @xcite .
finally it should be pointed out that the charm quark playing the dominant role in the heavy quark contributions to @xmath18 , is often described just by the lowest order bethe - heitler cross - section for the process @xmath42 and the additional contribution generated by the radiation @xmath43 @xcite .
the qcd describes the photon structure functions in the large @xmath1 region . in the low @xmath1 region
however one expects that the vmd mechanism is important . by the vector meson dominance mechanism in this case
we understand the model in which the virtual photon of virtuality @xmath1 fluctuates into vector mesons which next undergo interaction with the ( real ) photon of virtuality @xmath47 , see fig.4 . in order to be able to describe the photon structure function for arbitrary values of @xmath1
it would be very useful to have a unifed scheme which contains both the vmd and the qcd contributions , the latter suitably extended to the region of low values of @xmath1 .
[ fig : fig4 ] this may be achieved by utilising the dispersive representation in @xmath1 of the structure function . to this aim
let us notice that the @xmath44 collision can be viewed as the interaction of a real photon target with a photon with virtuality @xmath1 which fluctuates onto general hadronic state , cf .
we consider the virtual photon first fluctuating onto the @xmath48 state and then interacting with the real photon . like in the @xmath49 scattering one can write the dispersion relation for the @xmath44 scattering as follows @xcite : @xmath50 _ q ( m^2 , m^2 ) a_(q ) - ( w^2 , m^2 , m^2 ) [ eq : aa1 ] where @xmath51 and @xmath52 are the invariant masses of the incoming and outgoing @xmath48 pair . in eq .
( [ eq : aa1 ] ) , @xmath53 is the density matrix of the @xmath48 states and @xmath54 is an imaginary part of the corresponding forward scattering amplitude .
the above formula can be rewritten in a form of a single dispersion relation as follows , [ a2 ] f_2^(w^2 , q^2 ) = q^2 _ 0^ ( w^2 , q^2 ) where the spectral function ( w^2,q^2 ) & = & 1 4 ^2 _ 0 ^ 1 d dm^2 dm^2 ( q^2-m^2-(1- ) m^2 ) + & & ( m^2,m^2 ) a_(q|q)-(w^2,m^2,m^2 ) .
[ sindisp ] the centre - of - mass energy squared @xmath55 is related to the bjorken parameter @xmath2 in the following way : w^2 = q^2 ( 1 x - 1 ) .
one can now separate regions of low- and high values of @xmath56 in the integral ( [ a2 ] ) , by noticing that this integral corresponds to the ( generalised ) vector meson dominance representation of the @xmath5 .
[ fig : fig5 ] for low values of @xmath56 , @xmath57 , one uses the vector meson dominance model . in this approach
one assumes that the virtual photon forms a vector meson rather than a pair of well separated @xmath58 .
the integrand in the integral on the right hand side of equation ( [ sindisp ] ) defining the spectral function @xmath59 is then given by the following formula : @xmath60 @xmath61 where @xmath62 is the mass of the vector meson @xmath63 and @xmath64 denotes the @xmath65 total cross section .
the couplings @xmath66 can be estimated from the leptonic widths of the vector mesons , @xmath67 in equation ( [ spfvm ] ) we have included only diagonal transitions between the vector mesons having the same masses .
the corresponding spectral function @xmath68 thus reads : @xmath69 the resulting vmd contribution to the photon structure function @xmath0 , @xmath70 , is given by the following equation : [ eq : a2 ] f_2^vmd ( x , q^2 ) = _ v . in eq.([eq :
a2 ] ) we only consider mesons with masses @xmath71 .
the contribution coming from the region of high values of @xmath56 ( @xmath72 ) can be related to the photon structure function from the large @xmath1 domain .
it defines the partonic contribution @xmath73 to the structure function @xmath5 , extended to arbitrary low values of @xmath1 . for convenience ,
we adopt the approximation used in ref .
@xcite which gives : f_2^partons(x , q^2 ) = f_2^qcd(|x , q^2 + q_0 ^ 2 ) [ f2partons ] where [ xbar ] |x = .
the structure function @xmath74 is taken from the qcd analysis , valid in the large @xmath1 region , i.e. it is calculated from the existing parametrisations of the parton distributions .
modifications of the qcd contribution : replacement of the parameter @xmath2 by @xmath75 defined in equation ( [ xbar ] ) , shift of the scale @xmath76 and the factor @xmath77 instead of 1 , introduce power corrections which vanish as @xmath78 and are negligible at large @xmath1 .
the magnitude of @xmath79 is set to @xmath80 as in the case of the proton @xcite .
the @xmath73 thus defines a contribution to @xmath5 at arbitrary low values of @xmath1 .
an elaborated treatment of the partonic contribution to the proton structure function has been developed in ref .
@xcite , where long and short distance components have been carefully separated .
according to that paper the low mass region is dominated by the @xmath48 pairs with large transverse sizes in the impact parameter space ( thus corresponding to the vmd ) whereas the qcd part is dominated by pairs of small transverse size .
our representation of the photon structure function @xmath81 is based on the following decomposition : @xmath82 where @xmath70 and @xmath83 are defined by eq .
( [ eq : a2 ] ) and ( [ f2partons ] ) .
a total @xmath44 cross - section in the high energy limit is given by @xmath84 with @xmath85 .
the @xmath86 ( for fixed @xmath87 ) limit of eq . ( [ sigma ] ) gives the total cross - section @xmath88 corresponding to the interaction of two real photons .
> from ( [ dec ] ) , ( [ eq : a2 ] ) and ( [ f2partons ] ) we obtain the following expression for this cross - section at high energy : @xmath89 at large @xmath1 the structure function given by eq .
( [ dec ] ) becomes equal to the qcd contribution @xmath90 .
the vmd component gives the power correction term which vanishes as @xmath78 for large @xmath1 .
it should be noted that the vmd part contains only finite number of vector mesons with their masses smaller than @xmath79 . in the quantitive analysis of the photon structure function and of the total cross sections we have taken the structure function @xmath74 from the lo analyses presented in ref .
@xcite ( grv ) and @xcite ( grs),with a number of active flavours equal four . the latter parton parametrization is based on updated data analysis and holds for both virtual and real photons .
the vmd part was estimated using the following assumptions : 1 .
numerical values of the couplings @xmath66 are the same as those used in ref .
they were estimated from the relation ( [ gammav ] ) which gives the following values : @xmath91 2 .
cross - sections @xmath92 are represented as sums of the pomeron and reggeon contributions : @xmath93 where @xmath94 @xmath95 with @xmath96 and @xmath971 gev@xmath21 @xcite .
pomeron couplings @xmath98 are related to the corresponding couplings @xmath99 controlling the pomeron contributions to the total @xmath100 cross - sections .
we assume the additive quark model and reduce the total cross - sections for the interaction of strange quarks by a factor 2 .
this gives : @xmath101 + @xmath102 4 .
reggeon couplings @xmath103 are estimated assuming the additive quark model and the duality ( i.e. a dominance of planar quark diagrams ) .
we also assume that the quark couplings to a photon are proportional to the quark charge with the flavour independent proportionality factor .
this gives : @xmath104 + @xmath105 5 .
couplings @xmath99 and @xmath106 are taken from the fit discussed in ref .
@xcite which gave : @xmath107 since we are using the regge description of total cross sections @xmath108 our approach can only work for large values of @xmath87 , @xmath109 , away from the resonance region .
in this section we compare our results for the real photon structure function @xmath110 and the two - photon cross sections , @xmath111 and @xmath112 with corresponding measurements . in some cases our predictions
have been extended to the region @xmath113 where the model may not be applicable .
theoretical curves were obtained using two different parametrisations for the structure function @xmath74 , grv @xcite and grs @xcite . in fig.6
we show predictions for the photon structure function based on equation ( [ dec ] ) plotted as the function of @xmath2 for different values of @xmath1 in the region of small @xmath1 . in fig.7 the @xmath1 dependence of the photon structure function for different values of @xmath2
is presented . which corresponds to the threshold energy in the reaction @xmath114 .
] we confront our theoretical results with existing experimental data @xcite .
measurements of @xmath0 are scarce , especially for low values of @xmath1 .
however it can be seen that our prediction reproduces well the data independently of the parametrisation ( grv or grs ) of @xmath74 used in the model .
irregular behaviour of the dashed lines observed in figs 6 and 7 at high values of @xmath2 is connected with the teatment of the charm contribution to @xmath5 in the grs approach . in fig.8
we compare our predictions with the data on @xmath115 .
[ fig : fig6 ] theoretical curves were obtained from equation ( [ sgamgam ] ) .
we show experimental points corresponding to the low energy region ( @xmath116 10 gev ) @xcite and the recent preliminary high energy data obtained by the l3 , opal and delphi collaborations at lep @xcite .
the representation ( [ sgamgam ] ) for the total @xmath28 cross - section describes the data reasonably well .
the result of the calculation based on grs parametrisation of @xmath74 is slightly higher and has a shallower minimum as compared to that based on grv parametrisation .
calculations using the latter give a good description of the shape of the energy dependence of the cross section although the overall normalisation seems to be about @xmath117 too large .
[ fig : fig7 ] it should be stressed that our prediction is essentially parameter free .
the magnitude of the cross - section is dominated by the vmd component yet the partonic part is also non - negligible . in particular the latter term
is responsible for generating a steeper increase of the total cross - section with increasing @xmath87 than that embodied in the vmd part which is described by the soft pomeron contribution .
the decrease of the total cross - section with increasing energy in the low @xmath87 region is controlled by the reggeon component of the vmd part ( see eqs .
( [ rpgv ] ) , ( [ rgv ] ) and ( [ lambda ] ) ) and by the valence part of the partonic contribution .
[ fig : fig8 ] in fig.9 we show predictions for the total @xmath28 cross - section as a function of the total centre - of - mass energy @xmath87 in the wide energy range including the energies that might be accessible in the future linear colliders . in this figure
we also show a decomposition of @xmath118 into its vmd and partonic components ( only grv parametrisation was used in this analysis ) . at very high energies these two terms exhibit different energy dependence .
the vmd part is described by the soft pomeron contribution which gives the @xmath119 behaviour with @xmath120 = 0.0808 , eq .
( [ lambda ] ) .
the partonic component increases faster with energy since its energy dependence reflects the increase of @xmath121 with decreasing @xmath75 generated by the qcd evolution @xcite .
[ fig : fig9 ] this increase is stronger than that implied by the soft pomeron exchange . as a result the total @xmath28 cross - section , which is the sum of the vmd and partonic components
does also exhibit stronger increase with the increasing energy than that of the vmd component .
it is however milder than the increase generated by the partonic component alone , at least for @xmath122 gev .
this follows from the fact that in this energy range the magnitude of the cross - section is still dominated by its vmd component .
we found that for sufficiently high energies @xmath87 the total @xmath123 cross - section @xmath124 described by eq .
( [ sgamgam ] ) can be parametrized by the effective power law dependence , @xmath125 , with @xmath126 slowly increasing with energy ( @xmath127 for 30 gev @xmath128 10@xmath129 gev ) .
[ fig : fig10 ] [ fig : fig11 ] in fig.10 we show the @xmath26 cross section for different bins of the centre - of - mass energy @xmath87 plotted versus @xmath1 , the virtuality of the probing photon .
our theoretical predictions based on equation ( [ sigma ] ) are compared with measurements by the tpc/2@xmath130 collaboration @xcite and the agreement between the two is very good . in fig.11
we show the @xmath26 cross section for large energies @xmath87 plotted versus @xmath1 ( only grv parametrsation was used here ) . at medium and large @xmath1
, @xmath4 decreases as @xmath78 ( modulo logaritmic corrections ) and for very small values ( @xmath131 ) it exhibits a flattening behaviour .
we have presented an extension of the representation developed for the nucleon structure function @xmath132 for arbitrary values of @xmath1 , @xcite , onto the structure function of the real photon .
this representation includes both the vmd contribution and the qcd component , obtained from the qcd parton parametrisations for the photon , suitably extrapolated to the region of low @xmath1 . in the @xmath1=0
limit the model gives predictions for the total cross section @xmath3 for the interaction of two real photons .
we showed that our framework is fairly successful in describing the experimental data on @xmath133 and on @xmath134 and on the photon structure function @xmath0 at low @xmath1 .
we also showed that one can naturally explain the fact that the increase of the total @xmath28 cross - section with increasing cm energy @xmath87 is stronger than that implied by soft pomeron exchange .
the calculated total @xmath28 cross - section exhibits an approximate power - law increase with increasing energy @xmath87 , i.e. @xmath135 with @xmath126 slowly increasing with energy : @xmath127 for 30 gev @xmath128 10@xmath129 gev .
this research was partially supported by the polish state committee for scientific research grants no .
2 p03b 089 13 , 2 p03b 014 14 , 2 p03b 184 10 and by the eu fourth framework programme `` training and mobility of researchers '' , network quantum chromodynamics and the deep structure of elementary particles , contract fmrx - ct98 - 0194 .
ams thanks foundation for polish science for financial support .
9999 e. witten , nucl.phys .
* b120 * ( 1977 ) 189 . c. peterson , t.f .
walsh and p.m. zerwas , nucl.phys . *
b174 * ( 1980 ) 424 ; p. zerwas , phys.rev . *
d10 * ( 1974 ) 1485 .
h. abramowicz , m. krawczyk , k. charchua , a. levy and u. maor , int .
a8 * ( 1993 ) 1005 and references therein .
report of the working group on @xmath28 physics , p. aurenche and g.a .
schuler ( convenors ) , proceedings of the workshop on physics at lep2 , cern yellow report 96 - 01 and hep - ph/9601317 , g. altarelli , t. sjstrand and p. zwirner ( editors ) .
m. krawczyk , acta phys . polon .
* b28 * ( 1997 ) 2659 ; m. krawczyk , a. zembrzuski and m. staszel , hep - ph/9806291 ; an extended version of this report is in ift 99 - 15 and has been submitted to phys .
rep . ; + p. nisius , hep - ex/9912049 .
bauer et al . , rev .
* 50 * ( 1978 ) 261 and references therein .
j. kwieciski and b. badeek , z.phys .
* c43 * ( 1989 ) 251 .
b. badeek and j. kwieciski , phys . lett .
* b295 * ( 1992 ) 263 .
martin , m.g . ryskin and a.m. stasto ,
j. * c7 * ( 1999 ) 643 ; nucl .
* 74 * ( 1999 ) 121 .
e. gotsman , a. levy and u. maor , z. phys .
* c40 * ( 1988 ) 117 . g. a. schuler , comput .
* 108 * ( 1998 ) 279 .
a. donnachie , h. g. dosch and m. rueter , phys.rev .
* d59 * ( 1999 ) 074011 .
j.r . forshaw and j.k .
storrow , phys . rev .
* d46 * ( 1992 ) 4955 ; phys . lett .
* b278 * ( 1992 ) 193 .
g. a. schuler and t. sjstrand , z.phys .
* c73 * ( 1997 ) 677 .
a. donnachie , h. g. dosch and m. rueter , eur .
j. * c13 * ( 2000 ) 141 .
a. corsetti , r.m .
godbole and g. pancheri , phys . lett . *
b435 * ( 1998 ) 441 and references therein ; r. m. godbole , a. grau and g. pancheri , iisc - cts-7 - 99 , hep - ph/9908220 ; r. m. godbole , a. grau , g. pancheri , hep - ph/9912395 , iisc - cts-8 - 99 .
a. donnachie and s. soldner - rembold , hep - ph/0001035 , freiburg - ehep-2000 - 01 .
e. gotsman et al . , hep - ph/0001080 , taup-2618 - 00 . m. glck , e.reya and a.vogt , phys.rev . * d46 * ( 1992 ) 1973 . m. gluck , e. reya and i. schienbein , phys.rev .
* d60 * ( 1999 ) 054019 .
a. m. cooper - sarkar , r. c. e. devenish and a. de roeck , int.j.mod.phys .
* a13 * ( 1998 ) 3385 .
a. vogt , proceedings of the international conference on the structure and interactions of the photon , photon 99 , freiburg , 1999 , nucl .
* 82 * ( 2000 ) 394 ; e. accomando et al .
, phys . rep .
* 299 * ( 1998 ) 21 .
g. a. schuler and t. sjstrand , z. phys .
* c68 * ( 1995 ) 607 .
a. donnachie and p.v .
landshoff , phys . lett . *
b296 * ( 1992 ) 227 .
pluto collaboration ; ch .
berger et al . , phys .
lett . * b142 * , ( 1984 ) 111 ; nucl . phys . *
b281 * 365 .
tpc/2@xmath130 collaboration ; h. aihara et.al . , z. phys . * c34 * ( 1987 ) 1 .
l3 collaboration ; m. acciarri et al .
, phys . lett .
* b 436 * ( 1998 ) 403 .
opal collaboration , physics note pn389 ( 1999 ) .
opal collaboration , k. ackerstaff et al .
, z.phys . *
c74 * ( 1997 ) 33 .
pluto collaboration ; ch .
berger et al . , phys . lett . * b149 * ( 1984 ) 421 ; z. phys . * c26 * ( 1984 ) 353 .
tpc/2@xmath130 collaboration ; h. aihara et al .
, phys . rev . *
d41 * ( 1990 ) 2667 .
md-1 collaboration ; s.e .
baru et al . , z. phys .
* c53 * ( 1992 ) 219 .
l3 collaboration ; g. passaleva , talk presented at the ix lomonosov conference on elementary particle physics , moscow , september 20 - 26 , 1999 ( to appear in the proceedings ) .
opal collaboration ; g. abbiendi et al .
, cern - ep/99 - 076 .
delphi collaboration ; n. zimin , proceedings of the international conference on the structure and interactions of the photon , photon 99 , freiburg , 1999 , nucl .
* 82 * ( 2000 ) 139 . | a parametrisation of the real photon structure function @xmath0 in the low @xmath1 , low @xmath2 region is formulated .
it includes both the vmd and the qcd components , the latter suitably extrapolated to the low @xmath1 region and based on arbitrary parton distributions in the photon .
the parametrisation used together with the grv and grs parton densities describes reasonably well the existing high energy data on @xmath0 , @xmath3 and the low @xmath1 data on @xmath4 .
predictions for @xmath3 and for @xmath4 for energies which may become accessible in future linear colliders are also given .
* parametrisation of @xmath5 at low @xmath1 and of @xmath6 and @xmath7 at high energies * + , m. krawczyk@xmath8 , j. kwieciski@xmath9 and a. m. stato@xmath9 + @xmath10 _ department of physics , uppsala university , p.o.box 530 , 751 21 uppsala , sweden _
+ _ and institute of experimental physics , warsaw university , 00 - 681 warsaw , poland _ + @xmath8 _ institute of theoretical physics , warsaw university , 00 - 681 warsaw , poland _ + @xmath9 _ department of theoretical physics , h. niewodniczaski institute of nuclear physics , + 31 - 342 cracow , poland _ + |
there is much ongoing experimental effort by the atlas and cms collaborations at the cern large hadron collider ( lhc ) to search for the neutral higgs boson ( @xmath15 ) of the standard model ( sm ) @xcite .
this model of spontaneous symmetry breaking will be tested at the lhc over all of the theoretically preferred mass range , in an experimental programme which is expected to be completed by the end of the @xmath16 tev run of the lhc .
after analysing all of the data taken at @xmath17 tev @xcite there are only two regions for the mass of @xmath15 which have not been excluded at 95% c.l : i ) a region of light mass , roughly corresponding to @xmath18 , and ii ) a region of heavy mass , @xmath19 gev . at present
there is much speculation about an excess of events suggesting a mass of around 125 gev , especially in the channel @xmath0 @xcite where the excess is a bit larger than that expected for a sm higgs boson of the same mass .
more data is needed to clarify if this small excess is a fluctuation of the background , or if it is genuine production of a higgs boson .
several recent studies @xcite fit the current data in all the higgs search channels to the case of a neutral higgs boson with arbitrary couplings .
a sm - like higgs boson gives a good fit to the data , although a slight preference for non - sm like higgs couplings is emphasised in @xcite . the higgs sector of the sm , which consists of one fundamental scalar with a vacuum expectation value ( vev ) , might not be nature s choice .
there could be additional scalar fields which also contribute to the masses of the fermions and weak bosons , with a more complicated scalar potential which depends on several arbitrary parameters .
importantly , even in the event of no signal for a sm - like higgs boson at the lhc , the search for scalar particles should continue in earnest due to the fact that a non - minimal higgs sector can give rise to different experimental signatures , some of which are challenging to detect .
consequently , it will take much longer for the lhc to probe all of the parameter space of such models .
various interpretations of the excess of events at around 125 gev have been proposed in the context of models with two higgs doublets in @xcite , with a review of such models in @xcite .
early studies of non - minimal higgs sectors which can give rise to an enhanced br@xmath20 can be found in @xcite ( in which the enhancement is caused by increasing the partial width of @xmath0 ) and in @xcite ( in which the dominate tree - level decays modes such as @xmath21 are suppressed . ) the higgs triplet model ( htm ) @xcite is a model of neutrino mass generation with a non - minimal higgs sector .
the model predicts several scalar particles , including a doubly charged higgs boson ( @xmath1 ) and a singly charged higgs boson ( @xmath22 ) , for which direct searches are being carried out at the lhc @xcite . in a large part of the parameter space of the htm
the lightest cp - even scalar ( @xmath5 ) has essentially the same couplings to the fermions and vector bosons as the higgs boson of the sm @xcite .
therefore the ongoing searches for the sm higgs boson also apply to @xmath5 of the htm with very little modification .
an exception is the loop - induced decay @xmath7 which receives contributions from virtual @xmath1 and @xmath22 , and can have a branching ratio which is very different to that of the sm higgs boson . as shown recently in @xcite , the ongoing limits on br(@xmath7 ) constrain the parameter space of @xmath11 $ ] , where @xmath9 is a quartic coupling in the scalar potential ( see also @xcite for a related study ) .
the case of @xmath12 was studied in @xcite , which leads to destructive interference between the combined sm contribution ( from @xmath23 and fermion loops ) and the contribution from @xmath1 . in this work
we consider the case of @xmath10 , which leads to constructive interference and was not considered in @xcite .
the scenario of @xmath10 is more constrained by the ongoing searches for @xmath7 than the case of @xmath12 . moreover , the scenario of @xmath10 can provide enhancements of @xmath7 with smaller @xmath24 than the case of @xmath12 .
our work is organised as follows . in section
ii we briefly describe the theoretical structure of the htm . in section
iii we present the formulae for the contribution of @xmath1 and @xmath22 to @xmath7 , and we summarise the ongoing searches for these particles at the lhc .
section iv contains our numerical results , with conclusions in section v.
in the htm @xcite a @xmath25 complex @xmath26 isospin triplet of scalar fields , @xmath27 , is added to the sm lagrangian .
such a model can provide majorana masses for the observed neutrinos without the introduction of @xmath26 singlet neutrinos via the gauge invariant yukawa interaction : @xmath28 here @xmath29 is a complex and symmetric coupling , @xmath30 is the dirac charge conjugation operator , @xmath31 are the pauli matrices , @xmath32 is a left - handed lepton doublet , and @xmath33 is a @xmath34 representation of the @xmath25 complex triplet fields : @xmath35 where @xmath36 , @xmath37 , and @xmath38 .
a non - zero triplet vev @xmath39 gives rise to the following mass matrix for neutrinos : @xmath40 the necessary non - zero @xmath41 arises from the minimisation of the most general @xmath42 invariant higgs potential @xcite , which is written as follows @xcite ( with @xmath43 ) : @xmath44 here @xmath45 in order to ensure non - zero @xmath46 which spontaneously breaks @xmath47 to @xmath48 while @xmath49 . the scalar potential in eq .
( [ potential ] ) together with the triplet yukawa interaction of eq .
( [ trip_yuk ] ) lead to a phenomenologically viable model of neutrino mass generation . for small @xmath50 ,
the expression for @xmath51 resulting from the minimisation of @xmath52 is : @xmath53 for @xmath54 one has @xmath55 , which would naturally lead to a small @xmath51 even for @xmath56 of the order of the electroweak scale ( and is sometimes called the `` type ii seesaw mechanism '' ) .
however , in this scenario the triplet scalars would be too heavy to be observed at the lhc . in recent years
there has been much interest in the case of light triplet scalars ( @xmath57 ) within the discovery reach of the lhc , for which eq .
( [ tripletvev ] ) leads to @xmath58 , and this is the scenario we will focus on .
the case of @xmath59 is assumed in the ongoing searches at the lhc , for which the brs of the triplet scalars to leptonic final states ( e.g. @xmath60 ) would be @xmath61 . since @xmath58 for light triplet scalars then @xmath56 must also be small ( compared to the electroweak scale ) for the scenario of @xmath59 . moreover ,
if one requires that the triplet yukawa couplings @xmath62 are greater in magnitude than the smallest yukawa coupling in the sm ( i.e. the electron yukawa coupling , @xmath63 ) then from eq .
( [ nu_mass ] ) one has @xmath59 , and thus the decays of the triplet scalars to leptonic final states have brs which sum to @xmath61 . in extensions of the htm
the term @xmath64 ) , which is the only source of explicit lepton number violation , may arise in various ways : i ) it can be generated at tree level via the vev of a higgs singlet field @xcite ; ii ) it can arise at higher orders in perturbation theory @xcite ; iii ) it can originate in the context of extra dimensions @xcite , and iv ) in the context of other extensions of the htm @xcite . the branching ratios ( brs ) for @xmath65 depend on @xmath66 and are predicted in the htm in terms of the parameters of the neutrino mass matrix @xcite .
detailed quantitative studies of br(@xmath65 ) in the htm have been performed in refs .
@xcite with particular emphasis given to their sensitivity to the majorana phases and the absolute neutrino mass i.e. parameters which can not be probed in neutrino oscillation experiments .
a study on the relation between br(@xmath65 ) and the neutrinoless double beta decay was performed in ref .
@xcite .
an upper limit on @xmath51 can be obtained from considering its effect on the parameter @xmath67 . in the sm @xmath68 at tree - level , while in the htm one has ( where @xmath69 ) : @xmath70 the measurement @xmath71 leads to the bound @xmath72 , or @xmath73 .
therefore the vev of the doublet field @xmath74 is essentially equal to the vev of the higgs boson of the sm ( i.e. @xmath75 gev ) . at the 1-loop level @xmath51
must be renormalised @xcite , and the first explicit analysis in the context of the htm has recently been performed in @xcite , with bounds on @xmath51 similar to those of the tree - level analysis .
the scalar eigenstates in the htm are as follows : i ) the charged scalars @xmath1 and @xmath76 ; ii ) the cp - even neutral scalars @xmath5 and @xmath77 ; iii ) a cp - odd neutral scalar @xmath78 .
the doubly charged @xmath1 is entirely composed of the triplet scalar field @xmath79 , while the remaining eigenstates are in general mixtures of the doublet and triplet fields .
however , such mixing is proportional to the triplet vev , and hence small _ even if _ @xmath51 assumes its largest value of a few gev .
the expressions for the mixing angle in the cp - even sector @xmath80 and charged higgs sector @xmath81 are : @xmath82 we note that a large mixing angle ( including maximal mixing , @xmath83 is possible in the cp - even sector provided that @xmath84 @xcite , but we will not consider this parameter space in this work . the above approximation for @xmath85 is valid as long as @xmath86 is larger than @xmath87 by 5 gev or so .
therefore @xmath88 are predominantly composed of the triplet fields , while @xmath5 is predominantly composed of the doublet field and plays the role of the sm higgs boson . neglecting the small off - diagonal elements in the cp - even mass matrix , the approximate expressions for the squared masses of @xmath5 and @xmath77 are as follows :
@xmath89 the squared mass of the ( dominantly triplet ) cp - odd @xmath78 is given by : @xmath90 the squared mass of the ( dominantly triplet ) @xmath76 is given by : @xmath91 finally , the squared mass of the ( purely triplet ) doubly - charged scalar ( @xmath92 ) is given by : @xmath93 one can see that the squared mass of the ( dominantly doublet ) @xmath5 is simply given by @xmath94 , as in the sm . in the expressions for the masses of @xmath95 , @xmath96 , @xmath97 and @xmath98 there is a common term @xmath99 .
it is evident that the mass scales for @xmath5 and the dominantly triplet scalars ( @xmath100 ) are unrelated , the former being set by @xmath94 and the latter by @xmath99 . neglecting the terms which are proportional to the small parameter @xmath51 , one can see that there are only two possible mass hierarchies for the triplet scalars , with the magnitude of the mass splitting being controlled by @xmath101 ( and @xmath102 when @xmath51 is neglected ) : @xmath103 in our numerical analysis we choose the following seven parameters from the scalar potential to be the input parameters : @xmath104 therefore @xmath105 and @xmath56 are determined from the above parameters .
many previous works have studied the impact of singly charged scalars on the decay @xmath106 e.g. in the context of the minimal supersymmetric sm ( mssm ) @xcite , a two - higgs doublet model @xcite and the next - to - mssm @xcite .
the contribution of doubly charged scalars to this decay has received comparatively very little attention . in the context of a 3 - 3 - 1 model . ] a study in the context of a little higgs model was performed in @xcite , but the magnitude of the contribution from @xmath1 was shown to be much smaller than that of the singly charged scalar in the same model , due to the theoretical structure of the scalar potential .
recently the impact of the contribution from @xmath1 was studied in the htm in @xcite , and was shown to give a sizeable contribution to @xmath106 . in @xcite both enhancements and suppressions were discussed , while in @xcite only suppressions were studied .
we will closely follow the analysis of @xcite . note that the loop - induced decay @xmath107 would also receive contributions from @xmath1 and @xmath22 , but we will not consider this decay in this work because it is not expected to be observable in the early stages of operation of the lhc .
we will consider the parameter space where @xmath5 is essentially composed of the doublet field , while @xmath77 is essentially composed of the neutral triplet field , which is realised in a large region of the parameter space of the htm . in this case
both @xmath77 and @xmath78 will be difficult to observe in the 8 tev run of the lhc due to their small couplings to quarks and vector bosons , and so they can only be produced in reasonable amounts by mechanisms such as @xmath108 ( as discussed in @xcite ) .
consequently , the mass limits from the searches for the sm higgs in the interval @xmath109 gev do not apply to @xmath77 and @xmath78 since their production cross sections are much smaller than that for @xmath110 .
the explicit form of the decay width for @xmath111 is as follows @xcite : @xmath112 here @xmath80 is the fine - structure constant , @xmath113 is the number of quark colours , @xmath114 is the electric charge of the fermion in the loop , and @xmath115 @xmath116 .
the loop functions @xmath117 ( for the @xmath23 boson ) and @xmath118 ( for the fermions , @xmath119 ) are well known . for the contribution from the fermion loops we will only keep the term with the top quark , which is dominant .
the loop function for @xmath1 and @xmath22 is given by : @xmath120\ , \tau^{-2 } \
, , \label{eq : ascalar}\end{aligned}\ ] ] and the function @xmath121 is given by @xmath122 ^ 2 \hspace{0.5 cm } & \tau>1 \ , .
\end{array } \right .
\label{eq : ftau } \end{aligned}\ ] ] note that the contribution from the loop with @xmath1 in eq.([partial_width_htm ] ) is enhanced relative to the contribution from @xmath22 by a factor of four at the amplitude level .
the couplings of @xmath5 to the vector bosons and fermions relative to the values in the sm are as follows : @xmath123 from eq .([alpha_beta ] ) one has @xmath124 , and @xmath125 , and so it is clear that the above couplings of @xmath5 are essentially identical to those of the sm higgs boson because @xmath126 .
the scalar trilinear couplings are parametrised as follows : @xmath127 with the following explicit expressions in terms of the parameters of the scalar potential ( where @xmath128 etc ) @xcite : @xmath129 @xmath130 neglecting the terms which are suppressed by @xmath51 one has the simple forms @xcite : @xmath131 @xmath132 therefore one expects that the couplings @xmath133 and @xmath134 have similar magnitudes in the htm , and it is essentially @xmath9 which determines the value of @xmath133 . in fig .
[ amp ] we plot the amplitudes ( before squaring ) of the separate contributions to @xmath7 in eq .
( [ partial_width_htm ] ) for @xmath135 gev , @xmath8=250 gev , and three values of @xmath136 ( 300 gev , 250 gev and 200 gev ) . due to the enhancement factor of a factor of four from the electric charge of @xmath1 being twice that of @xmath76
, one expects the contribution of @xmath1 to @xmath7 to dominate that from @xmath76 , and this can be seen in fig .
in contrast , in the little higgs model of @xcite the theoretical structure of the model requires @xmath137 , and so the contribution from @xmath1 could be neglected in @xcite with respect to the that from @xmath76 .
note that the contribution from the @xmath1 loop interferes constructively with that of the @xmath23 loop for @xmath138 , while for @xmath139 the interference is destructive and its magnitude can be as large as that of the @xmath23 contribution for @xmath140 .
the @xmath76 loop is usually subdominant . for @xmath141
( i.e. @xmath142 ) one has @xmath143 and thus the loop from @xmath76 enhances the total amplitude from the charged scalars , with the contribution of the @xmath76 loop being roughly one quarter that of the @xmath76 loop .
if @xmath101 and @xmath9 have opposite signs and @xmath144 then the contribution from @xmath76 interferes destructively with that from @xmath1 . ) for the @xmath23 , @xmath145 , @xmath1 and @xmath22 contributions to @xmath7 as a function of @xmath9 for @xmath146 gev , and @xmath147 gev , 250 gev and 300 gev , with @xmath148 gev . ] theoretical constraints on the parameters of the scalar potential of the htm come from the requirements of perturbative unitarity in scalar - scalar scattering , stability of the potential , and positivity of the masses of the scalars .
a comprehensive study of all these constraints has been performed in @xcite , with earlier studies in @xcite .
the magnitude of @xmath9 plays a crucial role in determining the numerical value of the trilinear couplings in eq.([trilinhpp ] ) and eq.([trilinhp ] ) .
the main constraint on @xmath9 comes from the requirement of the stability of the scalar potential , and one of four such constraints derived in @xcite is : @xmath149 if @xmath150 and @xmath151 are taken to be zero , then the combined constraints on @xmath9 from perturbative unitarity in scalar - scalar scattering and from stability of the potential require @xmath12 , as shown in @xcite . however , for the case of sufficiently positive @xmath150 and @xmath151 the choice of @xmath10 is permissible , and is a possibility which was acknowledged in @xcite but its phenomenology was not studied .
we note that varying @xmath150 and @xmath151 has very little effect on the trilinear couplings in eq .
( [ eq : ghhpp ] ) and eq .
( [ eq : ghhp ] ) , and on the masses of the triplet scalars . in our numerical analysis
we will fix @xmath152 , and use eq.([stability ] ) to determine @xmath150 as a function of @xmath9 and @xmath153 . for the case of @xmath141 and @xmath154 ( on which we will focus ) one can show from eq .([stability ] ) that the stability of the vacuum requires @xmath155 .
the most negative value of @xmath9 that we will consider is @xmath156 , for which @xmath157 would be necessary . for
@xmath158 and @xmath159 one would require @xmath160 and @xmath161 respectively .
the strongest limits on @xmath8 come from the ongoing lhc searches , which assume production via @xmath162 @xcite and @xmath163 @xcite .
production mechanisms which depend on @xmath51 ( i.e. @xmath164 and fusion via @xmath165 @xcite ) have smaller cross sections than the above processes for @xmath166 , but such mechanisms could be the dominant source of @xmath1 at the lhc if @xmath167 and @xmath168 . the first searches for @xmath1 at the lhc with @xmath169 have been carried out by the cms collaboration @xcite ( with @xmath170 of integrated luminosity ) .
separate searches were performed for @xmath162 and @xmath163 , assuming the decay channels @xmath65 and @xmath171 with @xmath172 .
the atlas collaboration has carried out three distinct searches for the decay @xmath65 ( assuming production via @xmath162 only ) as follows : i ) two ( or more ) leptons ( for @xmath173 only ) , using @xmath174 of integrated luminosity @xcite ; ii ) three ( or more ) leptons ( @xmath175 ) , using @xmath176 of integrated luminosity @xcite ; iii ) four leptons ( @xmath175 ) , using @xmath176 of integrated luminosity @xcite .
the mass limits on @xmath8 from the lhc searches are stronger than those from the tevatron searches @xcite . in the cms search ( which currently has the strongest limits ) , the limit is @xmath177 for the decay with @xmath175 and assuming br=100% in a given channel . for final states with one @xmath178 or two @xmath178 the limits are weaker , being @xmath179 and @xmath180 respectively .
moreover , the limit of @xmath181 gev is derived in four benchmark points in the htm in which all six decay channels have a non - zero br(@xmath65 ) .
the above searches all assume dominance of the leptonic decay channels @xmath65 , which is the case in the htm for @xmath182 and degeneracy of the triplet scalars .
however , the decay channel @xmath183 dominates for @xmath184 ( and assuming degeneracy of the triplet scalars ) , for which there have been no direct searches .
therefore the above bounds on @xmath8 can not be applied to the case of @xmath184 , and in this scenario @xmath1 could be much lighter .
moreover , for the case of non - degeneracy of the triplet scalars the decay @xmath185 can be dominant over a wide range of values of @xmath186 gev @xcite , and thus the limits on @xmath8 from the searches for @xmath65 can be weakened . in our numerical analysis
we will consider values of @xmath8 as low as 150 gev .
two lhc collaborations have performed inclusive searches for @xmath0 ( cms in @xcite and atlas in @xcite ) .
assuming that the production cross section is that of the sm higgs boson , constraints are now being derived on the quantity br@xmath188 in any theoretical model : @xmath189 the dominant production process in the sm is @xmath110 ( which comprises around 87% of the inclusive production of @xmath190 ) while vector boson fusion @xmath191 , higgsstrahlung @xmath192 and @xmath193 give subdominant contributions .
moreover , both cms @xcite and atlas @xcite have performed searches for a fermiophobic higgs boson decaying to two photons .
these latter searches apply selection cuts which differ from those in @xcite .
this is to increase sensitivity to the vector boson fusion and the higgsstrahlung production mechanisms and to suppress any contribution from @xmath110 , which is expected to be very small or even absent for a fermiophobic higgs boson . in @xcite
it is estimated that the search strategy in @xcite is sensitive to the combination @xmath194 while the search strategy in @xcite is sensitive to the combination @xmath195 in all four of these searches for @xmath0 there is a small excess of events at @xmath196 gev , which leads to a weaker exclusion of @xmath197 at 95% c.l . the small excess could merely be a background fluctuation which will disappear with a larger integrated luminosity .
however , attributing this excess to genuine production of @xmath0 @xcite gives rise to the following best fit value of @xmath198 , which is derived from averaging the four measurements in @xcite : @xmath199 away from @xmath200 gev , values of @xmath198 in the interval @xmath201 are excluded for @xmath202 gev , with the sensitivity reaching @xmath203 in some very small mass intervals .
in the htm the couplings of @xmath5 to the quarks and vector bosons are almost identical to those of the sm higgs boson , and are given in eqs .([h1tt ] ) to ( [ h1zz ] ) . since @xmath204 and @xmath205 then the production rate of @xmath5 at the lhc is essentially the same as that of the sm higgs boson in all the standard search channels . as in @xcite
we define the simple ratio : @xmath206 this is approximately the quantity which is now being constrained by the lhc searches in @xcite and in @xcite , for which the dominant production mechanism comes from @xmath207 .
given the coupling in eqs .
( [ h1tt ] ) and ( [ h1bb ] ) , one can see that the ratio @xmath208 is essentially equal to one in the htm . as mentioned above
, the fermiophobic higgs searches in @xcite and atlas @xcite greatly reduce the contribution from @xmath209 .
for these searches one would need to replace @xmath210 in eq .
( [ eq : rgg ] ) by the couplings to vector bosons in eq .
( [ h1ww ] ) , and the production rate for @xmath5 in the htm would still be essentially the same as that of the sm higgs boson .
in this section we quantify the magnitude of the charged scalar loops ( @xmath1 and @xmath76 ) on the ratio @xmath198 .
the case of @xmath139 was studied in @xcite .
we confirm their results , and present results for the case of @xmath138 . in our numerical analysis
the following parameters are fixed : @xmath211 the choice of @xmath148 gev is motivated by the excess of events at this mass in the searches for @xmath7 at the lhc .
the value of @xmath212 gev satisfies the constraints on @xmath51 from the @xmath213 parameter .
moreover , this choice of @xmath51 ensures that br(@xmath60 ) is negligible , and so the strong limit of @xmath181 gev for the benchmark points in the htm does not apply due to the dominance of the decay mode @xmath214 ( or @xmath185 if there is a mass splitting between @xmath1 and @xmath76 ) .
we emphasise that the role of @xmath150 and @xmath151 in our numerical analysis is merely to ensure that the constraint from vacuum stability in eq .
( [ stability ] ) is satisfied for the novel case ( @xmath14 ) of interest to us .
our numerical results are essentially insensitive to the choice of @xmath215 , and we have explicitly checked that taking @xmath216 ( which is necessary to satisfy eq .
( [ stability ] ) with @xmath156 ) gives practically the same numerical results as for @xmath217 for our choice of @xmath212 gev .
this is because @xmath150 and @xmath151 have an almost negligible effect on the trilinear couplings @xmath3 and @xmath218 in eq .
( [ eq : ghhpp ] ) and eq .
( [ eq : ghhp ] ) , a consequence of the fact that their contribution is multiplied by the small parameter @xmath51 . for definiteness
we take @xmath219 .
the parameter @xmath101 determines whether there is a mass splitting among the triplet scalars , and @xmath101 also enters the expression for the @xmath218 coupling in eq .
( [ eq : ghhp ] ) .
we treat @xmath9 and @xmath8 as free parameters that essentially determine the magnitude of the @xmath1 contribution to @xmath7 .
we present results for the range : @xmath220 we note that the range @xmath221 was studied in @xcite . in fig .
[ rgam1 ] , @xmath198 is plotted in the plane of @xmath222 $ ] .
we fix @xmath141 and so @xmath1 and @xmath22 are essentially degenerate , and the couplings @xmath223 and @xmath224 in eq .
( [ trilinhpp ] ) and eq .
( [ trilinhp ] ) are approximately equal . therefore the contributions of @xmath1 and @xmath22 to the decay rate of @xmath7 in eq .
( [ partial_width_htm ] ) differ only by their electric charge , with the contribution of @xmath1 being four times larger at the amplitude level . in the left panel
the range @xmath225 gev is plotted , and in the right panel the range @xmath226 gev is plotted . for the case of @xmath139 one has destructive interference of the @xmath1 loop with that of the @xmath23 loop , leading to a significant suppression of @xmath198 i.e. in the region @xmath227 and @xmath228 gev there is a large parameter space for @xmath229 . for @xmath227 and @xmath230 gev
( i.e. the mass region which has yet to be probed in the direct searches which assume dominance of the decay @xmath60 ) the suppression is more mild , with @xmath231 .
consequently , for @xmath227 a statistically significant signal for @xmath7 would require considerably more integrated luminosity than for the case of the higgs boson of the sm , and so detection of @xmath7 might not be possible in the 8 tev run of the lhc .
however , the other lhc search channels which make use of the tree - level decays ( i.e @xmath232 etc ) would have the same detection prospects as those of the higgs boson of the sm .
we note that the scenario of a large mixing angle @xmath233 @xcite ( for which @xmath84 is necessary ) could further delay detection of @xmath5 in these latter channels because the production cross sections of @xmath5 ( and @xmath77 ) would have a suppression factor of @xmath234 relative to that of the higgs boson of the sm , as well as having masses which differ by a few gev .
the case of @xmath235 occurs for @xmath236 ( i.e. a negligible trilinear coupling @xmath3 ) , and also for a straight line which joins the points @xmath237 , @xmath238 gev and @xmath239 , @xmath240 gev .
hence any signal for @xmath7 with @xmath203 ( and assuming @xmath241 gev ) would restrict the parameter space of @xmath222 $ ] to two regions : i ) the region of @xmath236 , and ii ) the region of @xmath242 .
if @xmath1 is very heavy and out of the discovery reach of the lhc ( e.g. @xmath243 tev ) then @xmath203 could be accommodated with any positive and sizeable @xmath9 . as emphasised in @xcite ,
in the region of @xmath244 and @xmath245 gev the contribution of the @xmath1 loop is so large that @xmath246 occurs .
this region is still compatible with current lhc data , which excludes @xmath247 for @xmath87 around 125 gev , while @xmath248 is excluded for essentially all other choices of @xmath87 in the interval @xmath249 gev .
the excess of @xmath13 events at 125 gev , if assumed to originate from a higgs boson , roughly corresponds to @xmath250 .
if @xmath246 turns out to be preferred by lhc data then one interpretation in the htm would be the region of @xmath244 and @xmath245 gev @xcite .
we now discuss the case of @xmath138 , for which @xmath246 .
the current sensitivity to @xmath7 in the lhc searches is between @xmath251 in the mass range @xmath252 gev , and so the scenario of @xmath138 is now being probed by the ongoing searches .
one can see that the current best fit value of @xmath250 can be accommodated by values of @xmath24 which are much smaller than for the case of @xmath253 , e.g. @xmath254 can be obtained for @xmath255 ( or @xmath256 ) and @xmath238 gev .
importantly , any measured value of @xmath246 would be readily accommodated by the scenario @xmath10 , even for a relatively heavy @xmath1 , e.g. for @xmath181 gev one has @xmath257 .
if either of the decays @xmath214 or @xmath185 is dominant ( for which there have been no direct searches ) then @xmath258 gev is not experimentally excluded , and larger enhancements of @xmath198 are possible e.g. @xmath259 and 1.9 for @xmath260 and @xmath261 , with @xmath238 gev .
such large enhancements of @xmath198 would require a relatively light @xmath1 ( e.g. @xmath262 gev ) which decays dominantly to @xmath214 and/or @xmath185 , for which there have been no direct searches .
simulations of @xmath214 were performed in @xcite , with good detection prospects for @xmath262 gev .
a parton - level study of @xmath185 ( for the signal only ) has been carried out in @xcite .
the above discussion was for the case of @xmath141 and we now discuss the effect of @xmath263 . in this case the couplings @xmath223 and @xmath224 in eq .
( [ trilinhpp ] ) and eq .
( [ trilinhp ] ) are not equal ( as discussed in fig . [ amp ] ) , and there is a mass splitting between @xmath8 and @xmath264 given by eq .
( [ charged - mass ] ) and eq .
( [ doub - charged - mass ] ) .
therefore the contribution of @xmath1 to @xmath7 is not simply four times the contribution of @xmath76 at the amplitude level . in fig .
[ rgam - hp ] we plot @xmath198 as a function of @xmath9 , fixing @xmath146 gev and taking @xmath147 gev , 250 gev and 300 gev ( corresponding to @xmath265 and @xmath266 respectively ) .
one can see that the case of @xmath147 gev and @xmath10 leads to a value of @xmath198 which is roughly @xmath267 larger than the value for the case of @xmath268 .
this due to the an increase of the magnitude of the coupling @xmath134 in eq .
( [ trilinhp ] ) and also because of less suppression from the @xmath269 term in eq .
( [ eq : redgcalhhp ] ) .
note that @xmath270 when @xmath271 due to the non - vanishing coupling @xmath134 for @xmath263 ( see also fig . ( [ amp ] ) ) . for @xmath12
the magnitude of @xmath134 is less than that of @xmath133 due to destructive interference in eq .
( [ trilinhp ] ) , and at around @xmath272 the value of @xmath198 becomes equal to the case of @xmath268 . for @xmath273
one finds again values of @xmath198 which are slightly larger than for the case of @xmath268 . for @xmath274 gev ( i.e. @xmath275 )
the converse dependence of @xmath198 on @xmath9 is found . in the plane of @xmath222 $ ] for @xmath276 gev ( left panel ) and @xmath277 gev ( right panel ) , for @xmath148 gev and @xmath278.,title="fig : " ] in the plane of @xmath222 $ ] for @xmath276 gev ( left panel ) and @xmath277 gev ( right panel ) , for @xmath148 gev and @xmath278.,title="fig : " ] as a function of @xmath9 for @xmath146 gev , and @xmath147 gev , 250 gev and 300 gev , with @xmath148 gev . ]
ongoing searches at the lhc for the decay @xmath0 are approaching sensitivity to the prediction for the sm higgs boson in the range @xmath279 gev .
this data constrains models of new physics which enhance br(@xmath0 ) relative to the rate in the sm .
a local excess of events is seen by both the cms and atlas experiments , which is consistent with a signal for the higgs boson of the sm with a mass of around 125 gev .
doubly charged higgs bosons ( @xmath1 ) , which arise in the higgs triplet model ( htm ) of neutrino mass generation , can significantly alter the branching ratio of @xmath7 ( where @xmath5 is the lightest cp - even scalar in the htm ) , while the branching ratios of the dominant tree - level decays are essentially the same as those for the higgs boson of the sm .
the contribution of the loops involving @xmath1 is mediated by a trilinear coupling @xmath280 , and its magnitude essentially depends on two parameters : an arbitrary quartic coupling ( @xmath9 ) and the mass of @xmath1 ( @xmath8 ) .
consequently , the improving limits on br(@xmath7 ) from lhc data constrain the parameter space of @xmath222 $ ] .
an additional ( and subdominant ) contribution comes from a loop with @xmath76 , and its magnitude is determined by @xmath9 , @xmath101 and @xmath136 . as recently pointed out in @xcite , for the case of @xmath139 the contribution of @xmath1 interferes destructively with that of the @xmath23 loop , and it can lead to a significant suppression of the branching ratio of br(@xmath7 ) . in this scenario
much more data would be required to see a signal for @xmath7 in the htm at the lhc .
for very large positive @xmath9 ( i.e. @xmath281 ) and @xmath245 gev a sizeable enhancement of br(@xmath7 ) would be possible , and this parameter space is now being constrained by the lhc limits on br(@xmath7 ) , as well as being a possible explanation of the excess of events at 125 gev @xcite . in this work we pointed out that constructive interference of the @xmath1 contribution with the @xmath23 contribution occurs for @xmath138 , and such a parameter space
is consistent with theoretical constraints on @xmath9 from requiring the stability of the vacuum of the scalar potential . for @xmath282 gev ,
which is roughly the bound if the decays @xmath60 decays are dominant in the htm , an enhancement of up to @xmath283 and @xmath284 is possible for @xmath285 and @xmath261 .
conversely , if either of the decays @xmath214 or @xmath185 is dominant ( for which there have been no direct searches ) then @xmath258 gev is not experimentally excluded , and larger enhancements of @xmath286 and @xmath287 are possible for @xmath285 and @xmath261 with @xmath238 gev . consequently , the parameter space of @xmath14 in the htm is more tightly constrained by the ongoing searches for @xmath288 than the case of @xmath12 .
importantly , the case of @xmath14 would readily accommodate any signal for @xmath7 with a branching ratio which is higher than that for the higgs boson , for _ smaller _ values of @xmath24 than for the case of @xmath139 .
in such a scenario , dedicated searches at the lhc for the decay channels @xmath214 or @xmath185 with @xmath258 gev would be strongly motivated .
a.g.a was supported by a marie curie incoming international fellowship , fp7-people-2009-iif , contract no . 252263 .
sm is partially supported through the next institute .
g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 710 * , 49 ( 2012 ) .
s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 710 * , 26 ( 2012 ) .
s. chatrchyan _ et al . _
[ cms collaboration ] , cms pas hig-12 - 008 ; g. aad _ et al . _ [ atlas collaboration ] , atlas - conf-2012 - 019 .
g. aad _ et al . _ [ atlas collaboration ] , phys .
lett . * 108 * , 111803 ( 2012 ) .
d. carmi , a. falkowski , e. kuflik and t. volansky , arxiv:1202.3144 [ hep - ph ] ; a. azatov , r. contino and j. galloway , jhep * 1204 * , 127 ( 2012 ) ; j. ellis and t. you , arxiv:1204.0464 [ hep - ph ] .
e. gabrielli , b. mele and m. raidal , arxiv:1202.1796 [ hep - ph ] ; p. p. giardino , k. kannike , m. raidal and a. strumia , arxiv:1203.4254 [ hep - ph ] .
p. m. ferreira , r. santos , m. sher and j. p. silva , arxiv:1112.3277 [ hep - ph ] ; g. burdman , c. haluch and r. matheus , phys .
d * 85 * , 095016 ( 2012 ) ; p. m. ferreira , r. santos , m. sher and j. p. silva , arxiv:1112.3277 [ hep - ph ] ; a. arhrib , r. benbrik and n. gaur , arxiv:1201.2644 [ hep - ph ] ; p. m. ferreira , r. santos , m. sher and j. p. silva , phys .
d * 85 * , 035020 ( 2012 ) ; a. barroso , p. m. ferreira , r. santos and j. p. silva , arxiv:1205.4247 [ hep - ph ] ; w. mader , j. -h .
park , g. m. pruna , d. stockinger and a. straessner , arxiv:1205.2692 [ hep - ph ] ; l. basso , a. lipniacka , f. mahmoudi , s. moretti , p. osland , g. m. pruna and m. purmohammadi , arxiv:1205.6569 [ hep - ph ] ; a. arhrib , r. benbrik and c. -h .
chen , arxiv:1205.5536 [ hep - ph ] .
g. c. branco , p. m. ferreira , l. lavoura , m. n. rebelo , m. sher and j. p. silva , arxiv:1106.0034 [ hep - ph ] .
a. arhrib , w. hollik , s. penaranda and m. capdequi peyranere , phys .
b * 579 * , 361 ( 2004 ) .
v. d. barger , n. g. deshpande , j. l. hewett and t. g. rizzo , [ hep - ph/9211234 ] ; h. pois , t. j. weiler and t. c. yuan , phys .
d * 47 * , 3886 ( 1993 ) ; a. stange , w. j. marciano and s. willenbrock , phys .
d * 49 * , 1354 ( 1994 ) .
m. a. diaz and t. j. weiler , hep - ph/9401259 ; a. g. akeroyd , j. phys .
g * 24 * , 1983 ( 1998 ) .
w. konetschny and w. kummer , phys .
b * 70 * , 433 ( 1977 ) .
r. n. mohapatra and g. senjanovic , phys .
lett . * 44 * , 912 ( 1980 ) .
m. magg and c. wetterich , phys .
b * 94 * , 61 ( 1980 ) ; g. lazarides , q. shafi and c. wetterich , nucl .
b * 181 * , 287 ( 1981 ) .
j. schechter and j. w. f. valle , phys .
d * 22 * , 2227 ( 1980 ) .
t. p. cheng and l. f. li , phys .
d * 22 * , 2860 ( 1980 ) .
cms collaboration , cms pas hig-12 - 005 ( march 2011 ) .
p. dey , a. kundu and b. mukhopadhyaya , j. phys .
g * 36 * , 025002 ( 2009 ) .
a. g. akeroyd and c. w. chiang , phys .
d * 81 * , 115007 ( 2010 ) .
a. arhrib , r. benbrik , m. chabab , g. moultaka , m. c. peyranere , l. rahili and j. ramadan , phys .
d * 84 * , 095005 ( 2011 ) .
a. arhrib , r. benbrik , m. chabab , g. moultaka and l. rahili , jhep * 1204 * , 136 ( 2012 ) . s. kanemura and k. yagyu , arxiv:1201.6287 [ hep - ph ] . g. b. gelmini and m. roncadelli , phys .
b * 99 * , 411 ( 1981 ) .
e. ma , m. raidal and u. sarkar , phys .
lett . * 85 * , 3769 ( 2000 ) ; e. ma , m. raidal and u. sarkar , nucl .
b * 615 * , 313 ( 2001 ) .
e. j. chun , k. y. lee and s. c. park , phys .
b * 566 * , 142 ( 2003 ) .
j. schechter and j. w. f. valle , phys .
d * 25 * , 774 ( 1982 ) ; m. a. diaz , m. a. garcia - jareno , d. a. restrepo and j. w. f. valle , nucl .
b * 527 * , 44 ( 1998 ) ; a. g. akeroyd , m. a. diaz , m. a. rivera and d. romero , phys . rev .
d * 83 * , 095003 ( 2011 ) .
s. k. majee and n. sahu , phys .
d * 82 * , 053007 ( 2010 ) .
s. kanemura and h. sugiyama , arxiv:1202.5231 [ hep - ph ] .
a. g. akeroyd and m. aoki , phys .
d * 72 * , 035011 ( 2005 ) .
j. garayoa and t. schwetz , jhep * 0803 * , 009 ( 2008 ) .
a. g. akeroyd , m. aoki and h. sugiyama , phys .
d * 77 * , 075010 ( 2008 ) .
m. kadastik , m. raidal and l. rebane , phys .
d * 77 * , 115023 ( 2008 ) .
fileviez perez , t. han , g. y. huang , t. li and k. wang , phys .
d * 78 * , 015018 ( 2008 ) .
f. del aguila and j. a. aguilar - saavedra , nucl .
b * 813 * , 22 ( 2009 ) .
a. g. akeroyd and c. w. chiang , phys .
d * 80 * , 113010 ( 2009 ) .
s. t. petcov , h. sugiyama and y. takanishi , phys .
d * 80 * , 015005 ( 2009 ) .
t. blank and w. hollik , nucl .
b * 514 * , 113 ( 1998 ) ; m. c. chen , s. dawson and t. krupovnickas , phys . rev .
d * 74 * , 035001 ( 2006 ) ; p.
h. chankowski , s. pokorski and j. wagner , eur .
j. c * 50 * , 919 ( 2007 ) ; m. c. chen , s. dawson and c. b. jackson , phys .
d * 78 * , 093001 ( 2008 ) .
a. djouadi , v. driesen , w. hollik and j. i. illana , eur .
j. c * 1 * , 149 ( 1998 ) ; c. -s . huang and x. -h .
wu , phys .
d * 66 * , 075002 ( 2002 ) ; m. carena , s. gori , n. r. shah and c. e. m. wagner , jhep * 1203 * , 014 ( 2012 ) . i. f. ginzburg , m. krawczyk and p. osland , nucl .
instrum .
a * 472 * , 149 ( 2001 ) [ hep - ph/0101229 ] ; l. brucher and r. santos , eur .
j. c * 12 * , 87 ( 2000 ) ; a. g. akeroyd , m. a. diaz and m. a. rivera , phys .
d * 76 * , 115012 ( 2007 ) ; a. arhrib , r. benbrik , r. b. guedes and r. santos , phys .
d * 78 * , 075002 ( 2008 ) ; n. bernal , d. lopez - val and j. sola , phys .
b * 677 * , 39 ( 2009 ) ; p. posch , phys .
b * 696 * , 447 ( 2011 ) .
u. ellwanger , jhep * 1203 * , 044 ( 2012 ) ; s. moretti and s. munir , eur .
j. c * 47 * , 791 ( 2006 ) .
a. alves , e. ramirez barreto , a. g. dias , c. a. de s.pires , f. s. queiroz and p. s. rodrigues da silva , phys .
d * 84 * , 115004 ( 2011 ) .
m. a. shifman , a. i. vainshtein , m. b. voloshin and v. i. zakharov , sov .
j. nucl .
phys . * 30 * , 711 ( 1979 ) ; j. f. gunion , h. e. haber , g. l. kane and s. dawson , `` the higgs hunter s guide , '' front .
phys . * 80 * , 1 ( 2000 ) ; m. spira , fortsch .
phys . * 46 * , 203 ( 1998 ) ; a. djouadi , phys .
rept . * 459 * , 1 ( 2008 ) .
v. d. barger , h. baer , w. y. keung and r. j. n. phillips , phys .
d * 26 * , 218 ( 1982 ) .
j. f. gunion , j. grifols , a. mendez , b. kayser and f. i. olness , phys .
d * 40 * , 1546 ( 1989 ) ; j. f. gunion , c. loomis and k. t. pitts , econf * c960625 * , lth096 ( 1996 ) [ arxiv : hep - ph/9610237 ]
. m. muhlleitner and m. spira , phys .
d * 68 * , 117701 ( 2003 ) .
t. han , b. mukhopadhyaya , z. si and k. wang , phys .
d * 76 * , 075013 ( 2007 ) .
k. huitu , j. maalampi , a. pietila and m. raidal , nucl .
b * 487 * , 27 ( 1997 ) .
b. dion , t. gregoire , d. london , l. marleau and h. nadeau , phys .
d * 59 * , 075006 ( 1999 ) .
a. g. akeroyd , c. w. chiang and n. gaur , jhep * 1011 * , 005 ( 2010 ) .
j. f. gunion , r. vega and j. wudka , phys .
d * 42 * , 1673 ( 1990 ) .
r. vega and d. a. dicus , nucl .
b * 329 * , 533 ( 1990 ) ; j. maalampi and n. romanenko , phys .
b * 532 * , 202 ( 2002 ) .
t. aaltonen _ et al . _
[ cdf collaboration ] , phys .
lett . * 107 * , 181801 ( 2011 ) . v.
m. abazov _ et al .
_ [ d0 collaboration ] , phys . rev . lett .
* 108 * , 021801 ( 2012 ) .
s. chakrabarti , d. choudhury , r. m. godbole and b. mukhopadhyaya , phys .
b * 434 * , 347 ( 1998 ) ; j. f. gunion , int .
j. mod .
phys . * a13 * , 2277 - 2298 ( 1998 ) .
a. g. akeroyd and h. sugiyama , phys .
d * 84*,035010 ( 2011 ) .
a. melfo , m. nemevsek , f. nesti , g. senjanovic and y. zhang , phys .
d * 85 * , 055018 ( 2012 ) .
m. aoki , s. kanemura and k. yagyu , phys .
d * 85 * , 055007 ( 2012 ) .
cms collaboration , cms pas hig-12 - 002 ( march 2011 ) . | the higgs boson of the standard model is being searched for at the lhc in the channel @xmath0 . in the higgs triplet model ( htm )
there are contributions to this decay from loops of doubly charged scalars ( @xmath1 ) and singly charged scalars @xmath2 that are not present in the standard model .
these additional contributions are mediated by the trilinear couplings @xmath3 and @xmath4 , where @xmath5 is the lightest cp - even higgs boson .
we point out the possibility of constructive interference of the @xmath1 contribution with that of the @xmath6 contribution , which enables a substantial enhancement of the branching ratio of @xmath7 in the htm .
the magnitude of the contribution of @xmath1 is essentially determined by the mass of @xmath1 ( @xmath8 ) and a quartic scalar coupling ( @xmath9 ) , with constructive interference arising for @xmath10 .
consequently , the ongoing searches for @xmath0 restrict the parameter space of @xmath11 $ ] more stringently for @xmath10 than for the recently - studied case of destructive interference with @xmath12 . moreover , if the excess of @xmath13 events around a mass of 125 gev in the lhc searches for @xmath0 is substantiated with larger data samples , and the branching ratio is measured to be somewhat larger than that for the sm higgs boson , then such an enhancement could be readily accommodated by @xmath5 in the htm with @xmath14 . |
the development of quantum information theory [ 1 ] has drawn attention to fundamental questions about what is physically possible and what is not .
an example is the quantum no - cloning theorem [ 2,3 ] , which asserts that unknown pure states can not be reproduced or copied by any physical means .
wootters and zurek [ 2 ] shows that the cloning machines violates the quantum superposition principles , which applies to a minimum total number of three states , and hence does not rule out the possibility of cloning two non - orthogonal states .
yuen and dariano [ 4,5 ] show that a violation of unitarity makes the cloning of two non - orthogonal states impossible . recently ,
barnum et al . [ 6 ] have extended such results to the case of mixed states , and it was shown that two non - commuting mixed states can not be broadcast onto two separate quantum systems , even when the states need only be reproduced marginally .
more recently , koashi and imoto [ 7 ] extended the standard no - cloning theorem to the case of a subsystem correlated to others and derive a necessary and sufficient condition for two pure states , each entangled in two remote systems , to be clonable by the sequential access to the two systems .
identification and clone have close relation and also we can not identify an arbitrary unknown state . since
quantum states can not be cloned and identified faithfully , recently , the inaccurate coping and discriminating of quantum states have aroused great interest .
the inaccurate coping and discriminating of quantum states can be divided into two main categories : universal and state - dependent .
the first category is aimed at the approximate clone or identification of an arbitrary state [ 8 - 13 ] , of which the figuration of merit is the local fidelity . the second category is designed to identify and clone only a finite number of states and here we can identify two sub - categories .
approximate state - dependent cloning machines deterministically generate approximate clones of state belonging to a finite set [ 14 - 16 ] .
exact state - dependent cloning machines clone a set of non - orthogonal but liner - independent pure states with the limited success probability [ 17 - 20 ] . this job is done by duan and guo and they have found the maximum success probability if these copies are required to be exact clones .
duan and guo have provided a new possibility to copy states and demonstrated that these clones can get to the maximum success probability which is non - zero .
more recently , we [ 21 ] constructed general discriminating strategies of the quantum states secretly chosen from a certain set , given initial @xmath0 copies of each state , and obtained the matrix inequality , which describe the bound between the maximum probability of correctly determining and that of error .
we [ 22 ] also obtained the unitary representations and hamiltonians for the most general quantum probabilistic clone and identification of @xmath1 linear independent states with the given probability .
for application we realized it in @xmath2-state system with universal quantum logic gates under the condition that the maximum probability is chosen . in this paper we will apply our method , which was provided in [ 22 ] , and realize the probabilistic cloning and identifying linear independent quantum states of multi - particles system with universal logic gates .
entanglement is a special state of multi - particles system , so we realize the probabilistic clone and identification of entanglement states in actual . in section
ii , we will introduce our method in [ 22 ] and some results about universal logic gates . in section iii
, we execute this method on @xmath2-particles system and realize the probabilistic cloning and identifying four linear independent states of two particles .
it is obvious that entangled states of two particles are only special cases of our result .
we generalize our result to @xmath3particles system and we also provide the realization of the probabilistic clone and identification if we give @xmath0 initial copies for each state .
we begin our work by introducing the result in [ 22 ] , which is also the basis method used in this article .
consider the question of probabilistic identification if we have @xmath0 initial copies of each state .
we are concerned with a set of @xmath1 linear - independent quantum states@xmath4 , which span an @xmath1-dimensional hilbert space @xmath5 .
we use a general unitary - reduction operation and get @xmath6 where @xmath7 is the probability of success , @xmath8 , @xmath9 is the @xmath10-th cope of state@xmath11 . for @xmath12
are liner - independent , they span a @xmath1-dimensional hilbert space . we denote @xmath13 as the orthogonal basis states of that space .
@xmath14 is a set of orthogonal states in @xmath15-dimensional hilbert space , whose orthogonal basis states are @xmath16 , @xmath17 , @xmath18 , where @xmath19 are the orthogonal basis states of @xmath5 .
@xmath20 are orthogonal basis states of the probe system . the unitary evolution @xmath21 exists if and only if @xmath22 where @xmath23
$ ] , @xmath24 , @xmath25 _ { n\times n}$ ] .
equation ( 2.2 ) is equivalent to the following inequality @xmath26 inequality ( 2.3 ) determines maximum probability of success .
we choose the probability matrix @xmath27 and give a set of parameters that are determined by @xmath28 in the condition of eq .
( 2.4 ) and the orthogonal basis states @xmath29 , we can represent @xmath30 as @xmath31 where @xmath32 , @xmath33 , @xmath34 , @xmath35 .
now we shall be concerned with clone in the circumvent that if we have @xmath0 initial copies and wish to obtain @xmath36 copies , where @xmath37 .
we use the unitary evolution as following @xmath38 where @xmath39 and @xmath40 are normalized states of the probe system(not generally orthogonal ) but @xmath41 is orthogonal with each of @xmath40 .
the using of @xmath42 has the similar definition with @xmath12 and the basis states of space @xmath43 are @xmath44 @xmath30 exists if and only if @xmath45 where @xmath46 .$ ] eq .
( 2.7 ) is equivalent to the following @xmath47 given probability matrix @xmath27 , under the condition ( 2.4 ) but the different orthogonal basis vectors @xmath48 , we represent @xmath30 as the same form as that in eq.(2.5 ) . with the concrete form of the unitary representation of the evolution in cloning machine
, we can realize that probabilistic clone and identification with universal quantum logic gates , for deutsch [ 23 ] have proved that any unitary evolution could be executed with universal quantum logic gates , i.e. @xmath49 .
a. barenco et al .
[ 24 ] has demonstrated that any arbitrary @xmath50 could be realized by the combination of a set of gates of all one - bit quantum gates @xmath51 $ ] and the two - bit control - not gate(that maps boolean values @xmath52 to @xmath53 ) .
so our work is to realize our result with gate @xmath49 .
in the following part of this section we will introduce some basis ideas and notation about universal quantum logic gates . for any unitary matrix @xmath54 and @xmath55 ,
the matrix corresponding to the @xmath56-bit operator @xmath49 is @xmath57 , where the basis states are lexicographically ordered , i.e. , @xmath58 .
when @xmath59 , @xmath49 is so called toffoli gate [ 23 ] with @xmath60 input bits . for a general @xmath30 , @xmath49 can be regarded as a generalization of the toffoli gate , which , on the @xmath60 input bits , applies @xmath30 to the @xmath61th bit if and only if the other @xmath62 bits all are on @xmath63 .
we begin our discussion with introducing two lemma which tell us how to decompose a general unitary matrix to the products of the matrixes of the form @xmath49 .
_ lemma 1 : _ any unitary matrix @xmath64
_ { n\times n}$ ] can be decomposed into the following form @xmath65 where @xmath66_{n\times n}$ ] @xmath67 , @xmath68 is the unitary matrix which interchange between the @xmath69-th row and @xmath70-th row of @xmath71 and @xmath72 interchange the @xmath69-th collum and @xmath70-th collum of @xmath73 .
it is obvious that @xmath74 .
@xmath75 , where @xmath76 is a @xmath77 unitary matrix .
@xmath78 , where @xmath79 .
the meaning of this decomposition in math is that we use some unitary matrix to transfer the non - diagonal elements of @xmath30 into @xmath80 .
the decomposition is stylized and can be easily finished by any classical computer . @xmath81 and @xmath82 are completely determined by @xmath30 . when @xmath83 , we find @xmath84 , @xmath85 . we suppose in the subspace which is spanned by basis states @xmath86 and @xmath87 , for all @xmath88 .
when @xmath89 , there must exist @xmath10 making @xmath90 .
denote the minimum value of @xmath91 still as @xmath10 , we generally assume @xmath92 , then @xmath93 . for @xmath94 , if @xmath95 , we execute the @xmath10th bit control @xmath96 operation on the @xmath97th bit ( c - not gate ) . then for @xmath98 , when @xmath99 and @xmath100 , we execute @xmath101 on the @xmath97th bit . at
last we interchange the input sequence of the @xmath10th bit and the @xmath56th bit . with all the above operations , we transfer the operation @xmath68 into @xmath102 . with above lemma
, we can express a unitary evolution as the products of some controlled unitary operations . in this section
we first limit our discussion in two particles , which is four states system , and suppose @xmath103 , @xmath104 . for general @xmath1 particles and @xmath105 , @xmath36 , we will discuss that in the last part .
the realization of probabilistic clone and identification for two particles system has important value in practice , for the entanglement of two particles is a special state of two particles system .
we assume each particle reduce to either of two states @xmath106 or @xmath107 after we execute measurement @xmath108 on them .
so the states of the two particles span a @xmath109-dimension hilbert space @xmath5 , whose orthogonal basis states are @xmath110 .
consider four linear independent quantum states @xmath111 .
we express each state on the above basis states as @xmath112 where @xmath113 satisfy @xmath114 .
define @xmath115_{4\times 4}$ ]
. the determinant of @xmath116 should be non - zero because @xmath117 are linear independent .
contrast with the situation of @xmath118-dimension , we give following lemma _ lemma 2 : _ for any four states @xmath119 in hilbert space @xmath5 , there exist a unitary operator @xmath120 to make @xmath121 where @xmath122 and @xmath123 , @xmath124 , @xmath125 , if @xmath119 are linear independent , there must be @xmath126 .
@xmath127 is determined by @xmath128 so in the following we only consider the probabilistic clone and identification of states with the form as @xmath129 . with _
lemma 1 _ we can decompose @xmath120 to basis operation .
given a set of linear independent quantum states of two particles system @xmath130 , which have the forms we have introduced above .
we use a general unitary - reduction operation as that in eq.(2.1 ) to execute the probabilistic identification , where we let @xmath103 .
the maximum identification probability is determined by ineq.(2.3 ) .
denote @xmath131 and @xmath132 . on the given prior maximum probability matrix @xmath133 and orthogonal basis states
@xmath134 , @xmath30 can be determined by eq.(2.5 ) , where @xmath135 , @xmath136 , @xmath137 are determined by eq.(2.4 ) . @xmath138
represent the operate @xmath139 , where the matrix corresponding to the operator @xmath140 on the basis states @xmath141 is @xmath142 , @xmath143 represent unity operator of probe system . on new orthogonal basis states
@xmath144 , we express @xmath145 as @xmath146 , where @xmath147 .
so we obtain @xmath148 we have showed in _ lemma 1 _ that the unitary evolution @xmath120 and @xmath149 can be decomposed into the product of basis operations such as c - not and @xmath150 .
the decomposition of @xmath150 has been completed by barenco et al . [ 24 ] . for c - not gate we adopt the schematic convention of barenco et al . [
24 ] by representing the transformations of the target and control qubits by @xmath151 and @xmath152 respectively .
so we print the realization of quantum gates here @xmath153 where the s - gate is illustrated in fig
. 2 @xmath154 we still use a general unitary - reduction operation as that in eq .
( 2.6 ) to execute the probabilistic clone , where we let @xmath103 , @xmath104 .
the maximum clone probability is determined by ineq.(2.8 ) , where we let @xmath155 .
we denote @xmath156 represent the system of the two particles @xmath157 and @xmath158 for @xmath159 . to obtain the basis states @xmath160 ,
we first introduce a new operation which transfer all of the information describing state @xmath161 the @xmath36 clone are in into one state of two particles .
this operation must itself be performed using only pairwise and local interactions .
denote vector @xmath162 , @xmath163 , @xmath164 are denoted as same .
this operation acts as follows : @xmath165 the operator @xmath166 is unitary , so we must have @xmath167 where @xmath168 _ { 4\times 4}$ ] . for
the special form of @xmath169 , together with @xmath170 , @xmath124 , @xmath125 , suffices to determine @xmath164 uniquely through eq.(3.8 ) . to obtain an explicit expression for the operator @xmath166
, we must first specify how it transforms states in the subspace orthogonal to that spanned by @xmath171 .
a natural completion of the description of @xmath172 suggests itself if we take the sum and difference of the four equation in eq.(3.7 ) , giving @xmath173 where @xmath174 is the matrix representation of states @xmath175 , @xmath176 on the basis states @xmath177 , which are lexicographically ordered , i.e. , @xmath178 , of hilbert space @xmath179 .
we denote @xmath180 .
it is obvious that the states @xmath181 that @xmath182 represent are orthogonal and in the space spanned by @xmath183 .
so we can choose states @xmath184 in the subspace orthogonal to that spanned by @xmath185 and denote @xmath186 .
with eq.(3.9 ) , we let @xmath187 so we represent @xmath166 on the basis orthogonal states @xmath188 , @xmath189as @xmath190 . we find all the above steps for obtaining the unitary representation of @xmath166 can be easily performed with classical computer and the procedure is stylized
. denote @xmath191 , @xmath192 .
so we give the orthogonal basis states as @xmath193 , @xmath194 .
we introduce two new gates called control-@xmath172 and control-@xmath195 , whose function is that we execute @xmath166 or @xmath195 operates on the two target states when the control qubit on @xmath107 , otherwise we do not make influence on the two target states
. then we can transfer the above orthogonal basis vectors into the new orthogonal basis states @xmath196 with the c-@xmath197 gate and the inverse with the c-@xmath198 gate , where we take @xmath199 as the control qubit , and @xmath200 as the target states . on the new orthogonal basis vectors , with the given maximum clone probability matrix @xmath27 ,
combine with eq.(2.4 ) and eq.(2.7 ) , @xmath30 can be obtained through eq.(2.5 ) , where @xmath138 , @xmath137 , @xmath136 are determined by eq.(2.4 ) . @xmath138
represent the operate @xmath201 , where the matrix corresponding to the operator @xmath202 on the basis states @xmath203 is @xmath142 , @xmath143 represent unity operator of probe system , @xmath204 means only when @xmath158 is on @xmath205 we can execute the operator @xmath202 on system @xmath156 . on new orthogonal basis states
@xmath206 , we express @xmath207 as @xmath146 , where @xmath147 .
so we obtain @xmath208 where @xmath209 means only when @xmath158 is on @xmath205 we can execute the operator @xmath210 on system @xmath156 and @xmath199 .
with all the above operator , we realize the probabilistic clone of the states of two particles system . with the concrete form of @xmath166 , @xmath202 , @xmath211 and _ lemma 1 _ , we can decomposed these operators into the products of some basis operators such as c - not and controlled unitary operators .
we print the realization with quantum logic gates as here @xmath212 in this part we will give the results in the situation if we have @xmath0 initial copies and want to identify them or clone @xmath36 copies .
the methods are similar .
we first denote the operator which transfer the information of the possible states of the particles system @xmath213 to those of the particles system @xmath214 by @xmath215 .
we define @xmath216 and @xmath217 , where @xmath218 , to depict the operate which acts as the following : @xmath219 @xmath220 where we have defined the vector @xmath221 respectively by @xmath222 as that in eq.(3.8 ) , we can determine @xmath221 uniquely by above relation . for general @xmath223 , we still let @xmath224 , we still denote @xmath225 . for probabilistic identification , with eq .
( 3.12 ) , we can give the orthogonal basis states of eq .
( 2.1 ) as @xmath226 , @xmath227 . for probabilistic clone , with eq .
( 3.13 ) , we give @xmath228 .
we introduce two new gates called c-@xmath229-gate and c-@xmath230-gate , whose function is that we execute @xmath229 or @xmath230on the target @xmath36 states when the control bit on @xmath107 , otherwise we make no influence on the target @xmath36 states .
we can transfer the orthogonal basic states above into @xmath231 ( for probabilistic identification ) and @xmath232 ( for probabilistic clone ) with the c-@xmath233 gate , c-@xmath229 gate and @xmath234 gate ( the inverse with c-@xmath235 gate , c-@xmath230 gate and @xmath234 gate ) . on this
orthogonal basis states , we yield @xmath236 where @xmath237 is a unitary operator on system @xmath238 and its realization has been resolved by _
lemma 1_. it is obvious that @xmath138 is a controlled unitary operator . in the similar way , denote @xmath239 , where @xmath147 , we can express @xmath211 as @xmath240 in the discussion above , we do nt set any other limitation of probability matrix @xmath27 except the determining inequality ( 2.3 ) and ( 2.8 ) . when the probability changes , the main frame of the networks of probabilistic identification and clone remains the same but the rotation @xmath138 and @xmath241 have different forms .
the stability of the clone is another virtue of the networks like that in [ 22 ] .
we can also prove the capacity of error adaptation of our net .
we take the example in fig.2 and suppose to introduce an error in auxiliary system @xmath158 .
the input state in @xmath158 comes to @xmath242 where @xmath243 is a variable whose value is close to 0 . like that in [ 22 ] , we measure output port @xmath199 . if we get @xmath244 , in general it is regarded as a failure .
but if we measure output following c-@xmath245 gate of system @xmath158 and get @xmath246 which comes from the error @xmath243 of @xmath247 , it is obvious that the input state wo nt be destroyed and can be reset as another input of cloning machine . to the general situation , when @xmath36 is relatively large the error happens in the @xmath248 ports of @xmath158 with great probability .
if any of the ports comes to this situation , we can detect it and can avoid incorrect clones and identification . in the last part of this section we will discuss the realizing probabilistic cloning and identifying the states of @xmath1-particles system . for @xmath1-particles system , there exist @xmath249 linear independent states , which can be also represented like @xmath169 by a unitary evolution similar to that in _ lemma 2_. we can also define the similar operator @xmath166 and @xmath233 , @xmath229 that transfer the information of @xmath36 copies into one states of @xmath1-particles system , where @xmath250 . with _
lemma 1 _ we can realize @xmath166 with universal logic gates . with @xmath172 operator we can transfer the basis states into the controlled unitary evolution of @xmath251 system .
we still have @xmath252 where @xmath138 is @xmath253 control @xmath238 unitary operator .
we can also express @xmath211 like eq.(3.16 ) with @xmath254 and @xmath255 and the controlment of @xmath256 , where @xmath257 represent @xmath258 operator on the @xmath259th particle of @xmath238 system .
with all the above operator , we can realize probabilistic cloning and identifying the states of @xmath1-particles system given @xmath0 initial copies of each state .
we have realize the probabilistic cloning and identifying linear independent quantum states of multi - particles system , given prior probability , with universal logic gates using the method of unitary representation , which we have develope in [ 22 ] . in this paper , we do nt restrict the states of multi - particles system are separate or entanglement .
so our result is universal for both of them .
we still assume the state of each particle is in @xmath2-level , which means the state can be expressed with @xmath260 and @xmath107 , and the question of how to realize the probabilistic identifying and cloning the system with @xmath1-states of each particle is still interesting .
1 . first we measure the output port @xmath199 .
if we get @xmath244 , the identification is successful .
then we measure output port @xmath63 and @xmath2 .
if it is @xmath262 , we identify that the input state is @xmath263 , whose probability is @xmath7 . if we get @xmath264 the identification fails .
3.this quantum networks with logic gates contain a control-@xmath266-gate , a control-@xmath267-gate and a control-@xmath211-gate .
we have illustrated @xmath211 gates in fig .
2 . with the concrete form of @xmath197 , @xmath267 and _ lemma 1
_ we can also express @xmath197 and @xmath267 with quantum logic gates .
here we omit it for it is so complicated and should be finished by computer .
first we measure output port @xmath199 .
if we get @xmath268,the clone is successful
. then two copies of the input state in input port system @xmath156 are gained in output port @xmath156 and @xmath158 .
else if we get @xmath269 , the clone fails and we discard the output . | we realize the probabilistic cloning and identifying linear independent quantum states of multi - particles system , given prior probability , with universal quantum logic gates using the method of unitary representation . our result is universal for separate state and entanglement .
we also provide the realization in the condition given @xmath0 initial copies for each state .
pacs numbers : 03.67.-a , 03.65.bz , 89.70.+c |
let @xmath12 be a random vector whose index set is @xmath0 . throughout the paper
, we use the convention that @xmath13 when @xmath14 is a set of indices .
suppose that @xmath4 are distributed independently according to the poisson distribution and consider the distribution of @xmath15 when @xmath16 is given .
that is , @xmath15 follows the multinomial distribution with probability @xmath17 and total sum @xmath18 : @xmath19 let @xmath20 be subsets of @xmath0 satisfying @xmath21 .
the main technical result of the paper is to provide an algorithm to evaluate the conditional expectation @xmath22,\ ] ] where @xmath23 is an arbitrary function of @xmath24 .
we also consider a generalization where @xmath4 are distributed as a class of distributions including the multinomial distribution .
this problem arises from the evaluation of the multiplicity - adjusted @xmath11-value of scan statistics .
we begin with stating the scan statistics in spatial epidemiology , which is a typical example in this framework . in a certain country
, there are @xmath25 districts .
let @xmath26 be the set of districts . for each district @xmath5
, we suppose that the number @xmath4 for event we are focusing on ( e.g. , number of patients with some disease ) as well as its expected frequency @xmath27 estimated from historical data are available .
@xmath4 is assumed to be distributed according to the poisson distribution with parameter @xmath28 , @xmath29 , independently for @xmath5 .
the parameter @xmath30 is known as the standardized mortality ratio ( smr ) .
figure [ fig : yamagata - smr ] depicts an example of a choropleth map of smrs . for 44 districts ,
we indicate the values of the smrs with different colors . a set of adjacent districts with smrs higher than other areas is called a _
disease cluster_. the detection of such disease clusters is a major interest in spatial epidemiology . to detect a disease cluster ,
we settle a family of subsets @xmath31 as candidates of a disease cluster , and define a scan statistic @xmath32 for each @xmath33 .
@xmath34 is called the _
scan window_. the choice of the scan windows is an important research topic in spatial epidemiology @xcite .
when @xmath32 is larger than a threshold , say @xmath35 , we declare that @xmath34 is a disease cluster .
as such a scan statistic , @xcite proposed the use of the likelihood ratio test ( lrt ) statistic @xmath36 for the null hypothesis @xmath37 ( constant ) against the alternative @xmath38 under the conditional distribution with @xmath8 given . the conditional inference ( inference under the conditional distribution ) leads to a similar test independent of the nuisance parameter @xmath39 .
the expression of @xmath36 is given in section [ subsec : scan ] .
when the hypothesis @xmath40 holds , the disease cluster @xmath14 is called a hotspot .
this is a typical problem of multiple comparisons .
the @xmath11-value to assess the significance should be adjusted to incorporate the multiplicity effect .
one method is to define the @xmath11-value from the distribution of the maximum @xmath41 under @xmath42 : @xmath43 , \label{conditional}\ ] ] where @xmath44 .
this is of the form of ( [ expectation ] ) .
note that when @xmath18 is given , the distribution of @xmath15 is @xmath45 , where @xmath46 . in spatial epidemiology , the @xmath11-value
is usually estimated using monte carlo simulation .
although this is convenient and practical in most cases , when the true @xmath11-value is very small , it is difficult to obtain a precise value even when the number of random numbers in the monte carlo is large .
therefore , we have good reason to conduct exact computation according to the definition ( [ conditional ] ) by enumerating all possibilities .
however , this is generally difficult because of the computational complexity . in the area of multiple testing comparisons ,
many techniques to reduce computational time have been proposed .
for example , @xcite demonstrated that in a change point problem , the computational time for the distribution of the maximum could be reduced using the markov property among statistics .
see also @xcite , @xcite and references therein . in this paper
, we develop a similar computation technique by taking advantage of the markov structure among scan statistics .
the proposed method is based on the theory of a chordal graph , which is the foundation of the theory of graphical models @xcite .
the chordal graph theory provides rich tools , in not only statistics but also many fields of mathematical science .
in particular , in numerical analysis , this is a major tool used to conduct large - scale matrix computation ( e.g. , @xcite,@xcite ) .
we also apply the chordal graph theory to retrieve the markov structure to reduce the computational time by using the recursive summation / integration technique .
our technique is then similar to those used in the efficient computation of maximum likelihood estimator of graphical models ( e.g. , @xcite , @xcite , @xcite , @xcite , @xcite ) .
the remainder of the paper is organized as follows .
section [ sec : recursive ] provides the recurrence computational formula in the multinomial distribution under the assumption that the running intersection property holds .
we evaluate the computational complexity , and show that the recurrence computation methodology works for a class of distributions including the poisson distribution ( that is , the conditional distribution is the multinomial ) .
section [ sec : markov ] proposes a method to detect the markov property , and section [ sec : examples ] presents illustrative real data analyses to detect temporal and spatial clustering .
section [ sec : summary ] briefly summarizes our results and discusses further research .
in this section , we provide an algorithm to compute the expectation ( [ expectation ] ) when the sequence @xmath20 of subsets of @xmath0 has a nice property , which is called the running intersection property given in definition [ def : rip ] .
we will consider the general case in the next section . in this section ,
we use the symbol @xmath1 instead of @xmath34 .
we start with case @xmath47 . for @xmath48 ,
let @xmath49 and @xmath50 .
suppose that @xmath51 is a random vector distributed according to the multinomial distribution with summation @xmath18 and probability @xmath52 .
we consider the evaluation of the expectation @xmath53 \label{naive}\ ] ] exactly , where @xmath54 and @xmath55 are arbitrary functions .
obviously , @xmath56 and @xmath57 are not independent ; there is an overlap @xmath58 unless @xmath59 is empty .
moreover , there is a restriction that @xmath60 instead of the problem of evaluating the expectation , by a change of viewpoint , we first consider the problem of generating random numbers @xmath61 , where @xmath62 , @xmath63 .
@xmath64 can be generated according to the following three steps : @xmath65 where @xmath66 and @xmath67 are independent given @xmath68 .
correspondingly , we divide the expectation in ( [ naive ] ) into three parts as @xmath69 & = e^{(m_2,m_1)|n}\bigl [ e^{x_{r_2}|m_2}\bigl [ e^{x_{r_1}|m_1}[\chi_1(x_{b_1 } ) \chi_2(x_{b_2 } ) ] \bigr]\bigr ] \nonumber \\ & = e^{(m_2,m_1)|n}\bigl [ e^{x_{r_2}|m_2}\bigl [ \chi_2(x_{b_2 } ) e^{x_{r_1}|m_1}[\chi_1(x_{b_1 } ) ] \bigr]\bigr ] \nonumber \\ & = e^{(m_2,m_1)|n}\bigl [ e^{x_{r_2}|m_2 } [ \chi_2(x_{b_2 } ) \xi_1(m_1,x_{c_1 } ) ] \bigr ] , \label{e}\end{aligned}\ ] ] where @xmath70 \nonumber \\ & = e^{x_{r_1}|m_1}[\chi_1(x_{c_1},x_{r_1 } ) ] .
\label{xi}\end{aligned}\ ] ] the procedure for the numerical computation of ( [ e ] ) is as follows .
( i ) for possible values of @xmath71 and @xmath72 , compute @xmath73 in ( [ xi ] ) . here
, the expectation is taken over @xmath67 according to ( [ xr1 ] ) with @xmath72 fixed .
the results are stored in memory as a tabulation .
( ii ) compute @xmath74 $ ] according to ( [ e ] ) . here , @xmath75 , @xmath71 , and @xmath66 are random variables having distributions ( [ m2m1 ] ) and ( [ xr2 ] ) .
this technique substantially reduces the computational cost . to see this , we enumerate the number of required summations in detail . @xmath71 takes the values @xmath76 and @xmath72 takes the values of all nonnegative integer vectors whose sum is less than or equal to @xmath77 . in the expectation
@xmath78 $ ] in ( [ xi ] ) , the variable @xmath67 runs over all nonnegative vectors whose sum is @xmath71 .
because @xmath79 @xmath80 we see that the number of summations to prepare the table @xmath73 is @xmath81 in the expectations @xmath82 $ ] in ( [ e ] ) , the variable @xmath66 runs over all nonnegative vectors whose sum is @xmath75 , and @xmath83 .
hence , the number of summations for computing ( [ e ] ) is @xmath84 therefore , the number of summations is @xmath85 on the other hand , when we do not use this technique , the number of summations is @xmath86 which is of larger order than @xmath87 .
this technique for reducing computational time is available in a general setting . to describe it ,
we introduce several notions from the theory of chordal graphs and graphical models @xcite .
[ def : rip ] a sequence of sets @xmath89 is said to have the running intersection property ( rip ) if for each @xmath90 , there is a @xmath91 such that @xmath92 throughout this section , we suppose that sequence @xmath89 satisfies the running intersection property .
note that the indices @xmath93 of @xmath1 in definition [ def : rip ] are reversely ordered from the conventional definition .
the function @xmath94 defines a directed graph @xmath95 , where @xmath96 and @xmath97 .
the running intersection property is explained in detail in section [ sec : markov ] . for @xmath98 ,
write @xmath99 iff @xmath100 or @xmath101 let @xmath102 ( @xmath103 ) and @xmath104 , where @xmath105 are residual sets .
the disjoint union is denoted by @xmath106 .
the proposition below follows from the running intersection property .
@xmath107 obviously @xmath108 .
we prove @xmath109 for @xmath110 .
suppose that @xmath111 for @xmath112 .
let @xmath113 . since @xmath114 , @xmath115 .
on the other hand , @xmath116 implies @xmath117 . this is a contradiction .
write @xmath118 ( @xmath119 ) , @xmath120 , @xmath121 and @xmath122 .
@xmath123 is defined recursively as @xmath124 we state the proposed summation technique in general form in the theorem below .
[ thm : recursive ] suppose that a sequence of sets @xmath89 satisfies the running intersection property with respect to a function @xmath94 .
define @xmath125 , @xmath105 and @xmath123 as before .
let @xmath126 be an arbitrary function of @xmath127 .
suppose that @xmath128 is a multinomial random vector .
define @xmath129.\ ] ] in particular , define @xmath130.\ ] ] then @xmath131\bigr ] \\
\displaystyle & \hspace*{-20 mm } ( k^{-1}(i)\ne\emptyset ) , \\ \displaystyle e^{x_{r_i}|n_i } \bigl[\chi_i(x_{c_i},x_{r_i})\bigr ] & \hspace*{-20 mm } ( k^{-1}(i)=\emptyset ) .
\end{cases } \nonumber \\ & \label{recursive}\end{aligned}\ ] ] here , the conditional expectations in ( [ recursive ] ) are with respect to @xmath132 and @xmath133 we prove the case @xmath134 .
the case @xmath135 is similar and easier , and therefore omitted . noting that @xmath136
, we have @xmath137 \nonumber \\ & \mid x_{c_i},{\sum}_{j\in t_i } x_j = n_i \bigr ] . \label{xi1}\end{aligned}\ ] ] as @xmath138 and @xmath139 , the conditions on @xmath140 and @xmath141 implies the conditions on @xmath142 and @xmath143 .
that is , @xmath144 therefore , the inner conditional expectation on the right - hand side of ( [ xi1 ] ) is rewritten as @xmath145.\ ] ] moreover , when @xmath146 ( @xmath147 ) are given , @xmath148 ( @xmath147 ) are independent , and hence the expression above is rewritten as @xmath149 \\ & = { \prod}_{j\in k^{-1}(i ) } e\bigl [ { \prod}_{l\preceq j } \chi_l(x_{b_l } ) \mid x_{c_j},{\sum}_{l\in t_j } x_l = n_j \bigr ] \\ & = { \prod}_{j\in k^{-1}(i ) } \xi_i(n_j , x_{c_j}).\end{aligned}\ ] ] in the above , the second equality follows because of @xmath150 ( which will be proved later ) and hence @xmath151 is a function on @xmath140 through @xmath152 .
now , we have the formula @xmath153.\end{aligned}\ ] ] actually this is equivalent to ( [ recursive ] ) , because under the conditional distribution , @xmath154 ( @xmath155 ) and @xmath156 have the distributions given in ( [ condprob ] ) .
finally , we prove @xmath150 for @xmath157 .
obviously @xmath158 .
it suffices to prove @xmath159 for @xmath160 .
recall that @xmath160 implies @xmath161 .
also @xmath157 implies @xmath162 .
hence , @xmath163 , @xmath164 . by repeatedly applying this manipulation ,
we reach @xmath165 and @xmath166 follows .
theorem [ thm : recursive ] provides an algorithm to compute the exact expectation @xmath167 $ ] with @xmath168 by updating the tables @xmath169 , @xmath2 .
we can evaluate the number of required summations as before .
the result is summarized below without proof .
[ thm : no_sum ] the number of summations required in the algorithm of theorem [ thm : recursive ] is @xmath170 where @xmath171 note that when @xmath103 , the value @xmath169 is needed for @xmath172 . whereas , when @xmath173 , only the value for @xmath174 is needed .
this is the reason why the case @xmath173 is exceptional in ( [ order ] ) .
the number of summations ( [ order ] ) is smaller than in the absent of this recursive computation technique : @xmath175 as shown , the proposed algorithm has an advantage in time complexity .
whereas , it requires memory space to restore the tables . since in ( [ xi0 ] ) , @xmath176 and the elements of @xmath177 are arbitrary nonnegative integers such that @xmath178 , the size of the table for @xmath169 is @xmath179 however , in the process of computation of ( [ recursive ] ) for @xmath2 , the table @xmath169 can be deleted ( i.e. , the memory space can be released ) once @xmath180 is computed .
therefore , the problem of space complexity does not matter in practice .
although theorem [ thm : recursive ] is stated for the multinomial distribution , it also works for a class of distributions including the multinomial distribution .
let @xmath181 and @xmath182 be index sets such that @xmath183 , and let @xmath184 , @xmath185 and @xmath63 again .
suppose that @xmath4 is distributed independently for @xmath5 according to a certain distribution . under the conditional distribution where
@xmath8 is given , if we pose additional conditions that @xmath186 is given , then @xmath67 and @xmath66 become independent . therefore , the three steps ( [ m2m1])([xr1 ] ) for generating random numbers , and the corresponding decomposition of the expectation continue to hold for an arbitrary distribution of @xmath4 . if explicit expressions for probability density functions of the conditional distributions @xmath187 , @xmath188 , @xmath189 , and @xmath190 are available , we have the the computation formula of the type ( [ e ] )
. in general , if some explicit formula is available for the probability density function of the conditional distribution @xmath191 where the @xmath105 are subsets of @xmath0 such that @xmath192 , we can construct the recurrence computation formula of the type of theorem [ thm : recursive ] .
the class of distributions having the explicit conditional density function of ( [ generalcond ] ) includes the normal distribution , the gamma distribution , the binomial distribution and the negative binomial distribution .
the conditional distributions of ( [ generalcond ] ) corresponding to the above four distributions are the ( degenerate ) normal distribution , the dirichlet distribution , the multivariate hypergeometric distribution , and the dirichlet - multinomial distribution , respectively .
for these distributions , the recurrence computation formula of theorem [ thm : recursive ] works by replacing summations with integrations when the distribution is continuous .
as shown in the previous section , when the sequence @xmath193 of subsets of @xmath0 has the running intersection property , the expectation ( [ expectation ] ) can be evaluated efficiently with the recursive summation / integration technique proved in theorem [ thm : recursive ] . however , in general , @xmath193 does not have this property . to apply this technique to general cases ,
one method is to prepare another sequence @xmath89 having the running intersection property such that each @xmath194 is included into at least one of @xmath89 .
if we have such @xmath1 , by defining a function @xmath195 so that @xmath196 , the expectation ( [ expectation ] ) is written as @xmath197 , \quad \chi_i(x_{b_i } ) : = { \prod}_{j\in \tau^{-1}(i)}\chi_j(x_{z_j}),\ ] ] which can be dealt with by theorem [ thm : recursive ] .
such a sequence @xmath1 , @xmath2 , is known to be obtained through a chordal extension in algorithm [ alg : chordal_extension ] . here , we summarize some notions and basic facts on chordal graphs .
let @xmath198 be a connected undirected graph , where @xmath0 is a set of vertices and @xmath199 is a set of edges .
@xmath200 is called complete if every pair of vertices is joined by an edge . for a subset @xmath201 ,
let @xmath202 denote a subgraph induced by @xmath203 .
that is , @xmath204 when @xmath202 is complete , @xmath203 is called a clique .
a clique that is maximal with respect to inclusion is called a maximal clique .
@xmath200 is called chordal if every cycle with length greater than three has a chord .
let @xmath89 be a sequence of all maximal cliques of @xmath200 satisfying the running intersection property ( [ rip ] ) in definition [ def : rip ] .
the sequence is called perfect if for every @xmath2 , @xmath205 is a clique of @xmath200 .
denote by @xmath206 the set of maximal cliques of @xmath200 .
a vertex @xmath207 is called simplicial if its adjacent vertices form a clique in @xmath200 .
a perfect elimination ordering @xmath208 of @xmath200 is an ordering of vertices of @xmath200 such that for every @xmath93 , @xmath209 is a simplicial in @xmath210 .
let @xmath211 denote the set of simplicial vertices of @xmath200 in @xmath1 .
then @xmath1 is called a boundary clique if there exists @xmath10 , @xmath212 , such that @xmath213 the following proposition on the property of chordal graphs is well known ( e.g. , @xcite and @xcite ) .
let @xmath200 be an undirected graph .
the four statements below are equivalent .
\(i ) @xmath200 is chordal .
\(ii ) @xmath200 has a perfect sequence of the maximal cliques .
\(iii ) @xmath200 vertices can be ordered to have a perfect elimination ordering .
we generate a perfect sequence @xmath89 from @xmath193 according to the following procedure .
[ alg : chordal_extension ] * define an undirected graph @xmath214 with vertices @xmath26 and edges @xmath215 where @xmath216 . *
add edges of @xmath217 to @xmath200 so that the extended graph @xmath218 is a chordal graph .
* identify the perfect sequence of the maximal cliques @xmath89 of @xmath219 .
this sequence has the running intersection property and @xmath220 . the procedure for step 1 is referred to as the _ chordal extension_. constructing the chordal extension such
that the maximum size of the maximal clique is minimum is known to be a np - hard problem @xcite . in this section ,
we propose a heuristic method to construct an approximately best chordal extension . for step 2 ,
we provide a method proved in theorem [ thm : hara ] . the other method for the same purpose based on the maximum cardinality search procedure
is also known @xcite .
we now explain steps 02 in detail using a small example .
suppose that the vertex set is @xmath221 , and the family of subsets of @xmath0 is given by @xmath222 the associated undirected graph @xmath200 is shown in figure [ fig : g ] ( left ) .
step 1 is divided into three parts . in this substep , we renumber the vertices . consider the following procedure to make all vertices removed from the graph @xmath200 sequentially : first , find a vertex , say @xmath223 , that has the minimum number of adjacent edges in the graph @xmath200 .
then , remove @xmath223 as well as its adjacent edges @xmath224 from the graph @xmath200 . then , find a vertex , say @xmath225 , that has the minimal number of adjacent edges in the graph @xmath226 with @xmath227 vertices .
continuing with this procedure , we have a sequence of vertices @xmath228 .
this renumbering is called the minimum degree ( md ) ordering @xcite .
the matlab function symamd is available to obtain an md ordering @xcite . in our example , @xmath229 we illustrate the resulting renumbered graph in figure [ fig : g ] ( right ) .
[ cols= " < , < , < " , ] [ fig : part ]
in this paper , we proposed a recursive summation method to evaluate a class of expectation in multinomial distribution , and applied it to the evaluation of the @xmath11-values of temporal and spatial scan statistics .
this approach enabled us to evaluate the exact multiplicity - adjusted @xmath11-values .
our proposal has an advantage where the true @xmath11-value is too small and is barely estimated precisely by monte carlo simulations .
the proposed algorithm is easily modified to a class of distributions including the normal distribution , the dirichlet distribution , the multivariate hypergeometric distribution , and the dirichlet - multinomial distribution by replacing recursive summations with recursive numerical integrations if necessary . on the other hand ,
our proposed method has a limitation that it only works when the total data count @xmath18 and the window size are not large .
the limitation is due to @xmath18 , @xmath230 ( the sizes of the maximum clique ) , and @xmath231 . as shown in section [ subsec : yamagata ] , when the size of the scan windows is large , we can divide the whole scan windows into a number of groups , then compute the @xmath11-value for each group , and sum them .
we have implemented the proposed algorithms .
however , they are still under development .
our algorithm requires for loop calculations , where the numbers of the nests and the ranges of running variables depend on the input ( scan windows ) .
the dimensions of the arrays also depend on the data .
these make source coding complicated , and the resulting code inefficient .
one approach to overcome this difficulty may be the use of a preprocessor to generate the c program .
this approach has been successfully used in computer algebra ( e.g. , page 674 of @xcite ) .
the authors are grateful to takashi tsuchiya , anthony j. hayter , nobuki takayama , and yi - ching yao for their helpful comments .
this work was supported by jsps kakenhi grant numbers 21500288 and 24500356 .
blair , j.r.s . and
peyton , b.w .
an introduction to chordal graphs and clique trees , in `` graph theory and sparse matrix computation , '' a. george , j.r .
gilbert and j.w.h.liu , eds . ,
pp.129 , springer , new york .
buneman , p. ( 1974 ) . a characterization on rigid circuit graphs , _ discrete mathematics _ , * 9 * ( 3 ) , 205212 . fukuda , m. , kojima , m. , murota , k. and nakata , k. ( 2000 ) . exploiting sparsity in semidefinite programming via matrix completion i :
general framework , _ siam journal on optimizations _ , * 11 * ( 3 ) , 647674 .
hara , h. and takemura , a. ( 2006 ) .
boundary cliques , clique trees and perfect sequences of maximal cliques of a chordal graph , arxiv : cs/0607055 .
hara , h. and takemura , a. ( 2010 ) . a localization approach to improve iterative proportional scaling in gaussian graphical models , _ communications in statistics : theory and methods _ , * 39 * ( 8 - 9 ) , 16431654 .
hirotsu , c. and marumo , k. ( 2002 ) .
changepoint analysis as a method for isotonic inference , _ scandinavian journal of statistics _ , * 29 * ( 1 ) , 125138 .
koyama , t. , nakayama , h. , nishiyama , k. and takayama , n. ( 2014 ) .
holonomic gradient descent for the fisher - bingham distribution on the @xmath232-dimensional sphere , _ computational statistics _ , * 29 * ( 3 - 4 ) , 661683 .
kulldorff , m. ( 1997 ) . a spatial scan statistic , _ communications in statistics : theory and methods _ , * 26 * ( 6 ) , 14811496 .
kulldorff , m. ( 2006 ) .
tests of spatial randomness adjusted for an inhomogeneity : a general framework , _ journal of the american statistical association _ , * 101 * ( 475 ) , 12891305 . neff , n.d . and naus , j.i .
( 1980 ) . _
selected tables in mathematical statistics , vol.vi : the distribution of the size of the maximum cluster of points on a line _ , american mathematical society , providence , rhode island .
worsley , k.j .
confidence regions and tests for a change - point in a sequence of exponential family random variables , _ biometrika _ , * 73 * ( 1 ) , 91104 .
, guo , j.h . and
he , x. ( 2011 ) . an improved iterative proportional scaling procedure for gaussian graphical models , _ journal of computational and graphical statistics _ , * 20 * ( 2 ) , 417431 .
xu , p.f . , guo , j.h . ,
tang , m.l .
( 2014 ) . a localized implementation of the iterative proportional scaling procedure for gaussian graphical models , _ journal of computational and graphical statistics _ ,
* 24 * ( 1 ) , 205229 . | let @xmath0 be a finite set of indices , and let @xmath1 , @xmath2 , be subsets of @xmath0 such that @xmath3 .
let @xmath4 , @xmath5 , be independent random variables , and let @xmath6 . in this paper , we propose a recursive computation method to calculate the conditional expectation @xmath7 $ ] with @xmath8 given , where @xmath9 is an arbitrary function .
our method is based on the recursive summation / integration technique using the markov property in statistics . to extract the markov property
, we define an undirected graph whose cliques are @xmath10 , and obtain its chordal extension , from which we present the expressions of the recursive formula .
this methodology works for a class of distributions including the poisson distribution ( that is , the conditional distribution is the multinomial ) .
this problem is motivated from the evaluation of the multiplicity - adjusted @xmath11-value of scan statistics in spatial epidemiology . as an illustration of the approach , we present the real data analyses to detect temporal and spatial clustering .
_ keywords and phrases : _ change point analysis , chordal graph , graphical model , markov property , spatial epidemiology . |
even though there is ample observational evidence that the expansion rate of the universe is accelerating @xcite , there is still no convincing theoretical model that can explain the observations .
the simplest model in agreement with the data is the cosmological constant .
however , it suffers from size and coincidence problems .
in addition , it is likely that an earlier phase of accelerated expansion happened that was not due to a cosmological constant @xcite .
this period , inflation , is normally modeled as being due to a scalar field .
the same basic mechanism can explain the dark energy , and is often called quintessence in this context @xcite .
scalar field dark energy is often employed to model a general equation of state @xmath2 , although one can also build a perfect fluid with exactly the same first order perturbations @xcite . over the last few years
it has become clear that @xmath3 alone is not sufficient to distinguish between different models explaining the dark energy , and that the perturbations induced by the dark energy constitute a complementary probe @xcite .
the contribution of dark energy fluctuations to the perturbation dynamics is generally small and can be very difficult to measure . in a recent paper some of us derived theoretical approximations to the perturbations generated in scalar field dark energy models @xcite . in this paper , we will use these expressions to investigate how two cosmological probes that will be of great importance in the future , weak lensing ( wl ) and the full galaxy power spectrum , can be used to measure the typical features that appear in such models , specifically the existence of a sound horizon . exploiting the advantage of analytical approximations , we will try to isolate in some detail where the signal comes from . anticipating the conclusion ,
we will show that even a full sky , tomographic , deep redshift and imaging survey can not constrain significantly the sound speed of the dark energy unless it is smaller than , roughly , 0.01@xmath1 denotes the speed of light , and in the rest of the paper we will use units so that @xmath4 . ] .
small values of @xmath5 are not compatible with a scalar field description with standard kinetic term and require extensions like k - essence @xcite but are in general not forbidden : we will show that even with @xmath6 , dark energy , although `` cold '' , remains much less clustered than dark matter and will not develop non - linearities by the present time .
the special case of @xmath7 has some additional attractive features , in that such models can cross @xmath8 @xcite and can act as unified models of dark energy and dark matter @xcite ( but are also subject to the `` dark degeneracy '' @xcite ) . for such extreme cases
however dark energy becomes non linear along with dark matter and appropriate non - linear corrections are needed , especially for the weak lensing probe .
the sound speed of the dark energy could be measured also by other effects , like the integrated sachs - wolfe ( isw ) effect , either on the cmb or cross - correlated with large - scale structure . in @xcite
it has been shown that the cmb isw tail can put a lower limit to @xmath9 , while @xcite finds that @xmath10 is actually favored by the wmap 3-year data .
the isw is complementary to the probes we consider here since it is sensitive to perturbations on larger scales . as in @xcite
we set the anisotropic stress of the dark energy to zero , which is appropriate for scalar field dark energy and in general for standard gravity models . in order to investigate more general cases , for example modifications of general relativity
, a non - zero anisotropic stress would have to be taken into account @xcite , but this is left for future work .
in detail , the paper is organized as follows .
we begin by discussing the main results obtained in @xcite .
we then use these results to study the impact of the scalar field dark energy perturbations on the weak lensing signal and on probes of galaxy clustering , paying special attention to whether the presence of a sound horizon can be detected ( by measuring the sound speed @xmath9 ) .
we also discuss which aspect of these probes are most sensitive to the perturbations .
we finally investigate constraints on the integrated deviation from the poisson equation , which we call @xmath11 , which may be useful to quantify the power of probes beyond the conventional , @xmath3-based figure of merit used by the dark energy task force @xcite .
in this section we define our notation and present a short discussion of the perturbation equations and the approximate analytical solution of @xcite .
we refer the reader to that paper ( and citations therein ) for more details . in the following
, the dots will refer to the derivatives with respect to the conformal time @xmath12 which is related to the universal time @xmath13 by @xmath14 .
the physical hubble parameter that we will consider here is : @xmath15\label{eq : hubble}\ ] ] where @xmath16 $ ] and the subscript 0 denotes the present epoch .
this expression implies that we limit ourselves to a flat universe which is filled with matter ( designated by a subscript @xmath17 ) and a general fluid with average pressure @xmath18 . in general the equation of state parameter @xmath3
is a function of time , but we will take it to be constant later on
. we will consider linear perturbations about this spatially - flat background model , defined by the line element : @xmath19.\label{pert_newton_ds}\ ] ] as is apparent from the line element , we use the conformal newtonian ( longitudinal ) gauge and retain only scalar perturbations .
the perturbation equations for a dark energy fluid with sound speed @xmath5 and parameter of state @xmath3 are : @xmath20 where @xmath21 is a measure of the velocity perturbation , @xmath22 is the density contrast , and the prime here means the derivative with respect to the scale factor @xmath23 .
we parameterize the pressure perturbation as : @xmath24 where @xmath25 is the adiabatic sound speed and @xmath9 is the sound speed in the rest - frame of the dark energy fluid . as already mentioned , we assume that the parameter of equation of state @xmath3 stays constant so that @xmath26 .
we also assume a vanishing anisotropic stress @xmath27 as is the case for quintessence and k - essence models , and therefore we have @xmath28 .
finally , the gravitational potential can be found with the help of the einstein equations , @xmath29 if the universe is perfectly matter dominated , then @xmath30 is a constant . in @xcite
we used this result to derive analytical solutions to the perturbation equations for dark energy in different limits : * * scales larger than the sound horizon . *
the modes larger than the sound horizon are found for @xmath31 . in this case , we neglect all terms containing the sound speed in eq .
( [ vp ] ) and then in eq .
( [ deltap ] ) , effectively setting @xmath32 . neglecting a decaying solution for which @xmath33
we find : @xmath34 * * scales smaller than the sound horizon . * for modes smaller than the sound horizon , @xmath35 ,
we consider only the terms containing @xmath36 .
we find that the density perturbations become constant and that the velocity perturbations decay : @xmath37 in these equations , the factor @xmath38 sets the overall scale of the perturbations , chosen so that the matter perturbations are @xmath39 on sub - horizon scales .
in addition , the dark energy perturbations depend on their present density fraction ( equal to @xmath40 since the universe is taken to be flat ) but only through the combination @xmath41 , the equation of state parameter @xmath3 and the sound speed @xmath9 .
they are a function of both scale @xmath42 and scale factor @xmath23 .
let us quickly compare dark energy and dark matter fluctuations . for large scales the dominant term in eq .
( [ desolsmallsubfull ] ) is the term containing @xmath42 .
dark matter has a similar solution except that @xmath43 ; the ratio of the squares ( i.e. the power spectrum ratio ) for large scales then will be : @xmath44 if @xmath45 .
however , for small scales ( but still larger than @xmath46 ) , the first term in eq .
( [ desolsmallsubfull ] ) dominates ; in this case the ratio is : @xmath47 the dark energy perturbations are thus always suppressed relative to the dark matter perturbations , even for a very low sound speed ( note that the solution eq .
( [ desolsmallsubfull ] ) was obtained assuming @xmath32 ) . for the forecasts of this paper
, it is important to know whether we have to worry about non - linear effects : the non - linear clustering of dark energy models is still badly understood and may depend on the precise action of the model , and it is not clear what corrections we would need to apply . outside the sound horizon , the dark energy perturbation track those in the dark matter by the factors given above . these are roughly upper limits as @xmath48 is at the upper limit of what is still allowed , and values of @xmath3 closer to @xmath49 will lead to smaller dark energy perturbations . if @xmath6 the sound horizon at the present is of the order of @xmath50 . on these scales
dark matter fluctuations are still linear and therefore dark energy fluctuations are safely below the non - linearity threshold . on sub - sound - horizon scales ,
dark energy perturbations stop growing and become quickly negligible relative to the dark matter perturbations . if @xmath5 is smaller than @xmath51 then the dark energy will behave much like dark matter on the cluster scales and an appropriate non - linear correction would be needed , especially on the small scales probed by weak lensing ( the small scale cut - off in the galaxy clustering safely excludes non linear scales ) .
therefore we do not extend our analysis to such small sound speeds ; we include however for completeness the extreme case @xmath7 , for which we assume that we can apply the same non - linear correction as for the dark matter whether this is acceptable may in general depend on the precise action of the field .
in this section we discuss the effect of dark energy clustering on the galaxy and weak lensing probes . we derive approximate analytical expressions that show how the relevant quantities change with the sound speed .
this will help understanding qualitatively the numerical results of the next section .
matter domination was a necessary ingredient to derive the solutions given above .
however , dark energy comes to dominate eventually , and then the potential starts to decay and the perturbations grow more slowly or start to decrease .
it is difficult to capture this behavior accurately .
a way around this problem can be found by looking at the variable @xmath53 which we introduced in @xcite to describe the change of the gravitational potential due to the dark energy perturbations .
@xmath52 is defined through @xmath54 if the dark energy or modification of gravity does not contribute to the gravitational potential ( for example if the dark energy is a cosmological constant ) then @xmath55 .
otherwise @xmath52 will deviate from unity , and in general it is a function of both scale and time .
introducing the comoving density perturbation @xmath56 , we can write @xmath57 we neglect radiation in this paper as we are interested in observational tests at late times where it is subdominant .
then @xmath52 is defined as @xmath58 using simply the solution for the perturbations during matter domination , as discussed in the previous paragraph , we find that the resulting expression for @xmath52 is surprisingly accurate even at late times .
the reason is that both fluids , dark energy and matter , respond similarly to the change in the expansion rate so that most of the deviations cancel .
we find that the sub - soundhorizon expression below is accurate at the percent level , while on larger scales there are deviations of about 10 to 20% by today ( depending on @xmath3 ) .
the latter can be corrected `` by hand '' in order to obtain a more precise formula , but the expressions are sufficiently accurate for our purposes and we keep them as they are . as we have set the anisotropic stress to zero , the perturbations are fully described by @xmath52 . in @xcite
we provided the following explicit expression for the @xmath59 which captures the behavior for both limits ( above and below the sound horizon ) : @xmath60 here we used @xmath61 which we defined through @xmath62 so that @xmath63 counts how deep a mode is inside the sound horizon .
we see that during matter domination ( used in the first expression for @xmath63 ) a @xmath42 mode moves ever more deeply inside the sound horizon as @xmath23 grows , which is of course no surprise . from eq .
( [ eq : qdef3 ] ) we see that the deviation of @xmath52 from @xmath64 depends on the ratio of the energy density of the dark matter and the dark energy , and on the relative amount of perturbations .
the former scales as @xmath65 for a constant @xmath3 and the latter behaves as discussed in the last section . for @xmath66 and @xmath48
we have that @xmath67 on scales that are larger than the sound horizon , @xmath68 .
this is not a negligible deviation today , but it decreases rapidly as we move into the past , as the dark energy becomes less important .
as a scale enters the sound horizon , @xmath69 grows with one power of the scale factor slower ( since @xmath70 stops growing ) , suppressing the final deviation roughly by the ratio of horizon size to the scale of interest . in the observable range , @xmath71 . therefore if @xmath72 , @xmath73 and the dependence on @xmath5 is lost .
this shows that @xmath52 is sensitive to @xmath5 only for small values , @xmath74 .
we can characterize the dependence of @xmath52 on the main perturbation parameter @xmath9 by looking at its derivative , a key quantity for fisher matrix forecasts : @xmath75 where @xmath76 ( with the last expression being for @xmath48 ) . for the values we are interested here , this derivative has a peak just inside the sound horizon , the exact position is @xmath77^{1/4 } .\end{aligned}\ ] ] today the sound horizon is given by @xmath78 .
it lies in the observable range of @xmath42 for sound speeds of the order of @xmath79 .
we plot the derivative as the red curve in fig .
[ fig : der - g - q - b - p0-ink ] , there and from eq .
( [ eq : qdercs ] ) we can see that the shape is basically proportional to @xmath80 .
the derivative with respect to @xmath9 ( instead of @xmath81 ) contains an additional factor of @xmath82 which means that it will be boosted by several orders of magnitude for small sound speeds .
we will later forecast how well the deviation of @xmath52 from @xmath64 due to the dark energy perturbations can be measured by future cosmological surveys . in the @xmath84cdm model of cosmology , the dark matter perturbations on sub - horizon scales grow linearly with the scale factor @xmath23 during matter domination . during radiation domination
they grow logarithmically , and also at late times , when dark energy starts to dominate , their growth is suppressed .
it is well known that in @xmath84cdm the growth factor can be expressed as : @xmath85 where @xmath86 is called the growth index .
there are two ways to influence the growth factor : firstly at background level , with a different hubble expansion .
secondly at perturbation level : if dark energy clusters then the gravitational potential changes because of the poisson equation , and this will also affect the growth rate of dark matter .
all these effects can be included in the growth index @xmath83 and we therefore expect that @xmath83 is a function of @xmath3 and @xmath9 ( or equivalently of @xmath3 and @xmath52 ) . according to @xcite , the growth index depends on dark energy perturbations ( through @xmath52 ) as @xmath87 where @xmath88 however , dark energy perturbations ( hence @xmath52 ) are difficult to measure at least if the dark energy has a very low sound speed .
the growth index factor seems to be a more promising parameter and several experiments are planned to measure this quantity .
furthermore , we can invert eq .
( [ eq : gamma - q ] ) and ask the question : in the absence of anisotropic stress which @xmath52 is needed to generate a given @xmath83 ? using eqs .
( [ eq : gamma - q ] ) and ( [ eq : a - q ] ) we have : @xmath89\left[1-w-\frac{1}{3}\left(5 - 6w\right)\gamma\right].\label{eq : q - gamma}\ ] ] the last equation is worth another look : let us assume we measure the growth index @xmath90 ( the value usually associated to the cosmological constant ) then eq .
( [ eq : q - gamma ] ) becomes : @xmath91.\label{eq : q - gamma-1}\ ] ] if we are dealing with the cosmological constant then we have the expected result of @xmath92 . however ,
if @xmath93 then we can evaluate the corresponding value of @xmath52 ; for instance , if @xmath48 then we have @xmath94 , which is a positive number . however , this value is fairly big for dark energy perturbations , see @xcite .
moreover , for non - phantom models the quantity @xmath69 is always positive due to the relative increase of dark energy perturbations and the @xmath95 factor ; let us remind the reader that if the anisotropic stress is set to zero then @xmath52 fully describes the evolution of dark energy perturbations ; so , we can think to rephrase the sentence as : if @xmath96 which @xmath83 can we reach ?
setting eq .
( [ eq : q - gamma ] ) positive we have @xmath97 which sets a sort of upper bound to @xmath83 ; moreover , this is the limit we have if dark energy perturbations are set to zero ( which can be seen directly from eq .
( [ eq : gamma - q ] ) ) . in general we can say that if the anisotropic stress is zero then dark energy perturbations always decrease the value of the growth index .
if we want to allow for perturbations in the dark energy sector while increasing at the same time the growth index then we need a non zero anisotropic contribution ; this is the case for the dgp model ( treated as an effective dark energy perturbations , see @xcite ) where @xmath98 .
let us also consider the derivative of @xmath99 with respect @xmath9 to gain an idea of how strongly the growth factor depends on sound speed , and thus on the characteristics of the dark energy perturbations : @xmath100 we can see more clearly the dependence on the parameters by noting that the quantity @xmath101 can be approximated as @xmath102 for @xmath103 close to unity ( we use however the exact expressions in the calculations below ) .
eq.([eq : dg - dcs ] ) then reads : @xmath104 where @xmath105 is a constant during matter domination ( the value of @xmath106 today ) .
we first consider two different limits * @xmath107 ( sub - sound horizon regime ) : @xmath108 * @xmath109 ( super - sound horizon regime ) : @xmath110 both of these derivatives , like the one of @xmath52 discussed further above , have a strong dependence on the sound speed .
as before , the peak location is near the sound horizon , and again the derivative with respect to @xmath9 adds an additional @xmath82 factor so that small sound speeds will lead to much larger values of the derivative . the numerical factor ( including the @xmath3 dependence )
is @xmath111 and @xmath112 for @xmath48 , respectively .
we also notice that the first limit does not depend on @xmath69 like the derivative @xmath52 , but on @xmath113 as it is an integral .
the growth factor is thus not directly probing the deviation of @xmath52 from unity , but rather how @xmath52 evolves over time .
this is similar in the second limit , except that we probe the evolution of @xmath114 .
we can find a unified formula which accounts for both regimes : @xmath115}{\left(1 - 3w\right)\left(q - q_{0}\right)-\left(1 + 3w\right)\left[\left(q-1\right)x-\left(q_{0}-1\right)x_{0}\right]}\label{eq : gdercs - union}\ ] ] where @xmath116 and @xmath117 are evaluated at @xmath118 .
the derivative is shown as the blue dashed line in fig .
[ fig : der - g - q - b - p0-ink ] .
just as the @xmath52 derivative , the @xmath119 derivative peaks close to the sound horizon , but is smaller than the one of @xmath52 at this redshift .
it grows however with growing redshift due to its integral nature , while the other derivatives decrease . in order to quantify the impact of the sound speed on the matter power spectrum we proceed in two ways
: we first use the camb output @xcite ( which contains the full information on dark energy perturbations ) and then we consider the analytic expression from eisenstein & hu @xcite ( which does not include dark energy perturbations , i.e. does not include @xmath5 ) . as anticipated , we find here that the impact of the derivative of the matter power spectrum with respect the sound speed on the final errors is only relevant if high values of @xmath9 are considered ; for instance if @xmath120 then the errors on the sound speed increase by about @xmath121 times when the analytic formula in ref .
@xcite is considered instead of the camb output . by decreasing the sound speed ,
the results are less and less affected , and are completely unchanged when @xmath122 .
the reason is that for low values of the sound speed other parameters , like the growth factor , start to be the dominant source of information on @xmath123 , see fig .
[ fig - fisher - cs000001 ] . since for large @xmath5
our results show that the sound speed in practically not measurable with the probes here considered , we disregard this difference between the analytical fit and the camb output and use the analytical fit whenever we find it convenient , in particular in estimating the weak lensing fisher matrix . in fig .
[ fig : derlogpkdlogcs ] we show the derivative of the logarithm of the matter power spectrum ( from camb ) with respect to the logarithm of sound speed .
it is easy to understand this result : in fig . 3 of @xcite we showed that the power spectrum decreases by a few percent in a smooth step - like fashion at a scale just inside the sound horizon .
here we see the derivative of this transition . just like the step , the peak of the derivative is situated a little bit inside the sound horizon , as is the case for the other contributions as well . for two different values of the sound speed : @xmath124 and @xmath125 , red solid and blue dashed line , respectively .
the vertical dotted lines give the scale of the sound horizon for two different sound speeds , @xmath125 ( left ) and @xmath124 ( right).,width=249 ] the velocity field of matter is directly sourced by the gradient of the @xmath126 gravitational potential , and astronomers have been trying to measure it for a long time @xcite .
more recently constraints on the velocity perturbations from redshift space distortions became an important probe for dark energy phenomenology , since their combination with weak lensing allows to disentangle the two potentials , which in turn allows to put limits on the anisotropic stress of the dark sector @xcite .
this is an important test for modifications of gravity @xcite .
although there is no anisotropic stress in our scenario , we are nonetheless interested in how redshift space distortions are modified by the presence of dark energy perturbations . the distortion induced by redshift can be expressed in linear theory by the @xmath127 factor , related to the bias factor and the growth rate via : @xmath128 the derivative of the redshift distortion parameter with respect to the sound speed is : @xmath129 we see that the behavior versus @xmath9 is similar to the one for the @xmath52 derivative , so the same discussion applies . once again , the effect is maximized for small @xmath5 . in fig .
[ fig : der - g - q - b - p0-ink ] we compare the derivative of the redshift space distortion ( green dotted line , averaged over @xmath130 ) with those of the growth factor @xmath119 and the @xmath52 variable .
the @xmath127 derivative is comparable to @xmath119 at @xmath131 but becomes more important at low redshifts and , as mentioned , is similar to the @xmath52 derivative in shape . of the @xmath52 parameter eq .
( [ eq : qdercs ] ) ( red solid line ) , the growth factor eq .
( [ eq : gdercs - union ] ) ( blue dashed line ) , the redshift distortion parameter eq .
( [ eq : derbetadcs ] ) ( green dotted line ) and the matter power spectrum @xmath132 ( black dot - dashed line ) .
the derivatives are evaluated at a @xmath133 except the one for @xmath132 which is at @xmath131 ; for all the derivatives the sound speed is @xmath125 .
the vertical dotted line gives the scale of the sound horizon for @xmath125 at @xmath133.,width=249 ]
we want to investigate now the response of weak lensing ( wl ) to the dark energy parameters .
we proceed with a fisher matrix as in @xcite , to which we refer for the implementation details , the main difference here being that the parameter @xmath52 , which was general , now has an explicit form that describes how it changes as a function of scale and redshift .
since @xmath52 depends on @xmath3 and @xmath9 , we can forecast the precision with which those parameters can be extracted .
we can also try to trace where the constraints come from . for wl experiments
the important quantity is the lensing potential which is , using the metric eq .
( [ pert_newton_ds ] ) : @xmath134 for a vanishing anisotropic stress the wl potential becomes : @xmath135 in linear perturbation theory all @xmath42 modes evolve independently , so that the dark matter density contrast is often decomposed as : @xmath136 here @xmath137 determines the matter power spectrum today , @xmath138 and @xmath139 is the growth factor , which depends here not only time but also on @xmath42 since the dark energy perturbations in general introduce a scale dependence .
we can write eq .
( [ eq : wl - potential ] ) as : @xmath140 where we used eq .
( [ eq : deltamatter ] ) .
hence , the lensing potential contains three conceptually different contributions from the dark energy perturbations : * the direct contribution of the perturbations to the gravitational potential through the factor @xmath52 , see section [ sec : q ] .
* the impact of the dark energy perturbations on the growth rate of the dark matter perturbations , affecting the time dependence of @xmath141 , through @xmath139 , cf section [ sec : gamma ] . * a change in the shape of the matter power spectrum @xmath142 , corresponding to the dark energy induced @xmath42 dependence of @xmath141 , as discussed in section [ sec : pk ] .
we consider a typical next - generation tomographic weak lensing survey characterized by the sky fraction @xmath143 and by the number of sources per arcmin@xmath144 , @xmath145 .
we consider the range @xmath146 and we extend our survey up to three different redshifts : @xmath147 .
for the non linear correction we use the halo model by smith et al .
we choose as fiducial model @xmath148 , @xmath149 , @xmath150 , @xmath151 , @xmath152 , @xmath153 , @xmath154 , @xmath155 , and several values for @xmath9 .
the observationally borderline fiducial value of @xmath156 is chosen so as to maximize the impact on @xmath52 : values closer to @xmath49 reduce the effect and therefore increase the errors on @xmath5 .
we always plot the fully marginalized confidence ellipses ( and quote fully marginalized errors ) at @xmath157 , which in 2d correspond to semi - axes of length @xmath158 time the eigenvalues . in fig .
[ fig : ellipsesw0cs ] we report the confidence region for @xmath159 for two different values of the sound speed and @xmath160 . for high value of the sound speed ( @xmath120 ) we find @xmath161 and the relative error for the sound speed is @xmath162 .
as expected , wl is totally insensitive to the clustering properties of quintessence dark energy models when the sound speed is equal to @xmath64 .
the presence of dark energy perturbations leaves a @xmath3 and @xmath9 dependent signature in the evolution of the gravitational potentials through @xmath163 and , as already mentioned , the increase of the @xmath9 enhances the suppression of dark energy perturbations which brings @xmath164 .
once we decrease the sound speed then dark energy perturbations are free to grow at smaller scales . as an example we show in the lower panel of fig .
[ fig : ellipsesw0cs ] , the confidence region for @xmath159 for @xmath122 , we find @xmath165 , @xmath166 ; in the last case the error on the measurement on the sound speed reduced to the 50% of the total signal . in tab . [
tab : errors - wl ] we list the errors for @xmath156 and the relative errors of @xmath9 for six different values of the sound speed for a survey up to @xmath167 .
we should note here that the fisher matrix approach breaks down for such large errors since it only considers the local curvature of the likelihood at the location of the fiducial model @xcite .
if we performed an actual analysis for data with @xmath120 , we would find a lower limit near @xmath122 since those models are measurably different from @xmath120 according to the lower panel of figure [ fig : ellipsesw0cs ] , and the ellipse would be cut off there . on the other hand
, there would not be any upper limit to @xmath9 since for all those cases the dark energy perturbations are not detected .
for three different value of @xmath147 , blue dashed , green long - dashed and solid contour , respectively . the upper panel shows the confidence region when the sound speed is @xmath120 ; the bottom panel with the sound speed @xmath122 .
the parameter equation of state is for both cases @xmath155.,title="fig:",width=249 ] for three different value of @xmath147 , blue dashed , green long - dashed and solid contour , respectively . the upper panel shows the confidence region when the sound speed is @xmath120 ; the bottom panel with the sound speed @xmath122 .
the parameter equation of state is for both cases @xmath155.,title="fig:",width=249 ] + @xmath9 & @xmath168 & @xmath169 + @xmath170 & @xmath171 & @xmath172 + @xmath173 & @xmath174 & @xmath175 + @xmath176 & @xmath177 & @xmath178 + @xmath179 & @xmath180 & @xmath181 + @xmath182 & @xmath183 & @xmath184 + @xmath64 & @xmath185 & @xmath186 + to explore more directly which term influences the most the weak lensing signal we assume that there are two different @xmath52 s : one , @xmath187 , which enters directly in the growth index expression ( [ eq : gamma - q ] ) and the other , @xmath188 , which is the term that enters linearly in the expression of the gravitational potential ( [ eq : wl - potential ] ) .
we concentrate here only on these terms since , as we mentioned previously , the changes on the matter power spectrum are negligible in the interesting range ( i.e. for small sound speed ) . * @xmath187 only : here we consider only the contribution from @xmath187 setting effectively @xmath189 . *
@xmath188 only : here we assume that dark energy perturbations enter only through @xmath188 , effectively setting the function @xmath190 in eq .
( [ eq : gamma - q ] ) . in red and assuming only @xmath188 in blue.,width=249 ] we show the behavior of the errors in fig .
[ fig : rel - errors - wl - q - g ] as a function of the sound speed @xmath9 .
we notice that the main contribution comes from @xmath188 . keeping
only this term leads to the same errors as the full expression .
this is consistent with the discussion in sections [ sec : q ] and [ sec : gamma ] and the derivatives shown in fig .
[ fig : der - g - q - b - p0-ink ] where we see that the growth factor is always less sensitive to the dark energy perturbations than @xmath52 . as both enter in the same way in eq .
( [ eq : phiwl ] ) the result is no surprise .
at least for small sound speeds , gravitational lensing constrains directly the contribution of the dark energy perturbations to the metric , not their impact on the dark matter .
we now explore a second probe of clustering , the galaxy power spectrum .
following @xcite we write schematically the observed galaxy power spectrum as : @xmath191 where the subscript @xmath192 refers to the values assumed for the reference ( or fiducial ) cosmological model , i.e. the model at which we evaluate the fisher matrix . here
@xmath193 is the shot noise due to discreteness in the survey , @xmath130 is the direction cosine within the survey , @xmath194 is the present spectrum for the fiducial cosmology , @xmath195 is the linear growth factor of the matter perturbations , @xmath196 is the bias factor ( assumed scale independent ) and @xmath197 is the angular diameter distance .
the wavenumber @xmath42 is also to be transformed between the fiducial cosmology and the general one ( @xcite and see also @xcite and @xcite , for more details ) . to avoid non - linearity problems ( both in the spectrum and in the bias ) ,
the fisher matrix is calculated up to a limiting @xmath198 at @xmath199 : we choose values from @xmath200mpc for low-@xmath199 bins to @xmath201mpc for the highest @xmath199-bins . here again the galaxy power spectrum is affected by the dark energy perturbations in three different direct ways : * through the growth factor , see eq .
( [ eq : gamma - q ] ) ; * through the redshift space distortions , see [ sec : zdist ] . * through the present matter power spectrum @xmath194 , as discussed in [ sec : pk ] .
we consider a photometric survey from @xmath202 divided in equally spaced bins of width @xmath203 as our benchmark survey ; we assume an error on the measure of redshift of about @xmath204 and an area of @xmath205 .
these features are similar to those of proposed experiments like jdem and euclid , @xcite and @xcite , respectively ( see also detf report @xcite ) . here
too we also consider extended surveys to @xmath167 and @xmath206 .
moreover , we assume an overall radial distribution @xmath207 $ ] where @xmath208 and a @xmath209 .
the bias factor here is assumed to be only redshift dependent ; we assume as fiducial bias @xmath210 . in general
, however , dark energy perturbations could lead to a scale dependence of the bias factor .
our fiducial model is exactly the same as for the wl survey considered in the previous section . in fig .
[ fig : ellipsesw0cs - pk ] we report the confidence region for @xmath159 for two different values of the sound speed and @xmath160 . for high values of the sound speed ( @xmath120 ) we find , for our benchmark survey : @xmath211 , and @xmath212 .
here again we find that galaxy power spectrum is not sensitive to the clustering properties of dark energy when the sound speed is of order unity .
if we decrease the sound speed down to @xmath122 then the errors are @xmath213 , @xmath214 . in tab .
[ tab : errors - bao ] we listed explicitly the errors for @xmath156 and the relative errors of @xmath9 for six different values of the sound speed for a benchmark survey up to @xmath215 . for three different value of @xmath147 , blue dashed , green long - dashed and solid counter , respectively .
the upper panel shows the confidence region when the sound speed is @xmath120 ; the bottom panel with the sound speed @xmath122 .
the parameter equation of state is for both cases @xmath155.,title="fig:",width=249 ] for three different value of @xmath147 , blue dashed , green long - dashed and solid counter , respectively .
the upper panel shows the confidence region when the sound speed is @xmath120 ; the bottom panel with the sound speed @xmath122 .
the parameter equation of state is for both cases @xmath155.,title="fig:",width=268 ] + @xmath9 & @xmath168 & @xmath169 + @xmath170 & @xmath216 & @xmath217 + @xmath173 & @xmath218 & @xmath219 + @xmath176 & @xmath220 & @xmath221 + @xmath179 & @xmath222 & @xmath223 + @xmath182 & @xmath224 & @xmath225 + @xmath64 & @xmath226 & @xmath227 + we can ask again the question what part of the power spectrum is most sensitive to the dark energy perturbations , @xmath119 , @xmath127 or @xmath132 ? in order to disentangle their relative contributions , we plot the corresponding diagonal fisher matrix terms containing the sound speed as a function of redshift .
the results are shown in fig .
[ fig - fisher - cs1 ] for @xmath120 and in fig .
[ fig - fisher - cs000001 ] for @xmath122 . for large sound speed
the shape of the dark matter power spectrum is formally the most sensitive , but not sufficiently to constrain the perturbations . in the more interesting case where the sound speed is lower and
the perturbations can be detected , it is instead the growth factor that is the most sensitive . in both cases
the contribution from redshift space distortions to the overall constraints is sub - dominant except at redshifts below about @xmath228 and for small @xmath9 , cf figures [ fig - fisher - cs1 ] and [ fig - fisher - cs000001 ] . in the low - redshift region
the growth factor is less sensitive as it is an integral from @xmath131 . to further clarify the importance of the redshift space distortions
, we performed our analysis excluding the dependence of dark energy perturbations on the redshift distortion parameter .
the result , given in tab .
[ tab : errors - w - bao - nobeta ] , shows that for small sound speeds the contribution from @xmath127 is non - negligible when we compare it to table [ tab : errors - bao ] . ,
@xmath229 , @xmath230 , solid blue , red and green lines , respectively . for @xmath120,width=268 ] , @xmath229 , @xmath230 , solid blue , red and green lines , respectively . for @xmath122,width=268 ] in fig .
[ fig : errors ] the behavior in logarithmic scale of the sound speed errors @xmath231 are shown as a function of the sound speed itself .
we notice that the errors appear to scale as a power law for both probes , over the range of sound speeds given . for the fiducial surveys we find @xmath232 for the weak lensing and@xmath233 for the power spectrum . &
@xmath215 & @xmath167 & @xmath206 + @xmath9 & @xmath169 & @xmath169 & @xmath169 + @xmath170 & @xmath234 & @xmath235 & @xmath236 + @xmath64 & @xmath237 & @xmath238 & @xmath239 + in black , both for our benchmark survey.,width=268 ] we can also combine the two probes , the galaxy power spectrum and weak lensing , to improve the sensitivity to the dark energy perturbations , cf . tab .
[ tab : combined - bao - wl ] .
as expected the errors decrease by about @xmath240 for low values of the sound speed and remain unaltered when the sound speed is close to unity .
we find that the sound speed does not suffer of any strong degeneracies with the other parameters considered in this work .
we evaluated the dimensionless correlation matrix : @xmath241 where @xmath242 is the inverse of the fisher matrix @xmath243 .
for instance , for @xmath122 , we find for all parameters a correlation factor with @xmath9 of at most @xmath244 for galaxy power spectrum experiments and at most @xmath245 for wl experiments ; however , for the latter case we find for @xmath246 terms a correlation value of about @xmath247 .
increasing the sound speed the correlation terms decrease for galaxy power spectrum case and wl case ; for the latter only the term @xmath248 increases up to @xmath249 , for @xmath120 .
.correlation matrix for galaxy power spectrum experiments and @xmath122 .
[ cols="^,^,^,^,^,^,^",options="header " , ] so far we have been studying the errors on the dark energy parameters @xmath3 and @xmath9 . a subtly different question is whether we are able to detect dark energy perturbations or not
a question not about parameter constraints but about model probabilities @xcite .
it is not straightforward to use our constraints e.g. on the sound speed for this purpose , as it does not make sense to consider the speed of sound if there are no perturbations at all , i.e. a model without perturbations is not nested in the quintessence model considered here .
this is just another way to say that all quintessence models ( except those for which @xmath8 , equivalent to a cosmological constant ) necessarily have perturbations . but from the errors on @xmath3 and @xmath9 we can forecast the constraints on the amount of deviation of the poisson equation from the expectation due to the dark matter and other known contributions , given by @xmath52 . a model without dark energy perturbations would then be characterized by @xmath55 .
however , @xmath52 is a function which evolves and additionally can have a scale dependence .
it is then useful to define a compact way to express the deviation integrated over @xmath199 and @xmath42 .
we define therefore the average quantity @xmath250 where @xmath251 is the observed redshift range and @xmath252 is the momentum volume .
finding a deviation of @xmath11 from zero would signal some clustering of dark energy .
we note that it is straightforward to focus on certain scales or epochs by introducing an appropriate weight function into the integral above . in order to obtain constraints on @xmath11
we need to marginalize the fisher matrix over all parameters except @xmath253 and then project it over the new parameter set @xmath254 . defining @xmath255
we have @xmath256 the errors on @xmath11 for the @xmath142 case are reported in tab . [
tab : errors - w - pk ] and for wl in tab .
[ tab : errors - w - wl ] .
since the derivative of @xmath52 with respect the sound speed contains a term which goes like @xmath82 , the dominating source of error on @xmath11 comes from the errors on @xmath123 , as long as @xmath257 .
therefore a quick estimate can be obtained by approximating @xmath258 and we find that in the present case the relative errors on @xmath11 are similar to those on @xmath9 .
this supports the naive interpretation that we can only detect the perturbations if we can measure the sound speed , i.e. if the relative error on @xmath9 is less than unity . however , we are now in a position where we can use model comparison techniques to decide whether @xmath259 has been detected or not . using the savage - dickey density ratio ,
we find that the relative model probability is given by the value of the normalized posterior ( marginalized over all common parameters , i.e. all parameters except @xmath11 ) at @xmath260 , divided by the value of the normalized prior at the same place . to simplify the calculations , we choose a uniform prior in all variables except @xmath11 , and for that variable a gaussian pdf centered at @xmath260 with a width ( variance ) @xmath261 .
this then leads to a bayes factor of @xmath262 here @xmath263 is the value on which the posterior for @xmath11 is centered , and @xmath264 the variance of @xmath11 .
given the fiducial value @xmath265 and the fisher matrix error @xmath266 , which characterize the likelihood , we find for the posterior the shifted values ( due to the prior , see e.g. the appendix of ref .
@xcite ) @xmath267 however , it is not obvious how we should choose the width of the prior , for a general model with unknown sound speed .
one possibility is to use the most optimistic value , for which the model discrimination is maximal @xcite . to find this value ,
we maximize @xmath268 with respect to @xmath269 , which leads to @xmath270 . using this prescription ( valid only for @xmath271 )
we finally obtain @xmath272 this formula depends only on the ratio @xmath273 , and if we demand strong evidence in favor of the presence of perturbations , @xmath274 , we find that we need @xmath275 , i.e. not quite a 4-@xmath276 detection .
we remind the reader that this corresponds to the choice of prior that maximally favors the detection of perturbations , any other choice would need a stronger detection of @xmath11 to reach the same bayes factor . looking at tables [ tab : errors - w - pk ] , [ tab : errors - w - wl ] and [ tab : combined - bao - wl ]
we can see that in our class of quintessence models and for the observations that we consider here , we can only hope to strongly favor the presence of dark energy perturbations if the sound speed is @xmath277 .
@xmath9 & @xmath278 & @xmath169 + @xmath170 & @xmath279 & @xmath217 + @xmath173 & @xmath280 & @xmath219 + @xmath176 & @xmath281 & @xmath221 + @xmath179 & @xmath282 & @xmath283 + @xmath182 & @xmath284 & @xmath225 + @xmath64 & @xmath285 & @xmath227 + @xmath9 & @xmath278 & @xmath169 + @xmath170 & @xmath286 & @xmath172 + @xmath173 & @xmath287 & @xmath175 + @xmath176 & @xmath288 & @xmath178 + @xmath179 & @xmath289 & @xmath181 + @xmath182 & @xmath290 & @xmath184 + @xmath64 & @xmath291 & @xmath186 + @xmath9 & @xmath168 & @xmath169 & @xmath278 + @xmath170 & @xmath292 & @xmath279 & @xmath293 + @xmath173 & @xmath294 & @xmath236 & @xmath295 + @xmath176 & @xmath296 & @xmath297 & @xmath298 + @xmath179 & @xmath299 & @xmath175 & @xmath300 + @xmath182 & @xmath301 & @xmath302 & @xmath303 + @xmath64 & @xmath304 & @xmath305 & @xmath306 +
in this paper we extended the idea proposed in @xcite .
we investigated the effect of a generic modification of poisson equation @xmath53 in the galaxy power spectrum and in the weak lensing convergence power spectrum induced solely by dark energy clustering .
in other words , we are using perturbations in the metric and in the dark matter to detect the tiny traces of dark energy perturbations .
we identified several effects : on the power spectrum growth , on its shape , and ( for the galaxy clustering probe ) on the redshift distortion .
these effect are completely characterized by @xmath3 and @xmath123 .
we performed a fisher matrix analysis of future large - scale surveys to forecast the errors on @xmath3 and @xmath123 and therefore on @xmath53 or on its average version , @xmath11 .
as perhaps expected , we find that dark energy perturbations have a very small effect on dark matter clustering unless the sound speed is extremely small , @xmath307 .
for @xmath307 , we find that @xmath11 ( that is , the average deviation for a standard poisson equation ) could be constrained to within 66% or better .
let us remind the reader that in order to boost the observable effect , we always assumed @xmath48 : for values closer to @xmath49 the sensitivity to @xmath9 is further reduced ; as a test we performed the calculation for @xmath308 and @xmath122 and we find @xmath309 and @xmath310 for wl and galaxy power spectrum experiments , respectively .
such small sound speeds are not in contrast with the fundamental expectation of dark energy being much smoother that dark matter : even with @xmath10 , dark energy perturbations are more than one order of magnitude weaker than dark matter ones ( at least for the class of models investigated here ) and safely below non - linearity at the present time at all scales . models of `` cold '' dark energy with @xmath32 are interesting because they can cross the phantom divide @xcite and contribute to the cluster masses @xcite . small @xmath5 could be constructed for instance with scalar fields with non - standard kinetic energy terms .
here we showed that future large scale surveys have the potential to rule out or confirm this class of models .
as an example we report here the absolute errors for the sound speed @xmath32 : @xmath311 and @xmath312 for wl and galaxy power spectrum experiments respectively .
d. s. acknowledges support by the spanish @xmath313 under the project @xmath314 and the eu fp6 marie curie research and training network `` universenet '' ( @xmath315 ) .
m.k . acknowledges support by the swiss nsf . | the characterization of dark energy is a central task of cosmology . to go beyond a cosmological constant , we need to introduce at least an equation of state and a sound speed and consider observational tests that involve perturbations . if dark energy is not completely homogeneous on observable scales then the poisson equation is modified and dark matter clustering is directly affected .
one can then search for observational effects of dark energy clustering using dark matter as a probe . in this paper
we exploit an analytical approximate solution of the perturbation equations in a general dark energy cosmology to analyze the performance of next - decade large scale surveys in constraining equation of state and sound speed .
we find that tomographic weak lensing and galaxy redshift surveys can constrain the sound speed of the dark energy only if the latter is small , of the order of @xmath0 ( in units of @xmath1 ) . for larger
sound speeds the error grows to 100% and more .
we conclude that large scale structure observations contain very little information about the perturbations in canonical scalar field models with a sound speed of unity
. nevertheless , they are able to detect the presence of `` cold '' dark energy , i.e. a dark energy with non - relativistic speed of sound . |
an isolated stochastic population ( of molecules , bacteria , animals , parasites inhabiting a community of hosts , _
etc_. ) ultimately goes extinct .
the ultimate extinction is driven , even in the absence of unfavorable environmental variations , by large demographic fluctuations : chains of random events when population losses dominate over gains .
the risk of population extinction is a major issue in viability of small populations in the nature @xcite , whereas extinction of an endemic disease from a community @xcite is a desirable development .
calculating the extinction rate of a stochastic population is a challenging problem .
here one needs to evaluate the ( very low ) probability of a large fluctuation in a stochastic system which is far from equilibrium and therefore defies many standard assumptions and methods of statistical mechanics . in spite of this difficulty
a significant progress has been achieved in this type of problems via the use of a wkb ( wentzel - kramers - brillouin ) approximation borrowed from quantum mechanics ( or , more generally , wave mechanics ) and adapted to the dissipative , non - hermitian stochastic markov processes @xcite .
the wkb approximation employs , as a large parameter , the typical population size in the metastable quasi - stationary state . for a broad class of single - population stochastic systems
this approximation , complemented by additional perturbation techniques , yields accurate and controlled analytical results for the population extinction rate @xcite .
stochastic systems involving multiple populations present a much harder problem .
here one arrives , already in the leading order of the wkb approximation , at a generally non - integrable multi - dimensional problem of classical mechanics .
although the extinction rates and most probable paths of the system to extinction can be found numerically , analytical insight is usually limited .
the situation improves , however , if the multi - population system exhibits time - scale separation .
this may happen in two cases : ( i ) when the multi - population system is close to a bifurcation of the underlying deterministic rate equations @xcite , and ( ii ) when there is a wide difference in individual process rates .
the present paper deals with the second case .
the process rate disparity introduces an additional small parameter @xmath2 which enables one to separate , with controlled accuracy , a two - population system into a fast and slow subsystems .
each of these subsystems is one - dimensional and therefore amenable to analytical solution .
time scale separation was recently employed in ref .
@xcite for calculating the extinction rate of a biologically important component regulated by chemical reactions in a living cell . in that class of systems
the extinction probability flux sets in on the slow time scale , whereas the fast subsystem ( which rapidly adjusts to the slowly varying distribution of the slow subsystem ) modifies the effective production rate of the population on the way to extinction . in the present work we consider a different class of systems which enables us to generalize the standard notion of the quasi - stationary extinction rate by defining and calculating a _ time - resolved _ quasi - stationary extinction rate .
it turns out that this quantity smoothly changes in time from a short - time asymptote @xmath0 to a long - time asymptote @xmath1 .
the short - time quasi - stationary extinction rate @xmath0 sets in already on the fast time scale .
notably , @xmath0 coincides with the extinction rate in the case when the slow processes are absent altogether .
then , as the probability distribution evolves towards the long - time quasi - stationary distribution , the extinction rate undergoes a smooth but exponentially large change and approaches @xmath1 .
this evolution occurs on the time scale of the slow subsystem ( which is much longer than the time scale of the fast subsystem but much shorter than the mean time to extinction @xmath3 ) .
the exponentially large disparity of @xmath0 and @xmath1 is an instance of the recently discovered extinction rate fragility @xcite , where @xmath4 experiences a discontinuity when the rates of the slow processes are taken to zero .
essentially , the present work renders an alternative , time - resolved , description to the phenomenon of extinction rate fragility .
we will present the theory on the example of a well - known model of epidemiology : the stochastic sis ( susceptible - infected - susceptible ) model @xcite with a slow population turnover renewal and removal .
section [ govern ] presents the governing equations and explains how to exploit the time scale separation . in section [ solving ] we approximately solve the dynamics of the fast and slow subsystems , calculate the time - resolved disease extinction rate and compare our analytical predictions with a numerical solution of the exact master equation .
section [ solving ] also deals with the mean time to extinction of a population exhibiting a time - dependent extinction rate .
section [ discussion ] presents a discussion of our results .
the stochastic sis model @xcite describes a markov process involving susceptible and infected sub - populations . upon recovery
the infected individuals become susceptible again . the probability @xmath5 to observe , at time @xmath6 , @xmath7
susceptible and @xmath8 infected individuals is governed by the master equation with transition rates from table 1 .
one can always represent the renewal rate of susceptible individuals @xmath9 an independent parameter of the model @xmath9 as @xmath10 , where @xmath11 , an alternative independent parameter , scales as a typical average total population size in a steady state .
.stochastic sis model with population turnover [ cols="^,^,^",options="header " , ] rescaling time by the recovery rate , @xmath12 , one can write down the master equation for the sis model as @xmath13 where @xmath14 , @xmath15 , and we assume that @xmath16 if at least one of the indices @xmath7 or @xmath8 is negative
. we will assume throughout this work that @xmath17 and @xmath18 .
a slow population turnover implies smallness of @xmath19 .
for our theory to be accurate it is necessary that a strong inequality @xmath20 holds .
if one does not care , however , for pre - exponential factors , this condition can be relaxed to a much less restrictive one , @xmath21 , as we explain below .
the ultimate state of the sis model is infection - free .
when @xmath18 there is a quasi - stationary endemic state with a life time @xmath4 which is _ exponentially _ long with respect to @xmath11 .
to get an insight into how this quasi - stationary state is approached , let us consider the deterministic rate equations of the sis model which operate with the average numbers @xmath22 and @xmath23 of susceptible and infected individuals , respectively : @xmath24 these equations accurately describe the dynamics of @xmath22 and @xmath23 at times short compared with @xmath4 .
the attracting fixed point of eqs .
( [ n70 ] ) and ( [ n80 ] ) , @xmath25 describes the endemic state of the population which is typically reached on the long , _ demographic _ time scale @xmath26 .
( to remind the reader , our time is rescaled by the recovery rate @xmath27 . )
what happens on a much shorter _ epidemic _ time scale @xmath28
? it will be assumed that the initial average size of the total population , @xmath29 , is greater than @xmath30 , which is important for the further discussion . on the time scale
@xmath28 the terms proportional to @xmath19 can be neglected , and one can see that the average number of susceptible individuals approaches @xmath30 which is close to @xmath31 . in its turn , the average number of infected rapidly adjusts to the current value of the average total population size @xmath32 : @xmath33 . restoring the @xmath19-terms
, one can see that @xmath34 is a slow function of time , and its dynamics is described by the simple equation @xmath35 obtained by summing up eqs .
( [ n70 ] ) and ( [ n80 ] ) . as a result
, @xmath36 slowly flows to @xmath11 , @xmath37 and the average number of infected @xmath38 approaches @xmath39 from eq .
( [ fpm ] ) . on the phase plane @xmath40 ,
[ figphaseplane ] , the trajectory first rapidly reaches , the vertical line @xmath41 and then slowly , on the long time scale @xmath42 , approaches the ultimate fixed point ( [ fpn ] ) and ( [ fpm ] ) along this vertical line .
how is this determinisitic dynamics reflected in the actual evolution of @xmath43 ? at times @xmath44 the distribution @xmath43 is peaked at @xmath45 and evolves in time on the demographic time scale @xmath46 . at longer times , @xmath47 the distribution reaches its long - time asymptote .
it is important that a well - defined extinction probability flux sets in quite rapidly , at @xmath48 , and it varies in time on the time scale @xmath42 .
it is obvious that , in general , the disease extinction rate at @xmath49 should be very different from its value at @xmath50 . indeed , even the average numbers of susceptible and infected are generally very different at the earlier and later times .
it turns out , however , that an exponentially large disparity in the disease extinction rate at short and long times is observed even in the special case of @xmath51 , when the average numbers of susceptible and infected stay ( almost ) constant on the slow time scale @xmath42 .
the time - dependent disease extinction rate is defined as @xmath52 the limit of @xmath53 as @xmath54 is denoted by @xmath55 . for @xmath56 one
has @xmath57 . as a result , for these times @xmath58 where @xmath59 is the disease extinction probability current . to calculate @xmath59 we return to the master equation ( [ n10 ] ) and
notice that the disease can disappear from the population only via transitions from any of the states @xmath60 , where @xmath61 , to the state @xmath62 .
therefore , @xmath63 where @xmath64 as a result , for @xmath56 @xmath65 before exploiting the time - scale separation , let us first obtain some exact relations .
consider @xmath66 ( @xmath67 ) which is the probability to find @xmath68 individuals so that at least one of them is infected .
summing eq .
( [ n10 ] ) over @xmath8 while keeping the total number @xmath69 of individuals constant , we obtain an exact equation @xmath70 where @xmath71 .
an additional exact equation can be obtained once we represent the probability @xmath72 as @xmath73 where @xmath74^{-1}$ ] , defined for @xmath75 , is the probability to have @xmath8 infected individuals conditioned on @xmath76 and @xmath77 .
we will use the following shorthand : @xmath78 this conditional probability is identically equal to 0 for @xmath79 and @xmath80 , and it obeys the following exact equation : @xmath81 \nonumber \\ & & + \varepsilon\ , ( k+1)\ , \bar{p}_{k+1}\bigl[p_{k}(m)-p_{k+1}(m)\bigr]\nonumber \\ & & + \varepsilon\ , \bar{p}_{k+1}\ , \bigl[m p_{k+1}(m)-(m+1 ) p_{k+1}(m+1)\bigr]\biggl\}.\label{n60}\end{aligned}\ ] ] the disease extinction rate ( [ n20 ] ) in the new notation reads @xmath82 at @xmath56 , the exponentially small term @xmath83 in eqs .
( [ n30 ] ) can be neglected , and we obtain @xmath84 we see that the evolution of @xmath85 at @xmath56 proceeds on the slow time - scale @xmath42 and is decoupled from the evolution of @xmath86 .
equation ( [ n106 ] ) describes the relaxation of the probability distribution @xmath85 to the steady state distribution @xmath87 .
the latter is determined from the equation @xmath88 subject to normalization @xmath89 .
the evolution of @xmath86 is fast at @xmath90 and slow at @xmath91 .
the solution of eq .
( [ n60 ] ) at @xmath91 can be sought in the form @xmath92 where @xmath93 obeys the stationary equation @xmath94 where @xmath95 . for @xmath96 and @xmath97
we can confine ourselves only to the leading term @xmath93 in eq .
( [ n112 ] ) . in the following
we will omit the superscript @xmath98 .
importantly , eq .
( [ n100 ] ) does not include @xmath99 .
each of the decoupled equations ( [ n106 ] ) and ( [ n100 ] ) has a simple meaning .
equation ( [ n100 ] ) describes a one - dimensional _ quasi - stationary _ distribution ( qsd ) of the number of infected in the sis model without population turnover , where the total population size is conserved , @xmath100 .
this qsd forms relatively rapidly , at @xmath101 , so we can call this subsystem fast . once the qsd is found ,
@xmath102 yields the disease extinction rate for the given @xmath103 . in its turn , eq .
( [ n106 ] ) describes the evolution of a one - dimensional _ time - dependent _ distribution of the total population size @xmath103 .
the characteristic time scale of this time dependence is @xmath104 , so we can call this subsystem slow
. having found the slowly evolving distribution , one can calculate the time - resolved disease extinction rate from eq .
( [ n64 ] ) ( with the @xmath19-term dropped to avoid excess of accuracy ) : @xmath105 this rate is nothing but the average of the instantaneous extinction rate for a given @xmath103 ( found from the fast subsystem ) over the time - dependent @xmath103-distribution ( found from the slow subsystem ) .
this result is valid when @xmath97 and @xmath96 .
the next section deals with the solution of eqs .
( [ n106 ] ) and ( [ n100 ] ) , and with calculating @xmath106 , for a class of initial conditions for which the total number of individuals at @xmath107 is equal to @xmath11 : @xmath108 where @xmath109 is the kronecker s delta .
this special choice of initial condition will make it possible to relate the foregoing time - dependent picture of extinction to the phenomenon of extinction rate fragility .
disease extinction in the stochastic sis model without population turnover has been extensively studied starting from the pioneering paper by weiss and dishon @xcite .
the mean time to extinction of the disease in this case was first obtained by nsell @xcite , see also refs .
the extinction rate , rescaled to the recovery rate @xmath27 , is equal to @xmath110 , \label{n160a}\end{aligned}\ ] ] where @xmath111 ; here and in the following the superscript @xmath98 in @xmath112 is omitted .
equation ( [ n160a ] ) holds when the factor in the exponent is sufficiently large in absolute value @xcite : @xmath113 the particular value of the extinction rate for @xmath114 is nothing else but @xmath0 : the disease extinction rate observed at times @xmath115 , when the distribution of infected has already adapted to the current value of @xmath103 , but the @xmath103-distribution has not yet evolved and is still close to the kronecker s delta ( [ n200 ] ) : @xmath116 .
\label{qinitial}\end{aligned}\ ] ] equation ( [ n106 ] ) is exactly solvable @xcite with the help of the probability generating function @xmath117 where @xmath118 is an auxiliary variable , and @xmath119 .
once @xmath120 is found , the probabilities @xmath121 can be recovered from the taylor expansion : @xmath122 after a simple algebra the master equation ( [ n106 ] ) becomes an evolution equation for @xmath120 : @xmath123 this first - order pde can be solved by characteristics .
the general solution is @xmath124 where @xmath125 is an arbitrary function . to determine @xmath125
, one should use the initial condition . in terms of @xmath99
it is given by eq .
( [ n200 ] ) . in terms of @xmath126
we obtain @xmath127 this yields @xmath128 , and so the resulting solution , in terms of @xmath126 , is @xmath129.\ ] ] in the limit of @xmath130 we obtain @xmath131 $ ] .
this describes a poisson distribution with mean @xmath11 : @xmath132 using this poisson distribution we can calculate @xmath133 : the _ long - time _ asymptote of the disease extinction rate @xmath134 , observed when the @xmath103-distribution has already reached quasi - stationarity . as @xmath135
, we can use stirling s formula for @xmath136 and obtain @xmath137 where @xmath138 .
now we calculate @xmath139 by using the distributions ( [ n160a ] ) and ( [ finaldist ] ) . replacing the sum over @xmath103 by an integral over @xmath138 in eq .
( [ n64 ] ) , we obtain @xmath140.\end{aligned}\ ] ] as @xmath141 , the integral can be evaluated by the saddle - point method ( as a result , the exact location of the lower bound of integration is actually unimportant ) . the saddle point condition @xmath142 yields @xmath143 , corresponding to @xmath144 . by virtue of eq .
( [ n64a ] ) , @xmath144 is the most probable total population size when the number of infected is exactly one . performing the gaussian integration ,
we obtain @xmath145.\ ] ] this result , without the pre - exponential factor , was obtained by khasin and dykman @xcite who used ( the leading order of ) wkb approximation .
the important pre - exponent has not been known previously .
for the special class of initial conditions ( [ n200 ] ) , the short - time , @xmath0 , eq.([qinitial ] ) , and the long - time , @xmath1 , eq . ( [ qfinal ] ) , quasistationary extinction rates correspond to the quasistationary extinction rates observed without population turnover and with a very slow population turnover , respectively .
for @xmath146 the rate @xmath1 is _ exponentially _ larger than @xmath0 .
this exponential disparity reflects fragility of the extinction rate of the system without population turnover with respect to addition of slow population turnover .
although the fragility phenomenon was established in the @xmath19-domain @xcite , we see that a closely related phenomenon is also observed , for proper initial conditions , in the time domain .
now we discuss an important applicability criterion of our theory . the main contribution to the integral ( [ qint ] ) comes from a relatively narrow , @xmath147 , vicinity of the saddle point @xmath144 . in this vicinity , eq .
( [ n160a ] ) is valid if it describes a quasi - stationary distribution on the time scale @xmath42 . the corresponding criterion
, @xmath148 serves as the lower bound on @xmath19 for the validity of eq .
( [ qfinal ] ) . for @xmath17 and @xmath146
the quantity @xmath149 is exponentially larger than @xmath1 .
therefore , the criterion @xmath148 is much more restrictive than the obvious criterion @xmath150 .
returning to the fragility problem , we note that at @xmath151 the extinction rate experiences a gradual crossover from @xmath1 to @xmath0 in the @xmath19-domain .
let us return to eq .
( [ gsolution ] ) and find the time - dependent distribution @xmath121 from eq .
( [ probformula ] ) .
it is convenient to calculate the derivatives in the complex @xmath118-plane by using the cauchy theorem : @xmath152 where the integration contour encircles the pole @xmath153 of the complex plane @xmath118 . as @xmath154 and @xmath141
, we can evaluate the contour integral using the saddle - point approximation and deforming the contour so as to achieve the steepest descent .
using eq .
( [ gsolution ] ) , we rewrite eq .
( [ cauchy ] ) as @xmath155 where @xmath156 and @xmath138 .
the saddle point @xmath157 is determined from the equation @xmath158 . the suitable solution is @xmath159^{-1 } \left[1+x-2 \cosh \tau \right.\nonumber \\ \!\ !
( x-1 ) \cosh \tau+2 \cosh 2 \tau}\right ] .
\label{zstar}\end{aligned}\ ] ] now we expand @xmath160 at @xmath161 one has @xmath162 .
therefore , the steepest descent of function @xmath163 occurs along the straight line @xmath164 of the complex plane , see fig .
[ figcontour ] , so we deform the contour accordingly . in view of @xmath141 a small segment of the straight line
@xmath164 gives a dominant contribution to the integral , and the gaussian integration yields @xmath165 as a simple test of this result , one can go to the limit @xmath130 and obtain @xmath166 and @xmath167 . then
( [ pbar ] ) yields eq .
( [ finaldist ] ) as expected .
now we return to eq .
( [ n64a ] ) and calculate the time - resolved extinction rate @xmath168 of the disease by averaging the instantaneous extinction rate ( [ n160a ] ) over the slowly time - dependent distribution ( [ pbar ] ) . replacing the sum over @xmath103 by an integral over @xmath138 and evaluating the integral by the saddle - point method , we obtain after some algebra : @xmath169 ^ 2}{r x_s(\tau)z_*[x_s(\tau),\tau]}\sqrt{\frac{n x_s(\tau)}{2 \pi\phi^{\prime\prime } [ x_s(\tau ) , \tau]\lambda^{\prime\prime } [ x_s(\tau ) , \tau ] } } \nonumber \\ & \times&e^{-n\left[r^{-1}+x_s(\tau)\left(\ln x_s(\tau)r-1\right)+s_s(x_s(\tau ) , \tau)\right]}\,.\end{aligned}\ ] ] here the time - dependent saddle point @xmath170 is determined by the equation @xmath171=1\ , , \label{n290num}\ ] ] where @xmath172^{-1}\bigl[1-x+2\sinh\tau \nonumber \\ & & -\sqrt{(1-x+2\sinh\tau)^2 - 4(1-x)\left(e^\tau-1\right)}\ , \bigr]\,.\label{n250}\end{aligned}\ ] ] furthermore , the function @xmath173 is given by @xmath174-ce^{-\tau}\ln\left(1+ce^\tau\right)\bigr ] \nonumber\\ & + & c-\ln\left(1+c\right ) \label{n270}\end{aligned}\ ] ] with @xmath175 from eq .
( [ n250 ] ) . finally , @xmath176 and
we have written for brevity @xmath177\equiv \phi^{\prime\prime}[x_s(\tau ) , \tau]$ ] .
equation ( [ n320 ] ) is one of the main results of this work .
it describes a time - resolved disease extinction rate @xmath178 which smoothly changes from @xmath0 at @xmath49 to @xmath1 at @xmath50 .
figures [ rate ] and [ lograte ] show , by solid lines , typical examples of this behavior for two different sets of parameters . to test the theoretical time - dependent extinction rate @xmath178 predicted by eq .
( [ n320 ] ) we solved numerically a ( truncated version of the ) original master equation ( [ n10 ] ) , using the ode45 solver of matlab . a very good agreement between the theory and the numerical solution of the original master equation ( [ n10 ] ) was obtained for @xmath179 and @xmath180 , see fig .
[ rate ] . here
the time scale separation criterion @xmath96 was satisfied .
importantly , the ( quite restrictive ) criterion @xmath96 can be replaced by a much less restrictive one @xmath21 if one does not care about pre - exponential factors in eq .
( [ n320 ] ) .
indeed , criterion @xmath96 appears when one implements the time - scale - separation procedure directly in the master equation ( [ n10 ] ) .
one can follow a different strategy , however , and start with applying a time - dependent wkb approximation to the full two - dimensional master equation ( [ n10 ] ) .
a proper wkb ansatz is @xmath181\,,\ ] ] where @xmath182 and @xmath183 . in the leading wkb - order
one obtains a two - dimensional time - dependent hamilton - jacobi equation .
the corresponding hamiltonian @xmath184 is independent of @xmath11 .
the small parameter @xmath2 , present in the hamiltonian , allows one to perform a time - scale - separation procedure by seeking , for @xmath185 , @xmath186 where @xmath187 , and @xmath188 has a separable structure . our derivation , leading to eq .
( [ n320 ] ) , yields @xmath188 and @xmath189 but not @xmath190 .
the @xmath191 term makes a contribution of order @xmath192 in the exponent of eq .
( [ 2dwkb ] ) .
this contribution is negligible if @xmath193 . in this case
the pre - exponent in eq .
( [ n320 ] ) is accurate as we have already seen . on the contrary , if @xmath194 , the unknown correction @xmath195 becomes significant , and the account of the pre - exponent in eq .
( [ n320 ] ) would be in excess of accuracy .
now , what happens if we only need an accurate estimate of @xmath196 at @xmath141 ?
here both the pre - exponent in eq .
( [ n320 ] ) , and the unknown correction @xmath195 become negligible , and the result is described by the exponent in eq .
( [ n320 ] ) .
indeed , we observed an excellent agreement between theory and numerical calculations for @xmath197 when the parameter @xmath192 was comparable with , or even larger than @xmath198 , see for example fig .
[ lograte ] .
the numerical results , presented in figs . [ rate ] and [ lograte ] are converged both with respect to the truncation of the master equation , and with respect to the error tolerance of the matlab solver . versus the rescaled time @xmath199 for @xmath179 , @xmath200 and @xmath180 as predicted by eq .
( [ n320 ] ) ( solid line ) and obtained by a numerical solution of the ( truncated ) master equation ( [ n10 ] ) ( the dashed line ) . ] versus the rescaled time @xmath201 for @xmath202 , @xmath203 and @xmath204 as predicted by eq .
( [ n320 ] ) ( solid line ) and obtained by a numerical solution of the ( truncated ) master equation ( [ n10 ] ) ( the dashed line).,title="fig : " ] -10pt the time - resolved extinction rate , which we have calculated in this work , provides a sharp characterization of the stochastic population dynamics .
this characterization is lost if one is only interested in the average extinction quantities such as the mean time to extinction ( mte ) @xmath4 of the population .
to better understand this point , consider the disease extinction probability as a function of time : @xmath205 at times , @xmath206 , the growth rate of of @xmath207 obeys the relation @xmath208 which follows from eq.([ratecurrent ] ) and the conservation of the total probability . at these times
the extinction probability rate experiences a smooth but exponentially large change with time on the time scale of @xmath42 . at longer times , @xmath209 , eq .
( [ ext1 ] ) no longer holds , and should be replaced by the relation @xmath210 the mean time to extinction @xmath4 can be obtained by averaging @xmath6 over @xmath211 ( which is the probability distribution of extinction times ) : @xmath212 the dominant contribution to this integral comes from times @xmath50 where @xmath211 is determined by eq .
( [ ext2 ] ) .
therefore , up to exponentially small corrections , @xmath213 that is , the time - resolved extinction rate @xmath53 provides a much more detailed information about the system dynamics than the mte ( which only probes the late - time asymptote of the extinction rate ) .
we have addressed extinction of a population in a two - population system in the case when the population turnover renewal and removal is much slower than all other processes .
the ensuing time scale separation makes it possible to introduce a short - time quasi - stationary extinction rate @xmath0 and a long - time quasi - stationary extinction rate @xmath1 , and develop a time - dependent theory of the smooth transition between the two rates .
the quantities @xmath0 and @xmath1 coincide with the extinction rates when the population turnover is absent altogether , and present but very slow , respectively .
the exponential difference between the two rates reflects fragility of the extinction rate in the population dynamics without turnover @xcite .
the present work suggests an alternative picture of the extinction rate fragility : in the time domain instead of the @xmath214-domain where it was originally established .
our main results can be expressed in the following way .
we started out by presenting the probability distribution @xmath215 in a factorized form : @xmath216 , where @xmath85 is the probability to have the total population size @xmath103 , when at least one individual is infected , and @xmath86 is the probability to have @xmath217 infected individuals , when the total population size is @xmath103 . at @xmath218 and @xmath219 the @xmath86-distribution is independent of time to the zero order in @xmath192 ; the population turnover only affects @xmath85 . as a result ,
the time - dependent extinction rate @xmath53 is determined by the extinction rate for the total population size @xmath103 , obtained for @xmath220 and averaged over the slow - time - dependent @xmath85-distribution . the short - time quasi - stationary extinction rate @xmath0 corresponds to the initial probability distribution of the total population size @xmath103 , whereas the long - time quasi - stationary extinction rate @xmath1 corresponds to the steady - state @xmath103-distribution . under less restrictive conditions @xmath2 and
@xmath17 our theory accurately predicts the logarithm of the time - dependent extinction rate @xmath197 .
we have shown that the time - resolved quasi - stationary extinction rate encodes a more detailed information about the stochastic dynamics than the average quantities such as the mean time to extinction .
the latter quantity is determined by the late - time asymptote @xmath1 of the time - resolved extinction rate .
our analytical approach can be used in a host of two - population models which exhibit a long - lived quasi - stationary state on the way to extinction , and where a disparity between the process rates enables one to separate the system into two one - dimensional sub - systems : the fast and the slow .
one can envision two different scenarios in such systems . in one scenario
, extinction takes place only in the slow subsystem , whereas the fast subsystem merely modifies the effective process rates , as in ref . @xcite . in another scenario
the problem is reducible ( as in the sis model with a slow population turnover which we have considered here ) to averaging the instantaneous extinction rate , generated by the fast subsystem , over the time - dependent distribution of the slow subsystem . for a sufficiently large population size , the instantaneous extinction rate , generated by the fast subsystem , can be accurately calculated via wkb approximation @xcite .
how can one find the time - dependent population size distribution of the slow sub - system ?
for the sis model the one - dimensional master equation ( [ n106 ] ) for the slow population turnover is decoupled from the fast sub - system , and even exactly solvable . in general
one can not hope for such a dramatic simplification , and the fast subsystem will modify the effective process rates of the slow subsystem , as in ref .
it is important , however , that this effect can be described by a time - dependent wkb theory , once the corresponding large parameter is present .
we are grateful to mark dykman for useful discussions .
m. k. was supported in part by the army research office and by nsf grant no .
b. m. was supported by the israel science foundation ( grant no .
408/08 ) . 99 m.s .
bartlett , _ stochastic population models in ecology and epidemiology _ ( wiley , new york , 1961 ) .
s. r. beissinger and d. r. mccullough ( editors ) , _ population viability analysis _ ( university of chicago press , chicago , 2002 ) .
h. andersson and t. britton , _ stochastic epidemic models and their statistical analysis _ , lecture notes in statistics 151 ( springer verlag , 2000 ) .
r. kubo , k. matsuo , and k. kitahara , j. stat . phys . * 9 * , 51 ( 1973 ) ; hu gang , phys .
a * 36 * , 5782 ( 1987 ) ; m. i. dykman , e. mori , j. ross , and p. m. hunt , j. chem . phys . * 100 * , 5735 ( 1994 ) , m. assaf and b. meerson , phys . rev .
e * 81 * , 021116 ( 2010 ) . m. i. dykman , i. b. schwartz , and a. s. landsman , phys .
* 101 * , 078101 ( 2008 ) .
a. kamenev and b. meerson , phys .
e * 77 * , 061107 ( 2008 ) .
m. assaf and b. meerson , phys .
lett . * 100 * , 058105 ( 2008 ) .
m. khasin and m. i. dykman , phys .
* 103 * , 068101 ( 2009 ) .
weiss and m. dishon , math .
biosci . * 11 * , 261 ( 1971 ) .
i. nsell , adv .
appl . prob .
* 28 * , 895 ( 1996 ) .
h. andersson and b. djehiche , j appl . prob .
* 12 * , 662 ( 1998 ) ; i. nsell , j. theor . biol . * 211 * , 11 ( 2001 ) ; o. ovaskainen , j. appl .
prob . * 38 * , 898 ( 2001 ) , c.r .
doering , k.v .
sargsyan , and l.m .
sander , multiscale model . and simul .
* 3 * , 283 ( 2005 ) .
gardiner , _ handbook of stochastic methods _ ( springer , berlin , 2004 ) . | extinction of a long - lived isolated stochastic population can be described as an exponentially slow decay of quasi - stationary probability distribution of the population size .
we address extinction of a population in a two - population system in the case when the population turnover renewal and removal is much slower than all other processes . in this case
there is a time scale separation in the system which enables one to introduce a short - time quasi - stationary extinction rate @xmath0 and a long - time quasi - stationary extinction rate @xmath1 , and develop a time - dependent theory of the transition between the two rates .
it is shown that @xmath0 and @xmath1 coincide with the extinction rates when the population turnover is absent , and present but very slow , respectively .
the exponentially large disparity between the two rates reflects fragility of the extinction rate in the population dynamics without turnover . |
recent observations and high resolution numerical simulations of structure formation show that , on large scales , the universe is best thought of as a cosmic web @xcite : large scale sheets are traversed by filaments , which themselves intersect at knots @xcite , and there are vast relatively empty voids in between @xcite .
the knots where filaments intersect are the locations of rich galaxy clusters , and these are often identified with the massive virialized dark matter halos found in simulations .
halos are the best - studied features of the cosmic web ; they tend to be about two hundred times denser than the background universe , so they account for a small fraction of the volume of the universe .
simple analytical models have been very successful in understanding the properties of dark halos @xcite . however , with the exception of @xcite and @xcite , a framework for describing the properties and dynamics of filamentary and sheet - like structures in the cosmic web is still lacking .
the main goal of this paper is to provide such a framework : our goal is to discuss the cosmic web using language which extends naturally that currently used for discussing dark halos .
this complements recent work showing how the same framework can be used to discuss cosmic voids @xcite .
in current descriptions of large scale structure , the abundance by mass of dark matter halos and the evolution of this abundance , i.e. , the cosmological mass function @xmath3 , plays a fundamental role .
models of this mass function suggest that it encodes information about both gravitational dynamics and the statistics of the initial fluctuation field @xcite .
models of @xmath3 which assume that dark halos form from a spherical collapse are in reasonable but not perfect agreement with results from numerical simulations of hierarchical gravitational clustering @xcite . in reality , halos are triaxial @xcite , so they can not have formed from a spherical collapse .
the initial shear field and tidal effects almost certainly play some role in determining the evolution of an object , and models of the associated non - spherical collapse have been developed @xcite .
such non - spherical collapse models can be incorporated into models of the halo mass function in various ways @xcite . in particular
, @xcite argued that much of the discrepancy between model predictions and simulations is removed if one replaces the assumption of a spherical collapse with one where halos form from a triaxial collapse . in this paper
, we exploit a crucial difference between spherical and ellipsoidal collapse models of halo formation .
namely , if we think of a halo as being triaxial , then the formation of an object corresponds to the time when gravity has caused all three axes to turn around from the universal expansion and collapse @xcite . if the object is spherical , this collapse occurred at the _ same _ time for all three axes .
in contrast , a triaxial object has three critical times , corresponding to the collapse along each of the three axes the shortest axis collapses first , the intermediate axis later and the longest axis last .
if we identify ` halos ' with objects which have collapsed along all three axes , then it is natural to identify filaments with objects which have collapsed along only two axes , and sheets with objects which have only collapsed along one axis . given this identification , the remainder of this paper describes how to construct a model which naturally encapsulates the idea that halos form within filaments which themselves populate sheets .
in particular , we provide a framework for discussing the mass functions of sheets and filaments , and how these evolve , as well as the conditional mass functions of halos within filaments and sheets , and their evolution .
section [ model ] presents our model .
it begins with a discussion of our triaxial collapse model ( section [ ecmodel ] ) .
appendix a compares this model with others in the literature .
it then translates this model into a form which is most convenient for use in the ` excursion set ' approach ( sections [ excursionset][hittingtimes ] ) .
the excursion set approach is commonly used to estimate the mass function of halos @xcite .
our collapse model , when combined with the excursion set approach , has a rich structure .
we illustrate this by using it to generate analytic approximations to the mass functions of sheets and filaments ( section [ nmz ] ) , and to the conditional distributions of filaments in sheets , and halos in filaments ( section [ cmf ] ) .
comments on how mass accretion may depend on environment ( section [ accrete ] ) , and on the characteristic densities of sheets , filaments and halos , which may aid identification of such objects in the cosmic web ( section [ i d ] ) conclude this section .
section [ discuss ] summarizes our findings .
the gravitational collapse of homogeneous ellipsoids was studied by @xcite , @xcite , and @xcite .
although the exact evolution must be solved - for numerically , @xcite provide an elegant analytic approximation for the evolution which is remarkably accurate .
unfortunately , these early analysis did not reduce , as they should , to the zeldovich approximation in the linear regime .
@xcite noted that this was because they did not include the effects of the external tide self - consistently ; once these effects are included , the collapse model is indeed self - consistent . appendix a provides details , and also shows how the analysis of @xcite can be extended to this self - consistent case . in this model
, the initial collapse is expected to be dominated by the local strain tensor , which includes both internal and external tidal forces . for any given cosmology ,
the evolution of a patch in this model is determined by three quantities : its initial overdensity , @xmath4 , and two shape parameters , @xmath5 and @xmath6 .
it is this model which we will use in what follows . in the ` excursion
set ' approach @xcite , an approximation for the mass fraction in bound virialized halos of mass @xmath7 is obtained by mapping the problem to one which involves the first crossing distribution of a suitably chosen barrier by brownian motion random walks .
the choice of barrier is set by the collapse model and by the epoch for which one wishes to estimate the mass function , and the mapping between random walk variables and halo masses is set by the shape of the power spectrum of the initial fluctuation field . in particular ,
if @xmath8 denotes the fraction of walks which first cross a barrier at scale @xmath9 , where @xmath9 is the random walk variable , then the mass fraction in objects of mass @xmath7 is @xmath10 @xcite . here
@xmath11 is the variance in the initial density fluctuation field , smoothed on scale @xmath12 , and extrapolated using linear theory to the present time .
if the collapse is spherical , then the barrier is particularly simple : it has constant height , with higher barriers required to model the halo population at higher redshift . in particular
, this height is very simply related to the initial overdensity required for spherical collapse by redshift @xmath13 .
this value , which is usually denoted @xmath14 , will play a special role in what follows . in the ellipsoidal collapse model ,
the barrier height associated with halo formation by redshift @xmath13 depends on three numbers : the initial overdensity , and the shape parameters @xmath5 and @xmath6 . in principle , then , the problem of estimating the mass function is one of crossing a barrier in a higher dimensional space @xcite .
@xcite suggested that the computational complexity could be reduced significantly if one used suitably chosen representative values of @xmath5 and @xmath6 .
once these have been set ( for reasons given in appendix a , they suggested @xmath15 and @xmath16/\sqrt{5}$ ] ) , the excursion set approach corresponds to finding the first crossing distribution by a _
one - dimensional _ random walk in @xmath4 of a barrier whose height is _ mass - dependent_. in this respect , the estimation of the mass function associated with ellipsoidal collapse proceeds exactly as for spherical collapse
the only complication is that the barrier height is no longer constant .
our model for the mass function in filaments and sheets is entirely analogous . in principle , filaments at @xmath13 can form from some combination of @xmath4 , @xmath5 and @xmath6 ; the precise combination depends on the details of the collapse model . in our preferred model , the same one used by @xcite ,
this requires that collapse occurs along the intermediate length axis at @xmath13 , but along the longest axis at a later time .
thus , estimation of the mass fraction in filaments requires solution of a high - dimensional random walk , and , as for collapse along all three axes , we approximate by setting @xmath5 and @xmath6 equal to representative values , and considering a one - dimensional walk in the remaining variable @xmath4 . in this case , the barrier height associated with filament formation is that associated with collapse along two , rather than three axes . and
a similar argument means that the mass function in sheets can be approximated by studying the first crossing distribution of the one - dimensional barrier associated with collapse along just one axis .
figure [ randomwalk ] illustrates the basic idea .
the jagged line shows a random walk , and the three dotted lines show the barriers associated with collapse along one ( bottom ) , two ( middle ) and three ( top ) barriers . in our model
, the fraction of walks which first cross the lowest barrier on scale @xmath17 equals the mass fraction in sheets of mass @xmath7 , the fraction which first cross the second barrier at @xmath17 equals the mass fraction in filaments of mass @xmath7 , and the fraction which first cross the highest barrier at @xmath17 equals the mass fraction in triaxial halos of mass @xmath7
. we will provide estimates of these quantities shortly .
however , figure [ randomwalk ] shows that this model can also be used to provide significantly more information .
for instance , notice that all walks first - cross the barrier associated with sheets at a higher mass scale than when they first - cross the barriers associated with filaments or halos .
thus , in addition to providing a way to estimate the mass functions of sheets , filaments and halos , our approach also provides a framework for discussing the mass fraction in halos of mass @xmath18 that are embedded in filaments of mass @xmath19 which are themselves surrounded by sheets of mass @xmath20 .
figure [ randomwalk ] shows that this is simply related to the fraction of walks which are conditioned to pass through certain points before first crossing @xmath18 .
thus , our approach allows one to estimate how halo abundances correlate with the _ morphology _ and density of the large - scale structure which surrounds them .
so far , we have only discussed the model at a fixed redshift .
since the barrier shapes depend on redshift , our approach also allows one to estimate how sheet , filament and halo abundances evolve .
moreover , it allows one to address questions like : is the halo abundance in sheets of mass @xmath21 at @xmath22 any different from the mix of halos in @xmath21 filaments today ?
our model naturally incorporates the fact that the mass in high redshift sheets becomes partitioned up among filaments at later times , with filaments themselves being partitioned into halos . clearly , the model is rich in spatial , temporal and spatio - temporal information .
the key output from the triaxial collapse models is an estimate of the typical overdensity required for collapse along one , two and three axes by redshift @xmath13 .
the dotted curves in figure [ randomwalk ] show how these three ` barriers ' depend on mass . from bottom to top ,
the curves show @xmath23 ^{0.55 } \right \ } , \nonumber \\
\delta_{ec2}(\sigma , z ) & = & \delta _ { sc } ( z ) \left \ { 1 - 0.012\left[{{\sigma ^2 } \over { \delta _ { sc}^2 ( z)}}\right ] ^{0.28 } \right \ } \approx \delta _ { sc } ( z ) , \label{movingbarriers } \\
\delta_{ec3}(\sigma , z ) & = & \delta _ { sc } ( z ) \left \ { 1 + 0.45 \left[{{\sigma ^2 } \over { \delta _ { sc}^2 ( z)}}\right]^{0.61 } \right \}. \nonumber\end{aligned}\ ] ] these analytic approximations to the barriers associated with collapse along one , two and three axes are reasonably accurate ( c.f .
figure [ shethall ] ) .
they show that the critical overdensity for ellipsoidal collapse along all three axes , @xmath24 , is larger than in the spherical collapse model ( for which this number is @xmath25 ) .
however , collapse along just two axes requires an overdensity @xmath26 which is almost exactly that in the spherical model , and collapse along one axis only requires a smaller initial overdensity , @xmath27 . in this model ,
tidal forces enhance collapse along the first axis and delay collapse along the last axis relative to the spherical collapse model @xcite the expressions above quantify these effects .
note that the differences among the three critical overdensities are larger for larger values of @xmath28 , corresponding to ellipsoids of lower masses .
notice that these barrier shapes depend both on @xmath17 and on @xmath14 .
the presence of these two terms reflects the fact that the collapse depends on the expansion history of the universe , and on the initial spectrum of fluctuations .
what is of particular importance in what follows is that the barrier shapes actually depend , not on @xmath25 and @xmath28 individually , but on the scaling variable @xmath29 ^ 2 $ ] . in the excursion set approach
, this implies that the mass functions of sheets , filaments and halos at any given time , in any given cosmology , and for any given initial fluctuation spectrum , can all be scaled to universal functional forms .
our next step is to provide analytic approximations to these forms . in the simplest
excursion set model , object abundances are related to the first crossing distributions by one - dimensional random walks of the moving barriers in equation ( 2 ) .
these distributions are easily obtained by monte - carlo simulation .
( alternative numerical methods are also available , e.g. , zhang & hui 2006 . )
the monte - carlo results are simplified considerably by the fact that the barriers can be expressed in terms of the scaling variable @xmath29 ^ 2 $ ] .
the histograms in figure [ fcross ] show the results .
the abundances of sheets , filaments and halos are related to these first crossing distributions by equation ( [ nufnu ] ) .
the smooth curves show analytic approximations to these first crossing distributions , computed following @xcite .
namely , when the barrier is of the form @xmath30 , \nonumber \label{barrier}\ ] ] as is the case for all three barriers in equation ( [ movingbarriers ] ) , then the first crossing distribution is approximately @xmath31 ^ 2/2 } \left\{1+\frac{\beta}{{\nu}^{\alpha } } \left [ 1- \alpha + \frac{\alpha(\alpha -1)}{2 ! } + \dots \right ] \right\ } . \label{approx}\ ] ] alternative analytic estimates could be computed by noting that when @xmath32 , then the first crossing distribution can be written analytically in terms of sums of parabolic cylinder functions @xcite .
given that these would only provide approximations to the first crossing distributions we require ( because @xmath33 ) , we have chosen the considerably simpler approximation given in equation ( [ approx ] ) .
figure [ fcross ] indicates that these approximate solutions are sufficiently accurate to allow analytic estimates of a number of interesting quantities .
upon assumption of a cosmology [ which determines @xmath14 ] and a fluctuation spectrum [ which determines @xmath17 ] , the first crossing distributions shown in figure [ fcross ] can be converted to mass functions using equation ( [ nufnu ] ) . in what follows ,
we assume a spatially flat model with cosmological constant @xmath36 , where @xmath37 .
the dashed lines in the bottom - left panel of figure [ numfracall ] show the evolution of the number densities of halos having mass @xmath1 ( top ) to @xmath38 ( bottom ) .
except for the lowest mass range , these curves decrease monotonically with increasing redshift .
this reflects the well - known hierarchical nature of structure formation : at any given redshift , massive halos are less common than halos of lower mass , and they are even less common at earlier times . as a result ,
if one integrates the mass fraction in objects above some minimum mass , then this fraction is smaller at higher redshift .
if , instead , one fixes the mass fraction , then one must integrate down to smaller mass halos to obtain this fraction . since less massive halos are more abundant , plots of the abundance associated with a fixed mass fraction increase with increasing redshift @xcite .
the solid curves show such locii in the abundance - redshift plane for a range of mass fractions .
the top - left and right panels show analogous results , but for sheets and filaments respectively .
similar trends are seen in all three panels , indicating that sheets and filaments grow hierarchically in much the same way as halos .
these panels show that , at @xmath39 , more than half of the cosmic mass is in sheets with masses exceeding @xmath40 .
this fraction is about @xmath41 for filaments in the same mass range , and is only @xmath42 for halos . at the present time , more than @xmath43 of the cosmic mass is contained in sheets more massive than @xmath44 , in contrast to virialized halos which , in this mass range , contain only about half of the cosmic mass . at @xmath45 , more than 10% of the cosmic mass was already assembled in sheets more massive than that of the milky way halo ( @xmath46 ) , while only @xmath47 was assembled to halos in this mass range .
even at @xmath48 , more than @xmath49 of the cosmic mass was already in sheets more massive than @xmath50 , whereas the mass fraction in halos in the same mass range was ten times smaller .
this suggests that the structure at high @xmath13 is dominated by sheets . at
any given redshift , we may define a characteristic mass , @xmath51 , so that @xmath52 is equal to the critical overdensity @xmath14 . in our @xmath0cdm model ,
the value of @xmath53 drops rapidly with redshift , from just over @xmath40 at @xmath39 to @xmath54 at @xmath55 .
objects at @xmath13 whose mass satisfies @xmath56 , where @xmath57 , can be thought of as being increasingly unusual compared to objects of the characteristic mass @xmath51 .
the top two solid curves in the first three panels of figure [ numfracall ] show the abundance - redshift locii of sheets , filaments and halos whose masses satisfy @xmath58 and @xmath59 .
the panels show that the mass fraction in @xmath60 sheets , @xmath61 , is larger than that in @xmath60 filaments ( 0.32 ) and than that in @xmath60 halos ( 0.17 ) .
a similar scaling applies to the more massive @xmath62 objects , implying that , at any redshift , the typical structure is dominated by sheets .
the bottom right panel in figure [ numfracall ] compares the number densities of sheets ( dotted ) , filaments ( dashed ) and halos ( solid ) at different redshifts in the three other panels directly ( i.e. , these were the dashed curves in the other three panels ) : top - most curves show our model for masses of @xmath63 , bottom - most for @xmath21 , and the curves in between show results in which the mass changes by 1 dex .
these curves show that , at masses larger than @xmath64 , there are more sheets than filaments , and more filaments than halos , for almost all redshifts .
however , at small masses and late times this trend begins to reverse : at @xmath39 there are fewer sheets than filaments with masses smaller than @xmath65 , and there are more halos than sheets of mass @xmath66 . based on these mass functions
, we can estimate the average mass of objects more massive than some given minimum mass .
for instance , when the minimum mass is @xmath66 then , at the present time , this average mass is @xmath67 for sheets , @xmath68 for filaments , and @xmath69 for halos .
the average mass of sheets is about 2.5 times that of halos , again suggesting that large scale structure is dominated by sheets rather than by virialized halos .
halo abundances are expected to correlate with the overdensity of their surroundings massive halos populate dense regions @xcite .
however , our model predicts that halo abundances will also correlate with the morphology of their surroundings . to illustrate , the solid curve in figure [ cosmicweb ] shows the mass fraction of @xmath70 sheets that is in filaments of mass @xmath7 at @xmath39 .
the figure indicates that the most probable filament mass , @xmath71 , is a substantial fraction of that of its parent sheet .
although we do not show it , the analytic estimate of this quantity , given by inserting our barriers for filaments and sheets in equation ( 7 ) of @xcite , is in excellent agreement with this distribution .
this solid curve should be compared with the dashed one , which shows the mass fraction in halos ( rather than filaments ) of mass @xmath7 that are known to be in sheets of mass @xmath70 at @xmath39 .
( in this case , the analytic estimate is not as accurate . )
notice that the most probable halo mass is significantly smaller than that of the parent sheet , and the halo mass function in sheets is skewed significantly towards lower masses than that of filaments . as a final comparison ,
the dotted line shows the mass fraction of @xmath70 filaments ( rather than sheets ) that is in halos of mass @xmath7 , and the dot - dashed curve shows the mass fraction contained in halos of mass m within an average volume of the universe of the same mass ( @xmath72 ) . here again , the mass function within sheets and filaments is biased compared to the average universe and the most probable halo mass is a substantial fraction of the mass of the parent filament .
the significant differences between the dotted and dashed curves illustrate an important physical implication of our model : if the properties of a galaxy are correlated with the mass of its host halo , then the galaxy population in sheets of a given mass is expected to be different from that in filaments of the same mass . in other words , at fixed large - scale mass , galaxy properties are expected to be correlated with the morphology of the surrounding large scale structure . figure [ cosmicweb ] shows the distribution of masses of smaller scale objects when the large scale mass and morphology is fixed .
but our model also allows one to estimate the mass and morphology of the environment which surrounds a given smaller scale structure .
to illustrate , figure [ cosmicweb2 ] shows conditional mass functions in which the small - scale object is fixed , and we study the distribution of the surrounding large scale mass and morphology .
the solid curve shows the distribution of sheets around @xmath64 filaments , and the dotted and dashed curves show the distributions of filaments and sheets around @xmath64 halos .
( once again , the analytic estimates of these distributions are in reasonable but not perfect agreement . ) notice again that the mass of a sheet is well correlated with the mass of the filament it surrounds , as are the masses of filaments with those of the halos they surround . on the other hand
, halo and sheet masses can be very different .
these scalings indicate that our model can be used to estimate how the surrounding large scale structure correlates with different galaxy populations .
our model can be used to provide a simple estimate of how halos grow . to see why ,
consider a random walk with height equal to the rms fluctuation : @xmath28 .
the typical mass scales of sheets , filaments and halos are given by setting this value equal to the various barrier heights : @xmath73 .
to see what this implies , approximate the three moving barriers in equation ( 2 ) as @xmath74 then the typical mass scale of filaments is @xmath75 .
similarly , a characteristic mass scale for sheets is @xmath76 , implying @xmath77 , whereas the mass scale of halos is @xmath78 .
massive objects are associated with more extreme fluctuations .
we can approximate by studying the case in which the walk height is twice the rms value ( a @xmath62 fluctuation ) . in this case ,
@xmath79 , @xmath80 , and @xmath81 .
notice that , for massive objects , these factors are substantially closer to unity than they are for more typical objects .
this suggests that the most massive halos are a substantial fraction of the mass of the filaments they populate , and these filaments are a substantial fraction of the sheets in which they are embedded .
thus , massive halos will appear to be accreting most of their mass from filaments .
in contrast , a lower mass halo may be substantially less massive than the filament it inhabits .
the growth of such a halo will appear to be dominated by mergers with the other small halos which populate the same filament ( and possibly the other nearby filaments ) . in this respect ,
our model is qualitatively consistent with observations which suggest that massive clusters grow by accretion along filaments , whereas the growth of lower mass objects is less obviously anisotropic .
although our approach provides a framework for discussing the morphology of the cosmic web , we have , so far , not made any statements about precisely how the sheets , filaments and halos in our formalism are to be identified in cosmic density fields .
our collapse model is calibrated so that the object which forms after collapse has been completed along all three axes , i.e. , a halo , has the same density relative to the background as is expected in the spherical collapse model .
sheets in our model should be approximately this overdensity to the one - thirds power , whereas filaments are this to the two thirds power .
although , formally , the density of a halo depends on the background cosmological model , it is common practice to identify halos at a given epoch as objects which are 200 times denser than the critical density at that epoch .
therefore , filaments and sheets should be approximately 36 and 6 times denser than the critical density .
this provides a simple rule of thumb for identifying sheets and filaments in the dark matter density field .
note that , in the context of our model , it may be more appropriate to think of massive halos as lying at the centers of filaments , rather than defining their endpoints .
it will be interesting to see how these simple estimates compare with measurements in simulations .
we extend the excursion set approach to quantify how the cosmic web is made up of sheets , filaments and halos .
our model assumes that objects form from a triaxial collapse ; we define sheets as objects which have collapsed along only one axis , filaments as objects which have collapsed along two axes , and halos as objects where all three axes have collapsed .
therefore , our model requires specification of exactly how triaxial collapse occurs .
appendix a discusses our preferred collapse model , compares it with zeldovich s approximation , and shows how the analytic arguments of @xcite can be extended to provide an accurate analytic description of our ellipsoidal collapse model .
the details of how a patch collapses depends on its initial overdensity @xmath4 , and on its initial shape parameters @xmath5 and @xmath6 . to embed this collapse model in the excursion set approach requires study of three - dimensional random walks crossing a barrier @xmath82 . here ,
we follow @xcite and study the simpler problem in which @xmath5 and @xmath6 are replaced by representative values , and then study a one - dimensional boundary crossing problem .
we emphasize that this is only an approximation , albeit a useful one . for any redshift @xmath13 ,
insertion of the representative values of @xmath5 and @xmath6 in our collapse model provides estimates of the critical overdensities required for collapse along one , two and three axes .
we find that these overdensities , @xmath83 , @xmath84 and @xmath85 depend on both time and mass ( equation 2 ) . because of the dependence on @xmath28 , in the language of the excursion set approach ,
each of these critical overdensities is a ` moving barrier ' ( figure [ randomwalk ] ) .
insertion of each moving barrier into the excursion set approach provides estimates of the mass fraction in sheets , filaments and halos as a function of mass and time ( figure [ fcross ] ) . in our model , halos of a given mass @xmath18
populate filaments which are more massive , and the filaments themselves are surrounded by even more massive sheets .
hence , the characteristic masses of sheets are predicted to be substantially larger than of filaments or halos .
every halo at a given time was previously a filament of the same mass , and before that , a sheet .
a halo of mass @xmath86 today was a filament at redshift @xmath87 and a sheet at @xmath88 ( c.f .
equation 2 ) .
halo abundances are expected to correlate with the overdensity of their surroundings ( massive halos populate dense regions ) .
our model predicts that , at fixed large - scale overdensity , halo abundances will also correlate with the morphology of their surroundings ( figures [ cosmicweb ] and [ cosmicweb2 ] ) .
therefore , in models where the properties of a galaxy are correlated with the mass and formation history of its host halo , our model provides a framework for describing correlations between galaxy properties and the morphology of large scale structure .
for instance , distributions like those in figure [ cosmicweb ] may be used to study if morphological structures in the galaxy distribution , like the sdss great wall at @xmath89 , are unusual . and
distributions like those in figure [ cosmicweb2 ] provide a framework for understanding if galaxies at high redshift form preferentially in sheets or filaments @xcite . because our model exhibits various physically appealing features
, we anticipate that it will provide a useful framework for quantifying the relation between galaxies , halos , and the cosmic web .
we would like to thank gerhard brner for organizing a meeting at the ringberg castle , tegernsee , and the staff of the castle for their hospitality when this project began .
we thank joerg colberg and an anonymous referee for insightful comments and questions .
js thanks the university of massachusetts at amherst for hospitality during the summer of 2004 .
this work was supported by the nsf under career award ast-0239709 to ta , and grant ast-0520647 to rks .
bharadwaj , s. , sahni , v. , sathyaprakash , b.s . ,
shandarin , s.f . , yess , c. , 2000 , , 528 , 21 bharadwaj , s. , & biswajit , p. , 2004
, , 615 , 1 bhavsar , s.p . ling , e.n . , 1988 , , 331 , l63 bhavsar , s. , bharadwaj , s. , sheth , j. , 2003 , aip conference proceedings , 666 , 303 biswajit , p. , & bharadwaj , s. , 2005 , , 357 , 1068 breiman , l. , 1967 , 5th berkeley symp .
2(2 ) : 9 bond , j.r . , cole , s. , efstathiou , g. , kaiser , n. , 1991 , , 379 , 440 bond , j.r . , & myers , s. , 1996 , , 103 , 1 bond j. r. , kofman l. , pogosyan d. , 1996 , , 380 , 603 chiueh , t. , & lee , j. , 2001 , , 555 , 83 colberg j. , krughoff s. , connolly a. , 2005 , mnras , 359 , 272 doroshkevich , a.g . , 1970 , astrofizika , 3 , 175 efstathiou , g. , frenk , c. s. , white , s.d.m . , davis , m. , 1988 , , 235,715 eisenstein , d.j . ,
loeb a. , 1995 , , 439 , 520 gregory , s.a . , thompson l. a. , 1982 , scientific american , 246 , 106 gunn j.e .
, gott j.r .
iii , 1972 , , 176 , 1 gurbatov s. n. , saichev a. i. , shandarin s. f. , 1989 , , 236 , 385 icke , v. , 1973 , , 27 , 1 jing , y.p . , suto , y. , 2002 , , 574 , 538 kirshner r. p. , oemler a. , schechter p. l. , schectman s. a. , 1983 , iaus , 104 , 197 lee , j. , shandarin , s.f . , 1998 , apj , 500 , 14 mo h.j .
, white s.d.m . , 1996 , , 282 , 347 mo h.j . , white s.d.m . , 2002 , , 336 , 112 mo h.j .
, yang x. , van den bosch f.c . , katz n. , 2005 , , 363 , 1155 monaco , p. , 1995
, , 447 , 23 mller v. , arbabi s. , maulbetsch c. , 2003 , iaus , 216 , 79 peebles p.j.e . , 1980 , the large - scale structure of the universe , princeton university press press , w. , schechter , p. , 1974
, , 187 , 425 shandarin , s.f . , sheth , j.v . , & sahni , v. 2004 , , 353 , 162 sheth , j.v . , sahni , s. , shandarin , s.f . , & sathyaprakash , b.s .
, 2003 , 343 , 22 sheth r. k. , tormen g. , 1999 , , 308 , 119 sheth , r.k .
, mo , h.j . , tormen , g. , 2001 , , 323 , 1 sheth , r.k . , tormen , g. , 2002 , , 329 , 61 sheth , r.k . ,
van de weygaert , r , 2004 , , 350 , 517 vogeley m. s. , hoyle f. , rojas r. r. , goldberg d. m. , 2004 , iauc , 195 , 5 white , s.d.m . , silk , j. , 1979 , . , 231 , 1 zeldovich , ya.b . , 1970 , , 5 , 84
this appendix describes the triaxial collapse model we use in the main text .
although the evolution must be solved numerically in general , we also discuss a reasonably accurate analytic approximation to the evolution , and show that it is considerably more accurate than the zeldovich approximation .
let @xmath90 denote the scale factors for the three principal axes of the ellipsoid
. then @xmath91 } \label{ellipsoid}\ ] ] @xcite , where @xmath92 is the mean density of the universe , @xmath93 is the relative overdensity , and @xmath94 and @xmath95 denote the interior and exterior tidal forces . in particular , @xmath96 \prod ^3 _ { m=1 } [ a_m^2(t)+\tau]^{1/2}}},\ ] ] and the linear approximation for the external tides is @xmath97\,,\ ] ] where @xmath98 .
note that the @xmath99s are the initial eigenvalues of the strain tensor ; they are related to the initial density contrast @xmath100 and the shear ellipticity @xmath5 and prolateness @xmath6 by : @xmath101 the initial conditions are set by zeldovich approximation : @xmath102 , \ \
{ \rm and } \ \
\dot{a}_k(t_i ) = h(t_i)a_k(t_i ) -a_i h(t_i)\lambda_{k}(t_i).\ ] ] these expressions reduce to those of the spherical collapse model if the two tidal force terms are not included .
note that , as a result of the tidal forces , an initially spherical region will be distorted into a collapsing homogeneous ellipsoid .
the tidal gravitational forces accentuate the asymmetry of the ellipsoid , so that the collapse of the three axes can happen at very different times .
the shortest axis will collapse first , followed by the intermediate axis and then by the longest axis .
absent angular momentum , the collapse of any given axis continues to arbitrarily small sizes , a well - known feature of the spherical collapse model as well . given initial values of @xmath103 , ellipticity @xmath5 , prolateness @xmath6 , and epoch @xmath104 , equation ( [ ellipsoid ] )
must be solved numerically for each axis @xmath90 .
however , it is straightforward to extend the analytic approximation for ellipsoidal collapse provided by @xcite so that it reduces self - consistently to the zeldovich approximation at early times . in particular
, we write @xmath105 - { a(t)\over a_i}a_h(t_i ) \left [ 1-\frac{d(t ) \delta_i}{3 } - \frac{a_e(t)}{{a}(t ) } \right ] , \label{mowhitesilk}\ ] ] where @xmath106 , @xmath107 is the linear theory growth mode , and @xmath108 is the expansion factor of a universe with initial density contrast @xmath109 . note that the first term in the expression above is the zeldovich approximation to the evolution .
also note that , if all three axis are initially the same , then all the @xmath110s are the same , and @xmath111 . in this case , the perturbation is a sphere , and the expression above reduces to @xmath112 $ ] , so the approximation ( [ mowhitesilk ] ) is exact .
figure [ bondmo ] compares the full numerical evolution of the three axes in this model ( solid ) with the zeldovich approximation ( dotted ) and with equation ( [ mowhitesilk ] ) ( dashed ) , when the initial values are @xmath113 in an einstein de - sitter universe .
this shows clearly that , at late times , equation ( [ mowhitesilk ] ) is significantly more accurate than the zeldovich approximation .
the numerical solution allows each axis to shrink to arbitrarily small sizes . in studies of the spherical evolution model it is standard to assume that virial equilibrium leads to a non - negligible freeze - out radius . for an einstein de - sitter model , this freeze - out radius is @xmath114 , so the final object is @xmath115 times denser than the background universe .
@xcite suggested that this same factor could be used in the ellipsoidal collapse model as well : hence , we stop the collapse of axis @xmath116 by hand when @xmath117 .
thus , in this model , initial values of @xmath99 and @xmath104 yield initial values of @xmath118 and @xmath6 .
the collapse model then provides an estimate of the times at which each axis freezes - out .
in particular , note that if @xmath5 and @xmath6 are given , then there is a unique value of @xmath100 which will produce collapse of the @xmath116th axis at redshift @xmath13 .
the symbols in figure [ shethall ] show the critical densities for collapse at @xmath39 when @xmath15 and @xmath119 along one ( bottom ) , two ( middle ) and three ( top ) axes . in the random walk model discussed in the main text , these values of @xmath5 and @xmath6 , and this relation between the critical density required for collapse at @xmath13 and
the initial shape play a central role .
the solid curves show the simple fits to these relations given in equation ( 2 ) .
notice that the critical density for collapse of the second axis is _ very _ similar to that for a spherical model .
the approximation ( [ mowhitesilk ] ) provides an easy way to see why .
first , note that to second order in @xmath100 , @xmath120 second , recall that when @xmath15 then @xmath121 .
third , when @xmath15 and @xmath122 ( our representative values ) then the final term in the expression above is @xmath123 , so @xmath124 .
therefore , for @xmath125 , @xmath126 - { a(t)a_h(t_i)\over a_i a_2(t_i ) } \left [ 1-\frac{d(t ) \delta_i}{3 } - \frac{a_e(t)}{{a}(t ) } \right ] \approx { a_e(t)\over a_i};\ ] ] in this approximation , the second axis evolves _ exactly _ as in a spherical model with initial overdensity @xmath100 . | we discuss an analytic approach for modeling structure formation in sheets , filaments and knots .
this is accomplished by combining models of triaxial collapse with the excursion set approach : sheets are defined as objects which have collapsed along only one axis , filaments have collapsed along two axes , and halos are objects in which triaxial collapse is complete . in the simplest version of this approach , which we develop here , large scale structure shows a clear hierarchy of morphologies : the mass in large - scale sheets is partitioned up among lower mass filaments , which themselves are made - up of still lower mass halos .
our approach provides analytic estimates of the mass fraction in sheets , filaments and halos , and its evolution , for any background cosmological model and any initial fluctuation spectrum . in the currently popular @xmath0cdm model
, our analysis suggests that more than 99% of the cosmic mass is in sheets , and 72% in filaments , with mass larger than @xmath1 at the present time . for halos ,
this number is only @xmath2 .
our approach also provides analytic estimates of how halo abundances at any given time correlate with the morphology of the surrounding large - scale structure , and how halo evolution correlates with the morphology of large scale structure . |
very high energy ( @xmath3 tev ; vhe ) neutrinos are unique messengers of nonthermal phenomena in the universe related to the hadronic interactions of protons and nuclei in cosmic tevatrons and pevatrons - nature s masterly designed machines accelerating particles to tev and pev energies . in this regard
vhe neutrinos are complementary to @xmath1-rays which are produced both in electromagnetic and hadronic interactions . on the other hand , unlike @xmath1-rays , neutrinos are not fragile ; they interact only weakly with the ambient medium - gas , radiation and magnetic fields , and thus carry information about high energy processes occurring in hidden regions where the particle accelerators could be located .
this concerns , first of all , the regions associated with compact objects - black holes , pulsars , the initial epochs of supernovae explosions , _ etc_. the penetrating potential of neutrinos is important not only for extremely dense environments in which @xmath1-rays are dramatically absorbed , but also moderately opaque sources from which we can see @xmath1-rays , but after significant distortion due to internal and external absorption .
ironically , this nice ( from an astrophysical point of view ) feature of neutrinos makes , at the same time , their detection extremely difficult .
this explains why over several decades high energy neutrino astronomy has remained essentially a theoretical discipline with many exciting ideas and predictions but without the detection of a single vhe neutrinos source .
however , it is expected that , with arrival of the km@xmath4-volume class scale detectors like icecube and km3net ( see e.g. refs .
@xcite ) , the status of the field will be changed dramatically .
generally , predictions of vhe neutrino fluxes from astrophysical objects rely on assumptions with not well - constrained model parameters and , therefore , often contain large ( orders of magnitude ! ) uncertainties .
this leaves a significant freedom in speculations on the `` best - bet neutrino sources '' , and consequently allows a broad spectrum of opinions concerning the prospects for detecting the first astrophysical neutrinos - from very enthusiastic statements to rather careful predictions prevailed by a healthy scepticism ( see e.g. ref .
@xcite ) . presently extragalactic objects like active galactic nuclei ( agn ) and sources of gamma ray bursts ( grbs ) are believed to be the most likely objects to be detected as neutrino sources , and therefore the driving motivation of experimental vhe neutrino astronomy ( see e.g. ref .
the current models of agn and grbs indeed contain many attractive components ( concerning the conditions of particle acceleration and their interactions ; see e.g. ref .
@xcite ) that makes these objects _ potentially detectable _ sources of vhe neutrinos ( see e.g. refs .
@xcite ) . on the other hand , the poor understanding of certain aspects of the physics of agn and especially grbs , as well as the lack of constraints on neutrino productions rates from @xmath1-ray observations ( because of intrinsic and intergalactic absorption of vhe @xmath1-rays ) ,
formally allow calculations in extreme model - parameter segments which often lead to rather high ( in the case of blazars - overoptimistic ) neutrino flux predictions .
in contrast to extragalactic objects , the models of potential galactic neutrino sources , in particular the shell type supernova remnants ( snrs ) , pulsar wind nebulae ( pwne ) , star formations regions and the dense molecular clouds related to them , are robustly constrained by @xmath1-ray observations of the galactic disk in very - high ( @xmath5 tev ) @xcite and ultra - high energy ( @xmath6 tev ) @xcite domains . typically , the expected fluxes from these objects are below the detection threshold of the planned neutrino detectors .
however , the recent hess discoveries of several tev @xmath1-ray sources at the flux level of `` 1 crab '' , which can be interpreted within the hadronic models of gamma - ray emission , sustain a hope that first tev galactic sources will be detected in foreseeable future by km@xmath4-volume class instruments like icecube and km3net .
the recent performace studies of the km@xmath4-volume scale detectors show that the detection of persistent point - like neutrino sources ( for a typical angular resolution of vhe neutrino detectors the `` point - like '' source implies an object of angular size @xmath7 ) for a realistic exposure time ( typically , a few years continuous observations ) is limited by a flux @xmath8 ( see e.g. refs .
the corresponding energy flux is @xmath9 or somewhat less , depending on the spectrum in the most relevant energy interval between 1 and 100 tev .
this exceeds , by two orders of magnitude , the minimum @xmath1-ray flux detectable in the same energy band . on the other hand ,
the sensitivity of the km@xmath4-scale detectors is comparable or better than the minimum detectable energy flux achieved by the compton gamma ray observatory detectors ( comptel , egret ) in the mev / gev @xmath1-ray band . for an isotropic vhe source located at a distance @xmath10 , the luminosity of tev neutrinos
can be probed at the level @xmath11 at first glance , this is a quite modest luminosity , at least for a powerful hadronic source located in a dense environment . indeed , for production of tev neutrinos in _
p - p _ interactions with ambient gas of density @xmath12 , the required total energy in multi - tev protons is estimated @xmath13 , where @xmath14 is the radiative cooling time of protons due to inelastic _ p - p _ interactions , and @xmath15 is the fraction of average energy of a proton transferred to neutrinos .
one may conclude that even in a relatively low density environment ( @xmath16 ) the required total energy can be readily produced in young snrs through diffusive shock acceleration ( see e.g. ref .
@xcite ) or by a powerful pulsar assuming that a major fraction of the spin - down luminosity of the pulsar is converted to an ion - dominated wind ( see e.g. ref .
@xcite ) . however , this kind of estimates can be misleading since they are based on a silent assumption that all particles accelerated during the life - time of the source are effectively confined in a relatively compact region inside or nearby the accelerator .
in fact , production of tev neutrinos requires protons with energies well beyond 10 tev the escape of which from the source is difficult to prevent .
this , of course , would lead to a significant reduction of the neutrino production efficiency which in the case of _ p - p _ interactions can be expressed in the following form : @xmath17 \times c_{\rm
p \rightarrow \nu}$ ] , where @xmath18 is the escape time of nonthermal particles .
if the escape proceeds in the diffusion regime , then @xmath19 .
it is convenient to write the diffusion coefficient @xmath20 in the following form @xmath21 , where @xmath22 is the proton energy normalized to 1 tev , and @xmath23 is the magnetic field in units of mg .
generally the parameter @xmath24 is a function of energy . in the most effective confinement regime corresponding to the bohm diffusion , @xmath25 .
thus , in a source of a linear size @xmath26 , the production efficiency of tev neutrinos , assuming that a hard energy spectrum of protons extends effectively to 100 tev , is @xmath27 the maximum possible efficiency of a tev neutrino source , @xmath28 can in principle be achieved if @xmath29 .
such a condition can be best fulfilled in objects like giant molecular clouds , with a size @xmath30 pc , mass @xmath31 , and magnetic field @xmath32 mg , provided that the propagation of multi - tev protons proceeds close the bohm diffusion regime . however , bohm diffusion hardly can be realized in molecular clouds , thus @xmath33 , typically @xmath34 , which implies that the acceleration power should exceed @xmath35 .
this significantly reduces the number of potentially detectable neutrino sources to the most powerful representatives of nonthermal source populations in our galaxy .
in addition to young snrs and pwne , possible emitters of tev neutrinos are compact binary systems in which the compact object ( a black hole or a pulsar ) plays the role of particle accelerator , and the dense gas regions , e.g. the atmosphere of the companion star @xcite or the accretion plasma around the compact object @xcite , play the role of the target . moreover , in binary systems containing a luminous optical star and a compact object
, the photomeson interactions could provide an additional channel for neutrino production , provided that protons are accelerated to energies exceeding the interaction threshold , @xmath36 tev in the case of interactions with the starlight , and three orders of magnitute less for interactions with photons of the accretion disk . for a photon field with a thermal ( planckian ) distribution , the interaction and escape times of protons are @xmath37 , and @xmath38 , respectively .
it is remarkable that both timescales are proportional , although for completely different reasons ( ! ) , to the product @xmath39 , therefore the neutrino production efficiency appears independent , for a given luminosity of thermal radiation , of both the source size and the temperature of radiation , but strongly depends on the magnetic field @xmath40 thus we arrive at the conclusion that the neutrino production via photomeson interactions can be effective only in compact objects with strong magnetic fields ( @xmath41 kg ) and high turbulence ( for confinement of protons with @xmath42 ) .
the immediate proximity of the luminous star ( i.e. its photosphere ) or the compact object , e.g the accretion disk or the base of the jet @xcite , can be sites where the neutrino production proceeds with a reasonably high efficiency .
the _ high efficiency _ of neutrino production is a key condition for the detectability of potential vhe neutrino sources given the _ limited budget _ of available energy and the _ limitted sensitivity _ of detectors .
the above qualitative estimates show that this condition can be achieved only with certain combinations of a few key model parameters .
therefore , a careful inspection of these conditions before any detailed numerical calculations is highly advisable .
independent constraints on the detectability of a vhe neutrino source can be obtained also from analysis of multi - wavelength observations , in particular in the most relevant vhe @xmath1-ray band .
indeed , tev @xmath1-ray fluxes can be safely used as upper limits for neutrino fluxes , provided , of course , that the internal and external absorption of @xmath1-rays is negligible . generally ,
the production of vhe neutrinos is accompanied by production of @xmath1-rays , but not _ vice versa _ ; @xmath1-rays are copiously produced also in electromagnetic interaction both by electrons ( through bremsstrahlung , inverse compton scattering ) and protons ( e.g. through the synchrotron and curvature radiations ) . moreover , since the main channels of the neutrino and `` hadronic '' gamma - ray production are decays of charged @xmath43 and neutral @xmath44 mesons , with decay time of charged pions significantly longer than the decay time of neutral pions , at certain conditions the production of neutrinos can be suppressed compared to @xmath1-ray production . indeed , in the lab frame , the decay time of charged pions responsible for tev neutrino production is @xmath45 s. on the other hand , the cooling time of pions due to inelastic @xmath46 and @xmath47 interactions depends on densities of the ambient gas @xmath48 and x - ray photons @xmath49 : @xmath50 s , and @xmath51 s , respectively .
thus , charged pions would decay to @xmath52 and @xmath53 before interacting with the ambient photons and protons if @xmath54 and @xmath55 .
finally , the production of neutrinos from the subsequent muon decay would proceed with high probability as long as the magnetic field does not exceed @xmath56 g. this follows directly from the comparison of the decay time of muons , @xmath57 s , with their synchrotron cooling time , @xmath58 s. thus , the ratio of @xmath1-ray and neutrino production rates in a vhe source is expected to be of order of 1 or more , @xmath59 .
the absorption of @xmath1-rays can , of course , significantly change the initial @xmath60 ratio . while for galactic sources
the external @xmath1-ray absorption ( due to interactions with the interstellar ir photon fields ) is not dramatic up to several tens of tev @xcite , in compact galactic objects like x - ray binaries the internal absorption can be huge .
however , the absorption of gamma - rays does not imply that the information is lost ; the secondary electrons initiate a cascade in the same radiation field and/or cool via synchrotron radiation . in the first case ( which happens when the radiation density @xmath61 exceeds the energy density of the magnetic field @xmath62 )
, the initial @xmath1-rays are gradually reprocessed , down to energies at which the source becomes transparent . in binary systems containing luminous optical stars ( like ls 5039 ) the main energy is released at gev energies @xcite , while in the case of particle acceleration in the accretion disk around a black hole ( like cygnus x-1 ) , the initial energy is released mainly at mev energies @xcite .
the cascade development in the photon field can be suppressed at the presence of a strong magnetic field , @xmath63 ( typically 10 g or more ) .
interestingly , in this case the synchrotron radiation of pair - produced electrons extends well into the mev gamma - ray energy band , and therefore its detection is not prevented ( `` screened '' ) by the optical and x - ray backgrounds related to the optical star and the compact object .
thus , independent of the cooling regime ( via synchrotron radiation and/or the klein - nishina cascades ) , the pair - produced electrons leave a detectable imprint in the form of mev to gev gamma - ray photons .
so far the mev and gev @xmath1-ray sky has been explored at the depth corresponding to the energy flux @xmath64 .
therefore the data available in the mev - gev band from binary systems provide important , although not very restrictive upper limits on the energy flux of vhe neutrinos from these objects . with the arrival of agile and especially glast , the mev / gev @xmath1-ray observations will play a more decisive role in predictions of vhe neutrinos from compact objects .
the multiwavelength approach to the estimates of vhe neutrino fluxes expected from potential cosmic accelerators indicate that the fluxes of the best - candidate _
persistent _ galactic neutrino sources can not significantly exceed @xmath65 , i.e. most likely these sources are expected to be revealed at the level of statistically marginal signals . a possible exception
could be `` hidden '' sources - proton accelerators completely shielded from us by a very thick ( @xmath66 ) gas material in which the energy of ultrarelativistic protons is converted with 100 % efficiency to secondaries . at the same time ,
in such a thick shell of gas the high energy electromagnetic radiation would completely dissipate , and thus the source would becomes invisible because of much stronger background and forground thermal emission components . in such objects ,
the ( highly unknown ) total acceleration power of the source is the only model parameter that determines the neutrino flux ; the source would be detectable by vhe neutrino detectors if the power of the `` hidden '' pevatron exceeds , approximately by a factor of 10 , the estimate given by eq .
there is another ( more sophisticated ) scenario of realization of `` orphan '' neutrino sources , i.e. objects with neutrino fluxes not accompanied by electromagnetic radiation .
such a possibility is related to the features of acceleration and radiation of particles in optically thick ( with respect to the photon - photon pair production ) relativistic flows which can be formed by hot plasma left behind a relativistic shock or exist in the form of jets .
generally , it is silently assumed that nonthermal particles and hence their radiation are isotropically distributed in the comoving frame .
however , this assumption can easily be violated in relativistic shocks and jets with a strong impact on the emission properties , especially at very high energies .
namely , the beam pattern of relativistic jets with a bulk motion lorentz factor @xmath67 in this energy domain appears much broader than the inverse lorentz factor , @xmath68 .
this results in an _ off - axis _ high energy emission @xcite which is expected to be much brighter compared to the predictions derived from the standard doppler boosting considerations applied to an isotropic ( in the frame of the jet ) source . however , in optically thick sources the electromagnetic radiation from super - critical particles is reprocessed through the electron - photon cascades , thus becomes isotropic in the jet frame , and , therefore , strongly collimated in the lab - frame .
consequently , the reprocessed electromagnetic radiation can not be observable at large viewing angles .
this effect does not concern the high energy neutrinos , therefore the jet , when viewed off - axis , may appear as an _ over - bright _ neutrino source with an arbitrarily large ratio of the neutrino luminosity to the total electromagnetic luminosity .
it should be noted that acceleration of protons in relativistic shocks and shear flows can be significantly enhanced when it proceeds through the so - called _ converter _ mechanism @xcite .
this acceleration mechanism , which utilizes multiple conversions of protons to neutrons through photomeson reactions , has certain advantages compared to the standard diffusive shock acceleration scenario .
it essentially diminishes particle losses downstream and provides penetration of particles deep into the upstream region allowing a highly desirable energy boost by a factor of @xmath69 at _ each _ shock encounter @xcite .
since the copious neutrino production is an intrinsic feature of this scenario , the realization of the convertor acceleration mechanism in relativistic flows with large aspect angles would naturally lead to the appearance of an `` orphan '' vhe neutrino source .
a possible hadronic origin of gamma - radiation of some of the tev @xmath1-ray sources discovered by hess in the galactic plane @xcite makes them also potential emitters of high energy neutrinos ( see e.g. ref.@xcite ) .
recently , vissani @xcite kistler and beacom @xcite and kappes et al .
@xcite performed detailed calculations of the neutrino signal and background rates for the future 1@xmath2-volume scale neutrino telescopes based on the energy spectra and source - morphologies of galactic tev @xmath1-ray sources reported by hess .
the potential of the km@xmath4-volume class detectors is limitted , as mentioned above , by the detection of @xmath5 tev neutrino fluxes at the level of @xmath70 confined within an angle @xmath7 . for power - law energy distributions of parent protons with power - law indices
@xmath71 , the related @xmath1-ray flux in the same energy interval is slightly , by a factor of 1 to 2 , higher @xcite , i.e. quite close to the gamma - ray flux of the crab nebula - the standard candle of tev @xmath1-ray astronomy .
thus , the accompanied @xmath1-ray flux in units of `` 1 crab '' can be treated as the detection threshold of the _ galactic _ neutrino astronomy with km@xmath4-volume class instruments .
presently , in addition to the crab nebula itself , three more tev @xmath1-ray sources are detected at the `` 1 crab '' level - two young shell - type snrs rxj 1713.7 - 3946 @xcite and rxj
0852.0 - 4622 @xcite ( vela jr ) , as well as a nearby pwn - vela x @xcite .
young supernova remnants have been predicted , within a hadronic model , as extended tev @xmath1-ray and neutrino sources with shell type morphology and hard energy spectra extending to 100 tev @xcite .
the morphological and spectroscopic characteristics of young snrs rxj 1713.7 - 3946 and rxj 0852.0 - 4622 reported by the hess collaboration perfectly agree with these predictions , but ironically can not yet be considered as an ultimate proof the hadronic model .
the `` trouble makers '' are the alternative leptonic models which relate the tev @xmath1-rays to inverse compton scattering of electrons responsible also for the nonthermal x - ray emission .
although these models do not provide satisfactory explanations of the energy spectra of @xmath1-rays , and require very low magnetic field of order of 10 @xmath52 g @xcite , they unfortunately can not be safely excluded .
this prevent us from making a robust statement of detection of snrs as sources of cosmic ray protons and nuclei .
although it is believed that detailed theoretical studies of snrs in the context of their multiwavelength properties should allow us to arrive at certain conclusions concerning the origin of tev @xmath1-ray emission , formally only the detection of tev neutrinos from these objects can be considered as straightforward and unambiguous proof of acceleration of protons and nuclei .
the predicted detection rates of @xmath5 tev neutrinos ( of order of a few events per one year ) from the brightest @xmath1-ray snrs raj1713.7 - 3946 and raj0852.0 - 4622 by a km@xmath4-volume detector in the mediterranean sea , make the prospects of discovery of tev neutrinos from these snrs rather realistic .
unfortunately the locations of these two snrs are not favorable for icecube .
the existence of snrs of similar brightness in tev @xmath1-rays located in the northern hemisphere will be explored , hopefully soon , by the veritas and magic telescope systems .
the search for tev @xmath1-ray sources in the cygnus region - one of the most prominent and promising parts of the galactic plane - is of special interest .
the recent milagro observations of this region revealed a diffuse @xmath1-ray component with several hot spots @xcite , the strongest of which , mgro j2019 + 37 could be a neutrino source with a flux close to the detection threshold of icecube @xcite .
the close associations of some of the extended tev galactic sources discovered by hess with several well established synchrotron x - ray nebulae ( msh 15 - 52 , psr j1826 - 1334 , vela x , _ etc .
_ ) confirm the early theoretical predictions @xcite on tev gamma - ray visibility of young pwne with spin - down `` flux '' ( @xmath72 ) exceeding @xmath73 .
the broad - band spectral energy distributions of these sources are readily explained by the standard pwn model which assumes acceleration of ultrarelativistic electrons by the pulsar wind termination shock . yet , in some of these systems particle acceleration could be driven by ions present in the relativistic pulsar wind @xcite .
these ions are expected to produce @xmath1-rays and neutrinos via inelastic interactons with the ambient medium @xcite . in this regard , the extended tev source associated with the pulsar psr b0833 - 45 ( vela x ) is a possible candidate for such a `` hadronic plerion '' .
indeed , although the observed @xmath1-ray emission can be interpreted as inverse compton emission of nonthermal electrons @xcite , one needs to make some strong ( non - trivial ) assumptions in order to explain the rather unusual spectrum of this source with photon index @xmath74 and exponential cutoff around 14 tev .
the steady - state electron distribution constrained by @xmath1-ray data requires an @xmath75 type power - law spectrum with a sharp cutoff around 70 tev .
such a spectrum of electrons can be interpreted only in terms of negligible synchrotron cooling , which would be possible only in the case of unusually low nebular magnetic field ( a few @xmath52 g or less ) .
moreover , the total energy in relativistic electrons and in the magnetic field , which is required to match the observed x - ray and gamma - ray fluxes , is only a negligible fraction ( @xmath76 ) of the pulsar spin - down energy released over the pulsar s life - time @xmath77 yr .
this begs the question as to where the remaining energy has gone ?
interestingly if we assume that a large fraction of the spin - down luminosity of the pulsar is carried out by relativistic protons and nuclei , one can satisfactorily explain both the absolute flux and the spectrum of tev @xmath1-rays of this unusual source @xcite . remarkably , the tev neutrino flux expected within this scenario should be detectable by km3net @xcite .
this makes the vela x as one of the best - bet candidates to be the first _ detected _ astronomical tev neutrino source . for the icecube detector
an obvious target representing this source population is the crab nebula .
the recent detections of tev @xmath1-rays from two binary systems tentatively called microquasars
ls 5039 by hess @xcite and lsi 61 303 by magic @xcite are amongst the most exciting discoveries of observational gamma - ray astronomy in the very high energy regime .
this result clearly demonstrates that the galactic binary systems containing a luminous optical star and a compact object ( a black hole or a pulsar / neutron star ) , are sites of effective acceleration of particles ( electrons and/or protons ) to multi - tev energies . as usual , whether the @xmath1-rays are of hadronic or leptonic origin is a key question which however does not have a straightforward answer ( see e.g. ref .
the critical analysis of conditions of particle acceleration and radiation in these sources , based on the temporal and spectral behavior of tev @xmath1-ray emission , in particular on the modulation of the tev flux of ls 5039 with a period of 3.9 day , and the extension of its energy spectrum to 10 tev and beyond , reduces the possible interpretations to a few options .
one of them gives a preference to the hadronic origin of tev photons , especially if they are produced within the binary system @xcite .
if so , the detected @xmath1-rays should be accompanied by a flux of high energy neutrinos emerging from the decays of @xmath43 mesons produced by _
p - p _ and/or @xmath78 interactions .
the neutrino fluxes , which can be estimated on the basis of the detected tev @xmath1-ray fluxes , taking into account the severe internal @xmath79 absorption , depend significantly on the location of @xmath1-ray production region(s ) @xcite . the minimum neutrino flux above 1 tev is expected to be at the level of @xmath80
; however , it could be much higher - by a factor of 10 , or even more .
the detectability of the tev neutrino signals significantly depends on the high energy cutoff in the spectrum of parent protons .
if the spectrum of accelerated protons continues to 100 tev and beyond , the predicted neutrino fluxes of ls 5039 and lsi 61 303 can be probed by km3net @xcite and icecube @xcite high energy neutrino detectors .
9 halzen f 2007 _ science _ * 315 * 66 katz u f 2006 _ nimpa _ * 567 * 457 lipari p 2006 _ nimpa _ * 567 * 405 waxman e 2007 _ science _ * 315 * 63 rachen j p and mszros p 1998 _ phys . rev .
d _ * 58 * i d .
123005 mszros 2006 _ rep .
phys . _ * 69 * 2259 dermer c d and atoyan a 2006 _ new journal of physics _ * 8 * 122 hess collaboration 2006 _ astrophys j _ * 636
* 777 atkins r et al .
( milagro collaboration ) 2005 _ phys .
lett . _ * 95 * 251103 mckay t a et al .
( casa - mia collaboration ) 1993 _ astrophys .
j _ * 417 * 742 vissani f 2006 _ astropart . physics _
* 26 * 310 kistler m and beacom j f _ phys
d _ * 74 * i d .
063007 distefano c 2007 _
astrophys space sci _ , in press , astro - ph/0608514 kappes a , hinton j , stegmann ch and aharonian f a 2007 _ astrophys . j. _ * 656 * 870 malkov m a and drury l 2001 _ report on progress in physics _ * 64 * 429 arons j 1995 _ space sci .
rev . _ * 75 * 235 berezinsky v s 1977 , in proc .
_ `` ultra high energy neutrinos and detection possibilities by dumand '' _ ( fermilab , batavia ) 229 vestrand w t and eichler d 1982 _ astrophys .
j. _ * 261 * 251 hillas a m 1984 _ nature _ * 312 * 50 gaisser t k and stanev t 1985 _ phys .
* 54 * 2265 anchordoqui l a , torres d f , mccauley t p , romero g e and aharonian f a 2003 _ astrophys . j. _ * 589 * 481 levinson a and waxman e 2001 _ phys . rev .
lett . _ * 87 * i d .
171101 distefano c , guetta d , waxman e and levinson a 2002 _ astrophys .
j. _ * 575 * 378 moskalenko i v , porter t a and strong a w 2006 _ astrophys .
j. _ * 640 * l155 aharonian f , anchordoqui l , khangulyan d and montaruli t 2006__j of physics : conf . ser . _ _ * 39 * 408 bednarek w 2006 _ mnras _ * 368 * 579 aharonian f a and vardanian v v 1985 _ astrophys .
space sci . _ * 115 * 31 derishev e v , aharonian f a and kocharovsky , vl v 2007 _ astrophys .
j. _ * 655 * 980 derishev e v , aharonian f a , kocharovsky , v v and kocharovsky , vl v 2003 _ phys .
d _ * 68 * id.043003 dermer c d 2007 _ proc . of tev - particle astrophysics ii _
, in press aharonian f a 2007 _ science _ * 315 * 70 kelner s r , aharonian f a and bugayov v v 2006 _ phys .
d _ * 74 * id.034018 hess collaboration 2004 _ nature _ * 432 * 75 ; 2006 _ astron .
astrophys . _ * 449 * 223 hess collaboration 2005 _ astron .
astrophys . _ * 437 * l7 hess collaboration 2006 _ astron .
astrophys . _
* 448 * l43 drury l , aharonian f a and vlk h j 1994 astron .
astrophys * 287 * 959 abdo a a et al ( milagro collaboration ) 2006 , _ astrophys .
, submitted , astro - ph/0611691 beacom j f , kistler m d 2007 _ phys rev .
( submitted ) , astro - ph/0701751 aharonian f a , atoyan a m and kifune t 1997 _ mnras _ * 291 * 162 atoyan a m and aharonian f a 1996 _ mnras _ * 278 * 525 bednarek w and protheroe r j 1997 _ phys .
lett . _ * 79 * 2616 bednarek w 2003 _ astron
. astrophys _
* 407 * 1 amato e , guetta d and blasi p 2003 _ astron .
astrophys . _ * 402 * 827 horns d , aharonian f , santangelo a , hoffmann a and masterson c 2006 _ astron .
astrophys _ * 451 * l51 hess collaboration 2005 _ science _ * 309 * 746 ; 2006 _ astron . astrophys _ * 460 * 743 albert j et al ( magic collaboration ) 2006 _ science _ * 312 * 1771 paredes j m 2007 , in `` frontier objects in astrophysics and particle physics '' ; astro - ph/0609168 dubus g 2006 _ astron .
astrophys . _ * 451 * 9 boettcher m 2007 _ astropart .
_ , in press : astro - ph/0609136 christiansen h r , orellana m and romero g e 2006 _ phys .
d _ * 73 * 063012 torres d f and halzen francis 2006 astro - ph/0607368 | the undisputed galactic origin of cosmic rays at energies below the so - called knee around @xmath0 implies an existence of a nonthemal population of galactic objects which effectively accelerate protons and nuclei to tev - pev energies .
the distinct signatures of these cosmic accelerators are high energy neutrinos and @xmath1-rays produced through hadronic interactions . while @xmath1-rays can be produced also by directly accelerated electrons , high energy neutrinos provide the most straightforward and unambiguous information about the nucleonic component of accelerated particles . the planned @xmath2-volume class high energy neutrino detectors are expected to be sensitive enough to provide the first astrophysically meaningful probes of potential vhe neutrino sources .
this optimistic prediction is based on the recent discovery of high energy @xmath1-ray sources with hard energy spectra extending to 10 tev and beyond . amongst
the best - bet candidates are two young shell - type supernova remnants
rxj 1713.7 - 4946 and rxj 0852.0 - 4622 , and perhaps also two prominent plerions - the crab nebula and vela x. because of strong absorption of tev @xmath1-rays , one may expect detectable neutrino fluxes also from ( somewhat fainter ) compact tev @xmath1-ray emitters like the binary systems ls 5039 and ls i+61 303 , and , hopefully , from hypothetical `` hidden '' or `` orphan '' neutrino sources . |
the strong sensitivity of certain quantum states to small variations of external parameters opens up great opportunities for devising high - precision measurements , e.g. of length and time , with unprecedented accuracy .
a particularly important physical measurement technique is interferometry .
its numerous variations include ramsey spectroscopy in atomic physics , optical interferometry in gravitational wave detectors , laser gyroscopes and optical imaging to name but a few .
understanding limits on its performance in realistic situations under given resources is therefore of fundamental importance to metrology . in this paper
we examine the fundamental limits of the precision of optical interferometry in the presence of photon losses for quantum states of light with definite photon number .
optical interferometry aims to estimate the relative phase of two modes , or two `` arms '' , of the interferometer .
this estimation process requires a certain amount of resources which is typically identified to be the number of photons , @xmath0 , used for the measurement . the best precision which can be obtained using classical states of light scales like @xmath1 , the so - called standard quantum limit ( sql ) . using non - classical states of light this precision
can be greatly improved , ideally leading to heisenberg - limited scaling , @xmath2 @xcite .
indeed , recent years have seen many experimental proof - of - principle demonstrations of beating the sql using quantum strategies in various interferometric setups @xcite . unfortunately , highly non - classical states of light which potentially lead to heisenberg - limited sensitivity are very fragile with respect to unwanted but unavoidable noise in experiments . in quantum - enhanced optical interferometry ,
the loss of photons is the most common and potentially the most devastating type of noise that one encounters .
in particular it was noted that highly entangled quantum states , optimal for interferometry in the lossless case
n00n states @xcite are extremely fragile . even for moderate losses
they are outperformed by purely classical states @xcite .
a different approach has been taken in @xcite , where the noise arising from imperfect preparation of a state has been investigated . in @xcite ,
the first systematic approach was taken in order to determine the structure of optical states optimal for interferometry in the presence of losses .
the best possible precision using input states with definite photon number was given .
in this paper we elaborate and extend the ideas presented in @xcite .
our treatment is based on general quantum measurement theory @xcite .
the quantity of interest is the lowest possible uncertainty attainable in parameter estimation , inversely proportional to the square root of the quantum fisher information @xcite .
the quantum fisher information depends only on the state of the system and not on the measurement procedure .
we show that the optimization of the quantum fisher information can be done effectively over the class of input states which have a definite photon number .
since these input states are subject to unavoidable photon losses they will degrade into mixed states and their suitability for phase estimation is compromised .
our optimization takes this into account yielding the most suitable input states in the presence of photon losses leading to the highest possible quantum fisher information , and hence to the best possible precision .
we give a detailed description of the noise model and calculate an analytic expression for the quantum fisher information .
the latter is shown to be a concave function on a convex set and therefore suited for efficient convex optimization methods .
we numerically determine the optimal input states , compare them to alternative quantum and classical strategies , and show that they can beat the sql .
we note that the corresponding precision , which lies between the sql and the heisenberg limit ( depending on the loss rates ) , defines the best possible precision for optical two - mode interferometry .
in addition to this we discuss a measurement procedure that allows one to achieve the optimal precision , in terms of a positive operator - valued measure ( povm ) , and the possibility of using states with indefinite photon number or distinguishable photons .
we show that neither of these generalizations improves the estimation precision , and consequently the state with definite photon number and indistinguishable photons are optimal .
the paper is organized as follows . in sec .
[ sec : phasestimation ] a general scheme of quantum phase estimation is presented , and the notion of optimality is defined . in sec .
[ sec : fisher ] we introduce the quantum fisher information and discuss its most important properties . in sec .
[ sec : interferometry ] we derive an explicit formula for the quantum fisher information and prove that it is a concave function of input state parameters . in sec .
[ sec : optimalstates ] we discuss the structure of the optimal states for interferometry and compare them to alternative strategies and states . in sec .
[ sec : measurement ] we discuss a measurement with which it is possible to achieve optimal precision . finally , in sec .
[ sec : generalstates ] we discuss possible generalizations of the considered quantum states , particularly states with indefinite photon number and the case when photons are distinguishable .
acquires a phase @xmath3 relative to channel @xmath4 .
measurements are performed on the output state yielding an estimated value , @xmath5 , of the phase @xmath3 .
the beam splitters symbolize photon losses.,width=302 ] we consider a general interferometer with two arms as shown in fig .
a pure input state @xmath6 is fed into the interferometer and acquires a phase @xmath3 in the channel @xmath7 relative to the channel @xmath4 .
both channels , or `` arms '' , of the interferometer are subject to photon losses which can be modelled by fictitious beam splitters inserted at arbitrary locations in both channels .
the output of the interferometer therefore needs to be described in general by a mixed state @xmath8 . a measurement , represented by a positive operator valued measure ( povm ) @xmath9 which defines a probability distribution for the measurement outcomes , @xmath10,\ ] ] is subsequently performed on the output state @xmath8 .
an estimated value of the true phase @xmath3 is obtained by applying an estimator that assigns to a particular measurement result @xmath11 an estimated value @xmath12 .
we aim to estimate the phase @xmath3 as precisely as possible .
two more elements need to specified in order to make the problem of finding the optimal phase estimation strategy well defined : an a priori knowledge on the phase distribution @xmath13 and a cost function @xmath14 which can be seen as a measure for the uncertainty in the estimated phase . for phase estimation ,
the optimal choice of the input state , measurement and estimator is the one that minimizes the average cost function @xmath15 at this point two different approaches are most often pursued . in the _ global approach _
one assumes initial ignorance about the actual value of @xmath3 , which corresponds to the choice @xmath16 .
the solution of the problem then yields an estimation strategy which performs equally well irrespectively of the actual value of the estimated phase . in the _ local approach _
, on the other hand , the assumption is that the value of the actual phase lies in the vicinity of a known phase @xmath17 .
more precisely , the a priori probability is chosen to be @xmath18 , while the estimator is required to be _ _ locally unbiased _ _ @xmath19 the above condition is equivalent to a statement that the estimator will on average yield the true value of @xmath3 up to the first order in @xmath20 .
notice that without local unbiasedness the estimation problem would be trivial ( and also useless ) since in order to minimize @xmath21 in eq .
( [ eq : cost ] ) , with @xmath22 , the optimal choice for the estimator would be simply @xmath23 , and the choice of the measurement would be irrelevant .
the local approach is useful when we are interested in small deviations of the phase from a known one .
a significant advantage of the local approach over the global one is that for a natural choice of a quadratic cost function , there exist explicit lower bounds on @xmath21 based on the fisher information ( see sec .
[ sec : fisher ] ) . in many practical situations
these bounds are tight and the optimization over the measurement and the estimator can be avoided
. moreover , the local approach may also be useful in situations when there is no a priori knowledge on the phase @xmath3 .
if many copies of a state are given , one can first perform a rough measurement ( even not optimal ) on a small fraction of copies in order to narrow down the range of potential values of @xmath3 so that they lie in a vicinity of a known phase @xmath17 , and then perform the optimal estimation using the local approach .
this strategy will yield a high accuracy estimation for the global approach , without the need of optimizing the measurement and the estimator , since the rough measurement performed on small fraction of copies , even if not optimal , will not significantly influence the final accuracy @xcite .
in what follows we take the local approach . since in this case
we deal with small deviations of estimated phases from the true one , it is natural to choose @xmath24 as the cost function .
our goal is to find the optimal state , measurement and locally unbiased estimator minimizing the expression @xmath25
^ 2.\ ] ] to minimize the standard deviation @xmath26 given by the above formula we use an upper bound on @xmath26 based on the fisher information @xcite . for a given povm @xmath27 and state @xmath8 defining the probabilities @xmath28 ,
the cramr - rao inequality bounds the variance that can be obtained using any locally unbiased estimator , @xmath29 where the fisher information @xmath30 is given by @xmath31 if an experiment is repeated @xmath32 times the bound reads @xmath33 for large @xmath32 the cramr - rao bound is asymptotically achieved by the maximum likelihood estimator @xcite .
optimization over the measurements yields the quantum cramr - rao bound @xcite @xmath34 where the quantum fisher information @xmath35 is given by @xmath36.\ ] ] the hermitian operator @xmath37 is called the `` symmetric logarithmic derivative '' ( sld ) and is implicitly defined via the relation @xmath38.\ ] ] in the eigenbasis of @xmath8 , @xmath37 is given by @xmath39_{ij},\ ] ] where @xmath40 and the @xmath41 are the eigenvalues of @xmath8 ( whenever @xmath42 we set @xmath43 ) .
it has been shown that a measurement saturating the quantum cramr - rao bound exists and is given by a projective measurement on the eigenbasis of @xmath37 @xcite .
for the sake of completeness we state some important properties of @xmath35 ( see e.g. @xcite ) : \(i ) let @xmath8 , @xmath44 be two density matrices supported on orthogonal subspaces , @xmath45 , which do not cease to be orthogonal for an infinitesimal change of @xmath3 , i.e. @xmath46 , then @xmath35 is linear on the direct sum @xmath47 \nonumber\\ & = pf_q[\rho(\varphi ) ] + ( 1-p)f_q[\sigma(\varphi)].\end{aligned}\ ] ] ( ii ) @xmath35 is convex @xmath48 \nonumber\\ & \leq pf_q[\rho(\varphi ) ] + ( 1-p)f_q[\sigma(\varphi)]\end{aligned}\ ] ] ( iii ) for pure states @xmath49 , @xmath35 reads @xmath50,\ ] ] where @xmath51 .
property ( i ) is due to the fact that the sld for @xmath52 is a direct product of slds for @xmath53 and @xmath54 , respectively .
property ( ii ) is a consequence of the fact that by ( i ) the right hand side of ( ii ) can be viewed as the quantum fisher information of the state @xmath55 , where @xmath56 , @xmath57 are orthogonal ancillary states , while the left hand side is @xmath35 of the state after tracing out the ancillary system .
furthermore , @xmath35 is non - increasing under stochastic operations @xcite ( tracing out the ancilla is an example ) . property ( iii ) is a consequence of eqs .
( [ eq : qcr],[eq : sld ] ) , since for a pure state @xmath58 the measurement saturating the quantum cramr - rao bound in this case is a von neumann measurement projecting on any orthonormal basis containing two vectors @xmath59 where @xmath60 is the normalized vector orthogonal to @xmath61 lying in the space spanned by @xmath61 and @xmath62 .
assuming that we have @xmath0 photons at our disposal , we aim to find the input state that allows performing phase estimation with the best precision possible , i.e. yielding the highest value of the quantum fisher information @xmath35 .
in particular we consider the most general pure two - mode input state with definite photon number @xmath0 , @xmath63 where @xmath64 abbreviates the fock state @xmath65 .
this class of states includes the n00n state which , in the absence of losses , leads to heisenberg limited precision , but is very fragile in the presence of noise .
we are therefore looking for states which lead possibly to a lower precision than the n00n state , but which are more robust with respect to photon losses .
moreover , although states of the form ( [ eq : input ] ) seem to be a restriction , we show in sec .
[ sec : generalstates ] that our treatment effectively includes states with indefinite photon number . in the following subsections
we show how states of the form ( [ eq : input ] ) are influenced by photon losses , calculate its quantum fisher information and show that the latter can be maximized ( thus minimizing @xmath26 ) by means of convex optimization methods .
losses are modeled by fictitious beam splitters of transmissivity @xmath66 , @xmath67 in channels @xmath7 and @xmath4 respectively , and cause a fock state @xmath64 to evolve into @xmath68 where @xmath69 represents the state of two ancillary modes carrying @xmath70 and @xmath71 photons lost from modes @xmath7 and @xmath4 respectively , while @xmath72 including the phase accumulation @xmath73 and tracing out the ancillary modes results in the output density matrix @xmath74 where @xmath75 is the conditional pure state corresponding to the event when @xmath70 and @xmath71 photons are lost in modes @xmath7 and @xmath4 respectively , and @xmath76 is the normalization factor corresponding to the probability of that event .
equivalently , the loss process can be described by a master equation for two independently damped harmonic oscillators with loss rates @xmath77 , where @xmath78 is time , the solution of which is given by @xmath79 with kraus operators @xmath80 where @xmath81 is the annihilation operator for mode @xmath7 , and analogously for mode @xmath4 .
this state acquires a phase through the transformation @xmath82 .
notice that thanks to the relation @xmath83 we can commute the phase operator with the kraus operators since the phase terms @xmath84 cancels out .
it is therefore irrelevant if photons are lost before , during or after channel @xmath7 acquires its relative phase with respect to @xmath4 . using eq .
( [ eq : rhoout ] ) for the output state , one can calculate @xmath35 with the help of eqs .
( [ eq : qcr ] ) and ( [ eq : sld ] ) .
this requires diagonalization of @xmath8 which can be carried out in the case of one - arm losses . in the more general case of losses in both arms an analytic calculation of @xmath35
turns out to be infeasible . nevertheless , we are able to determine an upper bound to the quantum fisher information which , although not strictly tight for general input states , is very close to @xmath35 for the states we consider in sec . [
sec : optimalstates ] .
we consider first the case @xmath85 , @xmath86 , i.e. when losses are present in only one arm .
as can be seen from eq .
( [ eq : rhoout ] ) , in this case only states @xmath87 with @xmath88 contribute to @xmath8 . moreover , we have @xmath89 ,
hence we can write the output state as a direct sum @xmath90 making use of eqs .
( [ eq : lindir ] ) and ( [ eq : fishpure ] ) we get a formula for @xmath35 with explicit dependence on the input state parameters @xmath91 , @xmath92 where @xmath93 . in a more compact way the above formula can be rewritten as @xmath94 where @xmath95 is a vector containing variables @xmath96 , while the elements of the vector @xmath97 and the matrix @xmath98 are given by @xmath99 if losses are present in both arms , then , in the most general case , all @xmath87 contribute to @xmath8 .
states with different total number of lost photons , @xmath100 , are still orthogonal . using eq .
( [ eq : lindir ] ) we can therefore write @xmath101,\ ] ] where @xmath102 $ ] denotes the quantum fisher information of the state in brackets .
notice that states with the same @xmath103 are not necessarily orthogonal .
consequently , the calculation of the above expression requires solving an eigenvalue problem which is not feasible analytically . nevertheless , using the convexity of @xmath35 , eq .
( [ eq : convexity ] ) , we obtain a bound @xmath104,\ ] ] which can be calculated explicitly using eq .
( [ eq : fishpure ] ) .
it is given by @xmath105 or in a compact form , @xmath106 when losses are present only in one arm ( see previous paragraph ) @xmath107 .
when losses are present in both arms , however , the bound is not always tight .
the difference @xmath108 originates from the non - orthogonality of @xmath109 for a fixed @xmath110 , and physically corresponds to lack of knowledge about how many photons were lost from a particular mode . if this knowledge is not relevant then @xmath111 .
this happens , e.g. in the case of the n00n state @xmath112 , when loss of even a single photon renders the output states useless for phase estimation , hence the knowledge of which mode the photons were lost does not influence the value of @xmath35 . even with the explicit formulae for @xmath35 and @xmath113 given by eqs .
( [ eq : fisher1mode ] ) and ( [ eq : fisher2mode ] ) it is in general not possible to find an analytic solution for the optimal state , i.e. values @xmath114 , @xmath115 that maximize @xmath35 ( or @xmath116 ) .
however , we prove below that @xmath116 is a concave function of the @xmath117 .
consequently , the problem amounts to the maximization of a concave function @xmath116 on a convex set .
this allows for a feasible numerical constrained optimization using interior - point method routines ( e.g. implemented in _
mathematica _ 6.0 ) , and more importantly , any local maximum found is automatically the global maximum .
we prove concavity by showing that the hessian @xmath118 is negative semidefinite , i.e. for every vector @xmath119 we have @xmath120 .
using eq .
( [ eq : fisher2mode ] ) the hessian @xmath121 ( @xmath122 ) reads @xmath123 + \boldsymbol{x}^t \boldsymbol{b}^{(l_al_b ) } ( \mathbf{r}^{(l_al_b)})_{i , j } \boldsymbol{x}^t \boldsymbol{b}^{(l_al_b)}}{(\boldsymbol{x}^t \boldsymbol{b}^{(l_al_b)})^3}.\ ] ] since the denominator is always positive , it is sufficient to prove that for every @xmath70 , @xmath71 and for every vector @xmath119 we have @xmath124 introducing @xmath125 , @xmath126 the above condition can be written equivalently as @xmath127 since @xmath128 , it is sufficient to prove that @xmath129 is negative semi - definite on the set of vectors with coefficients summing up to @xmath130 . to this end
we define vectors @xmath131 ( @xmath132 ) , where @xmath133 , @xmath134 , @xmath135 ( for @xmath136 , @xmath137 ) , which span the space of all vectors with coefficients summing up to zero . writing @xmath138 and noticing that @xmath139 we arrive at @xmath140 which proves that @xmath116 is a concave function of the @xmath117 .
in this section we discuss optimization results based on eqs .
( [ eq : fisher1mode_a ] ) and ( [ eq : fisherbound ] ) derived in the previous sections .
the quantity we analyze is @xmath141 corresponding to the best possible precision for a fixed number of measurements @xmath32 ( cf .
( [ eq : qcr ] ) ) .
the only exception to this definition is in the case of the optimal state for losses in both arms where we set @xmath142 ( see sec . [ lossesintwoarms ] ) .
we compare the optimal precision to the precision which is obtainable using various alternative states and strategies .
in particular , we define the standard interferometric limit ( sil ) corresponding to the precision which can be achieved in a classical reference experiment .
this serves as a benchmark by which we can judge the advantage of using quantum states of light over classical ones . for a given photon number @xmath0
the sil is the precision of phase estimation using a mach - zehnder interferometer in which one arm gathers a phase @xmath3 , fed at one input port with a coherent state @xmath143 , where @xmath144 , and the vacuum at the other port .
the reflectivity of the two beam splitters can be adjusted to achieve the best precision , while the measurement consists of photon counting at the two output ports . without any additional
reference beams the input coherent state should be regarded as a state with unknown phase , and effectively described as a mixture of fock states with poissonian statistics ( see @xcite , and the discussion in sec .
[ sec : indefiniten ] ) . by eq .
( [ eq : lindir ] ) the quantum fisher information will be a weighted sum of quantum fisher information calculated for each input fock state after it passes through the first beam splitter .
taking the optimal value of transmissivity of the first beam splitter such that the ratio of the intensities in the arms @xmath7 and @xmath4 is @xmath145 the final uncertainty in phase estimation achievable with this classical strategy is given by @xmath146 note that the sil scales in the same way as the sql , particularly for equal losses in both arms we obtain @xmath147 , where @xmath148 . as a reference
let us also calculate the minimum uncertainty achievable using @xmath0 photons prepared in an unbalanced n00n state : @xmath149 the quantum cramr - rao bound yields : @xmath150 where the optimal amplitudes are given by @xmath151 and @xmath152 .
note that putting @xmath153 in eq .
( [ eq : noon_prec ] ) coincides with eq .
( [ eq : sil ] ) .
this implies that a coherent state performs equally well in phase estimation as the equivalent number of single photons sent one - by - one . in the following
we will concentrate on the two important scenarios of losses only in channel @xmath7 and equal losses in both channels .
the scenario of losses in only one arm of the interferometer , i.e. @xmath85 and @xmath86 , is relevant , for example , if losses are induced by a sample itself . figure [ fig : optstates10 ] shows the parameters @xmath96 of the optimal state for @xmath154 photons as a function of @xmath155 .
the corresponding precision is shown in fig .
[ fig : optfisher10 ] .
the shaded , grey area in this figure is bounded by the sil and the heisenberg limit @xmath2 , i.e. whenever a line is in this region there is an improvement over classical interferometry .
the precision obtainable with the n00n state is worse than the sil except for relatively low losses .
, for phase estimation with @xmath154 photons for the case of losses in one only arm , i.e. @xmath85 , @xmath86 . ]
photon states for losses in one arm , i.e. @xmath85 , @xmath86 .
red , dashed line : n00n state ; blue , solid line : optimal state ( cf .
[ fig : optstates10 ] ) ; bright green , dotted line : optimal two - component state ; dark green , dashed - dotted line : n00n chopping strategy .
the shaded area is bounded by the sil achievable with classical states and the heisenberg limit @xmath2 . ] for transmissivities exceeding a certain threshold @xmath156 the optimal state is equal to the n00n state ( [ eq : noon_state ] ) . note that for @xmath157 the n00n state is `` balanced '' , i.e. @xmath158 .
we will give an estimate for the threshold @xmath159 at the end of this subsection .
for @xmath160 the optimal state consists of more components . as seen in fig .
[ fig : optfisher10 ] over a wide range we can achieve almost the same precision by using two - component states of the form @xmath161 where we numerically optimize @xmath162 and the amplitude @xmath96 . with decreasing @xmath155 the optimal choice for @xmath162 increases .
this state is more robust in the presence of losses than a n00n state , since a loss of up to @xmath162 photons in the first mode does not destroy the superposition , while for the n00n state a loss of even a single photon renders the state useless for phase estimation . on the other hand , a large @xmath163 increases the sensitivity of the state with respect to an induced phase in channel @xmath7 .
the properties of large @xmath162 and large @xmath163 are therefore competing and the optimal result represents a trade - off between phase sensitivity and robustness . an alternative strategy that leads to an improvement over the sil , but uses states of a simpler structure than the optimal state is the n00n `` chopping '' strategy which was introduced in ref .
@xcite . instead of a single n00n state
, we use the same number of photons , but send them successively in @xmath164 smaller portions using @xmath165-photon n00n states . repeating an experiment @xmath164 times corresponds to an @xmath164 fold increase in fisher information , as shown in eq .
( [ eq : cramer_repeat ] ) . using eq .
( [ eq : noon_prec ] ) we find that the fisher information for the n00n chopping strategy reads @xmath166 treating both @xmath165 and @xmath164 as real numbers with @xmath167 $ ] , maximization of this expression over @xmath165 yields @xmath168 where the optimal choice of @xmath165 for the three regimes indicated above is @xmath169 , @xmath170 ( i.e. the solution of @xmath171 ) and @xmath172 .
we see that for very small transmissivities @xmath173 the precision is equal to the sil , specified in eq .
( [ eq : sil ] ) . for higher transmissivities the chopping strategy beats the sil , although only by a constant factor rather than in terms of scaling . for very high transmissivities
@xmath174 the best strategy is to use un - chopped n00n states .
we note that a similar strategy for multipartite qubit states has been devised in @xcite .
as mentioned above the n00n state ceases to be optimal below a threshold @xmath159 in which case the coefficient @xmath175 obtains a nonzero value . by adding an infinitesimal change @xmath176 to @xmath175 at the expense of @xmath177 ( which we assume are kept in the proportion which is optimal in the absence of @xmath175 )
, we can calculate the corresponding change in the quantum fisher information .
the change of @xmath175 by @xmath176 increases @xmath35 by @xmath178 , but at the same time due to decreasing weight of the other coefficients decreases it by @xmath179 . on the whole
the change of @xmath35 reads : @xmath180 substituting @xmath35 from eq .
( [ eq : fisher1mode ] ) at @xmath181 , @xmath152 , @xmath182 ( @xmath183 ) and calculating @xmath184 we determine the value @xmath159 below which @xmath185 is positive .
this implies that for @xmath160 an increase in @xmath175 will increase @xmath35 .
writing @xmath186 explicitly we get : @xmath187.\end{gathered}\ ] ] the roots of the expression in square brackets can be found numerically .
there is one real root in the interval @xmath188 $ ] which corresponds to the threshold @xmath159 .
the roots behave in very good approximation like @xmath189 where @xmath190 which is obtained by a fit to the roots between @xmath191 and @xmath192 . for example ,
for @xmath154 we find that the threshold at which the n00n state ceases to be optimal is at @xmath193 , which agrees with the numerical results obtained before ( see fig .
[ fig : optstates10 ] ) . in order to determine the optimal state in the case of equal losses in both arms of the interferometer , i.e. @xmath194 , we use @xmath116 as given by eq .
( [ eq : fisherbound ] ) as a basis for our numerical optimization .
it simplifies maximization drastically since it is a concave function of the @xmath117 . as pointed out in sec .
[ sec : calcfisherb ] , @xmath113 , is only an upper bound to the `` true '' quantum fisher information @xmath35 and therefore can not be reached by any measurement strategy if @xmath35 is strictly smaller than @xmath113 .
however , using @xmath116 can be still very useful : we maximize @xmath116 and use the resulting input state ( say , @xmath195 ) to calculate the corresponding @xmath196 .
on the one hand , the true maximum of @xmath197 ( corresponding to the truly optimal state , @xmath6 ) is certainly greater or equal to @xmath196 .
on the other hand we have @xmath198 due to convexity [ see eq .
( [ eq : boundconvexity ] ) ] .
but since @xmath199 is the global maximum we have @xmath200 .
consequently , the true maximum must lie _ in between _ @xmath196 and @xmath199 .
hence , if @xmath201 is small on a scale which is for this problem naturally given by the difference between the sil and the heisenberg limit , we obtain a very good approximation of the true maximal value of the quantum fisher information .
we showed numerically that for the examples discussed in this section this difference never exceeds @xmath202 .
, for phase estimation with @xmath154 photons for the case of equal losses in both arms , i.e. @xmath203 . due to this symmetry we have @xmath204 . ]
photon states for equal losses in both arms , i.e. @xmath203 .
red , dashed line : n00n state ; blue , solid line : optimal state ( cf .
[ fig : optstates2modes10 ] ) ; bright green , dotted line : the state optimal for lossless phase estimation in the global approach [ see eq .
( [ eq : stateoptglobal ] ) ] ; dark green , dashed - dotted line : the n00n chopping strategy .
the shaded area is bounded by the sil achievable with classical states and the heisenberg limit @xmath2 . ]
the result of optimization is shown in fig .
[ fig : optstates2modes10 ] for @xmath154 . due to the symmetry of the problem , @xmath205
, the solution has the property @xmath206 . for small losses
the optimal state is the balanced n00n state , i.e. @xmath158 .
however , below a threshold @xmath159 all other terms @xmath96 obtain non - vanishing values . with decreasing @xmath155 , @xmath207 and @xmath208
get smaller , and eventually the coefficients @xmath96 with @xmath209 are dominant .
interestingly , this structure is qualitatively similar to a two - mode generalization of the state which is optimal for one - mode lossless phase estimation in the global approach @xcite @xmath210{| k , n - k \rangle}.\ ] ] it was noted in ref .
@xcite that these states perform quite well in the presence of loss , and hence could be useful for the problem we consider .
the threshold @xmath159 can be obtained in a similar fashion as in the previous subsection . by adding an infinitesimal change @xmath211 to a component @xmath96 other than @xmath207 and @xmath208
we can derive a polynomial which has a root in @xmath188 $ ] defining @xmath212 .
since the polynomial is rather cumbersome we do not quote it here , however , the relevant root is again given in very good approximation by @xmath213 where @xmath214 is obtained by a fit to the roots between @xmath215 and @xmath192 .
for example for @xmath154 ( corresponding to figs .
[ fig : optstates2modes10 ] and [ fig : optfisher2modes10 ] ) we get @xmath216 .
analogous to the case of losses in only one arm we can examine a n00n chopping strategy by sending `` smaller '' n00n states through the interferometer .
the corresponding precision reads @xmath217 figure [ fig : optstates2modes10 ] shows the estimation precision for the optimal state ( blue line ; here we use @xmath218 ) , the n00n state ( red line ) , the n00n chopping strategy ( dark green line ) , and the state ( [ eq : stateoptglobal ] ) ( bright green line ) . in the latter case
the precision is indeed quite high in the regime of high losses , outperforming the n00n chopping strategy but for small losses it is significantly worse than both the optimal and the n00n chopping strategies . ) , outperforms even the optimal states we have presented , arise from taking a different perspective in quantifying the resources .
see sec .
[ sec : indefiniten ] for more discussion . ]
as discussed in sec . [ sec : fisher ] , a measurement saturating the quantum cramr - rao bound always exists . for the cases when
the states @xmath219 introduced in sec .
[ sec : interferometry ] are orthogonal for different @xmath110 ( i.e. when @xmath107 ) we can derive an explicit povm saturating the bound . to this end
we first assume that we perform a measurement determining the total number of photons lost , @xmath103 , which projects the system onto a particular pure state @xmath219 .
subsequently , we perform the measurement saturating the cramr - rao bound on pure states ( see sec . [
sec : fisher ] ) , i.e. a projection on a basis containing the vectors @xmath220 where @xmath221 , @xmath222 are the mean photon number and the variance of the photon number in mode @xmath7 for the state @xmath219 .
the above povm depends , in general , on the phase @xmath3 .
there are cases in which a single povm saturating the cramr - rao bound for all values of @xmath3 can be found . in a single photon case ( @xmath153 ) ,
states of the form @xmath223 , allow for a @xmath3-independent povm , whereas states with unbalanced weights of @xmath224 , and @xmath225 terms lead to optimal povms which are necessarily @xmath3-dependent .
moreover , for an @xmath0 photon state @xmath226 , a single , @xmath3-independent povm saturating the cramr - rao bound can be found provided that the state enjoys the path - symmetry property : @xmath227 , where @xmath228 is a fixed phase the same for all components @xcite . in general , due to losses the conditional pure states that appear in our problem lack the above mentioned symmetry and consequently the optimal povm varies with the change of @xmath3 .
one could argue that the @xmath0 photon state given by eq .
( [ eq : input ] ) does not represent the most general way of using @xmath0 photons to determine the relative phase in an interferometer : instead of considering a state with a _
definite _ photon number @xmath0 , one could consider a state which is a superposition of terms with different total photon number , and only imply a constraint that the _ mean _ photon number is @xmath0 .
furthermore , by using states of the form ( [ eq : input ] ) it is assumed that all photons in an arm of the interferometer are indistinguishable , i.e. occupy a common spatio - temporal mode .
this greatly limits the class of considered states since the hilbert space dimension of states sent into the interferometer would be of the order @xmath229 if the photons were distinguishable , while for indistinguishable photons it is @xmath230 . in the next two subsections
we show that none of these generalizations improves the phase estimation precision , and consequently it is sufficient to consider the states of the form given by eq .
( [ eq : input ] ) .
we emphasize that in this paper we consider closed systems , i.e. there are no additional reference beams , neither classical nor quantum , since any additional beam would contain photons and should therefore be explicitly taken into account for the determination of the required resources .
consider a superposition of input states of the interferometer with different , but definite photon numbers @xmath165 , @xmath231 where the mean total photon number is @xmath232 , i.e. terms with different @xmath165 evolve with different frequencies . in the absence of an additional reference beam which allows for clock synchronization between the sender and the receiver the relative phases between terms with different @xmath165 become unobservable and the state given by eq .
( [ eq : superpos ] ) is physically equivalent to a mixture @xcite , @xmath233 moreover , since the quantum fisher information is convex ( see sec .
[ sec : fisher ] ) , we have @xmath234 consequently it is always better to send a state with a fixed number of photons @xmath235 , with probability @xmath236 , rather than to use a superposition ( which is effectively a mixture ) .
the analysis can thus be restricted to states with definite photon number without compromising optimality . in this subsection
we consider general @xmath0 photon input states where the photons are distinguishable which is the case , e.g. , if they are sent in different time bins .
a state of this type can be written as @xmath237 where @xmath238 represents a binary sequence of length @xmath0 , and @xmath239 , @xmath240 .
summation from a sequence to a sequence should be understood using partial ordering between sequences , i.e. @xmath241 . therefore the summation is performed over all binary sequences .
furthermore @xmath242 denotes a state of @xmath0 photons , where @xmath243 means that the photon sent in the @xmath11-th time bin propagates in arm @xmath7 ( @xmath4 ) of the interferometer .
mathematically , we deal here with an arbitrary @xmath0-qubit state , which lives in a @xmath229 dimensional space . notice , that the indistinguishable case is recovered once we consider only the fully symmetric ( bosonic ) subspace of @xmath0 qubits .
@xmath116 can be calculated in a similar way as in sec .
[ sec : calcfisher ] and reads @xmath244 where boldface symbols denote binary strings of length @xmath0 , @xmath245 , @xmath246 denotes the number of @xmath247s in the sequence @xmath238 , and @xmath248 again @xmath249 provided that knowledge from which mode a photon was lost is not relevant or self - evident , e.g. when @xmath86 .
let us now symmetrize the state of @xmath0 photons .
this corresponds to replacing @xmath250 with @xmath251 , where the summation is performed over all permutations of an @xmath0 element set .
notice that the first term within the parenthesis in eq .
( [ eq : fishersm ] ) is not affected by symmetrization .
let us denote @xmath252 and @xmath253 .
then the subtracted term in eq .
( [ eq : fishersm ] ) reads @xmath254 .
performing symmetrization corresponds to replacing @xmath255 and @xmath256 with @xmath257 and @xmath258 respectively .
since @xmath256 are positive , the following inequality holds @xmath259 ^ 2}{1/2(b_{\boldsymbol{l}_a\boldsymbol{l}_b}+ b_{\boldsymbol{l}^\prime_a\boldsymbol{l}^\prime_b})},\ ] ] which shows that symmetrizing can only decrease the subtracted fraction , hence increase the fisher information .
this proves that the optimal states are symmetric states living in the bosonic subspace .
consequently , using indistinguishable photons is sufficient to obtain the optimal fisher information .
a physical intuition behind the above derivation is the following .
notice that if there are no losses , the optimal @xmath0 qubit state has the form @xmath260 it is the n00n state which lives in the fully symmetric subspace . in this case
there is no advantage in using distinguishable photons .
if there are losses , things get even worse for distinguishable photons .
there is still no advantage in terms of phase sensitivity , yet when a photon is lost , the _ knowledge of which photon was lost _ additionally harms the quantum superposition .
hence , it is optimal to use states from a fully symmetric subspace , where it is not possible to tell which photon was actually lost .
we have analyzed the optimal way of using @xmath0-photon states for phase interferometry in the presence of losses .
we have derived an explicit formula for the quantum fisher information in the case when losses are present in one arm of the interferometer , and have provided a useful bound for the case of losses in both arms . using the quantum fisher information as a figure of merit , we have found the optimal states , investigated their advantage over various quantum and classical strategies .
the optimal measurement saturating the cramr - rao bound has been presented , and it has been proven that in general it varies depending on the phase in the vicinity of which we perform estimation .
a close inspection of the properties of the quantum fisher information showed that it is optimal to use a quantum state with a definite number of indistinguishable photons .
neither allowing superpositions of different total photon - number terms , nor making photons distinguishable can improve estimation precision .
this research was supported by the epsrc ( uk ) through the qip irc ( gr / s82176/01 ) , the afosr through the eoard , the european commission under the integrated project qap ( contract no .
015848 ) , the royal society and the polish mnisw ( n n202 1489 33 ) . | we give a detailed discussion of optimal quantum states for optical two - mode interferometry in the presence of photon losses .
we derive analytical formulae for the precision of phase estimation obtainable using quantum states of light with a definite photon number and prove that maximization of the precision is a convex optimization problem . the corresponding optimal precision , i.e. the lowest possible uncertainty , is shown to beat the standard quantum limit thus outperforming classical interferometry .
furthermore , we discuss more general inputs : states with indefinite photon number and states with photons distributed between distinguishable time bins .
we prove that neither of these is helpful in improving phase estimation precision . |
we write @xmath15 for the real line and @xmath16 for the complex plane . moreover , we write @xmath17 for the extended complex plane ( the riemann sphere ) . for a complex variable @xmath18 ,
let @xmath19 denote the normalized arc length and area measures as indicated . moreover , we shall write @xmath20 for the normalized laplacian , and @xmath21 for the standard complex derivatives ; then @xmath22 factors as @xmath23 .
often we will drop the subscript for these differential operators when it is obvious from the context with respect to which variable they apply .
we let @xmath24 denote the open unit disk , @xmath25 the unit circle , and @xmath26 the exterior disk : @xmath27 we will find it useful to introduce the sesquilinear forms @xmath28 and @xmath29 , as given by @xmath30 where , in the first case , @xmath31 is required , and in the second , we need that @xmath32 . at times
we need the notation @xmath33 for the characteristic function of a subset @xmath34 , which equals @xmath35 on @xmath34 and vanishes off @xmath34 . the _ bloch space _ consists of those holomorphic functions @xmath36 subject to the seminorm boundedness condition @xmath37
let @xmath38 denote the group of sense - preserving mbius automorphism of @xmath24 .
by direct calculation , @xmath39 which shows that the bloch seminorm invariant under all mbius automorphisms of @xmath24 . the subspace @xmath40 is called the _ little bloch space_. we shall study growth properties of bloch function , where functions in the little bloch space are seen to grow too slowly .
in other words , the properties will take place in the quotient space @xmath41 .
an immediate observation we can make at this point is that provided that @xmath42 , we have the estimate @xmath43 which is sharp pointwise . for @xmath44 ,
let @xmath45 be its _
bergman projection_. it is the orthogonal projection to the holomorphic functions in @xmath46 , and acts boundedly on @xmath47 for each @xmath48 in the interval @xmath49 ( see , e.g. , @xcite ) . by appealing to the hahn - banach theorem
, we may identify the dual space of @xmath50 isometrically and isomorphically with the space @xmath51 , with respect to the sesquilinear form @xmath52 , provided @xmath51 is equipped with the canonical norm @xmath53 however , as for @xmath54 and @xmath55 , it may happen that @xmath56 fails to be in @xmath57 , the identification via the sesquilinear form requires some care .
the following calculation shows that that @xmath58 remains meaningful for @xmath54 and @xmath59 with @xmath60 ( @xmath61 denotes the @xmath62-dilate of @xmath63 ) : @xmath64 here , we use the facts that the bergman projection @xmath65 is self - adjoint on @xmath46 and preserves @xmath66 , and that we have the norm convergence @xmath67 as @xmath68 in the space @xmath50 .
it was shown by coifman , rochberg , and weiss @xcite that as a linear space , @xmath51 equals the bloch space @xmath69 , but actually , the endowed norm differs substantially from the seminorm .
recently , perl @xcite obtained the rather elementary estimate @xmath70 and showed that the constant @xmath71 is best possible . in the other direction , the best constant
is not known , but up to a little bloch function , the constant @xmath35 works . in conclusion ,
trying to understand @xmath51 in terms of the bloch seminorm involves a substantial loss of information .
we mention briefly the topic of optimal discretization of a given positive riesz mass as the sum of unit point masses .
the optimization is over the possible locations of the various point masses .
while this problem has a classical flavor , it seems to have never been pursued in the precise context we now present . for @xmath62 with @xmath72 and a polynomial @xmath63
, we consider the function @xmath73 which we call the _ hyperbolic discrepancy function_. the function @xmath74 can not vanish even locally , because @xmath75 means that @xmath76 , which is not possible for holomorphic @xmath63 because in the sense of distribution theory , @xmath77 is a sum of unit point masses , whereas @xmath78 , which is a smooth positive riesz density .
we are interested in the quantity @xmath79 where the infimum runs over all polynomials @xmath63 .
the number @xmath1 , which obviously is confined to the interval @xmath80 , will be referred to as the _ minimal discrepancy density for hyperbolic zero packing_. it measures how close the function @xmath74 can be to @xmath81 , on average .
there is also a more geometric interpretation ( compare with remark [ rem - geometry ] ) . in connection with the universal asymptotic variance @xmath5 defined below
, a variant of the density @xmath1 is more appropriate , which we denote by @xmath82 .
we write @xmath83 so that @xmath84 on @xmath85 while @xmath86 on the annulus @xmath87 .
the number @xmath82 is defined by @xmath88 and we call it the _ minimal discrepancy density for tight hyperbolic zero packing_. clearly , we see that @xmath89 .
we believe that @xmath90 .
this amounts to the belief that there are approximate minimizing polynomials @xmath63 for @xmath1 with modest growth and hence few zeros in the annulus @xmath87 . as for an actual extremal function @xmath91 for a given @xmath62 , it is sure to exist as a holomorphic function in @xmath85 by a normal families argument and fatou s lemma .
moreover , by a variational argument , we would also know that @xmath92=0,\ ] ] where @xmath93 denotes the bergman projection corresponding to the disk @xmath85 .
\(a ) the number @xmath94 describes the asymptotic minimal angle between the two vectors @xmath95 and @xmath35 along a family of weighted real hilbert spaces , as can be seen from lemma [ lem - basic0 ] below .
the analogous statement holds for @xmath96 as well .
( b ) to better explain zero packing , our exposition in section [ sec - imprcs ] begins with the planar case where the expression @xmath97 is the _ planar discrepancy function_. we believe that the equilateral triangular lattice has a good chance to be extremal for zero packing , and in section [ sec - imprcs ] , we explain how to evaluate the planar average of the corresponding @xmath98 as an integral over a single rhombus ( which is the union of two adjacent triangles ) .
\(c ) the hyperbolic zero packing problem considered here belongs to a more extensive family of problems .
indeed , it is equally natural to consider , more generally , for positive @xmath48 and @xmath99 , the hyperbolic @xmath100-discrepancy function @xmath101 .
the instance @xmath102 is related to the possible improvement in the application of the cauchy - schwarz inequality in @xcite and @xcite .
we now present the main result of this paper .
the minimal discrepancy density for hyperbolic zero packing enjoys the following estimate : @xmath103 .
[ thm - mddhzp ] the proof of this theorem is supplied in section [ sec - appendix ] .
the importance of theorem [ thm - mddhzp ] comes from its consequences .
suppose @xmath59 , where @xmath60 , and if @xmath104 denotes the dilate @xmath105 , then @xmath106 [ cor - main1 ] in other words , with @xmath107 as mcmullen s asymptotic variance @xcite , and @xmath108 as the universal asymptotic variance , we have that @xmath109 in fact , the first inequality is an equality .
we have that @xmath110 .
[ thm - equality ] in the paper @xcite by astala , ivrii , perl , and prause , the estimate @xmath111 was obtained . as a consequence of the inequality
, we obtain that @xmath112 .
this is where the estimate from above of theorem [ thm - mddhzp ] comes from .
this estimate is much smaller than the value @xmath113 which is the expected value of the discrepancy density for an appropriately tailored gaussian analytic function ( see subsection [ subsec - stochhyper ] ) .
the approximately extremal polynomial @xmath63 for the definition of the discrepancy density @xmath1 should have its zeros as hyperbolically equidistributed as possible , with a prescribed density . since it stands to reason that we may model these approximately minimizing polynomials by a single holomorphic function @xmath63 in the disk @xmath24 , we could try to look for @xmath63 which is a diffential of order @xmath35 ( or a character - diffential of the same order @xmath35 ) , periodic with respect to a fuchsian group @xmath114 such that @xmath115 is a compact riemann surface .
the most natural choice would be to also ask that the zeros of @xmath63 are located along a hyperbolic equilateral triangular lattice .
for instance , we may compare with the analogous planar case the bound achieved by the unilateral triangular lattice is @xmath14 .
however , the structure of hyperbolic lattices is more rigid than the corresponding planar one , and the relevant quantities are harder to evaluate .
mcmullen s notion of asymptotic variance is very much related to makarov s modelling of bloch functions as martingales @xcite , @xcite , @xcite .
compare also with lyons approach @xcite to understand bloch functions as maps from hyperbolic brownian motion to a planar brownian motion ( but for it , the speed of the local variance is variable but at least bounded ) @xcite .
we note in passing that in @xcite , the related notion of _ asymptotic tail variance _ was introduced . for @xmath116
, we consider the class @xmath117 of normalized @xmath12-quasiconformal mappings @xmath118 , where @xmath17 is the riemann sphere , which preserve the point at infinity and are conformal in the exterior disk @xmath26 .
the normalization is such that the mapping has a convergent laurent expansion of the form @xmath119 the _ integral means spectrum _ for the function @xmath120 ( which is defined in @xmath26 only ) is the function @xmath121 the _ universal _ integral means spectrum is obtained as @xmath122 , where @xmath123 and @xmath124 ranges over @xmath125 . in @xcite , ivrii obtains the following asymptotics for @xmath126 .
( ivrii ) the universal integral means spectrum enjoys the asymptotics @xmath127 [ thm - ivrii ] here , @xmath128 is the universal constant which appears in , so that @xmath129 .
hence a combination of theorems [ cor - main1 ] and [ thm - ivrii ] refutes the general conjecture to the effect that @xmath130 for real @xmath11 with @xmath131 @xcite , @xcite .
we now comment on ivrii s proof of his theorem .
it is an important for the proof that for small @xmath12 , the function @xmath132 can be modelled by @xmath133 for some @xmath60 with @xmath134 , where @xmath135 denotes the beurling transform @xmath136 moreover , after an inversion of the plane , @xmath133 essentially becomes @xmath137 .
while this is standard technology in quasiconformal theory , the first important observation ivrii makes is the `` box lemma '' , which says that for @xmath59 with @xmath138 , the control of the right - hand side integral in @xmath139 can be localized to a hyperbolic disk of large fixed radius instead .
this is a kind of weak control of square function type ( compare with e.g. bauelos @xcite ) , which tells us we are in the right ballpark .
a clever combination with the lipschitz property of bloch functions @xcite then gives the control from above and below , more or less simultaneously .
ivrii actually obtains slightly better control than stated above . in any case
, he also derives the following dimension expansion via the legendre transform formalism connecting the dimension and integral means spectra ( see , e.g. , @xcite , @xcite , and @xcite , p. 241 ) .
( ivrii ) the maximal minkowski ( or hausdorff ) dimension @xmath140 of a @xmath12-quasicircle has the asymptotic expansion @xmath141 [ cor - ivrii ] here , a _
@xmath12-quasicircle _ is simply the image of the unit circle @xmath142 under a @xmath12-quasiconformal mapping of the riemann sphere @xmath143 .
in particular , astala s well - known conjecture @xmath144 is incorrect .
in fact , prause made the observation that @xmath145 holds for every @xmath116 , based on a combination of corollary [ cor - ivrii ] and the methods developed by prause and smirnov @xcite , @xcite .
i would like to thank oleg ivrii for reading carefully a first version of this manuscript .
in addition , i would like to thank kari astala , alexander borichev , michael benedicks , istvn prause , and aron wennman for several valuable conversations .
the following identity interchanges dilations , and although elementary , it is quite important .
we write @xmath146 and @xmath104 for the dilates @xmath61 and @xmath147 , respectively .
suppose @xmath148 are two harmonic functions , which are are area - integrable : @xmath149 .
then we have that @xmath150 [ lem - basic1 ] this is lemma 5.1.1 in @xcite .
we also need the following identity .
suppose @xmath151 are functions , where @xmath152 is holomorphic and @xmath153 is harmonic . if @xmath154 and @xmath153 is the poisson integral of a function in @xmath155 , then we have that @xmath156 where we write @xmath157 for the coordinate function @xmath158 .
[ lem - basic2 ] this is lemma 5.2.1 in @xcite .
the classical isoperimetric inequality says that the area enclosed by a closed loop of length @xmath160 is at most @xmath161 .
torsten carleman ( see @xcite , @xcite ) found a nice analytical approach to this fact , which gave the estimate @xmath162 here , @xmath163 is the @xmath164 instance of the classical _ hardy space _
@xmath165 , for @xmath166 . for @xmath167 , @xmath165
consists all holomorphic functions @xmath168 subject to the norm boundedness condition @xmath169 as for , the geometrically relevant case is when @xmath63 is the derivative of the conformal mapping from disk @xmath24 to the domain enclosed by the loop .
there is of course no converse to the isoperimetrical inequality , since for a given enclosed area , the length of the boundary may be infinite .
however , if the boundary curve is regularized by replacing it with a level curve of the green function , the reverse problem starts to make sense .
we will not need here the appropriate regularized reverse version of , but instead the analogue where the hardy space @xmath163 is replaced by the corresponding bergman space @xmath50 of area - integrable holomorphic functions . for @xmath167 and @xmath170 , we introduce the scale of standard weighted lebesgue spaces @xmath171 of ( equivalence classes of ) borel measurable functions @xmath168 with @xmath172 we say that @xmath173 if and only if @xmath63 is holomorphic in @xmath24 and @xmath174 . in this case , we will often write @xmath175 in place of @xmath176 . the spaces @xmath177 are known as _ the standard weighted bergman spaces_. for @xmath178 , we recover the bergman spaces : @xmath179 . for @xmath180 , it is easy to see that the weighted bergman space is trivial : @xmath181 . on the other hand , for polynomials @xmath63 , @xmath182 so that in a sense , the hardy space @xmath165 appears as the limit of the space @xmath177 as @xmath183 .
this suggests that carleman s estimate might be part of a more general estimate comparing the norm in @xmath177 with that of @xmath184
. we shall be interested in obtaining a reverse inequality after dilation , with @xmath164 and @xmath178 : @xmath185 here , @xmath186 and @xmath72 .
the question at hand is to obtain in explicit form , or at least to estimate from above , the optimal constant @xmath187 , for @xmath72 . by the cauchy - schwarz inequality
, we have that @xmath188 this immediately shows that the optimal constant in is at most @xmath189 we intend to improve this estimate .
we need to analyze the degree of suboptimality in the cauchy - schwarz inequality in various situation . to this end , the following lemma is helpful
. if @xmath190 is an @xmath15-linear hilbert space , the following three conditions are equivalent for two given vectors @xmath191 and a real @xmath192 with @xmath193 : \(a ) @xmath194 , \(b ) @xmath195 , and \(c ) @xmath196 .
[ lem - basic0 ] if @xmath197 , all the three conditions are trivially met .
next , we assume @xmath198 . by expanding the square
, we find that @xmath199 which for @xmath200 attains its minimum for @xmath201 : @xmath202 the equivalence of ( a ) and ( c ) for @xmath203 is immediate from this formula . as for ( b ) , we note that if introduce the reciprocal constant @xmath204 , the inequality reads @xmath205 , which is the same as ( a ) if we switch the roles of @xmath206 and @xmath207 .
moreover , since ( c ) is preserved under such a switch , the equivalence of ( b ) and ( c ) now follows from the equivalence of ( a ) and ( c ) .
we now explain where the bound stated in theorem [ cor - main1 ] comes from .
we assume @xmath72 . from the definition of @xmath82
, we see that there exists a parameter @xmath208 with @xmath209 and @xmath210 as @xmath68 , such that @xmath211 it is immediate from and lemma [ lem - basic0 ] that @xmath212 where initially , @xmath63 is a polynomial , but by approximation , holds for any holomorphic @xmath63 in @xmath85 such that the right - hand side integral is finite .
next , we pick a bounded holomorphic function @xmath213 with @xmath214 , and apply lemma [ lem - basic2 ] combined with : @xmath215 it now follows that @xmath216 and we may put this together with the estimate , with @xmath217 , and arrive at @xmath218 by the elementary inequality @xmath219 and the standard paley identity ( @xcite , p. 236 ) for the @xmath220 norm ( which is a consequence of green s formula ) , we know that @xmath221 where in the last step , we used that @xmath214 .
we put this together with : @xmath222 we plug in @xmath223 , which gives that @xmath224 since @xmath225 as @xmath226 , the claimed estimate now follows .
we now turn to the proof of theorem [ thm - equality ] . in the same way as follows from the definition of @xmath1 , we read off from the definition of @xmath82 that @xmath227 for some positive @xmath228 with @xmath225 as @xmath68 .
moreover , the positive constant @xmath82 is the biggest universal constant so that holds . we will use duality to obtain the assertion that @xmath110 . to see this ,
let @xmath229 denote the dilation @xmath230 , and observe that expresses the operator bound @xmath231 with respect to @xmath52 , the dual space to the weighted bergman space @xmath232 is isometrically @xmath233 , which is just @xmath234 but equipped with the equivalent norm @xmath235 since the bound is asymptotically sharp , we then have @xmath236 with respect to the dual action @xmath52 , lemma [ lem - basic1 ] tells us that @xmath237 , and we recall that isometrically , the dual space to @xmath238 is @xmath239 while the dual to @xmath50 is @xmath51 .
since by basic functional analysis the norm of an operator and its adjoint are the same , we get from that @xmath240 for @xmath60 and @xmath59 , we observe that @xmath241 , which shows that @xmath242 as @xmath68 .
it follows from this combined with that @xmath243 now , the left - hand side expresses a uniform version of the asymptotic variance @xmath5 , which actually equals @xmath5 , by a generational construction of an approximately simultaneously extremal @xmath244 in the unit ball of @xmath6 using successive annuli which have hyperbolic width tending to infinity .
the details are left to the reader .
here , we consider a new type of extremal problems in complex analysis , which we call _ zero packing problems_. we first explain the planar zero packing problem , and then turn to the hyperbolic zero packing problem .
we first study a packing problem for zeros pertaining to the bargmann - fock space of entire functions .
it is well - known that there is no entire function @xmath245 such that @xmath246 .
the reason is that in the sense of distribution theory , @xmath77 equals the sum of unit point masses located at the zeros of @xmath63 ( counting multiplicities ) , so that off the zeros , @xmath247 is harmonic , while @xmath248 .
in particular , the nonnegative function @xmath249 can not vanish except on a one - dimensional set , and if @xmath250 is a polynomial in @xmath157 , then in particular @xmath251 as @xmath252 , and we would know that the _ discrepancy function _
@xmath253 is bounded .
note also that for the trivial function @xmath254 , the discrepancy @xmath255 equals the constant @xmath35 .
it is now a natural question to ask _ how small the discrepancy @xmath98 can be , on average _ , since it can not vanish even locally .
so , we consider the minimal average of @xmath98 in a disk @xmath256 of large radius @xmath257 : @xmath258 where the infimum is taken over all polynomials @xmath63 . here , the use of the origin as the base point is inessential since in , we can take the infimum over all entire @xmath63 without changing the value of @xmath259 , and , in addition , by the change - of - variables formula , we have for @xmath260 the translation invariance property @xmath261 where @xmath262 denotes the fock - space translate @xmath263 . in view of lemma [ lem - basic0 ] , this discrepancy density @xmath259 gives the best constant for the improved cauchy - schwarz inequality @xmath264 for the above problem , the the _ minimal discrepancy density for planar zero packing _
is @xmath265 .
[ defn - odp1 ] \(a ) the limsup might be considered as well , but we expect it to equal the liminf .
\(b ) in more geometric terms , the quantity @xmath266 is a measure of how well the geometries the following two geometries can approximate each other : ( i ) a surface with curvature form which is equal to a negative constant times planar area measure and ( ii ) a surface which is flat except for cone points of opening @xmath267 .
[ rem - geometry ] since polynomials are determined up to a multiplicative constant by their zeros , we feel that the terminology `` zero packing '' or `` packing of zeros '' is appropriate .
determine the value of @xmath0 .
for which configurations of zeros of the polynomial @xmath63 is it asymptotically attained ? is the equilateral triangular lattice optimal asymptotically ?
we illustrate with an equilateral triangular triangular tesselation in figure [ fig-1 ] .
complete_hex_motif.pdf the weierstrass sigma function @xmath268 can be used to analyze the asymptotic discrepancy density for the equilateral triangular lattice .
let @xmath269 be the half - periods associated with the lattice @xmath270 where it is assumed that @xmath271 are @xmath15-linearly independent .
we choose @xmath272 and @xmath273 , where @xmath274 is a positive real , and observe that the associated lattice @xmath275 is the equilateral triangular lattice with spacing @xmath276 between nearby points .
we recall the formula for the associated sigma function @xcite , @xcite : @xmath277 the function @xmath268 is entire , with periodicity - type formulae ( see @xcite , @xcite ) @xmath278 where @xmath279 denotes the logarithmic derivative ( the weierstrass zeta function ) .
we consider in the zero packing problem the function @xmath63 @xmath280 where @xmath281 is a positive amplitude constant , and @xmath282 is a parameter to be determined .
we would like the associated function @xmath283 to be periodic with the two complex periods @xmath284 . as it turns out ,
if we pick @xmath285 in the following way , @xmath286 this is the case . the asymptotic discrepancy density associated with this particular choice
can then be calculated over a single fundamental rhombus @xmath287 for the tiling @xmath288 , @xmath289 where we are free to minimize over the parameter @xmath281 . here
, @xmath290 is the normalized area of the fundamental rhombus .
the right - hand side of is in a natural sense the average of @xmath98 over the torus @xmath288 .
if we take logarithms in , we obtain the real - valued function @xmath291 which is @xmath275-periodic and has laplacian @xmath292 in the sense of distribution theory , where @xmath293 is the unit point mass at the point @xmath294 .
this means that on the torus @xmath288 , @xmath295 plays a role analogous to the fundamental solution for the laplacian in the plane .
we suggest the term _ logarithmic monopole _ for the function @xmath295 ( and any of its translates , whose laplacians has a point mass at some other point of the fundamental rhombus ) .
[ rem - monopole ] the zero packing problem minimizes the ratio of the integral of @xmath98 and the integral of the constant function @xmath35 .
an alternative approach would be to divide instead by the integral of @xmath296 times the appropriate gaussian weight .
to this end , we introduce a modified discrepancy function @xmath297 where @xmath298 is constant .
so , we consider the corresponding minimal weighted average of the discrepancy @xmath299 : @xmath300 here , the infimum is taken over all @xmath298 and all polynomials @xmath63 ( except the @xmath81 polynomial ) .
as it turns out , this does not give rise to a new concept .
we have that @xmath301 .
this follows from lemma [ lem - basic0 ] , since we are free to multiply @xmath63 by a scalar in the minimization .
it is difficult to know offhand what kind of packing of zeros would be optimal for the calculation of the asymptotic minimal discrepancy density @xmath302 .
a reasonable approach is to let a stochastic process do the digging for the optimal configuration , as in the so - called _ bellman function method _ , exploited repeatedly in harmonic analysis ( see , e.g. , the survey @xcite ) .
first , we note that the assumption that the function @xmath63 should be a polynomial in is excessive , since polynomials are dense in many spaces of holomorphic functions .
in particular , the density @xmath259 is unperturbed if we minimize e.g. over all entire functions @xmath63 . here
, we will replace @xmath63 by a gaussian analytic function ( gaf ) with close - to - optimal behavior . to set the notation
, we let @xmath303 stand for the standard rotationally invariant gaussian distribution with probability measure @xmath304 in the plane @xmath16 .
we pick independent copies @xmath305 for @xmath306 , and let @xmath307 be the gaf process @xcite @xmath308 the way things are set up , @xmath309 is a stationary process in the plane @xmath16 , with distribution from the standard normal @xmath303 .
given a positive amplitude constant @xmath281 , we observe that the associated density @xmath310 is stochastic , and we may ask for the number @xmath311 where @xmath312 stands for the probability of the event @xmath313 . then clearly , @xmath314 , and we conjecture the following .
we suggest that @xmath315 .
[ conj-1 ] in other words , since clearly @xmath316 , we believe that the given family of gafs is able to approximate with positive density the states @xmath63 which optimize asymptotically the discrepancy density @xmath0 . if the conjecture is true , this should offer a path to approximate @xmath0 with straightforward numerical methods .
let @xmath317 stand for the expectation , and observe that @xmath318 which tells us that the expected value of @xmath319 is minimized for the amplitude @xmath320 , and that the minimal expected value equals @xmath113 .
we obtain immediately an upper bound for @xmath0 : we have the following bounds : @xmath321 by the planar analogues of the methods we develop in section [ sec - appendix ] for the hyperbolic setting , it can established that @xmath322 .
we remark in addition that it is likely that @xmath323 .
numerical implementation of the suggested minimal configuration associated with the weierstrass sigma function , see was carried out by wennman @xcite using maple , which resulted in the estimate @xmath324 .
[ rem-4.3.3 ] we will work with an analogous packing problem coming from hyperbolic geometry , related to the possible improvement in the cauchy - schwarz inequality , in line with lemma [ lem - basic0 ] .
this time , the discrepancy is given by @xmath325 for a polynomial @xmath63 , or more generally , @xmath63 which is holomorphic in @xmath24 .
again , @xmath75 is the same as the equality @xmath326 is possible only when both @xmath63 and @xmath327 equal vanish .
the reason is the same as before : @xmath247 is harmonic off the zeros of @xmath63 , while @xmath328 . the average density of @xmath74 with respect to the hyperbolic area element
@xmath329 is the ratio @xmath330 and we could consider the inf over @xmath63 and then the liminf as @xmath68 .
however , in hyperbolic geometry , the length of boundary of @xmath85 is substantial the cutoff is a bit rough . to reduce the boundary effects , we instead average further before taking the ratio ( compare , e.g. with seip s densities @xcite ) , @xmath331 so , the minimal average discrepancy we are after is , for @xmath72 , @xmath332 where the infimum is over all polynomials @xmath63 , or , which gives the same result , over all holomorphic functions @xmath168 . in view of lemma [ lem - basic0 ] , this discrepancy is the best constant for the improved cauchy - schwarz inequality @xmath333 for the above problem , the the _ minimal discrepancy density for hyperbolic zero packing _ is @xmath334 .
[ defn - odp2 ] determine the value of @xmath1 . for which configurations of zeros of the function @xmath63
is it asymptotically attained ? is a lattice configuration optimal
asymptotically ? [ prob-2.001 ]
although the zero packing problem involves global issues , it probably has some analogies with the more local hyperbolic circle packing problems ( see , e.g. , @xcite ) .
one strategy for hyperbolic zero packing would be to pack according to a lattice configuration , for instance given by a tiling of the disk by hyperbolic regular @xmath48-gons with @xmath99-fold vertices ( provided @xmath335 ) .
we illustrate with an eightfold octagonal ( @xmath336 ) tiling of figure [ fig-3 ] .
such a _ schfli tiling _ exists provided that @xmath337 , and then the hyperbolic @xmath338-area of the @xmath48-gon is precisely @xmath339 .
a schfli tile is not always a fundamental domain for a fuchsian group @xmath114 , but it is nevertheless preferable to pass the setting where @xmath115 is a compact riemann surface of genus @xmath340 , and @xmath341 is a fundamental domain .
ngon8.pdf we leave setting of polynomial @xmath63 and suppose that @xmath168 is a nontrivial holomorphic function with @xmath342 which will happen if @xmath343 where @xmath344 is a character , that is , @xmath345 holds for @xmath346 .
we mention in passing that if @xmath347 , this would means that @xmath63 is a differential form of order @xmath35 .
note that the space of such differential forms has dimension equal to the genus of @xmath115 .
we mention in passing that @xmath347 may not the best choice , because we need to have more flexibility ( see remark [ rem - final ] below ) .
in addition , there may be a connection with mcmullen s solution of kra s theta conjecture @xcite , which tells us that the so - called averaging operator @xmath348 is a strict contraction of summable holomorphic quadratic diffentials .
since for @xmath349 we have that @xmath350 it follows from the assumption that @xmath351 which means that the function @xmath352 and hence @xmath74 is @xmath114-periodic . in this case , there is an easy way to compute the discrepancy density .
in fact , the formula analogous to the planar formula reads @xmath353 where @xmath341 is the fundamental domain for @xmath115 .
this can be shown by appeal to the ergodicity of geodesic flow in @xmath115 , which says that the time average over the flow converges to the space average with respect to the invariant measure ( the hyperbolic area element ) ( see , e.g. , hopf s expository paper @xcite ) . indeed ,
the `` time average '' of @xmath74 over the geodesic ray @xmath354 with @xmath355 and radial parameter @xmath11 with @xmath356 , is @xmath357 while the space average is expressed by the right - hand side of .
the limit of the ratio as @xmath68 is clearly unperturbed if we replace @xmath358 by @xmath359 , and hence results from by integration over the circle @xmath142 in @xmath360 . here
, we used the elementary observation that @xmath361 so , for a given fuchsian group @xmath114 , we would need to express a nontrivial holomorphic function @xmath63 such that holds , and then calculate the ratio . a natural approach to this issue is the construction of a hyperbolic logarithmic monopole ( compare with remark [ rem - monopole ] for the planar case ) . here , the _ hyperbolic logarithmic monopole _ ( a relative of the famous green function ) on the surface
@xmath115 is the real - valued function @xmath362 of @xmath363 such that @xmath364 in the sense of distribution theory , where @xmath365 denotes the measure which gives the unit point mass at @xmath366 .
the number @xmath367 appearing in the formula is the @xmath338-area of the fundamental domain @xmath341 .
it is needed to assure the existence of the function @xmath362 .
note that the property only determines @xmath362 up to an additive real constant . to determine it uniquely
, we might add the requirement that @xmath368 for some other point @xmath369 of the surface . by forming the difference
@xmath370 we obtain a dipole : @xmath371 the hyperbolic dipole function is better known in the literature on riemann surfaces than the monopole ( see , e.g. , @xcite , p. 213 ) .
in fact , it is possible to obtain the monopole @xmath362 from the dipole @xmath372 by integration over @xmath373 with respect to the ( hyperbolic ) surface area element .
we return to the issue of constructing a nontrivial holomorphic function @xmath63 such that holds . forming the logarithm @xmath374
, we have a function on the surface @xmath115 such that @xmath375 where @xmath376 are the zeros of the function @xmath95 on the surface @xmath115 ( they are essentially the zeros of @xmath63 in a fundamental domain @xmath341 , which is why there are finitely many of them ) . comparing with , we obtain that if @xmath377 , we can obtain a nontrivial @xmath63 from @xmath378 this settles the case when @xmath379 fairly constructively . actually , there is no remaining case , as the gauss - bonnet theorem gives that @xmath380 , an even integer , where @xmath381 is the genus of the compact surface @xmath115 .
\(a ) the first instance of the density formula which should be tried is probably when @xmath382 in and correspondingly @xmath383 . for instance , the schfli tile with @xmath384 and @xmath385 is the fundamental domain for a compact surface with hyperbolic area @xmath383 .
\(b ) the function @xmath63 that results from will generically satisfy with a character @xmath386 .
indeed , the nontrivial holomorphic @xmath63 with are parametrized by the zeros @xmath376 and an amplitude constant in @xmath387 , so they form an @xmath388-dimensional complex variety , whereas the holomorphic differentials form a @xmath16-linear subspace of dimension @xmath152 . as one dimension is bigger than the other , @xmath389 , the holomorphic differentials are rare among the solutions to .
[ rem - final ] as in the planar case , it is difficult to know offhand what kind of packing of zeros would be optimal for the calculation of the asymptotic minimal discrepancy density @xmath1 .
again , a reasonable approach is to let a stochastic process do the digging for the optimal configuration , and we look for an appropriate gaf process to supply random holomorphic functions in @xmath24 .
as before , we pick independent copies @xmath390 for @xmath306 , and let @xmath391 be the gaf process @xmath392 the way things are set up , @xmath393 is a stationary process in the disk @xmath24 , with distribution from the standard normal @xmath303 . given a positive amplitude constant @xmath394 , we observe that the associated density @xmath395 is stochastic , and we may ask for the number @xmath396 then clearly , @xmath397 , and in analogy with conjecture [ conj-1 ] , we conjecture the following .
we suggest that @xmath398 .
[ conj-2 ] as for the value of the asymptotic density @xmath1 , we observe that @xmath399 which tells us that the expected value of @xmath400 is minimized for the amplitude @xmath401 , and that the minimal expected value equals @xmath113 .
we obtain immediately an upper bound for @xmath1 , which is the same as in the planar case . this bound is weaker than the one found by astala , ivrii , perl , and prause in @xcite ( @xmath402 ) .
we have the following bounds : @xmath403 [ prop - estabove1.01 ] in section [ sec - appendix ] , we will obtain the estimate @xmath404 .
while this is far from optimal , it does give an improvement in the cauchy - schwarz inequality .
suppose @xmath168 is holomorphic and nontrivial
. then the function @xmath352 has local minima only at the zeros of @xmath63 .
moreover , for @xmath72 , the gradient of this function enjoys the estimate @xmath410 [ lem - pointwise1.03 ] since @xmath411=-(1-|z|^2)^{-2}<0\ ] ] holds off the zeros of @xmath63 , the critical points of the function @xmath412 can only be local maxima or saddle points , and this carries over to the function @xmath95 as well .
the estimate of the gradient uses the estimates of lemma [ lem - pointwise1.02 ] together with the product rule @xmath413=-|f(z)|\nabla|z|^2+(1-|z|^2)\nabla|f(z)|\ ] ] and the facts that @xmath414 and @xmath415 .
the necessary details are left to the reader .
we observe that we may replace @xmath63 by @xmath420 , where @xmath327 is a positive constant , and show instead that @xmath421 under the following normalization : @xmath422 by lemma [ lem - pointwise1.03 ] , we know from the normalization that @xmath423 where we decided to estimate on a slightly smaller disk .
next , a straightforward calculus exercise shows that the minimum over @xmath424 is attained at the value @xmath425 moreover , a well - known calculation shows that @xmath426 so that in particular , @xmath427
. next , we split our argument according to the size of @xmath428 .
_ suppose that @xmath431_. we let @xmath307 be the function @xmath432 , so that by , we know that @xmath433 we consider the set @xmath434 given by @xmath435 we divide case ii further according to the properties of the set @xmath434 .
case iia : _ suppose that _ @xmath436
. then @xmath63 has no zeros on @xmath437 , and , in addition , @xmath438 so that by the maximum principle applied to @xmath439 , @xmath440 as a consequence , we obtain that @xmath441 which gives the claimed estimate provided that @xmath442 .
case iib : _ suppose that _ @xmath443 .
then there exists a point @xmath444 with @xmath445 by lemma [ lem - pointwise1.03 ] , the function @xmath446 has no local minima off the zeros of @xmath63
. then we just follow the direction of steepest descent of the graph , which is well - defined by gradient flow except at the critical points . at the critical points ,
if it is a local maximum we could choose any direction at random ( we would never get there , only the starting point could be a local maximum ) , and at a saddle point , we go in the direction of the largest negative directional second derivative . in any case
, we obtain a path @xmath447 inside the disk @xmath448 , which starts at the point @xmath449 , and it leads us _ either to a zero of @xmath63 or to the boundary _ @xmath450 . for a given @xmath451 ,
we associate to the path @xmath447 the domain @xmath452 along the path @xmath447 , the function @xmath307 is decreasing , so that in particular , @xmath453 and hence , by , @xmath454 we pick @xmath455 such that @xmath456 , which guarantees that @xmath457 on @xmath458 .
then @xmath459 and we are led to estimate the normalized weighted area of @xmath458 . the estimate depends on the properties of the path @xmath447 . then the path @xmath447 takes us between two points that are at a distance greater than @xmath460 , and the weighted area of @xmath458 is at least @xmath461 in view of , then , we obtain that @xmath462 where , in the last step , we implemented the choice @xmath463 .
so , in this case , the assertion holds if @xmath464 .
case iib(b ) : _ the path @xmath447 leads from @xmath449 to a zero of @xmath63 in the disk @xmath448_. let @xmath465 denote the zero in question . then @xmath466 and @xmath467 , so that in view of the gradient estimate , we have that @xmath468 and at the same time , we know that the the area of the intersection @xmath469 equals at least a quarter of the area of @xmath470 , that is , @xmath471 .
it now follows that @xmath472 and , consequently , the assertion holds if @xmath473 .
there exists an absolute constant @xmath474 with @xmath475 , such that for all holomorphic @xmath168 and all points @xmath476 , @xmath477 [ prop-2.2 ] for instance , @xmath478 will do , where @xmath416 is the constant of proposition [ prop-2.1 ] . as mentioned in the intrduction , the estimate from above @xmath481 follows from the work of astala , ivrii , perl , and prause @xcite , so it remains to establish the estimate from below .
our starting point is proposition [ prop-2.2 ] , which tells us that there exists an absolute constant @xmath474 , with @xmath475 , such that for each @xmath482 and each holomorphic function @xmath213 , @xmath483 given @xmath482 , we introduce the mapping @xmath484 given by @xmath485 which is an involutive mbius automorphism of the unit disk @xmath24 ( so that @xmath486 ) .
moreover , a direct calculation shows that the derivative of @xmath484 equals @xmath487 we make the auxiliary observation that @xmath488 let @xmath489 denote the holomorphic function @xmath490 and observe that by and the change - of - variables formula , @xmath491 where @xmath492 , and , in the last step , we invoked with @xmath493 in place of @xmath153 .
if we write @xmath494 in place of @xmath489 , we obtain from that @xmath495 we are of course free to integrate both sides with respect to a positive finite measure : @xmath496 moreover , we calculate that @xmath497 for @xmath498 , provided @xmath499 is close enough to @xmath35 , where the bound by @xmath500 is a consequence of hyperbolic invariance , and the fact the left hand side vanishes is a consequence of a simple comparison of the hyperbolic lengths of the intervals @xmath501 $ ] and @xmath502 $ ] the latter interval is longer for @xmath498 ) .
it now follows from that @xmath503 since with @xmath478 and @xmath504 , the constant @xmath505 and since @xmath506 the claimed assertion follows from .
astala , k. , iwaniec , t. , martin , g. , _ elliptic partial differential equations and quasiconformal mappings in the plane_. princeton mathematical series , * 48*. princeton university press , princeton , nj , 2009 .
gradshteyn , i. s. , ryzhik , i. m. , _ table of integrals , series , and products_. corrected and enlarged edition edited by alan jeffrey . incorporating the fourth edition edited by yu .
v. geronimus and m. yu .
tseytlin . translated from the russian . academic press , new york - london - toronto , ont . , 1980 .
hough , j. b. , krishnapur , m. , peres , y. , virg , b. , _ zeros of gaussian analytic functions and determinantal point processes_. university lecture series , * 51*. american mathematical society , providence , ri , 2009 . | we introduce a new type of problem in complex analysis , referred to as _
zero packing_. the associated quantity is a density , denoted @xmath0 in the planar case , and @xmath1 in the case of the hyperbolic plane .
we refer to these densities as _ discrepancy densities for planar and hyperbolic zero packing _ , respectively , as they measure the impossibility of atomizing the uniform planar and hyperbolic area measures . a related density , @xmath2 , called the _ discrepancy density for tight zero packing _ , has the property that @xmath3 , and we conjecture that @xmath4 holds . the universal asymptotic variance @xmath5 associated with the image under the bergman projection of @xmath6 is related to one of these discrepancy densities : @xmath7 .
we obtain the estimates @xmath8 , where the upper estimate is derived from the estimate from below on @xmath5 obtained by astala , ivrii , perl , and prause , and the estimate from below is much more delicate .
in particular , it follows that @xmath9 , which in combination with the work of ivrii shows that the maximal dimension of quasicircles conjectured by astala can not be reached .
moreover , along the way , since the universal quasiconformal integral means spectrum has the asymptotics @xmath10 for small @xmath11 and @xmath12 , the conjectured formula @xmath13 is not true . as for the actual numerical values of the above discrepancy densities , we obtain the estimate from above @xmath14 by using the equilateral triangular zero packing , and suggest that equality should hold .
the values of the other discrepancy densities are expected to be somewhat close to the value of @xmath0 . |
the increasing amount of neutron data on quasi-1d antiferromagnets calls for creation of a theory capable of describing the behaviour of such strongly anisotropic nonlinear systems in greater detail . of particular interest with regard to recent experiments on @xmath2 and @xmath0
is the structure of the multiparticle continua . in the present paper
we discuss correlation functions of three closely related problems involving spin-@xmath3 heisenberg chains : ( i ) a single chain with a static alternating exchange , ( ii ) a single chain in a staggered magnetic field , ( iii ) a system of weakly coupled spin-@xmath3 heisenberg chains at low temperatures in the ordered phase . in the latter case
we consider two types of order : ( a ) spin - peierls @xcite and ( b ) antiferromagnetic @xcite .
an experimental realisation of situation ( i ) may be achieved in ( vo)@xmath4p@xmath4o@xmath5 .
the analysis of the recent experiments conducted by @xcite indicates that the strongest exchange occurs through alternatively arranged molecules of two different types such that the ratio of the exchange integrals is @xmath6 .
the second situation ( ii ) may be realized when a heisenberg magnet with a staggered land factor is placed in an external magnetic field .
this is the case for copper benzoate @xcite but the difference in land factors is very small and for reasons that will be discussed elsewhere the theory presented here does not apply .
the spin - peierls transition is a magnetoelastic transition which occurs in quasi-1d antiferromagnets due to an effective four - spin interaction .
such interactions may be generated by phonons which modify the exchange integrals ; however , the virtual absence of softening of the phonon spectrum in the spin - peierls material cugeo@xmath7 @xcite suggests that there may be also other mechanisms .
the interaction between the staggered parts of energy densities is strongly relevant and resolves in dimerization of the lattice at a certain temperature t@xmath8 and the formation of spectral gaps in the spectrum of magnetic excitations . in any realistic system phonons
are three - dimensional which determines a three dimensional nature of the spin - peierls transition . a simple model with a phonon mechanism has the following hamiltonian h & = & j_*r*,n(1+(-1)^nu_n,*r * ) _ n , * r*_n+1,*r * + _ n,*r**r***u_n,*r * d(*r - * r**)u_n,*r*&+&j_*r*,n_n , * r*_n+2,*r * + _ * r*,*c*,nk(*c * ) _ n , * r * _ n,*r+*c * * [ ha ] , where @xmath9 and @xmath10 label lattice sites along and perpendicular to the chains and @xmath11 are vectors connecting neighbouring chains . in order to make contact with experiments we have included n.n.n .
intrachain as well as interchain magnetic interactions .
since @xmath12 is usually much smaller than the phonon frequency , one can neglect the kinetic energy of the @xmath13-field and consider @xmath13 as commuting numbers .
one can denote ( n , * r * ) = ( -1)^n(_n,*r*_n + 1,*r * ) and integrate over the displacements .
the result is h_sp&= & j_*r*,n_n , * r*_n+1,*r*+ _ n,*r,*r**(n , * r*)j(*r*- * r*)(n , * r*)&+&j_*r*,n_n,*r*_n+2,*r * + _ * r*,*c*,nk(*c * ) _ n , * r * _ n,*r+*c * * , [ hsp ] where @xmath14 is proportional to the matrix inverse of @xmath15 . now one can forget about phonons and use ( [ hsp ] ) as a general model for the spin - peierls state . in their low temperature phases
both @xmath16 and the antiferromagnet discussed below have order parameters . for the spin - peierls model
this is the staggered energy density @xmath17 ( for simplicity we assume that there is no antiferromagnetic order in the spin - peierls model ) .
therefore in the low temperature phase it is convenient to subtract from @xmath18 its average value ( @xmath19 ) .
this leads to the following hamiltonian h_sp & = & _ * r*h^(0)_*r * + _ n,*r**r***:(n , * r*):j(*r*- * r*):(n , * r* ) : + _ * r*,*c*,nk(*c * ) _ n , * r * _ n,*r+*c * * , [ h1 ] + h^(0)_*r * & = & j_n_n , * r*_n + 1 , * r * + j_n_n , * r*_n+2,*r * + 2(j_a+j_b)_0_n(n,*r * ) - n_0 ^ 2(j_a+j_b ) , [ sp ] where @xmath20 are the couplings ( generated by the coupling to the phonons ) to the neighbouring chains in @xmath21 and @xmath22 directions ( the chain direction is chosen to be @xmath23 ) . for definiteness we assume in the following that @xmath24
. however , the general case can be treated by minor modifications of the formulas presented below .
sometimes the purely one - dimensional model ( [ sp ] ) is used to describe the low temperature phase .
we shall show below that such approximation always becomes very poor close to the spectral gap .
the second model we want to discuss consists of coupled spin-1/2 antiferromagnetic chains .
their hamiltonian is h_afm = j_n,*r*(_n,*r * _ n + 1,*r * ) + j__n,*r*,*a * _ n,*r*_n,*r + a * , [ hb ] where we take @xmath25 and * a * are lattice vectors in transverse directions . for simplicity we assume the transverse coupling to be isotropic in @xmath26 and @xmath27 directions .
taking the antiferromagnetic order ( which we assume to be along the @xmath28 direction ) into account by a mean - field analysis of the interchain interaction one obtains the effective single - chain problem @xcite h_0 = j_n_n_n + 1 - h_n(-1)^ns_n^z-2nj_m_0 ^ 2 , [ stchain ] where @xmath29 and @xmath30 .
the remaining interchain interactions will be treated in rpa ( see section vi ) .
we shall discuss the spin - peierls case of weak dimerization @xmath31 . in this case
one can use a continuous description of the spin-1/2 heisenberg chain which , for @xmath32 @xcite , is given by the gaussian model . in the framework of this model one can express spin operators in terms of bosonic exponents ( see _ e.g. _ @xcite ) .
thus , for instance , we have ( x ) = ( ) + .
[ boson ] the value @xmath33 was found by nakano and fukuyama @xcite by using the jordan - wigner transformation with a subsequent bosonization .
there is a simplier way to establish this value of @xmath34 , namely , we can use the fact that the initial hamiltonian has an su(2 ) symmetry which must be respected by the bosonized form . replacing @xmath35 in eq.([ha ] ) by the order parameter , that is treating it as a number and using eq.([boson ] ) we obtain the sine - gordon model .
since the su(2 ) symmetry in the sine - gordon model is not present for general values of @xmath34 its requirement imposes a restriction on the value of @xmath34 .
the corresponding point in the sine - gordon spectrum was discovered by coleman @xcite and haldane @xcite who pointed out that at @xmath33 there are only two breathers ; the first one has the same mass as kink and antikink ( let us call it @xmath36 ) and the second has the mass equal to @xmath37 .
therefore at @xmath33 kink , antikink and the first breather realize an su(2 ) triplet and the second breather becomes an su(2 ) singlet . in what follows
we also need the bosonized form of the spin density in the continuum ( @xmath38 ) , which is given by ( x ) & = & _ r ( x ) + _ l(x ) + ( -1)^n ( x ) , j^z_r , l&=&(_x),j^+_r , l= ( i ( ) ) , n^z ( x ) & = & - ( x ) , n^(x ) = .
[ boso ] here @xmath39 are the components of the staggered magnetization , @xmath40 are the currents of left and right moving fermions , @xmath41 and @xmath42 is a nonuniversal coefficient related to the bandgap for the charge excitations in the itinerant electron model that gives rise to the spin hamiltonian .
the field @xmath43 is the dual of the scalar field @xmath44 and obeys @xmath45 , where @xmath46 is the canonical conjugate of @xmath44
. we note that ( [ boso ] ) differ from the `` usual '' expressions ( see p. 270 - 271 of @xcite ) by a shift of the bosonic field by @xmath47 .
this operation interchanges @xmath48 and @xmath49 , but changes neither derivatives of @xmath44 nor the dual field @xmath43 .
notice that despite the fact that @xmath50 has the same dimension as the z - component of the staggered magnetization , it is given by a different operator ( sine instead of cosine ) .
in fact @xmath50 is the @xmath51-component of the charge density in the system with a frozen charge field .
as we shall see the situation is somewhat similar to that for the spin - ladder ( see for example @xcite ) , but there are also certain subtle differences . substituting ( [ boson ] ) into eqs.([h1 ] ) we get the following bosonized version of the spin - peierls hamiltonian @xmath52 h_*r*^(0 ) & = & x[_*r*^2 + ( _ x_*r*)^2 ] + x ( _ * r * ) -n_0 ^ 2 ( j_a+j_b ) , [ sg ] + v_phonon & = & _ * r*,*r*x:(_*r*(x)):j(*r*- * r*):(_*r*(x)):[vphonon ] + v_spin & = & _ * r*,*c * x k(*c * ) _ * r*(x)_*r+c*(x ) , [ vspin ] where @xmath53 .
note that we have kept only the most relevant terms and neglected _
e.g. _ the umklapp term ( for a discussion of the role of the umklapp term see @xcite ) .
our strategy is now to use exact results for the sine - gordon hamiltonian ( [ sg ] ) describing a single chain ( see also @xcite ) and to treat the residual interchain interaction @xmath54 in the random phase approximation ( rpa ) .
the hamiltonian ( [ sg ] ) describing a single chain is of the form @xmath55 .
the spectrum of @xmath56 is well - known @xcite : it is described in terms of a soliton , an antisoliton , a bound state ( called `` breather '' ) of mass @xmath36 and a second breather of mass @xmath57 .
the mass gap @xmath36 of the model is due to the dimerization caused by the coupling to the phonons .
it is related to the scale @xmath58 as follows @xcite = ^ .
the ground state energy density of @xmath56 is then given by @xmath59 , which in turn yields an expression for the ground state energy density of @xmath60 as a function of @xmath61 .
minimization with respect to @xmath61 yields the following mean field expressions for @xmath61 and mass gap @xmath62 of @xmath63 as functions of the dimensionless ( nonuniversal ) parameter @xmath64 and the couplings @xmath20 and @xmath12 _ 0&= & ( ) ^ ( ) ^3 ( ) ^2 ( ) ^= : c()^ , & = & v m= ( ) ^3 ( ) ^2 = : c^ .
[ gap ] here we have used @xmath65 for the fermi velocity , where @xmath66 is a function of the n.n.n .
coupling @xmath67 .
the ratio of the constants @xmath68 and @xmath69 is found to be = 0.175013 . [
ratio ] equation [ gap ] makes it clear that the gap originates from the interchain interactions .
in this section we derive exact results for various correlation functions of the sine - gordon model ( [ sg ] ) for @xmath71 .
we start by constructing a convenient basis of states for the sine - gordon theory by means of the zamolodchikov - faddeev algebra .
this is based on the knowledge of the exact spectrum and scattering matrix of the model @xcite .
we then formulate the problem of calculating correlation functions in terms of _ formfactors _ and finally give explicit results for the first few terms in the formfactor expansion .
the zamolodchikov - faddeev ( zf ) algebra for the sine - gordon model with @xmath70 was derived by affleck @xcite who suggested a representation which manifestly respects the su(2 ) symmetry . as mentioned above
there are three single particle states with mass @xmath36 which form a triplet under the @xmath72 symmetry .
the corresponding creation and annihilation operators are denoted by @xmath73 ( @xmath74 ) . here
@xmath75 denotes the breather state and @xmath76 denote soliton and antisoliton states respectively .
in addition there is one single - particle breather state with mass @xmath77 , which transforms as a singlet under @xmath72 .
its creation and annihilation operators are denoted by @xmath78 .
as usual the eigenstates are parametrized by a rapidity variable @xmath79 such that their momentum and energy are equal to p_j = m_j_j , _ j = m_j_j , where @xmath80 for the singlet state and @xmath81 for the triplet states . by definition
the zf operators ( and their hermitean conjugates ) satisfy the following algebra _
a(_1)z_b(_2 ) & = & s_a , b(_1 - _ 2)z_b(_2)z_a(_1 ) , a , b=,1_a(_1)z_2(_2 ) & = & s_a,2(_1 - _ 2)z_2(_2 ) z_a(_1 ) , a=,1z_2(_1)z_2(_2 ) & = & s_2,2(_1-_2)b(_2)b(_1 ) , [ fz1 ] where the two - particle scattering matrices @xmath82 are given by s_a , b()&=&= : s_0 ( ) , a , b=,1s_a,2()&=&s_0(+)s_0(- ) , a=,1s_2,2()&= & ( ) ^3 . [ smat ] for the creation and annihilation operators we have _
a(_1 ) z_b^+(_2 ) & = & s_0(_2 - _ 1 ) z_b^+(_2 ) z_a(_1 ) + 2_ab(_1 - _ 2 ) , a , b=,1_a(_1)z_2^+(_2 ) & = & s_a,2(_2 - _ 1)z_2^+(_2 ) z_a(_1 ) , + z_2(_1)z_2^+(_2 ) & = & s_2,2(_2 - _ 1)z_2^+(_2)z_2(_1 ) + 2(_1 - _ 2 ) .
[ fz2 ] from ( [ smat ] ) it follows that @xmath83 and @xmath84
. therefore particles with close momenta behave like free fermions and particles far apart in momentum space behave like free bosons .
we note that the soliton s - matrix @xmath85 has simple poles at @xmath86 and @xmath87 .
these poles correspond to the two breather bound states .
in other words a soliton and an antisoliton can form a bound state of either mass @xmath36 or mass @xmath77 .
although the s - matrix of light breathers is the same only the pole @xmath86 corresponds to a bound state
the heavy breather .
the pole at @xmath87 is _ redundant _ @xcite .
the soliton - breather s - matrices @xmath88 and @xmath89 exhibit soliton poles at @xmath90 and @xmath91 respectively .
all other poles are redundant and do not indicate the presence of bound states . > from the analytic properties of the s - matrices ( [ smat ] ) we deduce relations between the zf operators .
for example we find z_2()=__1-_2 z_1(_2 ) z_1(_1 ) .
such relations between the zf operators play an important role in what follows .
states in the fock space are constructed by acting with the operators @xmath92 on the vacuum state @xmath93 |_n _1__n
_1 = z^__n(_n) z^__1(_1 ) , [ states ] where @xmath94 .
we note that together with implies that states with different ordering of two rapidities _ and _ indices @xmath95 are related by multiplication with 2-particle s - matrices |_n
_k_k+1 _1 _ _ n _k_k+1
_1 = s__k,_k+1(_k-_k+1 ) the resolution of the identity is given by 11=_n=0^__i |_n _1__n
_1 ^_1 _n_1 _n| .
[ i d ] the formfactor approach is based on the idea of inserting between the operators in a correlation function ( x , t)o^(0,0)=_n=0^__i ( i_j=1^n p_jx-_jt ) |0| o(0,0)|_n _1__n
_1|^2 , and then determining the formfactors f^o(_1 _n)__1 _n:= 0| o(0,0)|_n _1__n _1 by taking advantage of their known analytic properties . from a physical point of view
we are interested in the fourier transforms of the connected retarded 2-point correlators of @xmath96 and @xmath97 .
their formfactor expansions are of the form d^cos(,q)&= & _
-^x_0^t e^i ( + i)t - ix & = & -2_n=0^__i |f^cos(_1 _n)__1 _n|^2 - .[dcos ] here we have reinserted the fermi velocity @xmath98 and lattice spacing @xmath99 .
the fourier transform @xmath100 of the connected retarded 2-point correlator of @xmath97 is the dynamical staggered susceptibility and will also be denoted by @xmath101 . in order to implement the formfactor expansion it is very useful to note that ( like for general values of @xmath34 ) operators from different representations behave differently under the charge conjugation transformation cc^-1 & = & - , + c z_()c^-1 & = & z _ ( ) , c z_1()c^-1 = - z_1 ( ) , + c z_2()c^-1 & = & z_2 ( ) .
[ chargeconj ] these transformation properties imply the following expansion & & |0= f_1^ - im(t - x ) z_1^()|0&&+ ^ -im[t(_1 + _ 2)-x(_1 + _ 2 ) ] & & u(_1,_2 ) [ z_^+(_1 ) z_-^+(_2 ) - z_-^+(_1 ) z _ ^+(_2 ) ] + ... where @xmath102 .
significantly , due to the su(2 ) symmetry transverse components of the staggered magnetization have the same correlation functions as @xmath103 ( this is clear from the su(2 ) symmetry of the hamiltonian ( [ sp ] ) ) .
consequently we conclude that sine - gordon 2-point correlation functions ( for @xmath104 ) of @xmath105 and @xmath106 are the same as 2-point correlators of @xmath97 .
the current operator @xmath107 is also odd in @xmath44 and therefore its expansion must begin with @xmath108 such that at small @xmath109 we have ( , q)(- , - q)~ + ... [ small ] where dots denote terms which have nonzero imaginary parts at higher energies . as we will see the threshold of the dynamical spin susceptibility is equal to @xmath36 for both @xmath110 and @xmath111 .
this is a distinct feature of the alternating chain .
it is related to the fact that kink and antikink create a bound state of the same mass .
recall that for the ladder chain ( or s = 1 antiferromagnet for this matter ) where particles do not have bound states , the value of the energy threshold at @xmath110 is twice that at @xmath111 ( see ref.@xcite ) . at frequencies smaller than @xmath112
the only contributions to the imaginary part of the magnetic susceptibility come from the first breather and kink - antikink pairs .
kink - antikink form factors can be calculated in the sine - gordon model ( for any value of the coupling @xmath34 ) along the following lines @xcite .
let us denote by @xmath113 ( @xmath114 ) the @xmath115-matrix eigenvalue corresponding to positive ( negative ) @xmath116-parity obtained by diagonalising the kink - antikink scattering : s_+=s_0 ( ) , s_-= s_0 ( ) , where @xmath117 and s_0 ( ) = - ( -i_0^ ) . then
, general unitarity and crossing arguments imply that the corresponding kink - antikink form factors @xmath118 are solutions of the following system of functional equations f_(-)&=&s_()f_()[23 ] + f_(-2)&=&f_(-).[24 ] [ 2partfun ] the `` minimal '' solutions of these equations are f_+()&= & f_0 ( ) + f_-()&= & f_0 ( ) , where @xmath119 is given by f_0 ( ) = ( _ 0^ ) . by minimal solution we mean a solution containing only the expected bound state poles in the physical strip and with the mildest asymptotic behaviour at infinity .
this prescription determines the minimal solution _ uniquely_. an infinite number of non - minimal solutions corresponding to all operators in the theory which are local with respect to the solitons
are obtained multiplying the minimal solution by an analytic function of @xmath120 . however , if we require the form factor to be power bounded in the momenta and to have only the bound state poles , we conclude that we can actually multiply the minimal solution only by a polynomial in @xmath120 . for a given operator , it is possible to put strong constraints on the asymptotic behaviour of its form factors , and then on the degree of the allowed polynomial @xcite . in the sine - gordon model
this procedure is complicated by a non - trivial behaviour of correlators in the ultraviolet limit .
nevertheless , the result is that for the operators cos and sin the allowed polynomial is of the zero degree , which means that their form factors coincide with the minimal ones .
the same conlusion can be reached in a simpler way going to the free fermion point @xmath121 , where the form factors of sin and cos can be easily computed remembering that ~| , _
^~j _ , and that the sin is related to the elementary field by the equation of motion . for the operators @xmath122 and @xmath123 , at the specific value of the coupling we are interested in , we find ( @xmath124 ) 0||_1,_2_-+ = ( 2d)z^1/2 ( _ 12 ) [ sin ] = f^sin()_-+ + 0||_1,_2_-+ = ( 2d)z^1/2 ( _ 12 ) [ cos]=f^cos(_12)_-+ + ( ) = c/2 \{2_0^x } c=(12)^ \{_0^x } 3.494607,d=0.136629 , where the relative normalisation between the two operators can be fixed exploiting the asymptotic factorisation of form factors discussed in ref .
we note that @xmath125 is to be analytically continued using the relation ( ) s_0 ( ) = ( - ) .
the additional factors @xmath126 and @xmath23 in and have been introduced in order to simplify the reduction of multiparticle formfactors using the annihilation - pole condition ( for soliton formfactors ) & & iresf^o(_1
_2n)__1 _2n |__2n-_2n-1=i = f^o(_1 _2n-2)__1 _2n-2 _ _ 2n^-_2n-1 & & \ { _ _ 1^_1 __2n-1^_2n-1 - s^_2n-1,_1__1,_1(_2n-1-_1) s^_2n-3,_2n-2__2n-1,_2n-2 ( _ 2n-1-_2n-2 ) } , [ annpole ] where @xmath127 and all other components are zero .
multiparticle formfactors are discussed in some detail in appendix * a*. the form factor ( [ sin ] ) has a pole at @xmath128 corresponding to formation of a bound state - the first breather .
the breather formfactor @xmath129 is given by the residue of ( [ sin ] ) divided by the three - particle coupling : |f_1|^2 = ( -2_0^ ) z 0.0533 z . similarly ( [ cos ] ) has a pole at @xmath130 corresponding to the second breather .
the absolute square @xmath131 of the breather formfactor is found to be |f_2|^2 = ( -4_0^ ) z 0.0262 z .
the expression for the imaginary part of the dynamical staggered susceptibility @xmath132 at @xmath133 is given by m(,q)= 2|f_1|^2(s^2 - m^2 ) + 2 , [ nearpi ] where @xmath134 .
note that all other formfactors do not contribute to this expression in the specified range of @xmath135 as their thresholds are above @xmath136 .
also the normalization @xmath137 enters only as an overall factor .
since the function @xmath125 vanishes at @xmath138 , the entire formfactor is also finite .
thus the two - particle contribution to @xmath139 exhibits a square - root singularity at the threshold as a function of @xmath135 .
the breather and @xmath140 contributions to the real part are found to be ( , q)= - - 2_0^ |f^sin()_+-|^2 , [ reals ] where the factor of @xmath141 stems from the sum over @xmath142 and @xmath143 . in fig .
[ fig : chipp ] we plot both the imaginary and real parts of @xmath144 = 0.45(a ) = 0.45(b ) it is straightforward to repeat the above analysis for the current operator @xmath107 using the explicit expressions for the formfactors given in @xcite . the contribution of the first breather leads to to with some normalization factor . as there is very little spectral weight at @xmath145
we concentrate on @xmath146 and do not repeat the above analysis for the current operator . for practical purposes
it is convenient to have an expression interpolating between the small @xmath109 ( [ small ] ) and @xmath146 ( [ nearpi ] ) behaviour .
such an expression giving the dynamical spin susceptibility in the entire range of @xmath109 at frequencies below the continuum may look like ( , q ) = + , where @xmath147 is a smooth function interpolating between the normalizations at @xmath148 and @xmath149 .
the mode @xmath150 is separated from the particle continuum by the gap of order of @xmath36 .
let us now turn to the two - point correlator of cosines .
the contributions of the second breather and the soliton - antisoliton continuum are given by m d^cos ( , q ) = 2|f_2|^2(s^2 - 3m^2 ) + 2 , [ secbr ] where the ratio of the single particle residues is universal : = = ( - _ 0^x ) 0.49131 .
[ univ ] note that the threshold of the breather - breather continuum is also at @xmath151 .
the corresponding contribution is taken into account in appendix * a*. the analogous contributions to the real part of @xmath152 are given by d^cos(,q)= - -2_0^|f^cos()_+-|^2 .
[ realc ] the remaining integrals in , , and have to be calculated numerically .
we find that at small @xmath135 the contributions of the two - particle continua to the real parts of both correlators are of the same magnitude as the single - particle contributions from the breather states .
as far as a single chain is concerned a single - mode approximation taking into account only the one - particle states is therefore very poor at small @xmath135 .
= 0.45(a ) = 0.45(b )
let us now take into account the interchain interactions ( both of spin and staggered energy densities ) in ( [ h1 ] ) .
this is accomplished through an rpa analysis along the lines of @xcite .
rpa becomes exact in the limit of an infinite number of neighbouring chains . in the rpa
we obtain the following expression for the correlation function of energy densities _
( s,)=(- , -q ; - ) ( , q ; ) = [ rpa ] where @xmath153 $ ] .
similarly the dynamical staggered susceptibility is given by ^zz(s,)=s^z(- , -q ; - ) s^z ( , q ; ) = , [ rpaspin ] where @xmath154 $ ] .
note that in the present approximation @xmath155 is only affected by the interchain interactions of staggered energy densities whereas @xmath156 only `` sees '' the interchain interactions of spin densities .
the reason for this decoupling is that in the sine - gordon theory describing the individual chains ( t , x ) s^z(0,0)=0 , because @xmath157 is even under charge conjugation whereas @xmath158 is odd . by rotational invariance
this implies @xmath159 .
note that the rpa is particularly simple as we have taken into account only interchain interactions of the staggered part of the spin density and neglected the smooth part as being less relevant .
if we take these subleading terms into account the rpa acquires a matrix structure like in @xcite as the sectors @xmath146 and @xmath145 become coupled .
an rpa analysis then requires the calculation of formfactors of the current operator .
we will discuss this refined rpa in a separate publication @xcite .
the response functions @xmath160 and @xmath155 can be easily calculated numerically for given values of @xmath20 and @xmath161 by using the expressions for @xmath162 and @xmath163 obtained in section [ sec:1d ] .
we note that the pole in the dynamical magnetic susceptibilities @xmath160 corresponds to the light breather which has quantum number @xmath164 , whereas the pole in @xmath155 is due to the heavy breather with @xmath165 .
the magnetic mode can be measured directly by neutron scattering whereas the @xmath165 excitation can be probed by measuring the phonon spectrum which will exhibit a softening . in order to visualize our results
we now plot them for a particular choice of parameters .
being aware that our theory probably can not be applied to @xmath2 where @xmath166 we nevertheless choose these parameters to reproduce the dispersions of magnetic excitations in that material .
we take @xmath167 , @xmath168 and @xmath169 , @xmath170 .
the value of @xmath66 is chosen such that the dispersion in @xmath28-direction reproduces the experimental fit of @xcite , whereas the other conditions follow from the experimental band - gaps for @xmath171-vectors @xmath172 ( @xmath173 ) , @xmath174 ( @xmath175 ) and @xmath176 ( @xmath177 ) . because of lorentz - invariance the energy @xmath178 and the z - component of the momentum @xmath109 only enter in the combination @xmath179 . in fig .
[ fig : dispy_sp ] ( a ) we plot the spin - wave dispersion in @xmath26-direction ( @xmath180 , @xmath181 ) and in fig . [ fig : dispy_sp ] ( b ) in @xmath26-direction ( @xmath182 , @xmath183 ) .
= 0.45(a ) = 0.45(b ) we see that the single - mode approximation ( sma ) in which all multiparticle contributions to the dynamical susceptibilities are neglected gives essentially the same result as the exact treatment .
we note that ( by construction ) the fits @xcite to the experimental results are essentially identical to the sma as far as dispersion relations are concerned .
let us now turn to the multiparticle continuum .
the imaginary part of the dynamical staggered susceptibility is directly measurable by neutron scattering .
the position of its poles yields the dispersion discussed above .
the incoherent part ( as a function of @xmath135 and @xmath184 ) is plotted in fig .
[ fig : imchix_sp ] ( a ) and ( b ) respectively .
= 0.45(a ) = 0.45(b ) in the rpa the 2-particle continuum starts at @xmath185 .
this is in disagreement with experiment for @xmath2 .
this may be because @xmath166 or because the dispersion in @xmath22-direction is rather strong , which in turn forces @xmath36 to be large in order to get a reasonable fit to the experimental data .
in the continuum limit the model is also equivalent to the sine - gordon model with @xmath70 @xcite .
the bosonization formulas are now ( x ) & = & _ r ( x ) + _ l(x ) + ( -1)^n ( x ) , j^z_r , l&=&(_x),j^+_r , l= ( i ( ) ) , n^z ( x ) & = & - ( x ) , n^(x ) = .
[ boso2 ] since the staggered magnetization is proportional to @xmath48 the bosonized single - chain hamiltonian is given by h = x - x - 2n j_m_0 ^ 2 .
following through the same steps as in the spin - peierls case we find the following mean - field results for staggered magnetization and mass gap m_0=c||^ , m = c^|j_| , where the ratio of @xmath68 and @xmath69 is given by ( [ ratio ] ) .
this relation can be used to determine the transverse coupling @xmath186 in terms of @xmath12 and the directly measurable quantities @xmath36 and @xmath187 as follows .
the gap @xmath36 is equal to @xmath188 which is found experimentally to be @xmath189 in @xmath190 @xcite .
the average magnetic moment is @xmath191 @xcite and @xmath192@xcite . using these values we find |j_|=()^20.96 mev .
let us now turn to the correlation functions for a single chain .
the correlator @xmath193 at small @xmath109 is still given by eq.([small ] ) , but for @xmath194 is given by eq.([secbr ] ) .
therefore around @xmath111 the pole is at @xmath195 ^zz ( , q ) d^cos(s)= + , [ afmczz ] where `` incoherent '' denotes multiparticle contributions .
some of these are determined exactly in appendix a. a plot of @xmath163 is shown in fig.[fig : cos ] .
correlation functions of transverse components of the staggered magnetization are given by eq.([nearpi ] ) and have a pole at @xmath196 ^xx ( , q)d^sin(s ) = + .
[ afmcp ] here we have used the fact that in the continuum model the correlation functions of @xmath106 , @xmath105 and @xmath97 are equal due to the su(2 ) symmetry present at @xmath104 . a plot of @xmath162 is shown in fig .
[ fig : chipp ] .
the difference in the correlation functions ( [ afmczz ] ) and ( [ afmcp ] ) is obviously related to the broken rotational symmetry of the hamiltonian ( [ stchain ] ) . next we take into account the interchain interaction by an rpa analysis .
this yields the following expression for the longitudinal dynamical susceptibility ^zz ( , q , ) & = & ^xx ( , q,)&= & .
here we again have taken into account only the staggered part of the spin density as it gives the most relevant contribution to the interchain interaction . as a result
the rpa expression for the susceptibilities are of scalar rather than matrix form @xcite .
the transverse susceptibility must have a pole at the neel wave vector @xmath197 as the spin @xmath72 symmetry is spontaneously broken .
this leads to the requirement that d^sin(0)=0.12509 , which fixes the normalization @xmath137 in terms of the transverse coupling and the breather mass as @xmath198 .
the normalization of the correlator of cosines then follows to be d^cos(0 ) .
the goldstone mode associated with the zero energy pole in @xmath199 is a spin - wave moving in @xmath28-direction and its dispersion is found from the singularities of @xmath200 . due to lorentz invariance @xmath201
depends only on @xmath135 rather than on @xmath178 and @xmath109 independently .
this immediately implies that the spin - wave dispersion for @xmath202 is ^2(q , ) = ( -q)^2+m^2(1-)(-q)^2 + m^2(1- ) .
[ swdisp ] this is in very good agreement with experiment being almost identical to the fit used in @xcite . in fig .
[ fig : chiperp ] ( a ) we plot the spin - wave dispersion for @xmath180 and @xmath181 .
we see that the sma works extremely well for all values of @xmath203 .
the imaginary part of the dynamical susceptibility is directly measurable by neutron scattering .
we find m^(,q , ) = ( -1+)+incoherent , where we have used the sma to get the delta - function part . the incoherent part is plotted in fig .
[ fig : chiperp ] ( b ) .
we see that there is in general no singularity at the threshold of the light breather - heavy breather continuum except at @xmath204 where @xmath205 so that we recover the pure 1-d result .
the situation for the soliton - antisoliton continuum is analogous .
let us now turn to the longitudial susceptibility . in the sma
there is a pole in @xmath206 at ^2(q,)=(-q)^2 + m^2[3-(k_x+k_y ) ] , [ disp ] where @xmath207 is given by .
this is compared to the exact result for the case where @xmath180 in fig .
[ fig : dispx ] ( a ) .
we see that the corrections to the sma result are very small . using the sma ( which we know from fig .
[ fig : dispx](a ) to be an excellent approximation ) to extract the coherent delta - function part we find m^zz(,q , ) = ( -3+(k_x+k_y))+incoherent . the incoherent part is plotted in fig .
[ fig : dispx](b ) .
.5 cm .5 cm in this appendix we consider multiparticle formfactors .
we start with 2-soliton 2-antisoliton formfactors and explicitly derive the related three and two - particle formfactors .
the extension to @xmath9-soliton @xmath9-antisoliton formfactors ( @xmath208 ) is straightforward and will not be discussed here . the formfactor expansion for 2-point correlation functions
is found to be rapidly converging and for small @xmath135 it is essentially sufficient to take into account 2-particle formfactors only .
note that most of the formulas below are to be understood in terms of analytic continuation of @xmath125 .
a useful formula is |(+i)(-i)|^2&=&(-)^2 \{4_0^x ( 1-x ) } & & .
multiparticle formfactors of ( quasi)local operators were studied in detail by smirnov @xcite . from his work the 2-soliton 2-antisoliton formfactor for @xmath96 is straightforwardly extracted & & f^cos(_1,_2,_3,_4)_++ = 0||_1,_2,_3,_4 _ ++ = & & 2(2d)^2 3(_l=1 ^ 4 e^_l ) ( _ m=1 ^ 4 e^-_m ) _
je^_i+_j_i < k(_i-_k ) e^-_j_j & & ( _ 3+_4-_1-_2 ) _
i=1 ^ 2_j=3 ^ 4 .
[ cos4 ] orderings other than @xmath209 are obtained from by using the generalization of f^cos(_1,_2,_3,_4)_-+-+ = s_0(_3-_2)f^cos(_1,_3,_2,_4)_++ .
it is easy to verify that the soliton - antisoliton formfactor of @xmath96 is obtained from ( [ cos4 ] ) _ via _ the annihilation pole condition ( [ annpole ] ) .
the 2-soliton 2-antisoliton formfactor for @xmath97 is very similar to the one for @xmath96 f^sin(_1,_2,_3,_4)_++ = -(_3+_4-_1-_2 ) f^cos(_1,_2,_3,_4)_++ .
the residue at the annihilation pole ( times @xmath210 ) now yields the soliton - antisoliton formfactor of @xmath97 .
breather formfactors are obtained from the residues of at its poles . in the soliton - antisoliton - even breather sector we find f^cos(_1,_2,_3)_-+2 & = & -2 ( _ 12 ) ( - ) ( _ 13 + ) & & ( _ 13- ) ( _ 23 + ) ( _ 23- ) [ e^_1+e^_2+e^_3 ] & & [ e^-_1+e^-_2+e^-_3 ] [ e^_1+_2+e^_3(e^_1+e^_2)+e^2_3 ] .
the corresponding formfactor for @xmath97 is f^sin(_1,_2,_3)_-+2 = -f^cos(_1,_2,_3)_-+2 , and different orderings are obtained by the appropriate generalization of f^cos(_1,_2,_3)_2-+ = f^cos(_2,_3,_1)_-+2 s_1,2(_21)s_1,2(_31 ) .
the residue at the annihilation pole ( times @xmath210 ) in @xmath211 gives the heavy breather formfactor @xmath212 .
the corresponding @xmath213 formfactor has no annihilation poles . in the soliton - antisoliton - odd breather sector
we obtain f^cos(_1,_2,_3)_-+1 & = & -2 ( _ 12 ) ( - ) ( _ 13 + ) & & ( _ 13- ) ( _ 23 + ) ( _ 23- ) [ e^_1+e^_2+e^_3 ] & & [ e^-_1+e^-_2+e^-_3 ] [ e^_1+_2+e^_3(e^_1+e^_2)+e^2_3 ] & = & f^sin(_1,_2,_3)_-+1 .
[ pm1 ] from the residues at the poles of ( [ pm1 ] ) we can derive the breather - breather formfactors : f^cos(_1,_2)_11 & = & -2 & & ^2(_21)^2(- ) ( _ 21 + ) ( _ 21- ) s_0(_21+)s_0(_21- ) .
this is identical to the soliton - antisoliton formfactor @xmath214 as can be proved by direct calculation .
some useful identities are @xmath215 , @xmath216 and ( _ 0^)= .
the other breather - breather formfactors are given by f^cos(_1,_2)_22 & = & -2 & & ^2(_12)^2(- ) ( _ 12+)(_12-),f^sin(_1,_2)_21 & = & -2 & & ( -)(- ) ( _ 12+)(_12- ) ( _ 12+)(_12- ) .
the special values of @xmath217 at the breather poles are given by @xmath218 and @xmath219 .
we note that @xmath220 .
it is apparent that @xmath9-particle formfactors depend only on @xmath221 independent rapidity variables .
this fact can be used to essentially simplify expressions like for correlation functions .
for example , 2-particle contributions for @xmath222 are given by d^cos(,q)|_2-part.&=&- _ 0^ + , d^sin(,q)|_2-part.&=&- _ 0^ + where we also have made use of various symmetry properties of the formfactors in order to perform the sum over @xmath223 .
similarly the contribution of formfactors involving one soliton , one antisoliton and one ( light ) breather of type 1 can be brought to the form -_-^ _
-^ 99 m. hase , i. terasaki and k. uchinokura , .
pouget , l.p .
regnault , a. ain , b. hennion , j.p .
renard , p. veillet , g. dhalenne and a. revcolevschi , .
k. hirota _
et al _ , . l.p .
regnault , a. ain , b. hennion , g. dhalenne and a. revcolevschi , .
satija , j.d .
axe , g. shirane , h. yoshizawa and k. hirakawa , .
tennant , s.e .
nagler , s. welz , g. shirane and k. yamada , , + d.a .
tennant , r. cowley , s.e . nagler and a.m. tsvelik , .
a. w. garrett , s. e. nagler , d. a. tennant , b. s. sales , t. barnes , cond - mat/9704092 .
d. c. dender _ et al _ , preprint cond - mat/9704034 .
k. hirota , b shirane , q. j. harris , q. feng , r. j. birgeneau , m. hase , and k. uchinokura , phys .
rev . b*52 * , 15 412 ( 1995 ) .
i. affleck , nucl .
b*265 * ( 1986 ) 448 .
s. coleman , ann .
( ny)*101 * ( 1976 ) 239 .
a.m. tsvelik , _
quantum field theory in condensed matter physics _ , cambridge university press ( 1995 ) .
t. nakano and h. fukuyama , j. phys .
* 50 * ( 1981 ) 2489 .
f. d. m. haldane , phys .
b*25 * ( 1982 ) 4925 .
d. g. shelton .
a. a. nersesyan and a. m. tsvelik , phys . rev .
* b53 * ( 1996 ) 8521 .
r. julien and f.d.m .
haldane , bull .
* 28 * ( 1983 ) 34 , + s. eggert and i. affleck , phys
* b46 * ( 1992 ) 10866 .
a.m. tsvelik , .
uhrig and h.j .
zamolodchikov , .
v.e . korepin and l.d .
faddeev , theor . mat . phys .
* 25 * ( 1975 ) 1039 , +
r. dashen , b. hasslacher and a. neveu , phys .
* d11 * ( 1975 ) 3424 .
zamolodchikov , .
m. karowski , h .- j .
thun , t.t .
truong , p.h .
a.b . zamolodchikov and al.b .
zamolodchikov , annals of physics * 120 * ( 1979 ) 253 .
m. karowski and p. weisz , nucl . phys .
* b139 * ( 1978 ) 455 .
f. a. smirnov , _ form factors in completely integrable models of quantum field theory _ , world scientific ( 1992 )
. g. delfino and g. mussardo , nucl . phys .
* b455 * , 724 ( 1996 ) . g. delfino , p. simonetti and j.l .
cardy , phys . lett .
* b 387 * ( 1996 ) 327 .
h. schulz , phys .
* 77 * ( 1996 ) 2790 .
f.h.l . essler and a.m. tsvelik , _ in preparation_. | we study weakly coupled antiferromagnetic spin chains in their ordered phase by combinining an exact solution of the single - chain problem with an rpa analysis of the interchain interaction .
a single chain is described by a quantum sine - gordon model and dynamical staggered susceptibilities are determined by employing the formfactor approach to quantum correlation functions .
we consider both antiferromagnetic order encountered in quasi-1d materials like @xmath0 and spin - peierls order as found in @xmath1 .
12h # 1([#1 ] ) # 1#2#3_nucl .
phys . _ * b#1 * ( # 2 ) # 3 # 1#2#3_phys .
rev . _
* d#1 * ( # 2 ) # 3 # 1#2#3_phys .
rev .
_ * b#1 * ( # 2 ) # 3 # 1#2#3_phys .
rev .
lett . _ * # 1 * ( # 2 ) # 3 # 1#2#3_phys .
lett . _ * # 1b * ( # 2 ) # 3 # 1#2#3_theor . mat .
phys . _ * # 1 * ( # 2 ) # 3 # 1#2#3_sov .
phys .
jetp lett .
_ * # 1 * ( # 2 ) # 3 # 1#2#3_j .
phys . cond .
matt . _ * # 1 * ( # 2 ) # 3 # 1#2#3_int .
j. mod .
phys . _ * a#1 * ( # 2 ) # 3 pacs : 74.65.+n , 75.10 .
jm , 75.25.+z |
spatially resolved x - ray spectroscopy is essential for mapping the distribution of mass , intracluster plasma , and heavy elements in galaxy clusters .
although chandra and xmm - newton have provided a wealth of information on these quantities in the central regions of clusters , relatively little is known about their outer regions , at radii greater than @xmath15 . here
@xmath16 is the radius within which the mean cluster density is 200 times the critical cosmic density . in the simplest case of the collapse of a spherical density perturbation in an einstein - desitter universe
, it is expected that the dark and bayronic matter within @xmath16 will have reached dynamical equilibrium @xcite .
cluster emission from the vicinity of @xmath16 is difficult to observe efficiently with chandra and xmm - newton because of their relatively high ( and , in the latter case , time - variable ) levels of instrumental background .
the study of cluster outskirts is important for several reasons .
first , a direct x - ray measurement of total cluster mass ( e.g. , within @xmath16 ) requires knowledge of the icm temperature profile at this radius .
this is especially important to facilitate comparison between x - ray and weak - lensing mass profiles , as the latter are difficult to measure in the inner regions of clusters .
second , structure and cluster formation models make definite predictions about total mass and icm distributions to these radii @xcite , so measurements in this radial range offer a test of these models .
finally , to the extent that @xmath16 marks the approximate boundary of the dynamically relaxed region of a cluster , one might expect to find inhomogeneities in the icm that could provide a direct view of the accretion processes by which clusters grow .
rosat has been used to characterize the x - ray surface brightness of the outer regions ( @xmath17 ) of individual clusters @xcite and , using stacking techniques , of ensembles of objects @xcite .
chandra surface brightness profiles in the range @xmath18 for a sample of relatively distant clusters have been studied by @xcite .
recent suzaku observations have successfully probed out to and beyond @xmath16 for a handful of clusters @xcite . in this paper
we present suzaku measurements of the surface brightness and temperature of the intracluster medium of the x - ray luminous ( @xmath19 erg s@xmath5 ) cluster abell 1795 .
this is a relatively hot ( spatially - averaged @xmath20 kev ) cluster with a cool core @xcite .
although chandra shows a 40-arcsec - long ( @xmath21 kpc ) plume `` trailing '' the brightest cluster galaxy @xcite , outside of the very center the cluster is quite regular and appears relaxed . from the observed temperature , and scaling relations derived from simulations @xcite we estimate @xmath22 mpc , consistent with the chandra measurements of @xcite , and @xmath23 mpc ( @xmath24 ) for our assumed cosmologycdm cosmology with @xmath25 km s@xmath5 mpc@xmath5 , @xmath26 and @xmath27 .
one arc minute corresponds to 73 kpc at the cluster redshift ( @xmath28 ) in this cosmology .
unless otherwise noted , quoted errors correspond to 90% confidence intervals . ] .
we exploit the relatively low and stable suzaku background to trace the temperature and icm density to the vicinity of @xmath16 . as noted by @xcite , if the virial radius @xmath29 is derived from the collapse of a spherical top - hat density perturbation , assuming that the cluster has just virialized , then for an einstein - desitter cosmology @xmath30 , independent of redshift . in general , however , @xmath31 is function of cosmology and redshift , and for our @xmath32cdm cosmology at the redshift of abell 1795 @xmath33 , so the virial radius is considerably larger than @xmath16 . for abell 1795
we expect @xmath34 . here
@xmath31 is the ratio of the mean cluster mass density within @xmath35 to the critical cosmic density .
abell 1795 was observed on 2005 dec 10 , during the suzaku performance verification phase , with a series of 5 overlapping pointings , illustrated in figure [ fig1 ] .
the results reported here were obtained with xis @xcite data from version 2.0.6.13 of the suzaku data processing pipeline .
data obtained with 3x3 and 5x5 editing modes for each pointing were merged and standard filters ( requiring earth elevation greater than 5 degrees , sunlit earth elevation angle greater than 20 degrees , cutoff rigidity greater than 6 gev / c , and excluding data obtained during and just after passages through the south atlantic anomaly ) were applied .
events in detector regions nominally illuminated by the xis calibration sources were also excluded .
the net good exposure times after filtering were 10.2 ks , 23.9 ks , 24.6 ks , 27.3 ks and 38.3 ks for the central , near - north , near - south , far - north and far - south pointings , respectively .
thus the total good exposure time was 124.3 ks .
( 60mm,150mm)f1.eps smoothed images of each field were inspected by eye and regions centered on nine obvious point sources were excluded , as indicated in figure [ fig1 ] . to obtain a properly normalized cluster surface brightness profile from our multiple pointings , we used the following procedure to generate exposure maps .
first , simulated monochromatic photon lists representing a uniform , 20 radius extended source were constructed , using 10@xmath36 photons at a number of energies for each detector and pointing .
the energy bins were selected to represent roughly constant regions of the instrumental effective area curve .
each simulated photon list was used as input to the xrt ray - tracing simulator ( xissim ; @xcite ) , which includes the effects of the optical blocking filter ( obf ) contamination , producing an image for each detector , pointing , and energy combination . to increase signal - to - noise
, these flat field images were smoothed by a two dimensional gaussian kernel with 72 fwhm .
the calibration source regions and regions of low effective area near the chip edges were clipped .
spectral weights were determined from a spectral model fit to the cosmic x - ray background ( described in section [ sec : cbg ] below ) .
these spectral weights were chosen under the null hypothesis that surface brightness fluctuations in the cluster outskirts are due to cosmic background variations . when applied to cluster emission , they will result in an overestimate of the cluster surface brightness , however we estimate this to be less than 10% ( 5% ) in the 0.52 kev ( 28 kev ) range for the cluster temperatures and abundances measured in this work .
errors in surface brightness between cluster regions of different temperatures ( ranging from 26 kev ) will be less than 2% due to the spectral weighting .
for each chip and pointing , the flat field frames were combined using the spectral weighting to produce an exposure map .
finally , the individual exposure maps were mosaicked onto a common sky coordinate ( wcs ) reference frame and summed .
images from each pointing toward the cluster were extracted in the 0.52 kev and 28 kev bands , and these were mosaicked in sky coordinates .
cutoff - rigidity - weighted particle background rates were computed for each spectral band , sensor and pointing from the archive of ` night - earth ' data ( i.e. , data obtained during orbit night while suzaku pointed at the earth ; @xcite ) . a ( spatially uniform ) particle background component
was then subtracted from each cluster image , and each resulting image was then divided by the appropriate exposure map . the exposure - corrected , 0.52 kev mosaic image is shown in figure [ fig1 ] . for purposes of this figure ,
the background- and exposure - corrected image has been smoothed with an adaptive , circular kernel .
the kernel radius is chosen to include at least 25 counts , up to a maximum radius of 10 binned pixels ( @xmath37 ) .
pulse - height spectra were accumulated for each in a series of annular sectors 2.54 in width , depending on radius .
the spectral extraction regions are illustrated in figure [ fig1 ] .
two ancillary response functions ( arf files ) were constructed for each region and xis detector using the monte carlo ftool xissimarfgen v2008 - 04 - 05 @xcite .
the first arf was produced from a 20 radius uniform surface brightness source ; this response file was used to fit uniform background emission in the spectral analysis that follows .
the second arf was produced from the cluster surface brightness distribution results described in section [ sec : sb ] ; this response was used to fit cluster emission in each region .
these files properly account for energy - dependent vignetting and scattering properties of the suzaku xrt , as well as for the time - dependent contamination layer on each of the xis optical blocking filters .
corresponding detector redistribution functions ( rmf files ) were constructed using the ftool xisrmfgen v2007 - 05 - 14 .
cutoff - rigidity - weighted particle - induced background spectra were determined from night - earth data for each sensor and pointing using the ftool xisnxbgen v2008 - 03 - 08 .
these background spectra were subtracted from each source spectrum before fitting .
xspec v12.5.0 @xcite was used to fit data from all four xis sensors simultaneously for all spectral fits reported here .
spectral fitting was confined to the 0.558 kev band where the response calibration is best .
calibration observations directly constrain the time - dependent , on - axis absorption by contamination of the xis optical blocking filter in this spectral band @xcite .
the emission from each cluster region was modelled by simultaneously fitting the pulse - height spectra from the cluster region and two background regions .
the sum of a thin thermal plasma model ( mekal in xspec ) plus a cosmic background model ( described in section [ sec : cbg ] below ) was fit to each cluster spectrum , while only the background model was fit to the background regions .
in all fits the cluster heavy element abundance , relative to cosmic values @xcite , was allowed to vary .
the best - fit redshift determined from the central @xmath38 region ( @xmath39 ) is consistent with the optical cluster redshift ( @xmath40 ) reported by @xcite .
the redshift was fixed at the optical value for other spectral fits .
the combined xis0,2,3 ( fi ) spectra and best - fit model components for the central 5 and the northern extension are shown in figure [ fig : northspec ] .
( .8,.8)f2.eps spectra of the celestial x - ray background outside the cluster were obtained from two trapezoidal background regions at the northern and southern extremes of the observed field , 2.0 to 2.3 mpc from the cluster center ; see figure [ fig1 ] . the background spectral model for each fit
included the sum of absorbed power - law and soft thermal components , to account for extragalactic and galactic backgrounds , respectively .
when modelling the spectra of the faintest cluster regions ( those with @xmath41 ) we also included a suite of emission lines to account for possible contributions from geocoronal solar wind charge exchange ( swcx ) .
we find that the contribution of swcx emission is uncertain and , though small , may be significant in the interpretation of the emission detected in these regions . in this section
we describe the background model used for the brighter ( @xmath42 ) cluster regions , for which swcx is neglected .
we discuss our treatment of swcx emission in the cluster outskirts in the following section .
lccc component & north & south & joint + + power law @xmath43 & 11.5 @xmath44 & 10.3 @xmath45 & 10.9 @xmath46 + thermal @xmath47 & 0.15 @xmath48 & 0.19 @xmath49 & 0.18 @xmath50 + power law @xmath51 & @xmath527.7 @xmath46 & @xmath527.0 @xmath46 & @xmath527.3 @xmath53 + thermal @xmath51 & @xmath522.9 @xmath54 & @xmath523.0 @xmath55 & @xmath523.0 @xmath56 + total @xmath51 & 10.6 @xmath54 & 10.0 @xmath54 & 10.3 @xmath56 + + power law @xmath43 & @xmath57 & 10.1 @xmath54 & ... + thermal @xmath47 & @xmath58 & @xmath58 & ... + power law @xmath51 & @xmath527.6 @xmath59 & @xmath526.8 @xmath56 & ... + thermal @xmath60 & @xmath522.8 @xmath61 & @xmath522.8 @xmath62 & ... + swcx @xmath60 & @xmath521.5 @xmath63 & @xmath521.3 @xmath64 & ... + total @xmath65 & 11.9 @xmath66 & 11.2 @xmath54 & ... + best - fit spectral model parameters for the background regions , neglecting swcx emission , are shown in the upper half of table [ tab : fluxes ] .
consistent results ( within @xmath67% in normalization ) were obtained for each model component in the two background regions . since our spectral analysis regions are quite small , expected poisson fluctuations in the extragalactic background are actually larger than @xmath67% , so this level of agreement must be regarded as fortuitous . to allow for these fluctuations ,
the normalization of the corresponding spectral component was allowed to vary separately when fitting each region of the cluster .
this point is discussed in detail in section [ secbgva ] below .
the 210 kev surface brightness of the power - law component , determined from the joint fit to both the north and south background regions , is @xmath68 erg s@xmath5 @xmath6 deg@xmath7 , in excellent agreement with the value reported by @xcite from xmm - newton observations ( @xmath69 erg @xmath6 s@xmath5 deg@xmath7 ) . as a check on
the estimated 0.52 kev background , we determined the rosat all - sky survey ( rass ) surface brightness in an annulus centered on abell 1795 with inner and outer radii of 0.5 and 1 degree , respectively .
we used the x - ray background tool @xcite provided by nasa s goddard space flight center to extract the rosat spectrum for this region .
we then fit the recommended x - ray background model @xcite , which includes two soft thermal components and a power law , to the rosat spectrum in the 0.12 kev band .
this yields a surface brightness in the 0.52 kev band of @xmath70 erg s@xmath5 @xmath6 deg@xmath7 , in excellent agreement with our ( more precise ) suzaku estimates listed in table [ tab : fluxes ] .
thus , even ignoring possible contributions of swcx emission , our suzaku background estimates are in excellent agreement with previous measurements of both the galactic and extragalactic components . in our chosen spectral bands
we can determine background ( and foreground ) flux in the vicinity of abell 1795 with a relative statistical uncertainty of just over 5% . as we discuss in section [ secbgva ] below
, the accuracy of our background subtraction is in fact limited by cosmic variations in the density of unresolved background sources rather than by statistical or instrumental systematic errors .
heavy ions in the solar wind undergo charge exchange with neutral atoms in the earth s geocorona and in the solar magnetosphere , resulting in a field - filling emission line spectrum which contaminates the more distant cluster emission .
lines typically produced by solar wind charge exchange ( swcx ) include those of highly - ionized c , o , ne , and mg ( e.g. , @xcite ) .
variations in the solar wind can produce variations in the geocoronal swcx flux on timescales of seconds , while heliospheric swcx contributes a stable or slowly - varying x - ray background ( e.g. , @xcite ) .
we have checked for swcx variability by constructing light curves for each pointing , excluding the center pointing where the cluster emission is expected to dominate any swcx emission .
light curves were extracted for all four xis detectors in the 0.42 kev range from the full detector field , excluding point sources .
particle background light curves were constructed from the ` night - earth ' data archive @xcite using the appropriate cut - off rigidity time series , and these were subtracted from the individual detector light curves , which were then combined for each field .
the resultant complete light curve is shown in the top panel of figure [ fig : ace ] .
any residual variations in count rate are a result of variations in the emission from the galactic foreground , the unresolved extragalactic background , the cluster , and/or swcx .
the extraction regions for corresponding pointings are identical in sky coordinates , so any differences between far north and far south ( `` fn '' and `` fs '' ) or between near north and near south ( `` nn '' and `` ns '' ) must be due to variations ( temporal or spatial ) in the surface brightness of these emission components .
( .9,.9)f3.eps there is clear variability in the light curve .
most notably , the far north count rate shows a strong spike ( marked by the green arrow ) .
count rates in the other three pointings are also inconsistent with the null hypothesis of constancy ( p - values @xmath71 0.1 ) , although they lack strong variation .
analysis of public data from the advanced composition explorer ( ace ) satellite @xcite reveals an enhanced flux of protons and heavy ions during much of the observation ( see figure [ fig : ace ] , bottom panel ) .
although there are gaps in the ace data , the proton , c@xmath72 , and o@xmath73 flux during the far north and near north pointings ( near mjd 53716 ) are similar to levels observed during times of swcx contamination in a suzaku observation of the north ecliptic pole @xcite .
the far north count rate spike occurs near a peak in both the proton and heavy ion flux .
there is an additional spike in the heavy ion flux near the end of the near north observation .
the solar particle flux near the end of the observations ( after mjd 53716.5 ) is consistent with that observed during quiescence in other suzaku observations @xcite .
we conclude that our observations are likely contaminated by variable swcx emission .
we estimate the level and variability of geocoronal swcx contamination by comparing the emission from the ( presumed cluster - free ) background regions . as in section
[ sec : cbg ] above , the non - swcx background is modeled as the sum of a soft thermal ( galactic ) background component and a power - law ( extragalactic ) component .
when fitting the spectra of the relatively small background regions alone , we found strong degeneracies between the swcx and thermal model parameters , so we fixed the temperature of the thermal model at @xmath74 kev , the value determined from the joint swcx - free model .
the photon index of the power - law component was fixed at @xmath75 .
the normalizations of both background components were allowed to vary independently in each region ( see section [ secbgva ] ) . to model the swcx in the background regions ,
we include a suite of unresolved gaussian emission lines typically observed in swcx emission between 0.52 kev @xcite , listed in table [ tab : swcxlus ] .
the line energies are fixed to their theoretical values , using the energy of the dominant forbidden transition for helium - like lines of o vii , ne ix , and mg xi @xcite .
the line fluxes are allowed to vary independently .
.modeled swcx emission lines [ tab : swcxlus ] [ cols="<,^ , < , < " , ]
we use markov chain monte carlo ( mcmc ) techniques to constrain the relatively large number of model parameters and their uncertainties .
the mcmc tools in xspec v12.5.0 @xcite were used to produce 8 chains of 7000 steps , of which the first 2000 were excluded as `` burn - in '' . the input covariance matrix was scaled to get sufficient mixing ( i.e. , so that the fraction of repeated locations in parameter space was @xmath76 0.8 ) .
these 40,000 chain steps were combined to produce probability distributions for each of the 22 free parameters , and from these the best fit value and 90% confidence intervals were estimated .
both the far north and far south background regions show an excess of line emission above the galactic and extragalactic components .
the spectra in figure [ fig : swcxspec ] suggest enhancement of the o vii triplet near 0.56 kev ( evident in xis1 ) and a @xmath76 3-@xmath77 feature near 1.5 kev ( evident in xis0,2,3 ) in the far north spectrum .
the latter is consistent with mg xii ly@xmath78 .
the far south spectra show much weaker upper limits to the enhancement near 0.56 kev ( in xis1 ) and 1.5 kev ( in xis0,2,3 ) , but they contain excess emission in the 0.91.0 kev region ( evident in all detectors ) that is consistent with ne ix and ne x. the line strengths and upper limits are shown in table [ tab : swcxlus ] .
the mg xii line is similar in energy to a strong instrumental al k@xmath78 line , however the al k@xmath78 line flux would have to be enhanced by more than a factor of 3 during the full far north exposure to produce such strong residual emission .
background variations of this sort have not been reported in previous suzaku observations , nor in the night - earth background analysis of @xcite .
the presence of a mg xii swcx line without additional swcx lines such as o viii , ne ix and ne x seems puzzling , although since detection of mg xii with xmm - newton is complicated by a much stronger and more variable instrumental al k@xmath78 line , there are few data from other sources .
mg xii ly@xmath79 has likely been observed in another xmm - newton swcx event @xcite .
@xcite claim a mg xi swcx detection at 1.34 kev .
the upper limits on the photon fluxes for other expected lines are low compared to other swcx detections , which suggests that our observations are less affected by geocoronal swcx .
( , ) f4a.eps ( , ) f4b.eps the contributions of swcx and other model components to the total surface brightness in the background regions are listed in the lower half of table [ tab : fluxes ] .
the table shows that swcx emission is detected at better than 90% confidence during our observations of both far north and far south fields , and that the swcx emission spectrum differs significantly between the two observations .
however , in each field swcx contributes less than 15% of the total , soft - band cosmic background brightness , and the difference between swcx contributions to the two fields is less than 5% of the total cosmic background brightness .
our spectral analysis of the outermost regions of the cluster ( @xmath41 ) follows that used for the interior of the cluster , except that we include the swcx model described above in the background model ( see figure [ fig : northspec ] ) . in particular , the soft thermal ( galactic ) background component was fit with a single free temperature and normalization for all regions ( 17.5@xmath24 north and south , and the north and south background regions ) .
the power - law ( extragalactic ) background component was fit with a fixed photon index @xmath75 and normalizations that were allowed to vary independently in each region ( see section [ secbgva ] ) . a mekal component , with temperature , abundance , and normalization free to vary ,
was used to model cluster emission .
we wish to measure cluster emission at very low surface brightness levels and it is important to understand the influence of various instrumental and other systmatic effects that may limit our ability to do so . in this section
we consider a number of potential systematic errors in more detail , including i ) the impact of our choice of passband on temperature measurement accuracy ; ii ) the accuracy of our knowledge of spatial response variations over the xis field of view ; iii ) the impact of scattered light , both from the bright cluster core and from point sources in the field of view , on the accuracy of our spectral modelling ; and iv ) the effects of counting statistics on the flux of unresolved sources .
a drawback of our decision to ignore data at energies below 0.55 kev in our spectral fits is that it is difficult to constrain the galactic absorption directly from the x - ray data unless the signal - to - noise ratio is quite high .
on the other hand , we have found that the best - fit model temperature and galactic column do exhibit a mild ( anti ) correlation .
thus , any systematic errors in determination of the galactic column ( such as might arise from spatially varying errors in the contamination correction ) can induce errors in the measured temperature .
we have therefore elected to fix the model galactic column density at the value obtained by fitting to the high - signal - to - noise ratio spectrum of the integrated cluster emission at @xmath80 .
fitting results are shown in table [ tab2 ] .
the best - fit value of @xmath81 is in excellent agreement with the nominal galactic column ( @xmath82 @xmath6 ) estimated for this field by nasa s hearsarc tool nh .
ccccc flux@xmath83 & @xmath84 & abundance & @xmath85 & @xmath86/dof + @xmath87 & @xmath88 & @xmath89 & @xmath90 & @xmath91 + as a check on the reliability of the suzaku calibration , we compared fitting results for regions of the cluster observed with multiple pointings .
two such ` overlap ' regions provide sufficient flux for a meaningful comparison .
each region is an annular sector with inner and outer radii of @xmath92 and @xmath93 , respectively .
the northern overlap region was observed with central and ` near - north ' pointings ; the southern overlap region was observed with the central and ` near - south ' pointings .
thus each overlap region was observed on opposite sides of the xis field of view . in order to gauge errors in the calibration of the soft response , we allowed the galactic column to vary when fitting individual spectra . for each overlap region ,
the best - fit model parameters from the two pointings are consistent with one another within 90% confidence statistical errors ( @xmath94 kev ) . the level of agreement , as well as the magnitude of calibration errors ,
is illustrated in figure [ fig_nhkt ] , which shows contours of @xmath86 in the @xmath81-@xmath95 plane for the northern overlap region .
figure [ fig_nhkt ] suggests residual systematic errors in the determination of the absorbing column may be of order @xmath96 @xmath6 .
as noted above , in subsequent analysis we fix @xmath82 @xmath6 and include a systematic temperature error of @xmath97 kev to account for this effect .
( , ) f5.eps we estimated the effect of x - rays scattered by the xrt from the bright , cool core of abell 1795 to the cluster outskirts with the aid of a number of simulations . for this purpose
we used the ftools xissim and xissimarfgen @xcite . for fields centered
@xmath98 from a bright source ( comparable to our far north and far south fields ) , actual stray intensity levels are expected to be less than @xmath99 times the values predicted by these tools @xcite . to gauge the effect of scattered x - rays from the bright cluster core on the surface brightness profiles at large radius , we simulated observations of a point source located at the cluster center with a spectrum and total flux observed within @xmath38 of the center ( see table [ tab2 ] ) . since the actual cluster emission is more diffuse than assumed in this model , our simulation overestimates the scattered flux from the ( unresolved ) bright cool core of the cluster .
we simulated the full mosaic observation of five pointings for each sensor , and normalized the simulated observations using the exposure maps discussed in section [ sexpmap ] above .
we then extracted surface brightness profiles for this point source using the same methods we used to extract the cluster profiles ( as discussed in section [ sec : sb ] ) .
in particular , all regions excluded from the cluster profiles because they contained cosmic sources were also excluded in the simulations . for clarity
, we did not add a cosmic x - ray background signal to the simulated point source profile .
the resulting point - source surface brightness profile in the 0.52 kev band is shown by the points in figure [ figsimps ] .
the figure shows that the expected scattered flux at @xmath100 exhibits a distinct oscillation at relatively low amplitude .
the oscillation is presumably due to the variations with field angle of the intensity of so - called ` secondary ' ( single bounce ) and ` backside ' ( three - bounce ) reflections @xcite .
( , ) f6.eps comparison of figure [ figsimps ] with the estimated celestial background brightness listed in table [ tab : fluxes ] shows that the surface brightness due to scattered flux in this radial range is always less than 7% of the cosmic background surface brightness we observe at the largest radii ( @xmath101 ) .
the actual scattered intensity from the cluster will be below that shown in figure [ figsimps ] at these radii because convolution of the point response function ( prf ) with the extended cluster flux distribution will reduce the amplitude of the oscillating wings .
@xcite note that , given alignment tolerances of the xrt reflectors and pre - collimator blades , the actual scattered flux at @xmath102 could in principle exceed the ray - tracing simulations ( which assume perfect alignment ) by as much as a factor of 2 .
allowing for this uncertainty , we expect that scattered cluster flux will contribute at most 10% of the total observed brightness at @xmath100 .
we confirmed the effect of scattered cluster flux on our spectral analysis in the outermost regions of our field ( @xmath103 ) by computing the effective area ( as a function of energy ) for detection of photons there originating from the central @xmath38 of the cluster .
here we used the measured cluster surface brightness distribution as input to the effective area calculation , rather than a simple point source .
note that this central region produces more than 80% of the total cluster flux reported for a1795 in the northern rosat all - sky survey ( noras ) catalog @xcite . using these effective area functions , and the cluster spectral model described in table [ tab2 ] , we employed xspec to simulate the flux and spectrum of x - rays scattered from the cluster center into each cluster ( and background ) region at @xmath104 .
the resulting fluxes are listed in the first two columns of table [ tab3 ] .
l c c c c c c & & & + region@xmath83 & @xmath105 & @xmath106 & @xmath105 & @xmath106 & @xmath105 & @xmath106 + 17.522 ( n ) & 0.36 & 0.25 & 0.38 & 0.02 & 0.74 & 0.27 + 17.522 ( s ) & 0.21 & 0.16 & ... & ... & 0.21 & 0.16 + & & & & & + 2226 ( n ) & 0.37 & 0.33 & 0.45 & 0.03 & 0.82 & 0.36 + 2226 ( s ) & 0.37 & 0.36 & ... & ... & 0.37 & 0.36 + & & & & & + 27.531.7 ( n ) & 0.47 & 0.35 & ... & ... & 0.47 & 0.35 + 27.531.7 ( s ) & 0.45 & 0.34 & ... & ... & 0.45 & 0.34 + we conclude from table [ tab3 ] that the scattered intensity from the cluster core contributes @xmath76 35% of the total observed surface brightness in the regions for which we perform spectroscopy .
both magnitude and variation with radius are in agreement with the foregoing analysis of the prf .
we reiterate that uncertainties in collimator alignment render these estimates uncertain by as much as a factor of 2 .
the estimated scattered fluxes in the northern and southern regions differ somewhat , particularly in the radial range 17.522 .
the differences are presumably due to the exclusion of point sources in the north ( see figure [ fig1 ] ) , which changes the effective radii of the annular bins there , coupled with the rapidly varying scattered flux in this region ( see figure [ figsimps ] ) . in any event ,
the scattered flux in all regions listed in table [ tab3 ] is comparable to or less than the statistical errors in the estimated background ( see table [ tab : fluxes ] ) .
we note that , because of the oscillating prf wings ( see figure [ figsimps ] ) , the simulations imply that spatial non - uniformity of the scattered flux will cause us to over - estimate the background at @xmath107 , and thus to ( slightly ) underestimate the cluster flux in the outer - most regions . finally , we investigated the effect of scattered flux from bright point sources on our cluster spectrophotometry .
the brightest such source is in the far - north field , with suzaku position @xmath108 , and we estimate its flux to be @xmath109 erg s@xmath5 @xmath6 ( 0.52 kev ) within an aperture of @xmath110 radius . we have excluded a 3.5 diameter region around this source from the cluster analysis .
we estimated the flux contributed by this source to each of our spectroscopic regions following methods described above for the cluster flux .
results are listed in table [ tab3 ] .
the wings of the image of this bright point source contribute less than 5% of the observed total flux to each of the two outermost spectral analysis regions ( extending from @xmath111 to @xmath24 from the cluster ) .
thus in these two regions , the bright point source contribution is comparable to that from scattered cluster flux . in all other regions ,
the bright point source contribution is smaller .
we note that point source contributions should be known more accurately than those from the cluster core because the the ray - tracing simulations are better calibrated at @xmath112@xmath113 than at @xmath114 @xcite .
wide and deep x - ray surveys conducted over the past @xmath115 years have revealed a great deal about the distribution in flux and space of the sources that constitute the extragalactic x - ray background ( @xcite and references therein ) .
we have used this information to estimate the expected variation of the extragalactic background not resolved by suzaku . on scales larger than the suzaku prf ( @xmath116 ) , the correlation of source angular positions is weak @xcite , and for purposes of estimating the expected variance in our background , we neglect it .
formally the expected background surface brightness @xmath117 due to sources with flux @xmath118 is @xmath119 , where @xmath120 is the differential number of sources per unit solid angle and flux .
the expected variance of the background brightness , @xmath121 , when measured over a solid angle @xmath122 , is @xmath123 .
we adopt a broken power - law model for @xmath120 using parameters derived by @xcite .
it may be instructive to note that if the integral number counts @xmath124 varied simply as @xmath125 , then the relative fluctuation in the the unresolved background within a solid angle @xmath122 would satisfy @xmath126 , where the factor @xmath127 varies between @xmath128 and @xmath129 for appropriate values of @xmath78 .
this relationship sets the flux limit @xmath130 to which point sources must be detected for accurate background estimation in a field of solid angle @xmath122 .
our suzaku data alone allow us to detect point sources to a limiting flux @xmath131 , where @xmath132 is the flux in units of @xmath133 erg s@xmath5 @xmath6 , in either the soft ( 0.52 kev ) or hard ( 28 kev ) bands .
for a significant fraction of our field , we can also use archival xmm - newton observations to constrain the number of unresolved sources contributing to the suzaku background at fluxes as low as @xmath134 in the soft band , and @xmath135 in the hard band ( see below ) . using our adopted source - count flux relation , we find the expected rms fluctuations in background due to unresolved sources are @xmath136 erg s@xmath5 @xmath6 deg@xmath7 for @xmath137 in the soft band .
the corresponding fluctuation amplitudes in the hard band are @xmath138 erg s@xmath5 @xmath6 deg@xmath7 for @xmath139 . here
@xmath140 is the solid angle of the measurement region in units of @xmath141 deg@xmath142 .
the field coverage of the available xmm - newton observations is shown in figure [ figxmm ] .
we searched the second xmm - newton serendipitous sky survey ( xmm2p ) catalog available at nasa s high energy astrophysics science archive research center for sources in our suzaku field at distance @xmath143 from the cluster center .
we found that the effective ( epic pn ) exposure in this region ranges from 1050 ksec .
our adopted xmm - newton flux thresholds ( @xmath144 and @xmath145 in the soft and hard bands , respectively ) exceed the nominal xmm - newton @xmath146 detection thresholds @xcite for the minimum effective exposure available in our field .
we summed the flux of sources brighter than these thresholds in each of our spectral analysis regions .
since the xmm2p catalog does not tabulate source flux in the 28 kev band , we estimated the 28 kev flux by scaling the tabulated xmm - newton 212 kev flux by a factor appropriate for power - law photon number distribution with index 1.4 ( that is , we assumed @xmath147 . ) the contributions of sources resolved by xmm - newton to the suzaku background , as well as the estimated fluctuations in the background due to fainter sources , are listed in table [ bgtab ] .
( 60mm,150mm)f7.eps l c c c c c c c c c c c & & & & & + radius@xmath83 & @xmath148 & & @xmath149 & @xmath150 & @xmath151 & @xmath152 & & @xmath153 & @xmath154 & @xmath155 & @xmath156 + north : & & & & & & & & & & & + 57.5 & 0.60 & & @xmath157 & 0@xmath158 & 7.8 & 4.9 & & @xmath157 & 0@xmath158 & 16.4 & 5.8 + 7.510 & 0.85 & & @xmath157 & 0@xmath158 & 7.8 & 4.2 & & @xmath157 & 0@xmath158 & 16.4 & 4.9 + 1012.5 & 0.93 & & @xmath159 & 2.2 & 3.1 & 0.96 & & @xmath160 & 2.7 & 12.0 & 2.6 + 12.515 & 0.98 & & @xmath159 & 3.9 & 3.1 & 0.94 & & @xmath160 & 3.8 & 12.0 & 2.6 + 1517.5 & 0.91 & & @xmath159 & 1.3 & 3.1 & 0.98 & & @xmath160 & 0 . & 12.0 & 2.6 + 17.5 26 & 1.58 & & @xmath159 & 0 & 3.1 & 0.74 & & @xmath160 & 7.4 & 12.0 & 0.8 + 27.531.7 & 1.54 & & @xmath157 & 0@xmath158 & 7.8 & 3.1 & & @xmath157 & 0@xmath158 & 16.4 & 3.6 + + south : & & & & & & & & & & & + 57.5 & 0.57 & & @xmath157 & 0@xmath158 & 7.8 & 5.1 & & @xmath157 & 0@xmath158 & 16.4 & 6.0 + 7.510 & 0.79 & & @xmath157 & 0@xmath158 & 7.8 & 4.3 & & @xmath157 & 0@xmath158 & 16.4 & 5.1 + 1012.5 & 0.93 & & @xmath159 & 0 & 3.1 & 0.96 & & @xmath160 & 0 & 12.0 & 2.6 + 12.515 & 0.83 & & @xmath159 & 1.3 & 3.1 & 1.0 & & @xmath160 & 4.6 & 12.0 & 2.7 + 1517.5 & 1.22 & & @xmath159 & 2.0 & 3.1 & 0.84 & & @xmath160 & 3.8 & 12.0 & 2.3 + 17.526 & 3.11 & & @xmath159 & 0 & 3.1 & 0.54 & & @xmath160 & 1.1 & 12.0 & 1.4 + 27.531.7 & 1.25 & & @xmath157 & 0@xmath158 & 7.8 & 3.5 & & @xmath157 & 0@xmath158 & 16.4 & 2.2 + in the outskirts of the cluster ( @xmath100 ) , the expected fluctuations in the unresolved background are @xmath161 erg s@xmath5 @xmath6 deg@xmath7 ( 0.52 kev ) , or about 10% of the total ( galactic plus extragalactic ) background . in our two background regions at @xmath107 , which lack complete xmm - newton coverage ,
the expected fluctuations are significantly larger , ranging from 3@xmath162 erg s@xmath5 @xmath6 deg@xmath7 .
we note also that the expected extragalactic background brightness for these regions is in excellent agreement with the measured values ( see table [ tab : fluxes ] ) .
finally , we stress that the expectation value of the extragalactic background brightness varies from region to region because the extragalactic source population ( as well as our knowledge of it ) varies ( see table [ bgtab ] ) .
thus , for example , it is appropriate when modelling the background to allow the extragalactic component to vary in brightness from region to region , and we do so in our fits .
radial surface brightness profiles were extracted from exposure - corrected images in the 0.52 kev and 28 kev spectral bands using the chandra x - ray center s ciao / sherpa tools .
point sources indicated in figure [ fig1 ] were excluded .
the center of the profile was determined from a fit of a circular beta - model ( @xmath163 with @xmath164 ) to the ( two - dimensional ) 0.52 kev image .
a one - dimensional model consisting of a @xmath165-model plus a constant background was then fit to each of the radial profiles .
the profiles and the best - fitting models are presented in figure [ figsb ]
. the count - rate to flux conversions for these profiles were derived from the swcx - free spectral model of the emission from the background regions ( see the upper panel of table [ tab : fluxes ] ) .
as explained in section [ sexpmap ] , this leads to a cluster flux overestimate of less than 10% ( 5% ) for the 0.52 kev ( 28 kev ) band .
it is appropriate for our null hypothesis that surface brightness variations in the a1795 outskirts are due to cosmic background fluctuations .
the northern and southern surface brightness profiles shown in figure [ figsb ] agree remarkably well with one another ( and with a simple @xmath165-model with @xmath166 ) for @xmath167 mpc ( 14 ) @xmath168 .
as expected , the cluster is symmetrical and evidently relaxed in this region . at larger radii ,
the profiles diverge , with the surface brightness apparently falling more rapidly than the @xmath165-model in the south , but actually rising in the north to an apparent peak near @xmath169 mpc ( 26 ) @xmath170 in the 0.52 kev band .
the surface brightness in this peak is about twice that in the nominal background regions , which are at @xmath171 mpc ( @xmath107 ) .
the southern profile actually drops @xmath172% below the level in the nominal background regions .
( 4.5in,3.375in)f8a.eps ( 4.5in,3.375in)f8b.eps ( 4.5in,3.375in)f9a.eps ( 4.5in,3.375in)f9b.eps the surface brightness profiles at radii @xmath100 are re - plotted in figure [ figsboutskirts ] , along with estimates of the cosmic x - ray background as a function of radius . to minimize the poisson fluctuations in the extragalactic background we have re - binned the data at @xmath173 .
the background estimates , shown as dashed histograms in figure [ figsboutskirts ] , include three components .
we assume a spatially uniform galactic contribution in the 0.52 kev band , with flux equal to the mean thermal flux obtained from the spectral fitting of the background regions ( see table [ tab : fluxes ] ) : @xmath174 erg s@xmath5 @xmath6 deg@xmath7 ( @xmath175 errors ) .
this estimate assumes no swcx emission ; we consider the effects of swcx below .
we adopt as the extragalactic background the sum of the total flux of any point sources detected by xmm - newton plus an estimate of the unresolved flux derived from our adopted model of number counts @xcite .
the latter two components are listed for each spectral analysis region in table [ bgtab ] .
error bars plotted on the background histograms include the expected rms poisson fluctuations , also as listed in table [ bgtab ] .
the global best - fit @xmath165-models ( assuming a spatially uniform background ) are shown as solid curves in figure [ figsboutskirts ] .
figure [ figsboutskirts ] shows that the background model agrees well with the measured surface brightness in the background regions ( at @xmath176 ) , which suggests that our extragalactic background model is reasonable . in the region @xmath177
( @xmath2 mpc ; @xmath18 ) , the observed soft - band emission significantly exceeds the background in both the north and the south .
the formal net soft - band cluster surface brightnesses in this radial range are @xmath178 erg s@xmath5 @xmath6 deg@xmath7 in the north ( south ) , quoting @xmath175 errors .
allowing for scattered flux contributions listed in table [ tab3 ] , the north detection is significant at a level greater than @xmath179 ; the net cluster surface brightness in the south falls below @xmath3 .
if we allow for possible swcx contributions , the estimated galactic plus geocoronal background is somewhat higher , but we still find significant cluster emission in the north . in this case , the estimated galactic plus geocoronal background rises to @xmath180 erg s@xmath5 @xmath6 deg@xmath7 in the north ( south ) , with @xmath175 errors . the corresponding results for cluster flux are @xmath181 erg s@xmath5 @xmath6 deg@xmath7 in the north ( south ) , with @xmath175 errors .
in the north , the signal exceeds the background by at least @xmath182 , allowing for the scattered light contributions listed in table [ tab3 ] .
the @xmath3 upper limit on the cluster emission in the south is @xmath183 erg s@xmath5 @xmath6 deg@xmath7 .
we also note that formally the measured net cluster surface brightness in the north exceeds that in the south by a factor of @xmath184 ( @xmath175 error ) ; the absolute brightness difference of @xmath185 erg s@xmath5 @xmath6 deg@xmath7 is significant at more than @xmath186 .
we conclude that suzaku has detected soft - band cluster emission with high confidence in a region with @xmath18 north of the cluster center , but not in the corresponding region in the south . in this radial range ,
the cluster is significantly brighter in the north than in the south .
the lower panel of figure [ figsboutskirts ] shows that suzaku detects no hard - band cluster emission at @xmath41 ( 1.3 mpc ) in either north or south .
indeed , the suzaku surface brightness measurements are remarkably consistent with the background estimates derived from the xmm - newton point - source catalog and the assumed number - counts model for @xmath177 . in the hard band , the significant north - south difference in measured suzaku surface brightness is attributable entirely to differences in the background source populations in these two regions . finally , comparison of the top and bottom panels of figure [ figsboutskirts ] shows that our background model implies spatial variations in the spectral shape of the background .
for example , the background is marginally harder in the north than in the south in the radial range @xmath177 , and in the north the background is harder at @xmath177 than at @xmath176 .
we allow for these variations in the spectral fitting used to extract the temperature profiles discussed below .
simulations of cluster x - ray emission in the vicinity of the virial radius @xcite suggest that the surface brightness profile follows a broken power law with the break at @xmath187 .
simple power - law fits to the a1795 surface brightness profiles at @xmath188 are roughly consistent with these simulations .
best - fit values for the power - law index @xmath78 , where @xmath189 , are listed in table [ pltab ] .
the two different radial ranges shown correspond to @xmath190 ( @xmath191 ) and @xmath192 ( @xmath193 ) . in the soft band , the northern and southern profile slopes are identical , within errors , over the smaller radial range , but the apparent excess emission at @xmath194 ( @xmath41 ) flattens the profile there . in all cases the hard - band profile slopes are steeper than the corresponding soft - band slopes .
l c c direction & + & 0.52 kev & 28 kev + + n & @xmath195 & @xmath196 + s & @xmath197 & @xmath198 + n+s & @xmath199 & @xmath200 + + n & @xmath201 & @xmath202 + s & @xmath203 & @xmath204 + n+s & @xmath205 & @xmath206 + the soft - band power - law index for the southern profile is quite similar to that reported by @xcite from stacked rosat profiles for a number of hot clusters .
for example , neumann finds @xmath207 for @xmath208 for a set of clusters similar to ( and including ) a1795 . on the other hand ,
the soft - band suzaku profile north of the cluster center is somewhat shallower than the stacked rosat profile .
we have fit the rolling power - law function used by @xcite to our data .
this function is @xmath209 with @xmath210 and @xmath211 . at small
@xmath212 , @xmath213 , while at large @xmath212 , @xmath214 , with the transition occurring around @xmath215 .
our data are confined to @xmath216 , so they do not constrain @xmath217 well . therefore in our fits
we fix @xmath218 for the soft ( hard ) band , values typical of the simulated clusters @xcite . with these constraints we find @xmath2193.5 , consistent with the simulations , for both hard - band profiles and for the soft - band profile in the south . again because of the excess emission at @xmath41 in the north , however , we find @xmath220 for the soft - band profile in this direction ; this is flatter than the simulated clusters .
we note that @xcite filter the emission from their simulated clusters before extracting surface - brightness and temperature profiles .
the aim of this filtering is to remove high - density , low - temperature features which are presumed to be too faint to detect , or perhaps even non - physical artifacts of the simulations .
we note that profiles of unfiltered , individual simulated clusters ( see , for example , their figure 2 ) show significant departures from the filtered soft - band surface brightness profile , especially at @xmath221 .
the magnitude of these deviations is at least as great as the north - south difference we see in abell 1795 .
similar fluctuations are not seen in the hard - band profiles of the simulated clusters .
we conclude that the north - south surface brightness difference we observe is similar to features present in these ( unfiltered ) simulated clusters .
the temperature profiles to the north and south of the cluster are shown in figure [ figtp ] , along with results from chandra and xmm - newton .
we are able to measure the cluster temperature past the limit of xmm - newton and chandra data ( about 1.1 mpc or 15 ) and find that the temperature continues to decline to the largest radius at which emission is detectable . at the outermost radial bin ,
spanning @xmath2 mpc , the temperature is about one - third of the peak cluster temperature observed by suzaku .
( , ) f10.eps the radial values and 90% error bars shown in figure [ figtp ] are estimated from the xissimarfgen ray - tracing simulations described in section [ sec : model ] .
these radial bins more accurately reflect the real locations of photons originating from the cluster , and we use these values for the profile fitting described in the remainder of this section .
such corrections are less straightforward for the surface brightness profiles shown in figures [ figsb ] and [ figsboutskirts ] , since the emission in that case is from a combination of sources ( i.e. , cluster and cosmic background ) that have different spatial distributions .
the suzaku data do not resolve the cool core at @xmath222 , but at larger radii we find good agreement between the suzaku and xmm - newton in the radial range covered by both observatories .
note that here we have used the most recent analysis of the xmm - newton data by snowden ( private communication ) , which uses the latest epic - pn calibrations and provides somewhat lower temperatures than those reported earlier @xcite .
the chandra temperatures reported by @xcite , however , are generally higher than the suzaku ( and xmm - newton ) temperatures outside the cool core , especially at @xmath223 .
the reason for this discrepancy is not entirely clear . in principle , scattering by the telescope of photons from the bright , cool core over the central few arcminutes would tend to reduce the apparent temperature .
one might expect the scattering to be more significant for suzaku , and , to a lesser extent , for xmm - newton than for chandra , given the latter s much sharper point response function . if this effect were important at radii as large as @xmath224 , however , it would be surprising if suzaku and xmm - newton data agreed as well as they do , given their very different point response characteristics .
a more plausible explanation may lie in the known discrepancy between chandra and xmm - newton temperatures for relatively hot ( @xmath225 kev ) clusters .
chandra reports temperatures that are systematically higher than those from xmm - newton for such clusters , and the magnitude of the discrepancy at @xmath226 kev is very close to that shown in figure [ figtp ] at @xmath227 .
the discrepancy is smaller at lower temperatures .
uncertainties in the thickness of a thin ( @xmath22820 ) contamination layer on the chandra mirrors likely account for this discrepancy .
we note that this effect is unlikely to explain the discrepancy at @xmath229 , where @xmath230 kev , a temperature at which this calibration error is thought be quite small .
we also note that the azimuthal range covered by the chandra observations @xcite differs from that we observed with suzaku at these large radii . in principle , therefore , the discrepancy between chandra and suzaku results at this large radius could reflect azimuthal temperature variations in the cluster .
in any event , we are principally concerned here with the cluster temperature profile at @xmath231 .
given the good agreement between the suzaku and xmm - newton temperature profiles in this radial range , we base our subsequent analysis on the suzaku temperatures shown in figure [ figtp ] . fitting a power law ( assuming @xmath232 ) to the projected temperature at @xmath191 yields consistent slopes in the north and the south , and @xmath233 for a joint fit to all the data in this radial range .
this is similar to the value @xmath234 reported by @xcite from suzaku observations of the outskirts of the cluster around pks 0745 - 191 .
we deprojected the temperature profiles to the north and south of the cluster separately using the smaug model in xspec .
the smaug algorithm @xcite fits analytic functions for the deprojected temperature , density , and abundance profiles , and compares the reprojected spectra to the data .
we fit spectra in the radial range 517.5@xmath235 ( 0.371.28 mpc ) , although the profiles were extended to 2 mpc for reprojection . since the emission north of @xmath111 is evidently not part of a spherically symmetric component , we excluded it from the deprojection .
we assumed profiles of the form @xmath236^{-\epsilon}$ ] , where @xmath237 , @xmath238 and @xmath239 were allowed to vary independently for the temperature and density profiles , while @xmath237 and @xmath239 were allowed to vary for the abundance profile .
the abundance profile @xmath238 parameter was frozen at 10 kpc , essentially producing a power - law dependence over the fitting region .
we assumed a mekal model for the plasma emission , and included galactic and extragalactic components identical to those described in section 2 .
parameter errors were estimated using mcmc techniques identical to those described in section [ swcx ] .
the best - fit deprojected power - law temperature profiles are shown as continuous curves in figure [ figdepro ] , along with the measured ( projected ) temperatures .
the shaded regions surrounding the curves indicate 90% confidence intervals .
the best - fit model parameters for deprojected temperature , hydrogen density , and abundance profiles are listed in table [ dptab ] .
figure [ figdepro ] shows that the model deprojected temperatures track the observed projected temperatures quite closely . as expected , the deprojected temperatures are slightly higher , but , as suggested by @xcite , the difference is relatively small , especially at large radius .
note that formally the fit quality is acceptable in the south but marginal in the north ( see table [ dptab ] ) .
llccc & & north & south & average + @xmath95 & @xmath237 ( kev ) & 12.8@xmath240 & 16.2@xmath241 & 14.5@xmath242 + & @xmath239 & 0.38@xmath243 & 0.40@xmath244 & 0.39@xmath243 + & @xmath238 ( mpc ) & 0.19@xmath245 & 0.13@xmath246 & 0.16@xmath247 + @xmath248 & @xmath237 ( 10@xmath249 cm@xmath249 ) & 3.03@xmath251 & 3.06@xmath252 & 3.04@xmath253 + & @xmath239 & 1.17@xmath254 & 1.22@xmath254 & 1.20@xmath254 + & @xmath238 ( mpc ) & 0.30@xmath255 & 0.30@xmath255 & 0.30@xmath255 + abund & @xmath237 ( solar ) & 4.31@xmath256 & 4.40@xmath257 & 4.36@xmath258 + & @xmath239 & 0.36@xmath259 & 0.36@xmath260 & 0.36@xmath261 + & @xmath238 ( mpc)@xmath262 & @xmath263 & @xmath263 & @xmath263 + @xmath264/dof & & 1875.93/1686 & 1851.83/1810 & ... + as expected from the projected temperatures ( see figure [ figtp ] ) , the deprojected suzaku temperatures are somewhat lower than those inferred from chandra data @xcite .
moreover , our deprojected temperature profile falls more rapidly with increasing radius than does the chandra profile .
for example , at @xmath265 mpc @xmath14 , our ( average ) model profile has a logarithmic slope @xmath266 while @xcite find @xmath267 .
( here we cite 68% confidence errors . ) in interpreting this comparison one must again bear in mind that the suzaku and chandra observations sample different ranges in azimuth around the cluster .
our measured slope is also steeper than those found for simulated clusters by both @xcite , who fit a functional form with logarithmic slope @xmath268 at @xmath269 , and by @xcite , who report @xmath270@xmath271 at @xmath272 . in the latter case
the steeper slope is associated with simulations which include radiative cooling .
our deprojected density profile approaches a power law at large radius , and the results in table [ dptab ] imply @xmath273 at @xmath274 mpc .
this is in agreement with the density profile slope reported by @xcite from chandra data : their fitting formula is reasonably well - approximated by a power law with index @xmath275 over the radial range we consider .
our assumed forms for density and temperature imply that the icm should be polytropic at large radius , with a best - fit index of @xmath10 .
finally , we have computed the implied entropy index , ( @xmath276 ) as a function of radius , assuming @xmath277 .
the result is shown , with 90%-confidence uncertainties , in figure [ fig_ent ] .
there we also show the entropy profile expected from hierarchical structure formation derived from ( non - radiative ) simulations by @xcite ; this curve increases as @xmath278 .
our average entropy profile increases more slowly , as @xmath279 .
( , ) f11.eps ( , ) f12.eps the relatively low temperatures and entropies we observe in the outskirts of abell 1795 may be an indication that the plasma we detect is not in hydrostatic equilibrium in the cluster s gravitational potential .
we note that our suzaku observations sample two restricted sectors of the cluster at this radius , and that data from other observatories sample different sectors .
if the icm is not in hydrostatic equilibrium at this radius , one might expect azimuthal variations in icm density and temperature .
for example , as we noted earlier in section [ sec : sb ] , the simulations of @xcite produce ( unfiltered ) clusters with distinct cool clumps at @xmath272 .
these may be the result of relatively cool , low - mass substructures falling toward the cluster center .
moreover , @xcite infer modest departures from hydrostatic equilibrium at radii @xmath280 from a comparison of cluster masses determined from x - ray and weak - lensing data .
this result is at least qualitatively in agreement with our interpretation of the temperature and entropy profiles of abell 1795 .
the x - ray surface brightness and temperature profiles can be used to estimate the cluster mass as a function of radius if we assume that the icm is hydrostatic equilibrium . in this case , following @xcite , @xmath281 \big[\frac{r}{{\rm 1\ : mpc}}\big ] \times \nonumber \\ & & \big [ \frac{d\log \rho_g}{d\log r } + \frac{d\log t}{d\log r } \big ] .
\label{eqn1}\end{aligned}\ ] ] here @xmath282 is the mass density of the icm and it is assumed that the mean molecular weight of the icm @xmath283 .
chandra @xcite and xmm - newton @xcite have been used to estimate mass profiles for abell 1795 at radii up to @xmath284 mpc . in this section
we compare mass constraints obtained from our suzaku observations with these measurements .
adopting the mean of the north and south profiles for the deprojected temperature and density ( see table [ dptab ] ) , together with the mean of the north and south temperatures measured in the final radial bin of each deprojected profile , we use equation [ eqn1 ] to estimate the cluster mass . we obtain @xmath285 within @xmath286 mpc , the value of @xmath287 reported by @xcite .
the 68% confidence uncertainties quoted here are dominated by uncertainties in the temperature measurement .
our mass estimate is somewhat lower than others reported for abell 1795 .
@xcite find @xmath288 at @xmath286 mpc , while @xcite report @xmath289 at @xmath290 mpc .
formally , these exceed our mass estimate by @xmath12 and @xmath291 , respectively .
the discrepancy is almost entirely due to the difference in temperature estimated at these radii .
for example , the best - fit model used by @xcite for the deprojected temperature predicts @xmath292 kev , while we estimate @xmath293 kev at @xmath294 mpc .
our lower mass estimate implies @xmath295 mpc , somewhat smaller than the chandra estimate of @xmath296 mpc @xcite , and consistent with the xmm - newton value @xmath297 mpc @xcite .
similarly , we find @xmath298 mpc , smaller than both the beppo - sax estimate of @xmath299 @xcite , and the value ( @xmath300 mpc ) we estimate from theoretical scaling relations @xcite .
finally , given that @xmath301 in our assumed cosmology @xcite , our mass profile implies @xmath302 mpc .
the rapidly falling temperature profile we observe implies a rapidly falling mass density profile and a slowly rising integrated mass profile ( see equation [ eqn1 ] ) . a power - law approximation to our best - fit integrated mass profile in the region
@xmath303 mpc yields @xmath304 .
this is much flatter , for example , than the slope expected from an nfw model with scale radius @xmath305 mpc reported by @xcite .
we have presented results of suzaku observations of two distinct regions in the outskirts of abell 1795 .
we detect x - ray emission to @xmath306 mpc in both regions , and in the north trace the hot icm to @xmath169 mpc @xmath307 .
we find that the x - ray surface brightness at @xmath2 mpc is significantly higher in the north than in the south .
we measure the run of temperature with radius at @xmath308 and find that it falls relatively rapidly ( @xmath309 ) and reaches a value about one - third of its peak at the cluster outskirts .
one possible interpretation of our observations is that the icm is not in hydrostatic equilibrium in the radial range @xmath18 , and that in the north we are seeing infalling plasma at relatively low ( @xmath310 kev ) temperature . in support of this picture
we note that the x - ray surface brightness is neither azimuthally symmetric nor ( in the north ) falling monotonically with increasing radius .
moreover , the rapidly falling deprojected temperature profile would require that there is relatively little gravitating mass at @xmath1 mpc if the icm is polytropic and in hydrostatic equilibrium , a result that conflicts with extrapolations of the best - fit nfw profiles derived from high - resolution chandra observations @xcite .
simulations @xcite predict , and recent lensing and x - ray observations @xcite provide evidence for , modest deviations from hydrostatic equilibrium in the icm at @xmath311 .
our observations of abell 1795 suggest that in some cases these deviations may be quite significant even in a cluster which appears to be relaxed at smaller radii .
additional observations of a variety of clusters , with more complete azimuthal coverage than we have obtained , will improve understanding of the state of the icm in cluster outskirts , and therefore of cluster masses , structure and formation mechanisms .
finally , we note that these and other observations of the low - surface - brightness outskirts of clusters have been made possible by suzaku s relatively low and stable non - x - ray background .
the surface brightness sensitivity of our observations , @xmath312 erg s@xmath5 @xmath6 deg@xmath7 ( @xmath3 ) , is less than 20% of the total cosmic background in this spectral band .
our sensitivity was ultimately limited by time - variable solar wind charge exchange emission and by poisson variations in the extragalactic source density .
we demonstrated here that variations in the geocoronal contribution can be modeled to this level from simultaneous observations of cluster - free background regions . to reduce background source density variations below this level in the relatively small cluster regions ( 0.010.03 deg@xmath142 ) we required information from the xmm - newton source catalog .
future suzaku observations of cluster outskirts would benefit from coverage of larger ( @xmath313 deg@xmath142 ) solid angle in each radial bin , especially if deep x - ray source catalogs are not available .
we thank alexey vikhlinin and steve snowden for providing temperature data from chandra and xmm - newton , respectively , and helen russell for useful discussions about deprojection .
mwb and edm were supported in part by nasa grant nng05gm92 g to mit .
hughes acknowledges support from nasa grant nngg05gp87 g . j.p .
henry acknowledges support from nasa grant nng06gc04 g .
99 anders , e. , & grevesse , n. 1989 , , 53 , 197 arnaud , k. a. 1996 , in astronomical data analysis software and systems v , ed jacoby , g. , & barnes , j. , asp conference series volume 101 , p17 arnaud , m. , & evrard , a. 1999 , , 305 , 631 bhringer , h. , et al .
2000 , , 129 , 435 borgani , s. , et al .
2004 , , 348 , 1078 brandt , w. n. , & hasinger , g. 2005 , , 43 , 827 bryan , g. , & norman , m. 1998 , , 495 , 80 , j. a. , & sembay , s. 2008 , , 489 , 837 , t. e. 2000 , , 532 , l153 deluca , a. , & molendi , s. 2004 , , 419 , 837 dickey , j. m. , & lockman , f. j. 1990 , , 28 , 215 ettori , s. , & balestra , i. 2008 , , in press ( astro - ph/0811.3556 ) ettori , s. , de grandi , s. , & molendi , s. 2002 , , 391 , 841 evrard , a. , metzler , c. , & navarro , j. 1996 , , 469 , 494 fabian , a. c. , sanders , j. s. , ettori , s. , taylor , g. b. , allen , s. w. , crawford , c. s. , iwasawa , k. , & johnstone , r. m. 2001 , , 321 , l33 fujita , y. , tawa , n. , hayashida , k. , takizawa , m. , matsumoto , h. , okabe , n. , & reiprich , t. 2008 , , 60 , s343 , r. , et al .
2007 , , 59 , s133 george , m. , fabian , a. , sanders , j. , young , a. & russell , h. 2008 , , 395 , 657 ( astro - ph/0807.1130 ) hallman , e. , burns , j. , motl , p. , & norman , m. 2007 , , 665 , 911 horner , d. 2001 , phd thesis , university of maryland ikebe , y. , bhringer , h. , & kitayama , t. 2004 , , 611 , 175 ishisaki , y. et al .
2007 , , 59 , s113 kalberla , p. m. w. , burton , w. b. , hartmann , d. , arnal , e. m. , bajaja , e. , morras , r. , & pppel , w. g. l. 2005 , , 440 , 775 , d. , acero , f. , lallement , r. , ballet , j. , & kharchenko , v. 2007 , , 475 , 901 koyama , k. , et al .
2007 , , 59 , s23 kuntz , k. , & snowden , s. 2000 , , 543 , 195 mahdavi , a. , hoekstra , h. , babul , a. , & henry , j. p. 2008
, , 384 , 1567 , e. d. , et al .
2008 , , 60 , s95 moretti , a. , campana , s. , lazzati , d. , & tagliaferri , g. 2003 , , 588 , 696 navarro , j. , frenk , c. , & white , s. 1997 , , 490 , 493 neumann , d. 2005 , , 439 , 465 piffaretti , r. , jetzer , ph . , kaastra , j. , & tamura , t. 2005 , , 433 , 101 , f. , molendi , s. , ghizzardi , s. , & de grandi , s. 2003 , , 592 , 62 reiprich , t. h. , hudson , d. s. , zhang , y .- y . , sato , k. , ishisaki , y. , hoshino , a. , ohashi , t. , ota , n. , & fujita , y. 2008 , , accepted ( arxiv:0806.2920 ) roncarelli , m. , ettori , s. , dolag , k. , moscardini , l. , borgani , s. , & murante , g. 2006 , , 373 , 1339 serlemitsos , p. et al .
2007 , , 59 , s9 smith , r. , et al .
2004 , , 128 , 1558 snowden , s. , egger , r. , freyberg , m. , mccammon , d. , plucinsky , p. , sanders , w. , schmitt , w. , & trmper , j. 1997 , , 485 , 125 , s. l. , collier , m. r. , & kuntz , k. d. 2004 , , 610 , 1182 snowden , s. , mushotzky , r. , kuntz , k. , & davis , d. 2008 , , 478 , 615 , e. c. , frandsen , a. m. , mewaldt , r. a. , christian , e. r. , margolies , d. , ormes , j. f. , & snow , f. 1998 , space science reviews , 86 , 1 tawa , n. , et al .
2008 , , 60 , s11 vikhlinin , a. , markevitch , m. , murray , s. s. , jones , c. , forman , w. , & van speybroeck , l. 2005 , , 628 , 655 vikhlinin , a. , kravtsov , a. , forman , w. , jones , c. , markevitch , m. , murray , s. s. , & van speybroeck , l. 2006 , , 640 , 691 vikhlinin , a. , forman , w. , & jones , c. 2003 , , 525 , 47 voit , m. 2005 , rev . mod .
phys , 77 , 207 watson , m. , et al .
2001 , , 365 , l59 yang , y. , mushotzky , r. , barger , a. , cowie , l. , sanders , d. , & steffen , a. 2003 , , 585 , l85 | we report suzaku observations of the galaxy cluster abell 1795 that extend to @xmath0 mpc , the radius within which the mean cluster mass density is 200 times the cosmic critical density .
these observations are the first to probe the state of the intracluster medium in this object at @xmath1 mpc .
we sample two disjoint sectors in the cluster outskirts ( @xmath2 mpc ) and detect x - ray emission in only one of them to a limiting ( @xmath3 ) soft x - ray surface brightness of @xmath4 erg s@xmath5 @xmath6 deg@xmath7 , a level less than 20% of the cosmic x - ray background brightness .
we trace the run of temperature with radius at @xmath8 mpc and find that it falls relatively rapidly ( @xmath9 ) , reaching a value about one third of its peak at the largest radius we can measure it . assuming the intracluster medium is in hydrostatic equilibrium and is polytropic , we find a polytropic index of @xmath10 and we estimate a mass of @xmath11 within 1.3 mpc , somewhat ( @xmath12 ) lower than that reported by previous observers
. however , our observations provide evidence for departure from hydrostatic equilibrium at radii as small as @xmath13 mpc @xmath14 in this apparently regular and symmetrical cluster . |
there is great and ongoing interest in nanophotonics to create arrays of resonant systems with high quality factors @xmath3 in order to control the propagation and emission of light at a deep and fundamental level .
new states occur when the state from an individual resonator interacts with the others in the array , such as hybrid collective states or states typical for tight - binding systems @xcite .
it has been predicted that periodic arrays of coupled cavities will form waveguides with unusual and tunable dispersion properties @xcite .
if a composite quantum resonance is considered that consists of a cavity strongly coupled to a single quantum dot , such a system will exhibit extreme non - linear behavior in response to even a single energy quantum ; arrays of such resonating systems have been predicted to exhibit unusual repulsive boson behavior for individual photons @xcite .
experimentally , great advances have recently been made in the design and fabrication of high-@xmath3 cavities in photonic crystals @xcite that may be assembled into arrays of resonators @xcite .
it is very difficult , however , to achieve resonant transport in such an array , since all resonators must mutually be tuned to within their linewidth .
this entails a tuning to typically one part in a million , which is extremely challenging with the state - of - the - art in nanotechnology .
therefore , we propose a new hybrid system exploiting alkali atoms in the gas phase as high - q resonators in the photonic crystal environment . atomic resonances hold several advantages , firstly they exhibit strong and extremely narrow resonances with relative inverse linewidths - or effective quality factors @xmath3 - in the range of @xmath4 and even higher , and secondly their physical properties are extremely well understood .
thirdly and most importantly , all atoms ( of the same isotope ) are identical , hence they do not need to be individually tuned .
this advantage is in contrast to engineered resonant nanostructures - such as cavity arrays - where each individual unit differs from all other ones @xcite . and concomitant dispersion .
the vapor is introduced in a photonic crystal with a photonic gap such as an opal . as a result ,
we anticipate narrowband resonances inside the photonic gap , shown as the red trough .
, title="fig : " ] + to have resonating atoms interact with the optical properties of the host photonic crystal , an ensemble of many atoms must be infiltrated in the crystal , see figure [ fig : concept ] .
the relative dielectric function @xmath5 of the atomic medium changes drastically around a resonance frequency @xmath6 , @xmath7 exceeding unity below resonance , and being below unity above the resonant frequency @xmath8 . to achieve @xmath9 ,
the atomic transitions must be isolated .
for this reason , we have chosen to study alkali atoms .
it is desirable for the atomic medium to be dense , in order to have @xmath7 change considerably near the resonance frequency @xmath10 .
therefore , we have chosen cesium , as it has the highest vapor density . since the imaginary part of @xmath11 - associated with extinction - is related to the dispersion by the kramers - kronig relations @xcite , @xmath7 can considerably differ from unity , while @xmath12 for a select frequency range outside resonance , see fig .
[ fig : concept ] .
thus , atomic systems have a large modulation of @xmath11 , which warrants a study of their effect on photonic materials .
it has already been realized that hybrid atom - photonic systems offer exciting prospects : the creation of new modes in the photonic band gap has been predicted @xcite , as well as induced transparency @xcite , nonlinearities @xcite , and modified pulse propagation @xcite .
previously , only few experiments have been reported of composite systems of photonic crystals with embedded resonant media .
it was observed that a strongly absorbing dye in a photonic crystal leads to intriguing braggoriton resonances @xcite .
unfortunately , however , the linewidth of dye is limited by quenching interactions with other dye molecules that increase with density @xcite , thus strongly limiting resonant effects . while semiconductor quantum dots are being pursued as internal probes inside photonic crystals , their broad inhomogeneous width ( e.g. , @xmath13 for cdse @xcite )
precludes large changes in @xmath7 . to reduce inhomogeneous width effects , van coevorden _
et al . _ theoretically studied a photonic crystal consisting of cold atoms on a lattice @xcite .
such systems have indeed been studied , although the low lattice site occupancy of @xmath14 limited the atomic density in the pioneering studies @xcite .
a 3d band gap is predicted to open when the resonance wavelength equals the lattice spacing @xcite
. the required condition of one atom per lattice site has been realized @xcite , although photonic crystal properties have not been reported to date .
recently , sub - doppler features were reported in the spectra of cs infiltrated in opal nanostructures @xcite , and a preliminary study of photonic crystals infiltrated with resonant alkali atoms has been reported @xcite . here
we report on an experiment in which we infiltrate an alkali gas - a saturated cesium vapor - in a thin opal photonic crystal .
we have decided to study a dense and hot vapor of @xmath15cs as a strongly polarizable resonant medium , as it has only one electron in the outer shell in the @xmath16 state .
the well - known d - transition from the ground state to the @xmath17 manifold is strong , in particular the d@xmath1 transitions of @xmath15cs with a frequency of 335.116 thz ( wavelength @xmath18 nm ) that consist of four isolated hyperfine transitions .
the excited state has a narrow intrinsic linewidth with @xmath19 mhz , and a large polarizability at resonance @xcite . in the ground state ,
the well - known hyperfine splitting of 9.193 ghz defines the unit of time .
and v@xmath20 connect the reservoir with the cell and the cell with the pump , respectively .
, title="fig : " ] + the need for a high vapor density requires heating of liquid cs .
therefore , a cell was built in which cs can both be heated as well as spatially confined , see fig .
[ fig : setup ] .
heating is accomplished by two heating elements ( thermocoax ) brazed at the top and the bottom into the cylindrical reservoir . upon the opening of valve v@xmath1 ( swagelok ss-4bg - v51 , heated ) , hot cs vapor diffuses into the experimental cell , which is also heated by two heating elements . because cs reacts strongly with both oxygen and water ,
the reservoir and experimental cell are kept in a vacuum of @xmath21bar , that is detection limited by the low conductance ( 0.08l / s ) of the valve and the thin tube . to remove moisture and oxygen ,
the assembly is baked and evacuated at @xmath22c via valve v@xmath20 by a turbo - molecular pump for at least 24 hours before admitting cs .
both the reservoir and the experimental cell are independently temperature controlled .
the temperature is measured by ring j - type thermocouples .
heating tape is wound around the tubes and valves to prevent cold spots that would otherwise determine the vapor pressure of cs .
the experimental cell , the valves , and the tubes are kept at a temperature 30 to 50k above that of the reservoir to prevent condensation . the experimental cell is placed in a vacuum tank kept at a pressure below @xmath23 mbar to prevent convection and thus stabilize the temperature .
to avoid light absorption from cs vapor between the photonic crystal and the window , sio@xmath20 opals were grown directly on the viewport windows by vertical controlled drying @xcite . to this end , the viewports ( vacgen , uk )
are cleaned overnight in a solution of 30 g naoh , 30ml of ultrapure water , and 200ml of ethanol .
all glassware is rinsed in deionized water and in ethanol , and dried in a stream of nitrogen .
an ethanol suspension of 0.1 to 0.5 vol% sio@xmath20 colloidal nanospheres with a diameter @xmath24 nm is placed in a vial .
the viewport is inserted at an angle between @xmath25 and @xmath26 into the vial , which is heated in an oven between @xmath27c and @xmath26c . upon evaporation of the suspension ,
the colloids move towards the area of fastest evaporation , which is the liquid meniscus , and are deposited on the substrate .
the structure of such an array is close packed , typically face - centered cubic ( fcc ) @xcite . as the evaporation proceeds , ordered arrays of spheres grow until finally there is no suspension left and the growth stops @xcite .
despite consistently scanning the parameter space in temperature , angle , and colloid density , the opals on the viewports were of moderate quality , with a reflectivity of @xmath28 , where @xmath29 is the reflectivity of viewport window .
in contrast , similar opals grown on microscope slides showed peaks in excess of @xmath30 .
thus the quality of the opals was limited by the surface quality of the viewports . from optical microscope images ,
the structures were estimated to be between 1 and 3 lattice spacings ( i.e. , fcc 111 spacings ) thick . since we observed greenish opalescence from the deposited nanostructures ,
likely related to second - order bragg diffraction ( see below ) , we refer to such a structure as opal .
the light for the high - resolution measurements is focused at an angle of @xmath31 onto the experimental cell , as shown in fig .
[ fig : setup ] . at this angle
the stop bands of the opal are blue shifted to match with the cs resonance . moreover
, this procedure avoids the 2 mm thick window to act as a fabry - prot etalon , with a fringe spacing of @xmath32@xmath33 , or 50ghz , which is comparable to the scan range of the laser .
waveplate ( wp ) , and is monitored by diode d@xmath20 .
light reflected by the sample in the experimental cell is detected by a photomultiplier tube ( pmt ) and lock - in amplifier . to determine the optical frequency of the laser ,
a coarse measurement is performed with a wavemeter , and a high - resolution frequency is measured by saturated absorption spectroscopy in a cs cell with diode d@xmath1 .
light from a white halogen lamp is used to measure the broadband optical properties of the sample .
the sample cell is placed in an evacuated tank ( volume @xmath34m@xmath35 ) to minimize air convection around all heated elements .
, title="fig : " ] + light for the high - resolution experiments is generated by a narrowband single - mode cw laser ( coherent mbr-110 ) , pumped by a frequency doubled nd : yag laser ( coherent verdi v-10 ) , see fig .
[ fig : setup_large ] . the laser can be scanned over 40ghz near @xmath18 nm , sufficient to excite all four cs hyperfine transitions in one scan .
the laser light is linearly polarized , modulated by a chopper , and the power is set by polarization means .
the beam is focused by an achromatic lens ( @xmath36 mm ) at an angle of @xmath31 onto the experimental cell containing the cesium vapor and the opal .
the elastically reflected light is collected by a photomultiplier tube ( hamamatsu r928 ) .
the photomultiplier is essentially insensitive to fluorescence from the sample due to its very small solid angle of @xmath37sr .
the photocurrent of the photomultiplier is measured by phase - sensitive detection .
the laser frequency is coarsely monitored by a wavemeter ( burleigh wa-10l ) with a resolution of 0.1@xmath33 .
the precise frequency is measured by doppler - free spectroscopy @xcite with a resolution of 26 mhz using a cs reference cell @xcite .
broadband photonic crystal properties are probed with a spatially - filtered halogen white light source and the reflectivity is measured with a 2048-channel spectrometer ( ocean optics , usb 2000 ) in the visible and near infrared range .
opal grown on the viewport normalized to the reflectance @xmath38 of the viewport , at oblique incidence ( @xmath39 , internally @xmath40 ) .
the first peak at @xmath41 @xmath33 is identified with the first - order ( 111 ) stop gap of a sio@xmath20 opal .
the second reflectivity peak is attributed to the range of flat bands and higher - order stop gaps .
the horizontal dashed line indicates the resonance frequency of the cs d@xmath1 transition .
( b ) bandstructure from u to l in the irreducible part of the brillouin zone of an fcc opal photonic crystal .
the gray horizontal bar indicates the forbidden gap at an angle of @xmath42 inside the crystal .
, title="fig : " ] + figure [ fig : reflectivity_and_bandstructure](a ) shows the broadband reflectivity spectrum of a thin opal grown on a viewport .
a broad peak is observed at @xmath41 @xmath33 ( 324thz ) .
the large relative bandwidth of 48 @xmath43 is attributed to finite - size effects of the thin opal @xcite , also in view of the low reflectivity . at @xmath44 @xmath33 ,
the cs d@xmath1 resonance is indicated , which is slightly blue shifted compared to the first stop band , as aimed for .
a second high - frequency peak is observed near @xmath45 @xmath33 or 555thz .
the high - frequency peaks seems to be split into 3 separate peaks that likely originate from multiple stop - gap interferences and concomitant flat bands , observed earlier in opals and inverse opals @xcite .
the bandstructure from u to l for a sio@xmath20 opal ( n = 1.45 ) is shown in fig .
[ fig : reflectivity_and_bandstructure](b ) .
the peak at @xmath41 @xmath33 corresponds well to the frequency range of the directional forbidden stop gap at an angle of @xmath42 inside the crystal .
the broad high - frequency peak at @xmath45 @xmath33 corresponds well with the range of second - order bragg diffraction that includes complex stop gaps and flat bands to which external light can not couple . .
on the left the reflection of the opal is shown as a function of probe frequency for four different temperatures , where the first line of spectra is a reference at room temperature without cs .
the dashed arrow indicates the red - shifting second - order stop band . on the right ,
the corresponding transfer - matrix calculations are shown with the used index of the opal material .
the inset is an extreme zoom in into the region where cs resonances are expected .
the temperature was increased by 10 to 20@xmath46c per hour .
, title="fig : " ] + broadband reflectivity spectra are shown in fig .
[ stopbandsfig ] for increasing temperature and thus cesium vapor pressure . at first contact with cs (
not shown ) , even before spectra are observed , the first - order opal peak has shifted down from 324thz to 299thz ( 9.5% relative shift ) , and the second peak has shifted from 555thz to 489thz ( 12% ) .
at a cell temperature of @xmath47 k , the first - order peak has already shifted out of the instrumental spectral range , and the second peak has shifted down to 460thz . at a temperature of @xmath48
k , the second reflectivity peak has further shifted to 447thz , or 20% down compared to the starting state . at higher temperatures the peaks can not be identified unambiguously . the overall reflectivity decreases with increasing temperature , which is caused by a thin oil film that settled on the optics in the vacuum tank .
the shifting of the reflectivity peaks is attributed to the chemical reduction of the sio@xmath20 nanospheres @xcite ; since cs is strongly caustic , it reduces sio@xmath20 to sio@xmath49 @xcite .
if @xmath50 varies from @xmath51 to @xmath52 , the refractive index of the nanospheres will increase from @xmath53 for sio@xmath20 to at most @xmath54 in case of pure si @xcite .
if a homogeneous reduction to sio occurs , the refractive index of the nanospheres will be equal to that of sio ( @xmath55 ) .
we surmise that the colloidal nanospheres consist of a mixture of sio@xmath49 with different @xmath50 , possibly mixed with cs oxides . by substituting the increased average refractive index of the opal layer into bragg s law @xcite , the peak shifts to a lower frequency . to get more insight in the shape and the shifting of the observed resonances , a transfer - matrix calculation was performed on a single crystal layer of nanospheres surrounded by an atomic vapor , and probed at an angle of @xmath56 . at the right hand side of fig .
[ stopbandsfig ] , a corresponding transfer - matrix calculation of a single crystal layer is shown .
the nanosphere s refractive index is chosen to give the best overal match .
for the reference spectrum , before contact with cs , the agreement is reasonable ( first row ) .
the d@xmath1 resonance is just blue of the first peak . on first contact with cs , both the first and second peak red shift ( not shown ) . at @xmath57k ,
the second peak has shifted to 460thz , which is matched by the calculation for a nanosphere refractive index @xmath58 .
the d@xmath1 resonance is at the blue foot of the first - order peak . at @xmath48k ,
the d@xmath1 resonance is exactly in the trough between two peaks , obtained for @xmath59 .
as the temperature is increased even further , @xmath60 is fitted , shifting the second peak so that the d@xmath1 resonance is now on its red side .
we see that the increase in cesium density ( moderated by the increasing temperature ) and progression in time increase the nanosphere refractive index . as a result ,
the opal reflectivity peaks are tuned with respect to the atomic resonances .
compared to the broadband reflectivity , we increased the spectral resolution by @xmath61 to resolve the cs hyperfine structure .
therefore , intermixed with the measurements of the broadband spectra we have scanned the narrowband excitation laser and recorded the reflected intensity , as displayed in fig . [ fig : explanationfig ] , that zoom in on the cs d@xmath1 transitions . at a low cs density ( at @xmath62 k )
, the background reflection of the opal is low and the cs resonances appear as red - shifted peaks in reflectivity . to determine the linewidth , we modeled the peaks with a gaussian function and obtain about @xmath63 ghz fwhm . from earlier work @xcite
, we deduced that the peaks are both collision and doppler broadened .
if we consider the atoms to behave as tiny resonators , the width correspond to an effective quality factor @xmath64 .
while similar - or even higher - quality factors have been reported @xcite , the advantageous feature of the cs atoms studied here is that their center frequency varies by less than 1 part in @xmath65 (= 335 thz/0.4 ghz ) , which is beyond state - of - the - art in nanocavity fabrication .
therefore , atoms warrant attention as resonant systems in nanophotonic arrays . figure [ fig : explanationfig](a ) shows measured high - resolution reflectivity spectra at four moments in time while the temperature and hence the cs density was raised .
the two hyperfine resonances @xmath66 and @xmath67 are clearly observed . at @xmath47
k , the background reflectivity of the opal is around @xmath68 , or @xmath69 .
close to the two resonances , the reflectivity shows two clear peaks . below the @xmath70 transition frequency , the spectrum shows a short tail with a trough , and beyond the resonances the spectrum shows a broad and asymmetric tail .
when the second spectrum is taken , both peaks have become more pronounced .
the low - frequency tail has expanded , and the low - frequency trough has disappeared .
the background reflectivity of the opal has decreased to 0.013 .
the third spectrum has dramatically changed : after the first peak a trough has appeared that clearly dives below the opal s background .
similarly , a second trough with a minimum below the background has appeared at the second resonance . the initial trough below resonance has completely disappeared , and the baseline has recovered to 0.02 . the last spectrum , at @xmath71 k , shows another striking change : the peaks seen at the earlier spectra have completely vanished and given way to two marked troughs .
the spectra is completely without peaks , exactly the opposite of the second spectrum .
moreover , the opal background has recovered even further to 0.05 .
our observations can be qualitatively understood by considering a polarizable atomic medium inside a photonic structure , whose reflectivity peaks shift due to a changing opal refractive index .
these features have been incorporated in our transfer - matrix model , whose results are shown in fig .
[ fig : explanationfig](b ) .
we consider the real and imaginary parts of the atomic refractive index @xmath72 and @xmath73 , respectively , and their frequency relative to that of the opal s stopband ( see fig .
[ fig : explanationfig ] ) .
we distinguish 4 characteristic cases as illustrated in fig .
[ fig : explanationfig](c ) . 1 .
if the atomic resonance is at the blue edge of the reflectivity peak , an increase in @xmath72 from unity below resonance will effectivly cause the stop gap to red shift at that frequency .
a red shift at the blue edge of a peak reduces the reflectivity . above resonance ,
the opposite happens since @xmath72 is below unity .
the behavior of @xmath72 versus frequency across the resonance thus results in an decreasing and increased reflectivity , as seen at @xmath47 k. on resonance , the enhanced @xmath73 will destroy the destructive interference , thereby increasing the reflectivity .
2 . for resonances at the trough between two opal reflectivity peaks ,
the increase in @xmath73 on resonance will dominate the reflectivity and cause the two peaks , as seen on the second row where we fitted n=1.58 .
the gradient of the opal s reflectivity vanishes , and any change in @xmath72 will hardly change @xmath74 , since the opal s reflectivity peaks will be little shifted .
3 . at the red edge of the second opal peak ,
the increase in @xmath72 below resonance will red shift the opal peak . as a result the reflectivity will increase . above resonance ,
the decrease in @xmath72 will blue shift the opal peak , resulting in a decrease of the reflectivity .
hence , the measured reflectivity has a positive correlation with @xmath72 .
the result of increase in @xmath73 depends on whether the resonance is at the foot of the opal reflectivity peak or near the maximum : in the former case , the reflectivity increases , while in the latter case the constructive bragg interference is suppressed , resulting in a reflectivity trough .
4 . an atomic resonance near the center of the reflectivity peak will lead to following behavior : since the gradient of the reflectivity peak is small at its maximum , the induced shift by the change in @xmath72 will have little effect on the opal s reflectivity .
in contrast , the absorption expressed by the @xmath73 will remove the constructive bragg interference .
hence an increase in @xmath73 will correlate with a trough at resonance , as seen in the last spectrum , at @xmath71 k. it appears that the resonance frequencies in our experiments are systematically lower from calculated ones .
the difference is attributed to a pressure shift of remnant n@xmath20 in the reservoir . using the shift due to a n@xmath20 buffer gas of -10.97mhz / mbar from ref .
@xcite , a partial pressure of 27 mbar explains the difference .
this partial pressure agrees excellently with a measurement of @xmath75 mbar at evacuation after the experiment . from the good match between the measured spectra and the theory
curves we conclude that the system is well described by an atomic vapor filling the voids around the opal spheres .
moreover , from the agreement we deduce that the real part of the refractive index must have been less than unity for probe frequencies above resonance . to the best of our knowledge ,
this is the first observation of sub - unity refractive index in a photonic crystal , inside a stop band .
the striking spectral features reported above reveal that a hybridization occurs between the optical properties of the cesium atoms and the opal structure .
the result of the hybridization depends strongly on the tuning of the atomic resonances relative to the opal s gap structure . the simple case anticipated in fig .
[ fig : concept ] occurs when the atomic resonances are centred in the gap .
this case corresponds to the observations at @xmath71 k in fig .
[ fig : explanationfig ] .
in addition , intriguing hybridizations are observed when the atomic resonances are tuned to either the red or the blue edge of a stop gap . in these cases ,
the real parts of the atomic dispersion come into play , leading to asymmetric reflectivity features in the observed spectra .
this situation differs from arrays of cavities in photonic crystals . in the latter case , the cavities
must necessarily be tuned deep within the gap , in order to profit from the confinement offered by the gaps .
conversely , a cavity resonance tuned to the edge of a gap will obviously be lossy , since the cavity resonance is derived from the surrounding photonic crystal , in contrast to an atomic resonance , that obviously also exists outside a photonic crystal .
in contrast to the one - dimensional ( 1d ) cases of long hollow optical fibers @xcite or waveguides on a chip @xcite , three - dimensional ( 3d ) photonic crystals have the demonstrated potential to significantly alter the local density of states ( ldos ) @xcite , allowing to control the excited - state lifetime of emitters more in 3d @xcite than in 1d @xcite .
our results demonstrate that the required high atomic densities can be achieved .
several improvements merit investigation in future studies .
first of all , other combinations of gases and photonic crystals will have to be more stable against chemical interactions .
another option is to coat the opal with a thin protective layer or to use light - induced adsorption to increase the alkali pressure .
once these issues are resolved , we can investigate the influence of the ldos on the lifetime of excited states .
the high-@xmath3 resonances in a photonic crystal could even allow to study the thermalization of a photon gas @xcite .
finally , our experiment could offer an alternative experimental route to complex experiments that are being done with atomic lattices @xcite .
the measured peaks are somewhat broader than those calculated with our model .
the model does not take into account collisional broadening or transit time broadening .
while collisional self - broadening @xcite does not play a significant role at our experimental conditions , the n@xmath20 background broadens a delta - function line to about 400mhz fwhm @xcite .
in addition , the atoms between the opal spheres will have only a small interaction time with the light before hitting a material wall or before being outside the optical focus , leading to transit - time broadening @xcite . since the magnitude strongly depends on geometry , it is hard to estimate , yet it could easily amount to a few 100mhz . in view of these broadening mechanisms ,
the observed resonances are surprisingly narrow , suggestive of selective reflection effects @xcite .
in selective reflection , slow atoms with only a small velocity component parallel to the propagation direction of light contribute disproportionately to the signal , leading to sharp spectral features .
the prospect of exploiting selective reflection to yield even sharper features in future makes this system even more attractive .
we report on an experiment in which we infiltrate a dense alkali gas - a saturated cesium vapor - in a thin opal photonic crystal .
the stop bands of the nanostructure are tuned to overlap with the narrow atomic resonances of the alkali atoms , leading to an intriguing hybridization of the atomic resonances and the crystal spectrum .
as the temperature is increased , we observe a shift of the peak opal reflectivity frequency in excess of @xmath76 , which is attributed to the chemical reduction of the silica by the cesium . as a result
, the frequency of the photonic bands are scanned through the near - infrared cs d@xmath1-transition .
simultaneously , the cs resonances undergo marked changes in strength , off - resonance reflectivity , and lineshape . when the atomic resonance is tuned to the red edge of the stop gap
, the reflectivity reveals a striking inversely dispersive shape .
we successfully interpret all observed spectral features by using a transfer - matrix model that includes the dispersion and absorption of two of the hyperfine transitions of cs , as well as the ongoing change of the opal s refractive index due to the chemical reduction .
our study is a step forward in the understanding of strongly resonant systems inside photonic crystals , and shows that a hybrid combination of atomic and condensed matter holds interesting physics .
it is a pleasure to thank cock harteveld , and leon woldering for expert technical support .
we thank javier garca de abajo , tom hijmans , ad lagendijk , and gerhard rempe for encouragements on the project .
this work is part of the research program of the stichting voor fundamenteel onderzoek der materie ( fom ) that is financially supported by the nederlandse organisatie voor wetenschappelijk onderzoek ( nwo ) , notably the ppom program and `` inrichting leerstoelpositie '' grant ( 02ilp012 ) .
we also thank erc ( grant no 279248 ) , nwo ( vici ) , and stw for support .
a. yariv , y. xu , r. k. lee , and a. scherer , opt . lett . * 24 * , 711 ( 1999 ) .
m. j. hartmann , f. g. s. l. brandao , and m. b. plenio , nat .
* 2 * , 849 ( 2006 ) .
a. d. greentree , c. tahan , j. h. cole , and l. c. l. hollenberg , nat .
* 2 * , 856 ( 2006 ) y. akahane , t. asano , b .- s .
song , and s. noda , nature ( london ) * 425 * , 944 ( 2003 ) .
e. kuramochi , m. notomi , s. mitsugi , a. shinya , t. tanabe , and t. watanabe , appl .
88 * , 041112 ( 2006 ) .
u. p. dharanipathy , m. minkov , m. tonin , v. savona , and r. houdr http://arxiv.org/abs/1311.0997 h. altug and j. vukovi , appl .
phys . lett . *
84 * , 161 ( 2004 ) .
m. notomi , e. kuramochi , and t. tanabe , nature photon .
* 2 * , 741 ( 2008 ) .
a. f. koenderink , a. lagendijk , and w. l. vos , phys .
b * 72 * , 153102 ( 2005 ) .
y. takahashi , h. hagino , y. tanaka , b .- s .
song , t. asano , and s. noda , opt .
expr . * 15*. 17206 ( 2007 ) . c. f. bohren and d. r. huffman , _ absorption and scattering of light by small particles _
( wiley , new york , 1983 ) .
d. v. van coevorden , r. sprik , a. tip , and a. lagendijk , phys . rev . lett . * 77 * , 2412 ( 1996 ) . a. y. sivachenko , m. e. raikh , and z. v. vardeny , phys .
a * 64 * , 013809 ( 2001 ) .
y. q. ye , j. q. shen , and y. jin , appl .
a * 93 * , 505 ( 2008 ) .
j. tidstrm , c. w. neff , and l. m. andersson , j. opt .
* 12 * 035105 ( 2010 ) .
m. soljacic and j. d. joannopoulos , nat . mater .
* 1 * , 3 ( 2004 )
. r. m. camacho , m. v. pack , j. c. howell , a. schweinsberg , and r. w. boyd , phys .
lett . * 98 * , 153601 ( 2007 ) .
n. eradat , a. y. sivachenko , m. e. raikh , z. v. vardeny , a. a. zakhidov , and r. h. baughman , appl .
phys . lett . *
80 * , 3491(2002 ) .
a. imhof , m. megens , j. j. engelberts , d. t. n. de lang , r. sprik , and w. l. vos , j. phys .
b * 103 * , 1408 ( 1999 ) . d. j. norris and m. g. bawendi , phys . rev .
b * 53 * , 16338 ( 1996 ) .
m. weidemller , a. hemmerich , a. grlitz , t. esslinger , and t. w. hnsch , phys .
* 75 * , 4583 ( 1995 ) .
g. birkl , m. gatzke , i. h. deutsch , s. l. rolston , and w. d. phillips , phys .
lett . * 75 * , 2823 ( 1995 ) .
m. greiner , o. mandel , t. esslinger , t. w. hnsch , and i. bloch , nature ( london ) * 415 * , 39 ( 2002 ) .
p. ballin , e. moufarej , i. maurin , a. laliotis , and d. bloch , appl .
lett . * 102 * , 231115 ( 2013 ) .
p. j. harding , _ photonic crystals modified by optically resonant systems _
thesis , university of twente , 2008 ) , available from www.photonicbandgaps.com .
l. young , w. t. hill iii , s. j. sibener , s. d. price , c. e. tanner , c. e. wieman , and s. r. leone , phys .
a * 50 * , 2174 ( 1994 ) .
r. j. rafac and c. e. tanner , phys .
a * 58 * , 1087 ( 1998 ) .
r. j. rafac , c. e. tanner , a. e. livingston , and h. g. berry , phys .
a * 60 * , 3648 ( 1999 ) .
h. miguez , f. meseguer , c. lpez , a. mifsud , j.s .
moya , and l. vazquez , langmuir * 13 * , 6009 ( 1997 ) .
m. megens and w.l .
vos , phys .
* 86 * , 4855 ( 2001 ) .
p. jiang , j. bertone , k. hwang , and v. colvin , chem .
mater . * 11 * , 2132 ( 1999 ) .
a. hartsuiker and w. l. vos , langmuir * 24 * , 4670 ( 2008 ) .
j. f. bertone , p. jiang , k. s. hwang , d. m. mittleman , and v. l. colvin , phys .
* 83 * , 300 ( 1999 ) w. l. vos and h. m. van driel , phys .
lett . a * 272 * , 101 ( 2000 ) .
s. g. romanov , t. maka , c. m. sotomayor torres , m. mller , r. zentel , d. cassagne , j. manzanares - martinez , and c. jouanin , phys .
e * 63 * , 056603 ( 2001 ) .
j. f. galisteo - lpez and c. lpez , phys .
b * 70 * , 035108 ( 2004 ) .
e. jahier , j. guna , p. jacquier , m. lintz , and m.a .
bouchiat , europhys .
j. d * 13 * , 221 ( 2001 ) .
a. andalkar and r. b. warrington , phys .
a * 65 * , 032708 ( 2002 ) .
w. l. vos , r. sprik , a. van blaaderen , a. imhof , a. lagendijk , and g. h. wegdam , phys .
b. * 53 * , 16231 ( 1996 ) . c. l. chen and a. v. phelps , phys .
rev . * 173 * , 62 ( 1968 ) .
w. demtrder , _ laser spectroscopy _ , ( springer , berlin , 2003 ) , 3rd ed .
j. p. woerdman and m. f. h. schuurmans , opt
* 14 * , 248 ( 1975 ) ; m. f. h. schuurmans , j. phys .
( france ) * 37 * , 469 ( 1976 ) ; a. l. j. burgmans and j. p. woerdman , j. phys .
( france ) * 37 * , 677 ( 1976 ) ; g. nienhuis , f. schuller , and m. ducloy , phys .
rev . a , * 38 * , 5197 ( 1988 ) ; m. ducloy and m. fichet , j. phys .
( france ) ii * 1 * , 1429 ( 1991 ) ; m. chevroller , m. fichet , m. oria , g. rafmat , d. bloch , and m. ducloy , j. phys .
( france ) ii * 2 * , 631 ( 1992 ) .
p. wang , a. gallagher , and j. cooper , phys .
rev . a * 56 * 1598 ( 1997 ) .
d. mller , e. a. cornell , d. z. anderson , and e. r. i. abraham , phys .
rev . a , * 61 * , 033411 ( 2000 ) .
s. ghosh , a. r. bhagwat , c. k. renshaw , s. goh , a. l. gaeta , and b. j. kirby , phys .
* 97 * , 023603 ( 2006 ) . p. s. light , f. benabid , f. couny , m. maric , and a. n. luiten , opt . lett . *
32 * , 1323 ( 2007 ) .
a. d. slepkov , a. r. bhagwat , v. venkataraman , p. londero , and a. l. gaeta , opt . expr .
* 16 * , 18976 ( 2008 ) .
w. yang , d. b. conkey , b. wu , d. yin , a. r. hawkins , and h. schmidt , nat .
* 1 * , 331 ( 2007 ) .
p. lodahl , a. f. van driel , i. s. nikolaev , a. irman , k. overgaag , d. vanmaekelberg , and w. l. vos , nature ( london ) * 430 * , 654 ( 2004 ) .
s. noda , m. fujita , and t. asano , nat .
* 1 * , 449 ( 2007 ) . n. vats , s. john , and k. busch , phys
a * 65 * , 043808 ( 2002 ) . c. hooijer , d. lenstra , and a. lagendijk , opt
. lett . * 25 * , 22 , 1666 ( 2000 ) j. klaers , f. vewinger , and m. weitz , nat . phys .
* 6 * , 512 ( 2010 ) .
i. bloch , j. dalibard , and w. zwerger , rev .
phys . * 80 * , 885 ( 2008 ) . | we study a hybrid system consisting of a narrowband atomic optical resonance and the long - range periodic order of an opaline photonic nanostructure . to this end , we have infiltrated atomic cesium vapor in a thin silica opal photonic crystal . with increasing temperature
, the frequencies of the opal s reflectivity peaks shift down by @xmath0 due to chemical reduction of the silica .
simultaneously , the photonic bands and gaps shift relative to the fixed near - infrared cesium d@xmath1 transitions . as a result
the narrow atomic resonances with high finesse ( @xmath2 ) dramatically change shape from a usual dispersive shape at the blue edge of a stop gap , to an inverted dispersion lineshape at the red edge of a stop gap .
the lineshape , amplitude , and off - resonance reflectivity are well modeled with a transfer - matrix model that includes the dispersion and absorption of cs hyperfine transitions and the chemically - reduced opal .
an ensemble of atoms in a photonic crystal is an intriguing hybrid system that features narrow defect - like resonances with a strong dispersion , with potential applications in slow light , sensing and optical memory . |
we imagine that the strong flare in mcg-6 - 30 - 15 originates in a compact reconnection site elevated to a height @xmath0 above the surface of the accretion disk .
the radiation from this primary source partly shines toward the disk and creates a hot spot . the distance of the spot s center to the disk center
is denoted by @xmath1 .
the flare is supposed to be in keplerian co - rotation with the disk and we assume that the primary illumination sets on and fades out instantaneously .
the irradiation of the disk then evolves across the hot spot , starting from the spot center and progressing toward the border .
therefore we expect the lightcurve of the reprocessed radiation to be curved even if the time evolution of the primary is box - shaped .
we want to model the exact shape of the lightcurve expected from such a flare setup at different orbital phases of the disk .
we first conduct detailed local radiative transfer computations .
the varying intensity of the irradiation across the hot spot is taken into account as well as the angular dependence of the reprocessed emission .
the vertical structure of the disk is assumed to remain in the same hydrostatic equilibrium as before the onset of the flare .
the profile is computed with an extended version of the code described in ( * ? ? ? * raska ( 2002 ) ) .
the local spectra across the spot are then computed by the codes titan and noar ( ( * ? ? ?
* dumont 2000 , 2003 ) ) .
we assume a black hole mass @xmath2 , a disk accretion rate @xmath3 ( in units of the eddington accretion rate ) , and @xmath4 ( with @xmath5 ) . based on the local spectra , the relativistic ray - tracing code ky ( ( * ? ?
* doviak 2004 ) ) computes the time evolution of the spectrum seen by a distant observer .
we set the disk inclination to @xmath6 and assume a kerr parameter @xmath7 having known characteristics of mcg-6 - 30 - 15 in mind .
in ky we define the time - dependent disk emission according to the subsequent illumination and fade - out of the spot from its center toward the border .
the duration of this time sequence is normalized by @xmath0 , which we constrain from the observed soft - to - hard time delay of @xmath8 s ( ( * ? ? ?
* ponti 2004 ) ) .
we assume that this delay is entirely due to the light traveling time between the source and the disk leading to @xmath9 .
the flare lasts for 2000 sec , which at @xmath4 corresponds to @xmath10 of the keplerian time scale .
more details about this type of flare model are given in ( * ? ? ?
* goosmann ( 2006 ) ) . in figure
[ fig : lcs ] we show the model lightcurves obtained for flares at 8 different orbital positions .
they show significant differences in shape and in normalization when comparing different orbital phases .
all curves have a broadened maximum which does not match the observed , peaked shape ( see fig . 4 in ( * ? ? ?
* ponti 2004 ) ) .
thus , even if the flare is reflection - dominated , neither the geometrically evolving irradiation across the spot nor relativistic and doppler modifications can account for the observed shape of the lightcurve .
an intrinsic time evolution of the primary source is therefore required . ]
* ponti ( 2004 ) ) conducted a timing analysis for the observed flare and computed cross - correlation functions and time delays between six different energy bands @xmath11 ( see their fig .
we attempt to reproduce the observed time delays using the following , simple mathematical parameterization for the primary radiation @xmath12 and the reprocessed component @xmath13 : these terms account for the intrinsic change of the primary that is suggested by the modeling presented in section [ sec : lcs ] .
the overall factor @xmath20 normalizes @xmath13 against the primary and @xmath21 denotes an intrinsic delay between the two components
. the time evolution of the reprocessing can be broadened by choosing @xmath22 .
this represents a spread of the light travel - time between the source and the reprocessing site . for further details on this model see ( *
* goosmann ( 2007 ) ) . where the lightcurve @xmath27 belongs to the energy bin @xmath28 .
we adopt here the very same method as was used in the analysis of ( * ? ? ?
* ponti ( 2004 ) ) .
the values for @xmath29 sec and @xmath30 sec are adjusted to the flare lightcurve plotted in fig . 4 of ( * ? ? ?
* ponti ( 2004 ) ) .
the spectral slope of the reprocessed component has a major impact on the obtained energy - dependent time delays .
we choose a rather hard slope of @xmath13 and set @xmath31 .
such a spectral shape corresponds to the reprocessing in a medium of magnetically confined , cold clouds as analyzed by ( * ? ? ?
* kuncic , celotti , & rees(1997 ) ) . their fig .
2 shows reprocessed model spectra that are much harder than the primary due to multiple soft x - ray absorption by the individual clouds . for the primary slope we set @xmath32 having in mind that for the higher flux state during the flare the spectrum should steepen ( ( *
* shih , iwasawa , & fabian 2002 ) ) .
the model can reproduce the observed time lags for the observed flare of mcg-6 - 30 - 15 . in fig .
[ fig : mcg - delay - fit ] we show two examples of satisfactory data representation
. the results account for the slight flattening of the delay curve toward higher energies .
it turns out that different choices of parameters can lead to similar delay curves , as there are three free parameters involved ( @xmath20 , @xmath33 , and @xmath21 ) and the error bars of the data are large .
to distinguish between different parameter sets some fine tuning of the delay curves shape would be necessary .
this is not meaningful with the current data .
the observed time - delay can not be reproduced when the spectral slope of @xmath13 is softer , i.e. adjusted to an ionized reflection scenario as proposed by ( * ? ? ?
* ballantyne , vaughan , & fabian(2003 ) ) .
if the primary source is located within @xmath39 of the black hole , the relative variability of the primary and the reflection component can be reproduced ( ( * ? ? ?
* miniutti & fabian 2004 ) ) .
such a setup also explains the high equivalent width of the iron k@xmath40 line , which requires a relatively large amount of reflection . while these ionized reflection and light - bending models
are certainly important to explain the behavior of mcg-6 - 30 - 15 on longer time - scales , the data obtained during the 2000 sec flare period does not allow to conclude on the time evolution of the iron k@xmath40 line .
from the results of our delay modeling , we therefore suggest the possibility that the strong individual flare does not originate in the inner accretion flow , where most of the x - ray energy is dissipated .
we rather imagine that the flare occurred farther away from the black hole , in a region where the disk is fragmented and replaced by magnetically confined , cold clumps . | we discuss implications of a strong flare event observed in the seyfert galaxy mcg-6 - 30 - 15 assuming that the emission is due to localized magnetic reconnection .
we conduct detailed radiative transfer modeling of the reprocessed radiation for a primary source that is elevated above the disk .
the model includes relativistic effects and keplerian motion around the black hole .
we show that for such a model setup the observed time - modulation must be intrinsic to the primary source . using a simple analytical model
we then investigate time delays between hard and soft x - rays during the flare .
the model considers an intrinsic delay between primary and reprocessed radiation , which measures the geometrical distance of the flare source to the reprocessing sites .
the observed time delays are well reproduced if one assumes that the reprocessing happens in magnetically confined , cold clouds
. keywords .
galaxies : active , galaxies : seyfert , x - rays : individual ( mcg-6 - 30 - 15 ) the seyfert galaxy mcg-6 - 30 - 15 was observed for 95 ksec with _ xmm - newton _ in the year 2000 ( ( * ? ? ?
* wilms 2001 ) ) .
the x - ray lightcurve of this observation reveals a strong flare event of which ( * ? ? ?
* ponti ( 2004 ) ) conducted a detailed analysis . in this
proceedings note we discuss the possibility that the flare is produced by localized magnetic reconnection and we constrain some details of such a flare setup . |
worldwide there is a great interest of governments and organizations involved in public safety and security ( pss ) towards the evolution of existing wireless systems for critical communications based on professional mobile radio ( pmr ) technologies such as terrestrial trunked radio ( tetra ) , tetra for police ( tetrapol ) , or association of public - safety communications officials - project 25 ( apco p25 ) .
it is widely recognized that efficient communications are of paramount importance to deal with emergency or disaster situations .
the possibility of pss operators to use a wide range of data - centric services , such as video sharing , files transmission ( e.g. , maps , databases , pictures ) , ubiquitous internet and intranet access , has a strong impact on the efficiency and the responsiveness of the emergency services . however , pmr systems are based currently on 2 g networks , thus offering a wide range of voice services , but having a limited possibility to support data services . even if some efforts have been done to enhance pmr systems and to offer higher communication capacity , achievements are still behind those made in the commercial world that recently has developed the 3gpp long term evolution ( lte ) technology .
hence , there is a great consensus in adopting the commercial lte framework to answer to the pss communication needs . a common standard for commercial and pss environments can open the door to new opportunities and synergies , offering advantages to both worlds .
however , the lte technology needs some specific enhancements to be fully compliant with the mission and safety critical requirements .
indeed , pmr networks must be reliable , secure and resilient , guaranteeing service accessibility and wide coverage as well .
in addition pss operators need some specific applications and functionalities such as push - to - talk , dispatch services , priority management , group communications and direct communications @xcite .
therefore , adopting lte as pmr broadband technology needs that these features are included in the future releases of the 3gpp standard also guaranteeing interoperability with actual narrowband pmr systems . in this paper ,
[ req ] analyses the main requirements of pmr communications .
the use of lte for critical communications and the description of critical services currently not supported by the lte standard is given in sec .
[ lte - pss ] underling the need of further research activity .
[ ap ] proposes an evolution path of the pmr lte - based network architecture .
finally the attention is devoted to group communications in sec .
[ gc ] , where the recommendations that are under evaluation by the 3gpp are analysed and on their basis our proposal is detailed .
conclusions are drawn in sec .
mission critical communications are characterized by different and more severe requirements respect to commercial communications . in particular
, we can distinguish between typical narrowband pmr and new broadband requirements and functionalities : * * pmr narrowband requirements and functionalities * @xcite * * _ high reliability and availability .
_ the system shall be available for 99% of time in 24 hours and 99.9% of time in a year .
it shall cover 96% of the area in outdoor environment and 65% in indoor .
* * _ half - duplex and full - duplex voice calls_. * * _ fast call setup time _ lower than 300 @xmath0 @xcite . * * _ push to talk _ management in half - duplex communications . * * _ call priority and preemption_. the system shall assign different levels of priority to calls and interrupt low priority calls on arrival of high priority calls that do not find available resources . * * _ direct mode communications _ between terminals without the support of network infrastructure . * * _ text message service _ , e.g. , the short data service of tetra system . * * _ network interoperability_. communications with users located on _ external networks _ or pstn ( public switched telephone network ) . * * _ emergency calls_. * * _ group calls_. gcs allow a user to talk simultaneously with several users belonging to a group that can be predefined or formed on - demand
. the network shall be able to permit the coexistence of many active groups at the same time .
a real life scenario comprises an average of 36 voice groups corresponding at least to 2000 users in an area .
it is expected that up to 500 users can participate in a group @xcite . moreover ,
each user can be registered to many groups at same time . * * pmr broadband requirements and functionalities * @xcite * * _ data communications _ for fax and image transfer . * * _ synchronous video transmission _ that consists of a bidirectional communication composed of different data , audio and signaling flows at 256 kbps . in order to fulfil all the above requirements and functionalities
it has been evaluated that a bandwidth of 10mhz is needed @xcite . as a consequence
one open point is to identify an harmonised frequency band allocation for the new pmr systems that takes into account different national allocation policies , interferences towards other existing systems and economic convenience .
this section describes the main communication features that shall be gradually introduced in the future lte standard releases in order to satisfy the typical needs of public safety organization .
indeed , currently the lte system does not provide services considered vital for the pss context .
direct connection among devices ( direct mode operation - dmo ) is a mandatory feature for pmr systems .
it allows pss users to communicate without the involvement of the network infrastructure , e.g. , if the network is not available due to a failure or lack of coverage .
current pmr systems foresee the use of dmo for direct connection between user terminals ( back - to - back ) , but also between the users and special terminals that operate as relay and/or gateway towards the trunked mo ( tmo ) network .
the 3gpp is working to introduce in the lte standard the direct communication as a new service named _
proximity service _ ( prose ) .
in particular the technical report @xcite deals with the definition of an advanced network architecture able to support prose service either for commercial or pss applications .
the service definition is on going and several proposals are under investigation .
in general we can state that the prose foresees two different operative modes : * network assisted * not network assisted in the first case the network assistance is required to authenticate terminals , allocate resources and manage the security .
the network provides to terminals the set of parameters needed for the call management . in the second case ,
the connection is activated directly by the terminals without any network involvement , using parameters already known by the user equipment ( ue ) and pre - allocated resources . for commercial applications
only the network assisted mode is considered , while for pss applications both modes are allowed because in this case the prose shall guarantee ues in proximity to communicate under any network condition . in network assisted solutions we can also distinguish between : * _ full network control _ :
the direct link among the ues is handled by the network including control ( connection set - up , maintenance ) and data planes .
the communications occur on licensed bands and the resources can be allocated dynamically or in a semi - static mode by the network ; * _ partial network control _ :
the network is involved only in the authentication and authorization phase , but the direct connections between the ues are initialized in autonomous way .
usually the resources are pre - allocated to this kind of service .
even if broadband services are gaining importance in emergency and rescue operations , voice call still remain a fundamental service even for future pmr systems
. however , this appears to be in conflict with the lte architecture that is full - ip based and mainly developed for data communications , hence , unable to support the circuit switching mode typical for classical voice calls . in lte
the implementation of voice service is following an evolution path starting from circuit switching fall back approach that relies on legacy networks @xcite . in pmr environment
this implies a transition phase during which voice service will be supported by the existing pmr systems . at a later stage
, voice calls will rely on voice over lte ( volte ) , an advanced profile defined by global system for mobile communications association ( gsma ) , based on ip multimedia subsystem architecture ( ims ) .
volte supplies basic functions as call establishment and termination using session initiation protocol ( sip ) , call forwarding , calling i d presentation and restriction , call - waiting and multiparty conference .
however , volte does not support primary functions of mission critical communications , i.e. , group and direct communications .
although ims supports open mobile alliance ( oma ) push - to - talk over cellular ( poc ) , which allows up to 1000 multicast groups and 200 users in one group , it is not suitable for critical communications because of the high call set - up time that reaches 4000 @xmath0 in case of pre - established session and 7000 @xmath0 if the session has to be created @xcite
. thus the adoption of poc needs some specific network solutions@xcite .
another important aspect of volte is the lack of a suitable security level due to absence of end - to - end encryption mechanism typical of pmr communications @xcite .
communication security is of paramount importance in pss networks whose main aspects are encryption , authentication , provisioning and user management
. however , the details concerning security are out of the scope of this paper .
the interested reader can refer to @xcite for some examples of promising approaches on how to face the main security issues in a public safety network .
the group call ( gc ) represents one of the most important and indispensable service of pmr networks .
it enables an efficient management of the rescue teams and permits to send commands to all the pss operators in a disaster area and share information . for the above reasons ,
the remaining part of this paper will be devoted on the discussion and proposal of approaches on how to manage gcs in a public safety lte - based network .
the communications addressed to more than one user can be distributed with a point - to - point ( p2p ) approach , using one _ unicast _ transmission for each involved user , or with point - to - multipoint ( p2 m ) flows .
+ in lte the distribution of _ multicast _ communications is demanded to the evolved multimedia broadcast multicast service ( embms ) , introduced in the 3gpp standard starting from release 10@xcite . in particular , it t has been mainly developed for multicast and broadcast distribution of multimedia data , such as video streaming provided by external sources that act as service providers . for this reason only a downlink channel is considered .
providing mission critical services in a public safety lte network involves a proper network architecture solution in order to achieve and maintain the required performance and reliability levels .
hence , it easy to foresee that several research efforts and further releases of the lte standard will be needed before this task will be fully accomplished . as a consequence , in the short term
the simplest way for enabling high data rate services for pmr networks based on lte , is the use of already deployed commercial networks .
3gpp provides recommendations and technical specifications for the sharing of network devices and modules between operators , commonly used in commercial networks .
lte network is composed by several modules , which can be collected in different logical domains : * the _ services domain _ , managing the contents and the services to be provided to ues , injecting the traffic in the network through a service delivery platform ; * the _ evolved packet core _ ( epc ) network , mainly responsible of the control functions ; * the _ radio access network _ ( ran ) , composed by the enodebs ; * the _ users equipment _ domain .
while the service domain shall be dedicated to pmr communications , due the specific services to be provided , the core network and the radio access network can be fully or partially shared between the professional and the public networks , as shown in fig .
[ network_evol ] . in an initial phase the pmr network can rely completely on a commercial network , limiting to provide non mission critical professional services .
this is mainly due to the fact that the public safety agencies can not control the fulfilment of the quality of service requirements , as well as the reliability and availability of the network , that are fully managed by carrier operators @xcite . from the network architectural point of view , in the first two configurations represented in the figure , the pmr operator acts as a mobile virtual network operator ( mvno ) , in core network sharing or full ran sharing configuration , respectively .
the two networks partially sharing their epc can follow the gateway core network ( gwcn ) specifications of 3gpp @xcite . in the third configuration
the professional network expands to the eutran , with a partial coverage of the territory .
the deep diffusion of the network on the area of interest is one of the major obstacle on the upgrade of current pmr networks over new and more efficient communication technologies .
lte allows the sharing of part of the eutran between operators , following the multiple operator core network ( mocn ) 3gpp specification .
the last presented configuration represents an autonomous lte pmr network , with full radio coverage of the territory . in this case
the pmr operator is the owner of the eutran , but some forms of passive sharing , such as mast sharing or site sharing , can be anyway implemented for reducing costs and environment impacts of the new network deployment .
as stated before future lte - based pmr networks are expected to support bandwidth consuming applications such as video , imaging and data communications .
in addition , multiple gcs shall be able to handle a great number of ues at the same time and in geographical areas served by several neighbour enbs ( i.e. , contiguous cells ) . in this section
we analyse recommendations provided by 3gpp @xcite , @xcite for the deployment of the gc service in lte networks .
these focus mainly on multicast transmission mode , in particular on the lte embms framework , even if unicast transmission is taken also into account .
the advantage of the unicast transmission scheme for the delivery of a gc service is that each unicast link can be tailored to the propagation conditions experienced by the served ue through the channel state information feedback and/or by acknowledge ( ack ) messaging .
conversely , in the multicast transmission the most adverse propagation conditions must be always considered in order to meet specific service requirements . however , the use of the unicast mode is not recommended in scenarios involving the delivery of multimedia contents to large groups of users , as the ones expected for future pmr networks . in this case , the use of the multicast mode provides a more efficient and reliable solution .
indeed , the main benefit of using p2 m flows is that the same content can be received by many users at the same time with a bandwidth and power consumption not dependent on the number of simultaneous users .
+ the embms service combines the advantages of p2 m transmissions with geographical and temporal flexibility .
indeed , the multicast distribution over embms is not extended necessarily on the entire network , but its scope can be limited to a small geographical area , such as a city center or a stadium , as well as large areas or regions . + in particular , the embms framework provides two kinds of transmission schemes , namely single cell ( sc ) and single frequency network ( sfn ) . in sc - embms mode , each enb delivers the mbms data flow independently from the others . on the other hand , in sfn - embms mode multiple enbs
are synchronized in order to transmit the same physical signal at the same time .
the enbs involved in the distribution of the same multicast / broadcast flow form a _
mbms area_. one enb can belong to more than one mbms area simultaneously .
the sfn - embms transmission mode leads to a significant improvement in terms of cell coverage and spectral efficiency , since that at the ue receiving side the signals coming from multiple enbs are combined resulting in a higher useful power .
[ fig.rate_gc ] shows a performance comparison between sc and sfn ( with 4 enbs ) embms transmission schemes in terms of system throughput as a function of the number of active group calls .
performance has been derived for two types of service ( video flow @ @xmath1 kbps at the application layer and voice call @ @xmath2 kbps @xcite ) and bandwidths .
we can see that the throughput increases with the number of active gcs up to a maximum value .
then it decreases rapidly due to a saturation of the system .
the benefits of using sfn are evident .
another advantage of the sfn - embms scheme for pmr networks is that the mbms area can be adapted in order to cover the whole critical area involved in pss operations .
the lte standard schedules unicast and multicast / broadcast services in separated subframes . in particular , in each lte radio frame up to 6 out of 10 subframes can be reserved for the delivery of embms services .
each embms service is mapped into a mbms bearer service and into a logical _ multicast traffic channel _ ( mtch ) .
then the mtchs are multiplexed with a logical multicast control channel ( mcch ) and mapped into a physical multicast channel ( mch ) . from an architectural point of view
, the embms framework introduces new modules for the set - up and management of multicast and broadcast contents .
in particular , the _ broadcast - multicast service center _ ( bm - sc ) represents the interface between the core network ( epc ) and the _ content provider _ , and it is responsible for the scheduling of mbms services over the lte network .
the _ mbms gateway _ ( mbms - gw ) is a logical node that transmits the multicast flow towards the enbs that belongs to the mbms area . finally ,
the _ multicell coordination entity _
( mce ) is a logical node that ensures that all the enbs of the mbms area use the same radio resource configuration .
the embms framework in the lte standard is designed only for downlink distribution of multimedia data flows .
hence , it has to be modified in order to support different services such as the gc .
to this goal 3gpp proposes to introduce a new logical interface , called _ group communications system enabler _ ( gcse ) , which is in charge of the creation , management and deletion of gcs .
the proposals under evaluation @xcite are mainly based on the multicast communication approach , the unicast mode is enabled whenever a multicast transmission can not be accomplished . in both cases
all the group members belonging to a gc perform uplink communications through the establishment of unicast bearers .
the gc management by the gcse must ensure also the meeting of the pmr network requirements such as short call latencies , service continuity for ues moving from one cell to another , security and mechanisms for priority and pre - emption .
assuming the embms framework and gcse as basis , 3gpp indicates three different architectural approaches that could be adopted @xcite .
the difference among these depends on the network element that manage the gcs .
the first defines that the epc handles the gcs , while the second approach specifies a decentralized solution where the management of gcs is in charge of the enb .
finally , the third approach is based on ims .
in accordance with the indications of the 3gpp , we outline here a promising solution for gcs in the future lte - based pmr networks and support our vision by providing suitable performance evaluations and comparisons .
the envisaged solution is based on embms framework and gcse as previously described , and assumes a centralized approach for the call management ( i.e. , performed by the epc ) that leads to benefits in terms of coverage and reliability . in particular
, we will discuss herein two possible approaches .
the first is fully based on multicast transmissions , named _
static embms activation _
, where the unicast mode is considered only as backup solution when the multicast transmission is not available , e.g. , due to bad propagation conditions or because the ue moves out of a mbms area .
unicast mode is activated only to provide service continuity when the ue detects an increasing packet loss .
a second , more advanced solution , named _ dynamic embms activation _
, is considered in order to increase the system flexibility and efficiency . _ dynamic embms activation _ manages group communications over both unicast and multicast transmissions .
the idea is based on the ability to perform a counting procedure for the gcse , with the aim to transmit downlink media contents through multiple unicast bearers until a certain number of users requires the same gc service , hence to activate a multicast bearer .
the advantage is that the gc is never broadcast in a cell with no active group members , and , hence , the spectrum efficiency is maximized .
for both solutions , the network architecture is represented in fig .
[ fig.network ] , where the gcse acts as an _ application server _
( gcse as ) providing the functionality for the management of gcs .
in addition , the gcse uses the existing lte standardized interfaces in order to communicate with the epc and selects the proper transmission scheme for the delivery of gcs .
finally , multicast bearers are considered for downlink communications whenever possible . on the other hand , uplink traffic
is always sent via unicast bearers .
gcse provides a multicast to unicast switching mechanism depending on the selected approach .
hence , the gcse is connected to both the pgw ( for the distribution of unicast traffic ) and to the bm - sc ( for the multicast traffic ) through the sgi and gc2 interfaces , respectively . in the _
static embms activation _
solution , we assume that the gcs are always managed by resorting to multicast transmission except when it is not available .
this approach exploits the basic structure of the embms , for which it is needed to verify the satisfaction of the pss requirements . from fig .
[ fig.rate_gc ] it is possible to see that this is true for the expected requirements in terms of number of gcs supported for each type of service @xcite .
however , the most critical requirement is represented by the gc set - up time that must be lower than 300 @xmath0 ( sec .
[ req ] ) .
this depends on the time requested to establish the embms bearer and to start - up the call .
two options have been analysed : * _ pre - established embms bearer_. the network establishes the embms bearer over preconfigured mbms areas before the gc starts .
this implies that the bm - sc pre - establishes in advance all the information related to the gc , such as the _ temporary multicast group identifier _ ( tmgi ) , qos class and the enbs belonging to the mbms area .
in particular , the gcse as requests the creation of an embms bearer to the bm - sc by means of the pcrf interface , which is in charge of the exchange of the information related to the embms session .
as soon as a ue requests a gc , the downlink traffic is transmitted using one of the pre - established embms bearer .
this solution provides a fast total set - up time for the gc service that depends only on the time to start - up the call .
3gpp estimates that it is nearly 220 - 250 @xmath0 @xcite , which is consistent with the requirement for the pmr voice communication ( see sec .
[ req ] ) . *
_ dynamic bearer setup at group call start - up_. in this case the embms bearer is established only when needed .
it means that in addition to the 220 - 250 @xmath0 for the call set - up , also the delay for the downlink bearer establishment shall be considered for evaluating the user experience .
the additional latency is assessed on the order of 115 @xmath0 @xcite , taking into account 10 @xmath0 for radio interface delay , 5 @xmath0 for network interface delay ad 5 @xmath0 of request processing delay .
this additional latency is not negligible and the total delay experienced by users , for call set - up and downlink bearer set - up , also exceeds 300 @xmath0 .
[ bearer_gr ] shows the delay contributions composing the overall end - to - end latency at call start - up , for both options . even if dynamic bearer set - up is more flexible and permits an optimization of the user resources , pre - established bearer is the option selected in our solution , because it permits to satisfy the gc set - up time requirement .
however , it could be not sufficient because when the mbms service is mapped on the embms bearer and on the logical mtch channel , it has to be multiplexed with the control information ( mcch ) and then sent on the physical mch to be transmitted . in the current lte standard
, the mcch can be updated with a minimum period of 5.12 @xmath3 ( called _ mcch modification period _ ) , it means that the gc should wait also this time before to be transmitted exceeding the set - up delay requirement . hence , we propose also a shorter mcch modification period that should be about 50 @xmath0 ( once every 5 lte frames ) . in this case
the use of unicast and multicast is more flexible and also the gc set - up time requirement can be relaxed . indeed in this solution ues
are always served through unicast transmissions at the communication start - up , and the use of pre - established unicast bearer shall be considered in order to accomplish the required end - to - end call set - up time .
on the other end the activation of embms bearer can be dynamic , since users are already receiving the service and will be able to switch on multicast reception as soon as embms has been completely activated , as a consequence the mcch modification period can be neglected and unmodified . the number of users associated to a gc is provided on a per - cell basis by the pgw by means of the _ user location information _ ( uli ) procedure .
consequently , the gcse selects the most efficient transmission scheme .
an alternative to the uli procedure is represented by the ue that sends a message to the gcse whenever moves from one enb to another one .
this solution implies a higher data transmission over the core network , but at the same time allows the ue to switch to idle mode in order to save battery life delivering high burst of data in short period of time . in the proposed
_ dynamic embms activation _
solution the gcse exploits the information regarding the ues associated to a gc service , to detect the most efficient solutions .
in particular , the total amount of resources requested by the ues if they would be served with unicast mode is evaluated and compared with the resources reserved for the multicast mode . as an example fig .
[ tx_eff ] shows the spectral efficiency in terms of throughput of a group member , normalized to the resources allocated to the multimedia gc service when the number of group(s ) members increases .
we can see that the multicast solution is independent on the number of group members , the resource usage depends only on the number of subframes allocated to the multicast service .
conversely , the unicast spectral efficiency decreases with the number of served ues . with
the proposed _ dynamic embms activation _
procedure the system is able to switch from unicast to multicast when the number of ues overcomes a given value represented by the intersection of the curves , hence the spectral efficiency is always the maximum value ( i.e. , the envelope of the unicast and multicast curves ) .
the development of the lte technology offers an excellent opportunity to improve both performance and capabilities of the actual pmr communication systems .
however , towards this goal , substantial research efforts are needed mainly to support the specific critical features of pss systems currently not available in the lte standard . after a critical review of the state - of - the - art concerning group call standardization proposals in 3gpp , the paper outlined a viable architecture solution and validated its efficiency by providing performance evaluations .
k. zhang , x. liang , x. shen , and r. lu , `` exploiting multimedia services in mobile social networks from security and privacy perspectives , '' _ communications magazine , ieee _ , vol .
52 , no . 3 , pp .
5865 , march 2014 .
p. stavroulakis , _ terrestrial trunked radio - tetra : a global security tool _ , ser . signals and communication technology.1em plus 0.5em minus 0.4emspringer , 2007 .
[ online ] .
available : http://books.google.it/books?id=zgb721nzzg4c , `` tr 23.979 3rd generation partnership project ; technical specification group services and system aspects ; 3gpp enablers for open mobile alliance ( oma ) ; push - to - talk over cellular ( poc ) services ; stage 2 ( release 11 ) , '' tech .
rep . , 09 2012 .
lorenzo carl ( s13 ) received a m.sc . in telecommunications engineering from the university of florence in 2013 , and is currently a ph.d .
student the department of information engineering , university of florence .
his research interest are mainly focused on radio resource allocation strategy for broadband wireless networks and digital network coding .
he is an author of technical papers published in international conferences .
romano fantacci ( m84 , sm90 , f05 ) received the phd degree in telecommunications in 1987 from the university of florence where he works as full professor since 1999 .
romano has been involved in several research projects and is author of more than 300 papers .
he has had an active role in ieee for several years .
he was associate editor for several journals and funder area editor for ieee transactions on wireless communications .
francesco gei received the master s degree in telecommunications engineering in 2008 from the university of florence , where he currently collaborate as researcher .
after initial studies on uwb ( ultra wide band ) communication systems ( 2005 - 2008 ) , francesco has been involved in industrial research project on sdr ( software defined radios ) and mesh networks for military and professional purpose ( 2008 - 2013 ) . in last years , he focuses his activity on pmr ( professional mobile radio ) evolution towards 4 g and portability of pmr services over lte ( long term evolution ) mobile networks .
dania marabissi ( m00 , sm13 ) received the phd degree in informatics and telecommunications engineering in 2004 from the university of florence where she works as an assistant professor .
she has been involved in several national and european research projects and is author of technical papers published in international journals and conferences .
she is principal investigator of the firb project `` held '' .
she was associate editor for ieee transaction on vehicular technology .
luigia micciullo received her degree in telecommunication engineering from the university of florence in 2007 . in 2007
she joined the department of electronics and telecommunications of the university of florence as a research assistant .
her research interests include physical layer for wireless communications , ofdm , intra - systems coexistence , interference management , and multi - tiered networks , broadband professional mobile radio evolution . | currently public safety and security communication systems rely on reliable and secure professional mobile radio ( pmr ) networks that are mainly devoted to provide voice services . however , the evolution trend for pmr networks is towards the provision of new value - added multimedia services such as video streaming , in order to improve the situational awareness and enhance the life - saving operations .
the challenge here is to exploit the future commercial broadband networks to deliver voice and multimedia services satisfying the pmr service requirements .
in particular , a viable solution till now seems that of adapting the new long term evolution technology to provide ip - based broadband services with the security and reliability typical of pmr networks . this paper outlines different alternatives to achieve this goal and , in particular , proposes a proper solution for providing multimedia services with pmr standards over commercial lte networks .
professional mobile radio , lte , group call , critical communications |
starless cores that are on the verge of or undergoing gravitational collapse provide important information on the initial stage of star formation . in order to understand the dynamics of such starless cores ,
it is necessary to obtain information on at least density , temperature , and velocity distributions .
the density and temperature have been determined by submillimeter / nir and nh@xmath7 observation with a certain degree of accuracy @xcite .
however , the velocity field has not yet been elucidated .
the observations of starless cores in the molecular lines of cs , @xmath8 , and @xmath9 @xcite show that the inward motion is extended to @xmath10 pc with a speed of @xmath11 km s@xmath0 , which is derived using a two - layer model @xcite .
the analysis of the cs j = 3 - 2 and j = 2 - 1 lines with the same model for several cores such as l183 , l1521f , l1689b , l1544 , and l694 - 2 , suggests that the infall speed appears to increase toward the center @xcite .
since the two layer model is rather simple , the derived velocity may be an average value along the line of sight and its variation with depth should be considered with caution .
moreover , it should be noted that in general , cs is significantly depleted in the central regions of starless cores @xcite . on the other hands , by applying an improved version of the two layer model @xcite to the interferometric observations in @xmath9 , @xcite found that the infall velocity increases toward the emission peak for l694 - 2 and l1544 .
this transition might trace the deep central region of the core because the emission of @xmath9 primarily comes from the central part due to its low optical depth and centrally condensed matter distribution @xcite .
therefore , it is highly likely that these independent observations do not trace the entire range of the velocity fields in the collapsing cores .
similar shortcomings are evident in @xcite , @xcite , and @xcite , although they have assumed that the infall speed varies monotonically with the radial distance and have performed more sophisticated radiative transfer calculations . in order to derive a reliable velocity field , by overcoming these problems , we selected hcn j = 1 - 0 hyperfine lines as a probe and observed several candidates of collapsing cores .
then , we performed radiative transfer calculations by considering different types of velocity fields .
an advantage of using the hcn j = 1 - 0 lines is that three hyperfine lines are well separated and they cover a wide range ( a factor of 5 ) of the optical depth .
in addition , a direct comparison between the integrated intensity map of hcn and that of the @xmath9 reveals that hcn is not depleted significantly @xcite . for radiative transfer analyses , we used the one - dimensional monte carlo code by considering the line overlap effect due to the hyperfine splitting of hcn energy levels @xcite .
therefore , the capability of this code is almost the same as that of code used by @xcite , @xcite , and @xcite .
the only difference is that we explored more complicated velocity fields , by relaxing the condition of the monotonic increase or decrease of velocity field , which is an implicit assumption of the previous studies .
we analyzed two round - shaped collapsing starless cores , l694 - 2 and l1197 , located at @xmath12 , @xmath13 and @xmath14 , @xmath15 , respectively .
the observations were performed in september 2003 by using the iram 30 m at a spacing of 11@xmath16 @xcite .
the beam size of the telescope was @xmath17 for the hcn j = 1 - 0 and the velocity resolution of the backend was 0.033 km s@xmath0 .
data were converted to the t@xmath18 scale using a main beam efficiency of 0.82 .
we set the frequency of the three hcn j = 1 - 0 hyperfine lines according to the cologne database for molecular spectroscopy @xcite as follows : 88.6304157 ghz for f = 0 - 1 , 88.6318473 ghz for f = 2 - 1 , and 88.6339360 ghz for f = 1 - 1 . in order to perform a comparison with the one - dimensional radiative transfer calculations
, we assumed that two cores are spherically symmetric .
it is justified by the spherically symmetric density distribution from the observation of dust extinction and emission for l694 - 2 @xcite .
the hcn j = 1 - 0 line profiles are distributed concentrically and they all exhibit blue asymmetry for l694 - 2 .
this is the case for l1197 except for red asymmetry in south - west region occupying only @xmath19 10% of the observed area , which has a negligible contribution to the analysis .
the observations will be described in detail in a separate paper ( sohn , in preparation ) .
the observed spectra were azimuthally averaged at concentric radial annuli with a step of 10@xmath16 .
we set the centers of the cores to the peak intensity positions of @xmath9 , which are ( 0@xmath16 , 0@xmath16 ) for l694 - 2 and ( 0@xmath16 , 11@xmath16 ) for l1197 relative to the map centers mentioned above .
since the spectra were at rectangular grid points , they were weighted by a factor of @xmath20 , where @xmath21 is the distance in arcseconds between the radial distance of the grid point and the specified radius and 14@xmath16 is one half width of telescope beam .
the resulting spectra are shown in fig .
[ fig1 ] and will be used as templates for the comparison with synthesized line profiles .
we synthesized line profiles by using the monte carlo radiative transfer model and compared them with the observed ones in order to derive the model parameters that best explain the observations .
we adopted distances of 250pc and 400 pc to l694 - 2 and l1197 , respectively @xcite .
velocities relative to the local standard of rest were selected as 9.61 km s@xmath0 and @xmath22 km s@xmath0 for l694 - 2 and l1197 , respectively .
the model cores with a radius of 0.15 pc for both l694 - 2 and l1197 were composed of 30 concentric shells at regular intervals . additionally , these cores were surrounded by an envelope with an adjustable radius ( see below ) .
we used @xmath23 model photons for the monte carlo code .
the kinetic temperature was assumed to be spatially constant at 10k , on the basis of the values derived for l1498 and l1517b @xcite .
for the density distribution , we used the following form adopted by @xcite , @xmath24 where @xmath25 is the central density , @xmath26 is the radius of the inner flat region , and @xmath27 is the asymptotic power index . in dynamic models , @xmath26 decreases with an increase in the central density @xcite . to consider this effect , we defined @xmath28 such that it depended on @xmath29 , @xmath30 where @xmath31 is the scale radius of the bonnor - ebert sphere , @xmath32 is the sound speed , @xmath33 (= 2.33 ) is the mean molecular weight , and @xmath34 is the hydrogen mass . if @xmath27 is 2.5 in equation ( [ model ] ) , it represents approximately the density distribution of the bonnor - ebert sphere @xcite . in our model , we tested two cases of sound speed , @xmath32 = 0.2 and 0.3 km s@xmath0 , and finally adopted the latter .
since the integrated intensity map of the thinnest hcn j = 1 - 0 , f = 0 - 1 line is similar to that of the n@xmath3h@xmath35 j = 1 - 0 line , it appears that hcn is not significantly depleted for these two cores at least @xcite . in order to take into account any minor depletion or enhancement of hcn
, we assumed the hcn abundance variation of @xmath36 @xmath37 @xmath38^\beta$ ] , where @xmath39 is the power index and @xmath40 and @xmath41 are the density and hcn abundance relative to h@xmath3 near the boundary , respectively .
@xmath40 were fixed lastly as @xmath42 @xmath43 for l694 - 2 and @xmath44 @xmath43 for l1197 , respectively .
the line profiles are most sensitive to the infall velocity field .
we characterized the velocity field as a @xmath45-shaped field .
( the reason why we adopted this shape will be discussed in next section . )
the innermost region within @xmath46 is static , and the infall velocity increases with a linear function of the radial distance toward the peak velocity position ( @xmath47 ) with a maximum magnitude of @xmath48 , and then linearly decreases to the outermost layer .
we fixed the size of the outermost layer and its velocity to @xmath49 and a constant , respectively .
in fact , the size of the outermost layer is determined on the basis of the first several trials .
the velocity of the outermost layer can be directly derived from the line profile .
the fitting procedure will be described in detail in next section . in summary , we have seven free parameters in total : @xmath29 and @xmath27 for density , @xmath41 and @xmath39 for abundance , and @xmath46 , @xmath50 , and @xmath51 for velocity .
we made reference to the values of previous studies for density and abundance @xcite . besides these , a few auxiliary parameters were necessary to fit the line profiles .
the e - folding widths of the absorption coefficient profile in the inner static region were specified as 0.15 km s@xmath0 and 0.13 km s@xmath0 for l694 - 2 and l1197 , respectively ; they are larger than the pure thermal width by a factor of @xmath52 . in the other layers ,
these widths were automatically adjusted such that the line widths of the model spectra closely resembled the observed ones .
to explain the hyperfine anomalies of hcn , an envelope with larger microturbulence width and lower density must be introduced @xcite .
the diffuse and turbulent molecular clouds surrounding the cores traced by co may be considered as the envelope @xcite .
we selected the microturbulence width of 0.5 km s@xmath0 and the abundance of @xmath53 , which were the same for both cores .
the density was set to @xmath54 @xmath43 for l694 - 2 and @xmath55 @xmath43 for l1197 .
then , we varied the size of the envelope to adjust the relative intensity ratio among the hyperfine lines .
it is not necessary for the envelope parameters to be unique since various combinations of the density , abundance , and size are acceptable if the optical depth is almost the same and the excitation temperature is sufficiently low and close to 2.7 k. it should be noted that the envelope is introduced to account for only the intensity ratio among the hyperfine lines .
the line profiles produced by this method were convolved with the telescope beam for a direct comparison with the observations .
the difference between the model line profile and the observed one was evaluated by @xmath56 @xcite .
we considered all channels of three hyperfine transitions on 5 positions , in the velocity range from @xmath57 km s@xmath0 to @xmath58 km s@xmath0 .
in the first phase , we adjusted the density distribution and constant hcn abundance without infall motion so that the intensities of f = 0 - 1 at all locations are similar to those observed . further , we attempted to fit the line profiles with the infall velocity fields of the constant or monotonically increasing / decreasing function of the radius and found that they can not reproduce the observed spectra .
the infall velocities at every grid points were then varied by trial and error .
it was immediately evident that the infall motion is dominant in the middle and outer layers and other parts of the cores are rather static or exhibit a small amount of motion .
therefore , we approximated the velocity field with the @xmath45-shaped one .
the @xmath45-shaped velocity field could be qualitatively understood by inspecting the characteristic features of the observed line profiles shown in fig .
[ fig1 ] as follows . on the basis of the fact that the position of the self - absorption dip in the thickest f = 2 - 1 line indicates the velocity at the outermost part
, we can estimate the infall velocity in that region .
these infall velocities for l694 - 2 and l1197 are @xmath59 and @xmath60 km s@xmath0 , respectively . on the other hand ,
the velocity in a region close to the central part can be derived from the line width of an optically thin n@xmath3h@xmath35 line , since it is not depleted significantly and , therefore , its emission mainly originates from the central part in which the density is the highest . from @xcite , the fwhm of l694 - 2 and l1197
are found to be 0.27 km s@xmath0 and 0.28 km s@xmath0 , respectively .
these values can be used to derive the contribution of non - thermal motion after removing the thermal width .
if this motion is attributed to the systematic inward motion , the maximum infall velocity will be 0.12 km s@xmath0 for both l694 - 2 and l1197 .
since cores with a pure thermal line width are extremely rare , the systematic motion will be even smaller .
the line shape also provides information on the velocity field in the middle layer .
the increasing rate of the brightness from the absorption dip to red part is considerably slower than to blue part , particularly for l694 - 2 , thereby resulting in a weaker red peak .
this suggests that the region of @xmath61 remains close to the surface layer in the red part , implying that the infall velocity increases steeply toward the deeper layer .
then , it must decrease again toward the center to satisfy the constraints of the n@xmath3h@xmath35 lines . in the second phase , in order to derive the velocity field in a more systematic manner , we reduced the number of parameters describing the velocity field to three , @xmath46 , @xmath50 , and @xmath48 as mentioned in the previous section .
first , we adjusted the parameters of the velocity field , while fixing the other density and abundance parameters .
thereafter , the density parameters were modified followed by the abundance parameters , while the other parameters were unchanged .
we repeated this process a few times .
the spectra of the best fit models are displayed in fig .
[ fig1 ] , where one can note that almost all the details are reproduced , particularly for l694 - 2 . the density and velocity distribution of the models
are shown in fig .
the minimum @xmath62 s are estimated to be 2.7 and 9.2 for l694 - 2 and l1197 , respectively .
the distributions of the @xmath62 around the best fit model are displayed in fig .
[ fig3 ] and fig .
as expected , the abundance parameters are not completely independent , since both are related to the column density of hcn .
there seems to be ( anti-)correlations among the velocity parameters , too .
to summarize , the infall velocity begins to increase from 0 near the @xmath63 pc for both the cores , and reaches the maximum infall velocity of @xmath64 km s@xmath0 at @xmath65 pc for l694 - 2 and @xmath66 km s@xmath0 at @xmath67 pc for l1197 as shown in fig .
the central density and power index of l694 - 2 are around @xmath68 @xmath43 and @xmath69 , while those of l1197 are @xmath70 @xmath43 and @xmath71 , respectively .
our result of l694 - 2 is similar to that derived from the near - infrared extinction and dust emission maps @xcite .
the abundance is uniform with @xmath72 for l694 - 2 ( @xmath73 ) , while it increases slightly toward the center for l1197 ( @xmath74 ) from @xmath75 to @xmath6 .
since we optimized several parameters cyclically , it is uncertain whether they are a set of best fit parameters .
this may be true because we did not explore sufficient parameter spaces .
however , since our optimization is based on the qualitative understanding of line profile formation , it is unlikely that a quite different solution will exists .
the degree of coincidence of the synthesized and the observed line profiles also supports this argument , as shown in fig .
[ fig1 ] . although a new solution may be derived , but characteristic features such as the @xmath45-shaped velocity field will remain unchanged . the fact that the velocity distribution is concentrated in the middle and outer layers is attributed to the selection of hcn hyperfine lines that can probe a wide range of core interiors and to the relaxation of the condition of a constant or monotonic decrease / increase in the velocity field .
our model naturally explains the `` extended inward motion '' observed for several starless cores since the infall motion is dominant in the middle and outer layers .
the size of the infalling layer is @xmath76 pc , which is consistent with the observations @xcite .
the extended inward motion probed by cs toward l1544 and l1551 appears to be exceptional in that it extends over very wide area .
this may be due to other types of motion @xcite . the velocity field that we found does not contradict the observation of @xcite , in which increasing inward motion toward the center in the central 30@xmath16 region of l694 - 2 is derived .
for this , they used the optically thin n@xmath3h@xmath35 line to probe the infall velocity of @xmath77 km s@xmath0 and therefore the more dominant motion in the middle and outer layers has been missed . in our case , we assumed that the velocity in the central region is zero in order to reduce the number of parameters .
in fact , we have tested several sets of models whose velocity fields have non - zero values in the deeper part , and have found that the velocity of @xmath78 km s@xmath0 is acceptable . the distribution of the infall motion of our model is similar to that of the ambipolar diffusion model and isothermal core collapse model @xcite in the sense that the models also have the @xmath45-shaped velocity field and their infall velocity peaks lie just outside the central density - flat region . however , the velocity gradient of our model inside the velocity peak position is steeper than that predicted by the dynamic models .
we can observe such a sudden decrease in the infall motion in the deeper layer in the so - called first collapse phase of the model used in @xcite .
however , this occurs when the central part is very dense and opaque at an evolutionary stage later than that of our sample . some mechanisms resisting against gravity such as rotation may work ; however , they are easily excluded since rotation is very slow @xcite .
an oscillation of starless cores around equilibrium may be able to explain this , as shown in @xcite , where pressure wave reflected from center combined with inward motion due to perturbation from outside results in the steep velocity gradient in the middle .
however the magnitude of the motion is much less than sound speed . in the case of l694 - 2 and l1197 ,
the inward motion is similar to or greater than the sound speed , suggesting that they are undergoing dynamical collapse .
the mechanism of a steeper pressure gradient should be sought , but it is inappropriate to describe it in greater detail at this stage .
the motion of the ambipolar diffusion model @xcite is very slow and therefore it can not be used to explain the velocity field derived in this study .
this is because the model of @xcite is developed for l1544 when the infall velocity of this core was known to be less than 0.1 km s@xmath0 . if the magnetic field is reduced to explain the infall motion of the cores with ambipolar diffusion , the model will be supercritical and reduced to the usual isothermal collapse model .
by combining high spatial resolution hcn observation toward l694 - 2 and l1197 with the monte carlo radiative transfer calculation , we have found that the infall velocity is most dominant in the middle and outer layers of the cores and the infall velocity can be as large as 0.3 km s@xmath0 .
the most important feature of our study is that the velocity field is determined with such accuracy and fidelity that it is possible to compare dynamics and observations for the first time .
an extensive investigation of the velocity distribution in collapsing starless cores is in progress ( sohn et al .
, in preparation ) . the result will clarify the dynamic status of starless cores in greater detail .
this study was supported by the ministry of science and technology , korea , under the grant r14 - 2005 - 058 - 01002 - 0 .
s.h.l . used the pc cluster facility of korea astronomy and space science institute .
c.w.l acknowledges support by kosef r01 - 2003 - 000 - 10513 - 0 program . | in an attempt to understand the dynamics of collapsing starless cores , we have conducted a detailed investigation of the velocity fields of two collapsing cores , l694 - 2 and l1197 , with high spatial resolution hcn j = 1 - 0 maps and monte carlo radiative transfer calculation .
it is found that infall motion is most active in the middle and outer layers outside the central density - flat region , while both the central and outermost parts of the cores are static or exhibit slower motion .
their peak velocities are 0.28 km s@xmath0 for l694 - 2 and 0.20 km s@xmath0 for l1197 , which could not be found in simple models .
these velocity fields are roughly consistent with the gravitational collapse models of the isothermal core ; however , the velocity gradients inside the peak velocity position are steeper than those of the models .
our results also show that the density distributions are @xmath1 and @xmath2 in the outer part for l694 - 2 and l1197 , respectively .
hcn abundance relative to h@xmath3 is spatially almost constant in l694 - 2 with a value of @xmath4 , while for l1197 , it shows a slight inward increase from @xmath5 to @xmath6 . |
the field studying atomic gases at extremely low temperatures has seen riveting developments over the past decade @xcite .
this has pushed the field into a rather unique position where cold atoms in various geometries can be used to do quantum simulation of widely used models from other fields including condensed - matter physics @xcite , particle and high - energy physics @xcite , nuclear physics @xcite , and even chemistry @xcite . a remarkably useful feature of cold atomic gas experiments is the control over external trapping parameters that makes it possible to explore physics in different dimensionalities @xcite . in particular ,
a number of impressive experiments have studied various aspects of interacting bosons in one dimension @xcite . using the atomic interaction resonances caused by one - dimensional confinement @xcite it has thus become possible to realize many interesting one - dimensional systems including the famous hard - core bosonic tonks - girardeau gas @xcite ( see refs .
most recently , one - dimensional systems of interacting fermions have been also realized in experiments @xcite . in the last few years it has become possible to control the particle number in one - dimensional experiments with great accuracy @xcite , allowing one to build a controllable few - body system of fermions and study fermionization for strong interactions @xcite , pairing @xcite , impurity physics @xcite , a two - site hubbard model @xcite , and heisenberg spin models @xcite .
these developments have sparked a great deal of excitement in the community studying few - body physics and its relation to many - body phenomena . over the last decade a lot of new aspects of these one - dimensional systems
have been covered by different authors .
this includes the physics of small trapped bosonic systems ( single- and two - component ) @xcite , details of few - fermion systems with two values of an internal degree of freedom ( spin ) @xcite and various aspects of the transition from few- to many - body physics @xcite .
furthermore , a number of studies have looked into mixed systems of bosons and fermions and systems with particles of unequal mass @xcite . in the present paper
we are interested in studying one - dimensional fermions or bosons with strong repulsive short - range interactions .
as has been recently discussed , in the strongly interacting limit the system may be described by an effective hamiltonian which has the form of an ( anisotropic ) heisenberg spin model @xcite .
it was predicted theoretically that by tuning the interaction strength from the weak ( repulsive ) to the strongly interacting regime for a two - component fermi system in one dimension one may arrive in the ground state of an antiferromagnetic heisenberg model @xcite and recently this was confirmed in experiments @xcite .
however , the heisenberg model obtained is non - trivial in the sense that the exchange couplings are determined by the local geometry of the trapping potential .
computing the local exchange coupling constants is a formidable numerical task and thus several papers have discussed the possibility to use various approximations to obtain these quantities .
a very neat approximation is the strong - coupling ansatz of levinsen _ et al . _
@xcite which allows one to get an extremely accurate set of exchange coefficients for arbitrary system sizes in the case where the external confinement is given by a harmonic oscillator .
it has also been conjectured that for smooth potentials , the well - known local density approximation ( lda ) should give a very accurate value for the local exchange coefficients @xcite and numerical results for up to six particles show that the deviations are at most a few percent @xcite .
this is a very reasonable expectation and many studies in the past have used the lda when studying the properties of many - particle systems in one dimension @xcite .
however , this does not necessarily imply that the lda works equally well for smaller systems as large system sizes can sometimes average out some of the finer details that the lda may not capture .
in the present paper we test the performance of the lda for strongly interacting one - dimensional systems with up to six particles by comparing it to exact calculations .
we do this for several different potential forms , including some experimentally relevant double - well geometries . to the best of our knowledge
, no previous study has presented results for the effective spin models with up to six particles in non - harmonic confinement .
we gauge the performance of the lda against exact results not only for statics ( producing the local exchange constants needed for the effective hamiltonian ) but also for dynamics .
for the latter case we study quantum state transfer using the spin models that one obtains with the lda and with exact calculations .
transfer of quantum states in two - component spin systems is a delicate process that depends sensitively on the local exchange couplings and thus provides a difficult challenge for the lda .
the dynamical propagation of information and correlations in one - dimensional setups with cold atoms is a focal point of research at the moment @xcite and it is thus theoretically important to have accurate models for these systems also in the strongly interacting regime where time - dependent exchange and non - equilibrium quantum magnetism can be studied @xcite . the paper is organized as follows . in sec .
[ formal ] we outline the model and its assumptions , and then discuss how to compute exchange coefficients exactly and within the lda .
section [ results ] presents a comparison of exact and lda results for double - well potentials . in sec .
[ qst ] we introduce the spin model picture of strongly interacting 1d systems and we apply the local exchange coefficients in the context of quantum state transfer to investigate some consequences of the differences between different computational schemes .
section [ summary ] contains a summary , discussions , and outlook .
we consider particles that are confined to move in one dimension ( 1d ) and assume that along the direction of motion there is a trapping potential , @xmath0 , which is the same for all the particles .
the particles have short - range interactions that we model by a dirac delta - function and in turn the hamiltonian may be written in the following way @xmath1 + g_{1d } \sum_{i > j } \delta(x_i - x_j),\ ] ] where @xmath2 and @xmath3 are the momentum and coordinate operators of particle @xmath4 , @xmath5 is the mass of a particle and @xmath6 is the trapping potential for the @xmath4th particle .
the interaction strength is parametrized by @xmath7 . in experimental setups
the 1d confinement is achieved by applying a very tight transversal trap .
it may then be shown that @xmath7 can be directly related to the scattering properties of the non - trapped atoms and is a function of the low - energy three - dimensional scattering length @xmath8 and the transverse trapping length @xmath9 @xcite .
what is very interesting is that one finds resonances where @xmath7 diverges due to the presence of the transverse confinement .
this has been clearly demonstrated in recent experiments @xcite . in particular , the regime where @xmath10 is very large
is accessible experimentally @xcite . for simplicity we will use the notation @xmath11 from now on since this can give rise to no ambiguities in the present context .
in the present paper we will consider two - component fermi systems with @xmath12 atoms , where @xmath13 and @xmath14 are number of atoms of components with spin projection up and down , respectively .
the zero - range two - body interaction in the hamiltonian will act only between pairs with opposite spin projection . for pairs with the same spin projection the pauli principle requires antisymmetry upon exchange of the two particles . as the wave function must also be continuous
, we may infer that when two particles with identical spin projections are close to each other the wave function has a continuous first derivative . in turn , the dirac delta - function two - body interaction has no effect . in experiment with atoms
there are interactions between pairs of atoms with identical spin projection , but they are highly suppressed due to their short - range nature and thus can be safely neglected for our purposes .
the general hamiltonian above takes this implicitly into account as we ensure antisymmetry among like components in our @xmath15-body wave functions .
we note , however , that many of our results may be easily transferred to two - component bosons with uniform interactions , i.e. a single @xmath16 controlling the interactions between both identical and different components @xcite .
we now consider the strongly interacting regime where @xmath17 @xcite .
the most general eigenstate of the hamiltonian in the limit @xmath17 is @xcite @xmath18 where the sum is over the @xmath19 permutations , @xmath20 , of the coordinates , @xmath21 are the coefficients that depend on the ordering of the particles , @xmath22 , when @xmath23 and zero otherwise .
@xmath24 is a fully antisymmetric @xmath15-particle slater determinant wave function constructed from @xmath15 single - particle wave functions that are obtained by solving the corresponding single - particle schrdinger equation with the potential @xmath0 . in this paper we are interested in the lowest energy manifold of @xmath15-body states which are obtained by taking the single - particle states with the @xmath15 lowest energies .
the wave function @xmath24 describes the non - interacting @xmath15-fermion system with the energy @xmath25 .
it is important to note that this @xmath15-body energy is @xmath26 times degenerate , and thus there is a manifold of @xmath27 degenerate @xmath15-body states in the limit @xmath17 which is a quasi - degenerate manifold at large but finite @xmath16 ( which is where we will be working below as we consider quantum state transfer ) .
this degeneracy arises from the fact that in the @xmath17 limit the particles become essentially impenetrable . yet , there are still @xmath27 distinguishable ways that the particles may be ordered on a line which all have the same energy . in practice
one may think of these various orderings of the spins along a line as a set of basis states @xcite .
in the limit @xmath28 we can write the @xmath15-particle energy , @xmath29 , to the linear order in @xmath30 as @xcite @xmath31 where @xmath32 .
the important observation is that there is an @xmath21-independent coefficient , @xmath33 , in this expression .
it is a geometric factor that depends solely on the total number of particles @xmath15 and on the potential @xmath0 and its single - particle eigenstates .
remarkably , it does _ not _ depend on what system one is considering as long as the interactions are strong , i.e. as long as we consider the limit @xmath17 .
if one is able to compute @xmath33 for a given @xmath0 , then this may be used to study multi - component fermi or bose systems with all possible combinations of internal components among the particles .
the geometric factor @xmath33 can in a certain sense be thought of as the local exchange coupling in the system .
this interpretation is very useful when mapping the system onto a spin model @xcite .
if one thinks of the @xmath15 particles sitting on a line , then the index @xmath34 on @xmath33 corresponds to a pair of particles and the exchange coupling for that pair is proportional to @xmath33 . since strongly interacting 1d systems
are governed by exchange processes , the statics and dynamics of such systems is essentially dictated by the @xmath33 coefficients and we would thus like to compute these in as general circumstances as possible .
therefore , the coefficients @xmath33 will therefore be the main focus of our discussion .
an exact expression for @xmath33 was first derived in @xcite @xmath35 where the derivative in the integral denotes @xmath36_{x_1 = x_n},\ ] ] where one first takes the derivative of the @xmath15-body antisymmetric function @xmath24 with respect to @xmath37 and then subsequently sets @xmath38 .
the exact expression is an @xmath39-dimensional integral and the numerical calculation of @xmath33 is by no means an easy task .
it is therefore desirable to consider whether appropriate approximations can be made to access these quantities also for larger values of @xmath15 .
an important observation was made in ref .
@xcite where it was noticed that for @xmath40 , a local density approximation can be used in computing @xmath33 and this yields results that are off by only a few percent for the benchmark case where @xmath0 is a simple harmonic oscillator potential . using a highly accurate ansatz wave function for the @xmath15-body problem , it later became possible to get a highly accurate approximation to @xmath33 for the harmonic trap for any @xmath15 given as a ratio of quadratic polynomials in @xmath15 @xcite .
here we are concerned with the question of how well the local density approximation ( lda ) does for different potentials .
we must therefore define and discuss how the lda may be applied in the context of strongly interacting particles in 1d .
this discussion closely parallels that of ref .
the main inspiration for the lda method in the present context comes from earlier work on the hubbard model using the bethe ansatz where ogata and shiba have shown that in the strongly interacting limit the spin and charge dynamics decouple @xcite .
the spin degrees of freedom may correspondingly be described by a spin model of the heisenberg type ( we return to this later on ) with an exchange coupling that is proportional to the third power of the density @xcite .
this result is derived for a homogeneous system with periodic boundary conditions which are the typical basic conditions needed to solve the bethe ansatz equations @xcite . in order to transfer these results into the present context with non - homogeneous confinement of the 1d system , ref .
@xcite suggested to use the density from the lda in the expression for the exchange coupling in the spin model .
this yields the expression @xmath41 here @xmath42 is the 1d thomas - fermi density @xcite which is given by @xmath43 where @xmath44 is the chemical potential of the system .
@xcite proposed to calculate the thomas - fermi density in the center - of - mass positions @xmath45 of the @xmath4th and @xmath46th particles .
the position @xmath45 can be written as @xcite @xmath47 , where @xmath48 taking this position makes sense in light of the interpretation of @xmath49 as a local exchange coupling of a pair of particles and thus by symmetry one should take the density at their common center - of - mass which may be determined within the @xmath15-body system .
one may argue that any discrepancy between the exact value of @xmath49 and the lda approximation can be alleviated by simply picking the right values of @xmath45 .
however , finding the right @xmath45 appears as difficult as calculating the exact @xmath49 directly .
note also that for large @xmath15 finding @xmath50 is an equivalently difficult task and thus one needs some method of determining densities accurately .
here we will ignore these problems and be concerned with how well the lda performs in cases where we can compute both the exact and the lda expression for @xmath49 easily .
we therefore work with @xmath40 here . for a commonly used harmonic oscillator trapping potential @xmath51 the thomas - fermi density
can be written down in a simple form @xmath52 in this case the approximation is rather accurate compared to the exact calculation and the relative error is not larger than @xmath53 percent and decreases rapidly with increasing particle number @xmath15 @xcite .
however , for potentials of more complicated shape the calculation of the center - of - mass positions becomes much more difficult . in the following
we discuss the coefficients @xmath54 and @xmath55 , with @xmath56 for different potential profiles .
the numerical calculations of eqs . and consist of two main parts .
first , we construct the slater determinant @xmath24 with the eigenstates , @xmath57 , of a particle trapped in the potential @xmath6 .
we assume that our system is confined in a box potential with length @xmath58 . to handle this numerically
, we use the normalized eigenstates @xmath59 where @xmath60 is an integer , as an expansion basis for the functions @xmath61 to evaluate integrals in eqs .
and we utilize a recursive many - dimensional integration procedure with the one - dimensional integrals being calculated using a standard trapezoidal integration routine .
even though more advanced integration methods of course are available they are not readily available for the case of nested integrals ( integration limits that depend on subsequent integrations ) .
we find that the trapezoidal integration converges rapidly with the relative error for the values of the coefficients not exceeding @xmath62 .
we find that @xmath63 basis vectors @xmath64 and @xmath63 integration points are sufficient to achieve this accuracy in a reasonable time . in the following sections we dismiss the superscript @xmath65 for simplicity and make sure that it is clear from the context of the discussion and in the figures whether we are discussing the exact expressions for @xmath49 or the lda approximations . from eq .
( top panel ) and @xmath66 from eq .
( bottom panel ) for different values of parameters.,title="fig : " ] from eq .
( top panel ) and @xmath66 from eq .
( bottom panel ) for different values of parameters.,title="fig : " ]
to perform a comparison between the lda results and the exact solutions for the exchange coupling constants @xmath54 we use two specific forms of a double well potential which have flexibility to explore the similarities and differences between the lda and the exact calculation .
the first potential has the form @xmath67 where @xmath68 and @xmath69 define the energy scale , and the height of the central barrier .
notice that this potential has a cusp ( discontinuity of the first derivative ) at @xmath70 .
this is perfectly well allowed in the formalism as this will not cause any troubling discontinuities in the single - particle wave functions or their first derivatives .
the second potential is a symmetric trap of the form @xcite @xmath71 - u\sin^2 \left [ \pi \left ( \frac{2x}{l } + 1\right)\right ] , \ ] ] where the values of @xmath72 and @xmath73 control the shape of the potential .
figure [ fig : potentials ] shows the shapes of the potentials we use in the article for different values of the control parameters . for the double well potential in eq .
with @xmath74 as function of the central barrier height @xmath69 .
the exact values based on the expression in eq . are shown as solid lines , while the corresponding approximate values based on the lda formula in eq . are the dashed lines .
the top panels shows @xmath75 and 4 , the middle panel has @xmath76 , and the bottom panel has @xmath77 .
note that the parity symmetry of eq .
implies that @xmath78 .
the dashed horizontal lines in the @xmath79 and @xmath77 cases correspond to the value @xmath80 for @xmath81 and @xmath82 , respectively .
this is the excepted asymptotic values for large @xmath69 as discussed in the text.,title="fig : " ] for the double well potential in eq .
with @xmath74 as function of the central barrier height @xmath69 .
the exact values based on the expression in eq . are shown as solid lines , while the corresponding approximate values based on the lda formula in eq . are the dashed lines .
the top panels shows @xmath75 and 4 , the middle panel has @xmath76 , and the bottom panel has @xmath77 .
note that the parity symmetry of eq .
implies that @xmath78 .
the dashed horizontal lines in the @xmath79 and @xmath77 cases correspond to the value @xmath80 for @xmath81 and @xmath82 , respectively .
this is the excepted asymptotic values for large @xmath69 as discussed in the text.,title="fig : " ] for the double well potential in eq .
with @xmath74 as function of the central barrier height @xmath69 .
the exact values based on the expression in eq . are shown as solid lines , while the corresponding approximate values based on the lda formula in eq . are the dashed lines .
the top panels shows @xmath75 and 4 , the middle panel has @xmath76 , and the bottom panel has @xmath77 .
note that the parity symmetry of eq .
implies that @xmath78 .
the dashed horizontal lines in the @xmath79 and @xmath77 cases correspond to the value @xmath80 for @xmath81 and @xmath82 , respectively .
this is the excepted asymptotic values for large @xmath69 as discussed in the text.,title="fig : " ] we note that we consider values in the interval @xmath83 $ ] where @xmath58 is the length of a box which confines the whole system .
this is required for our numerical procedure where we use @xmath84 in our calculations .
energies are measured in units of @xmath85 , with a factor of 4 coming from the box extending to @xmath86 for the potential in eq .
we set @xmath74 ( in units of @xmath58 and @xmath87 ) while for eq .
we have used @xmath88 in units of @xmath87 . in deep enough potentials the single - particle states
do not feel the effect of this outside boundary . however , for shallow potentials some of the states might reside in the whole box and thus be influenced by the outer wall .
we will see how it affects the exchange constants in the subsections below .
we now consider the coefficients @xmath54 using eqs . and for different values of the control parameters ;
the height of the central barrier @xmath69 for the potential in eq . and the parameter @xmath73 for the potential in eq . .
[ fig : dwcomp ] and [ fig : apcomp ] show both the lda and exact values of the coefficients @xmath54 for different numbers of particles @xmath89 . in this article
we consider only spatially symmetric 1d potentials and so @xmath78 . in fig .
[ fig : dwcomp ] we show the results for the potential in eq . comparing the exact results obtained from eq . to the lda results obtained with the formula in eq . for particle numbers
note that we do not need to plot all the @xmath54 coefficients as the parity invariance of the double well in eq .
gives us the convenient symmetry relation @xmath91 .
the results demonstrate that the lda does very well for small values of the central barrier height @xmath69 which should not be too surprising as the potential for small @xmath69 resembles very much that of a harmonic oscillator which was previously shown to be quite accurately described within lda @xcite .
one notices that for the even particle numbers the middle coefficient , i.e. @xmath80 , @xmath92 , and @xmath93 for @xmath81 , @xmath79 , and @xmath77 , respectively , has similar behavior , and that this behavior is not captured well by the lda for @xmath94 .
in fact , this middle coefficient , @xmath95 , calculated via eq .
goes to zero much faster than the exact value .
this @xmath95 pertains to the exchange coupling in the middle of the trap , i.e. to the exchange taking place right at the central barrier in the potential .
it is therefore determined by the probability for particle tunneling through this central barrier , and we clearly see the lda fail to capture this effect for larger barriers , i.e. larger values of @xmath69 , where lda underestimates exchange .
this is connected to the fact that lda and in particular the thomas - fermi density in eq .
is only well - defined in between the classical turning points , where @xmath96 .
hence for large barriers where @xmath97 , the thomas - fermi density goes to zero and thus in turn the coefficient @xmath95 goes to zero .
for example , in the case of the potential in eq .
@xmath98 for @xmath99 ( when @xmath74 as we have here ) .
the termination values of @xmath100 , @xmath101 , and @xmath102 can be clearly seen on the horizontal axis for @xmath80 ( @xmath81 ) and @xmath103 ( @xmath79 ) in the top panel of fig .
[ fig : dwcomp ] , and for @xmath93 ( @xmath77 ) in the bottom panel of fig .
[ fig : dwcomp ] . for the smooth double well potential in eq . as function of the parameter
@xmath73 that controls the central barrier . here
we have set @xmath88 .
the exact values based on the expression in eq . are shown as solid lines , while the corresponding approximate values based on the lda formula in eq .
are shown as dashed lines .
the top panel shows @xmath75 and 4 , the middle panel has @xmath76 , and the bottom panel has @xmath77 .
note that the parity symmetry of eq .
implies that @xmath78.,title="fig : " ] for the smooth double well potential in eq . as function of the parameter @xmath73 that controls the central barrier .
here we have set @xmath88 .
the exact values based on the expression in eq . are shown as solid lines , while the corresponding approximate values based on the lda formula in eq .
are shown as dashed lines .
the top panel shows @xmath75 and 4 , the middle panel has @xmath76 , and the bottom panel has @xmath77 .
note that the parity symmetry of eq .
implies that @xmath78.,title="fig : " ] for the smooth double well potential in eq . as function of the parameter @xmath73 that controls the central barrier .
here we have set @xmath88 .
the exact values based on the expression in eq . are shown as solid lines , while the corresponding approximate values based on the lda formula in eq .
are shown as dashed lines .
the top panel shows @xmath75 and 4 , the middle panel has @xmath76 , and the bottom panel has @xmath77 .
note that the parity symmetry of eq .
implies that @xmath78.,title="fig : " ] when the height of the barrier is large ( large @xmath69 ) each of the two wells can be approximated by the harmonic oscillator potential .
we see that in this case the exact values of the coefficients becomes almost constant as function of the barrier height . for even number of particles
we intuitively expect that half of the particles will be located in each well . in the case with @xmath79 this would imply that @xmath80 should go to a constant for large @xmath69 , reaching the value corresponding to the @xmath104 case with @xmath81 ( two particles in a single harmonic well ) .
this is clearly seen in the top panel of fig .
[ fig : dwcomp ] where the horizontal dashed line marks the latter value .
likewise , for @xmath77 , we expect a split into two three - body systems and thus that both @xmath80 and @xmath93 should be approaching the @xmath80 value for @xmath104 and @xmath82 .
again this is seen very nicely in the bottom panel of fig .
[ fig : dwcomp ] the first two coefficients for @xmath77 approach the dashed horizontal line corresponding to @xmath82 and @xmath104 ( notice that there are different vertical scales on the three panels in fig .
[ fig : dwcomp ] ) . for the spatially symmetric potentials considered here with an odd number of particles
, we can not use the same logic of division of the particles into the two wells as the barrier grows large as for even particle numbers .
the overlap across the barrier of the single - particle wave functions will not vanish for odd particle numbers . hence ,
the coefficients @xmath54 which depend solely on these single - particle wave functions will also not vanish .
we clearly see in the top and the middle panel of fig .
[ fig : dwcomp ] that for @xmath105 the lda result does not do a good job in describing the exchange couplings .
again , this is caused by the fast decrease of the thomas - fermi density on which the lda result relies as the barrier increases .
this decrease of density clearly takes place much faster than seen in the exact results .
we thus see that odd - even effects can be considerable in the comparison of the lda and the exact method . in order to further explore the case with odd particle numbers
we have tried to apply a small tilt to the potential in order to explicitly break the parity invariance .
this could for instance be done by an additional term that is linear in @xmath106 in the potential . when this is done for an odd number of particles the values of the exchange couplings approach the values of the couplings of a smaller system .
for instance , for the system consisting of @xmath76 particles , two of the particles will be located in one of the wells , with the interacting coefficient approaching the value of @xmath80 for two particles in a harmonic trap , and the other three will occupy the other well , with @xmath107 approaching those of a three - particle system in a harmonic trap . at the same time we find the intuitive result , namely that the interaction between these two subsystems will approach zero as the barrier increases .
we will not discuss the introduction of slight symmetry - breaking terms any further here .
all in all , for the double well potential the lda approach provides a reliable approximation to the exact values of the coefficients @xmath54 for small central barrier heights ( small values of @xmath69 ) and then becomes poor for larger barrier heights .
an exception is found for even particles numbers where the system splits into two equal size groups that can be described as particles in two separate harmonic wells .
here lda does approach the exact results for the split system asymptotically . to further explore the robustness and/or failures of applying the lda to our problem system , we now change the potential into a different kind of double well shape which has the form given in eq . .
what one should notice here is that the potential in eq .
is in a sense shallow , i.e. it does not increase to infinity around its edges
. therefore the box potential that we have as a hard - wall boundary around the system could be felt by the particles .
this is in fact the case for larger particle numbers as we will now discuss .
.comparison of the exact calculation , @xmath108 ( first column ) , to that of the lda , @xmath109 ( second column ) , for a potential consisting of a flat box with infinite walls .
notice that in this case the exchange coefficients are independent of @xmath4 , i.e. @xmath110 . the percentages in parenthesis in the second column give the deviations of the lda from the exact results .
notice that the lda always underestimates the exchange coupling coefficients .
[ tab : box ] [ cols="<,<,<",options="header " , ] first we consider the cases with @xmath81 and @xmath82 as shown in the top panel of fig .
[ fig : apcomp ] .
here we see that the qualitative behavior of the @xmath54 coefficients is similar to the potential of eq .
( shown in the top panel of fig .
[ fig : dwcomp ] ) with the two - particle case shown a steady decrease with the barrier parameter , @xmath73 , while the three - particle case first decreases and then has a slight increase again at large values of @xmath73 .
this is analogous to the behavior in the top panel of fig .
[ fig : dwcomp ] .
we also notice that the lda does a very good job for the two- and three - particle systems for most values of @xmath73 in the top panel of fig .
[ fig : apcomp ] . for the @xmath82 case this should be compared and contrasted to the case in the top panel of fig .
[ fig : dwcomp ] where the same @xmath82 case for the potential in eq .
demonstrates that the lda deviates significantly from the exact result for larger barrier heights ( large @xmath69 values in that case ) . for more particles ( @xmath111 )
the lda results start to deviate significantly from the exact results as we see in all three panels in fig .
[ fig : apcomp ] .
we see that already for four particles the coefficient @xmath54 differs significantly in the lda and exact calculations .
this difference only increases for larger particle numbers .
the reason for this discrepancy can be traced back to the form of the potential in eq . and its shallow nature as compared to that in eq . .
for the parameters we have chosen , the potential does not become deep enough for five or six single - particle eigenstates to be confined completely inside of the potential profile .
this can be rephrased by stating that not all the five or six lowest single - particle energy eigenvalues are in fact negative .
the most energetic single - particle wave functions that we require to build the totally antisymmetric function , @xmath24 , are thus feeling the outside of the potential .
in particular , they are influenced by the box potential that surrounds our system . while one may alleviate this problem by going to deeper potentials ( increasing @xmath72 in eq . )
we choose @xmath88 in order for us to study quantum state transfer in the last part of our paper and make contact with recent results that utilize the same parameters @xcite . to gain further insights into the influence of the box and the behavior of lda in this respect we may consider just the box potential on its own . in tab .
[ tab : box ] we present the results of applying both the exact formula ( first column ) and the lda version ( second column ) to the box potential .
we clearly see a large discrepancy for the particle numbers we have studied . the deviation of the lda from the exact result is given in percentages in the parenthesis in the second column of tab . [
tab : box ] . by doing a crude fit to the percentages we find that the deviations scale approximately with the particle number as a power law @xmath112 .
we thus find a quite slow convergence and for smaller system sizes the deviations can be significant .
we can see that as the single - particle wave functions are distributed over the whole confining infinite square well potential the local density approach becomes rather inadequate .
the large discrepancies in the values of the lda and the exact results for the @xmath54 coefficients of the shallow potential in eq . for particle numbers @xmath113 can now be better understood .
they are in large part the result of the fact that the outer box boundaries do become important for these particles numbers and that the lda does a poor job of describing particles in a box with a flat bottom .
we do see that the lda approximation gives qualitatively similar results to the exact calculation even for @xmath114 , but quantitatively the lda may differ substantially .
we have checked that as one increases the depth of the potential in eq .
one does indeed see better agreement .
however , the deviations we identified as common for both potential profiles within the lda remain . , for an @xmath79 two - component fermi system with @xmath115 obtained from the potential in eq . for @xmath116 ( first row ) , @xmath117 ( second row ) , @xmath118 ( third row ) , and @xmath119 ( fourth row ) where @xmath73 is in units of @xmath87 .
the other potential parameter is @xmath88 for all panels .
the ( blue ) solid line corresponds to the @xmath115 coefficients obtained from the exact calculation and the ( brown ) dashed line corresponds to the lda approach .
time is measured in units of @xmath120.,title="fig : " ] , for an @xmath79 two - component fermi system with @xmath115 obtained from the potential in eq . for @xmath116 ( first row ) , @xmath117 ( second row ) , @xmath118 ( third row ) , and @xmath119 ( fourth row ) where @xmath73 is in units of @xmath87 .
the other potential parameter is @xmath88 for all panels . the ( blue ) solid line corresponds to the @xmath115 coefficients obtained from the exact calculation and the ( brown ) dashed line corresponds to the lda approach .
time is measured in units of @xmath120.,title="fig : " ] , for an @xmath79 two - component fermi system with @xmath115 obtained from the potential in eq . for @xmath116 ( first row ) , @xmath117 ( second row ) , @xmath118 ( third row ) , and @xmath119 ( fourth row ) where @xmath73 is in units of @xmath87 .
the other potential parameter is @xmath88 for all panels .
the ( blue ) solid line corresponds to the @xmath115 coefficients obtained from the exact calculation and the ( brown ) dashed line corresponds to the lda approach . time is measured in units of @xmath120.,title="fig : " ] , for an @xmath79 two - component fermi system with @xmath115 obtained from the potential in eq . for @xmath116 ( first row ) , @xmath117 ( second row ) , @xmath118 ( third row ) , and @xmath119 ( fourth row ) where @xmath73 is in units of @xmath87 .
the other potential parameter is @xmath88 for all panels . the ( blue ) solid line corresponds to the @xmath115 coefficients obtained from the exact calculation and the ( brown ) dashed line corresponds to the lda approach .
time is measured in units of @xmath120.,title="fig : " ]
in order to understand some potential effects that could be implied by using the lda instead of the exact solutions for the exchange couplings , we now consider a dynamical protocol that has recently been discussed in the context of strongly interacting systems in 1d . as mentioned earlier , one may indeed map these systems into effective spin models @xcite .
this can be done for both fermi systems @xcite and bose systems @xcite . to keep the discussion concise
, we will mainly consider the example of the two - component fermi system here .
in that case , the spin mapping is into the famous heisenberg spin-@xmath121 model whose hamiltonian ( up to a constant energy shift ) is given by @xmath122 where @xmath123 is a spin operator where @xmath124 is the vector of the pauli matrices . here
the nearest - neighbor interaction coefficients are related very simply to the local exchange coefficients and the coupling constant as @xmath125 .
this is another justification for using the term local exchange coupling for the @xmath54 above , i.e. that under the spin mapping they appear as nearest - neighbor couplings in what is equivalent to a spin chain hamiltonian . for bose systems , one may have more involved spin models as the coefficients for @xmath106- , @xmath126- , and @xmath127- direction spin operators are not necessarily the same ( see ref .
@xcite for further details ) .
below , we will make one detour from the uniform heisenberg spin model in eq . in order to consider the so - called xx model which is a special case of the hamiltonian in eq . where all the terms with @xmath127-component operators are eliminated .
but for @xmath76.,title="fig : " ] but for @xmath76.,title="fig : " ] but for @xmath76.,title="fig : " ] but for @xmath76.,title="fig : " ] spin-@xmath121 chains have been proposed as media for quantum state transfer about a decade ago @xcite .
the quantum state transfer in these chains essentially corresponds to flipping a single spin at one end of the chain and then dynamically evolving the state such that one may project onto the state at which the single spin flip has reached the other end for all subsequent times .
the probability of finding the flipped spin at the other end is known as the fidelity of the quantum state transfer and if it reaches unity at some subsequent time we say that the system supports perfect quantum state transfer @xcite .
in the original proposal @xcite a spin chain with constant @xmath128 was considered and it was shown that the fidelity could not reach unity in this case .
later studies demonstrated that if the coefficients are chosen as @xmath129 with @xmath130 some overall constant , then perfect state transfer is possible ( although only within the so - called xx model ) .
one would then have an ideal communications channel for quantum information .
however , it turns out to be exceedingly difficult to produce a system that fulfills these requirements on @xmath115 . in ref .
@xcite volosniev _
proposed strongly interacting 1d atomic systems as a possible realization of perfect state transfer and found that the potential in eq .
can give rise to perfect state transfer in the xx model for @xmath117 .
we will return to this below .
first we need to define the spin state space and the fidelity that we consider . for a spin chain with one impurity ( a single spin that is flipped ) and @xmath131 majorities ( in other words @xmath132 and @xmath133 ) it is natural to use the basis of spin functions @xmath134 . in this basis
we define the fidelity of the quantum state transfer which is a function of time and is given by @xmath135 the fidelity can be straightforwardly interpreted as the hamiltonian acting on the initial state on the right ( with the single flipped spin on the left edge of the systems ) and then projection on the final state on the left ( with the single flipped spin on the right edge ) . in practice ,
one expands the initial state on the set of eigenstates of the hamiltonian , then constructs the state at all later times , and then projects onto the final state which also has some expansion in terms of the eigenstates of the hamiltonian .
notice that the overall time scale of the transfer depends on @xmath54 and on @xmath16 . in the strict limit where @xmath136
, there is no dynamics as the particles are completely impenetrable and all orderings are eigenstates ( all of which are degenerate ) .
however , in the more realistic case where @xmath16 is large but finite , our effective spin models work to linear order in @xmath30 and can thus be used to study dynamics in the strongly interacting regime .
the timescale of transfer will then depend linearly on @xmath16 .
one may think of the state transfer process as a set of subsequent flips of pairs of spins along the chain .
each of these local exchanges depend on @xmath54 , i.e. if @xmath54 is large this happens fast and vice versa .
we thus see that large barriers will suppress the transfer as expected .
however , the process can depend rather delicately on the actual values of the local exchanges as we will now demonstrate .
for instance , in the exact results we may see slow suppression of the local exchange with barrier height as in fig .
[ fig : dwcomp ] in regimes where lda would give exactly zero ( when the chemical potential is below the barrier ) .
this is one source of error in the lda that could carry into a dynamical protocol like state transfer in a severe way .
but for @xmath77.,title="fig : " ] but for @xmath77.,title="fig : " ] but for @xmath77.,title="fig : " ] but for @xmath77.,title="fig : " ] here we plot the fidelity of eq . as a function of time for different potentials of the form given in eq . for the cases with @xmath79 ( fig .
[ fig : apfidelityn=4 ] ) , 5 ( fig .
[ fig : apfidelityn=5 ] ) , and 6 ( fig .
[ fig : apfidelityn=6 ] ) particles , respectively .
all the three cases we present show the characteristic oscillatory behavior of state transfer fidelities as function of time @xcite .
however , one does indeed notice that the oscillations become more prominent and more irregular as @xmath15 increases ( system size or chain length in the spin chain language ) , and also as the control parameter for the central barrier @xmath73 is increased .
perfect state transfer is never achieved for these parameters in a heisenberg model of the type in eq . without an external magnetic field .
this is in accordance with previous results @xcite .
the calculations demonstrate that even when there are only small differences in the exact and lda results , the fidelity can show large variations particularly for longer time intervals .
this is probably due to an accumulation of phase factors in the system over time that tends to drive the exact and lda results apart .
however , we do see some instances of very good agreement as for instance in the second row of fig .
[ fig : apfidelityn=4 ] where @xmath117 and @xmath88 in eq . .
even in this case , noticeable differences between exact and lda results are seen , but the overall agreement is very good . for the larger particle numbers in figs .
[ fig : apfidelityn=5 ] and [ fig : apfidelityn=6 ] we may even notice a clear tendency for the lda results to predict large fidelity spikes that are either at different times as compared to similar spikes using the exact result , or in some instances the lda shows spikes where the exact results have none ( see for instance the fourth row in fig .
[ fig : apfidelityn=5 ] or the third row in fig .
[ fig : apfidelityn=6 ] ) .
we thus conclude that in most cases , the lda results can provide large deviations from the exact results in a dynamical process such as quantum state transfer .
in closing this section , we want to consider how well the lda results do in the case where perfect state transfer is achieved .
this can be achieved with parameters corresponding to the second row in fig .
[ fig : apfidelityn=4 ] in a setup where we can discard the @xmath137 interaction in the hamiltonian in eq . and
thus reduce the problem to the heisenberg xx model .
this can be realized by either using a tailored external magnetic field applied to the system or in specific models with strongly interacting two - component bose systems @xcite . in fig .
[ fig : apperftransfern=4 ] we compare the fidelities computed using the exact and the lda values of the exchange coupling coefficients .
we see that the perfect transfer is indeed obtained for the values calculated via eq .
however , the fidelity of the spin transfer decreases with time for the coefficients obtained via eq . .
the relative differences in the @xmath54 exchange coefficients are no more than @xmath138 percent when comparing the exact to the lda method .
however , we still see that it affects the quantum transfer properties on longer time scales significantly .
, for an @xmath79 two - component fermi system with @xmath115 obtained from the potential in eq . for @xmath117 and @xmath88 ( in units of @xmath87 ) for the case
where the @xmath127-components in eq .
can be discarded so that we are in the so - called heisenberg xx spin model . in this case
perfect state transfer can be achieved @xcite .
the ( blue ) solid line corresponds to the @xmath115 coefficients obtained from the exact calculation and the ( brown ) dashed line corresponds to the lda approach .
time is measured in units of @xmath120 . ]
we have considered strongly interacting two - component systems in one dimension held in place by an external confinement .
such systems have an effective hamiltonian that can be completely specified by computing a set of local exchange coefficients which may be interpreted as nearest - neighbor spin exchange interactions when the system is mapped onto a spin model hamiltonian of the heisenberg type .
computing these exchange coefficients as accurately as possible is an important yet also computationally difficult task . in the present paper
we have explored two different approaches .
one is a brute force calculation of a multi - dimensional integral which gives the exact result ( but is prohibitive for larger particle numbers ) and the other is an approach inspired by the local density approximation ( lda ) that could in principle be used to reduce the computational complexity . while previous studies have shown that the local density approach can be accurate at the level of a few percent for the case of a harmonic oscillator potential , we explore more complicated geometries consisting of two instances of a double - well potential .
our findings demonstrate that while the lda does rather well with potentials that resemble a harmonic oscillator , the exchange couplings do not have the right qualitative and quantitative behavior in the lda when there are significant barriers as is typical of a double - well potential .
in particular , the lda can not capture the right quantum tunneling processes across such barriers and may thus leave out important effects .
we have shown that for small systems this can lead to an underestimation of exchange by the lda for potentials which have a more interesting structure than the simple single - well harmonic oscillator .
we expect this observation to have influence on experiments with small system sizes in 1d optical lattices .
there one typically also has an external overall smooth potential ( approximately harmonic in shape ) .
thus , the potential seen by the particles is the superposition of harmonic trap and optical lattice , i.e. a smiling lattice potential .
this is a very structured potential and in the strongly interacting regime one could be in dire straits with a simple lda approach as compared to the exact exchange couplings . in order to test out the lda in a concrete physical process , we considered quantum state transfer of single spin flips in systems with four , five , and six particles .
here we found that while the lda performs reasonably for some double - well realizations there is amplification of the deviations of the lda compared to the exact results in the transfer fidelity that can be significant and lead to large errors in both maximum fidelity values and the specific times at which these are attained .
the current study has concentrated on small particle numbers where very accurate results can be obtained for the local exchange coefficients using multi - dimensional integration so that a comparison between exact and lda results is possible for arbitrary potentials . incidentally , as we discussed in the introduction , the system sizes used here are also of great current experimental interest in cold atoms .
one would , however , like to study how these results scale to larger particle numbers .
this most likely requires alternative approaches not only to the exact formula but also to the lda formula . in the current implementation
it depends on the density of the system which is not easy to compute for arbitrary potentials .
one may thus pursue an agenda of finding an alternative lda method that obtains the density by some other and computationally much simpler approach , and simultaneously explore how to get a computational reduction of the exact formula .
the authors gratefully acknowledge discussions with and feedback from a. g. volosniev , m. valiente , d. petrosyan , n. j. s. loft , a. e. thomsen , and l. b. kristensen .
this work was supported by the danish council for independent research dff natural sciences and the dff sapere aude program .
99 i. bloch , j. dalibard , and w. zwerger , rev .
* 80 * , 885 ( 2008 ) .
m. lewenstein _
et al . _ ,
* 56 * , 243 ( 2007 ) .
t. esslinger , ann .
cond . mat .
* 1 * , 129 ( 2010 ) .
y. nishida and d. t. son , phys .
d * 76 * , 086004 ( 2007 ) ; d. t. son , phys .
d * 78 * , 046003 ( 2008 ) .
k. balasubramanian and j. mcgreevy , phys .
* 101 * , 061601 ( 2008 ) ; j. maldacena , d. martelli , and y , tachikawa , j. high energy phys . 10 ( * 2008 * ) 072 ; a. adams , k. balasubramanian , and j. mcgreevy , j. high energy phys . 11 ( * 2008 * ) 059 .
n. t. zinner and a. s. jensen , j. phys .
g : nucl . part .
phys . * 40 * , 053101 ( 2013 ) .
m. a. baranov , m. dalmonte , g. pupillo , and p. zoller , chem .
rev . * 112 * , 5012 ( 2012 ) .
h. moritz , t. stferle , m. khl , and t. esslinger , phys .
* 91 * , 250402 ( 2003 ) .
t. stferle , h. moritz , c. schori , m. khl , and t. esslinger , phys .
* 92 * , 130403 ( 2004 ) . t. kinoshita , t. wenger , and d. s. weiss , science
* 305 * , 1125 ( 2004 ) .
b. paredes _ et al .
_ , nature * 429 * , 277 ( 2004 ) . t. kinoshita , t. wenger , and d. s. weiss , nature * 440 * , 900 ( 2005 ) .
e. haller _ et al .
_ , science * 325 * , 1224 ( 2009 ) .
e. haller _ et al .
_ , nature * 466 * , 597 ( 2010 ) .
m. olshanii , phys .
lett . * 81 * , 938 ( 1998 ) .
l. w. tonks , phys . rev . * 50 * , 955 ( 1936 ) .
m. d. girardeau , j. math . phys . * 1 * , 516 ( 1960 ) .
g. pagano _
_ , nature phys . * 10 * , 198 ( 2014 ) .
f. serwane _
et al . _ , science * 332 * , 336 ( 2011 )
. g. zrn _
et al . _ ,
* 108 * , 075303 ( 2012 ) .
et al . _ ,
* 111 * , 175302 ( 2013 ) .
et al . _ , science * 342 * , 457 ( 2013 )
. s. murmann _
rev . lett . * 114 * , 080402 ( 2015 ) .
s. murmann _
et al . _ ,
lett . * 115 * , 215301 ( 2015 ) .
s. zllner , h .- d .
meyer , and p schmelcher , phys .
a * 74 * , 063611 ( 2006 ) ; phys . rev .
a * 75 * , 043608 ( 2007 ) ; phys .
* 100 * , 040401 ( 2008 ) .
f. deuretzbacher , k. bongs , k. sengstock , and d. pfannkuche , phys .
a * 75 * , 013614 ( 2007 ) .
e. tempfli , s. zllner , and p. schmelcher , new j. phys .
* 11 * , 073015 ( 2009 ) .
m. d. girardeau , phys .
a * 83 * , 011601(r ) ( 2011 ) .
i. brouzos and p. schmelcher , phys .
. lett . * 108 * , 045301 ( 2012 ) .
i. brouzos and a. frster , phys .
rev . a * 89 * , 053632 ( 2014 ) .
b. wilson , a. frster , c. c. n. kuhn , i. roditi , and d. rubeni , phys .
a * 378 * , 1065 ( 2014 ) .
n. t. zinner _
et al . _ ,
europhys . lett .
* 107 * , 60003 ( 2014 ) .
m. a. garcia - march _ et al .
a _ 92 _ , 033621 ( 2015 ) .
a. s. dehkharghani _
_ , scientific reports * 5 * , 10675 ( 2015 ) .
p. massignan , j. levinsen , and m. m. parish , phys .
* 115 * , 247202 ( 2015 ) .
l. guan , s. chen , y. wang , and z .- q .
ma , phys .
lett . * 102 * , 160402 ( 2009 ) . c. n. yang , chin .
phys . lett .
* 26 * , 120504 ( 2009 ) .
guan and z .- q .
ma , phys .
a * 85 * , 033632 ( 2012 ) .
d. rubeni , a. frster , and i. roditi , phys . rev . a * 86 * , 043619 ( 2012 ) . i. brouzos and p. schmelcher , phys .
a * 87 * , 023605 ( 2013 ) .
p. o. bugnion and g. j. conduit , phys .
a * 87 * , 060502(r ) ( 2013 ) .
s. e. gharashi and d. blume , phys .
* 111 * , 045302 ( 2013 ) .
t. sowiski , t. gra , o. dutta , and m. lewenstein , phys .
a * 88 * , 033607 ( 2013 ) .
s. e. gharashi , x. y. yin , and d. blume , phys .
rev . a * 89 * , 023603 ( 2014 ) .
a. g. volosniev _
_ , few - body syst . * 55 * , 839 ( 2014 ) .
x. cui and t .- l . ho , phys . rev .
a * 89 * , 023611 ( 2014 ) .
n. j. s. loft _
et al . _ ,
j. d * 69 * , 65 ( 2015 ) .
r. lundmark , c. forssn , and j. rotureau , phys .
rev . a * 91 * , 041601(r ) ( 2015 ) .
s. e. gharashi , x. y. yin , and d. blume , phys .
rev . a * 91 * , 013620 ( 2015 ) . f. nur nal , b. hetnyi , and m. . oktel , phys .
rev . a * 91 * , 053625 ( 2015 ) .
a. g. volosniev , h .- w .
hammer , and n. t. zinner , arxiv:1507.00186 ( 2015 ) .
m. d. girardeau , phys .
a * 82 * , 011607(r ) ( 2010 ) .
l. guan and s. chen , phys .
* 105 * , 175301 ( 2010 ) .
g. e. astrakharchik and i. brouzos , phys .
a * 88 * , 021602(r ) ( 2013 ) .
a. g. volosniev _
et al . _ , nature commun . * 5 * , 5300 ( 2014 ) .
e. j. lindgren _
et al . _ , new j. phys . * 16 * , 063003 ( 2014 ) .
f. deuretzbacher _ et al .
a * 90 * , 013611 ( 2014 ) .
a. g. volosniev _
et al . _ ,
rev . a * 91 * , 023620 ( 2015 ) .
l. yang , l. guan , and h. pu , phys . rev .
a * 91 * , 043634 ( 2015 ) .
j. levinsen , p. massignan , g. m. bruun , and m. m. parish , science adv .
* 1 * , e1500197 ( 2015 ) t. sowiski , m. gajda , and k. rzaewski , europhys
* 109 * , 26005 ( 2015 ) .
t. grining _ et al .
_ , phys . rev .
a * 92 * , 061601 ( 2015 ) .
m. d. girardeau and m. olshanii , phys .
a * 70 * , 023608 ( 2004 ) .
m. d. girardeau and a. minguzzi , phys .
lett . * 99 * , 230402 ( 2007 ) .
s. zllner , h .- d .
meyer , and p schmelcher , phys .
a * 78 * , 013629 ( 2008 ) ; x .- w .
guan , m. t. batchelor , and j .- y .
lee , phys .
a * 78 * , 023621 ( 2008 ) .
f. deuretzbacher _ et al .
_ , phys . rev
. lett . * 100 * , 160405 ( 2008 ) .
b. fang , p. vignolo , m. gattobigio , c. miniatura , and a. minguzzi , phys .
a * 84 * , 023626 ( 2011 ) .
n. l. harshman , phys .
a * 86 * , 052122 ( 2012 ) .
m. a. garcia - march and th .
busch , phys .
a * 87 * , 063633 ( 2013 ) .
m. a. garcia - march _ et al .
a * 88 * , 063604 ( 2013 ) .
n. l. harshman , phys .
rev . a * 89 * , 033633 ( 2014 ) .
s. campbell , m. a. garcia - march , t. fogarty , and th .
busch , phys .
a * 90 * , 013617 ( 2014 ) .
m. a. garcia - march _ et al .
_ , new j. phys .
* 16 * , 103004 ( 2014 ) .
p. damico and m. rontani , j. phys .
b * 47 * , 065303 ( 2014 ) ; phys . rev . a * 91 * , 043610 ( 2015 ) .
n. p. mehta , phys .
rev . a * 89 * , 052706 ( 2014 ) .
m. a. garcia - march _ et al .
a * 90 * , 063605 ( 2014 ) .
r. e. barfknecht , i. brouzos , a. frster , phys .
rev . a * 91 * , 043640 ( 2015 ) .
a. g. volosniev , h .- w .
hammer , and n. t. zinner , phys .
a * 92 * , 023623 ( 2015 ) .
a. s. dehkharghani , a. g. volosniev , and n. t. zinner , phys .
a * 92 * , 031601(r ) ( 2015 ) .
n. l. harshman , few - body syst .
* 57 * , 11 ( 2016 ) ; few - body syst . *
57 * , 45 ( 2016 ) .
d. pcak , m. gajda , and t. sowiski , new j. phys . *
18 * , 013030 ( 2016 ) .
t. gra , phys .
a * 92 * , 023634 ( 2015 ) .
a. s. dehkharghani , a. g. volosniev , and n. t. zinner , arxiv:1511.01702 ( 2015 ) .
d. s. petrov , g. v. shlyapnikov , and j. t. m. walraven , phys .
lett . * 85 * , 3745 ( 2001 ) .
a. recati , p. o. fedichev , w. zwerger , and p. zoller , phys .
lett . * 90 * , 020401 ( 2003 ) .
g. e. astrakharchik , d. blume , s. giorgini , and l. p. pitaevskii , phys .
* 93 * , 050402 ( 2004 ) .
i. v. tokatly , phys .
lett . * 93 * , 090405 ( 2004 ) .
g. e. astrakharchik , phys .
a * 72 * , 063620 ( 2005 ) ; g. e. astrakharchik , j. boronat , j. casulleras , and s. giorgini , phys .
lett . * 95 * , 190407 ( 2005 ) .
g. orso , phys .
lett . * 98 * , 070402 ( 2007 ) .
liu , h. hu , and p. d. drummond , phys . rev . a * 78 * , 023601 ( 2008 ) .
m. colom - tatch , phys .
a * 78 * , 033612 ( 2008 ) . g. de rosi and s. stringari , phys .
a * 92 * , 053617 ( 2015 ) .
m. chenaeu _
et al . _ ,
nature * 481 * , 484 ( 2012 ) .
t. fukuhara _ et al . _ , nature physics * 9 * , 235 ( 2013 ) ; t. fukuhara _ et al . _ ,
nature * 502 * , 76 ( 2013 ) .
lett . * 113 * , 147205 ( 2014 ) .
t. fukuhara _ et al .
lett . * 115 * , 035302 ( 2015 ) .
s. trotzky _
et al . _ ,
science * 319 * , 295 ( 2008 ) . m. ogata and h. shiba ,
b * 41 * , 2326 ( 1990 ) .
k. a. matveev , phys .
b * 70 * , 245319 ( 2004 ) .
guan , m. t. batchelor , and m. takahashi , phys .
a * 76 * , 043617 ( 2007 ) .
k. a. matveev and a. furusaki , phys .
* 101 * , 170403 ( 2008 ) .
s. bose , phys .
lett . * 91 * , 207901 ( 2003 ) s. bose , contemporary physics , volume * 48 * , issue 1 , 13 - 30 ( 2007 ) m. christandl , _ et al .
a * 71 * , 032312 ( 2005 ) m. christandl , _ et al .
lett . * 92 * , 187902 ( 2004 ) g. m. nikolopoulos , _ et al .
_ , europhys . lett . *
65*(3 ) , 297 - 303 ( 2004 ) | one - dimensional multi - component fermi or bose systems with strong zero - range interactions can be described in terms of local exchange coefficients and mapping the problem into a spin model is thus possible . for arbitrary external confining potentials
the local exchanges are given by highly non - trivial geometric factors that depend solely on the geometry of the confinement through the single - particle eigenstates of the external potential . to obtain accurate effective hamiltonians to describe such systems one needs to be able to compute these geometric factors with high precision which is difficult due to the computational complexity of the high - dimensional integrals involved .
an approach using the local density approximation would therefore be a most welcome approximation due to its simplicity .
here we assess the accuracy of the local density approximation by going beyond the simple harmonic oscillator that has been the focus of previous studies and consider some double - wells of current experimental interest .
we find that the local density approximation works quite well as long as the potentials resemble harmonic wells but break down for larger barriers . in order to explore the consequences of applying the local density approximation in a concrete setup we consider quantum state transfer in the effective spin models that one obtains . here
we find that even minute deviations in the local exchange coefficients between the exact and the local density approximation can induce large deviations in the fidelity of state transfer for four , five , and six particles . |
the reionization epoch is now within the direct observational reach thanks to rapid recent observational advances in two fronts optical quasar absorption from sloan digital sky survey ( sdss ) ( fan et al .
2001 ; becker et al .
2001 ) and the wilkinson microwave anisotropy probe ( wmap ) experiment ( kogut et al . 2003 ) .
the picture painted by the combined observations , perhaps not too surprisingly , strongly suggests a complex cosmological reionization process , consistent with the double reionization scenario ( cen 2003 ) .
it may be that this is the beginning of a paradigm shift in our focus on the high redshift universe : the star formation history of the early universe can now be observationally constrained .
it thus becomes urgent to theoretically explore galaxy and star formation process at high redshift in the dark age ( @xmath14 ) . in the context of the standard cold dark matter model
it is expected that stars within halos of mass @xmath15 at high redshift play an important , if not dominant , role in determining how and when the universe was reionized .
furthermore , these fossil halos may be seen in the local universe as satellites of giant galaxies .
this linkage may potentially provide a great leverage to nail down the properties of the high redshift galaxies . in this paper , as a step towards understanding galaxy formation at high redshift
, we investigate the properties of dark matter halos at @xmath14 , using very high resolution tpm n - body ( bode et al .
2001 ; bode & ostriker 2003 ) simulations .
while there is an extensive literature on properties of halos at low redshift , there is virtually no systematic study of dark halos at @xmath14 .
the lcdm simulation has a comoving box size of @xmath16mpc with @xmath17 particles , a particle mass of @xmath18 , and comoving gravitational softening length of @xmath19kpc .
these resolutions allow us to accurately characterize the properties of halos down to a mass @xmath20 ( having about @xmath21 particles within the virial radius ) .
the outline of this paper is as follows .
the simulation details are given in 2 . in
3 we quantify properties of dark matter halos in the mass range @xmath0 , including the mass function , bias and clustering properties , density profile distribution , angular momentum spin parameter distribution , internal angular momentum distribution and peculiar velocity distribution .
we conclude in 4 .
a standard spatially flat lcdm cosmology was chosen , with @xmath22 and @xmath23 ; the hubble constant was taken to be 70 km / s / mpc .
the initial conditions were created using the grafic2 package by bertschinger ( 2001 ) .
the matter transfer function was calculated with the included boltzmann integrator ( ma & bertschinger 1995 ) , using @xmath24 for the baryon fraction and @xmath25 for the normalization of the matter power spectrum .
the simulation contained @xmath26 particles in a comoving periodic box 4@xmath27mpc on a side , making the particle mass @xmath28 .
the starting redshift was @xmath29 , and the system was evolved down to @xmath30 .
the evolution was carried out with the parallel tree particle
mesh code tpm ( xu 1995 ; bode , ostriker , & xu 2000 ; bode & ostriker 2003 ) , using a @xmath31 mesh .
the evolution took 1150 pm steps , with particles in dense regions taking up to 19,500 steps .
the run was carried out using up to 256 processors on the terascale computing system at pittsburgh supercomputing center .
the spline softening length in the tree portion of the code was set to @xmath32kpc . with this softening length , relaxation by @xmath30 inside the core of a collapsed halo ( assuming an nfw density distribution with @xmath33 )
will not be significant over the course of the simulation for those objects containing more than 100 particles .
the opening angle in the barnes - hut criterion used by tpm was @xmath34 , and the time step parameter @xmath35 ; also , the initial values for locating trees were @xmath36 and @xmath37 see bode & ostriker ( 2003 ) for details . in the tpm code ,
not all regions are treated at full resolution . the limiting density ( above which all cells are put into trees for increased resolution ) rises with time . by the end of this run ,
all cells containing more than 18 particles are still being followed at full resolution .
thus this factor is not important if the analysis is limited to halos with over 100 particles .
dark matter halos are identified using denmax scheme ( bertschinger & gelb 1991 ) , smoothing the density field with a gaussian length of @xmath38kpc . in computing all quantities
we include all particles located inside the virial radius of a halo .
* m @xmath39 * 0.01 in , respectively , at @xmath30 .
[ f2a],title="fig : " ] * m @xmath40 * 0.01 in * m @xmath39 * 0.01 in * m @xmath41 * 0.01 in * m @xmath42 * 0.01 in first , we present visually a distribution of the dark matter mass and dark matter halos of varying masses .
figure 1 shows the distribution of dark matter particles projected onto the x - y plane .
figures ( 2a , b , c , d , e ) show the distributions of dark matter halos with masses greater than @xmath43 , respectively , at @xmath30 . the progressively stronger clustering of more massive halos is clearly visible in the display but we will return to the clustering properties more quantitatively in 3.3 .
it is also noted that voids are progressively more visible in the higher mass halos than in low mass halos .
rrrrrrrr @xmath44 ( @xmath45 mpc@xmath46 ) & 0.85 & 1.20 & 1.75 & 2.40 + @xmath47 & 1.9 & 2.0 & 2.1 & 2.2 + @xmath48 ( @xmath27 mpc ) & @xmath49 & @xmath50 & @xmath51 & @xmath52 + figure 3 shows the halo mass functions at four redshifts .
table 1 summarizes the fitting parameters for a schechter function of the form @xmath53 we see that the schechter function provides a good fit to the computed halo mass function .
the faint end slope is @xmath54 , consistent with the expectation from press - schechter theory ( press & schechter 1974 ) .
while there appears to be a slight steepening of the slope at the low mass end from @xmath55 to @xmath56 from @xmath30 to @xmath57 , it is unclear , however , how significant this trend is , given the adopted , somewhat degenerate fitting formula .
the turnover at @xmath58 indicates the loss of validity of our simulation at the low mass end .
0.05 in rrrrrrrr @xmath5910@xmath60 & [email protected] & [email protected] & [email protected] & [email protected] + @xmath5910@xmath62 & [email protected] & [email protected] & [email protected] & [email protected] + @xmath5910@xmath63 & [email protected] & [email protected] & [email protected] & [email protected] + @xmath5910@xmath64 & [email protected] & [email protected] & [email protected] & [email protected] + @xmath5910@xmath65 & [email protected] & [email protected] & [email protected] & [email protected] + rrrrrrrr
@xmath5910@xmath60 & [email protected] & [email protected] & [email protected] & [email protected] + @xmath5910@xmath62 & [email protected] & [email protected] & [email protected] & [email protected] + @xmath5910@xmath63 & [email protected] & [email protected] & [email protected] & [email protected] + @xmath5910@xmath64 & [email protected] & [email protected] & [email protected] & [email protected] + @xmath5910@xmath65 & [email protected] & [email protected] & [email protected] & [email protected] + we characterize the relative distribution of halos over the total dark matter distribution by the following relation : @xmath66 where @xmath67 and @xmath68 are the halo density and mean halo density ; @xmath69 and @xmath70 are the mass density and mean mass density ; @xmath71 is fixed to be unity at @xmath72 .
this empirical fitting formula is motivated by the found result that there appears to be a break in @xmath73 at @xmath74 .
tables 2 and 3 list the parameters @xmath75 and @xmath71 .
the smoothing length used here is 0.3 @xmath27 mpc .
figure 4 shows four typical cases to indicate the goodness of the fitting formula . at @xmath76 our results ( table 2 )
indicate @xmath77 , a rather rapid drop .
as expected , the drop - off is more dramatic for larger halos , as visible in figure 2 .
this implies that at @xmath78 halos are unlikely to be found in underdense regions ( on a scale of @xmath79 ) .
the increase of @xmath71 with redshift implies that voids are emptier at higher redshifts .
another way to characterize the relative distribution of dark matter halos over mass is to compute the ratio of the correlation functions , which are shown in figure 5 .
the correlation function @xmath80 is calculated by counting the number of pairs of either particles or halos at separation @xmath81 ( using logarithmically spaced bins ) and comparing that number to a poisson distribution .
it can be seen that the bias falls in the range @xmath2 , with the trend that the more massive halos are more biased and at a fixed mass halos at higher redshifts are more biased , as expected .
figure 6 recollects the information in figure 5 and shows the bias as a function of halo mass at four different redshifts at the scale chosen to be @xmath82mpc .
the agreement between our computed results and that using analytic method of mo & white ( 1996 ) is good , indicating that the latter is valid for objects at scales and redshifts of concern here .
0.01 in 0.01 in we use a variant fitting formula based on the nfw ( navarro , frenk , & white 1997 ) profile : @xmath83 note that for nfw profile , @xmath84
. an important difference , however , is the scaling radius used .
we use the radius where the logarithmic slope of the density profile is @xmath85 , @xmath86 , instead of the more conventional core " radius .
this is a two parameter fitting formula , @xmath47 and @xmath86 , while @xmath87 is a function of @xmath47 and @xmath86 at a fixed redshift , since the overdensity interior to the virial radius @xmath88 is assumed to be known .
this fitting formula is intended for the range in radius @xmath89 only .
we fit the density profile of each halo using the least squares method .
four randomly selected examples of such profiles along with the fitted curve using equation ( 3 ) are shown in figure 7 , indicating reasonable fits in all cases . both fitting parameters , @xmath47 and @xmath86 , however , display broad distributions .
we found that there is only weak correlation between @xmath47 and @xmath86 .
figure 8 shows histograms for the distributions of @xmath47 , for four typical cases .
we fit @xmath47 distributions using a gaussian distribution function : @xmath90 the gaussian fits are shown as smooth curves in figure 8 , demonstrating that the proposed gaussian fits are good .
tables 4,5 list fitting parameters @xmath9 and @xmath91 , respectively .
we recollect the data in table 4 and shows in figure 9 the median inner density slope as a function of dark matter halo different mass at four different redshifts ( symbols ) .
the curves in figure 9 are empirical fits using the following formula @xmath92 it is seen that this fitting formula provides a reasonable fit for the simulated halos .
rrrrrrrr 10@xmath63 - 10@xmath64 & [email protected] & [email protected] & [email protected] & [email protected] + 10@xmath64 - 10@xmath65 & [email protected] & [email protected] & [email protected] & [email protected] + @xmath59 10@xmath65 & [email protected] & [email protected] & [email protected] & [email protected] + rrrrrrrr 10@xmath63 - 10@xmath64 & [email protected] & [email protected] & [email protected] & [email protected] + 10@xmath64 - 10@xmath65 & [email protected] & [email protected] & [email protected] & [email protected] + @xmath59 10@xmath65 & [email protected] & [email protected] & [email protected] & [email protected]
+ rrrrrrrr 10@xmath63 - 10@xmath64 & [email protected] & [email protected] & [email protected] & [email protected] + 10@xmath64 - 10@xmath65 & [email protected] & [email protected] & [email protected] & [email protected] + @xmath59 10@xmath65 & [email protected] & [email protected] & [email protected] & [email protected]
+ rrrrrrrr 10@xmath63 - 10@xmath64 & [email protected] & [email protected] & [email protected] & [email protected] + 10@xmath64 - 10@xmath65 & [email protected] & [email protected] & [email protected] & [email protected] + @xmath59 10@xmath65 & [email protected] & [email protected] & [email protected] & [email protected] + figure 10 shows histograms for the distributions of @xmath86 for four typical cases .
we fit @xmath86 distributions using a lognormal function : @xmath93 which are seen to provide reasonable fits to the data in figure 10 .
tables 6,7 list fitting parameters @xmath94 and @xmath95 , respectively . we find no visible correlation between @xmath47 and @xmath86 , as shown in figure 11 .
0.01 in an important point to note is that the average slope of the density profiles of small halos ranges from @xmath96 to @xmath97 from @xmath30 to @xmath98 ( table 4 and figure 8) , which is somewhat shallower than the universal density profile found by navarro et al .
our results at @xmath99 are in good agreement with ricotti ( 2003 ) , who simulated a @xmath82mpc box for small halos at @xmath100 .
if the theoretical argument for the dependence of halo density profile on the slope of the initial density fluctuation power spectrum ( syer & white 1988 ; subramanian , cen , & ostriker 2000 ) is correct , as adopted by ricotti ( 2003 ) to explain the dependence of inner density slope on halo mass , it then follows that the neglect of density fluctuations on scales larger than his box size of @xmath82mpc in ricotti ( 2003 ) would make the density profiles of the halos in his simulation somewhat shallower than they should be in the cdm model .
thus , the difference in the box size ( @xmath82mpc in ricotti 2003 versus @xmath16mpc for our simulation box ) would have expected to result in a slightly steeper inner slope in our simulation , which is indeed the case .
another point to note , which is not new but not widely known , is that there is a large dispersion in the inner slope of order @xmath101 due to the intrinsically stochastic nature of halo assembly .
this was found earlier by subramanian et al .
( 2000 ) . therefore , while a universal " profile is informative in characterizing the mode , a dispersion would be needed to give a full account .
this is particularly important for applications where the dependence on the inner slope is very strong , e.g. , strong gravitational lensing .
more relevant for our case of small halos is that , for example , the fraction of small halos with inner slope close to zero ( i.e. , flat core ) is non - negligible at @xmath30 .
a proper statistical comparison with observations of local dwarf galaxies , however , is not possible with the current simulation without evolving small galaxies to @xmath102 . , versus @xmath86 for all the halos with @xmath103 at @xmath30 .
[ f10 ] ] , versus the halo central density , defined as the density at @xmath104 .
for all the halos with @xmath105 at @xmath30 .
[ f10a ] ] , versus the environmental density , defined as the dark matter density smoothed by a gaussian window of radius @xmath106mpc , for all the halos with @xmath105 at @xmath30 .
[ f10b ] ] also worthwhile is to understand whether or not there is some dependence of the inner slope of the density profile on the central density of a halo or the environmental density where a halo sits . in figures 12 and 13
we show the correlation between the inner slope of the density profile , @xmath47 , and the central density of the halo , and between @xmath47 and the environmental density , respectively .
we find no visible correlations between either pair of quantities . but
a relationship between halo shape and environment might have been missed due to stochastic variations .
thus , we have checked to see if deviation @xmath107 , between computed and predicted ( based on mass and epoch using equation 5 ) slope exists .
figure 14 shows this and no correlation is seen .
we note that the abundances of halos in the mass range of interest here are on the rise in the redshift range considered ( lacey & cole 1993 ) , as evident in figure 3 . during this period halos considered
the inner density profiles of halos steepen with time , consistent with the increase of logarithmic slope of the power spectrum with time , that corresponds to the evolving nonlinear mass scale .
however , at some lower redshift not probed here , low mass halos will cease to form .
subsequently , the evolution of halo density profile may show distinct features and some conceivable correlations , not seen in figures 12 and 13 , may show up .
we will study this issue separately . since the central density of a halo
may be considered a good proxy for the formation redshift of the central region , the non - correlation between @xmath47 and the central density indicates that , for halos of question here , the subsequent process of accretion of mass onto halos is largely a random process , independent of the density of the initial central seed " .
the fact that halos of different masses show a comparable range of central density ( not shown in figures ) suggest that halos of varying masses form nearly simultaneously , dictated by the nature of the cold dark matter power spectrum at the high - k end ; i.e. , density fluctuations on those scales involved here depend weakly ( logarithmically ) on the mass .
the non - correlation between the inner slope of the halo density profile and the environmental density may be interpreted in the following way .
one may regard regions of different overdensities as local mini - universes of varying density parameters .
the independence of the inner slope on local density is thus consistent with published results that the halo density profiles in universes of different @xmath108 do not significantly vary .
0.01 in , versus the halo central density , defined as the density at @xmath104 . for all the halos with @xmath105 at @xmath30 .
[ f10b ] ] versus spin parameter @xmath109 .
no correlation between the two is observed .
[ f11 ] ] rrrrrrrr 10@xmath110 - 10@xmath111 & [email protected] & [email protected] & [email protected] & [email protected] + 10@xmath111 - 10@xmath112 & [email protected] & [email protected] & [email protected] & [email protected] + 10@xmath112 - 10@xmath113 & [email protected] & [email protected] & [email protected] & [email protected] + @xmath59 10@xmath113 & [email protected] & [email protected] & [email protected] & [email protected] + rrrrrrrr 10@xmath110 - 10@xmath111 & [email protected] & [email protected] & [email protected] & [email protected] + 10@xmath111 - 10@xmath112 & [email protected] & [email protected] & [email protected] & [email protected] + 10@xmath112 - 10@xmath113 &
[email protected] & [email protected] & [email protected] & [email protected] + @xmath59 10@xmath113 & [email protected] & [email protected] & [email protected] & [email protected] + rrrrrrrr 10@xmath63 - 10@xmath64 & [email protected] & [email protected] & [email protected] & [email protected] + 10@xmath64 - 10@xmath65 & [email protected] & [email protected] & [email protected] & [email protected] + @xmath59 10@xmath65 & [email protected] & [email protected] & [email protected] & [email protected] + rrrrrrrr
10@xmath63 - 10@xmath64 & [email protected] & [email protected] & [email protected] & [email protected] + 10@xmath64 - 10@xmath65 & [email protected] & [email protected] & [email protected] &
[email protected] + @xmath59 10@xmath65 & [email protected] & [email protected] & [email protected] & [email protected] + we compute the spin parameter defined as @xmath114 ( peebles 1969 ) , where @xmath115 is the gravitational constant ; @xmath116 is the total mass of the dark matter halo ; @xmath117 is the total angular momentum of the dark matter halo ; @xmath118 is the total energy of the dark matter halo ; all quantities are computed within the virial radius .
we fit the @xmath109 distributions using a modified lognormal function : @xmath119 ^ 2\over 2\sigma_{\lambda}})},\ ] ] where @xmath120 is fixed to be @xmath121 , which is determined through experimentation .
the modified lognormal fits for @xmath109 are shown as smooth curves in figure 15 ; the goodness of the fits is typical .
tables 8,9 lists fitting parameters @xmath122 and @xmath123 , respectively .
note that the median value of @xmath109 is @xmath124 .
we see that the typical spin parameter has a value @xmath12
. however , the distribution of the spin parameter among halos is very broad , with a lognormal dispersion of @xmath13 .
this implies that consequences that depend on the spin of a halo are likely to be widely distributed even at a fixed dark matter halo mass .
such consequences may include the size of a galactic disk and correlations between dark matter halo spin ( conceivable misalignment between spin of gas and spin of dark matter would complicate the situation ) and other quantities . in figures 16 and 17
we show the correlation between @xmath109 and the central density of the halo , and between @xmath109 and the environmental density , respectively .
we find that there may possibly exist a weak correlations between @xmath109 and the central density of the halo , in the sense that halos with higher central densities ( or equivalently earlier formation times ) have lower @xmath109 , with a very large scatter , whereas no correlation is discernible between @xmath109 and the environmental density .
finally , figure 18 show the relation between @xmath109 and @xmath47 , where no correlation is visible .
0.01 in 0.01 in next , we compute the angular momentum profiles for individual dark matter halos .
we then fit the each angular momentum profile by the following function in the small @xmath125 regime : @xmath126 where @xmath127 and @xmath128 are two fitting parameters ; @xmath129 is the virial mass and @xmath130 . in order to compute the angular momentum profile an appropriate smoothing window needs to be applied to dark matter particles .
we find that @xmath131 varies for different smoothing scales , with typical values around 0.2 , while @xmath127 remains roughly constant for each individual halo .
however , @xmath127 is broadly distributed for all dark matter halos .
figure 19 show histograms for the distributions of @xmath132 , for four randomly selected cases .
we fit the @xmath127 distribution using a modified lognormal function : @xmath133 ^ 2\over 2\sigma_{a}})},\ ] ] where @xmath134 is fixed through experimentation .
tables 10,11 list fitting parameters @xmath132 and @xmath135 , respectively . in general we find our fitting formula ( equation 10 ) provides a good fit for each individual halo .
the distribution of matter at small @xmath136 is most relevant for the formation of central objects , such as black holes ( e.g. , colgate et al .
2003 ) or bulges ( e.g. , donghia & burkert 2004 ) .
our calculation indicates that the fraction of mass in a halo having specific angular momentum less than a certain value is roughly @xmath101 times the ratio of that value over the average specific angular momentum of the halo . finally , we compute the peculiar velocity of dark matter halos .
we find that the distribution once again can be fitted by lognormal distributions as equation ( 9 ) . in order to provide a good fit for the results at @xmath30 shown in figure 20
, it is found that @xmath137km / s in equation ( 8) with median velocity @xmath138km / s and lognormal dispersion @xmath139 .
we caution , however , the absolute value of the peculiar velocity of each halo , unlike the quantities examined in previous subsections , may be significantly affected by large waves not present in our simulation box . adding missing large waves should increase the zero point @xmath140 to a larger value .
therefore , the peculiar velocity shown should be treated as a lower limit . in other words ,
expected peculiar velocity of dark matter halos at these redshifts are likely in excess of @xmath141km / s .
using a high resolution tpm n - body simulation of the standard cold dark matter cosmological model with a particle mass of @xmath28 and a softening length of @xmath32kpc in a @xmath16mpc box , we compute various properties of dark matter halos with mass @xmath0 at redshift @xmath1 . we find the following results .
\(3 ) the central density profile of dark matter halos are found to be in the range @xmath142 well fitted by @xmath10 with a dispersion of @xmath11 , in rough agreement with the theoretical arguments given in ricotti ( 2003 ) and subramanian et al .
( 2000 ) .
\(4 ) the median spin parameter of the dark matter halos is @xmath12 but with a lognormal dispersion of @xmath13 .
the angular momentum profile at the small end is approximately linear with the fraction of mass in a halo having specific angular momentum less than a certain value is roughly @xmath101 times the ratio of that value over the average specific angular momentum of the halo .
this research was supported in part by ast-0206299 and nag5 - 13381 .
the computations were performed on the national science foundation terascale computing system at the pittsburgh supercomputing center .
becker , r.h .
, et al .
2001 , aj , 122 , 2850 bertschinger , e. , & gelb , j.m .
1991 , computers in physics , 5 , 164 bertschinger , e. 2001 , , 137 , 1 bode , p. , & ostriker , j.p .
2003 , apjs , 145 , 1 bode , p. , ostriker , j.p .
, & xu , g. 2000 , , 128 , 561 cen , r. 2003 , apj , 591 , 12 colgate , s.a . , cen , r. , li , h. , currier , n. , & warren , m.s .
2003 , apjl in press , astro - ph/0310776 donghia , e. , & burkert , a. 2004 , astro - ph/0402504 fan , x. , et al .
2001 , aj , 122 , 2833 lacey , c. , & cole , s. 1993 , mnras , 262 , 627 .
ma , c .- p . , & bertschinger , e. 1995 , , 455 , 7 mo , h.j .
, & white , s.d.m .
1996 , mnras , 282 , 347 navarro , j. , frenk , c.s .
, & white , s.d.m .
1997 , apj , 490 , 493 peebles , p.j.e .
1969 , apj , 155 , 393 press , w.h . , & schechter , p. 1974 , apj , 187 , 425 peebles , p.j.e .
1993 , principles of physical cosmology ( princeton : princeton university press ) ricotti , m. 2003 , mnras , 344 , 1237 syer , d. , & white , s.d.m .
1998 , mnras , 293 , 337 subramanian , k. , cen , r. , & ostriker , j.p .
2000 , apj , 538 , 528 xu , g. 1995 , , 98 , 355 | we compute the properties of dark matter halos with mass @xmath0 at redshift @xmath1 in the standard cold dark matter cosmological model , utilizing a very high resolution n - body simulation .
we find that dark matter halos in these mass and redshift ranges are significantly biased over matter with a bias factor in the range @xmath2 .
the dark matter halo mass function displays a slope of @xmath3 at the small mass end .
we do not find a universal dark matter density profile .
instead , we find a significant dependence of the central density profile of dark matter halos on halo mass and epoch with @xmath4 ; the high - mass ( @xmath5 ) low - redshift ( @xmath6 ) halos occupy the high end of the range and low - mass ( @xmath7 ) high - redshift ( @xmath8 ) halos occupy the low end .
additionally , for fixed mass and epoch there is a significant dispersion in @xmath9 due to the stochastic assembly of halos .
our results fit a relationship of the form @xmath10 with a dispersion about this fit of @xmath11 and no systematic dependence of variance correlated with environment .
the median spin parameter of dark matter halos is @xmath12 but with a large lognormal dispersion of @xmath13 .
various quantities are tabulated or fitted with empirical formulae . |
this article is a companion article for @xcite . in that article we discussed intermittent dynamics associated with boundary crisis ( homoclinic ) bifurcations in families of unimodal maps . in the present work we treat saddle node bifurcations from the same perspective . by jakobson
s celebrated work @xcite , the logistic family @xmath0 , @xmath1 $ ] , admits absolutely continuous invariant measures ( a.c.i.m.s ) for @xmath2 from a set of positive measure .
in fact , @xmath3 is a ( lebesgue or full ) density point of this set .
a different argument for this result was given by benedicks and carleson @xcite , @xcite .
their reasoning was generalized to unfoldings @xmath4 of unimodal misiurewicz maps @xmath5 , with eventually periodic critical point @xmath6 ( and possessing negative schwarzian derivative ) .
it was shown that the bifurcation value @xmath7 is a density point of the set of parameter values for which @xmath8 admits an absolutely continuous invariant measure @xcite , @xcite .
we generalize these results to unfoldings of saddle - node bifurcations in families @xmath4 of unimodal maps with negative schwarzian derivative .
we establish that a saddle - node bifurcation value occurs as a point of positive density of the parameter set for which there are absolutely continuous invariant measures . under the assumption that @xmath5 is not more than once renormalizable
, we construct parameters for which @xmath8 possesses a.c.i.m.s supported on the maximal interval @xmath9 $ ] , see theorem [ existence ] below .
in contrast to the misiurewicz bifurcation values , the saddle - node bifurcation value is not a full density point of this parameter set .
this is because it is known that the parameter set for which the map has a periodic attractor has positive density at any saddle - node .
following the construction of a.c.i.m.s , we continue with a detailed discussion of the intermittency that occurs due to the saddle - node bifurcation .
that saddle - node bifurcations can give rise to intermittency is known since @xcite , who called intermittency associated with a saddle - node bifurcation type i intermittency .
pomeau and manneville studied type i intermittency in connection with the lorenz model . in the model ,
simplifying ( hyperbolicity ) assumptions on the dynamics outside a neighborhood of the saddle - node periodic orbit are made . in perhaps the most basic example of type i intermittency , in families of unimodal maps , such simplifications are not justified .
this is due to the presence of a critical point .
our discussion of absolutely continuous invariant measures allows us to give a rigorous treatment of intermittent time series , where we explain and prove quantitative aspects earlier discussed numerically in @xcite , see theorems [ intermittency ] and [ saddlenode ] below .
diaz et .
@xcite studied the unfoldings of saddle - node bifurcations in higher dimensional diffeomorphisms and their results imply in the present context that there exists a subset @xmath10 which has positive density at @xmath11 , such that for each @xmath12 , @xmath8 has an absolutely continuous invariant measure .
the measures produced there are supported on small periodic domains on which the map renormalizes to a hnon - like family . in the present context
this corresponds to parameter values inside periodic windows , for which @xmath8 is renormalizable .
however , the invariant measures produced in jakobson s work are supported on the maximal possible interval , @xmath13 $ ] and the invariant measures that we construct are also supported on @xmath13 $ ] . while editing the final draft of this paper , we learned of the results of maria joo costa @xcite , corresponding to part of her 1998 thesis , on related work in a similar context .
she focused on the sink - horseshoe bifurcation in which a sink and a horseshoe collapse , see @xcite , and studied families @xmath14 of unimodal interval maps to describe bifurcations .
the class of families of interval maps studied in @xcite differs from ours ; it consists of unfoldings of unimodal maps @xmath15 \to [ 0,1]$ ] , that possess a saddle node fixed point at @xmath16 , so that the critical point @xmath17 satisfies @xmath18 . for such families
costa derived the analogue of theorem [ existence ] below , stating that absolutely continuous invariant measures supported on the maximal interval occur with positive density at the bifurcation point .
the set - up of costa s proof corresponds to ours in that she also combines arguments originating from benedicks and carleson @xcite with the introduction of induced maps , as discussed below .
let @xmath19 be a family of unimodal maps of the interval @xmath20 $ ] , with critical point at @xmath6 .
suppose that each @xmath8 is at least @xmath21 smooth and that @xmath22 , @xmath23 , and @xmath24 , are @xmath25 w.r.t . @xmath26 .
suppose that each @xmath8 has negative schwarzian derivative ( see @xcite ) and that @xmath27 .
further , suppose that @xmath28 and that the fixed point at 0 is hyperbolic repelling .
we say that @xmath19 _ unfolds a ( quadratic ) saddle - node _ if , * there is a @xmath29-periodic point @xmath30 , with @xmath31 , and , * @xmath32 at @xmath33 . for the sake of clarity
, we will assume that @xmath34 with this convention , for @xmath35 the saddle - node point disappears and complicated dynamics may occur . for a set @xmath36 of parameter values , let @xmath37 be its lebesgue measure .
a unimodal map @xmath38 is called _ renormalizable _ if there is a proper subinterval @xmath39 $ ] containing the critical point @xmath6 , so that @xmath40 for some @xmath41 and @xmath42 for @xmath43 .
a renormalizable map is called _ once renormalizable _ if the above property defines @xmath44 uniquely . [ existence ] let @xmath45 be as above , unfolding a saddle - node bifurcation at @xmath46 .
assume that @xmath5 is once renormalizable .
consider the set @xmath47 such that for each @xmath48 , the map @xmath8 has an absolutely continuous invariant measure @xmath49 whose support is the maximal interval @xmath13 $ ] .
the set @xmath47 has positive density at @xmath46 : @xmath50 it does not have full density : @xmath51 careful numerical studies of the quadratic family , @xmath52 , predict that between the period doubling limit and @xmath53 , excluding the period three window , less then about 15% of the parameter values correspond to the periodic windows @xcite .
numerical simulations also suggest that near a saddle - node bifurcation , there is a large set of parameter values for which the a.c.i.m.s are supported on the maximal possible interval .
this is illustrated in figure [ fig : quad ] .
existence of the parameter set @xmath47 follows from a similar approach as used by benedicks and carleson in their treatment of jakobson s result .
below we comment in more detail on the construction of the parameter set @xmath47 .
we now first consider intermittency that results from the saddle - node bifurcation and introduce our second main result .
intermittent dynamics manifests itself by alternating phases with different characteristics . in one phase ,
referred to as the laminar phase , the dynamics appear to be nearly periodic . while in the other phase , the relaminarization phase , the orbit makes large , seemingly chaotic excursions away from the periodic region .
these excursions are called chaotic bursts .
let @xmath54 be a neighborhood of the orbit , @xmath55 of @xmath30 , not containing a critical point of @xmath56 .
let @xmath57 be defined as @xmath58 whenever the limit exists , where @xmath59 is the usual indicator function of the set @xmath60 .
that is , @xmath61 is the relative frequency with which the orbit @xmath62 visits @xmath54 ( for those @xmath63 for which the limit exists ) .
the following theorem discusses @xmath61 for @xmath64 near 0 .
denote by @xmath65 the atomic measure supported on the orbit of @xmath30 , given by @xmath66 we will denote the usual weak convergence of measures by the symbol @xmath67 .
[ intermittency ] let @xmath45 and @xmath47 be as in theorem [ existence ] . there exist sets @xmath68 of parameter values with positive density at @xmath7 , so that @xmath69 restricting to @xmath70
, @xmath71 is a constant , @xmath72 , almost everywhere on @xmath20 $ ] and @xmath72 depends continuously on @xmath64 at @xmath11 .
there exist @xmath73 so that @xmath74 each set @xmath75 from the above theorem is constructed as a union @xmath76 of sets @xmath77 , where each @xmath77 has , as a lebesgue density point , a parameter value @xmath78 for which @xmath79 hits a hyperbolic repelling periodic point @xmath80 , for some @xmath81 ( the map @xmath82 is a misiurewicz map ) .
different sets @xmath75 result from considering points @xmath80 in different periodic orbits .
for the construction of @xmath77 , we adapt benedicks and carleson s treatment of jakobson s result ( we follow in fact the approach taken by luzzatto @xcite ) and construct parameter values for which @xmath83 has a positive lyapunov exponent .
the main obstacle in the present case is the fact that when orbits fall into the saddle - node region they remain there for many iterations and the derivative of @xmath84 in that region is close to one .
we get around this problem by defining an induced map which skips over the saddle - node region .
we show that the induced map has uniform exponential growth of derivatives and that this implies exponential growth of derivatives ( with weaker constants ) for the original map .
careful estimates bound the measure of the sets @xmath77 for increasing values of @xmath85 , thus bounding the measure of their union @xmath75 . for comparison
we include a result from @xcite , which builds on results in @xcite , @xcite , and @xcite , showing that also periodic attracting orbits are found for parameters from a set with positive density at the bifurcation point .
the frequency with which the dynamics is in the laminar phase behaves in a similar way to the frequency in theorem [ intermittency ] .
note that the fact that the set @xmath47 in theorem [ existence ] does not have full density at @xmath46 , follows from the following result .
[ saddlenode ] let @xmath45 be as above , unfolding a saddle node bifurcation at @xmath46 . there exist sets @xmath86 of parameter values of positive measure and positive density at 0 , i.e. @xmath87 so that for each @xmath88 , @xmath8 has an attracting periodic orbit .
further , @xmath89 where @xmath90 , @xmath88 , is the invariant measure supported on the periodic orbit . restricting to @xmath91 , @xmath71 is a constant , @xmath72 , almost everywhere on @xmath20 $ ] and @xmath72 depends continuously on @xmath64 at @xmath11 .
there exists @xmath92 so that @xmath93 different sets @xmath86 lead to different limit values @xmath94 in theorem [ saddlenode ] . in fact , the proof of theorem [ saddlenode ] makes clear that arbitrary large numbers occur as the limit values @xmath94 .
this fact , together with theorems [ existence ] and [ saddlenode ] lead us to conjecture that there is a parameter set @xmath95 which has @xmath46 as a lebesgue density point , so that @xmath96 it was shown in @xcite that such a limit can not hold without restricting the parameter set .
denote by @xmath97 a small neighborhood of @xmath30 on which @xmath56 is invertible .
let @xmath98 and @xmath99 denote the usual local stable and local unstable sets for @xmath30 .
[ embedding ] let @xmath45 be a @xmath25 family of @xmath100 , @xmath101 , maps unfolding a saddle - node .
then there exists a family of @xmath100 flows , @xmath102 , on @xmath97 such that @xmath103 for each @xmath104 .
further , @xmath105 in the @xmath25 topology on @xmath97 and in the @xmath100 topology on compact intervals away from the fixed point . the flow @xmath106 is uniquely determined by @xmath5 . proof .
the @xmath107 version of this theorem is due to takens @xcite .
the @xmath100 result follows from part 2 of @xcite .
the case @xmath46 follows from appendix 3 of @xcite .
the case @xmath108 and the convergences as @xmath109 follow from theorem iv.2.5 and lemma iv.2.7 of the same .
@xmath110 [ il ] this result
is known as the takens embedding theorem . a version of it appears in @xcite .
they proved that one may obtain @xmath111 which depends @xmath100 smoothly on both @xmath63 and @xmath26 , even at the fixed point , if one requires that @xmath112 be @xmath113 smooth , where @xmath114 may be larger than @xmath115 .
proposition [ embedding ] allows for our weaker hypotheses and its implications are sufficient for our purposes .
choose @xmath116 and let @xmath117.\ ] ] also , choose a point @xmath118 so that @xmath119 \subset e.\ ] ] for the sake of convenience we restrict @xmath97 to be the interval @xmath120,\ ] ] so that @xmath121 and @xmath122 $ ] .
we will use the embedding flow on the interval @xmath123 $ ] . given @xmath104 and @xmath124 ,
define @xmath125 to be the unique number for which @xmath126 for @xmath104 and @xmath127 , let @xmath128 be defined by @xmath129 it follows from the smoothness of @xmath111 that for each @xmath104 , the functions @xmath130 are @xmath100 diffeomorphisms from @xmath131 to [ 0,1 ] .
we will use @xmath130 as coordinates on @xmath131 . in the following , we will associate @xmath20 $ ] with the unit circle @xmath132 by making the identification @xmath133 .
we also treat @xmath131 as circles through the coordinates @xmath134 . given @xmath135 and @xmath136 as above ,
let @xmath137 be the sequence , @xmath138 , defined by : @xmath139 for each @xmath140 let @xmath141 \rightarrow [ { \gamma}_{l+1},{\gamma}_l]$ ] , be the reparameterization map defined by @xmath142 we have that @xmath143 and @xmath144 .
we may invert @xmath145 , for each @xmath85 , to obtain maps @xmath146 \rightarrow [ 0,1]$ ] .
[ distortion ] the reparameterization maps @xmath147 are smooth monotone decreasing functions with uniformly small distortion : given @xmath148 there is @xmath149 so that for every @xmath150 and every @xmath151 $ ] , @xmath152 proof .
diaz et .
al . proved this result under the hypothesis that @xmath112 is @xmath113 , using ilyashenko and li s embedding result ( remark [ il ] ) .
a proof of this result under the current hypotheses appears in @xcite based on @xcite .
@xmath110 + we remark that @xmath153 converges as @xmath154 ( see @xcite ) , so that @xmath155 as @xmath156 .
this fact , together with proposition [ distortion ] imply the next proposition @xcite .
[ density ] let @xmath157 be a measurable subset of @xmath158 and denote @xmath159 $ ] . if the limit @xmath160 exists and equals @xmath161 , then @xmath162 we remark that , with @xmath77 satisfying the assumption in the above proposition , @xmath163 converges as @xmath164
let @xmath165 denote the local ( first hit ) map from @xmath166 to @xmath167 induced by @xmath168 .
the convenience of using @xmath169 and @xmath170 as coordinates on @xmath171 and @xmath172 is seen in the following proposition .
[ localmap ] for each @xmath140 and each @xmath173 $ ] @xmath174 where @xmath175 denotes a rigid rotation by angle @xmath176 . proof .
this follows from proposition [ embedding ] and the definitions of @xmath177 and @xmath178 as the time variables for the embedding flow for @xmath179 .
@xmath110 + in addition to @xmath180 $ ] defined above , it is also convenient to define an extension , @xmath181 , by @xmath182 thus @xmath183 and @xmath184 \rightarrow [ 0,\infty)$ ] for @xmath108 .
furthermore , we will denote @xmath185 wherever @xmath181 is defined .
similarly , @xmath186 is given by @xmath187 let @xmath188 , @xmath189 .
it follows from the assumptions that there is an integer @xmath190 such that @xmath191 . by choosing @xmath135 so that @xmath192 is in the interior of @xmath193 we have that @xmath194 for all @xmath26 small .
denote this point by @xmath195 .
it then follows that for any @xmath85 , either @xmath196 denote this intersection of @xmath197 with @xmath172 by @xmath198 .
note that for a fixed @xmath85 the function @xmath199 will have a jump discontinuity at which the value will jump from one endpoint of @xmath200 to the other .
denote by @xmath201 the point at which the discontinuity takes place .
there exists a limit @xmath202 as @xmath154 the sequence of maps @xmath199 converges in the @xmath100 topology on compact sets not containing @xmath203 . proof .
this easily follows from proposition [ embedding ] and the definition of @xmath204 .
@xmath110 + specifically , we will make use of the implication that the derivatives of @xmath205 with respect to @xmath176 converge uniformly as @xmath154 for @xmath176 in compact intervals away from the discontinuity . later
, this will allow us to make estimates of derivatives along @xmath197 which are uniform in @xmath85 .
in this section , we apply the local analysis near the saddle - node from the previous section to obtain expressions for global return maps
. we study the occurrence of misiurewicz maps for small parameter values .
we begin with a simple but , useful lemma .
[ lem_over ] @xmath206 $ ] . proof .
let @xmath207 be the interval bounded by @xmath30 such that @xmath208 .
consider @xmath209 note that @xmath210 is an invariant set .
we have @xmath211 .
if @xmath210 does not cover @xmath212 $ ] , then for some @xmath213 , @xmath214 is disjoint from @xmath215 .
this contradicts the assumption that @xmath5 is once renormalizable .
@xmath110 + we will use the freedom in the choice of @xmath136 and take @xmath136 so that @xmath6 is not in @xmath216 . by this choice
we have that @xmath6 will be in the interior of @xmath217 for some @xmath218 . for @xmath46 ,
let @xmath219 be the first hit map from @xmath220 to @xmath221 .
define @xmath222 \rightarrow { \mbox{\bb r}}$ ] by @xmath223 and define @xmath224 \rightarrow [ 0,1]$ ] by @xmath225 . by identifying the endpoints of @xmath20 $ ]
we may consider @xmath226 as a map from a subset @xmath227 of @xmath132 into @xmath228 and @xmath229 as a map from @xmath230 into @xmath132 .
following @xcite , we call @xmath226 the _ mather invariant _ for @xmath5
. one may show that @xmath229 is a modulus of smooth conjugation , in other words , it is invariant under differentiable changes of variables . for each @xmath231 in the domain , @xmath232 , of @xmath229 and @xmath233 such that @xmath234 associate the integer @xmath235 some @xmath233 will return ( perhaps many times ) to @xmath236 $ ] before eventually being trapped in @xmath237 and the function @xmath238 \rightarrow { \mbox{\bb z}}^-$ ] measures the depth of the deepest return .
finally , to each @xmath231 in the domain of @xmath229 and the corresponding point @xmath233 associate a positive integer @xmath239 given @xmath190 , denote by @xmath240 the subset of @xmath232 defined by @xmath241 the points in @xmath242 are those whose forward orbits enter @xmath98 in a bounded number iterations and which do not come too close to @xmath30 in the process , either by re - entering @xmath99 too close to a or by landing in @xmath98 too close to @xmath30 .
since the forward orbit of almost every @xmath243 has @xmath55 as its omega limit set , the measure @xmath244 may be made close to @xmath245 by taking @xmath190 to be large .
we will let @xmath246 , so that @xmath54 is a neighborhood of the orbit of @xmath30 for @xmath46 .
denote by @xmath247 and @xmath248 note that orbits associated with @xmath249 do not reenter @xmath250 before they hit @xmath251 .
also note that for @xmath218 fixed , the intervals @xmath252 approach @xmath250 as @xmath253 .
consider the first return map , @xmath254 , of the interval @xmath255 and let @xmath256 be the normalized map given by @xmath257 identifying the endpoints of @xmath20 $ ] , we may consider @xmath256 as a map on the circle .
[ approx ] given any @xmath218 , @xmath258 for each @xmath259 . proof .
denote @xmath260 and @xmath261 .
let @xmath262 denote the global ( first hit ) map from @xmath263 to @xmath264induced by @xmath265 .
let @xmath266 denote the local map .
note that @xmath254 is not equal to @xmath267 since some points in @xmath263 will return to @xmath263 before hitting @xmath264 ( by landing in @xmath268 ) .
however , the two maps do agree when restricted to @xmath269 since those points will in fact hit @xmath264 before returning to @xmath263 .
since we are only considering a finite number of iterations from @xmath263 to @xmath264 , it follows from the construction of @xmath229 that @xmath270 converges to @xmath229 on the restricted set @xmath249 .
proposition [ localmap ] then implies that , @xmath271 converges to @xmath272 on @xmath249 .
@xmath110 in this section we will identify parameter values @xmath176 for which @xmath168 maps the critical point @xmath6 onto some repelling hyperbolic periodic point and thus @xmath6 does not return to a neighborhood of itself , i.e. @xmath168 satisfies the _
misiurewicz condition_. there are two ways that hyperbolic periodic points can occur : as periodic points whose orbits pass through @xmath54 , or as continuations of periodic points for @xmath7 ( outside of @xmath273 . in the former case ,
the periodic orbits exist for parameter values within subintervals of @xmath274 $ ] for each @xmath85 . in the latter case
the periodic orbits exist for all @xmath108 sufficiently small .
we confine ourselves to a discussion of this case .
we show that parameters for which @xmath6 is mapped onto the periodic orbits under consideration occur in decreasing sequences @xmath275 of values approaching @xmath11 as @xmath154 .
later we will show that such a sequence typically gives rise to the parameter sets predicted in theorem [ existence ] .
note that the assumptions on @xmath5 imply that the nonwandering set , @xmath276 , of @xmath5 restricted to @xmath20 \setminus \bar{e}$ ] is a hyperbolic set ( see theorem iii.5.1 in @xcite ) .
if a point @xmath233 is mapped onto @xmath276 then @xmath229 will have a discontinuity at @xmath277 . by lemma [ lem_over ] , for any @xmath278
there is a point @xmath279 which is mapped onto @xmath280 by @xmath5 .
recall that @xmath281 is the first point in the orbit of @xmath6 that hits @xmath171 .
let @xmath282 .
[ mis2 ] suppose that @xmath283 is a periodic point contained in the nonwandering set @xmath276 of @xmath20 \setminus \bar{e}$ ] .
denote by @xmath284 the continuation of @xmath283 for @xmath104 .
let @xmath285 be a point of discontinuity of @xmath229 corresponding to @xmath283 and suppose that @xmath286 is such that @xmath287 is mapped onto @xmath285 by @xmath288 for some @xmath289 .
suppose further that @xmath290 then there is a sequence of parameter values @xmath275 , @xmath291 $ ] , such that @xmath6 is mapped onto @xmath80 by iterates of @xmath82 .
the point @xmath292 is in @xmath293 for all small @xmath26 , and , @xmath294 approaches @xmath287 as @xmath109 .
thus , @xmath295 .
let @xmath296 be the point in @xmath220 which is mapped onto @xmath283 by an iterate of @xmath5 .
if @xmath297 , then @xmath298 approaches @xmath296 as @xmath154 and ( [ partialmb ] ) implies the result .
if @xmath299 then it follows there is an @xmath81 such that @xmath300 for all @xmath301 .
applying proposition [ approx ] to @xmath302 and @xmath303 then implies that @xmath304 is mapped by the @xmath305-return map into @xmath306 arbitrarily close the @xmath307-th preimage of @xmath296 for @xmath85 large .
the result then follows from ( [ partialmb ] ) .
@xmath110 + the following corollary provides the parameter values which will be used in the proof of theorem [ existence ] .
[ oml ] given any @xmath278 , there exists integers @xmath308 and @xmath309 and a sequence of parameter values @xmath291 $ ] , @xmath310 such that @xmath311 further , there is a number @xmath312 such that @xmath313 as @xmath154 . proof . for each @xmath278
there is a point @xmath279 such that @xmath296 is mapped to @xmath280 by an iterate of @xmath5 .
let @xmath314 .
if we let @xmath315 then @xmath316 and @xmath317 then lemma [ mis2 ] implies the result .
we start the proof of theorem [ existence ] . in this section
we construct a set @xmath75 with positive density at @xmath46 , so that @xmath8 for @xmath318 has bounded recurrence ( see definition [ def_br ] ; the definition is in terms of an induced map which is defined below ) .
we deduce that @xmath8 for @xmath318 has an absolutely continuous invariant measure .
considering @xmath8 for @xmath319 $ ] , we reparameterize and use the parameter @xmath173 $ ] defined by @xmath320 , see section [ sec_snlocal ] .
denote the resulting family of unimodal maps by @xmath321 . by proposition [ oml ] ,
there exists a converging sequence of parameter values @xmath322 so that @xmath323 is a hyperbolic periodic point .
the map @xmath324 is a misiurewicz map and by an extension of jakobson s theorem , see theorem v.6.1 in @xcite , @xmath325 is a lebesgue density point of a set of parameters @xmath326 for which @xmath321 supports an absolutely continuous invariant measure .
theorem [ existence ] can be proved by establishing that the measure of @xmath326 is bounded away from 0 uniformly in @xmath85 .
proposition [ density ] guarantees that the union @xmath327 has positive lebesgue measure and positive density at @xmath46 .
combined with the statement that @xmath321 , @xmath328 , admits an absolutely continuous invariant measure , this proves theorem [ existence ] . in the following section we define induced maps .
the actual construction of the sets @xmath326 is described in section [ sec_indconstructions ] .
as before , @xmath30 denotes a saddle - node periodic point of @xmath5 , of period @xmath29 .
we may assume that @xmath30 is nearest to the critical point @xmath6 of all points in @xmath329 , so that @xmath56 is a homeomorphism on @xmath330 .
we may suppose that there is a periodic point @xmath331 and an integer @xmath190 such that @xmath332 .
let @xmath333 be the periodic continuation of @xmath334 and @xmath335 the continuation of @xmath336 such that @xmath337 .
we will suppress the @xmath26 dependence of @xmath136 in the notation .
let @xmath338 be the fundamental intervals given by @xmath339 .
denote @xmath340 note that @xmath341 . before proceeding with the definition and construction of @xmath326
, we introduce an induced map @xmath342 .
namely , @xmath342 will be defined from @xmath321 in the following way , @xmath342 will equal @xmath321 , except for points in @xmath343
. points which fall in this set will be mapped ahead to the image of @xmath344 .
that is , @xmath345 thus points in @xmath346 are mapped by the first hit map to the interval @xmath347 . for @xmath348
, @xmath342 will have @xmath349 discontinuities .
recall that for @xmath108 and @xmath350 the function @xmath351 is defined by @xmath352 where @xmath353 observe that @xmath342 equals @xmath354 write @xmath355 let @xmath356 denote the map defined from @xmath5 in the following way .
if @xmath357 then @xmath358 . if @xmath359 then @xmath360 is given be the first hit map from @xmath361 to @xmath362 induced by @xmath5 . if @xmath363 then let @xmath364 be defined as above and let @xmath365 in other words , if @xmath366 then @xmath63 is mapped to @xmath362 using the time coordinate shifted by @xmath367
. note that @xmath360 will have countably many discontinuities accumulating at @xmath30 from both sides . in @xmath361
these occur at @xmath368 . in @xmath369
these occur at those @xmath63 for which @xmath370 [ approx2 ] for each @xmath176 the sequence of maps @xmath371 converges to the map @xmath356 , uniformly on compact sets which are disjoint from the discontinuities of @xmath356
. proof . given that @xmath342 satisfies
( [ tildef ] ) and @xmath356 is defined by ( [ tildefinfty ] ) , this follows from proposition [ localmap ] .
@xmath110 + observe that , with @xmath286 and @xmath81 as in proposition [ oml ] , @xmath372 . in the following we will consider iterates of intervals .
the map @xmath373 , @xmath374 , is discontinuous along backward iterates of @xmath136 in @xmath97 , so that the intervals in the image of an interval under @xmath373 might be arbitrarily small , regardless of the size of the original interval .
we therefore slightly adjust the definition of @xmath373 to avoid this problem .
consider an interval @xmath39 $ ] .
for @xmath375 , denote @xmath376 and let @xmath377 be the component of @xmath378 adjacent to @xmath379 and @xmath380 the component of @xmath378 adjacent to @xmath381 .
this yields a partition @xmath382 of @xmath383 ( elements of the partition can be empty ) .
if the leftmost or rightmost nonempty intervals of this partition do not contain a fundamental domain @xmath338 , @xmath384 , join them to the adjacent intervals .
note that if @xmath383 is partitioned into two elements @xmath385 neither of which is a fundamental domain , this leaves a choice in coding the resulting interval after @xmath386 or @xmath387 . this way an interval @xmath383 that covers one or more fundamental domains is partitioned into subintervals which are at least as large as a fundamental domain .
given @xmath388 , define @xmath389 note that as long as @xmath383 does not cover a fundamental domain in @xmath97 , @xmath390 equals some fixed iterate of @xmath321 .
also note that @xmath391 consists of at most two components .
we further remark that , if @xmath392 maps @xmath383 homeomorphically onto @xmath381 , then there is a fixed number @xmath207 of iterates after which @xmath393 contains @xmath6 in its interior .
define maps @xmath394 and @xmath395 by @xmath396 we would like to consider images @xmath397 .
as for single maps we come across the difficulty that @xmath342 is discontinuous along the backward orbit of @xmath136 in @xmath97 .
we consider the set @xmath398 . by a fundamental strip we mean a set @xmath399 , @xmath400 .
consider a curve @xmath401 that projects injectively to @xmath20 $ ] by the projection @xmath402 .
define @xmath403 to be the intersection of @xmath404 with the fundamental strip @xmath399 , @xmath400 , and let @xmath405 be the connected component of @xmath404 that is adjacent to @xmath406 and @xmath407 the connected component adjacent to @xmath408 .
this defines a partition @xmath409 of @xmath404 with possibly empty elements .
if the leftmost or rightmost nonempty element of this partition does not cross a fundamental strip , join it to the adjacent element . this way a partition of a curve @xmath404 that crosses at least one fundamental strip
is obtained all of whose elements cross a fundamental strip .
define @xmath410 in this section the actual construction of the set @xmath326 is described .
we make use of the induced maps introduced in section [ sec_ind ] and for the rest we closely follow @xcite .
consider @xmath176 near @xmath325 for a fixed value of @xmath85 .
write @xmath411 [ def_br ] for @xmath412 , we say that @xmath342 satisfies the bounded recurrence condition @xmath413 if for all positive integers @xmath414 , @xmath415 we say that @xmath342 satisfies @xmath416 if it satisfies @xmath417 for all @xmath44 . similarly , we say that @xmath321 satisfies @xmath417 if it satisfies the equivalent condition on @xmath418 .
it is easy to see that if @xmath342 satisfies @xmath417 then so does @xmath419 , and vice versa .
the next proposition is the main result of section [ sec_br ] and will be shown to imply theorem [ existence ] .
[ jakobson ] for every @xmath420 there exists @xmath421 , so that there is a set @xmath326 of parameter values for which @xmath342 , @xmath328 , satisfies @xmath416 and @xmath422 for some @xmath423 , uniformly in @xmath85 . by proposition [ density ] , @xmath327 has positive density at @xmath46 .
theorem [ existence ] is shown by establishing that @xmath321 possesses an absolutely continuous invariant measure for @xmath328 .
this is done more formally in section [ sec_proof ] , but the idea is straightforward : one shows that @xmath342 , @xmath328 , is a collet - eckmann map ( meaning that @xmath424 for some @xmath92 , @xmath425 ) .
this implies that also @xmath321 , @xmath328 , is a collet - eckmann map , albeit with weaker expansion .
collet - eckmann maps are known to admit absolutely continuous invariant measures . to prove proposition [ jakobson ] , we adapt the reasoning in @xcite , where jakobson s result is proven using a variant of benedicks - carleson s proof .
below we will describe the constructions and show key estimates and computations that differ from those in @xcite .
after some preliminary work , the reasoning will follow @xcite closely
. we will therefore be able to simply refer to @xcite for several of the more technical parts of the proof .
note that we must consider a sequence of families @xmath342 that tends to the family @xmath356 as @xmath164 , see proposition [ approx2 ] .
it is instructive to compare with @xcite where the analysis leads to the study of a sequence of families @xmath426 of smooth unimodal maps ( renormalizations of @xmath38 ) , converging to the logistic family as @xmath164 .
as they remarked , there are uniform lower bounds on the measure of the set of parameter values for which absolutely continuous invariant measures occur , for families from a neighborhood of the logistic family ( and so for all large enough values of @xmath85 ) .
a similar situation was considered in @xcite .
following the reasoning below one can construct a set of parameter values for which @xmath356 has bounded recurrence and show that this set has positive measure .
in fact , the reasoning below constructs such sets , and uniformly bounds their measures , for families nearby @xmath356 , in particular for the families @xmath342 with @xmath85 large .
we concentrate on the families @xmath342 and formulate the results in terms of this sequence of families .
we now start the constructions . for a positive integer @xmath115 , let @xmath427 and @xmath428 $ ] .
let @xmath429 be a small positive number .
given @xmath421 , write @xmath430 and @xmath431 .
we can suppose that @xmath432 and @xmath433 are integers .
let @xmath434 subdividing each interval @xmath435 into @xmath436 subintervals , @xmath437 , of equal length provides partitions @xmath438 of @xmath439 and @xmath440 of @xmath441 .
given @xmath442 , write @xmath443 ( or @xmath444 ) and let @xmath445 . define the _ binding period _ of @xmath63 as @xmath446 .
suppose @xmath447 and define the binding period associated with @xmath448 as @xmath449 for a fixed @xmath85 , let @xmath450 be a subinterval of @xmath451 .
denote @xmath452 where @xmath453 is the projection @xmath454 .
if @xmath455 intersects @xmath439 , @xmath456 , then @xmath305 is called a _ return time _ for @xmath450 .
define the binding period associated with a return time @xmath305 of a parameter interval @xmath450 as @xmath457 for each @xmath85 let @xmath458 be the trivial partition @xmath459 of the parameter interval @xmath460 .
inductively we will define parameter sets @xmath461 and partitions @xmath462 thereof . in order to define @xmath462 given @xmath463
, we first construct a refinement @xmath464 of @xmath463 .
we say that a return of @xmath450 at time @xmath305 is a _ bound return _ if there is a return time @xmath465 of @xmath450 and @xmath466 .
let @xmath467 .
* chopping times . *
: : we say that @xmath44 is a chopping time for @xmath450 if + 1 .
@xmath468 intersects @xmath441 in at least three elements of @xmath440 , and 2 .
@xmath468 is not a bound return for @xmath450 .
* non - chopping times . *
: : we say that @xmath44 is a non - chopping time for @xmath450 in all other cases , that is if one or more of the following occurs : + 1 .
@xmath469 . 2 .
@xmath468 is a bound return of @xmath450 .
@xmath468 intersects no more than two elements of the partition @xmath440 of @xmath441 . in case @xmath44
is a non - chopping time for @xmath450 , we let @xmath470 .
if @xmath44 is a chopping time for @xmath450 we partition @xmath450 as follows .
write @xmath471 , so that each @xmath472 fully contains one and at most one element of @xmath440 . if @xmath473 contains an interval of length less then @xmath474
, we include this interval in the adjacent interval of @xmath440 .
otherwise an interval of @xmath473 is an element of the partition of @xmath468 .
write the resulting partition of @xmath468 as @xmath475 .
there is a corresponding partition @xmath476 of @xmath450 , given by @xmath477 .
let each element of this partition be an element of @xmath478 .
note that an element of @xmath478 partitioning @xmath450 need not be connected , but can consist of several intervals .
this may happen if @xmath479 intersects @xmath343 in at least two fundamental domains , for some @xmath480 .
let @xmath467 and consider @xmath481 .
we speak of a bound , essential or inessential return time or an escape time for @xmath482 in the following situations .
+ _ bound return time : _ the interval @xmath483 intersects @xmath439 and @xmath44 is a bound return time for @xmath450 .
+ _ inessential return time : _ the interval @xmath483 intersects @xmath439 and @xmath44 is a non - chopping time for @xmath450 that is not a bound return time , but @xmath468 intersects at most two elements of the partition @xmath440 .
+ _ essential return time : _ the interval @xmath483 intersects @xmath439 and @xmath44 is a chopping time for @xmath450 . + _ escape time : _ the return time @xmath44 is a chopping time for @xmath450 , but @xmath483 does not intersect @xmath439 . in this case
we call @xmath482 an escape component of @xmath450 . + any interval @xmath484 belongs to a unique nested sequence of intervals @xmath485 where @xmath486 for @xmath487 . if @xmath190 is a chopping time for @xmath488 , then @xmath489 is strictly contained in @xmath488
. chopping times are either escape times or essential return times .
the return depth of @xmath482 at time @xmath305 is defined if @xmath490 intersects @xmath439 , as @xmath491 define functions @xmath492 and @xmath493 which associate to @xmath494 the sum of the return depths and the sum of the essential return depths , over the first @xmath44 iterates @xmath495 .
define @xmath496 and @xmath497 the sets @xmath498 will be shown to satisfy the stated properties in proposition [ jakobson ] . here
expansion properties of the maps @xmath321 , @xmath342 and @xmath499 are discussed .
we relate expansion properties of these maps . in section [ sec_mane ]
we prove a ma type result for @xmath342 , that is , we show that there is expansion along orbits outside a neighborhood of @xmath6 .
the relation between expansion along orbits of @xmath321 , @xmath342 and @xmath499 is discussed in the next two lemma s .
[ lemma_withouttilde ] if there exist @xmath500 , @xmath501 such that @xmath502 for all @xmath503 , then there are @xmath504 and @xmath505 , so that @xmath506 , for all @xmath41 .
write @xmath507 with @xmath508 minimal such nonnegative integer .
compute @xmath509 now @xmath510 implies that the piece of orbit @xmath511 is in @xmath54 . since @xmath512 , the term @xmath513 is bounded below in @xmath54
further , @xmath508 is bounded above by @xmath514 since any point in @xmath54 is mapped outside of @xmath54 in @xmath514 or fewer iterations .
therefore the quantity @xmath515 is bounded from below by a constant @xmath516 .
thus , @xmath517 . we can let @xmath518 .
since there is a minimum number of iterations of @xmath321 needed for an orbit to enter @xmath54 after leaving @xmath519 , and the number of consecutive iterations in @xmath54 is bounded above by @xmath520 , it follows that the fraction @xmath521 is bounded below by a constant @xmath522 .
hence , @xmath523 is strictly larger than some number @xmath524 .
@xmath110 + similarly one derives the following lemma relating expansion of @xmath499 to expansion of @xmath342 .
[ breve - tilde ] if @xmath525 for some @xmath526 , @xmath527 , then there are constants @xmath528 and @xmath501 so that @xmath529 .
the converse statement holds as well .
given a subinterval @xmath530 $ ] , write @xmath531 . then @xmath532 is bounded and bounded away from 0 .
in fact , if @xmath343 is small , the quotient @xmath533 will be close to 1 for large values of @xmath190 , since there is then a large number of iterates between any two passages through @xmath343 .
observe that for @xmath343 small , @xmath534 in lemma [ breve - tilde ] is close to @xmath535 .
write @xmath536 and @xmath537 .
let @xmath538 .
then the bounded recurrence condition @xmath539 where @xmath540 , translates into @xmath541 for a constant @xmath404 which bounds @xmath542 from below .
if @xmath343 is small and @xmath190 is large , then @xmath404 is close to 1 .
hence , when translating the bounded recurrence condition for @xmath342 into one for @xmath499 , an almost identical condition is obtained .
a similar remark can be made for @xmath395 and @xmath543 .
the following proposition is central and provides exponential expansion of iterates of @xmath342 that stay away from the critical point @xmath6 .
the estimates are uniform in @xmath544 .
the proposition is modeled after theorem iii.6.4 in @xcite , which treats families of smooth unimodal maps .
[ prop_exp ] there are constants @xmath545 and @xmath501 , and a neighborhood @xmath546 of @xmath6 so that for any neighborhood @xmath547 of @xmath6 with @xmath548 and each large enough integer @xmath85 , the following holds . for each @xmath549 , if @xmath550 for @xmath551 and @xmath552 , then @xmath553 if @xmath550 for @xmath551 , but not necessarily @xmath554 , then @xmath555 moreover , if @xmath556 for @xmath557 , then @xmath553 we will make use of koebe s principle , which we quote here .
see @xcite for the proof and additional information .
let @xmath558 be two intervals .
we say that @xmath559 contains a @xmath560-scaled neighborhood of @xmath547 if both components of @xmath561 have at least length @xmath562 .
the distortion of a diffeomorphism @xmath563 on an interval @xmath383 is defined as @xmath564 * [ koebe principle]*[koebe ] let @xmath38 have negative schwarzian derivative and let @xmath565 be a neighborhood of @xmath38 in @xmath21 . then for each @xmath421 , there exists @xmath566 so that @xmath567 , if @xmath568 are intervals , @xmath569 is a diffeomorphism and @xmath570 contains a @xmath560-scaled neighborhood of @xmath571 then the distortion of @xmath572 is bounded by @xmath94 .
a useful property of maps with negative schwarzian derivative is the following principle , see @xcite . *
[ minimum principle]*[minimum ] let @xmath38 be a map with negative schwarzian derivative on a closed interval @xmath573 $ ] .
if @xmath574 does not vanish on @xmath383 , then @xmath575 .
the following lemma is similar to theorem iii.6.2 in @xcite .
so is its proof .
[ lemma_size ] there are constants @xmath92 , @xmath576 , so that for all large enough @xmath85 the following holds .
let @xmath577 be a maximal interval with @xmath578 a homeomorphism .
then @xmath579 proof .
denote @xmath580 .
the number of elements in @xmath581 is fixed .
there is also a minimum distance between any two points in @xmath581 , uniformly in @xmath85 .
thus we may let @xmath546 be a neighborhood of @xmath6 such that @xmath582 for all @xmath85 .
similarly , @xmath583 is a finite set .
let @xmath584 be a maximal interval on which @xmath585 is a homeomorphism , but @xmath586 not .
let @xmath587 be the subintervals of @xmath584 on which @xmath586 is a homeomorphism .
the boundary points of @xmath588 are contained in @xmath589 . since @xmath581 and @xmath590 finite , all intervals @xmath591 have lengths which are bounded below uniformly in @xmath85 . applying the koebe principle ( to @xmath592 )
one checks that there is a constant @xmath593 with @xmath594 further , since @xmath581 and @xmath590 are each finite it is clear that @xmath595 is not a homeomorphism for some uniformly bounded @xmath305 .
the result follows and it is clear that the constants can be chosen uniformly in @xmath85 .
@xmath110 + the next proposition discusses expansion properties of @xmath592 , for @xmath596 .
[ prop_ga ] for any small enough neighborhood @xmath546 of @xmath6 , there are constants @xmath500 and @xmath501 , so that the following holds for all @xmath85 sufficiently large . if @xmath597 for @xmath551 , then @xmath598 if @xmath599 , then @xmath598 without any condition , @xmath600 proof .
let @xmath546 be a neighborhood of @xmath6 , small enough so that @xmath601 for @xmath602 .
we claim that there are @xmath500 , @xmath501 so that for all large enough integers @xmath85 , if @xmath603 for @xmath604 , then @xmath605 .
it suffices to show that there exists @xmath210 with @xmath606 ( compare the proof of theorem iii.3.3 in @xcite ) .
assume there exist points @xmath607 $ ] with @xmath608 , @xmath551 , and @xmath609 .
since @xmath592 has negative schwarzian derivative we have that on one side of @xmath610 , @xmath611 .
let @xmath612 be the maximal interval bounded by @xmath610 so that @xmath613 is a homeomorphism on @xmath612 and @xmath614 on @xmath612 .
this implies that @xmath615 .
let @xmath616 be the other boundary point of @xmath612 . then either @xmath617 or @xmath618 , for some @xmath619 . in both cases
, @xmath620 is bounded away from @xmath11 . by lemma [ lemma_size ] ,
@xmath621 is small as @xmath81 is big .
since @xmath615 , also @xmath622 is small for large @xmath81 .
the interval @xmath623 has length bounded away from @xmath11 , say @xmath624 .
it follows that @xmath625 tends to @xmath626 as @xmath627 .
in fact , if @xmath628 denotes the minimum of @xmath629 over @xmath20 \backslash w$ ] , then @xmath630 lemma [ lemma_size ] yields that @xmath631 , so that @xmath632 now ( [ largerbound ] ) and ( [ smallerbound ] ) contradict each other , proving the claim .
observe that @xmath210 and hence the constants @xmath633 and @xmath534 can be chosen uniformly in @xmath85 .
this proves the first estimate .
next , suppose @xmath634 .
let @xmath612 be the maximal interval containing @xmath63 such that @xmath635 is a homeomorphism on @xmath612 .
because the orbit of @xmath6 is finite , the interval @xmath636 extends a positive distance away from @xmath637 to both sides .
koebe s principle implies @xmath638 for some @xmath639 which gives the second estimate . finally , not assuming any condition , split the iterates @xmath640 into a part that ends in @xmath546 , one iterate starting in @xmath546 , and a part that stays outside @xmath546 . combining the first two estimates for the first and last part ,
proves the last estimate .
@xmath110 + proof of proposition [ prop_exp ] .
as in the proof of theorem iii.6.4 in @xcite .
@xmath110 + [ prop_parameterstate ] there are constants @xmath641 so that for all large enough @xmath85 the following holds . if @xmath642 , @xmath643 and @xmath644 , @xmath645 proof .
let @xmath646 be the continuation of @xmath647 for @xmath176 near @xmath325 given by @xmath648 ( where @xmath649 is the hyperbolic periodic point from the definition of @xmath650 ) .
then proposition [ localmap ] implies that @xmath651 for some constant @xmath652 .
writing @xmath653 , the chain rule gives @xmath654 now @xmath655 is arbitrarily small for @xmath85 sufficiently large . by ( [ estder ] ) and the exponential growth of @xmath656 , the statement of the proposition holds for @xmath596 . by the chain rule ,
@xmath657 it follows that @xmath658 by assumption , @xmath659 for @xmath660 . for each positive integer @xmath94
, there are constants @xmath661 so that @xmath662 if @xmath663 . hence , @xmath664 for @xmath663 .
the constant @xmath665 is small if @xmath94 is large .
the proposition follows .
@xmath110 + the next proposition is also used in section [ sec_measure ] to show the existence of absolutely continuous invariant measures of @xmath321 for @xmath666 .
the binding period of @xmath667 is defined by ( [ binding1 ] ) , ( [ binding2 ] ) . [ lem_binding ] there exist constants @xmath668 and @xmath669 such that the following holds .
let @xmath456 , @xmath667 and suppose that @xmath305 is an essential return time for @xmath450 with return depth @xmath115 . then the binding period @xmath670 satisfies @xmath671 .
we have @xmath672 furthermore , @xmath673 , or @xmath674 , and @xmath675 .
the next lemma implies bounded distortion of iterates of @xmath676 on @xmath677 during the binding period .
the lemma is an ingredient for the proof of proposition [ lem_binding ] , as in @xcite .
the proof of the lemma differs from @xcite ; distortion estimates for a passage through @xmath678 have to be treated separately .
the remainder of the proof of proposition [ lem_binding ] is as in @xcite . for @xmath679 , @xmath680 for @xmath681 and for some @xmath92 .
write @xmath682 and @xmath683 . by the chain rule , @xmath684
if @xmath685 lies outside @xmath343 , then @xmath686 is bounded by a constant @xmath687 and @xmath688 if @xmath689 , we can write @xmath690 . thus @xmath691 . using this , it follows that again @xmath692 for some @xmath687 .
hence , @xmath693 we proceed to estimate @xmath694 . by the definition of binding period , @xmath695 .
write @xmath696 . if @xmath343 is small enough
, we have @xmath697 ( see section [ sec_relate ] ) .
we may hence assume that @xmath698 .
it follows from this and ( [ etaj ] ) that @xmath699 .
further @xmath700 for some @xmath687 , so that @xmath701 for some @xmath687 . combining ( [ etaj ] ) and ( [ zj ] ) shows that @xmath702 is bounded , thus proving the lemma .
we remark that the distortion bound is close to 1 if @xmath703 is small .
this follows from the observation that @xmath704 is outside a neighborhood of @xmath6 , where it undergoes exponential expansion , for a large number of iterates .
@xmath110 the main results from the previous sections are propositions [ prop_exp ] , [ prop_parameterstate ] , [ lem_binding ] .
these results have their counterparts in proofs of the work of benedicks and carleson . from this point on , we can follow @xcite closely .
for completeness we sketch the remaining steps leading to the proof of proposition [ jakobson ] in the next two sections . in the inductive constructions ,
the following two propositions are shown to hold .
the proofs are as in @xcite , relying on propositions [ prop_exp ] , [ prop_parameterstate ] , and [ lem_binding ] .
* [ bounded recurrence ] * each point in @xmath461 satisfies @xmath417 .
one shows that in fact @xmath705 , @xmath706 , if @xmath707 .
that is , a substantial proportion of the returns are chopping times . by assumption ,
hence , @xmath709 . as in @xcite
one shows that this bound implies @xmath417 . *
[ bounded distortion ] * restricted to a connected component of @xmath667 , the map @xmath710 is a diffeomorphism with uniformly bounded distortion for all @xmath711 where @xmath712 is the last essential or inessential return time of @xmath450 and @xmath713 is the associated binding period . if @xmath714 then the same statement holds for all @xmath715 restricted to any subinterval @xmath716 such that @xmath717 .
for the proof of proposition [ jakobson ] , combinatorial properties of @xmath462 are studied .
escape times play a central role in this study .
the combinatorial properties described next are used at the end of the section to prove proposition [ jakobson ] . to each @xmath718
is associated a sequence @xmath719 , @xmath720 of escape times and a corresponding sequence of escaping components @xmath721 with @xmath722 .
let @xmath723 for @xmath724 and @xmath725 for @xmath726 .
this defines @xmath727 for each @xmath728 .
observe that for @xmath729 and @xmath728 , the sets @xmath727 and @xmath730 are either disjoint or coincide .
define @xmath731 and let @xmath732 be the natural partition of @xmath733 into sets of the form @xmath727 .
observe that @xmath734 and @xmath735 .
for @xmath736 , @xmath737 , let @xmath738 denote by @xmath739 the partition @xmath740 restricted to @xmath450 .
define @xmath741 by @xmath742 and @xmath743 for @xmath744 .
let @xmath745 [ comb ] @xmath746 for some @xmath747 .
one can take @xmath748 , which is negative for @xmath749 small enough .
the proof divides into two parts .
one bounds the cardinality of @xmath750 by @xmath751 and one shows that for any @xmath752 , @xmath737 , and @xmath753 , one has @xmath754 .
combining the two statements proves the proposition . for the proofs one can follow @xcite .
@xmath110 + [ lem_estcale ] @xmath755 proof .
the equality follows immediately from the definitions . for the inequality , let @xmath728 , @xmath752 and write @xmath756 proposition [ comb ] ( with the remark from its proof that @xmath757 with @xmath758 ) implies @xmath759 so that @xmath760 assuming that @xmath761 has been chosen small enough and @xmath432 large enough .
since @xmath762 and @xmath763 is constant on elements of @xmath764 , we have @xmath765 applying ( [ repeatedly ] ) repeatedly gives @xmath766 @xmath110 + observe @xmath767 chebyshev s inequality and lemma [ lem_estcale ] yield @xmath768 if @xmath432 is large enough .
this implies @xmath769 write @xmath770 . for @xmath771 small
, there exists @xmath207 so that @xmath772 for all @xmath773 . hence @xmath774 noting that @xmath775 , for all @xmath85 , it follows that a uniform lower bound for @xmath776 exists .
this concludes the proof of proposition [ jakobson ] .
since @xmath207 goes to @xmath626 as @xmath777 , we have also shown that @xmath778 is a lebesgue density point of @xmath326 . in the previous sections
it was proved that @xmath342 has bounded recurrence for a set of parameter values @xmath328 with measure bounded from below , uniformly in @xmath85 . by proposition [ density ] , this implies that @xmath779 has positive measure and has positive density at @xmath46 .
the following proposition implies that @xmath321 , @xmath328 , has exponential expansion along the orbit of @xmath780 .
the proof is as in @xcite .
[ prop_br > exp ] if @xmath176 is close to @xmath325 and @xmath342 satisfies @xmath781 , then @xmath782 for some @xmath528 , @xmath783 .
combining proposition [ prop_br > exp ] with lemma [ lemma_withouttilde ] , gives [ prop_coleck ] for each @xmath318 , there are @xmath500 , @xmath784 , so that @xmath785 thus @xmath8 is a collet - eckmann map if @xmath318 .
collet - eckmann maps are known to admit absolutely continuous invariant measures , see theorem v.4.6 in @xcite .
this concludes the proof of theorem [ existence ] , except for the conclusion that @xmath786 $ ] , which we postpone until the next section .
in this section we study intermittent time series of @xmath787 at parameter values @xmath26 for which @xmath8 admits an absolutely continuous invariant measure .
a proof of theorem [ intermittency ] is in section [ sec_inter ] . for the proof one needs to know the ( average ) length of the relaminarization phase .
the relaminarization phase is studied in section [ app ] . in the previous section we constructed a set @xmath75 of parameter values with positive density at @xmath46 , so that @xmath8 has bounded recurrence ( see definition [ def_br ] ) for @xmath318 . by proposition [ prop_coleck ] , the bounded recurrence condition implies the collet - eckmann condition stating that @xmath788 for some @xmath687 , @xmath789 .
it is known that a map @xmath8 satisfying the collet - eckmann condition , possesses an absolutely continuous invariant measure , see @xcite . because we need bounds on the density of the invariant measures in our discussion of intermittency , we give an alternative way to produce invariant measures following @xcite .
( see also @xcite and @xcite ) [ prop_tildeacip ] for @xmath318 , @xmath787 possesses an absolutely continuous invariant measure @xmath790 .
there is a constant @xmath92 not depending on @xmath26 so that for any borel set @xmath36 , @xmath791 the support of @xmath790 equals @xmath792 $ ] .
outside the domain of definition of @xmath713 , let @xmath799 .
define the return map @xmath800 on @xmath20 $ ] by @xmath801 for any @xmath92 and all @xmath318 we may assume that @xmath802 is bounded from below by @xmath94 , by taking @xmath439 small enough . by proposition
[ prop_exp ] , if @xmath803 and @xmath804 , then @xmath805 for some @xmath806 . as stated in proposition
[ prop_exp ] , the constant @xmath633 does not depend on @xmath439 . by taking @xmath439 small , @xmath807 .
it follows that some power of @xmath800 is expanding .
the first part of the proposition follows as in the proof of theorem 1 in @xcite .
we sketch the necessary arguments to make clear that the constant @xmath94 does not depend on @xmath26 .
details are left to the reader .
the measure @xmath790 is constructed by finding its density as a fixed point of a perron - frobenius operator @xmath808 .
define @xmath809 note that @xmath810 let @xmath811 \to { \mbox{\bb r}}$ ] be the function @xmath812 where @xmath813 is continuous and @xmath814 elsewhere .
one shows that the variation of @xmath815 is bounded , uniformly in @xmath44 and @xmath26 .
this relies on the negative schwarzian derivative ( see @xcite ) and the analysis of the local saddle - node bifurcation in section [ sec_snlocal ] ( see proposition [ embedding ] ) . by induction the variation of @xmath815
is bounded .
using the uniform bound on the variation of @xmath815 , one bounds the variation of @xmath816 for a density @xmath817 with bounded variation .
it follows that @xmath808 has a fixed point with uniformly bounded variation .
if @xmath818 denotes the measure whose density is the fixed point of @xmath808 , then @xmath790 is obtained by pushing forward @xmath818 , @xmath819 where @xmath820 is the set on which @xmath821 .
the uniform bound for @xmath790 follows from the properties of @xmath800 , as in @xcite , see also @xcite .
that the support of @xmath790 equals all of @xmath792 $ ] follows as in theorem 2 in @xcite .
it relies on the fact that , if @xmath318 , then for each interval @xmath39 $ ] , there exists @xmath44 so that @xmath13 \subset f_{\gamma}^n(i)$ ] .
the necessary topological expansion at @xmath46 is guaranteed by lemma [ lem_over ] .
@xmath110 + the invariant measure @xmath90 for @xmath8 is constructed by pushing forward @xmath790 .
recall that @xmath822 with @xmath823 .
let @xmath824\backslash { { \tilde{e}}}_{\gamma } ) ) + \sum_{k=0}^{l } \sum_{i=0}^{kq-1 } \tilde{\nu}_{\gamma}(f_{\gamma}^{-i}(a ) \cap e^{-k}).\end{aligned}\ ] ] this measure is obviously finite and can thus be rescaled to a probability measure @xmath90 .
proof . to prove that @xmath90 is invariant for @xmath8 ,
recall that @xmath825 and @xmath787 equals @xmath8 outside @xmath678 .
compute @xmath826\backslash { { \tilde{e}}}_{\gamma } ) ) + \sum_{k=0}^{l } \sum_{i=1}^{kq-1 } \tilde{\nu}_{\gamma}(f_{\gamma}^{-i}(a ) \cap e^{-k } ) + \\ \nonumber & & \sum_{k=0}^{l } \tilde{\nu}_{\gamma}(f_{\gamma}^{-kq}(a ) \cap e^{-k } ) \\ \nonumber & = & \tilde{\nu}_{\gamma}({{\tilde{f}}}^{-1}_{\gamma}(a ) \cap ( [ 0,1]\backslash { { \tilde{e}}}_{\gamma } ) ) + \sum_{k=0}^{l } \sum_{i=1}^{kq-1 } \tilde{\nu}_{\gamma}(f_{\gamma}^{-i}(a ) \cap e^{-k } ) + \\ \label{eqn_inv } & & \sum_{k=0}^{l } \tilde{\nu}_{\gamma}({{\tilde{f}}}_{\gamma}^{-1}(a ) \cap e^{-k}).\end{aligned}\ ] ] by @xmath787 invariance of @xmath790 , @xmath827\backslash { { \tilde{e}}}_{\gamma } ) ) $ ] equals @xmath828\backslash { { \tilde{e}}}_{\gamma } ) ) $ ] and @xmath829 equals @xmath830 .
the right hand side of ( [ eqn_inv ] ) therefore is @xmath831 , showing @xmath8-invariance of @xmath832 and thus of @xmath90 .
@xmath110 + the following proposition serves several purposes , it is used to provide estimates for the average length of laminar and relaminarization phases .
its proof will be postponed to section [ app ] .
[ prop_outside ] let @xmath559 be a small neighborhood of @xmath6 . for @xmath833 $ ] ,
let @xmath834 be the number of iterations under @xmath787 required for @xmath63 to enter @xmath559 .
then there is @xmath835 so that for all @xmath318 , @xmath836 } \tilde{l}_{v } d\tilde{\nu}_{\gamma}/ \tilde{\nu}_{\gamma}([0,1]\backslash v ) & \le & l.\end{aligned}\ ] ] corresponding statements for the study of the boundary crisis bifurcation are contained in @xcite .
note that @xmath559 need not contain @xmath678 ; a similar statement where @xmath787 is replaced by @xmath8 is therefore untrue .
however , proposition [ prop_outside ] has the following corollary which deals with @xmath8 . paraphrasing , it shows that a typical ( with respect to the invariant measure @xmath90 ) point @xmath1 $ ] needs a bounded number of iterates to enter @xmath678 .
[ prop_relaminar ] for @xmath833 $ ] , let @xmath837 be the number of iterations under @xmath8 required for @xmath63 to enter @xmath678 .
then there is @xmath835 so that for all @xmath318 , @xmath836 } l_{{{\tilde{e}}}_{\gamma } } d\nu_{\gamma}/ \nu_{\gamma}([0,1]\backslash { { \tilde{e}}}_{\gamma } ) & \le & l.\end{aligned}\ ] ] proof . at @xmath46
, @xmath5 is renormalizable : there is an interval @xmath559 containing @xmath6 in its interior and @xmath30 in its boundary , so that @xmath838 . by slightly extending @xmath559
, we may assume that it contains @xmath678 for small values of @xmath26 .
the result follows from proposition [ prop_outside ] by noting that outside @xmath678 , @xmath787 equals @xmath8 .
@xmath110 + we will now show how theorem [ intermittency ] is proved by combining propositions [ prop_tildeacip ] , [ prop_outside ] and [ prop_relaminar ] .
+ proof of theorem [ intermittency ] .
let @xmath75 be as constructed in section [ sec_br ] and take @xmath318 .
let @xmath90 be the absolutely continuous invariant measure for @xmath8 , obtained in lemma [ prop_mb ] .
the measure @xmath90 is ergodic , so that by birkhoff s ergodic theorem , @xmath839 for any borel set @xmath840 $ ] and almost all @xmath833 $ ] .
it follows that also @xmath841 hence , for almost all @xmath1 $ ] , the distribution with which points in the orbit @xmath842 are in @xmath678 is given by @xmath790 .
let @xmath383 be a compact interval in @xmath251 .
if @xmath559 is a neighborhood of @xmath6 , then by ( [ ergodictilde ] ) and applying proposition [ prop_outside ] , the measure @xmath843 of @xmath559 is bounded from below by a positive constant , uniformly in @xmath26 .
observe that by invariance of @xmath790 , @xmath844 , @xmath289 .
therefore , the measure @xmath845 of @xmath383 is bounded from below by a positive constant , uniformly in @xmath26 .
an easy computation shows that for any compact interval @xmath383 inside @xmath251 , the number of iterates needed for a point @xmath846 to leave @xmath678 is bounded from below by @xmath847 for some @xmath92 .
it follows that the average duration of orbit pieces of @xmath8 in @xmath678 is bounded from below by @xmath847 for some @xmath92 . combining this with proposition [ prop_relaminar ] proves the upper bound on @xmath72 in theorem [ intermittency ] .
the lower bound is a trivial consequence of the fact that there is always a positive number of iterates between two laminar phases and that the maximum number of consecutive iterations in @xmath848 is bounded above by @xmath849 for some constant @xmath850 .
the argument to show that @xmath90 converges weakly to @xmath65 is similar . by definition of @xmath832 ,
@xmath851 reasoning as above one shows that @xmath852 , @xmath853 , is bounded from below , uniformly in @xmath26 .
it follows that @xmath854 gets arbitrarily large as @xmath855 ( because then @xmath156 ) .
since @xmath856\backslash { { \tilde{e}}}_{\gamma } ) = \tilde{\nu}_{\gamma } ( [ 0,1]\backslash { { \tilde{e}}}_{\gamma})$ ] is obviously bounded , @xmath857 as @xmath855 .
because this holds for any neighborhood @xmath678 , this shows that @xmath858 as @xmath855 .
@xmath110 purpose of this section is to indicate a proof of proposition [ prop_outside ] . to introduce the reasoning
, we start with an alternative proof for proposition [ prop_relaminar ] , which does not derive it as a corollary to proposition [ prop_outside ] . + proof of proposition [ prop_relaminar ] .
we claim that @xmath859\backslash { { \tilde{e}}}_{\gamma})$ ] is bounded away from 0 , uniformly in @xmath26 .
to establish the claim , note that @xmath860\backslash { { \tilde{e}}}_{\gamma})$ ] and @xmath20\backslash { { \tilde{e}}}_{\gamma}$ ] together cover @xmath20 $ ] . by @xmath787-invariance of @xmath790 ,
the measures of both sets are the same , and , since they add up to at least 1 , are bounded from below by @xmath861 . the normalized measure
@xmath862\backslash { { \tilde{e}}}_{\gamma})$ ] equals @xmath863\backslash { { \tilde{e}}}_{\gamma})$ ] outside @xmath678 .
therefore , applying proposition [ prop_tildeacip ] , it follows that a uniform bound of the form @xmath864\backslash { { \tilde{e}}}_{\gamma } ) & \le & k \sqrt { m(a)}\end{aligned}\ ] ] holds for borel sets @xmath865\backslash { { \tilde{e}}}_{\gamma})$ ] . at @xmath46 , @xmath5 is renormalizable : there is an interval @xmath559 containing @xmath6 in its interior and @xmath30 in its boundary , so that @xmath838 . by slightly extending @xmath559
, we may assume that it contains @xmath678 for small values of @xmath26 .
write @xmath866 .
write @xmath867 for the union of @xmath868 and those intervals in @xmath869 that contain a critical point for @xmath84 , see figure [ relaminar ] . of course , also for borel sets @xmath840\backslash o$ ]
a uniform bound of the form @xmath871\backslash o ) \le k \sqrt { m(a)}$ ] holds .
let @xmath872\backslash o \to o$ ] denote the restriction of @xmath8 to @xmath20\backslash o$ ] .
by @xcite , @xmath873 possesses a conditionally invariant measure @xmath818 ; @xmath818 is characterized by @xmath874\backslash o)\ ] ] for borel sets @xmath36 .
moreover , @xmath818 is absolutely continuous with respect to lebesgue measure and has density bounded and bounded away from 0 on its support , uniformly in @xmath26 .
say @xmath875 for borel sets
write @xmath876 for some @xmath877 .
since @xmath878 is bounded from below , @xmath879 is bounded away from 1 . applying ( [ unifbound ] ) , @xmath880 ) ) & \le & k \sqrt{m(h_{\gamma}^{-n}([0,1 ] ) ) } \\ & \le & k \sqrt{c } \sqrt{\zeta_{\gamma}(h_{\gamma}^{-n } ( [ 0,1 ] ) ) } \\ & = & k \sqrt{c } \sqrt{\beta_{\gamma}}^n.\end{aligned}\ ] ] it follows that @xmath881 } l_o d\nu_{\gamma}= \sum_{i\ge 1 } \nu_{\gamma}(h_{\gamma}^{-i}([0,1]))\ ] ] is bounded uniformly in @xmath26 .
@xmath110 + proof of proposition [ prop_outside ] .
let @xmath207 be the interval containing @xmath6 so that @xmath882 .
we may assume that @xmath883 and @xmath884 .
given @xmath559 , let @xmath885 be the union of the interval in @xmath886 that are mapped onto @xmath559 by iterates of @xmath8 .
write @xmath867 for the union of @xmath868 and those intervals in @xmath869 that contain a critical point for @xmath84 .
restricted to @xmath868 , @xmath8 equals @xmath787 .
include into @xmath867 the set of points @xmath887 , where @xmath628 is the set of discontinuities of @xmath787 .
write @xmath888 for the restriction of @xmath787 to @xmath20\backslash o$ ] . by proposition [ prop_exp ] , some iterate of @xmath888 is expanding .
observe that the number of branches of @xmath888 is constant for @xmath319 $ ] , but increases with @xmath85 since the branches of @xmath889 are included . for fixed @xmath85 , proposition [ prop_outside ] follows for @xmath319 \cap \omega$ ] as in the above indicated alternative proof of proposition [ relaminar ] .
to obtain uniform bounds in @xmath85 , we must investigate properties of conditionally invariant measures for @xmath888 .
following @xcite , there is a conditionally invariant measure @xmath890 for @xmath888 ; @xmath890 is characterized by @xmath891\backslash o).\ ] ] the proof of proposition [ prop_outside ] is identical to the above proof of proposition [ prop_relaminar ] , once lemma [ lemma ] is proved .
@xmath110 + proof . by proposition [ prop_exp ] ,
an iterate @xmath892 of @xmath888 is an expansion .
hence , the existence @xcite of the conditionally invariant measure follows from @xcite . to derive bounds on its density
, we must examine the existence proof .
the conditionally invariant measure @xmath890 is constructed by finding its density as a fixed point of a perron - frobenius operator .
write @xmath893)$ ] for the set of positive continuous functions @xmath563 on @xmath20 $ ] with @xmath894 } g dm = 1 $ ] . near a point
@xmath1 $ ] , @xmath888 has a number of inverse functions @xmath895 .
define a perron - frobenius operator @xmath808 on @xmath893)$ ] by @xmath896 ) } g dm\ ] ] a fixed point of @xmath808 is the density of a conditionally invariant measure .
it is shown in @xcite that @xmath808 possesses a unique fixed point @xmath897 and that for each @xmath898)$ ] , @xmath899 the fixed point @xmath897 is the density of the measure @xmath890 .
for a lipschitz density @xmath563 , let the regularity of @xmath563 be given by @xmath900 , g'(x)\mbox { is defined } , g(x)>0 \}.\ ] ] we will show that @xmath901 for some @xmath902 independent of @xmath563 .
we remark that from @xcite one concludes that such a bound holds when @xmath26 is restricted to an interval @xmath274 $ ] ( see also @xcite ) .
their arguments do not imply ( [ boundedregularity ] ) , the difficulty being the growth of the number of branches of @xmath888 as @xmath855 . to evaluate @xmath903 , compute @xmath904 ' \right| } { \sum_i \psi_i'(x ) g\circ \psi_i ( x ) } \\ & \le & \frac { \left| \sum_i \psi_i''(x ) g\circ \psi_i ( x ) \right| } { \sum_i \psi_i'(x ) g\circ \psi_i ( x ) } + \frac { \left| \sum_i \psi_i'(x ) g'\circ \psi_i ( x ) \psi_i'(x ) \right| } { \sum_i \psi_i'(x ) g\circ \psi_i ( x ) } \\ & \le & \max_i \frac{| \psi_i''(x)|}{\psi_i'(x ) } + \max_i |\phi_i'(x)| \frac{g'\circ \psi_i ( x)}{g\circ \psi_i(x)}.\end{aligned}\ ] ] in the last step , we used that @xmath905 for numbers @xmath906 and positive numbers @xmath907 .
we claim that @xmath908 is bounded , uniformly in @xmath26 .
this is clear for @xmath63 outside @xmath678 , where @xmath909 is bounded away from 0 and @xmath910 is bounded .
take @xmath911 and consider the inverse branch @xmath895 provided by the flow @xmath912 of the adapted vector field @xmath913 ( see proposition [ embedding ] ) .
thus @xmath914 .
to bound @xmath915 we consider the flow @xmath912 for negative @xmath916 .
let @xmath917 .
then @xmath918 is a solution of @xmath919 for @xmath26 small , @xmath920 is close to @xmath921 , which is nonzero by assumption .
further , @xmath922 is a solution of @xmath923 . using
this one can show that @xmath918 is bounded , uniformly in @xmath26 .
we sketch a possible reasoning .
one may assume that @xmath924 .
let @xmath925 and denote the corresponding flow by @xmath926 .
for arguments sake , take @xmath927 .
then , for @xmath928 , @xmath929 for some @xmath930 close to 1
. solving for @xmath926 shows that @xmath931 note that @xmath932 .
it follows that @xmath933 differentiating ( [ diffeqn ] ) yields @xmath934 these bounds and ( [ j ] ) prove that @xmath918 is bounded , uniformly in @xmath26 . replacing @xmath888 by @xmath892 , which is an expanding map
, the above reasoning shows that @xmath935 for constants @xmath936 and @xmath784 .
now ( [ boundedregularity ] ) follows by iterating the bound on the regularity of @xmath937 .
we can now basically follow the proof of proposition 7 in @xcite to conclude the lemma . take @xmath898)$ ] with @xmath938 . then , for @xmath1 $ ] , @xmath939 where @xmath895 are the inverse branches of @xmath940 near @xmath63 .
since @xmath894 } g dm = 1 $ ] , there is a component @xmath941 of @xmath20\backslash o$ ] on which @xmath942 .
we may also assume that @xmath943 on a component @xmath941 with @xmath944 for some constant @xmath945
. then @xmath946 .
if @xmath44 is large enough , @xmath947 \subset \tilde{h}_{\gamma}^n ( a_j)$ ] for all small @xmath26 .
hence , for each @xmath948 $ ] , there exists @xmath949 with @xmath950 .
this gives latexmath:[\[\begin{aligned } p^n_{\gamma}g ( x ) & \ge & |\psi_{i_0 } '' ( x)| g\circ \psi_{i_0 } ( x ) , \\ & \ge & @xmath944 , we may take @xmath949 so that @xmath952 is bounded from below . therefore , @xmath953 for some @xmath954 which is independent of @xmath563 and @xmath26 . by ( [ converges ] ) , this proves the lemma . | we discuss one parameter families of unimodal maps , with negative schwarzian derivative , unfolding a saddle - node bifurcation .
it was previously shown that for a parameter set of positive lebesgue density at the bifurcation , the maps possess attracting periodic orbits of high period . we show that there is also a parameter set of positive density at the bifurcation , for which the maps exhibit absolutely continuous invariant measures which are supported on the largest possible interval .
we prove that these measures converge weakly to an atomic measure supported on the orbit of the saddle - node point . using these measures
we analyze the intermittent time series that result from the destruction of the periodic attractor in the saddle - node bifurcation and prove asymptotic formulae for the frequency with which orbits visit the region previously occupied by the periodic attractor . |
bose - einstein condensates with dipole - dipole interaction have drawn significant interests over the past few years @xcite .
in particular , after the experimental observations of dipolar effects in cr condensates @xcite and in spin-1 rb @xcite condensates , the study of dipolar quantum gases is at the forefront of both theoretical and experimental research .
armed with the technique of feshbach resonance which tunes the short - range @xmath0-wave scattering length , experimentalists have investigated the stability and the anisotropic collapse of the cr condensate @xcite .
in addition , dipolar effects have also been observed in condensates of k @xcite and li @xcite atoms which only possess small magnetic dipole moments .
owing to the large electric dipole moment of the polar molecules , recent success in making high phase - space - density gas of polar molecules @xcite provides an even more promising platform for the study the dipolar quantum many - body system .
the long - range and anisotropic characters of the dipolar interaction have profound impact on the properties of quantum gases .
for example , when loaded into a optical lattice , the long - range repulsive dipole - dipole interaction breaks the translational symmetry of the lattice such that the density wave phases can be formed @xcite . in the presence of vortices @xcite or a small perturbation in the trapping potential @xcite
, the radial density profile of the condensate was found to develop oscillations .
of particular interest is that , even for a singly trapped dipolar condensate , a biconcave density profile can be found in a pancake - shape trap @xcite . in this case
, the maximum density occurs at the periphery of the condensate , resembling the shape of a red blood cell .
et al_. have proposed to detect the biconcave density profile by tuning the contact interaction to induce local collapse @xcite . in the present paper ,
we systematically investigate the ground state structure and the free expansion dynamics of a harmonically trapped dipolar condensate . for a given trap geometry ,
we map out the stability diagrams on the parameter plane of the interaction strengths .
in addition to the biconcave dipolar condensates in pancake - shaped traps , a dumbbell density profile can be induced in cigar - shaped trap when the dipolar interaction is properly tuned .
for simplicity , both biconcave and dumbbell density profiles are referred to as density oscillations in the condensates . on stability diagram ,
the density oscillation occurs only when the interaction strengths fall into certain isolated areas , which is called as density oscillation island ( doi ) .
we also find the relation between the position of doi and the trap geometry by fitting the numerical results . finally , by studying the free expansion of the condensate with density oscillation
, we show that the density oscillation can be identified in the time - of - flight image .
we will also consider dipolar condensate trapped in box potentials .
we show that the rigid walls of box potentials in general induce more pronounced density oscillations over a much larger region of the parameter space than harmonic potentials .
the remainder of this paper is organized as follows . in sec .
[ stheo ] , we briefly outline our model . in sec .
[ snum ] , we introduce the numerical method used to compute the dipolar mean - field . the numerical results on both the ground state wave function of harmonically trapped dipolar condensates and the corresponding free expansion dynamics are presented in sec . [ sresu ] . in sec .
[ box ] , we discuss the enhanced density oscillation in a box potential . finally , we conclude in sec .
[ sconc ] .
we consider a trapped condensate of @xmath1 bosons with dipole moments @xmath2 .
particles interact with each other via contact interaction @xmath3 and dipole - dipole interaction @xmath4 .
the total interaction potential takes the form @xmath5 where @xmath6 with @xmath7 being the mass of the particle and @xmath8 being the @xmath0-wave scattering length tunable through feshbach resonance , @xmath9 is the polar angle of @xmath10 , and @xmath11 $ ] is a parameter continuously tunable via an orienting field which rotates rapidly around @xmath12-axis @xcite .
the inclusion of the factor @xmath13 makes the dipolar interaction between polarized dipoles tunable . for positive @xmath13 ,
the dipolar force is repulsive ( attractive ) if two dipole moments aligned side by side ( head to tail ) .
however , the opposite becomes true for negative @xmath13 .
moreover , the trapping potential is assumed to be axially symmetric @xmath14 where @xmath15 is the radial frequency and @xmath16 is trap aspect ratio such that the axially frequency of the trap is @xmath17 .
the condensate wave function satisfies the non - local gross - pitaevskii equation , which , in dimensionless form , becomes @xmath18\psi({\mathbf r}),\label{gp3d}\end{aligned}\ ] ] where @xmath19 and @xmath20 are the dimensionless parameters , characterizing the strengths of the contact and dipolar interactions , respectively .
the dimensionless units adopted here are @xmath21 for length , @xmath22 for energy , @xmath23 for time , and @xmath24 for wave function . in eq .
( [ gp3d ] ) , the control parameters of the system are reduced to @xmath16 , @xmath25 , and @xmath26 . we shall show how the condensate stability and the structure of the wave function depend on these parameters .
. ] following the standard procedure ,
the numerical solution of the ground state wave function can be obtained by evolving eq .
( [ gp3d ] ) in imaginary time .
compared to the gross - pitaevskii equation with only contact interaction , the difficulty of solving eq .
( [ gp3d ] ) lies at the calculating the mean - field potential of the dipole - dipole interaction @xmath27,\label{phidd}\end{aligned}\ ] ] where @xmath28 denotes the inverse fourier transform , @xmath29 is the density of the condensate with @xmath30 being its fourier transform , and @xmath31 is the fourier transform of the dipolar interaction potential with @xmath32 being the polar angle of @xmath33 . from the second line of eq .
( [ phidd ] ) , it becomes apparent that @xmath34 can be efficiently calculated in momentum space using fast fourier transform ( fft ) @xcite .
moreover , taking advantage of the cylindric symmetry of the system , ronen _
et al_. @xcite further reduce the 2d fourier transform in @xmath35 plane into an 1d hankel transform of order @xmath36 for the ground state wave function by integrating over the azimuthal variable .
this procedure greatly simplifies the 3d calculation to a 2d one in the @xmath37-@xmath12 plane .
the fact that fft implicitly treats the system as a periodic one makes our system a 3d periodic lattice of condensates .
therefore , the dipolar interactions between the artificial periodic copies of condensates introduce error to the mean - field potential @xmath38 . to solve this problem , ronen _
et al_. @xcite suggest to truncate the dipolar interaction such that @xmath4 is nonzero only within a sphere of radius @xmath39 , i.e. , @xmath40 this truncation of the dipolar interaction has no physical consequence as long as @xmath39 is larger than the extend of the actual condensate , but it eliminates the interaction between artificial copies of condensates .
this procedure works well in nearly spherical traps .
however , for cigar- or pancake - shaped trapping potential , one can not find a sphere which only contains a single condensate . here
, a natural generalization is to introduce cylindrical truncation on the dipolar interaction potential @xmath41 the fourier transform of @xmath42 takes the form @xmath43\nonumber\\ & & -4\pi\int_r^\infty \rho
d\rho \int_0^zdz \cos(k_zz)\nonumber\\ & & \times\frac{\rho^2 - 2z^2}{(\rho^2+z^2)^{5/2}}j_0(k_\rho\rho),\label{cylcut}\end{aligned}\ ] ] where @xmath44 , @xmath45 , @xmath46 is the zeroth order bessel function .
the first two lines at the r.h.s . of eq .
( [ cylcut ] ) are previously obtained in ref .
@xcite , and the remaining integral term can be easily worked out using numerical integration . with a gaussian wave function
, we have confirmed that the precision of the dipolar interaction energy is improved significantly by introducing cylindrical truncation on @xmath4 .
before we conclude this section , let us briefly summarize the numerical procedure employed to find the ground state wave function of a dipolar condensate .
for a given set of control parameters @xmath47 , we first determine the proper values of @xmath39 and @xmath48 .
then , using eq .
( [ cylcut ] ) , we numerically compute the fourier transform of the modified dipolar interaction potential eq .
( [ cylcut ] ) . finally , following the standard procedure for imaginary time evolution , we obtain the ground state wave function .
in this section , we shall first map out the stability diagram of a trapped dipolar condensate in the parameter space ( @xmath49 ) .
the value of the trap aspect ratio @xmath16 ranges from @xmath50 to @xmath51 , which covers both highly prolate and oblate trap geometries . for given parameters @xmath16 and @xmath26 ,
increasing the repulsive contact interaction will stabilize the condensate , we therefore manually set an upper limit on @xmath25 such that @xmath52 .
the lower limit of @xmath25 , on the other hand , is introduced naturally as the system always becomes unstable by tuning the contact interaction to attractive with sufficiently large strength .
similarly , the lower and upper limits of the dipolar interaction strength are also determined by the stability of the condensate . in the second part of this section , we turn to study the free expansion dynamics of an initially trapped dipolar condensate , aiming at finding signatures of the density oscillation in time - of - flight image . of the condensate in the presence of density oscillation for @xmath53 ( a ) and @xmath54 ( b).,width=307 ] in previous numerical studies @xcite ,
the stability diagram is plotted in @xmath16-@xmath26 parameter plane for a given contact interaction strength @xmath25 ( which is usually taken to be zero ) . here , for a fixed trap aspect ratio @xmath16
, we map out the stability diagram in @xmath25-@xmath26 parameter plane .
as we shall show below , this scheme makes it more convenient for us to understand the behavior of the density oscillation in a trapped dipolar condensate .
figure [ cril01 ] plots the stability diagram in @xmath25-@xmath26 parameter plane of a dipolar condensate in a spherical trap , which also represents the basic structure of the stability in other trap geometries .
since the dipolar force is partially attractive , a condensate always becomes unstable as one increases the absolute value of @xmath26 .
in fact , for given @xmath25 and @xmath16 , there exist two critical dipolar interaction strengths , @xmath55 and @xmath56 , such that the condensate is stable only if @xmath57 . alternatively ,
if @xmath26 is fixed , the instability can also be induced by decreasing the scattering length ( even to negative ) , similar to what have been done experimentally @xcite .
moreover , when @xmath25 ( @xmath58 ) is so large that the system falls into the thomas - fermi regime , the variational calculation shows that the stability boundary for given @xmath16 only depends on the ratio of @xmath26 and @xmath25 @xcite .
this properties is clearly reflected in fig .
[ cril01 ] , as for @xmath59 , the critical @xmath26 values linearly depend on @xmath25 with slopes @xmath60 and @xmath61 for @xmath62 and @xmath63 , respectively .
the reason that the absolute value of @xmath64 is larger than @xmath65 is because , for spherical traps , the dipolar force with negative @xmath26 is more attractive " than that with positive @xmath26 .
in general , the stability boundaries should depend on the trap aspect ratio @xmath16 .
however , we find that , for positive ( negative ) @xmath26 in prolate ( oblate ) trap , the critical dipolar interaction strength is roughly independent of the trap geometry , i.e. , @xmath66 and @xmath67 .
in fact , under these conditions , the condensate shrinks as the dipolar interaction strength grows , such that it only occupies a small spatial volume near the stability boundary .
therefore , the critical dipolar interaction strength becomes insensitive to @xmath16 .
for this reason , we shall focus on the upper stability boundary ( @xmath62 ) for oblate traps and the lower stability boundary ( @xmath63 ) for prolate traps . ) with fitting parameters listed in tab .
[ tabl].,width=259 ] .the values of fitting parameters for the positions of dois .
[ cols="^,^,^,^",options="header " , ] in fig .
[ crioblate ] , we summarize the main results for the stability diagrams in oblate traps .
density oscillations occur when the interaction parameters fall into the shaded areas , namely , the condensate has a biconcave density profile in oblate trap .
the origin of the density oscillation can be intuitively understood as follows : the long - ranged dipolar interaction is repulsive in @xmath68-plane , therefore , when the dipolar force exceeds certain threshold value the density at the center of the trap is depleted to lower the total energy . on the other hand , as one increases the dipolar interaction strength @xmath26 , the attractive dipolar interaction along @xmath12-axis also becomes stronger , which destablizes the system . in a slightly oblate trap ,
the system becomes unstable before such threshold dipolar force required to from the density oscillation is reached .
therefore , for @xmath69 , we do not find any density oscillation in condensate wave function for @xmath25 up to @xmath70 . to get a better understanding of the density oscillation , we plot , in fig .
[ effpot ] ( a ) , the effective potential of the system @xmath71 apparently , when density oscillation occurs in pancake - shaped traps , the effective potential takes a shape of mexican hat . for @xmath72 and @xmath73 .
( b ) energies as functions of @xmath25 for @xmath72 and @xmath74 .
the solid , dashed , and dash - dotted lines correspond to , respectively , the kinetic , contact interaction , and dipolar interaction energies .
the shaded area marks the range of the interaction parameters where the density oscillation appears.,width=240 ] the first density oscillations island ( denoted as doi-1 ) appears when @xmath75 .
the shape of the doi clearly indicates that the stability of the system is slightly increased by forming density oscillation in condensate , and away from doi , the @xmath76 roughly linearly depends on @xmath25 .
as we increase @xmath16 , doi-1 moves along the stability boundary to negative direction of the @xmath25-axis .
in addition , the area of doi-1 shrinks . for even larger
@xmath16 , the second ( doi-2 ) and third ( doi-3 ) density oscillation regions appear .
they also move along the stability boundary by following the similar pattern as doi-1 when the trap aspect ration is increased .
a doi can be roughly characterized using the contact interaction strength : the range of @xmath25 parameter covered by the doi , @xmath77 , represents its width ; while the @xmath25 value at the center of this range , @xmath78 , is the position of the doi .
figure [ center ] shows the trap aspect ratio dependence of the positions of the three dois ; the corresponding width is plotted as error bar .
one immediately notes that the @xmath16 dependence of @xmath78 can be fitted using equation @xmath79 where the values of the fitting parameters @xmath80 , @xmath81 , and @xmath82 for different dois are listed in tab .
[ tabl ] . as shown in fig .
[ center ] , the fitted curves agree very well with the numerical results for all dois .
moreover , the values of the fitting parameter @xmath82 also suggest that increasing @xmath16 will eventually pushes all three dois to the region with @xmath83 . for large @xmath16 ,
all three dois are squeezed into a small region of @xmath25 value , but they remain disconnected .
we point out that even though eq .
( [ fit ] ) can not explain why the dois appear under those contact interaction strengths , it can be used to predict the position of dois .
once the position and width of a doi is known , one may access the doi by simply increasing the strength of dipolar interaction strength to the value close to the critical value . in fig .
[ energy ] ( a ) , we present the typical @xmath26 dependences of the energies per particle .
one immediately notices that the energies vary smoothly as the system enters the doi , indicating that condensates with and without density oscillation can not be distinguished from each other by measuring the release energy . as shown in fig .
[ energy ] ( b ) , for fixed @xmath26 , one may also enters the doi by lowering the contact interaction strength .
again , the energies are smooth functions of @xmath25 across the boundary of a doi . for a condensate in cigar - shaped traps .
the shaded areas mark the dois .
the inset shows the typical result for the isodensity surface of the condensate .
the broken arrows denote the direction to which the dois move as @xmath16 is decreased.,width=259 ] the stability diagram of a dipolar condensate in cigar - shaped traps is presented in fig .
[ cricigar ] , where the dois appear when @xmath84 .
in contrast to @xmath85 case , the dipolar interaction becomes repulsive ( attractive ) along axial ( radial ) direction for @xmath86 .
therefore , along @xmath12 direction , the repulsive dipolar force depletes the density at the center of the trap such that the isodensity surface of the condensate with density oscillation takes dumbbell - shaped form . corresponding to dumbbell structure , the effective potential [ fig .
[ effpot ] ( b ) ] takes the form of a double - well potential . as the trapping potential becomes more pancake - shaped , the dois move along the stability boundary to the negative direction of the @xmath25-axis ; in addition , the area of the doi shrinks .
we remark that our results do not rule out the possibility that there might exist more dois for larger @xmath25 values , since our numerical calculations have only covered limited ranges of the parameters @xmath16 and @xmath25 .
now we turn to study the expansion of an initially trapped dipolar condensates .
compared the contact interaction , the anisotropic nature of the dipolar interaction makes the expanded cloud behave quite differently @xcite .
indeed , it has been used experimentally as a diagnostic tool for the detection the dipolar effects in bose - einstein condensate @xcite .
( a ) , @xmath87 ( b ) , @xmath88 ( c ) , and @xmath89 ( d ) .
the initial state are obtained using @xmath90 ; during the expansion , @xmath25 is tuned to zero.,width=307 ] to study the free expansion dynamics , we consider an initially trapped condensate with control parameters ( @xmath49 ) falling into a doi . at time @xmath91 , we switch off the trapping potential , the expansion dynamics of the condensate is then described by eq .
( [ gp3d ] ) with @xmath92 .
if the interactions are ignored completely during expansion , the condensate will expand ballistically such that the time - of - flight image represents the momentum distribution of the trapped condensate . from the fourier transform of the initial wave function with density oscillation
, one can easily deduce that two side peaks would appear in the time - of - flight image , which can be used as the signature of the density oscillation in the initial condensate . to demonstrate this
, we consider the expansion of an initially cigar - shaped condensate obtained using the control parameter @xmath90 . in stead of tuning both @xmath25 and @xmath26 to zero , here
we only switch off the contact interaction for @xmath93 .
figure [ expcigar ] shows the column densities of the expanded cloud , i.e. , @xmath94 due to the density oscillation , the initial column density has two peaks , in analog to that of the condensate trapped in a double - well potential .
after the condensate expands such that these two peaks overlap , the third peak appears as a result of the interference . at time @xmath95 , three peaks have roughly the same height .
when the system evolves continuously , the heights of the two side peaks become lower and lower , but they are still visible at @xmath96 . however , if the contact interaction remains unchanged during free expansion , the above scenario will be spoiled such that the side peaks vanish very quickly for a cigar - shaped condensate . ( a ) ,
@xmath87 ( b ) , and @xmath97 ( c ) .
the curves at the upper row show the corresponding column density on @xmath98-axis , i.e. , @xmath99 .
the initial wave function is obtained using the control parameters @xmath100.,width=307 ] fortunately , for a pancake - shaped condensate , the side peaks of the column density remain visible for a long time if initially the dipolar interaction is very close to the stability boundary . in fig .
[ expoblate ] , we demonstrate the free expansion of a pancake - shaped condensate with initial wave function obtained using the control parameters @xmath100 . to make it easier for visualizing the structure of the expanded cloud
, we also plot the column density on @xmath98-axis .
it can be seen that , even for @xmath101 , the side peaks are still distinguishable from the column density , indicating that the time - of - flight image may be used to detect the density oscillation in a dipolar condensate .
in this section we briefly discuss the density oscillations of a dipolar condensate trapped in a box potential . specifically , we consider a potential which is harmonic along the @xmath12-axis with frequency @xmath102 , but is represented by infinite square wells with length @xmath103 along the @xmath98- and @xmath104-axis .
such potential can be realized using , for example , tightly focused light sheet @xcite .
-@xmath104 plane and a harmonic potential in the @xmath12 axis .
the region between the lower and the upper solid lines represents regions where density oscillation will occur in the ground state .
the inset shows an isodensity surface of a dipolar condensate with density oscillation .
the length @xmath103 along the @xmath98- and @xmath104-axe is taken to be 28@xmath105 , where @xmath106 .
the dimensionless interaction strengths are defined as @xmath107 and @xmath108.,width=307 ] figure [ boxf ] displays the stability diagram of such a system . as in the previous case
, there exists an critical dipolar interaction strength beyond which the system will collapse .
however , a rather significant difference here is the much enlarged region of density oscillation .
in fact , here there is no disconnected dois and density oscillation dominates a major portion of the parameter space .
the inset of fig .
[ boxf ] exhibits a typical density plot of a condensate trapped in such a potential .
one can easily see the pronounced density oscillations near the walls . that a hard wall induces density oscillations in dipolar condensate can be understood from the roton excitation of the system @xcite . here
the effect of the wall is not unlike a vortex line @xcite or a localized repulsive potential @xcite : they all create a local superfluid density depression , in the vicinity of which the roton mode becomes soft and manifests itself as density oscillations @xcite . the same physics also underlies the predicted density ripples in superfluid helium near a boundary @xcite or a vortex line @xcite . from this study , one can conclude that density oscillations in dipolar condensates can be very easily induced by a rigid wall .
however , in case the condensate is not confined in the @xmath98-@xmath104 plane ( which can be simulated using the periodic boundary condition in the plane ) , we do not observe any density oscillations as dipolar interaction strength increases , the system remains uniform in the transverse plane until it collapses .
to conclude , we have systematically investigated the density oscillations in trapped dipolar condensates by mapping out the stability diagram of the system on @xmath25-@xmath26 parameter plane .
in addition to the biconcave density profile in an oblate trap , the dipolar interaction can also generate dumbbell - shaped density profiles in prolate traps . on stability diagram ,
the density oscillation occurs only when the interaction strengths fall into the dois .
the relation between the positions of dois and the trap geometry is explored by fitting the numerical results .
we also study the free expansion of the condensate with density oscillation , it is shown that the density oscillation can be identified in the time - of - flight image . finally , to improve the numerical precision for the calculation of dipolar interaction energy , we introduce the truncated dipolar interaction potential which is nonzero only within a properly chosen cylinder .
this work was supported by nsfc through grants 10974209 and 10935010 and by the national 973 program ( grant no .
2006cb921205 ) .
sy acknowledges the support from the bairen " program of chinese academy of sciences .
hp acknowledges the support from nsf . | we investigated the ground state wave function and free expansion of a trapped dipolar condensate .
we find that dipolar interaction may induce both biconcave and dumbbell density profiles in , respectively , the pancake- and cigar - shaped traps . on the parameter plane of the interaction strengths , the density oscillation occurs only when the interaction parameters fall into certain isolated areas .
the relation between the positions of these areas and the trap geometry is explored . by studying the free expansion of the condensate with density oscillation
, we show that the density oscillation is detectable from the time - of - flight image . |
centrality measures the degree to which network structure contributes to the importance , or status , of a node in a network .
over the years many different centrality metrics have been defined .
one of the more popular metrics , betweenness centrality @xcite , measures the fraction of all shortest paths in a network that pass through a given node .
other centrality metrics include those based on random walks @xcite and path - based metrics .
the simplest path - based metric , degree centrality , measures the number of edges that connect a node to others in a network . according to this measure ,
the most important nodes are those that have the most connections . however , a node s centrality depends not only on how many others it is connected to but also on the centralities of those nodes @xcite .
this measure is captured by the total number of paths linking a node to other nodes in a network .
one such metric , @xmath0-centrality @xcite , measures the total number of paths from a node , exponentially attenuated by their length . the attenuation parameter sets the length scale of interactions . unlike other centrality metrics , which do not distinguish between local and global structure
, a parameterized centrality metric can differentiate between locally connected nodes , i.e. , nodes that are linked to other nodes which are themselves interconnected , and globally connected nodes that link and mediate communication between poorly connected groups of nodes .
studies of human @xcite and animal @xcite populations suggest that such ` bridges ' or ` brokers ' play a crucial role in the information flow and cohesiveness of the entire group .
one difficulty in applying @xmath0-centrality in network analysis is that its key parameter is bounded by the spectrum of the corresponding adjacency matrix of the network . as a result ,
the metric diverges for larger values of this parameter .
we address this problem by defining _
normalized @xmath0-centrality_. we show that the new metric avoids the problem of bounded parameters while retaining the desirable characteristics of @xmath0-centrality , namely its ability to differentiate between local and global structures .
in addition to ranking nodes , parameterized centrality can be used to identify communities within a network @xcite . in this paper
, we generalize modularity maximization - based approach @xcite to use normalized @xmath0-centrality . rather than find regions of the network that have greater than expected number of edges connecting nodes @xcite
, our approach looks for regions that have greater than expected number of weighted paths connecting nodes .
one advantage of this method is that the attenuation parameter can be varied to identify local vs. global communities .
normalized @xmath0-centrality is a powerful tool for network analysis . by differentiating between locally and globally connected nodes , it provides a simple alternative to previous attempts to quantify fine - grained structure of complex networks , such as the motif - based @xcite and role - based @xcite descriptions .
the former measures the relative abundance of subgraphs of a certain type , while latter classifies nodes according to their connectivity within and outside of their community . applying either of these descriptions to real networks is computationally expensive : role - based analysis , for example , requires the network to be decomposed into distinct communities first .
normalized @xmath0-centrality , on the other hand , measures node connectivity at different length scales , allowing us to resolve network structure in a computationally efficient manner
. we use normalized @xmath0-centrality to study the structure of several benchmark networks , as well as a real - world online social network .
we show that this parameterized centrality metric can identify locally and globally important nodes and communities , leading to a more nuanced understanding of network structure .
bonacich @xcite defined _ @xmath0-centrality _ @xmath1 as the total number of attenuated paths between nodes @xmath2 and @xmath3 , with @xmath4 and @xmath0 giving the attenuation factors . in communication and information networks
we are considering , @xmath5 . ] along direct edges ( from @xmath2 ) and indirect edges ( from intermediate nodes ) in the path from @xmath2 to @xmath3 , respectively , and @xmath6 is the length of the longest path . given the adjacency matrix of the network @xmath7 , @xmath0-centrality matrix is defined as follows : @xmath8 the first term gives the number of paths of length one ( edges ) from @xmath2 to @xmath3 , the second gives the number of paths of length two , etc .
although @xmath9 along different edges in a path could in principle be different , for simplicity , we take them all to be equal : @xmath10 . in this case , the series converges to @xmath11 , which holds while @xmath12 , where @xmath13 is the largest characteristic root of @xmath7 @xcite .
the computation of @xmath13 is difficult , especially for large networks , which include most complex real - world networks . to get around this difficulty
, we define _ normalized @xmath0-centrality _ matrix as : @xmath14 as we show in the appendix , in contrast to @xmath0-centrality , normalized @xmath0-centrality is not bounded by @xmath13 .
also , we prove that , assuming @xmath15 is strictly greater than any other eigenvalue , @xmath16 exists ; and as @xmath0 is increased , @xmath17 converges to this value and is finite for @xmath18 . just like the original @xmath0-centrality , normalized @xmath0-centrality contains a tunable parameter @xmath0 that sets the length scale of interactions . for @xmath19 , ( normalized
) @xmath0-centrality takes into account direct edges only . as @xmath0 increases
, @xmath17 becomes a more global measure , taking into account ever larger network components .
the expected length of a path , the radius of centrality , is @xmath20 .
much of the analysis done by social scientists considered local structure , i.e. , the number @xcite and nature @xcite of an individual s ties . by focusing on local structure , however , traditional theories fail to take into account the macroscopic structure of the network .
many metrics proposed and studied over the years deal with this shortcoming , including pagerank @xcite and random walk centrality @xcite .
these metrics aim to identify nodes that are ` close ' in some sense to other nodes in the network , and are therefore , more important .
pagerank , for example , gives the probability that a random walk initiated at node @xmath2 will reach @xmath3 , while random - walk centrality computes the number of times a node @xmath2 will be visited by walks from all pairs of nodes in the network .
normalized @xmath0-centrality , @xmath21 , also measures how ` close ' node @xmath2 is to other nodes in a network and can be used to rank the nodes accordingly .
the presence of a tunable parameter turns normalized @xmath0-centrality into a powerful tool for studying network structure and allows us to seamlessly connect the rankings produced by well - known local and global centrality metrics . for @xmath19 ,
normalized @xmath0-centrality takes into account local interactions that are mediated by direct edges only , and therefore , reduces to _ degree centrality_. as @xmath0 increases and longer range interactions become more important , nodes that are connected by longer paths grow in importance . for @xmath22 ,
the rankings produced by normalized @xmath0-centrality are equivalent to those produced by @xmath0-centrality .
also as shown in the appendix , for symmetric matrices , as @xmath23 , normalized @xmath0-centrality converges to _ eigenvector centrality _ @xcite .
the rankings no longer change as @xmath0 increases further , since @xmath0 has reached some fundamental length scale of the network .
girvan & newman @xcite proposed _ modularity _ as a metric for evaluating community structure of a network .
the modularity - optimization class of community detection algorithms @xcite finds a network division that maximizes the modularity , which is defined as @xmath24 ( connectivity within community)-(expected connectivity ) , where connectivity is measured by the density of edges .
we extend this definition to use normalized @xmath0-centrality as the measure of network connectivity . according to this definition , in the best division of a network ,
there are more weighted paths connecting nodes to others within their own community than to nodes in other communities .
modularity can , therefore , be written as : @xmath25\delta(s_i , s_j)}\ ] ] @xmath26 is given by eq .
( [ eq : inf1 ] ) . since @xmath4 factors out of modularity , without loss of generality we take @xmath27 .
@xmath0 can be varied from 0 to 1 .
@xmath28 is the expected normalized @xmath0-centrality , and @xmath29 is the index of the community @xmath2 belongs to , with @xmath30 if @xmath31 ; otherwise , @xmath32 .
we round the values of @xmath33 to the nearest integer .
to compute @xmath34 , we consider a graph , referred to as the null model , which has the same number of nodes and edges as the original graph , but in which the edges are placed at random . to make the derivation below more intuitive , instead of normalized @xmath0-centrality , we talk of the number of attenuated paths . in normalized @xmath0-centrality
, the number of attenuated paths is scaled by a constant , hence the derivation below holds true .
when all the nodes are placed in a single group , then axiomatically , @xmath35 .
therefore @xmath36 = 0 $ ] , and we set @xmath37 therefore , according to the argument above , the total number of paths between nodes in the null model @xmath38 is equal to the total number of paths in the original graph , @xmath39 .
we further restrict the choice of null model to one where the expected number of paths reaching node @xmath3 , @xmath40 , is equal to the actual number of paths reaching the corresponding node in the original graph .
similarly , we also assume that in the null model , the expected number of paths originating at node @xmath2 , @xmath42 , is equal to the actual number of paths originating at the corresponding node in the original graph @xmath43 .
@xmath44 , @xmath42 and @xmath40 are then rounded to the nearest integers .
next , we reduce the original graph @xmath45 to a new graph @xmath46 that has the same number of nodes as @xmath45 and total number of edges @xmath44 , such that each edge has weight 1 and the number of edges between nodes @xmath2 and @xmath3 in @xmath46 is @xmath47 . now
the expected number of paths between @xmath2 and @xmath3 in graph @xmath45 could be taken as the expected number of the edges between nodes @xmath2 and @xmath3 in graph @xmath46 and the actual number of paths between nodes @xmath2 and @xmath3 in graph @xmath45 can be taken as the actual number of edges between node @xmath2 and node @xmath3 in graph @xmath46 .
the equivalent random graph @xmath48 is used to find the _ expected _ number of edges from node @xmath2 to node @xmath3 . in this graph
the edges are placed in random subject to constraints : ( _ i _ ) the total number of edges in @xmath48 is @xmath44 ; ( _ ii _ ) the out - degree of node @xmath2 in @xmath48 = out - degree of node @xmath2 in @xmath49 ; ( _ iii _ ) the in - degree of a node @xmath3 in graph @xmath48 = in - degree of node @xmath3 in graph @xmath50 . thus in @xmath48
the probability that an edge will emanate from a particular node depends only on the out - degree of that node ; the probability that an edge is incident on a particular node depends only on the in - degree of that node ; and the probabilities of the two nodes being the two ends of a single edge are independent of each other . in this case , the probability that an edge exists from @xmath2 to @xmath3 is given by _
edge in @xmath51 emanates from i _ @xmath52 _ _ edge in @xmath51 incident on j__=@xmath53 . since the total number of edges is @xmath44 in @xmath48 , therefore the expected number of edges between @xmath2 and @xmath3
is @xmath54 , the expected the expected @xmath0 centrality in @xmath45 .
once we compute @xmath55 , we have to select an algorithm to divide the network into communities that maximize @xmath55 .
brandes et al .
@xcite have shown that the decision version of modularity maximization is np - complete . like others @xcite
, we use the leading eigenvector method to obtain an approximate solution . in this method
, nodes are assigned to either of two groups based on a single eigenvector corresponding to the largest positive eigenvalue of the modularity matrix .
this process is repeated for each group until modularity does not increase further upon division .
we can generalize the centrality metric presented above to a notion of path - based connectivity and relate it to other centrality metrics .
let @xmath56 be a @xmath57 matrix such that @xmath58 is the number of paths of length @xmath6 connecting nodes @xmath2 and @xmath3 .
the number of paths of length one connecting @xmath2 and @xmath3 is @xmath59 ; the number of paths of length two is @xmath60 , etc .
the expected number of paths connecting nodes @xmath2 and @xmath3 is @xmath61 where : @xmath62 where @xmath63 can be a scalar or a vector . the proximity score @xmath61 can be used to find out how connected , or close , two nodes are .
several path - based centrality metrics can be expressed in terms of @xmath61 , including random walk models @xcite , degree centrality , katz score @xcite , as well as @xmath0-centrality . in a random walk model ,
a particle starts a random walk at node @xmath2 , and iteratively transitions to its neighbors with probability proportional to the corresponding edge weight . at each step ,
the particle returns to @xmath2 with some restart probability ( @xmath64 ) .
the proximity score is defined as the steady - state probability @xmath65 that the particle will reach node @xmath3 @xcite . * if @xmath66 where @xmath67 is a constant and @xmath68 is an @xmath69 matrix with @xmath70 if @xmath71 and @xmath72 otherwise ; then , @xmath61 reduces to proximity score in random walk models @xcite . *
if @xmath73 , where the scalar @xmath74 is the attenuation factor along the @xmath3-th link in the path , then @xmath61 reduces to @xmath0-centrality score from @xmath2 to @xmath3 ( normalizing leads to normalized @xmath0-centrality ) . for ease of computation ,
we have taken @xmath75 and @xmath76 , @xmath77 .
@xmath0-centrality holds for @xmath78 ( @xmath13 is the largest eigenvalue of @xmath7 ) .
however , as shown in the appendix , normalized @xmath0-centrality holds for all values of @xmath0 . *
when @xmath79 , this in turn reduces to the katz status score @xcite . *
if @xmath80 , and adjacency matrix @xmath7 is symmetric , then as @xmath81 , @xmath82 is proportional to the inner product of the eigenvector corresponding to @xmath83 , with itself .
it would lead to eigenvector centrality as shown in appendix .
* when @xmath84 and @xmath85 , then @xmath61 is the degree centrality used in modularity - maximization approaches @xcite .
we have implemented @xcite a simple algorithm to compute @xmath17 using an alternative formulation of @xmath0-centrality : @xmath86 for any given value of @xmath0 , this method iteratively computes @xmath87 and consequently @xmath88 until convergence .
experimentally we have observed , that this method reaches convergence very quickly . considering a network with @xmath89 nodes and @xmath90 links , each iteration of this algorithm ( for a given value of @xmath0 ) has a runtime complexity of @xmath91 and space complexity of @xmath92 . in order to study variation in this metric for @xmath93 , we must choose a step size of order @xmath94 , since by the gershgorin circle theorem , @xmath95 , where @xmath96 and @xmath97 are the maximum out- and in - degree of the network respectively .
since the formulation of normalized @xmath0-centrality is very similar to that of pagerank , similar block based strategies can be used for fast and efficient computation of both pagerank and @xmath17 @xcite . like pagerank , normalized @xmath0-centrality
can easily be implemented using the map - reduce paradigm @xcite , guaranteeing the scalability of this algorithm .
we apply the formalism developed above to benchmark networks studied in literature and a network extracted from the social photosharing site flickr . [
cols="^,^,^ " , ] [ tbl : mi ] the number and purity of the communities found in networks as a function of the parameter @xmath0 are shown in table [ tbl : mi ] .
the case @xmath19 corresponds to edge - based modularity method .
as @xmath0 increases , the number of groups discovered in all networks goes down , while their purity increases .
this is consistent with our hypothesis that using smaller values of @xmath0 allows us to identify more local network structure , while larger values of @xmath0 lead to more global structure . in the karate club network , for example , at @xmath19 , there are four small communities , as shown in fig .
[ fig : karate](a ) .
these local communities coalesce into two large groups as @xmath0 increases ( fig .
[ fig : karate](c ) ) , which are identical to the groups identified by zachary @xcite . to evaluate the communities discovered in the florentine families network , we use the tight constraint of party loyalty .
hence the families could be either medicean , oligarch , or have split loyalties .
we note that this is a very conservative evaluation criterion , since there were factions present within the parties themselves @xcite .
since families with split loyalties would be correctly classified as belonging to either of the two parties , we remove them from purity calculation , focusing instead on identifying community of party loyalists only .
purity is further reduced by the presence of the two isolated groups .
however , purity of discovered communities increases with @xmath0 .
the small local communities found at lower value of @xmath0 could indicate factions within parties .
a variety of metrics have been proposed to measure node s centrality in a network @xcite , yet few studies systematically evaluated their performance on real - world networks .
liben - nowell and kleinberg @xcite compared the performance of several commonly used centrality metrics on the link prediction task and found katz score @xcite to be the most effective measure for this task , outperforming pagerank @xcite and its variants .
the @xmath0-centrality metric modifies the katz score by introducing a parameter @xmath0 , that gives a weight to indirect links and also sets the length scale of interactions in the network .
we showed recently @xcite that normalized @xmath0-centrality outperforms other centrality metrics on the task of predicting influential nodes in an online social network .
guimera and collaborators @xcite proposed role - based description of complex networks .
they define a role in terms of the relative within - community degree @xmath98 ( which measures how well the node is connected to other nodes in its community ) and participation coefficient @xmath99 ( which measures how well the node is connected to nodes in other communities ) .
they proposed a heuristic classification scheme based on where the nodes lie in the @xmath98@xmath99 plane .
this classification scheme is similar to the local vs. globally - connected distinction we are making , with connector nodes being more globally connected nodes while provincial hubs and peripheral nodes are more locally connected .
role - based analysis requires community decomposition of the network to be performed first .
this is a computationally expensive procedure for most real - world networks .
our approach , on the other hand , allows us to differentiate between roles of nodes in a more computationally efficient way .
community detection is another active area in networks research ( see @xcite for a comprehensive review ) . like us , arenas et al .
@xcite have generalized modularity to find correlations between nodes that go beyond nearest neighbors .
their approach relies on the presence of motifs @xcite , i.e. , connected subgraphs such as cycles , to identify communities within a network .
for example , higher than expected density of triangles implies presence of a community , and a triangle modularity may be defined to identify it .
the motif - based modularity uses the size of the motif to impose a limit on the proximity of neighbors .
our method , on the other hand , imposes no such limit .
the measure of global correlation computed using @xmath0-centrality is equal to the weighted average of correlations for motifs of different sizes .
our method enables us to easily calculate this complex term .
in this paper , we introduced normalized @xmath0-centrality as a metric to study network structure .
like the original @xmath0-centrality @xcite on which it is based , this metric measures the number of paths that exist between nodes in a network , attenuated by their length with the attenuation parameter @xmath0 .
this parameter sets the length scale of the interaction .
when @xmath19 , the centrality metric takes into account direct edges only and is equivalent to degree centrality . as @xmath0 increases
, the metric takes into consideration more distant network interactions , becoming a more global measure .
normalized @xmath0-centrality allows us to smoothly interpolate between local metrics , such as degree centrality , and global metrics , such as eigenvector centrality @xcite .
unlike the original @xmath0-centrality , which bounds @xmath0 to be less than the reciprocal of the largest eigenvalue of the adjacency metric of the network , normalized @xmath0-centrality sets no such limit .
we used normalized @xmath0-centrality to study the structure of networks , specifically , identify important nodes and communities within the network .
we extended the modularity maximization class of algorithms @xcite to use ( normalized ) @xmath0-centrality , rather than edge density , as a measure of network connectivity . for small values of @xmath0 smaller , more locally connected communities
emerge , while for larger values of @xmath0 , we observe larger globally connected communities .
we also used this metric to rank nodes in a network . by studying changes in rankings that occur
when parameter @xmath0 is varied , we were able to identify locally important ` leaders ' and globally important ` bridges ' or ` brokers ' that facilitate communication between different communities .
we applied this approach to benchmark networks studied in literature and found that it results in network division in close agreement with the ground truth .
we can easily extend this definition to multi - modal networks that link entities of different types , and use approach described in this paper to study the structure of such networks @xcite .
this work is supported in part by the nsf under award 0915678 and in part by afosr .
in this section we prove some properties of the normalized @xmath0-centrality metric proposed in this paper .
the _ @xmath0-centrality matrix _
@xmath100 @xmath101 $ ] is defined as : @xmath102 the _
normalized @xmath0-centrality matrix _ is then given by : @xmath103 the normalized @xmath0-centrality vector is @xmath104 where @xmath105 is a @xmath106 unit vector and @xmath89 is the number of nodes in the network . if @xmath107 is an _ eigenvalue _ of @xmath7 , then @xmath108 invertibility of @xmath109 would lead to the trivial solution of eigenvector @xmath110 ( @xmath111 ) .
hence for computation of eigenvalues and eigenvectors , we require that no inverse of @xmath109 should exist , i.e. @xmath112 equation ( [ eq : char_eq ] ) is called the _ characteristic equation _ , solving which gives the _ eigenvalues _ and _ eigenvectors _ of the adjacency matrix @xmath7 .
the adjacency matrix @xmath7 can be written as : @xmath113 where @xmath114 is a matrix whose columns are the eigenvectors of @xmath7 , and @xmath115 is a diagonal matrix whose diagonal elements are the eigenvalues of @xmath7 , @xmath116 , arranged according to the ordering of the eigenvectors in @xmath114 . without loss of generality we assume that @xmath117 . the matrices @xmath118 can be determined from the product @xmath119 where @xmath120 is the _ selection matrix _ having zeros everywhere except for element @xmath121 @xcite .
adjacency matrix @xmath7 raised to the power @xmath122 is then given by @xmath123 using equation ( [ eq : lambda_k ] ) , [ eq : bc_eq ] reduces to @xmath124 where @xmath125 if @xmath126 , and @xmath127 if @xmath128 .
for the equations [ eq : bc_eq ] and [ eq : cm ] to hold non - trivially , @xmath129 .
1 . @xmath132 : if @xmath132 , @xmath87 ( and @xmath17 ) would be independent of @xmath0 , since @xmath133 2 .
@xmath134 : the sequence of matrices @xmath135 _ converges _ to @xmath136 as @xmath137 if all the sequences @xmath138 for every fixed @xmath2 and @xmath3 converge to @xmath139 @xcite . if @xmath134 , @xmath140 _ converges _ to @xmath141 .
@xmath142 3 .
@xmath143 and @xmath144 , @xmath145 dominates in the equation ( [ eq : cm ] ) .
@xmath146 since centrality score due to @xmath0-centrality is @xmath148 and that due to normalized @xmath0-centrality is @xmath149 , from equations [ eq : b_alpha1 ] and [ eq : b_alpha ] , the induced ordering of nodes due to @xmath0-centrality ( @xmath147 ) would be equal to induced ordering of nodes due to normalized @xmath0-centrality ( @xmath147 ) . under the assumption that @xmath155 is strictly greater than any eigenvalue , as @xmath158 , equation ( [ eq : b_alpha ] ) reduces to @xmath159
this is because all other eigenvectors shrink in importance as @xmath158 @xcite .
therefore , as @xmath158 , we have @xmath160 under the assumption that @xmath155 is strictly greater than any other eigenvalue , @xmath161 dominates in the equation ( [ eq : cm ] ) , [ eq : maxa ] .
@xmath162 for symmetric matrices @xmath171 therefore equation [ eq : z ] reduces to @xmath172 where @xmath173 is the column of @xmath114 representing the eigenvector corresponding to @xmath174 . hence , in case of symmetric matrices : @xmath175 where @xmath176 and @xmath177 . since @xmath178 corresponds to the eigenvector centrality vector @xmath169 ,
hence for symmetric matrices , the induced ordering of nodes given by eigenvector centrality @xmath169 is equivalent to the induced ordering of nodes given by normalized centrality @xmath179 . | a variety of metrics have been proposed to measure the relative importance of nodes in a network .
one of these , @xmath0-centrality @xcite , measures the number of attenuated paths that exist between nodes .
we introduce a normalized version of this metric and use it to study network structure , specifically , to rank nodes and find community structure of the network . specifically , we extend the modularity - maximization method @xcite for community detection to use this metric as the measure of node connectivity .
normalized @xmath0-centrality is a powerful tool for network analysis , since it contains a tunable parameter that sets the length scale of interactions . by studying how rankings and discovered communities change when this parameter is varied allows us to identify locally and globally important nodes and structures .
we apply the proposed method to several benchmark networks and show that it leads to better insight into network structure than alternative methods . |
preparing a desired pure state is a crucial task in quantum information technologies .
however , in a realistic situation any quantum system unavoidably interacts with the surrounding environment and is often described by the following markovian master equation : @xmath0 \nonumber \\
+ \sum_{i=1}^m & \left ( l_i\rho(t ) l_i^\dagger -\frac{1}{2 } l_i^\dagger l_i
\rho(t ) -\frac{1}{2}\rho(t ) l_i^\dagger l_i \right ) , \end{aligned}\ ] ] where @xmath1 is the system hamiltonian and @xmath2 is the coupling operator representing the interaction between the system and the @xmath3th environment channel .
usually the state @xmath4 in eq .
dissipatively evolves in time towards a mixed state and never recover its purity .
however , it has been shown in several papers @xcite that by engineering suitable pairs of the system operators @xmath1 and @xmath2 , the state @xmath4 can be uniquely moved to a pure steady state .
this means that a desired pure state can be deterministically generated without specific initialization of the system .
we actually find some applications of this environment engineering approach to entangled state generation @xcite and further some advanced quantum information processing such as quantum computation @xcite , memories @xcite , and distillation @xcite .
denotes the @xmath3th oscillator .
( b ) extended system that yields the same pure gaussian steady state as that produced by the system in ( a ) .
the auxiliary system couples to the environment locally , whereas it couples to the target system in a quasilocal manner through a certain hamiltonian . ,
width=302 ] here we mention the quasilocal interaction
. an operator is called quasilocal if it acts only on a given site of the system and its neighborhood ; particularly in this paper we call the interaction quasilocal if it occurs between only two nodes . in the case of a quantum oscillator network
whose @xmath3th node is a single - mode oscillator of variable @xmath5 , an example of a quasilocal coupling is @xmath6 .
however , to generate a desired pure state such as a highly entangled state via the above - mentioned dissipation - based method , the system is often required to have non - quasilocal ( global ) interaction with the environment .
let us consider a seven - node quantum oscillator network depicted in fig .
this network has internal couplings among the nodes @xmath7-@xmath8 and @xmath9-@xmath10 through a hamiltonian @xmath1 in addition two global couplings of the nodes @xmath11 and @xmath12 to the environments , the operators of which are , for instance , @xmath13 and @xmath14 . from a practical viewpoint , clearly
this kind of global interaction is hard to implement .
actually , there have been several proposals to engineer a dissipative system having a quasilocal interaction with the environment , which yet deterministically produces a useful pure state . for instance , it was shown in ref @xcite that a qubit two - dimensional cluster state can be generated via a quasilocal dissipative process and ref . @xcite demonstrated the case of a uniformly distributed entangled state .
moreover , for general finite - dimensional systems , ticozzi and viola provided a condition to determine whether a given pure state can be stabilized under fixed locality constraints @xcite .
however , in the infinite - dimensional case , only a few specific results have been reported @xcite .
the result obtained in ref @xcite is that for any pure gaussian state we can always engineer an open gaussian system whose unique steady state is identical to that pure state .
however , as mentioned above , this open gaussian system often has to have a global interaction with the environment , such as the system shown in fig .
1(a ) . in this paper ,
nonetheless , we show that an arbitrary pure gaussian state can be dissipatively generated in a certain extended system having only local interaction with the environment .
the idea is described as follows .
first , we couple the underlying target system , via a set of two - body hamiltonians , to a certain auxiliary system having local interactions with the environment .
that is , the extended open system composed of the target and the auxiliary systems contains only quasilocal interactions [ see fig . 1 ( b ) ]
then , based on the result of ref @xcite , we will show that a gaussian pure state generated in the original system of interest , which can include global interactions with the environment , is also generated in the above - mentioned extended system .
this configuration is actually a generalization of the result of ref @xcite , where dissipation - induced generation of a two - mode squeezed state in atomic ensembles was demonstrated .
in addition , we present a simple schematic that only utilizes a single - mode auxiliary system .
more specifically , even in the case where the above - mentioned result requires us to have a multimode auxiliary system , it will be shown that only a single - mode auxiliary system can serve to achieve the same goal .
the key idea is the use of a switching scheme of the auxiliary system ; that is , in the case of the system depicted in fig . 1 , a single - mode auxiliary system first couples to the subsystems @xmath15 and then switches to couple to the subsystems @xmath16 .
a particularly important result is that , with this switching scheme , an arbitrary approximate gaussian cluster state @xcite can be generated dissipatively .
this is a generalization of the switching scheme proposed by li _ et al . _ in ref @xcite , which demonstrated dissipative generation of several types of four - mode cluster states in an actual physical setup .
finally , we examine how much a specific dissipation - induced pure state is robust against some perturbation .
the system is a pair of two atomic ensembles that dissipatively generates a two - mode squeezed state @xcite ; here damping decoherence and parameter uncertainty are further taken into account .
then we find the optimal system parameters that maximize the entanglement between the atomic ensembles .
a point worth noting is that these optimal parameters are out of the range where the auxiliary cavity mode can be adiabatically eliminated ; that is , this is an example where the extended system is really robust against some perturbation compared to only the target system obtained through the adiabatic elimination . _
notation_. let @xmath17 be a matrix whose entry @xmath18 is an operator on a hilbert space or a complex number .
then @xmath19 denotes the transpose of @xmath20 .
also , @xmath21 means the hermitian conjugate of @xmath20 ; i.e. , @xmath22 for an operator @xmath18 and @xmath23 for a complex number @xmath18 .
we use @xmath24 and @xmath25 to denote the real and imaginary parts of @xmath20 . for a vector of operators
@xmath26 , we define the anti - commutator of @xmath26 by @xmath27
in this section , we describe a general gaussian dissipative system in terms of a quantum stochastic differential equation ( qsde ) and present the condition for the steady state of this system to be pure @xcite .
the time evolution of a general open quantum system is generated by the unitary operator @xmath28 subjected to the following qsde ( in ito form ) @xcite : @xmath29 with @xmath30 ( we have not included the scattering term ) . here
@xmath31 is the annihilation process on the @xmath3th environment vacuum field ; this satisfies the quantum it rule , e.g. , @xmath32 .
let @xmath33 be the initial state of the system and @xmath34 be the vacuum state of the whole environment field .
the master equation is obtained by differentiating @xmath35 and tracing out the field state .
next , let @xmath36 be the canonical conjugate pair of the @xmath3th quantum oscillator , which satisfies the canonical commutation relation ( ccr ) @xmath37 . defining the vector of observables @xmath38^\top$ ] ,
we can write the ccr as @xmath39 where @xmath40 denotes the @xmath41 identity matrix ( we often drop the subscript @xmath42 ) . a general linear open system can be characterized by the following hamiltonian @xmath1 and the coupling operators @xmath43^\top$ ] : @xmath44 where @xmath45 and @xmath46 . from eq .
, the qsde of @xmath47^\top$ ] is given by @xmath48 where @xmath49 and @xmath50 with @xmath51^\top$ ] .
also , we have defined the field operators @xmath52 , @xmath53 , and @xmath54^\top$ ] .
a gaussian system can be fully characterized by only the mean vector @xmath55^\top $ ] and the covariance matrix @xmath56 with @xmath57 .
note that a gaussian state is pure if and only if @xmath58 . the importance of the linear system is in the fact that , if the initial state of this system is gaussian , then it preserves the gaussianity of the state for all time . in particular , the covariance matrix @xmath59 obeys the following lyapunov differential equation : @xmath60 hence the covariance matrix of the steady state is given by the solution of the algebraic lyapunov equation @xmath61 this equation has a unique solution if and only if @xmath62 is a hurwitz matrix ; i.e. , all eigenvalues of @xmath62 are in the open left half complex plane .
now we are interested in a linear gaussian system whose steady state is uniquely pure . it has been shown in ref .
@xcite that an arbitrary pure gaussian state can be dissipatively generated , if the system matrices @xmath63 and @xmath64 can be freely chosen .
in particular , the following result is useful in this paper : _ theorem 1 @xcite_. suppose that eq . has a unique solution @xmath65 .
then , this is the covariance matrix of a pure gaussian state if and only if the following matrix equations are satisfied : @xmath66
to generate a certain desirable pure gaussian state dissipatively , the system is often required to have a global interaction with the environment ; that is , for instance , the coupling operator must be of the form @xmath67 , which corresponds to @xmath68 . in this section , we show that such a global interaction between the system and the environment can always be avoided by constructing an auxiliary system that quasilocally couples to the target system and locally couples to the environment .
in addition to the underlying target system consisting of @xmath42 quantum oscillators , we consider an auxiliary system consisting of @xmath69 oscillators .
let @xmath38^\top$ ] and @xmath70^\top$ ] be the vectors of canonical conjugate pairs of each system .
we assume that the decoherence of the target system is negligible , i.e. , the system operators are @xmath71 where @xmath45 .
the auxiliary system has no self - hamiltonian and dissipates through typical damping channels , hence the system operators are given by @xmath72 where @xmath73^\top$ ] is a vector of annihilation operators , i.e. , @xmath74 .
the two systems couple via the following interaction hamiltonian : @xmath75 with @xmath76 .
note that @xmath77 describes a quasilocal interaction between the two systems , because it can always be decomposed into a sum of two - body hamiltonians as follows : @xmath78 where @xmath79^\top$ ] and @xmath80^\top$ ] .
we will see later that this quasilocal hamiltonian has the same effect as the coupling operator of the system originally of interest , i.e. , @xmath81 , when choosing @xmath63 and @xmath64 so that they satisfy the condition of theorem 1 . defining @xmath51^\top$ ] , we can rewrite eq . as @xmath82 then , the overall system vector @xmath83^\top$ ] obeys the following linear qsde : @xmath84 where @xmath85 the covariance matrix @xmath86 evolves in time through eq . and its steady solution , if it exists , is obtained by solving eq . . to avoid confusion
, we again present the same equation @xmath87 where @xmath62 and @xmath88 are now given in eq . .
recall here that we were originally interested in the system with the hamiltonian @xmath89 and the coupling operator @xmath81 , which thus obeys the linear dynamics @xmath90 where @xmath91 and @xmath92 .
note that the dissipation channel @xmath81 can be global .
its steady covariance matrix in particular is our concern , which is subjected to @xmath93 the following theorem states that , if the system has a pure steady state , then the extended system can dissipatively produce the same pure gaussian state . _
theorem 2_. suppose that eq . has a unique solution @xmath94 that corresponds to a pure gaussian state .
. has a unique solution @xmath95 .
first , we prove that eq .
has a unique solution , which is equivalent to that @xmath62 is a hurwitz matrix . for this purpose ,
let us set @xmath96 ; i.e. , @xmath97 is an eigenvector of @xmath98 and @xmath99 is the corresponding eigenvalue .
then , multiplying eq .
by @xmath100 from the left and by @xmath101 from the right , we have @xmath102 .
this further becomes @xmath103 , where we have defined @xmath104^\top$ ] with @xmath105 and @xmath106 .
now let us assume @xmath107 ; then , from @xmath108 we obtain @xmath109 moreover , noting the assumption that eq . has a unique solution @xmath110 we find that @xmath111 satisfies @xmath112 however , this takes zero due to eq . and
the cayley - hamilton theorem .
this conclusion contradicts @xmath113 , hence we have @xmath114 ; this further leads to @xmath115 , implying that @xmath62 is a hurwitz matrix .
next let us prove that @xmath116 is the solution of eq . .
when representing @xmath65 in a block matrix form @xmath117 $ ] , these entries satisfy @xmath118 now the assumption is that @xmath94 corresponds to a unique pure steady state , hence from theorem 1 we have @xmath119 note that the former condition comes from the fact that theorem 1 yields @xmath120 and @xmath121 .
then , we find that the set of matrices @xmath122 , @xmath123 , and @xmath94 satisfying eq .
is the solution to eq . .
because eq . has a unique solution as shown in the former part of this proof , @xmath95 is the unique solution .
this theorem states that , once we find a suitable pair of matrices @xmath63 and @xmath64 such that the system dissipatively and uniquely generates a desired pure gaussian state , the same goal can be achieved by alternatively constructing the extended system .
note again that the system can couple to the environment globally , while the extended system locally couples to the environment through the operator @xmath124 and its internal modes quasilocally couple with each other through the hamiltonian @xmath125 .
that is , the target system stabilizing a desired pure state can be realized as a subsystem of an extended system having local coupling to the environment and quasilocal internal couplings among the nodes . _
remark 1_. as an auxiliary system , we usually take a system with very fast modes that can be adiabatically eliminated . in our case , even when the assumption of theorem 2 does not hold , by taking @xmath126 sufficiently large , the auxiliary system rapidly converges to the vacuum and the mode @xmath127 can be adiabatically eliminated .
then , the target system is approximated by the system whose coupling operator is @xmath128 @xcite . in this sense ,
theorem 2 implies that we can treat the system as if the auxiliary modes were heavily damped as long as the pure steady state condition is satisfied .
we consider here a two - mode gaussian system studied in ref @xcite .
although for this system the requirement of quasilocality is already satisfied , this example clearly illustrates our idea .
the physical setup is depicted in fig .
[ fig : par ] ; the target system is two atomic ensembles trapped in a ring - type cavity , while the auxiliary system corresponds to this optical cavity composed of two propagating modes .
these two systems interact with each other via external pulse lasers with rabi frequencies @xmath129 and @xmath130 @xmath131 . here
we assume that the number of atoms in each ensemble is sufficiently large ; then the collective spin component of the atomic ensemble can be approximated by an annihilation operator and consequently the interaction hamiltonian is given by @xmath132 where @xmath133 and @xmath134 are the annihilation operators of the @xmath3th atomic ensemble and the @xmath3th cavity mode , respectively .
here @xmath135 denotes the number of atoms of the @xmath3th ensemble , and also @xmath136 where @xmath137 and @xmath138 denote the coupling strength and the detuning , respectively .
if we set the parameters as @xmath139 , and @xmath140 , then the interaction hamiltonian can be simply written as @xmath141.\end{aligned}\ ] ] further , let us set the rabi frequencies as @xmath142 and @xmath143 , where @xmath144 is a parameter .
then , the interaction hamiltonian is of the form with @xmath145 where @xmath146 . for @xmath147rb atoms ,
the spontaneous emission of each atom is negligible and thus we can set @xmath148 , for typical values of the parameters , @xmath149 100 khz and @xmath150 @xcite . also , the number of atoms is large enough so that the self hamiltonian of the atomic ensembles is assumed to be 0 , i.e. , @xmath151 or equivalently @xmath152 .
the coupling operator of the cavity is given by @xmath153 , where @xmath126 is the damping rate , and we assume that the detuning of the cavity is 0 , i.e. , @xmath154 .
now , for the matrix @xmath64 given by eq . and
@xmath152 , the lyapunov equation has the following unique solution : @xmath155 where @xmath156 .
this is the covariance matrix of a pure two - mode squeezed state .
therefore , from theorem 2 , the lyapunov equation has a unique solution @xmath157 ; that is , the pair of atomic ensembles acquires the two - mode squeezed state at steady state , while the auxiliary cavity mode becomes a trivial coherent state .
equivalently , the whole four - mode atom - cavity system generates the same atomic steady state as that generated in the two - atomic- ensemble system with the hamiltonian @xmath151 and the coupling operators @xmath158 another physical realization of this purely dissipative system was proposed in refs .
@xcite . and
@xmath130 are the rabi frequencies of the @xmath3th external laser fields .
, width=207 ]
as shown in the preceding section , an arbitrary pure gaussian state can be generated in the target system , by introducing a certain @xmath69-mode auxiliary system that locally couples to @xmath69 environment channels . from an engineering viewpoint
, it is clearly convenient if we can achieve the goal with the auxiliary system having a small number of modes @xmath69 . in particular , let us consider the simple case @xmath151 ; i.e. , the system hamiltonian @xmath1 is negligible compared to the interaction hamiltonian , as in the case of sec . [
sec : epr ] . in this case
, we need the condition @xmath159 , which is understood from theorem 1 together with the fact that , for a covariance matrix @xmath65 corresponding to a pure gaussian state , @xmath160 has an @xmath42-dimensional kernel @xcite
. however , this requirement is very demanding especially when @xmath42 is large . in this section
we present a general solution to the above - posed problem .
that is , we prove that even for a general @xmath42-mode system with @xmath151 , only a single - mode auxiliary system coupling to a single environment channel introduces a dissipative mechanism that drives the system state to an arbitrary approximate gaussian cluster state @xcite .
this result is significant in the sense that a most simple dissipative system deterministically generates a most useful quantum state from the quantum information viewpoint .
the key idea is the use of a switching scheme of the interaction hamiltonian between the target and the auxiliary systems , which is a generalization of the idea proposed by li _
@xcite .
as introduced above , in this section we consider a most simple system ; that is , the system s hamiltonian is negligible ( @xmath151 ) and the auxiliary system is single - mode ( @xmath161 ) , in addition to the assumptions @xmath162 and @xmath163 .
we represent the interaction hamiltonian in the following form : @xmath164 where @xmath165 denotes the @xmath166th annihilation operator of the target system . @xmath167 and @xmath168 are parameters .
first let us consider the case @xmath169 and set the parameters as @xmath170 and @xmath171 .
then , the corresponding @xmath64 matrix in eq . is given by @xmath172/\sqrt{2}$ ] and eq .
has the following unique solution : @xmath173 where @xmath156 .
this is the covariance matrix of a pure squeezed state , hence from theorem 2 we find that eq .
has a unique solution @xmath174 . as a result ,
when @xmath169 , the target system acquires a pure squeezed state while the auxiliary state becomes vacuum , through the interaction hamiltonian .
next we consider the case of @xmath175 . in this case , as mentioned above , eq .
does not have a unique solution , hence theorem 2 can not be directly applied .
now let us consider a unitary matrix @xmath176 and the unitary coordinate transformation @xmath177 of the vector of system variables @xmath178^\top$ ] . in terms of the quadratures @xmath179 and @xmath180 , we find that @xmath181^\top$ ] is the symplectic and orthogonal transformation of @xmath26 as follows : @xmath182 note that , since @xmath183 is symplectic , @xmath184 satisfies the ccr : @xmath185 .
we explain here the idea of switching . at the @xmath186th switching stage ,
the parameters @xmath187 of the interaction hamiltonian are chosen as follows : @xmath188 where @xmath189 is the @xmath190 element of the unitary matrix introduced above . with this choice
, the interaction hamiltonian is written as @xmath191 this is no more than the hamiltonian discussed in the case @xmath169 .
hence , in the @xmath186th switching stage the @xmath186th mode @xmath192 deterministically changes to a pure squeezed state with covariance matrix .
note that during the @xmath186th stage , the other system variables @xmath193 do not change at all .
therefore , applying the interaction hamiltonian repeatedly by changing @xmath194 , the corresponding system variable @xmath195 gets squeezed in this order , as schematically shown in fig .
[ fig : cartoon2 ] . in particular ,
if the initial state is the ground state with respect to @xmath184 ( and thus @xmath26 as well ) , all the off - diagonal elements of the covariance matrix remain 0 and then the system s covariance matrix @xmath196 becomes @xmath197 . as a result ,
the steady state of the target system of the mode @xmath198 is a pure gaussian state with covariance matrix @xmath199 this is a unitary transformed pure squeezed state ; we will see later that this state can represent any approximate gaussian cluster state by appropriately choosing the unitary matrix @xmath200 or equivalently the switching parameters @xmath187 .
a gaussian cluster state is an entangled state of great importance , particularly in one - way quantum computation @xcite .
hence , this state should be a target that is dissipatively generated with the scheme presented in this paper , particularly with the switching scheme .
thus we show here how to chose the unitary matrix @xmath200 so that the covariance matrix represents a given target gaussian cluster state .
here is the definition of a gaussian cluster state @xcite : let @xmath201 be the adjacency matrix representing the graph structure of a cluster state of interest ; i.e. , the @xmath202 element of @xmath62 represents the weight of the coupling between the @xmath3th and the @xmath166th nodes of the network .
then the approximate gaussian cluster state is defined as a state satisfying @xmath203 where @xmath204 is a certain parameter contained in the state ; it often corresponds to a squeezing parameter .
also we have defined @xmath205 .
it is of course impossible in reality to take the limit @xmath206 , hence we call the state with finite @xmath207 the approximate gaussian cluster state .
now we describe a relation that connects a given adjacency matrix @xmath62 and the unitary matrix @xmath200 characterizing the covariance matrix .
as will be mentioned later in remark 2 , one such relation has already been obtained in refs .
@xcite particularly for the aim of constructing a concrete optical process corresponding to @xmath200 , but we here provide an alternative method in rather an abstract way . _
proposition 1_. [ pro2 ] let @xmath201 be the adjacency matrix of a given gaussian cluster state and define @xmath208 .
then the polar decomposition @xmath209 yields a unitary matrix @xmath200 and a real matrix @xmath210 .
the gaussian state with covariance matrix characterized by this unitary @xmath200 then satisfies @xmath211 hence it is an approximate gaussian cluster state converging to the idel one in the limit of @xmath212 .
first , @xmath210 is real because @xmath213 .
next , for the gaussian state with covariance matrix we have @xmath214\ , { \frac{1}{2}}s^t \begin{bmatrix } e^{-2\xi}i_n & 0 \\ 0 & e^{2\xi}i_n \end{bmatrix } s \begin{bmatrix } -a \\ i_n \end{bmatrix } \nonumber \\ & \hspace{1.8 cm } = { \frac{1}{2}}e^{-2\xi}f_1^\top f_1 + { \frac{1}{2}}e^{2\xi}f_2^\top f_2 , \end{aligned}\ ] ] where @xmath215 and @xmath216 .
noting that @xmath217 and @xmath210 is real , we have @xmath218 and @xmath219 .
hence we have @xmath220 and @xmath221 . furthermore , noting that @xmath222 , we have @xmath223 , thus @xmath224 . as a result , we have @xmath225 .
the merit of this result is that the unitary matrix @xmath200 is straightforwardly constructed from a given @xmath62 compared to the result of refs .
@xcite , although in this case @xmath200 does not have a clear correspondence to some optical realizations .
however , eq . clarifies how to physically implement the interaction hamiltonian . _
remark 2_. the approximate gaussian cluster state with covariance matrix can be deterministically generated on optical fields , by the following method @xcite : we prepare @xmath42 independent and identical squeezed light fields , and mix them via some passive optical devices such as a beam splitter and a phase shifter in a specific order determined from the adjacency matrix @xmath62 .
the collection of these transformations is totally represented by a unitary matrix ; if we denote that unitary matrix as @xmath226 , the covariance matrix of the output fields is identical to eq . .
that is , the unitary transform in the switching scheme corresponds to the scattering process on optical fields , and the dissipation - induced pure squeezed states correspond to the initially - prepared optical squeezed states .
hence , the presented switching scheme can be interpreted as a dissipative counterpart to the optical scheme proposed in refs .
the biggest difference between these two schemes is that in the dissipative case we consider a state generated in matter whereas the optical state exists in a flying light field ; the former can be later manipulated or stored , while the latter is suited for propagating quantum information .
a specific relation between these two regimes was discussed in ref .
@xcite . _
remark 3_. in proposition 1 the polar decomposition is used to find the appropriate unitary matrix @xmath200 , but it is clear from the proof that only a certain decomposition of the form @xmath227 with @xmath200 unitary and @xmath210 real gives the same relation .
here we see that , besides the polar decomposition , the gram schmidt procedure also serves as a convenient method to obtain such a decomposition .
let @xmath200 be a unitary matrix whose row vectors are obtained from the gram schmidt procedure of the row vectors of @xmath228 .
then we immediately have the relation @xmath227 with a real lower triangular matrix @xmath210 , which is called the rq decomposition .
hence the gram - schmidt procedure also yields the unitary matrix satisfying the condition in proposition 1 . and
@xmath130 are the rabi frequencies of the @xmath3th external laser fields.,width=226 ] we can implement the presented switching scheme for @xmath42 atomic ensembles ( target system ) trapped in a single - mode optical cavity ( auxiliary system ) in a similar configuration studied in sec .
[ sec : epr ] . the physical setup in the case @xmath229
is depicted in fig .
[ fig : setup2 ] .
let @xmath230 be the annihilation operator approximating the collective spin component of the @xmath166th atomic ensemble and @xmath231 the annihilation operator of the cavity mode .
then the interaction hamiltonian is given by @xcite @xmath232 , \end{aligned}\ ] ] where @xmath233 is the laser phase . the switching scheme shown in sec .
[ sec : switch ] suggests that , at the @xmath186th switching stage , we choose the parameters as @xmath234 where @xmath235 is a parameter .
the unitary matrix @xmath236 is determined from the target gaussian cluster state . as proven in sec .
[ switching scheme ] , the whole state of the atomic ensembles deterministically reaches this target , if they are all initially set to the ground states .
let us consider the problem of dissipatively generating a four - node square cluster state , depicted in fig .
[ fig : square ] , in the atomic ensemble system discussed in the preceding section .
the connecting edges between the nodes are equally weighted and thus the adjacency matrix @xmath62 of this graph state is given by @xmath237 we follow proposition 1 to determine a unitary matrix @xmath200 characterizing the switching law .
in particular , here we utilize the gram - schmidt procedure ( see remark 3 ) . to make the calculation simple ,
we orthogonalize the following matrix : @xmath238 then , @xmath200 is given by @xmath239/\sqrt{2 } \\
-i[0,0,1,-1]/\sqrt{2 } \\
-[i , i,2,2]/\sqrt{10 } \\
\mbox{}[2,2,i , i]/\sqrt{10 } \end{bmatrix}.\end{aligned}\ ] ] actually with this choice we have @xmath240 thus the state approximates well the target square cluster state when large @xmath241 is taken .
the switching law with the unitary matrix clarifies how to choose the laser parameters as shown below . at the first switching stage ,
they are determined from the first row vector of @xmath200 as @xmath242 through the interaction hamiltonian with these parameters , the ccr pair @xmath243 gets squeezed in the long time limit .
then we switch the parameters and set : @xmath244 then the second node @xmath245 gets squeezed . by repeating a similar procedure , each ccr pair @xmath246 , @xmath247 becomes a squeezed state with covariance matrix . as a result , if each atomic ensemble is in the ground state at the initial time , the whole state changes to a pure gaussian state with covariance matrix , implying that the system reaches the target approximate gaussian cluster state satisfying eq . .
all of the above results , including the switching law of the parameters of the interaction hamiltonian , have been obtained in ref .
@xcite ; here we have shown the same result from a general standpoint .
that is , from the general theory developed in secs .
[ switching scheme ] and [ cv cluster state ] , now we know that any approximate gaussian cluster state can be deterministically generated in atomic ensembles trapped in a single - mode cavity .
moreover , as demonstrated here , the appropriate switching law can be systematically constructed , once the target cluster state is specified .
in this section , we reconsider the two atomic ensembles discussed in sec . [
sec : epr ] , taking into account some specific perturbations added to the system .
actually , an atomic ensemble constructed in a cavity often loses coherence due to spontaneous emission .
in addition to this kind of quantum effect , in practice any system contains some parameter uncertainties , which can bring a serious loss of coherence as well when aiming to dissipatively generate a pure state . in the case of an atomic ensemble ,
the number of atoms is usually never determined exactly .
we take these two typical perturbations and evaluate how much the steady state is affected by these losses . in particular , we find the optimal squeezing level and the cavity damping rate that maximize the entanglement .
although these investigations do not straightforwardly provide new insight into the quasilocality discussed throughout this paper , the result will clarify a merit of enlarging the system rather than focusing only on the target system obtained by adiabatic elimination . versus the squeezing level @xmath156 and the cavity damping rate @xmath126 for the cases ( a ) @xmath248 , ( b ) @xmath249 , and ( c ) @xmath250 .
, width=226 ] first we assume that the atomic ensembles are subjected to a loss of atomic coherence , in which case the corresponding coupling operator is represented by @xmath251^\top$ ] , with @xmath252 the decoherence rate .
the coefficient matrix of the driving term of the system , given in eq .
, is then changed to @xmath253 where @xmath254 is defined via eq . .
the eigenvalues of @xmath62 are given by @xmath255 a notable point is that , when large squeezing is introduced , i.e. , @xmath256 , the eigenvalue @xmath257 approaches zero , even in the case @xmath258 .
this loss of stability of the system implies that a desirable convergence of the state is prevented .
thus the perturbation @xmath259 is not negligible , particularly when aiming to generate a large entangled state .
now the coefficient matrix of the diffusion term of the system is @xmath260 and the lyapunov equation @xmath261 has the following explicit solution : @xmath262 where @xmath263 . here , for simplicity , we have replaced @xmath264 and @xmath265 by @xmath126 and @xmath252 , respectively .
the entanglement of this two - mode gaussian state can be quantified by the logarithmic negativity @xcite and is given by @xmath266 } { 2(\kappa + \gamma)[\kappa \gamma + 4(1-r^2)]}. \nonumber\end{aligned}\ ] ] figure [ fig : neg ] shows @xmath267 versus the squeezing level @xmath156 and the cavity damping rate @xmath126 , for the cases where the decoherence rate @xmath252 takes the values ( a ) @xmath248 and ( b ) @xmath249 . in the former case ( a ) , since there is no decohering effect , the steady state is pure with maximal entanglement @xmath268 without respect to @xmath126 ; i.e. , large squeezing directly means large entanglement .
however , once the system is subjected to the decoherence , the steady - state entanglement drastically degrades , even for a very small decoherence rate @xmath249 , as depicted in fig .
[ fig : neg ] ( b ) .
in particular , it is apparent from the figure that large squeezing brings about a large loss of entanglement , which is yet consistent with the fact that a large squeezed state is usually very fragile to decoherence .
it is also reasonable that now @xmath126 affects @xmath267 in two ways : large @xmath126 induces rapid leaking of the photons through the cavity light field , while small @xmath126 means that the cavity mode interacts with the atomic ensembles many times , implying many emissions of the photons into the cavity field .
consequently , there exists an optimal set of the parameters @xmath126 and @xmath269 that maximizes @xmath267 : @xmath270 where @xmath271^{1/3}$ ] . particularly in the case
@xmath249 , these optimal values are given by @xmath272 and @xmath273 , and @xmath267 then takes the value of about @xmath274 . _
remark 4_. it should be pointed out that the above optimal parameters are out of the range where the auxiliary cavity mode can be adiabatically eliminated ; that is , as discussed in remark 1 in sec .
[ sec : extended ] , the auxiliary system can be adiabatically eliminated only when @xmath126 is sufficiently large , but clearly in this case @xmath267 goes down to zero . in this sense ,
the system studied here provides an example where the extended system is really robust against decoherence compared to only the target system obtained through adiabatic elimination .
since the condition imposing the system to have a pure steady state is described by a set of algebraic equations , it is easily violated by some parameter changes . in the case of the atomic system under consideration
, the numbers of atoms of each ensemble must be exactly the same , i.e. , @xmath275 , but it is fairly unrealistic . hence let us examine here how much the difference between @xmath276 and @xmath277 affects the entanglement of the steady state .
we particularly set @xmath278 ; as usual , the number of trapped atoms is of order @xmath279 , and such relatively large uncertainty ( 10@xmath280 ) can actually happen .
figure [ fig : neg]-(c ) shows the logarithmic negativity @xmath267 , where the decoherence due to the spontaneous emission discussed in the preceding section is additionally taken into account . as expected , further degradation of the entanglement is observed and in almost all ranges of the parameters @xmath241 and @xmath126 the state is no longer entangled .
nevertheless , surprisingly , it is not a uniform degradation ; actually , as in the previous case , by engineering the system with optimal parameters @xmath281 , we obtain a steady state that still has a relatively large entanglement of @xmath282 , which is only a 17@xmath280 loss of entanglement compared to the ideal value of @xmath283 , where there is no uncertainty ( i.e. , @xmath284 ) . in other words , by constructing the system with these parameters @xmath281 , we can guarantee the entangled state with at least @xmath282 , against the uncertainty of the difference of the number of atoms up to 10@xmath280 .
this robustness property indicates the possible effectiveness of the dissipation - based method for state preparation even in a realistic situation , as long as the system parameters are appropriately determined .
the main results of this paper are twofold : first , we have shown that an arbitrary gaussian pure state can be deterministically generated via the local dissipative environment , by constructing an appropriate auxiliary system .
second , we have shown that , even when only a single - mode auxiliary system is available , a well - tuned switching scheme allows us to stabilize any approximate gaussian cluster state in a dissipative way .
the former is a generalization of the scheme proposed in ref .
@xcite that yields a dissipation - induced two - mode squeezed state , while the latter is that of ref .
@xcite where deterministic generation of several four - mode cluster states was demonstrated .
the essential mechanism for bringing quasilocality to the system considered in this paper is that , for a gaussian system , any interaction hamiltonian is always a sum of two - body ( hence quasilocal ) hamiltonians .
this implies that as long as this kind of interaction hamiltonian is taken when constructing an auxiliary system , any non - gaussian ( target ) system couples to the environment quasilocally .
this would be an interesting approach to explore a general method of constructing a desired quasilocal dissipative environment for general non - gaussian systems . | this paper shows that an arbitrary gaussian pure state can be deterministically generated in a dissipative open system that has quasilocal interactions between the subsystems and couples to the surrounding environment in a local manner .
a quasilocal interaction , which means that the interaction occurs among only a few subsystems , is a crucial requirement for practical engineering of a dissipative system .
the key idea is that first an auxiliary system having a local interaction with the environment is prepared and then that auxiliary system is coupled to the underlying target system via a set of two - body hamiltonians in such a way that a desired pure state is generated .
moreover , we show that even with a simple single - mode auxiliary system , the deterministic generation of an arbitrary approximate gaussian cluster state is possible , by devising an appropriate switching scheme .
we discuss in a specific example how much a dissipation - induced pure gaussian state can be perturbed by decoherence and parameter error . |
controlling the self - assembly of colloidal particles is an important aspect of pure and applied colloid science . in this context , assembling novel structures in liquid crystal matrices has become a very active field of research as a result of important theoretical and experimental advances @xcite .
liquid crystals ( lcs ) are characterized by anisotropic mechanical and optical properties due to the long - range orientational molecular ordering .
consequently , in addition to the usual isotropic colloidal interactions , colloids in liquid crystals interact through strongly anisotropic effective forces , and have been observed to self - assemble into ordered aggregates in two ( @xmath1 ) @xcite and three ( @xmath0 ) spatial dimensions @xcite .
a distinctive feature of these systems is the presence of topological defects , the type of which determines the symmetry of the far - field director around an isolated colloidal particle , which in turn governs the asymptotic behavior of the nematic - induced colloidal interactions @xcite .
the behavior of the interactions at short distances is more complex since the defects can move when the colloidal particles are brought close together , rendering the description of the self - assembly of liquid crystal colloids a challenging theoretical problem .
theoretical and simulation methods have been developed to study lcs on various scales .
on the macroscopic scale , the long - range forces between colloidal particles can be calculated using the frank - oseen ( fo ) elastic free energy approach @xcite . on the mesoscopic scale where non - linear effects become important , a description based on the landau - de gennes ( ldg ) tensorial order - parameter theory is necessary @xcite . this is the case when the interactions between the defects dominate , as revealed by the changes in their equilibrium structure when the distance between the particles decreases @xcite . on the microscopic scale ,
presmectic ordering and density variations can not be ignored and density functional approaches @xcite or computer simulations are usually employed @xcite .
the anchoring at the surfaces of colloidal particles can be controlled , for instance , through the adsorption of surfactants or deposition of thin organic or inorganic films @xcite .
when the anchoring on a spherical colloid is planar , a tangential nematic director configuration is expected at the colloidal surface . at the fo level
the boundary conditions are met by the creation of a pair of antipodal surface topological defects , called boojums @xcite .
the boojums are aligned with the far - field director , ensuring that the far - field lc configuration is homogeneous as required by the global uniform boundary conditions @xcite . recently ,
experimental measurements of the effective forces between two spherical colloids with planar anchoring in a @xmath0 nematic have been reported @xcite .
the authors have found that the forces at short distances deviate considerably from the asymptotic quadrupole - quadrupole ones , and that the equilibrium configuration of a pair of particles corresponds to close contact implying the absence of a repulsion at short distances .
by contrast , experimental results obtained for inclusions in free - standing smectic @xmath2 films @xcite revealed the presence of a short distance repulsion which keeps the inclusions at a well - defined equilibrium separation larger than contact . again , as in the 3d system of ref .
@xcite , the asymptotic quadrupole - quadrupole approximation for the pair - wise force breaks down at short distances . a recent theoretical study based on the numerical minimization of the @xmath0 ldg free energy functional @xcite is in line with the experimental results of ref .
@xcite concerning the equilibrium configuration of two colloidal spheres and the absence of colloidal short - range repulsion
. theoretical analysis of the @xmath1 fo model @xcite , based on an _ ad hoc _
ansatz for a single particle solution and the superposition approximation , have correctly described the configuration of two inclusions in a smectic @xmath2 film at short distances , but have failed to provide a consistent mechanism for the short - range repulsion observed experimentally .
a subsequent superposition analysis @xcite , based on the exact single particle solution @xcite , has predicted the existence of two regimes as a function of the anchoring strength . for weak anchoring the inclusions coalesce while for strong anchoring , the effective interaction exhibits a well defined minimum and a strong short - distance repulsion that prevents coalescence @xcite . although the @xmath0 studies of ref .
@xcite are in qualitative agreement with the experimental results of ref .
@xcite , the question of the structure of the boojum cores remains open .
the authors of ref .
@xcite could not resolve the structure the defect cores because the meshes which have been used in the minimization of the ldg free energy functional were too coarse . in the @xmath1 systems , studied in refs .
@xcite , the theoretical results are far from satisfactory since the superposition approximation is not valid at short distances where the interaction between defects is expected . motivated by these open questions , and by the experimental results which reveal a qualitative difference between @xmath1 and @xmath0 systems @xcite , we address in this article the issues of the structure of the boojum cores and the effective interactions between circular and spherical colloids in @xmath1 and @xmath0 , respectively .
we concentrate on the mesoscale and present the results of systematic numerical analysis of the ldg free energy functionals in @xmath1 and @xmath0 for colloidal particles characterized by a tangential ( degenerate in @xmath0 ) anchoring .
we use finite element methods ( fem ) with adaptive meshes in order to minimize the ldg functionals and obtain pair - wise effective interactions at all distances .
we find that the defects are point like in @xmath1 but acquire a rather complex structure in @xmath0 .
the structure depends on the combination of the anchoring potential , the particle radius , the temperature , and the lc elastic anisotropy .
more specifically , we identify defect core transitions between , as we call them , single - core , double - core , and split - core structures .
the split - core structure is favoured by increasing the anchoring strength or the particle radius and by decreasing the temperature .
the lc elastic anisotropy responsible for a preferred planar anchoring at the nematic - isotropic interface ( twist elastic constant smaller than bend and splay ) @xcite also favours the split - core structure .
finally , the split - core structure is favored by surface potentials characterized by a well - defined uniaxial minimum @xcite . in @xmath1
the boojums have a single - core point - like structure for arbitrarily strong surface potentials .
the paper is organized as follows : in sec .
[ 3d ] we discuss the @xmath0 system .
first we review the experimental results and define the ldg functional .
then we discuss the results for a single spherical colloid , focusing on the structures of the boojum core as a function of the anchoring strength , the colloid size and the temperature .
we consider two types of lcs differing by their anisotropy .
finally we consider the effective interaction between two colloidal particles and compare the results with the experimental measurements @xcite . in sec .
[ 2d ] we turn our attention to @xmath1 systems and calculate the interaction between colloids , where a short - range repulsion is found for strong anchoring .
a comparison with the experimental results of @xcite is carried out . in sec .
[ conclusions ] we discuss and compare our results in @xmath1 and @xmath0 . for completeness
, the numerical techniques used throughout this article are described in appendix [ appendix ] .
in the linear regime the effective force between two colloidal particles can be computed within the one - elastic - constant approximation by using the electrostatic analogy @xcite . for the case of spherical colloids with
degenerate tangential anchoring the leading asymptotic term in the multipole ansatz for the director field is given by a quadrupole term @xmath3 @xcite , where @xmath4 is the distance to the center of the colloidal particle .
the superposition approximation then leads to the quadrupolar effective interaction between two particles @xcite .
smalyukh et al .
@xcite have studied the anisotropic interactions between two colloids with tangential anchoring by using laser tweezers . at ( relatively )
large distances @xmath5 the radial dependence of the measured force was found to comply with the expected @xmath6 quadrupolar behaviour .
however , the angular dependence of the force disagreed qualitatively with the quadrupolar one at all distances , indicating the importance of the non - linear effects and a breakdown of the superposition approximation . in what follows
, we shall use the ldg theory in order to calculate effective interactions between two colloids with tangential anchoring .
we shall consider two types of lcs differing by their elastic anisotropies .
strong emphasis is placed on short distance behaviour and strong anchoring limit , where the non - linear effects are expected to dominate .
the problem is challenging from the theoretical point of view because the defect structure and dynamics are complex , both in the single - particle case as well as for interacting particles , as it was predicted a decade ago @xcite in @xmath1 and described recently in @xmath0 @xcite . in the following
we shall show that the colloidal interaction at short distances results , not only from the re - arrangement of the defect positions , but also from structural changes of the defect cores
. ultimately , we aim at describing quantitatively the experimental results of smalyukh et al .
@xcite , for the angular dependence of the elastic force at moderate and short inter - particle distances . within the landau - de gennes ( ldg )
theory @xcite nematic liquid crystals are characterised by a traceless symmetric order - parameter tensor with components @xmath7 , which can be written as @xmath8 where @xmath9 are the cartesian components of the director field @xmath10 , @xmath11 is the uniaxial order - parameter , which measures the degree of orientational order along the nematic director , and @xmath12 is the biaxial order parameter , which measures the degree of orientational order along the directions perpendicular to @xmath10 , characterized by the eigenvectors @xmath13 and @xmath14 .
the corresponding ldg free energy functional is @xmath15 with @xmath16 and @xmath17 the bulk and elastic free energy densities , given by @xmath18 where @xmath19 depends linearly on the temperature @xmath20 and is usualy written as @xmath21 , with @xmath22 a material dependent constant and @xmath23 the supercooling temperature of the isotropic phase . @xmath24 and @xmath25 are positive ( material dependent ) constants , and @xmath26 and @xmath27 are phenomenological parameters which can be related to the frank - oseen ( fo ) elastic constants .
the first integral in eq .
( [ free_energy ] ) is taken over the @xmath0 domain , @xmath28 , occupied by nematic , while the second integral is over the surfaces @xmath29 ( in our case the surfaces of the colloidal particles ) and accounts for non - rigid anchoring conditions .
depending on the preferred orientation of the director ( anchoring direction ) with respect to the surface normal @xmath30 , the surface anchoring is i ) homeotropic , when the anchoring direction is parallel to @xmath30 , ii ) planar , when the anchoring direction is orthogonal to @xmath30 , and iii ) tilted , when the anchoring direction and @xmath30 form an angle smaller than @xmath31 ( and larger than zero ) .
the last two cases can be classified futher as monostable , multistable or degenerate ( in some articles the term `` random '' is used instead of `` degenerate '' ) , depending on whether the surface imposes one , a finite or an infinite number of equivalent anchoring directions @xcite , respectively .
the simplest quadratic surface free energy , @xmath32 , favouring monostable nematic ordering @xmath33 , i.e. , with a well - defined director , scalar and biaxial order - parameters was proposed by nobili and durand @xcite : @xmath34 when @xmath35 , the surface free energy @xmath36 has a unique minimum at @xmath37 . in general
this is not necessary , the requirement being that the total free energy is bounded from below . in this paper
we consider only planar degenerate anchoring described by a family of covariant surface potentials originally proposed by fournier and galatola @xcite : @xmath38 where @xmath39 , @xmath40 , @xmath41 is the anchoring strength favouring tangential orientation of the director @xmath42 , and @xmath43 guarantees the existence of a minimum for a scalar order - parameter at the surface equal to its bulk value , @xmath44 . at a flat surface
the nematic is uniform and uniaxial everywhere . in the original formulation @xcite the surface scalar order - parameter
is allowed to vary and the biaxiality at a flat surface increases with the difference between the surface and bulk scalar order - parameters . the quartic surface potential given by eq .
( [ fournier ] ) may be viewed as the minimal biaxiality potential characterised by a well - defined degenerate planar minimum .
the quadratic surface potential ( @xmath45 ) is the covariant version of eq .
( [ nobili ] ) the minimum of which depends on the coupling to the bulk nematic . in the strong anchoring regime , the nematic director is parallel to the spherical colloidal surface everywhere , with an orientation determined by the far field director . in sec .
[ single ] we will see that the quartic term @xmath46 has a profound effect on the structure of the topological defects , by controlling the coupling to the bulk nematic through the deviation of the surface and bulk scalar order - parameters . in the flat surface limit
the biaxiality vanishes in the strong anchoring regime @xcite , while on the surface of large spherical colloids point - like singularities with charge + 1 split into pairs of point - like singularities with charge + 1/2 connected by disclination lines .
it is convenient to define the dimensionless temperature @xmath47 . at @xmath48
the uniaxial nematic is stable and the degree of orientational order is given by @xmath49 the nematic becomes unstable at @xmath50 . at @xmath51
both the nematic and the isotropic phases coexist .
typical values of the bulk parameters for 5cb are @xcite @xmath52 j @xmath53km@xmath54 , @xmath55 j@xmath53m@xmath54 , and @xmath56 j@xmath53m@xmath54 , @xmath57j@xmath53 m , @xmath58 k. the spatial extension of inhomogeneous regions and the cores of topological defects is of the order of the bulk correlation length , which is given by @xmath59 at the nematic - isotropic ( ni ) transition @xcite .
we define the elastic constant anisotropy @xmath60 , and consider two cases i ) @xmath61 with @xmath62 corresponding to 5cb , and @xmath63 with @xmath64 .
note that stability arguments require @xmath65 @xcite . depending on the lc material , @xmath27 can be positive or negative , and its sign controls the molecular orientation at the ni interface , which is planar for @xmath66 and homeotropic for @xmath67 @xcite .
the ldg elastic constants @xmath26 and @xmath27 may be related to the fo elastic constants @xcite , @xmath68 and @xmath69 , through the uniaxial ansatz @xmath70 , yielding @xmath71 and @xmath72 . in general @xmath73 and @xmath74 are different , but in most cases the difference is small and the ldg free energy is deemed adequate . in the following the ldg free energy eq .
( [ free_energy ] ) is minimised numerically using fem with the adaptive mesh refinement .
a detailed description of the numerical procedures is given in appendix [ appendix ] .
when spherical colloids with strong planar anchoring are dispersed in nematic lcs a pair of topological surface defects , named _ boojums _
, appears at the antipodes of the particles @xcite .
the understanding of the structure of defects in terms of the order - parameter distribution in the cores is beyond mere topological arguments and has been the subject of investigation since the early 1930s @xcite . for infinitely strong anchoring and within the class of axially symmetric fields , hedgehog defects exhibit three different structures : the radial hedgehog , a small ring or loop disclination @xcite and a third structure , a split - core defect , which was found to be metastable @xcite .
phase and bifurcation diagrams indicate that the transition from the hedgehog to the ring structure is first order , as predicted in ref .
although several studies addressed the structure of hedgehog defect cores @xcite the structure of boojums has not been fully understood @xcite .
a homotopy classification of surface topological defects is given by volovik in ref .
@xcite . according to volovik s classification a point defect at the boundary of a nematic lc may be viewed as the combination of a bulk hedgehog and a surface boojum and is therefore characterised by two topological charges the charge @xmath75 of the bulk hedgehog and the index @xmath76 of the projection @xmath77 of the director field onto the surface with normal @xmath30 .
the charges @xmath75 and @xmath76 are related to a continuous topological charge @xmath78 defined as @xcite @xmath79 where @xmath80 is a hemisphere surrounding the defect , and @xmath81 are arbitrary coordinates on @xmath80 .
evaluation of the integral in eq .
( [ volovik ] ) gives an explicit relation between the continuous charge @xmath78 and the charges @xmath75 and @xmath76 @xcite @xmath82 which provides a means of calculating the charge @xmath75 as the index @xmath76 can be calculated independently @xcite . if @xmath78 is an integer the defects can detach from the surface and
if @xmath83 they may vanish altogether .
by contrast , for non - integer charges @xmath78 the defects are topologically " bound to the surface . for instance , a single - core boojum , which is discussed below , has @xmath84 and @xmath85 , i.e. , the defect can not vanish nor detach itself from the surface . in this section
we investigate the dependence of the boojum core structure on i ) the strengths of the anchoring potential , @xmath86 , ii ) the reduced temperature @xmath87 , and iii ) the colloidal radius @xmath88 .
we introduce the dimensionless anchoring strengths @xmath89 ( @xmath90 ) , and consider lcs with positive ( e.g. 5cb ) and negative elastic anisotropies .
we pay particular attention to the nature and degree of the nematic order within the boojum cores . in agreement with ref .
@xcite we find that the apparent singularities of the director field are replaced by uniaxial order - parameter distributions with negative scalar order - parameter , corresponding to oblate nematic order , surrounded by biaxial layers .
we shall show that boojum cores are axially symmetric point - like with index @xmath91 on small colloids .
we call this structure single core boojum .
it is stable at high temperatures and relatively weak anchoring . in ref .
@xcite a similar boojum core structure ( named by the authors `` fingered '' boojum ) has been reported at a flat surface . on large colloids , at low temperatures and strong anchoring , the axial symmetry is broken and the boojum @xmath92 point - like cores split into pairs of @xmath93 point - like surface defects which are connected by disclination lines . a structure without a fully developed disclination line , the double core boojum , is also found .
the detailed structure of the boojums as well as the transitions between the different configurations depend in detail on the colloid and the lc parameters as we shall discuss below .
we start by describing the three distinct stable configurations of boojum cores , namely single , double and split core boojums .
the cores differ both in their surface and bulk structures .
typical core configurations are illustrated in the lower panels of fig .
[ core_types ] where the degree of biaxiality defined as @xmath94 is shown by a color map .
@xmath95 characterises the uniaxial nematic while @xmath96 corresponds to the maximal biaxiality nematic state .
@xmath96 is obtained when one ( and only one ) eigenvalue of @xmath97 vanishes .
note that in the isotropic phase all of the eigenvalues of @xmath97 vanish @xcite .
figure [ core_types ] illustrates the effect of @xmath87 on the structure of boojums for @xmath98 and strong quadratic ( @xmath99 ) surface potential .
as @xmath87 decreases the boojum core transforms from the single to the double core , and then to the split core structure with the accompanying disclination half - ring .
the core is fairly large close to the ni transition and shrinks as the temperature decreases .
single core , @xmath100 double core , and @xmath101 split core configurations .
the transitions between different configurations are driven by decreasing temperature @xmath87 : a ) @xmath102 , @xmath100 @xmath103 , c ) @xmath104 , at fixed anchoring strengths , @xmath105 , radius of the colloidal particle , @xmath106 , and elastic constant anisotropy , @xmath61 .
the scale bars represent the nematic correlation length @xmath107 at the ni transition . in the upper panels and in the lower panel @xmath101 the iso - surface of the scalar order - parameter corresponding to @xmath108
is shown . in the lower panel
@xmath100 the proto - ring of the double core boojum is represented by the iso - surfaces @xmath109 and @xmath110 .
the lower panels depict the color - coded biaxiality parameter eq .
( [ biaxiality_param ] ) in the plane @xmath111 .
the biaxiality parameter in the plane @xmath112 is also shown for the single core boojum , in the lower panel @xmath113 .
the short lines on the surface of the colloidal particle represent the eigenvector of @xmath97 corresponding to the largest eigenvalue .
this eigenvector corresponds to the nematic director only in the case of prolate nematic order .
see the main text for further details.,scaledwidth=50.0% ] a description of the boojum core configurations in terms of the eigenvalues of @xmath97 along the @xmath114-axis ( the direction of the far - field director ) is illustrated in fig .
[ core_types_eigenvalues_along_z ] , where the eigenvalues @xmath115 are plotted as functions of the distance @xmath114 to the colloidal surface at fixed @xmath116 ( the @xmath117 coordinates of the center of the colloidal particle ) .
single core boojums are uniaxial along the @xmath118axis , as revealed by the two degenerate eigenvalues .
the scalar order - parameter vanishes at @xmath119 above the colloidal surface , where the three eigenvalues vanish and the system is isotropic . below this point
the scalar order - parameter is negative , since the principal ( non - degenerate ) eigenvalue is negative and the nematic order is oblate . above @xmath119
the scalar order - parameter is positive indicating that the nematic becomes prolate .
by contrast , double core boojums are biaxial along the @xmath118axis .
the surface of maximal biaxility intersects the @xmath118axis at three distinct points ( see fig . [ core_types_eigenvalues_along_z]@xmath24 and fig . [ handled_hemisphere ] ) separating regions of low nematic oblate and prolate order .
split core boojums are also biaxial along the @xmath118axis .
the surface of maximal biaxiality intersects the @xmath118axis at two distinct points , @xmath120 and @xmath121 , that delimit the region of oblate nematic order . in the outer region , @xmath122 ,
two of the eigenvalues are degenerate and the nematic liquid crystal is uniaxial . as @xmath114 decreases the two negative eigenvalues depart from each other , implying non - zero biaxiality .
one eigenvalue remains negative .
the other two exchange places at a point that can be identified with the center of the boojum core . at this point
the nematic is uniaxial with @xmath123 .
all such points form a @xmath0 curve which is surrounded by the surface of maximal biaxiality ( @xmath124 ) .
this surface encloses the region with oblate uniaxial order ( @xmath123 ) . in all cases ,
the boojum cores are non - singular and biaxial over extended regions , with linear dimensions of the order of the bulk correlation length , @xmath125 .
on the @xmath118axis for the three types of boojum cores : @xmath113 single core , @xmath100 double core , and @xmath101 split core configurations , respectively .
the model parameters are as in fig .
[ core_types ] .
the far - field director is oriented along the @xmath118axis.,scaledwidth=48.0% ] we proceed to describe in detail the structure of the three types of boojum cores .
single core boojums are characterised by an axially symmetric distribution of the order - parameter @xmath97 .
consequently , @xmath126 on the @xmath114 axis , where two eigenvalues are degenerate , as shown in fig.[core_types_eigenvalues_along_z]@xmath19 .
the iso - surface @xmath127 is a distorted hemisphere that is biaxial ( @xmath96 ) everywhere ( see fig.[core_types]@xmath19 ) except at the point where it intersects the @xmath114 axis , @xmath128 , where the fluid is isotropic . at @xmath129 , the location of the putative surface point defect , the scalar order - parameter is negative ,
i.e. , the molecules are perpendicular to the director @xcite ( oblate nematic order ) . , and the color - coded biaxial parameter eq .
( [ biaxiality_param ] ) in the plane @xmath111 .
@xmath19 : the model parameters are as in fig .
[ core_types]@xmath24 ; @xmath24 : @xmath130 , @xmath131.,scaledwidth=48.0% ] in double core boojums the axial symmetry is broken . as a result
the system becomes biaxial along the @xmath114 axis , figs .
[ core_types]@xmath24 and [ core_types_eigenvalues_along_z]@xmath24 . the most prominent feature of this structure however , is the appearance of a proto - ring resembling the fully developed half - ring disclination of the split core boojum .
the proto - ring is formed in a region , of the order of @xmath107 , of low nematic order ( the scalar order - parameter is @xmath132 ) and is depicted in fig .
[ core_types]@xmath24 by the iso - surface @xmath110 .
the double core structure shares a number of features with the split core boojum .
its innermost core is a half - ring starting and ending on the colloidal surface , with very low biaxiality , @xmath133 , and negative scalar order - parameter .
further inspection of the eigenvalues of @xmath97 reveals that the iso - surface @xmath124 ( @xmath96 ) is a hemisphere with a handle at the top , see fig .
[ handled_hemisphere ] .
this iso - surface intersects the colloidal surface in a single closed loop . in the hole , below the handle
, the biaxiality is vanishingly small and the scalar order - parameter is positive , i.e. , the nematic is prolate uniaxial .
finally , the split core boojum is characterised by a @xmath134 disclination half - ring connected to the colloidal surface .
the inner core of the ring has zero biaxiality and oblate nematic order .
this region is surrounded by the iso - surface @xmath124 , forming a half - torus see figs .
[ core_types]@xmath25 and [ core_types_eigenvalues_along_z]@xmath25 , which intersects the colloidal surface in two distinct closed loops . the nematic is prolate uniaxial everywhere except in a region of order @xmath125 in the core of the half - ring .
how does the stability of these structures depend on the colloidal and lc parameters ?
all three types of boojums are topologically equivalent , in the sense that any of them can be transformed smoothly into any other .
the split and double core boojums result from breaking the axial symmetry of the single core , which splits on the colloidal surface .
this splitting is complete in the split core boojum in the sense that it is accompanied by the splitting of the @xmath124 surface , and only partial in the double core boojum .
the dependence of the equilibrium ( stable ) boojum core structure on the strength of the anchoring potential , @xmath86 , the reduced temperature @xmath87 , and the radius @xmath88 is complex and will be addressed below for lcs with positive and negative elastic anisotropies .
usually the anchoring on the surface of the colloidal particle is assumed to be infinitely strong .
here we investigate in detail the effects of the strength of the quadratic ( @xmath135 ) and quartic ( @xmath136 ) terms of the surface potential , eq .
( [ fournier ] ) .
when the anchoring strengths @xmath135 and @xmath136 are large the split core structure is favoured independently of @xmath137 , and @xmath138 .
when @xmath99 and @xmath135 is large the split ( double ) core boojum is found only for sufficiently large colloids and low temperatures in lcs with @xmath139 ( @xmath140 ) . at flat surfaces
the quartic potential induces uniform uniaxial nematic profiles @xcite . on
spherical colloids , in the strong anchoring regime , this potential stabilizes the split core boojums where the biaxial regions are reduced .
by contrast , in systems with a quadratic surface potential only , @xmath99 , the minimum is determined by the coupling of the surface to the bulk nematic @xcite . in this case
the stability of the boojum cores depends in detail on @xmath41 , @xmath88 , @xmath87 , and @xmath138 . minus the surface free energy @xmath141 ) as a function of the dimensionless quadratic anchoring strength , @xmath142 , @xmath143 , at several values of reduced temperature @xmath87 .
the inset shows the elastic ( full line ) excess ( dashed line ) and surface ( dash - dotted line ) free energies as functions of @xmath144 .
the lower panel depicts the configurations of ( 1 ) the single core boojum for @xmath145 , and ( 2 ) the split core boojum for @xmath146 , @xmath147 .
the black lines on the surface of the colloidal particle represent the eigenvectors corresponding to the largest eigenvalue of @xmath148 .
the iso - surfaces corresponding to @xmath108 are also shown .
the lc anisotropy @xmath149 , and the radius of the colloid @xmath150.,title="fig:",scaledwidth=45.0% ] + minus the surface free energy @xmath141 ) as a function of the dimensionless quadratic anchoring strength , @xmath142 , @xmath143 , at several values of reduced temperature @xmath87 .
the inset shows the elastic ( full line ) excess ( dashed line ) and surface ( dash - dotted line ) free energies as functions of @xmath144 .
the lower panel depicts the configurations of ( 1 ) the single core boojum for @xmath145 , and ( 2 ) the split core boojum for @xmath146 , @xmath147 .
the black lines on the surface of the colloidal particle represent the eigenvectors corresponding to the largest eigenvalue of @xmath148 .
the iso - surfaces corresponding to @xmath108 are also shown .
the lc anisotropy @xmath149 , and the radius of the colloid @xmath150.,title="fig:",scaledwidth=45.0% ] we start by investigating the quadratic surface potential , @xmath99 . in fig.[effect_of_w ] we plot the elastic free energy @xmath151 , where @xmath152 is the excess ( over the uniform uniaxial nematic ) free energy and @xmath141 is the surface free energy , as a function of the dimensionless anchoring strength , @xmath142 , for several values of @xmath87 from just below the ni transition ( @xmath153 ) to deep in the bulk nematic ( @xmath154 ) , at fixed values of @xmath138 and @xmath88 .
@xmath28 is the lc volume .
two regimes are observed . in the strong anchoring regime , @xmath155 ,
the elastic free energy exhibits a plateau that starts with the nucleation of the boojums at opposite poles of the colloidal particle .
the inset shows that in this regime the excess , @xmath156 , and the elastic , @xmath151 , free energies approach each other asymptotically , since the surface free energy @xmath141 decreases roughly as @xmath157 . in the weak anchoring regime , @xmath158 ,
the surface free energy increases linearly with @xmath144 and dominates over the elastic term , which grows as @xmath159 .
this behaviour is in line with the case of the homeotropic anchoring , where the elastic free energy was also found to vary as @xmath159 in the weak and to saturate in the strong anchoring regimes @xcite .
figures [ effect_of_w](1 ) and ( 2 ) illustrate how the structure of the boojum cores evolves as the strength of quadratic surface potential increases , at low temperatures .
as the anchoring strength increases the competition between elastic and anchoring energies leads to conflicting director orientations at the poles of the colloidal particles where the regions of reduced uniaxial and increased biaxial order develop .
inspection of the director configurations confirms that the system starts by developing a single core point - like surface defect , see fig.[effect_of_w]@xmath160 .
the projection @xmath161 of the director field onto the surface of the particle is characterised by an index @xmath162 . in this regime of small
@xmath144 , the surface director field has a finite normal component as can be seen in fig.[effect_of_w]@xmath160 .
deep in the strong anchoring regime the normal component of the surface director vanishes , see fig.[effect_of_w]@xmath163 , and the region of reduced uniaxial nematic order opens into a half - ring which is attached to the colloidal surface , and resembles a @xmath134 disclination line defect . .
circles correspond to a quadratic anchoring potential , @xmath164 , while squares and diamonds correspond to a quartic anchoring potential , @xmath135 and @xmath165 .
the surface core splitting is the distance between two @xmath134 point - like surface defects in the double and split core boojums .
single core boojums are stable at small anchoring strengths and high temperatures .
split core structures are stable at large anchoring strengths and low temperatures .
@xmath61 , and the colloid radius @xmath150 .
solid symbols correspond to double core boojums that are stable over a narrow range of anchoring strengths.,scaledwidth=45.0% ] for the quadratic surface potential ( @xmath164 ) boojum core structural transitions driven by the anchoring strength are observed at low temperatures only .
the boojum core structure diagram is shown in fig . [ cores_by_anchoring ] by circles .
the single core is stable at weak anchoring @xmath166 , the double core is stable in a narrow range of anchoring strengths , @xmath167 , while the split core is stable at strong anchoring @xmath168 . at high temperatures ( @xmath169 , not shown in fig.[cores_by_anchoring ] )
the quadratic anchoring potential is incapable of stabilizing the double or spilt core boojums .
only the single core structure is observed in the range of anchoring strengths @xmath170 .
we now proceed with the analysis of the quartic potential , eq .
( [ fournier ] ) .
we set @xmath149 , @xmath171 and assume , for simplicity @xmath172 .
we find three types of stable cores , at all temperatures . as the strength @xmath144 of the anchoring potential increases , the sequence of single , double and split core structures is observed .
two cases corresponding to @xmath173 ( squares ) and @xmath153 ( diamonds ) , are shown in the core structure diagram of fig .
[ cores_by_anchoring ] . at high temperature
the single core is stable for @xmath174 while the split core is stable for @xmath175 .
the double core is stable for intermediate anchoring strengths @xmath144 .
the transition between these core structures occurs at lower values of @xmath144 and the surface core splitting decreases as @xmath87 decreases . when compared with the quadratic potential at the same temperature and colloidal radius , the transitions for the quartic potential
occur at lower @xmath144 . in summary , at
fixed colloidal size and temperature , a strong quartic potential stabilizes the split core structure .
a strong quadratic potential is capable of stabilizing the split core boojums , only for lcs with positive elastic anisotropy .
for lcs with negative elastic anisotropy the split core boojums do not appear . in this article
we address only two cases @xmath176 .
presumably , there exist a positive threshold value of @xmath138 above which quadratic surface potential favours the split core boojums .
the double core structure is an intermediate structure that is stable over a limited range of anchoring strengths , for both types of surface potential . in the previous section we have investigated the role of the anchoring strength on the stability of the boojum core structures , and found evidence that the temperature plays a major role on the stability of the boojum cores .
here we consider the temperature as the control parameter of the structural boojum core transitions .
we start by investigating a colloidal particle with a quadratic surface potential , @xmath99 .
we consider reduced temperatures in the range @xmath177 and set @xmath178 , which is @xmath179 in lcs with positive elastic anisotropy @xmath61 , and @xmath180 in lcs with @xmath181 , since the bulk correlation length @xmath107 depends on the value of @xmath138 . which is @xmath179 for @xmath61 , and @xmath180 for @xmath181 .
split core boojums are stable for the quartic , @xmath135 and @xmath182 , potential at all temperatures and for both @xmath149 ( diamonds ) and @xmath183 ( squares ) . for the quadratic potential , @xmath99 and lcs with @xmath149 the sequence single - double - split cores is observed as the temperature decreases.,scaledwidth=45.0% ] a quadratic potential ( @xmath99 ) with @xmath184 stabilizes the split core structure at low temperatures only in lcs with positive elastic anisotropy , @xmath61 . the core structure diagram is depicted in fig .
[ cores_by_tau ] by circles . for @xmath185
the split core structure is stable , while at high temperatures @xmath186 the boojum has a single core structure .
the transition between these structures seems to occur continuously through the intermediate double core structure , which is observed in the range @xmath187 .
the surface splitting of the split core structure increases slowly as the temperature decreases and saturates at a value close to @xmath107 .
the split core structure was not observed in lcs with negative anisotropy , @xmath183 . recalling that @xmath139 corresponds to @xmath188 ,
while @xmath140 to @xmath189 , we note that unfavourable twist distortions are responsible for the suppression of the split core structures in lcs with @xmath183 .
twist deformations vanish when the azimuthal component of the director field , @xmath190 , vanishes @xcite , which is the case for single core boojums . by contrast , for double and split core boojums @xmath190 acquires a non - zero value that increases from the double to the split core structures . because the twist distortions are energetically unfavourable in lcs with @xmath140 the split core boojums are unstable with respect to the double core structure .
this is in line with a twist transition that occurs at the core of a hyperbolic hedgehog defect ( see figs . 4 , 5 in ref .
@xcite ) on a colloidal particle with homeotropic anchoring , and in general , with the effect of elastic anisotropy on the structure of distortions around topological defects in @xmath1 @xcite . for a quartic potential with @xmath191
the split core is stable for both types of lc elastic anisotropy , at all temperatures .
the core splitting increases with temperature up to @xmath192 for @xmath61 , and @xmath193 for @xmath181 , close to the ni coexistence temperature ( see fig .
[ cores_by_tau ] diamonds and squares , respectively ) .
the quartic term in the surface potential is quite effective in stabilizing the split core structure , even when twist distortions are unfarourable ( @xmath181 ) .
in fact , the split core is not only stabilized , but its splitting is even larger than that observed in systems with @xmath139 .
we note that the boojum size is proportional to the splitting parameter shown in fig .
[ cores_by_tau ] and it decreases as the temperature decreases . in summary , strong quadratic surface potentials and low temperatures
favour split core boojums only for lcs with positive elastic anisotropy . when @xmath194 the double core structure is stable .
strong quartic potentials favour split core boojums , which are the only stable structures , regardless of the lc anisotropy .
finally , we investigate the effect of the colloidal radius @xmath88 on the structural transitions of the boojum cores . for this purpose
we consider colloids with sizes in the range @xmath195 . , at @xmath173 . for quadratic potentials ,
single core boojums are stable at small and spilt core boojums are stable at large colloidal sizes , @xmath88 . the lc anisotropy @xmath61 . for quartic potentials ,
the split core structure is stable over the whole range of @xmath88 .
the solid lines are guides to the eye.,scaledwidth=45.0% ] boojum core transitions driven by the colloidal size , for quadratic ( circles ) and quartic ( diamonds ) potentials are illustrated in fig .
[ cores_by_r ] . for lcs with @xmath61
the quadratic potential stabilizes the split core boojum on large colloids , @xmath196 , the single core boojum on small colloids , @xmath197 , and the double core boojum at intermediate colloid sizes .
by contrast , in lcs with @xmath183 , and the same quadratic potential , the split core is not observed and the single - double core transition occurs at @xmath198 ( not shown ) .
this behaviour is similar to the one described in previous sections , for quadratic potentials . when twist distortions are energetically unfavourable they are suppressed rendering the split core structure unstable . for a quartic potential with @xmath199
the split core is stable for both types of elastic anisotropy in the whole range of colloidal sizes .
the size of the half - ring , given by the core splitting on the colloidal surface , increases with @xmath88 and saturates at @xmath200 for @xmath61 , and at @xmath201 for @xmath181 .
again , the quartic surface potential eq .
( [ fournier ] ) stabilizes split core boojums in lcs with negative elastic anisotropy .
we end this section by noting that although , it may be difficult to observe the reported boojum core structures using standard optical methods , their influence on the effective interaction between colloidal particles at short distances is significant , as will be discussed in subsection [ 3d_two_colloids ] . the far - field nematic configuration , however , is not affected by the mesoscopic structure of the boojum cores . axis . for large @xmath5 pairs of boojums
are aligned parallel to the far - field director .
[ 2_colloids ] , scaledwidth=25.0% ] in this section we shall calculate the two - body effective interaction potential @xmath202 @xcite for colloids with a quadratic surface potential , eq .
( [ fournier ] ) .
for the definition of @xmath5 and @xmath203 , see fig . [
2_colloids ] .
@xmath204 is defined by decomposing the excess ( over the free energy of the uniform nematic ) free energy @xmath205 for the system of two particles as follows @xmath206 here @xmath156 is the excess free energy calculated independently for one isolated colloid . by definition @xmath204 tends to zero when @xmath207 .
in particular , we shall demonstrate how the non - linear effects become dominant , leading to a qualitative change of the effective interaction , e.g. , the potential @xmath208 changes from repulsive at large @xmath5 to strongly attractive at moderate to short distances . as we shall see this
is driven by the complex behaviour of the interaction between the topological defects that results in a re - arrangement of their structure at short distances .
analogous changes in the effective interaction have been reported experimentally @xcite and confirmed within ldg theory @xcite where it has been shown that the axial symmetry of the location of the boojum pairs is continuously broken , resulting in an effective colloidal attraction . here
we shall re - examine the effective colloidal interaction based on the results of the previous section where we have shown that the axial symmetry of the boojum cores may be broken at the level of a single colloidal particle . in what follows we use the notation set in fig .
[ 2_colloids ] .
the far - field director is aligned with the @xmath114-axis . due to the spherical symmetry of the colloids
, @xmath204 depends on the colloidal separation @xmath5 and on the polar angle @xmath203 only . in this section
we consider colloidal particles of equal size , @xmath178 , the anisotropy of the elastic constants @xmath149 , the reduced temperature @xmath130 , and the other model parameters corresponding to the ldg parametrisation of 5cb , described in subsection [ ldgmodel ] .
for this set of parameters , the boojum cores exhibit a single core configuration over the whole range of anchoring strength @xmath209 .
ultimately , we aim at explaining the experimental results of smalyukh et al .
@xcite , namely the force versus @xmath203 curves , which disagrees strongly with the asymptotic quadrupole - quadrupole result , at moderate and short colloid - colloid distances . for several values of @xmath203 .
the surface potential is quadratic , eq . ( [ fournier ] ) , and the anchoring strength is @xmath210 .
elastic anisotropy @xmath149 , radius of the particles @xmath211 , and reduced temperature @xmath130 . [ interact_w_0.0001 ] , scaledwidth=45.0% ] in fig .
[ interact_w_0.0001 ] we plot @xmath212 in the weak anchoring regime . in this regime
the interaction is repulsive when the inter - colloidal vector is either parallel or perpendicular to the far - field director , @xmath213 or @xmath31 respectively , and is attractive at intermediate colloidal orientations , as expected for quadrupole - quadrupole interactions . for several values of @xmath203 .
the anchoring strength is @xmath214 .
elastic anisotropy @xmath149 , radius of the particles @xmath211 , and reduced temperature @xmath130 .
the insets depict the boojum configurations for several distances @xmath5 and @xmath215 . in the attractive range , @xmath216 ,
the inner boojums are @xmath134 disclination rings represented by the iso - surface of the scalar order - parameter @xmath108 .
[ interact_w_0.01 ] , scaledwidth=50.0% ] the character of the effective interaction changes significantly in the strong anchoring regime , as shown in fig .
[ interact_w_0.01 ] . for oblique , @xmath217 , and perpendicular , @xmath218 , orientations of the colloidal pair
the interaction remains qualitatively the same : attractive when the colloids are at an oblique angle with the far - field director and purely repulsive when the colloids are perpendicular to it .
however , at @xmath213 the effective interaction is no longer purely repulsive . at short distances , @xmath219 ,
the particles start to attract each other .
this change results from a symmetry break of the boojum - pair configuration as shown in the insets of fig .
[ interact_w_0.01 ] . at large distances the boojums
are located at the poles of the particles , see inset ( 3 ) in fig .
[ interact_w_0.01 ] , and aligned parallel to the far - field director , as in the case of an isolated particle . as the separation between the colloids decreases , the repulsion of the inner defects increases .
the change from repulsive to attractive colloidal interaction is driven by a change in the position of the inner defects , see inset ( 2 ) in fig .
[ interact_w_0.01 ] .
this mechanism for attractive colloidal interactions the re - arrangement of the inner defects with quadrupolar symmetry was reported almost 10 years ago for @xmath1 colloids with homeotropic anchoring @xcite .
we note that although in the single particle configuration the boojums are in the single core state , the inner boojums undergo a transition to the split core state as the distance between the colloidal particles decreases , see the inner boojums in fig .
[ interact_w_0.01](2 ) .
the symmetry break in the position of the inner boojums is accompanied by a symmetry break of the inner boojums core structure , which changes from the axially symmetric single core to the asymmetric split core configuration .
the global minimum of @xmath202 corresponds to close contact , @xmath220 , at an oblique angle @xmath203 close to @xmath221 .
this is in line with the experimental results of @xcite where the boojum - colloids were reported to coalesce . in sec .
[ 2d ] we shall show that this is not the case for boojum - colloids in @xmath1 systems . between two spherical colloidal particles , aligned with the far - field director , @xmath213 .
the surface anchoring potential is quadratic , eq . ( [ fournier ] ) .
elastic anisotropy @xmath149 , radius of the particles @xmath211 , and reduced temperature @xmath130 .
the anchoring strength @xmath144 varies in the interval @xmath222 $ ] .
the inset depicts the height of the free energy barrier as a function of @xmath144 , exhibiting quadratic behaviour @xmath223 in the weak anchoring regime .
[ pair_theta_0_barrier_vs_w ] , scaledwidth=50.0% ] the effect of the anchoring strength @xmath144 is trivial for colloidal orientations @xmath224 , and thus we focus on @xmath215 . in fig .
[ pair_theta_0_barrier_vs_w ] we plot @xmath208 for several values of anchoring strengths @xmath144 .
the dependence of the free energy barrier on the anchoring strength is illustrated in the inset .
the barrier height is weakly discontinuous at @xmath225 delimiting two types of behaviour . for strong
anchoring , @xmath226 , the free energy at distances close to the free energy maximum , @xmath227 , is a smooth function of @xmath5 .
this implies that as the colloidal separation decreases the boojums rearrange smoothly , driving the change from repulsive to attractive interactions . for weak
anchoring , @xmath228 , the nematic order at the poles of the particles is suppressed , but there are no defects .
the system behaves almost linearly in this regime , i.e. , @xmath229 and the resulting free - energy @xmath230 .
the effective interaction is repulsive at all @xmath5 , as expected for two quadrupoles oriented at @xmath215 .
we conclude that the interaction between colloids aligned with the far - field director changes from repulsive to attractive at short - distances as the result of the re - arrangement of the positions and of the core configurations of the inner boojums . finally , we shall give a detailed description of the angular dependence of the effective force acting on the colloids .
the experimental measurement of this force has been reported in ref .
@xcite . according to this reference
brownian motion dominates the dynamics for @xmath231 , but at shorter distances ( @xmath232 ) the dynamics is governed by the nematic mediated force . in this regime
the force of the order of the elastic constant @xmath233 has been found .
@xcite . since the measured elastic force was significant only for relatively small distances , strong deviations from the asymptotic quadrupolar interaction have been reported .
for instance , the orientation @xmath203 corresponding to the strongest attraction has been found to depend on the distance between colloidal particles , while the interaction of quadrupolar type exhibits maximal attraction at @xmath234 , for any @xmath5 .
in addition , the attraction for @xmath235 and repulsion for @xmath236 has been found , while quadrupoles repel at @xmath215 . finally , the measured radial dependence of the force ( at fixed @xmath237 ) deviated systematically from the quadrupolar power law @xmath238 for @xmath239 .
these findings suggest that non - linear effects dominate and that the superposition approximation does not hold in the range of distances that have been investigated .
we have calculated the absolute value @xmath240 of the elastic force , @xmath241 , as well as its radial and polar components @xmath242 and @xmath243 by numerical differentiation . , for several values of @xmath5 .
@xmath98 , @xmath244 , @xmath173 and anchoring strength @xmath146 .
[ force ] , scaledwidth=45.0% ] in fig .
[ force ] the resulting magnitude of the force is plotted as a function of the relative orientation , @xmath203 , for several values of the particles separation @xmath5 .
we note that as the separation between the particles increases the magnitude of the force decreases and its structure as a function of @xmath203 decreases . in fig .
[ force ] the angular dependence of the force , at separation @xmath245 , resembles that of the quadrupolar force @xcite with principal maxima at @xmath213 and @xmath246 .
in addition , the force exhibits three secondary maxima at intermediate values of @xmath203 in line with the angular dependence of the quadrupole - quadrupole force @xcite .
as the distance between the colloids decreases the angular dependence of the force changes quite rapidly and quite drastically . just below @xmath247
the maximum at @xmath246 becomes a minimum ( the force weakens ) and the difference between the secondary maxima at intermediate values of @xmath203 becomes more pronounced .
a similar angular dependence was observed in the force reported in ref .
@xcite at colloidal separations @xmath248 . however , our theoretical results appear to be more sensitive to the inter - colloidal separation and predict different angular behaviours at colloidal separations @xmath247 and @xmath245 at @xmath213 . at this orientation
the experiments report a local minimum for colloidal separations in the range @xmath248 @xcite . , for several values of @xmath5 .
the values of the model parameters are the same as in fig .
[ force ] .
[ force_r ] , scaledwidth=45.0% ] , at several values of the particles separation @xmath5 .
the values of the model parameters are the same as in fig .
[ force ] .
[ force_theta ] , scaledwidth=47.0% ] in an attempt to rationalise the discrepancies between the theoretical and the experimental results we have plotted the radial and polar components of the force as a function of @xmath203 in figs .
[ force_r ] and [ force_theta ] .
the radial component of the force is found to be two orders of magnitude larger than the polar component , and thus it dominates the behaviour of the total force .
this is in sharp contrast with the measured force , which is characterised by a radial component that differs significantly from the total force at short distances , around @xmath213 ( see fig . 4 of reference @xcite , in particular the inset c ) where the radial component of the force at @xmath247 appears to vanish at @xmath213 ) .
the results for the polar component of the force plotted in fig .
[ force_theta ] reveal that the angular dependence changes abruptly at @xmath213 , by contrast to the behaviour of the force at @xmath217 and @xmath31 .
this is related to the head - on interaction of the intermediate boojums and the associated discontinuity of @xmath249 at @xmath250 .
this , probably , has had an impact on the experimental results as the colloidal configuration is unstable at @xmath215 rendering an accurate force measurement , at this orientation , very difficult .
finally , we note that the colloidal particles in the experiment are of the order of @xmath251 , while we considered colloids one order of magnitude smaller
. this may also account for some discrepancies between the theoretical and experimental results .
dispersed colloidal particles in smectic @xmath2 ( sm@xmath2 ) films are realised either as inclusions of a lower order phase ( isotropic , nematic , or cholesteric ) or as smectic islands with a higher number of layers than the surroundings @xcite .
planar boundary conditions at the colloidal particle may lead to the nucleation of a pair of surface defects @xcite .
planar anchoring may also lead to the nucleation of a satellite ( bulk ) defect @xcite , which will not be considered here . at large distances
the interaction between these particles is of quadrupolar type , and deviates from it at short - distances .
by contrast to the observations in 3d systems , where the spherical colloids with planar anchoring come into contact and coalesce @xcite , inclusions in sm@xmath2 films maintain a well - defined separation @xmath252 , where @xmath88 is the inclusion radius .
the distortions in sm@xmath2 films due to an isolated inclusion with planar anchoring , as well as the asymptotic effective pair interaction , have been described by invoking the electrostatic analogy and the superposition approximation . in the simplest approach @xcite the distortions around an inclusion
are modelled by the superposition of three solutions to the laplace equation corresponding to one topological defect with charge @xmath92 in the center of the inclusion and two @xmath253 surface defects . then by using the superposition approximation
the quadrupole - quadrupole effective interaction is obtained .
however , this solution does not satisfy the boundary conditions at the inclusion boundary , and the superposition approximation fails to describe the repulsion observed experimentally at small distances @xcite .
a related approach @xcite , based on the exact solution for the @xmath254director ( defined below ) around an inclusion with arbitrary anchoring strength @xcite reveals that the repulsion appears simultaneously with the nucleation of the defects .
nevertheless , neither approach describes the non - linear effects responsible for the re - orientation of the defects at short distances . in what follows , we shall use the landau description of sm@xmath2 films in order to calculate the interaction between two inclusions with planar anchoring .
we aim at a detailed description of the defect structure , and we will show that the short distance colloidal interactions result from the re - arrangement of the defect positions , as in other @xmath1 and @xmath0 systems with quadrupolar symmetry @xcite . .
@xmath255 is the normal to the layers , @xmath10 is the average local molecular orientation , and @xmath256 is the in - plane projection of @xmath10.,scaledwidth=30.0% ] in sm@xmath2 phases the lc molecules are organized into @xmath1 layers with the local average molecular orientation @xmath10 at a characteristic angle @xmath257 with respect to the layer normal @xmath255 . when layer deformations are negligible there are no variations of @xmath257 in the direction of @xmath255 , @xmath258 . in this case
, the relevant order - parameter is the in - plane projection of the director @xmath10 .
this is an ordinary ( variable length ) @xmath1 vector field @xmath256 . the simplest mesoscopic model for sm@xmath2 films
is obtained by expanding the free energy in terms of the invariants of the order - parameter @xmath256 and its derivatives @xmath259 @xcite .
the free energy is then written as @xmath260 , where @xmath261 is the thickness of the film , and the bulk and elastic free energy densities are , at lowest order , @xmath262 where , for simplicity , we assume the one - elastic constant approximation ( @xmath263 ) .
@xmath264 is assumed to be a linear function of temperature @xmath20 and @xmath265 is a constant .
dimensionless free energy densities ( the free energy @xmath266 has units of volume ) are defined through the rescaling @xmath267 and @xmath268 , @xmath269 where @xmath270 is the correlation length .
( [ smcbulk ] ) predicts non - zero values for the bulk scalar order - parameter @xmath271 at all temperatures .
topological defects correspond to regions of reduced orientational order @xmath272 , and their topological charge is defined by the winding - number of the vector field @xmath273 @xcite . due to the vector nature of the @xmath256 director ,
it is possible to distinguish clockwise , anticlockwise or mixed planar anchoring on circular inclusions in sm@xmath2 films @xcite .
the orientation that is realised results from the interplay of the anchoring potential , the @xmath256 director field in the vicinity of the inclusion , and the lc properties .
we consider the case of mixed planar anchoring which favours the nucleation of boojum pairs @xcite , the @xmath1 counterpart of the @xmath0 boojums investigated in subsection [ single ] .
we use a surface potential of the rapini - papoular form @xcite which is the @xmath1 version of the quadratic surface potential , eq .
( [ fournier ] ) , @xmath274 where @xmath275 is the anchoring strength , and @xmath255 is the normal to the surface @xmath276 of the inclusion .
we note that in @xmath1 quartic terms do not change the surface potential qualitatively and will not be considered . rescaling the variables we obtain @xmath277 where we introduced the dimensionless anchoring strength @xmath278 . in this section we consider a single circular inclusion in a sm@xmath2 film .
the mixed planar anchoring is enforced by fixing the far - field @xmath279director parallel to the @xmath280axis . in this case
a pair of boojums may nucleate at the particle surface . within the fo elastic theory
the @xmath279director is a unit vector @xmath281 .
the analytic solution for the orientational field , @xmath282 , at arbitrary anchoring strength @xmath144 , can be written as @xcite @xmath283 , \label{phi}\ ] ] where @xmath284 are polar coordinates , and @xmath285 .
the corresponding free energy is @xmath286 .
the solution , eq .
( [ phi ] ) , describes non - singular surface disclinations with a core size @xmath287 @xcite , which for strong anchoring behaves as @xmath288 . for any finite @xmath144 the elastic
free - energy is also regular , with the leading large @xmath144 behaviour @xmath289 .
singular topological defects appear only in the limit @xmath290 . in the strong anchoring regime @xmath291
the exact solution , eq . ( [ phi ] ) , may be approximated by the sum of a finite number of solutions @xmath292 to the laplace equation , where @xmath292 represents a point singularity ( or topological defect ) of some winding number @xmath293 @xcite .
the number of these topological defects , their locations , and winding numbers , are chosen to comply with the boundary conditions .
the behaviour of the @xmath279director near the boojums suggests that these are @xmath253 topological defects . since the colloidal particle with rigid planar anchoring accounts for a @xmath92 virtual defect in the center
, a naive guess would be to represent @xmath282 as the sum of three point singularities , which we write as @xmath294 @xcite .
two @xmath295 singularities are located on opposite poles of the inclusion and are oriented along the far - field @xmath279director .
however , a more careful analysis reveals that this ansatz violates the boundary condition at the surface of the inclusion , and additional virtual defects must be taken into account .
the asymptotic @xmath290 behaviour of the solution , eq .
( [ phi ] ) , indicates that the correct ansatz corresponds to the triplet @xmath296 , where a @xmath297 virtual defect is placed in the center of the inclusion and two @xmath298 surface defects may be thought of as the sum of two @xmath295 defects , one of which is virtual and the other real @xcite . , excess ( over the free energy of the uniform system ) free energy @xmath156 , and surface free energy @xmath141 as a function of the planar anchoring strength @xmath278 .
points correspond to the numerical minimization of the landau free - energy functional ( eqs .
( [ smcbulk ] ) , ( [ smcelastic ] ) , ( [ rapini - papoular ] ) ) .
dashed lines are obtained using the exact solution , eq .
( [ phi ] ) .
the radius of the particle @xmath299.,scaledwidth=50.0% ] in fig .
[ energy2dsingle ] we compare the excess free energy @xmath156 , the surface free energy @xmath141 , and the elastic free energy @xmath151 , obtained by numerical minimization of the landau free energy functional ( eqs .
( [ smcbulk ] ) , ( [ smcelastic ] ) , and ( [ rapini - papoular ] ) ) with the analytic results obtained by using the exact solution , eq .
( [ phi ] ) , to the frank - oseen theory .
details of the numerical method may be found in the appendix . in the weak anchoring regime
the agreement is excellent . in the strong anchoring regime , however , the solutions start to deviate . inspection of fig .
[ energy2dsingle ] reveals that the discrepancies may be attributed to the differences in the surface free energies at large @xmath144 .
indeed , the fo free energy is inaccurate in the presence of defects and ultimately diverges as @xmath300 . in the weak anchoring regime , @xmath301 ,
the excess free energy is dominated by the surface term which is linear in @xmath144 , while the elastic free energy is quadratic , @xmath159 , as in @xmath0 systems , see fig .
[ effect_of_w ] .
director configurations for anchoring strengths , @xmath113 @xmath302 , @xmath100 @xmath303 , @xmath101 @xmath304 , and @xmath305 .
the grey - scale map corresponds to the light intensity @xmath306 as seen in experiments under cross - polarisers .
the radius of the particle is @xmath307.,scaledwidth=45.0% ] figure [ single2d ] illustrates the equilibrium @xmath256-director configurations for increasing values of the anchoring strength @xmath144 .
the grey - scale map corresponds to the optical transmittance , @xmath308 , as seen in experiments under cross - polarisers . as the anchoring strength increases
the inclusion - induced distortions , seen as bright fringes , grow . for @xmath309
a pair of topological defects nucleates at opposite poles of the particle ( see fig.[single2d]@xmath101 and @xmath310 ) . , eq . ( [ chi_schopol ] ) , of the 2d boojum core as a function of the anchoring strength @xmath278.,scaledwidth=45.0% ] in order to quantify the spatial extension of the boojums , we follow ref . @xcite and introduce an effective core radius as follows @xmath311^{\frac{1}{2 } } , \label{chi_schopol}\ ] ] where the integral is over the region with one boojum . @xmath312 as a function of the anchoring strength @xmath144 is plotted in fig .
[ schopohl ] .
we distinguish three regimes . as the anchoring increases the lc molecules align along the preferred surface orientation creating small regions of low orientational order . at higher anchoring strength , @xmath313 , @xmath314 , and the regions of low orientational order exhibit pre - nucleation of defects . finally , in the nucleation regime , @xmath315 , the effective radius increases with @xmath144 slowly , and eventually saturates at @xmath316 .
by contrast to the @xmath0 case , the @xmath1 boojums are always point - like and never split .
this is a direct consequence of the topological constraint which forbids half - integer defects in sm@xmath2 films . in @xmath0 nematics ,
however , half - integer ring disclinations are not only topologically allowed , but are even energetically favourable @xcite .
this is one of the reasons why single core boojums in @xmath0 may split into half - ring disclinations . in this subsection
we shall analyse the pairwise effective interaction between inclusions with planar anchoring in sm@xmath2 films . at @xmath5
the inclusions `` decorated '' by boojums interact as @xmath1 quadrupoles @xcite @xmath317 this interaction is repulsive for @xmath318 and attractive for @xmath319 , with @xmath320 . at large distances
the single - particle distribution of the @xmath256-director is only slightly perturbed by the presence of the other particle . as the separation decreases the perturbation increases and non - linear effects take over , changing the character of the interaction .
the threshold distance @xmath227 at which the non - linear regime sets in depends strongly on the anchoring strength . for several values of @xmath203 , anchoring strengths @xmath321 , and radius @xmath322 .
[ 2d_inclusions_1 ] , scaledwidth=50.0% ] , but for anchoring strengths @xmath323 .
[ 2d_inclusions_2 ] , scaledwidth=50.0% ] director configurations for interacting colloidal particles with the anchoring strength @xmath324 at relative orientation @xmath213 and separations @xmath19 ) @xmath325 and @xmath24 ) @xmath326 , @xmath307 .
@xmath25 ) close view of the configuration between the colloidal particles , for parameters as in @xmath24 ) .
the greyscale map corresponds to the light intensity @xmath327 as seen in experiments under cross - polarisers .
[ 2d_inclusions_3 ] , scaledwidth=45.0% ] we minimize the landau free energy functional , eqs .
( [ smcbulk ] ) , ( [ smcelastic ] ) , ( [ rapini - papoular ] ) numerically .
details may be found in the appendix . in figs .
[ 2d_inclusions_1 ] and [ 2d_inclusions_2 ] we plot the effective interaction free - energy @xmath328 as a function of the particle separation , @xmath5 , for anchoring strengths @xmath329 and @xmath330 .
@xmath331 contains the contributions to the free - energy from the distortions of the @xmath279director , @xmath332 is the free energy of the uniform bulk system , and @xmath333 is the free energy of an isolated inclusion . at weak anchoring , @xmath302
, @xmath204 is always repulsive for perpendicular @xmath334 or parallel @xmath215 orientations , and attractive for oblique orientations , @xmath335 . when the anchoring strength increases regions of reduced order develop around the poles of the particles leading to an increase of the elastic free energy , see fig .
[ energy2dsingle ] .
this also results in an increase of the interaction strength at short distances , as can be seen in fig .
[ 2d_inclusions_1 ] for @xmath336 and fig .
[ 2d_inclusions_2 ] for @xmath304 .
after the nucleation of topological defects the effective interaction changes drastically as shown in fig .
[ 2d_inclusions_2 ] , @xmath324 , at separations @xmath337 . in this regime
the repulsion becomes even stronger for orientations @xmath338 , but at @xmath335 the free - energy develops a well defined local minimum at @xmath339 , which is followed by a repulsion for @xmath340 .
the latter prevents the inclusions from coalescing in agreement with the experimental observations @xcite .
we recall that this is in sharp contrast with the behaviour of @xmath0 boojum - colloids , where no such repulsion is observed . for strong
anchoring and @xmath213 , @xmath212 changes from repulsive to attractive , with a discontinuous slope , at @xmath341 , see fig .
[ 2d_inclusions_2 ] , @xmath304 .
similar behaviour is observed for @xmath324 at @xmath342 , and is not shown in fig .
[ 2d_inclusions_2 ] because the energy scale was chosen in order to emphasize the short range repulsion and the local minimum for @xmath217 .
figure [ 2d_inclusions_3 ] illustrates the equilibrium @xmath256-director configurations for inclusions oriented along the far - filed @xmath279director , @xmath213 , and for the separations @xmath19 ) @xmath325 , @xmath24 ) @xmath326 ; the anchoring strength is @xmath324 . for @xmath343 ( @xmath344 for @xmath304 )
the inner boojums are aligned along the far - field @xmath279director , giving rise to an overall repulsion between the inclusions .
as the separation decreases , the inner boojums slide away from the @xmath280axis in opposite directions , as shown in fig .
[ 2d_inclusions_3]@xmath25 ) .
this leads to a strong attraction between the particles .
we emphasize that in @xmath1 the re - arrangement of the inner boojums at @xmath215 proceeds discontinuously , while in @xmath0 continuously , see fig .
[ interact_w_0.01 ] black circles .
the contact value @xmath345 is , however , larger than @xmath346 which corresponds to the global minimum of @xmath202 for the case of strong anchoring .
we have carried out a detailed study of the effective colloidal interactions for particles dispersed in a nematic host ( @xmath0 ) or in a smectic @xmath2 film ( @xmath1 ) . in both cases planar ( degenerate in @xmath0 ) anchoring is imposed on the surfaces of the colloidal particles . within the frank - oseen formalism in the strong anchoring regime ,
the boundary conditions are met by the creation of a pair of antipodal surface defects , boojums @xcite . for an isolated particle the vector connecting two boojums
aligns itself with the far - field director , as required by the global boundary conditions @xcite .
the resulting director field has quadrupolar symmetry , and the ensuing effective colloidal interactions exhibit quadrupole - quadrupole asymptotic behaviour @xcite , both in @xmath1 and in @xmath0 .
however , at short distances , where the superposition approximation fails and the non - linear effects dominate , the experimental results report significant deviations from this quadrupolar behaviour .
the most important of these is the crossover from a repulsive to an attractive interaction at some threshold distance for particles aligned along the far - field director , @xmath250 .
the threshold distance is @xmath347 in @xmath0 , see fig .
[ interact_w_0.01 ] , and @xmath348 in @xmath1 , see fig .
[ 2d_inclusions_2 ] .
this crossover is driven by the configurational reorientation of the inner boojums as the distance between the particles decreases .
we have shown that the short - range colloidal interactions in @xmath0 result not only from the re - arrangement of the defect positions , as predicted in @xmath1 a decade ago @xcite and described recently in @xmath0 @xcite , but also from changes in the structure of the boojum cores .
the description of the structural transitions between boojum cores is a challenging theoretical problem as distinct structures , characterized by tensor order - parameters that vary in regions of the order of the bulk correlation length , have similar free energies .
their stability results from a delicate balance of various contributions and structural transitions may be driven by the anchoring strength , the temperature or the colloid radius .
we have concentrated on the mesoscale and addressed the structure and dynamics of the boojums and the resulting colloidal interactions based on the ldg free energy in @xmath0 and @xmath1 , for spherically symmetric colloids .
we have used finite elements methods with adaptive meshes in order to resolve the structure of the defect cores and establish the nature of the short - distance effective colloidal interactions .
we have found that the defects are point like in @xmath1 , but acquire a rather complex structure in @xmath0 , which depends on the combination of the anchoring potential , the colloid radius , the temperature and the lc elastic anisotropy .
in addition , we have found defect - core transitions between @xmath349 single core , @xmath350 ) double core , and @xmath351 ) split core structures of the boojum pairs .
single core boojums are uniaxial along the @xmath118axis ( see figs .
[ core_types]@xmath19 , [ core_types_eigenvalues_along_z]@xmath19 ) .
the scalar order - parameter vanishes at one point above the colloidal surface , where the system is isotropic . below this point
the scalar order - parameter is negative and the nematic order is oblate . above that point
the scalar order - parameter is positive indicating that the nematic order is prolate .
by contrast , double core boojums are biaxial along the @xmath118axis ( see figs . [ core_types]@xmath24 , [ core_types_eigenvalues_along_z]@xmath24 ) .
the surface of maximal biaxiality intersects the @xmath118axis at three distinct points separating regions of low nematic oblate and prolate order .
split core boojums are also biaxial along the @xmath118axis .
the surface of maximal biaxiality intersects the @xmath118axis at two distinct points that delimit the region of oblate nematic order , figs .
[ core_types]@xmath25 , [ core_types_eigenvalues_along_z]@xmath25 . as a general conclusion
we find that on small colloids boojum cores are axially symmetric ( see fig .
[ cores_by_r ] ) , point - like with index @xmath92 the index referring to the charge of the projected director field on the surface , which are stable at high temperatures ( see fig . [ cores_by_tau ] ) and relatively weak anchoring ( see fig . [ cores_by_anchoring ] ) . on large colloids at low temperatures and strong anchoring , the axial symmetry is broken and the boojum @xmath92 point - like cores split into a pair of @xmath93 point - like surface defects , connected by a disinclination line ( see fig . [
core_types]@xmath25 ) . a structure without a fully developed disinclination line , the double core boojum ,
was also found ( see figs .
[ core_types]@xmath24 , [ handled_hemisphere ] ) .
the detailed structure of the cores as well as the transitions between the different configurations depend in detail on the colloid and lc parameters .
we stress that the details of the core structure do not affect the far field configuration , i.e. , on a spherical surface surrounding the boojum cores the order - parameter distribution resembles that of a point boojum . their effect ,
however , on the short - distance interaction may be significant .
indeed , we have shown that the short - distance attraction at @xmath213 , which results from the reorientation of the inner boojums , as the distance between the colloids decreases , corresponds to a structural change from single core to split core for the inner boojums pair .
we have shown how the non - linear effects become dominant , both in @xmath1 and @xmath0 , leading to a qualitative change of the effective colloidal interaction , namely , the crossover from the large - distance quadrupole - quadrupole repulsion for a range of orientations around @xmath213 , to an attraction at moderate to small distances , see figs .
[ interact_w_0.01 ] , [ 2d_inclusions_2 ] .
this results from the complex dynamics of the topological defects that rearrange their position at short distances both in @xmath1 and @xmath0 .
this attraction , along the far - field direction , agrees with experimental observations @xcite , and with earlier calculations @xcite where it is shown that the axial symmetry of the boojum - pairs is broken when the repulsion changes into the attraction .
the mechanism for attractive colloidal interactions driven by the re - arrangement of defects with quadrupolar symmetry
was reported in @xmath1 for homeotropic anchoring nearly a decade ago @xcite .
we emphasize an important difference between the @xmath1 and @xmath0 effective pair interaction potentials . in @xmath0
the global minimum occurs at contact @xmath352 for orientations at oblique angle with the far - field director . in @xmath1
the optimal orientation of two inclusions is also oblique , but there is a free energy barrier @xmath353 keeping the particles apart .
this distinction is in line with experimental observations in @xmath1 and @xmath0 @xcite .
finally , we have calculated the force between boojum colloids in @xmath0 , see fig .
[ force ] .
we have found , that as the distance decreases the angular dependence of the force changes rapidly and quite drastically . just below @xmath247
the maximum at @xmath246 becomes a minimum and the difference between the secondary maxima at intermediate values of @xmath203 becomes more pronounced .
this angular dependence was actually observed by @xcite at larger colloidal separations , @xmath354 .
the theoretical results appear to be too sensitive to the inter - colloidal distance , when compared to the experimentally measured forces .
we acknowledge partial financial support by fct - daad transnational cooperation scheme under the grant n : 50108964 , fct grants pestoe
/ fis / ui0618/2011 , ptdc / fis/098254/2008 and ptdc / bpd/50327/2007 ( nms ) , fp7 irses marie - curie grant pirses - ga-2010 - 269181 .
in this section we discuss briefly how the scheme for numerical minimization of the free energy functionals is implemented . in order to discretize the continuum models we resort to a finite element method @xcite with adaptive meshes . for @xmath0 systems the nematic and the colloidal particles of radius @xmath88 are confined in a cubic box of linear size @xmath355 . for smectic @xmath2 films the system is confined in a @xmath356 square box .
the surfaces of the spherical colloidal particles are discretized using the open source _ gnu triangulated surface library _ @xcite .
the library uses a recursive subdivision algorithm in order to triangulate the surface of a unit sphere . starting from the icosahedron as the first approximation the next levels of refinement
are constructed by subdividing each triangle into four smaller triangles .
we have used the sixth refinement level corresponding to a surface mesh with @xmath357 triangles . in the next step
the domain @xmath28 accessible to the nematic is triangulated . for @xmath0 systems
the triangulation is carried out using the _ quality tetrahedral mesh generator _ ,
tetgen , @xcite library .
tetgen generates boundary conforming delaunay @xcite meshes of @xmath0 domains with a piecewise linear boundary .
the boundary can also be specified as a surface mesh , which is the case for the triangulated surfaces of the spherical particles .
tetgen supports an isotropic size conforming triangulation , i.e. , the resulting elements conform to a given local size map . for @xmath1 systems
the corresponding nematic domains are triangulated using the _ inria _
@xmath358 software @xcite .
the @xmath358 package supports the creation of anisotropic meshes , where the triangulation process is governed by specifying a metric map .
we use linear elements both in @xmath1 and @xmath0 , i.e. the values of the nematic field @xmath359 are specified only at the vertices @xmath360 of the mesh and a linear interpolation is used in order to determine the values of @xmath361 at some other point @xmath362 of @xmath28 . due to the discretization
the volume and surface integrals in @xmath0 are replaced by sums of integrals over tetrahedral and triangular elements , respectively , and in @xmath1 by sums of integrals over triangular and line elements ( the discretized perimeters of the colloids ) .
each integral is evaluated numerically using generalized gaussian quadrature rules for multiple integrals @xcite . for integration over tetrahedra a fully symmetric cubature rule with 11 points , which is exact for polynomials of degree 4 ,
is used @xcite . for integration over triangles
a fully symmetric quadrature rule with 7 points , which is exact for polynomials of degree 5 @xcite , is used .
finally , one - dimensional integrals are evaluated using the 5 points gauss - legendre quadrature , which is exact for polynomials of degree 9 . during the minimization
the values of the tensor order - parameter @xmath363 at the vertices of the bounding box are kept fixed , equal to the values of the uniaxial bulk nematic .
the values of @xmath359 at all other vertices are obtained by numerical minimization .
we used the _ inria _ s * m1qn3 * @xcite optimization routine . the routine implements a limited memory quasi - newton technique ( limited memory broyden - fletcher - goldfarb - shannon method ) of nocedal @xcite ) .
the systems are characterised by two widely different length scales .
one is given by the nematic correlation length and the other by the radius of colloidal particles .
therefore , we resort to adaptive re - meshing technique in order to obtain a sufficiently good approximation for the `` exact '' solution in a reasonable amount of time and with the finite computing resources at hand .
the main objective of the adaptive refinement is to construct a mesh which meets the given interpolation error tolerance for a minimal number of elements .
such meshes are called optimally - efficient @xcite .
let @xmath364 be one of the components of the exact solution @xmath365 to our problem , and @xmath366 its linear interpolation on some mesh .
it can be shown that the interpolation error associated with some element @xmath367 ( either a tetrahedron or a triangle ) satisfies the following inequality @xcite @xmath368 where @xmath369 is some vertex of the element @xmath367 , @xmath362 is the position within element @xmath367 measured relative to @xmath369 , and @xmath370 is the hessian of the exact solution @xmath364 evaluated at the vertex @xmath371 .
if we define the absolute value of the @xmath1 hessian @xmath372 as @xmath373 where @xmath374 is an orthogonal matrix that diagonalizes @xmath375 , then the estimate for the interpolation error on the element @xmath367 can be written in a simpler form @xmath376 @xmath377 is the diameter ( the length of the longest edge ) of the element @xmath367 .
it is straightforward to write down similar expressions for the @xmath0 case
. therefore , if the sizes of the elements are chosen such that @xmath378 for all the elements , then the resulting interpolation error will be approximately equally distributed among all the elements . as it was mentioned above the _ bl2d _
package supports anisotropic meshes .
the triangulation is governed by specifying a metric map @xmath379 , and the resulting mesh is characterised by edges of unit length according to the metric @xmath379 @xcite .
we assume that the required metric is proportional to the hessian , @xmath380 . in order to estimate @xmath375 at some vertex @xmath381 we use the weak definition of the hessian @xcite @xmath382 where the integral is over the elements which share the vertex @xmath369 , and
@xmath383 is the piecewise linear hat - function associated with vertex @xmath384 : @xmath385 at @xmath369 and @xmath386 at any other vertex .
we use linear interpolation in order to obtaine @xmath375 at some internal point @xmath362 of an element .
tetgen implements an isotropic mesh refinement strategy which is based on a maximum local element volume constraint .
the constraints on the volume of elements are obtained by applying the equidistributing principle , where the nodes of the mesh are chosen such that for each element @xmath367 the following condition holds : @xmath387 in ref .
@xcite , it is shown that an optimally - efficient @xmath1 triangulation which minimizes the interpolation error fulfills asymptotically the equidistributing principle ( [ equidistr_princ ] ) .
we assume that the same is valid in the case of @xmath0 triangulation . | colloidal particles dispersed in a liquid crystal lead to distortions of the director field .
the distortions are responsible for long - range effective colloidal interactions whose asymptotic behaviour is well understood .
the short distance behaviour of the interaction , however , is sensitive to the structure and dynamics of the topological defects nucleated near the colloidal particles in the strong anchoring regime .
the full non - linear theory is required in order to determine the interaction at short separations .
spherical colloidal particles with sufficiently strong planar degenerate anchoring nucleate a pair of antipodal surface topological defects , known as boojums .
we use the landau - de gennes formalism in order to resolve the mesoscopic structure of the boojum cores and to determine the pairwise colloidal interaction .
we compare the results in three ( @xmath0 ) and two ( @xmath1 ) spatial dimensions .
the corresponding free energy functionals are minimized numerically using finite elements with adaptive meshes .
boojums are always point - like in @xmath1 , but acquire a rather complex structure in @xmath0 which depends on the combination of the anchoring potential , the radius of the colloid , the temperature and the lc elastic anisotropy .
we identify three types of defect cores in @xmath0 which we call single , double and split core boojums , and investigate the associated structural transitions .
the split core structure is favoured by low temperatures , strong anchoring and small twist to splay or bend ratios . for sufficiently strong anchoring potentials characterised by a well - defined uniaxial minimum , the split core boojums are the only stable configuration . in the presence of two colloidal particles
there are substantial re - arrangements of the defects at short distances , both in @xmath0 and @xmath1 .
these re - arrangements lead to qualitative changes in the force - distance profile when compared to the asymptotic quadrupole - quadrupole interaction . in line with the experimental results ,
the presence of the defects prevents coalescence of the colloidal particles in @xmath1 , but not in @xmath0 systems . |
azimuthal correlations between particles and the reaction plane , observed in non - central collisions , constitute one of the main subjects of heavy ion collision studies .
the primary nucleon - nucleon collisions being believed to be insensitive to the reaction plane , the latter phenomena are considered to be the source of information on _ collective effects _ present in the heavy ion reaction . a rich phenomenology @xcite has therefore been developed . among others ,
the success of the hydrodynamical approach @xcite in describing these phenomena is to be quoted here .
the azimuthal correlations are usually quantified in terms of the fourier coefficients of the azimuthal distribution of the outgoing particles w.r.t . the reaction plane .
these are defined by @xcite : @xmath2 \rangle \ ; \ ; , \ ] ] where @xmath3 denotes the azimuthal angle of the emitted particle , while @xmath4 is the orientation of the reaction plane defined by the impact parameter vector @xmath5 . specifically , the first - order coefficient @xmath6 reflects the _ sidewards collective motion _ of the particles and is known as directed flow .
a very sizeable amount of experimental data on directed flow is at present available from the low energy regime up to the lhc @xcite .
it is known that @xmath1 depends on collision centrality , particle type , transverse momentum and rapidity .
for instance , the rapidity dependence of pion directed flow at sps energies is known to consist of a smooth passage through zero at mid - rapidity @xcite , while large values of @xmath1 , of the order of 0.2 - 0.25 , are reported in the vicinity of target rapidity , in semi - central collisions @xcite . in the present paper we point at the possible _ electromagnetic component _ of directed flow .
we estimate the size of the spectator - induced electromagnetic interaction on the collective motion of positive and negative pions . the spectator - induced electromagnetic effect on charged pion ratios
@xmath7 was discussed in our earlier paper @xcite .
this effect is in fact observed in experimental data @xcite , and results in very large distortions in ratios of charged pions observed at high rapidities and low transverse momenta .
measurements of directed flow offer a new way of spotting the influence of the spectator charge on the emission of final state particles .
they also provide a new source of information on the space - time evolution of the non - perturbative process of pion production .
our present analysis will be focussed on peripheral pb+pb collisions occurring at the top energy available to sps heavy ion experiments ( @xmath0 gev ) .
we choose peripheral collisions because they are characterized by the largest spectator charge and also large values of directed flow .
we focus on the sps energy range by virtue of the abundance of experimental data on the rapidity dependence of @xmath1 , which also gives the possibility of comparison with our results as explained in section [ results ] .
we note that the energy range of the sps partially overlaps with that of the rhic beam energy scan , addressed in short in section [ bes ] .
this paper is organized as follows .
section [ model ] contains the description of our monte carlo model . in section [ results ]
we present the results of our simulation , together with the comparison to experimental data .
the discussion of our results is made in section [ discussion ] . in section [ bes ]
, we comment on the experimental data from the rhic beam energy scan .
our conclusions are presented in section [ conclusions ] .
a detailed description of our model can be found in @xcite . only the aspects that are relevant for the present analysis will be listed below .
our aim is to obtain a realistic estimate of the influence of the spectator - induced electromagnetic interaction on the directed flow of pions .
on the other hand , we wish to avoid the detailed discussion of the complex and poorly known mechanism of soft particle production .
therefore we decide on a maximally simplified approach : 1 . we assume , as an example , a peripheral pb+pb collision involving @xmath8 participating nucleons at @xmath0 gev . this
is shown in fig .
[ fig : model ] .
the two spectator systems are modelled as two homogeneous , lorentz - contracted spheres .
the reaction plane is defined by the collision axis ( dashed line in the figure ) and the impact parameter vector @xmath5 .
charged pions are assumed to be emitted from a single point in position space , that is , the original interaction point .
the time of pion emission @xmath9 is a free parameter of our model ; this parameter sets the initial conditions for the electromagnetic interaction . for peripheral pb+pb collisions studied here ,
the initial spectra of charged pions are assumed to be similar to those in nucleon - nucleon collisions .
charged pions are then numerically traced in the electromagnetic field of the spectator charges until they reach the distance of 10,000 fm away from the interaction point and from each of the two spectator systems .
the fragmentation of the spectator systems is neglected .
we do not consider the effects of participant charge , strong final state interactions , etc .
several clarifications should be added to the above . 1 . the simplified geometry of the peripheral collision presented in fig .
[ fig : model ] was determined on the basis of a dedicated study , discussed in detail in @xcite .
a geometrical monte carlo simulation served this purpose .
this used proper nuclear density profiles @xcite and assumed the elementary nucleon - nucleon cross - section equal to 31.4 mb , in good agreement with experimental p+p data at this collision energy @xcite .
the geometrical impact parameter corresponding to 60 participating nucleons was found to be @xmath10 fm .
additionally , the center of gravity of each of the two spectator systems was found to be displaced by @xmath11 fm relative to that of the original nucleus .
the average spectator charge was @xmath12 elementary units . + considering the exact shape of the spectator system as unimportant for our subsequent studies , and for the sake of clarity , we modelled the two spectator systems as homogeneously charged spheres . in the rest frame of each sphere , its density was the standard nuclear density @xmath13/@xmath14 .
the center of each sphere was additionally displaced by 0.76 fm in order to match the center of gravity of the spectator system .
thus our effective impact parameter ( distance of closest approach between the two spheres centers ) was @xmath15 fm .
2 . the simplification of initial conditions for pion emission ( reduction of the emission zone to a single point in space and time , as discussed above ) gives a convenient way to estimate the sensitivity of the electromagnetic effect to the basic characteristics of pion production ( the pion formation time , the distance between the pion formation zone and the two spectator systems , etc ) .
the initial momentum spectra of pions ( before the action of the electromagnetic field ) are assumed to obey wounded nucleon scaling @xcite and to be similar to underlying nucleon - nucleon events . as such , they are described by an analytical parametrization of average pion @xmath16 distributions in p+p collisions , recorded by the na49 experiment at the sps @xcite .
more details on this parametrization are given in @xcite .
for simplicity , distributions of positive and negative pions are assumed to be identical .
the experimental data cited above are expressed in terms of the feynman variable @xmath17 and of transverse momentum @xmath18 , and cover the region from @xmath19 to @xmath20 and from @xmath21 to @xmath22 gev / c . only a small extrapolation towards higher @xmath23
needs to be applied in the present analysis .
the uncertainty of this extrapolation has little or no effect on the results presented in section [ results ] .
it should be underlined that full azimuthal symmetry is assumed for the initial emission of pions .
3 . charged pions , with their initial momentum vector defined above in point ( b ) , and weighted by their initial distribution , are subjected to the electromagnetic field of the two spectator systems .
the pion trajectory @xmath24 is given by the classical relativistic equation of motion : + @xmath25 + where @xmath26 is the pion mass , and the pion momentum is defined by the lorentz force acting on the pion : + @xmath27 + here , @xmath28 is the pion charge , while @xmath29 and @xmath30 are standard superpositions of fields from the two spectator systems .
our equations take account of relativistic effects , including retardation .
technically , the propagation of the pion is performed numerically by means of an iterative procedure made in small steps in time , with variable step size .
the iteration proceeds until the pion is 10,000 fm away from the interaction point and from each of the two spectator systems in their respective rest frames .
negative pions which do not escape from the potential well induced by the spectator system are rejected and do not enter into our final state distributions .
in this section we present the results of our analysis .
these will be displayed as a function of the scaled rapidity @xmath31 , where @xmath32 and @xmath33 are the rapidity of the pion and of the incoming nucleus in the collision c.m .
we decide on this particular variable in order to simplify the interpretation of our results , and for an easier possible comparison with other collision energies ( see also the discussion made in @xcite ) .
the green solid curve shows the directed flow induced electromagnetically by the right ( r ) spectator .
the blue solid curve shows the directed flow induced electromagnetically by the left ( l ) spectator .
black dots show the result of the addition of these two curves .
the red solid curve displays the result of the simulation including both spectators .
note : all the simulations assume @xmath34 fm / c ; the blue solid curve is obtained by reflection of the green solid curve about midrapidity.,scaledwidth=95.0% ] we start by discussing the basic features of the spectator - induced electromagnetic effect on the directed flow of pions . in particular , this will include the role of each of the two spectator systems in the overall effect .
[ fig : onespec ] shows the directed flow @xmath1 of @xmath35 and @xmath36 , induced by their electromagnetic interaction with the spectator charges .
the value of @xmath1 is integrated over the transverse momentum of the pion , from @xmath21 to 1 gev / c .
all the simulations presented in the figure assume the simplest situation where the pion emission time @xmath9 is equal to zero ( immediate pion creation ) .
it should be underlined that as our simulation contains no initial azimuthal anisotropy ( sect .
[ model ] ) , any non - zero value of @xmath1 apparent in fig .
[ fig : onespec ] is indeed _ solely due to the electromagnetic interaction with the spectator systems_. therefore , we will refer to it as `` electromagnetically - induced directed flow '' in order to differentiate it from the `` standard '' flow phenomena which are caused by the strong interaction . in this context ,
several remarks are in order . 1 .
as it is clearly apparent from the figure , the electromagnetic field induced by the spectator charges exerts a strong influence on the directed flow of charged pions .
it can be held responsible for large values of @xmath1 , exceeding 0.2 ( 20% ) for negative pions close to beam rapidity .
the directed flow induced electromagnetically by the two spectators ( red solid curve , which goes through the black dots in the figure ) exhibits a characteristic structure as a function of rapidity .
qualitatively , this structure recalls the overall behaviour of pion directed flow as it is reported from experimental data ( see e.g. @xcite ) .
the directed flow displays a smooth transition through zero at @xmath37 , with large values of flow ( anti - flow ) at large negative ( positive ) rapidity .
the sign of electromagnetically - induced directed flow for negative pions is opposite with respect to positive pions . for absolute values of @xmath1 ,
differences in the shape of the two curves are seen close to beam rapidity .
what specifically follows from the above is that the spectator - induced electromagnetic interaction may result in the _ splitting _ of values of @xmath1 observed for positive and negative pions .
this splitting exhibits a strong rapidity - dependence , and reach very large values close to beam rapidity .
the directed flow which is induced electromagnetically by each of the two spectators _ separately _ is also presented in the figure ( the two spectator systems are denoted `` left '' and `` right '' as in fig .
[ fig : model ] ) .
the contribution of a single spectator does not remain confined to its hemisphere of the collision . on the contrary
, it extends into the opposite hemisphere ; thus the `` right '' spectator moving at positive rapidity exerts its influence on the pion @xmath1 down to and beyond @xmath38 , and vice - versa for the `` left '' spectator .
the value of @xmath1 obtained as the result of the electromagnetic interaction with both spectators ( red solid curve , which goes through black dots in the figure ) appears equal to the result of direct addition of the two single - spectator curves ( displayed by the black dots ) .
thus , as it it becomes evident from the above , the spectator - induced electromagnetic interaction appears strong enough to result in sizeable effects , and to play an important role in flow phenomena involving charged pions in heavy ion collisions .
this imposes the necessity of more detailed studies which will be presented below .
the top and bottom panels correspond to different values of the pion emission time @xmath9.,scaledwidth=95.0% ] the central issue of this paper is the sensitivity of the spectator - induced electromagnetic effect to _ the space - time evolution of the collision_. this will be addressed in the present section .
[ fig : initial ] shows the rapidity - dependence of the electromagnetically - induced directed flow of positive and negative pions , obtained assuming two different values of the pion emission time : @xmath34 and @xmath39 fm / c .
the values of @xmath1 are obtained by integration over @xmath18 of the pion , with integration limits defined as in section [ onespec ] above .
the curves in panels ( a ) and ( b ) of the figure are the same as the respective curves in fig .
[ fig : onespec ] .
the following remarks are to be made . 1 .
the electromagnetically - induced directed flow displays a clear dependence on the pion emission time . in the region of beam rapidity , the position of the minimum ( maximum ) in the valley ( peak ) in pion @xmath1 shifts from @xmath40 at @xmath34 fm / c to above @xmath41 at @xmath42 fm / c . at the collision energy discussed here , this shift in @xmath31 corresponds to a shift of about 0.6 units in rapidity .
the actual shape of the valley ( peak ) also exhibits sensitivity to @xmath9 .
2 . also in the region of lower rapidities , and down to midrapidity , the change in @xmath9 results in changes of @xmath1 . for instance
, at @xmath43 the absolute values of @xmath1 go from about 2.4% at @xmath34 fm / c down to about 0.5% at @xmath42 fm / c .
what follows from the above is that the electromagnetically - induced splitting of @xmath1 for @xmath35 and @xmath36 , mentioned in section [ onespec ] , will also exhibit sensitivity to @xmath9 .
these observations demonstrate that the spectator - induced electromagnetic effect is indeed sensitive to the initial conditions , set for the electromagnetic interaction by the emission time @xmath9 .
changes in the pion emission time , or equivalently in the position of the pion formation zone with respect to the two spectator systems , will result in changes of @xmath1 which should be observable in experiment .
in other terms , the electromagnetic effect on directed flow _ is indeed sensitive to the space - time evolution of pion production _ , and provides new information on the dynamics of the collision .
gev , shown at fixed values of pion transverse momentum : @xmath44 mev / c ( solid ) , @xmath45 mev / c ( dashed ) , @xmath46 mev / c ( dotted ) .
the top and bottom panels correspond to different values of the pion emission time @xmath9.,scaledwidth=95.0% ] up to now , for simplicity , only the rapidity - dependence of @xmath18-integrated directed flow was discussed in this paper . in the present section we study its dependence on the transverse momentum of the pion .
this is shown in fig .
[ fig : pt ] .
the electromagnetically - induced directed flow exhibits a strong dependence on pion @xmath18 .
in particular , this is valid close to beam rapidity where very large absolute values of @xmath1 ( of the order of 0.6 for positive pions ) are attained at the lowest considered value of transverse momentum , @xmath44 mev / c .
the absolute value of the electromagnetically - induced @xmath1 rapidly decreases with increasing @xmath18 . on the other hand , little or no dependence on pion
@xmath18 is apparent in the region of lower absolute values of @xmath31 , down to midrapidity .
the overall pattern of dependence on initial conditions , discussed in section [ initial ] above , repeats itself also at fixed values of @xmath18 . .
the three curves correspond to three different values of the assumed pion emission time @xmath9 .
note : the data points at positive @xmath31 are obtained by reflection about midrapidity.,scaledwidth=62.0% ] taking into account the charge - asymmetric nature of the electromagnetic effect discussed here , any comparison of our work to experimental data must involve measurements of directed flow of pions of _ a given charge _
, either @xmath35 or @xmath36 .
this constitutes an important practical difficulty as most of results published on the rapidity - dependence of @xmath1 involve ( if particle identification is available at all ) only `` charged pions '' , i.e. , summed @xmath35 and @xmath36 .
one data set which fulfills our requirement comes from the wa98 experiment , and contains a measurement of positive pions near target rapidity made by means of the plastic ball detector , in pb+pb collisions at @xmath0 gev .
this measurement is presented in @xcite , but a more detailed description of the analysis can be found in @xcite .
the latter reference specifies the corresponding centrality definition as 40 - 80% of the pb+pb cross - section . following the discussion made therein
, we expect that differences w.r.t . our definition of the peripheral collision ( section [ model ] )
would only have a small effect on the actual values of @xmath1 .
[ fig : comp ] shows the experimental data from wa98 , superimposed with the rapidity - dependence of the electromagnetically - induced directed flow as obtained from our work . for the latter , three values of the pion emission time
are assumed : @xmath34 , @xmath47 , and @xmath39 fm / c .
our results are integrated over @xmath18 from 0 to 1 gev / c .
the curves corresponding to @xmath34 and @xmath39 fm / c are the same as in fig .
[ fig : initial ] . as it can be seen in the figure
, the values of @xmath1 obtained from our simulation , and resulting exclusively from the spectator - induced electromagnetic interaction , appear comparable to those measured by experiment . the above statement should clearly be taken with some caution .
account should be taken of the simplicity of our model , and of various detailed issues related to the measurement @xcite .
these problems are beyond the scope of the present paper .
nevertheless , it seems clear that a very sizeable part of the measured directed flow of @xmath35 _ may come from the electromagnetic origin_. we note that this observation is consistent with our analysis of the influence of spectator charge on @xmath48 ratios in peripheral pb+pb collisions at the same energy . here
, pion emission times assumed in the range @xmath49-@xmath39 fm / c also result in good agreement between our model and experimental data @xcite .
summing up the principal results of our work , the following can be said : 1 .
we devised a simple model of the pb+pb collision ( section [ model ] ) .
this model does not contain any initial azimuthal anisotropy in pion emission but includes the electromagnetic interaction between the pion charge and the two spectator systems .
2 . this model predicts sizeable values of @xmath1 ( electromagnetically - induced directed flow ) for positive and negative pions in the final state of the collision . for positive pions ,
the predicted values are comparable to those seen in experimental data . as such , a very large part of positive pion directed flow may come from the electromagnetic interaction .
the sign of this effect is opposite for positively and negatively charged pions .
this will result in a splitting of values of @xmath1 observed for @xmath35 and @xmath36 .
the size of electromagnetically - induced directed flow , and therefore also the size of the electromagnetic splitting addressed above , depends on the pion emission time @xmath9 .
the latter parameter defines , in our model , the details of the space - time evolution of pion production , like e.g. the time of pion formation , or the position of the pion formation zone with respect to the spectator system(s ) .
5 . as it follows from the above , the electromagnetically - induced directed flow , and the resulting splitting of @xmath1 for @xmath35 and @xmath36 , _
bring new information on the space - time dynamics of the pb+pb reaction_. this has , in our view , important consequences for future studies of directed flow .
it is of course clear that our model contains important simplifications . among others
, it intentionally neglects the initial pion flow induced by the strong rather than the electromagnetic force .
the inclusion of such phenomena is the domain of more detailed models and we leave it for future studies .
however , it is clear from our work that the electromagnetic splitting of @xmath50 and @xmath51 addressed above will remain present also in more sophisticated phenomenological descriptions of the heavy ion collision .
this inspires a set of proposals and remarks listed below . 1 .
the study of rapidity - dependence of directed flow made _ separately _ for pions of a given charge , i.e. , @xmath35 and @xmath36 , will bring additional information on the dynamics of the heavy ion collision . this information will add up to the already rich phenomenological content of summed charged pion flow studies .
the electromagnetic splitting of @xmath1 for positive and negative pions will provide information about the position of the pion formation zone with respect to the spectator systems . specifically , also the _ lack _ of such splitting at a given rapidity would imply that the formation of pions occurs relatively far from the nearest spectator .
2 . the situation should be similar for other produced particles , like e.g. , charged kaons .
the situation for final state protons should be at least partially different as protons at forward rapidity will come also from spectator fragmentation @xcite . however , also in this case the role played by the electromagnetic repulsion of protons from the spectator systems should be investigated in detail .
3 . last but not least , our analysis brings attention to _ asymmetric _ nuclear collisions , be it a+b , d+a , or p+a reactions . in such collisions , the electromagnetic influence of the two spectators would not cancel out at midrapidity , and therefore the role of single spectators could be isolated , yielding further information on the space - time evolution of the collision ( see fig . [
fig : onespec ] for comparison ) .
all of the above observations are consistent with the results of our study of spectator - induced electromagnetic forces acting on charged pion ratios @xcite .
however , we feel that the precision of the present experimental measurements of the size of directed flow , the key role played by flow measurements in the present phenomenology of heavy ion collisions , and the overall interest of the community in such studies , make the consideration of the role played by electromagnetic interactions important for future analyses .
in the context of the discussion made above , and based on the comparison between the results of our simulation and existing sps experimental data , it is important to note that the rhic beam energy scan ( bes ) program can also address the search of the spectator - induced electromagnetic effects discussed here .
specifically , preliminary data on directed flow of positive and negative pions from the star experiment have already been reported in @xcite .
the measured values of @xmath50 and @xmath51 indeed display a signal of splitting which increases with increasing rapidity . at positive rapidity , @xmath50 remains below @xmath51 , which follows our predictions for the electromagnetically - induced directed flow , formulated in sections [ onespec ] and [ initial ] above .
presenting final conclusions on the nature of this effect probably requires more precise modelling of the space - time mechanism of charged pion production , including also its centrality dependence .
however , we note that the overall magnitude of the splitting of @xmath1 values seen by experiment looks ( roughly ) comparable to that predicted in the preceding sections , for the largest assumed values of the pion emission time @xmath9 and in the relatively limited range of @xmath31 available to the measurement @xcite . _
thus , we consider that the observed splitting is indeed a signature of electromagnetically - induced directed flow_.
the electromagnetic field induced by the presence of spectator systems may result in sizeable distortions of the directed flow pattern observed for pions of a given charge ( that is , @xmath35 and @xmath36 ) produced in heavy ion collisions .
this implies the presence of the electromagnetic splitting of @xmath1 for particles of opposite charges .
the size of this splitting depends on the space - time scenario of pion production and therefore brings new information on the collision dynamics .
the comparison to experimental data suggests that a very large part of directed flow observed for positive pions may come from the electromagnetic origin .
new experimental and phenomenological studies are needed in order to elucidate these questions , and possibly bring new insight into the evolution of the heavy ion reaction in space and time . on the experimental side , this implies measurements of particles of specific charges , for symmetric and asymmetric nuclear reactions . on the theoretical side , the possibility of verification of existing models of the heavy ion reaction by taking into account spectator - induced electromagnetic effects should be investigated .
furthermore , studies of the role of spectator - induced electromagnetic interactions for higher harmonics of the azimuthal distribution would also be highly indicated .
+ * acknowledgments * + this work was supported by the polish national science centre ( on the basis of decision no .
dec-2011/03/b / st2/02634 ) .
r. j. m. snellings , h. sorge , s. a. voloshin , f. q. wang , n. xu , phys .
* 84 * , 2803 ( 2000 ) .
l. p. csernai , g. eyyubova , and v. k. magas , phys .
c * 86 * , 024912 ( 2012 ) .
m. luzum and j .- y .
ollitrault , phys .
lett . * 106 * , 102301 ( 2011 ) .
e. retinskaya , m. luzum , and j .- y .
ollitrault , phys .
lett . * 108 * , 252302 ( 2012 ) .
w. florkowski , _ phenomenology of ultra - relativistic heavy - ion collisions , _
singapore , world scientific ( 2010 ) .
p. boek , phys .
c * 85 * , 034901 ( 2012 ) , + p. boek and i. wyskiel , phys .
c * 81 * , 054902 ( 2010 ) .
a. m. poskanzer and s. a. voloshin , phys .
c * 58 * , 1671 ( 1998 ) .
a. andronic _ et al . _
( fopi collaboration ) , phys .
c * 64 * , 041604 ( 2001 ) .
j. barrette _ et al . _
( e877 collaboration ) , phys .
b * 485 * , 319 ( 2000 ) .
j. barrette _ et al . _
( e877 collaboration ) , phys .
c * 59 * , 884 ( 1999 ) .
j. adams _ et al . _ ( star collaboration ) , phys .
c * 73 * , 034903 ( 2006 ) . c. alt _ et al . _
( na49 collaboration ) , phys .
c * 68 * , 034903 ( 2003 ) .
b. b. back _ et al . _
( phobos collaboration ) , phys .
* 97 * , 012301 ( 2006 ) .
b. i. abelev _ et al . _
( star collaboration ) , phys .
* 101 * , 252301 ( 2008 ) .
l. adamczyk _ et al . _ ( star collaboration ) , phys . rev
. lett .
* 108 * , 202301 ( 2012 ) .
i. selyuzhenkov , j. phys .
g * 38 * , 124167 ( 2011 ) .
h. schlagheck ( wa98 collaboration ) , nucl .
a * 663 * , 725 ( 2000 ) .
a. rybicki , a. szczurek , phys .
c * 75 * , 054903 ( 2007 ) .
t. osada , s. sano , m. biyajima and g. wilk , phys .
c * 54 * , 2167 ( 1996 ) .
s. mizutori , j. dobaczewski , g. a. lalazissis , w. nazarewicz and p. g. reinhard , phys .
c * 61 * , 044326 ( 2000 ) , + a. trzciska , ph .
d. dissertation , heavy ion laboratory , warsaw university , 2001 .
m. kusek - gawenda , e. kozik , a. rybicki , i. sputowska , a. szczurek , arxiv:1303.6423 [ nucl - ex ] . h. appelshuser _ et al . _
( na49 collaboration ) , eur .
phys . j. a * 2 * , 383 ( 1998 ) .
b. baatar _
( na49 collaboration ) , arxiv:1207.6520 [ hep - ex ] .
y. pandit ( star collaboration ) , acta phys .
supp . * 5 * , 439 ( 2012 ) . | we estimate the electromagnetic effect of the spectator charge on azimuthal anisotropies observed in heavy ion collisions . for peripheral pb+pb reactions at the top energy of the cern super proton synchrotron , @xmath0 gev , we predict this effect to bring very large distortions to the observed directed flow , @xmath1 , of positive and negative pions emitted close to beam rapidity .
the overall magnitude of this effect is comparable to values of @xmath1 reported by the wa98 experiment .
we argue that also at lower rapidities , the spectator induced electromagnetic effect may result in the splitting of values of @xmath1 observed for positive and negative pions .
such a splitting is visible in the data reported by the star collaboration from the rhic beam energy scan .
both effects are sensitive to the space - time scenario assumed for pion emission .
therefore , they bring new information on the collision dynamics . |
lyman break galaxies ( lbgs ) are currently among our best windows on the universe of galaxies at redshifts @xmath11 .
lbgs are especially useful for investigating galaxy formation and evolution because they are relatively easy to find in large numbers ( approaching 2 arcmin@xmath12 down to @xmath13 ; steidel et al 2003 ; giavalisco et al 2002 ) and because they are relatively bright in the optical , thus permitting optical spectroscopic followup .
spectroscopic redshifts of several thousand lbgs have been measured to date , with photometric redshifts of tens of thousands also available @xcite , and luminosity functions beyond @xmath14 have been measured @xcite . lbgs account for almost all of the star formation at @xmath15 observable in optical windows , and roughly half of the total star formation at those redshifts once dust - obscured systems such as sub - mm galaxies are included @xcite .
although a complete picture of lbgs and their relation to galaxies today remains elusive , many pieces of the picture are now available .
spectra of lbgs clearly show evidence of strong bursts of star formation , ranging from @xmath16 , confirmed with deep radio and x - ray imaging with the vla and _ chandra _ by @xcite , who estimate a mean star formation rate ( sfr ) @xmath17 for uv - selected galaxies at @xmath18 .
deep images with hst reveal small sizes @xmath19 kpc ( with @xmath20 ) , high luminosities @xmath21 , and diverse morphologies , including multiple small knots , diffuse halos , and asymmetrical linear features .
images detect only @xmath22 of lbgs , implying that the agn fraction is at most that much @xcite .
the spectra of lbgs also show evidence of strong outflows , e.g. ly@xmath23 emission lines red - shifted and interstellar absorption lines blueshifted with respect to systemic redshifts @xcite and a paucity of qso h i absorption lines near lbg sightlines @xcite , although @xcite find that only smaller outflow bubbles are needed to explain the spectra , i.e. 0.5 mpc comoving , implying that metals observed in the ly@xmath23forest are unlikely to come from lbgs . measuring and understanding the dynamical state and masses of lbgs is important for placing them in the context of galaxy formation . a galaxy s mass affects its rate of accretion of new material , its ability to retain gas against the expulsive force of supernova winds , and possibly its eventual morphological type .
the need to constrain lbg dynamics motivates the present work . in the local universe , total galaxy mass is usually measured kinematically by deriving rotation curves or emission line widths for disk galaxies @xcite , measuring internal velocity dispersions for elliptical and other spheroidal systems , and measuring velocity distributions of satellite galaxies or other test particles for all galaxy types @xcite , and then translating the observed velocity field into a dynamical or virial mass .
the masses of larger systems such as galaxy groups and clusters , but not of their constituent galaxies , are likewise constrained by direct measurement of their velocity dispersions , as well as by x - ray emission and gravitational lensing .
three kinds of direct evidence to date have been used to study the dynamics of lbgs and lbg host halos : emission line widths and spatially - resolved kinematics ( for total mass and dynamical state ) , spectral energy distributions ( for stellar mass only ) , and clustering analysis .
lbg mass measurements from redshifted h@xmath23 emission line widths and spatially - resolved kinematics generally fall in the range @xmath24 @xcite .
some signs of rotation curves are seen at @xmath25 @xcite , and there is also evidence for superwinds , especially ly@xmath23 emission lines blue - shifted with respect to stellar and interstellar absorption lines @xcite .
@xcite found that one - third of 62 galaxies at @xmath26 they studied with the sinfoni integral field unit ( ifu ) on the very large telescope ( vlt ) showed rotation - dominated kinematics , one - third are interacting or merging systems , and one - third are dominated by random motions ; derived h@xmath23 dynamical masses for the whole sample are @xmath27 , with median @xmath28 .
estimates of the stellar mass of lbgs have benefited from deep imaging with _
spitzer space telescope _ s infrared array camera ( irac ) , including especially the goods and extended groth strip ( egs ) fields .
stellar masses @xmath29 of lbgs at @xmath30 generally range from @xmath31 for luminosities @xmath32 , with typical median values @xmath33 @xcite .
ly@xmath23 emitting galaxies ( lae s ) show similar stellar masses @xcite .
samples of lbgs in the redshift range @xmath34 are now large enough to constrain star formation history in lbgs : the median lbg mass is reported to be lower by around a factor of 10 at @xmath35 than at @xmath15 @xcite , implying significant buildup of stellar mass through star formation and merging over that time , and mirroring the observed increase by roughly a factor of 10 in the comoving stellar mass density @xmath36 from @xmath37 to @xmath38 measured in the distant red galaxy ( drg ) population @xcite .
finally , the observed clustering of lbgs has been used to constrain the mass of dark matter halos in which the galaxies presumably reside .
two - point correlation analysis reveals that lbgs cluster with each other with typical correlation length @xmath39 mpc and implied halo masses @xmath40 @xcite , while still leaving the masses of individual lbgs within those halos uncertain .
clustering of lbgs with damped ly@xmath23 qso absorption line clouds ( dlas ) has also been detected , with implied dla halo masses on the order of @xmath41 @xcite .
lbgs are also observed to cluster on small scales : @xcite and @xcite found that the two - point correlation function of faint objects in the hdf , consisting largely of lbgs at high redshift , peaks between 025 and 04 ( @xmath42 kpc at @xmath43 ) with amplitude greater than 2 , i.e. lbgs appear as multiple clumps of emission rather than monolithic sources .
they interpret that multiplicity as evidence of starbursting regions within otherwise normal gas - rich galaxies .
many of the targets discussed in the current paper fall into that category of close pairs or multiple - knot sources .
galaxy formation theorists have used observations of lbgs extensively to check and revise their predictions and detailed descriptions of mass assembly , star formation , feedback , morphology , and clustering over cosmic time .
hydrodynamic @xmath44-body simulations @xcite and semi - analytic models ( sams ; cole et al 2000 ; somerville et al 2001 ; bower et al 2006 ) of cold dark matter ( cdm ) structure and galaxy formation have generally succeeded in describing the observed photometric , stellar mass , size , star formation , luminosity function , and clustering properties of lbgs , especially by including varying amounts of dust extinction . in particular , the maximum stellar masses of lbgs predicted in smoothed particle hydrodynamics ( sph ) models are @xmath45 at @xmath46 and @xmath47 at @xmath37 @xcite ; sizes are predicted to be @xmath48 kpc and halo and internal velocity dispersions to be @xmath49 and @xmath50 , respectively ; star formation rates are expected to be @xmath51 @xcite ; and sfr is predicted to correlate only weakly with stellar mass @xcite .
the available information therefore suggests that lbgs at @xmath52 range in stellar mass from @xmath53 with total mass from dynamical constraints @xmath54 , and that they reside in dark matter halos with mass @xmath55 , with significant clustering .
their star formation rates range from @xmath56 , with gas content and therefore star formation duty cycles and gas depletion timescales poorly constrained .
their comoving number density correponds roughly to that of luminous galaxies today , and their optical sizes and morphologies are diverse and compact .
the general picture then is that lbgs are a heterogeneous set of actively star - forming galaxies or sub - galactic clumps , building blocks that will likely coalesce with other lbgs and/or non - lbg clumps to form more massive galaxies , groups , and clusters by @xmath38 . to address further the open question of the dynamics of lyman break galaxies and their host halos and help place them with more detail in the larger context of galaxy formation and evolution ,
we have obtained spatially resolved optical ( rest - uv ) spectra with the 10-m keck telescope of a sample of lbgs and searched for kinematic information that we can use to constrain individual lbg and lbg host halo dynamics directly .
our sample is more than one magnitude fainter and is at higher redshift than the kinematic targets previously studied by others .
nir ifu observations , especially with adaptive optics ( ao ) @xcite , now provide direct access at @xmath57 to h@xmath23 and other rest - optical emission lines not strongly affected by dust , in contrast to ly@xmath23 . however , optical spectroscopy such as that presented here has the advantage of lower sky and telescope background emission and higher sensitivity instruments and detectors , as well as broader redshift coverage .
the observations are described in [ sec : obs ] , evidence for kinematic information is presented in [ sec : kin ] , and we discuss the results in the light of current models , theory , and other observations in [ sec : disc ] . throughout the paper we adopt the currently favored cosmological parameters @xmath58 , and @xmath59 .
our sample was drawn from 46 candidate and confirmed lbgs in the hubble deep field north ( hdf - n ) .
we used the same sample selection as @xcite : blue in `` @xmath60 '' ( @xmath61 ) but extremely red in `` @xmath62 '' ( @xmath63 ) , the colors expected for a blue , star - forming galaxy at @xmath64 with a lyman continuum break .
this color criterion is similar to that adopted by @xcite , but is slightly less restrictive in the blue @xmath62 , red @xmath60 region , which @xcite found allowed detection of `` ly@xmath23-break galaxies '' at higher redshift in addition to pure lyman break galaxies .
seven of the 46 sources were included via that relaxed criterion .
the galaxy magnitudes fell in the range @xmath65 , with a median value of 25.34 .
the sample is thus more than 1 magnitude fainter than that of @xcite and more than 2 magnitudes fainter than that of @xcite .
we chose both targets with confirmed spectroscopic redshifts and those without , since the success rate for lbgs selected in this way is typically so high ( @xmath66 for bright sources , @xmath13 ) .
the median reported redshift was @xmath67 .
many of the galaxies discussed here were studied by @xcite and @xcite but with lower spectral resolution and/or less attention to the position angle of the spectrograph slit .
we took full advantage of the fine pixel scale and high resolution of the hdf wfpc2 images , 0.04pixel@xmath68 with @xmath69 014 .
we preferentially chose lbgs that showed clear signs of one or more of the following : ( 1 ) significant spatial extent @xmath70 in the hdf - n image ; ( 2 ) multiple knots and/or close pairings of lbgs in the hdf - n ; and ( 3 ) previous spectroscopic evidence for some kinematic features .
we also prioritized bright targets over faint ones . in all we observed 32 kinematic lbg targets in the hdf - n .
of the 32 total lbg targets observed , 21 showed some promise in the data of allowing dynamical mass estimates ; the remaining discussion will focus exclusively on those 21 objects plus one serendipitously discovered lbg , which together we call the `` high - priority targets '' .
a complete list of all 22 targets is given in table [ tab : targs ] , and the 11 unused targets are listed for completeness in table [ tab : unusedtargs ] .
most of those 11 were not useful for kinematic measurements because they had insufficient signal - to - noise ratio , no strong emission line , or no detectable spatial extent in our spectra .
we compared our target list to the list of lbgs hosting agn in the hdf according to _ chandra _ x - ray flux as reported by @xcite ; no sources match within 10 down to 0.5 - 2 kev flux levels as low as @xmath71 , so we conclude that none of our targets hosts a luminous agn .
our observations were made during six observing runs from may 1997 to april 1999 at the 10-meter keck - ii telescope using the low resolution imaging spectrograph ( lris ; oke 1995 ) .
we designed six slit masks around the lbg targets ; high - priority targets were observed with multiple masks to increase integration time .
each target was examined closely in the hdf image , and then its corresponding slit was tilted to match the long axis of any extended emission from the target galaxy .
slits were also tilted to cover multiple targets in close pairs or groupings .
the masks were designed so that the position angle ( pa ) of most slits was within @xmath72 of the pa of the mask , meaning the slits were aligned close to perpendicular to the dispersion direction .
but some slits were tilted as much as @xmath73 from the mask pa and the normal to the dispersion direction .
each mask had between 23 and 25 slits total ( not including holes for alignment stars ) .
about half the slits on each mask were designed around targets in the hdf and flanking fields for other programs .
slit lengths for our lbg targets ranged from 8 - 35 , with a typical length of 15 .
all slits were 11 wide , although the slit tilts cause that width to be projected to a dimension smaller on the sky by a factor of cos(@xmath74 ) , where @xmath74 is the relative angle between the slit and the mask .
the mask pas were optimized for the anticipated mean hour angle of the hdf at the time of observation , in the sense that we tried to keep the majority of slits aligned with the parallactic angle to minimize slit losses due to differential refraction .
we relied on our own astrometric measurements of the hdf and flanking fields to derive target positions . for deriving wavelength solutions ,
hgnearkr arc lamps were observed through each mask .
strong night sky lines were used to fine - tune the wavelength solution of some spectra .
we used a 600 l / mm grating blazed at 5000 , resulting in a dispersion of 1.28 pixel@xmath68 , a typical resolution of 4 pixels @xmath75 5 full - width - at - half - maximum ( fwhm ) or 300 @xmath76 .
this represented a compromise between a lower - resolution grating that might have delivered higher signal - to - noise per pixel and a higher - resolution grating that would have provided more detailed kinematic information .
the grating angle was adjusted so that central wavelengths for slits at the mask centers were between 5000 and 6000 , giving a spectral coverage of 2620 and spectra ranging from a minimum of 3456 to a maximum of 6978 .
the spatial pixel scale for slits perpendicular to the dispersion direction was 0.215 arcsec pixel@xmath68 , although this scale is compressed ( more arcsec pixel@xmath68 ) by 1/(cos @xmath74 ) for tilted slits , where @xmath74 is the relative pa between the slit and the mask .
individual exposure times ranged from 1800 sec to 2700 sec , with at least two exposures through each mask to help reject cosmic rays .
spatial dithering was not possible due to the tilted slits .
targets received between two and 17 exposures each , depending on priority and repeat placement on multiple masks .
this resulted in total exposure times ranging from 9,900 to 50,000 sec per target .
the weather was mostly clear and photometric but some exposures were affected by cirrus .
the seeing was typically 1 .
images of all the targets are shown in fig .
[ fig : images ] , with the location and pa of each slitlet superposed .
targets observed at multiple pas are shown once for each pa .
we also included in our analysis some data from an earlier keck / lris observing run in april 1996 @xcite .
the slits for that run s slitmasks were tilted only to accommodate multiple objects , not to align to the major axis of extended objects .
nevertheless , some of those slits did fall at the same pa as the slits in the observing runs aimed at kinematic study , allowing us to add the datasets together .
the new observations are summarized in table [ tab : obs ] , and properties of the targets are listed in table [ tab : targs ] .
we reduced the keck / lris spectra using a combination of iraf tasks and custom iraf scripts and c routines .
raw images were bias - subtracted , cosmic - ray cleaned , flat - field corrected , and geometrically corrected to account for optical distortion .
the next stage was carried out in one of two ways : ( 1 ) the wavelength solutions for all the slits in a single exposure were measured simultaneously , the image was rebinned so that all slits had the same pixel - wavelength mapping and orientation perpendicular to the dispersion , and then background sky emission was subtracted and individual slits extracted as single two - dimensional spectra ; or ( 2 ) individual slits were extracted from each exposure and then treated as long - slit spectra : a wavelength solution was derived from a corresponding arc lamp image , the data were rectified , the wavelength solution was fine - tuned using night - sky emission lines , and background sky emission was subtracted .
finally , for each slit all individual exposures were combined using exposure - time weighting and spatial registration to make a single two - dimensional , sky - subtracted spectrum .
one - dimensional ( 1d ) spectra of most targets were extracted from the two - dimensional ( 2d ) image .
representative 2d and all the 1d spectra are shown in fig .
[ fig:2da ] and fig .
[ fig:1da ] , respectively .
in this section we present several new redshifts in the hdf , and then describe our measurements of kinematic features in the reduced spectra and how we use them to derive estimates of dynamical masses , along with some significant caveats on those estimates .
we examined each one - dimensional spectrum visually to measure or confirm the target s redshift . for difficult cases we tried cross - correlating the one - dimensional spectrum with various template spectra , including an average of confirmed lbg spectra from @xcite and a high - signal - to - noise ratio ( s / n ) spectrum of the gravitationally - lensed lbg ms 1512-cb-58 @xcite kindly provided by m. pettini . in no case
did the cross - correlation produce a reliable new redshift . in all
, we found three tentative and one robust new redshifts ; confirmed , refined , or constrained four more ; and were unable to obtain useful redshift constraints for two .
we were also able to measure the redshift for a new lbg the image of which fell on one of the slits but that was not in our original target list , for a total of five new redshifts .
table [ tab : zs ] gives the redshift obtained for every object in the sample .
the redshift quality @xmath77 is given on a scale from 1 to 4 such that @xmath78 means there is little hope of assigning a redshift given the signal - to - noise ratio in our data , 2 means real features are evident but the redshift is not secure , 3 means the redshift is probable , and @xmath79 means the redshift is definitely secure , with multiple spectral features identified .
we next searched for evidence in our lbg spectra that would allow us to constrain their dynamical state or even estimate their masses or their host halo masses .
measuring robust dynamical masses , however , is not straightforward , and the mass estimates we derive are subject to several major caveats .
we discuss some of the most significant of these below .
the velocities and velocity limits we measure may simply not reflect orbital motion under gravity .
peculiar motion of galaxies or subgalactic clumps in unvirialized systems such as loose groups or very young clusters could play a significant role .
chance near superposition of physically unrelated sources could also lead to erroneous mass estimates .
since many of our targets are indeed clumps of knots or close pairs , these factors may strongly affect our results and interpretation . to try to assess the significance of the non - virialization scenario , we turn to the `` phenomenological '' n - body and semi - analytic @xmath80cdm galaxy formation models of @xcite .
their `` model 3 '' represents a combination of two scenarios : star formation dominated by quiescent star formation in disks , usually just the central galaxy , and star formation dominated by merger - induced starbursts ( both galaxy - galaxy and halo - halo mergers ) .
one run of that model produced 13,257 simulated lbgs at @xmath81 `` detected '' in 1 deg@xmath82 with criteria similar to those used by @xcite .
we searched the catalog for projected close pairs of lbgs and examined their characteristics .
we find 1206 pairs with separation @xmath83 . in fig .
[ fig : deltav ] , we plot the relative velocity @xmath84 in each pair vs. the projected separation @xmath85 on the sky , where the velocities are calculated using the two galaxies redshifts and their peculiar velocities .
we focus on the simulated pairs with redshift difference @xmath86 ( where `` redshift '' means cosmological redshift only , without considering peculiar velocity ) , which we take to be truly physically related .
there are 103 such pairs in the simulated catalog , or 8.5% of all pairs with @xmath87 ; these are represented as open triangles in fig .
[ fig : deltav ] .
we also plot the values for the four close pairs in our sample ( see [ sec : disc ] ) .
we calculate for each simulated close pair of lbgs a dynamical mass @xmath88 where @xmath89 is the maximum observed circular velocity , which in this case we take as @xmath84 , and @xmath90 is the radius at which the velocity is measured , in this case the separation between the pair .
we also calculate the total ( dark matter ) halo mass of each pair as the sum of the two individual halo masses listed for the galaxies in the simulation catalog .
[ fig : mhmv ] shows calculated dynamical mass vs. total halo mass for the close pairs in the simulation of @xcite .
surprisingly , there is no strong correlation seen .
this implies that the velocities in fact do not ( or not always ) reflect true virial velocities
. they may instead reflect infall velocities of subclumps or galaxies streaming into a potential well ; such infall velocities are likely larger than virial velocities by at least a factor of two .
we conclude that , if the simulation of @xcite is correct , the pair - wise velocities in our lbg sample may provide at best weak constraints on the true total masses of our lbg systems .
a second caveat is that we may be sampling the kinematics of only a small region , either central to a larger potential well or embedded but not centered in a large massive proto - galactic cloud or even a large but uv - faint galaxy .
as discussed above in [ sec : intro ] , several recent groups have used optical and nir photometry to study the rest - optical seds of lbgs in the hdf - n , and found that their stellar masses were typically @xmath91 @xcite .
these stellar masses provide a firm lower limit to the total dynamical mass . in the local universe ,
rotation curves of massive disk galaxies rapidly reach maximum rotational velocity @xmath92 , typically within a few kpc @xcite , similar to the half - light radii of the lbgs studied here .
therefore , if lbg kinematics behave at all like those of local disks then we might not in fact be missing higher rotation velocities from faint extended regions below our detection limit ( although we could be missing faint emission at the same velocity from larger radii , which would raise the mass estimates ) . similarly , the central velocity dispersions @xmath93 of massive local elliptical galaxies are generally excellent indicators of the total galactic virial mass as measured by independent methods @xcite .
again this implies that even if lbgs reside in larger systems , the total mass may be well - sampled by the lbg kinematics .
finally , we can compare lbgs at @xmath15 to luminous compact blue galaxies ( lcbgs ) at redshifts @xmath4 , which have many properties similar to those of lbgs such as small half - light sizes @xmath94kpc , high luminosities @xmath95 ( where @xmath96 is the characteristic luminosity in a schechter luminosity function of galaxies today ) , blue optical colors , diverse , irregular morphologies , asymmetries @xmath97 , star formation rates @xmath98 , and narrow optical emission lines , @xmath99 @xcite .
@xcite found from deep optical imaging that more than half of the 27 compact narrow emission line galaxies ( cnelgs , close cousins of lcbgs ) at @xmath100 in their sample had sizes consistent with small local dwarfs , arguing against massive underlying host galaxies .
similar results derive from studies of nearby compact uv - luminous galaxies using galex @xcite .
the overall picture is that lcbgs at intermediate redshift are galaxies with luminosity @xmath95 but masses only 1/10th that of a typical @xmath96 galaxy , and dynamical / stellar mass ratios @xmath101 .
constraints on stellar mass do not address the question of extended massive dark matter or gas halos surrounding the lbgs , of course .
however , like measurements of local galaxy halo masses using satellite galaxy kinematics @xcite , our constraints on the dynamical masses of close pair lbg systems are insensitive to mass - to - light - ratio or stellar population uncertainties .
the third significant caveat regarding our mass estimates is that the ly@xmath23 emission linewidths we use to estimate dynamical masses of individual lbgs and sub - clumps are subject to the influence of bulk inflows and outflows of gas , strong absorption by dust , and multiple scattering by neutral hydrogen , all of which can change the line profile .
such effects have been studied theoretically by several groups ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
observational evidence includes the systematic asymmetry and redshifting of the ly@xmath23 emission line - presumably due to a combination of outflows and absorption by foreground dust in lbgs and local starbursts @xcite , and @xcite find that the ly@xmath23 emission of local starburst galaxies is spatially more extended than the continuum emission , presumably due to resonant scattering .
we searched for but could not find any published or unpublished direct comparison between observed linewidths of ly@xmath23 and those of h@xmath23 or other non - resonance nebular emission lines .
therefore our mass estimates based on ly@xmath23 emission line widths must be regarded as tentative .
finally , we generally lack constraints on the inclination angle @xmath74 of our targets , so each mass has associated with it an unknown sin @xmath74 correction factor .
apart from estimating or measuring inclinations of individual lbgs ( as done by , e.g. , bouch et al 2007 ) , this problem can be addressed only by assuming an average inclination or else assembling a large enough sample of kinematic measurements that all inclinations are statistically well represented .
given the caveats discussed above , we will proceed with caution in analyzing lbg dynamics from our sample . here
we make an empirical distinction between _ individual lyman break galaxies _ , with any multiple knots of emission connected by continuous flux visible in the hdf images and usually separated by @xmath1021 or less , and _ lbg systems _ , consisting of two or more lbgs without obvious extended emission bridging the gap(s ) between them .
both categories may correspond to star - forming galaxies or sub - galactic clumps embedded in low - mass or massive dark matter halos , and both may yield useful constraints on dynamical mass .
the lbg systems offer an additional opportunity to measure the total host halo mass as opposed to masses of subclumps or individual lbgs , although given the results of our examination of the simulation of @xcite , we will not assume that such results are robust .
we examined visually the two - dimensional spectra of all 22 of the high - priority targets for evidence of any of the following : ( 1 ) spatially extended emission lines , especially ly@xmath23 but also including c iv and he ii ; ( 2 ) spectrally unresolved , resolved , or multiple - peaked emission lines , especially from lbgs with multiple knot morphologies ; ( 3 ) extended continuum against which extended absorption lines might be detected ; and ( 4 ) emission and/or absorption lines from close pairs of galaxies .
the features detected are listed in table [ tab : zs ] . for each lbg or lbg system with evidence of kinematic information ,
we then estimated the dynamical mass .
for simple gravitational rotation , the dynamical mass would be @xmath103 as in eq .
[ eq : mdyn ] above .
we divide our mass estimates into two categories : ( 1 ) masses estimated from extended emission or absorption or multiple sources ; and ( 2 ) masses estimated solely from the ly@xmath23 emission line velocity width and observed half - light radius . the measured velocity and size parameters and derived mass estimates for our sample
are listed in table [ tab : mass ] .
we compare in fig .
[ fig : deltav ] the distribution of velocities , which range from @xmath104 to nearly 5000 @xmath76 , and projected separations of the four close pair systems with the prediction of model 3 of @xcite .
the pair with the largest @xmath84 appears to lie outside the locus of simulated pairs with small @xmath105 ( triangle symbols ; assumed to be physically associated , rather than just chance projection ) , but there are some simulated pairs with small @xmath105 and even larger @xmath84 than that target
i.e. , two lbgs within a few @xmath106 kpc of each other but with very high relative velocity , @xmath107 . to assess the probability @xmath108 that each pair in our sample is a true physical association , we can compare the density of small-@xmath105 pairs ( triangles ) to the density of large-@xmath105 pairs ( dots ) in different sections of the simulated @xmath109 plane .
we divide fig .
[ fig : deltav ] into eight equal segments bounded by 1 , 4 , and 7 and 0 , 2000 , 4000 , 6000 , and 8000 km / s .
pairs with separation @xmath110are excluded because we would be unlikely to resolve such pairs spatially given the typical seeing in our data .
we calculate the probability @xmath111 , where @xmath112 and @xmath113 are the number of small-@xmath105 and total pairs , respectively , in each section .
we find that @xmath108 ranges from 57% for the two real lbg pairs ( @xmath114 symbol ) in the lower left corner of fig .
[ fig : deltav ] to 6% for the pair in the upper middle section .
in other words , according to the simulation , two of the four pairs have a better than 50% chance of being true physical associations , while one has a 94% probability of being a chance superposition .
the value of @xmath108 for each pair is shown in fig .
[ fig : m * ] . for each of the eight lbgs with ly@xmath23 emission ( six of which
are also included in the spatially extended emission category ) , we fit a gaussian function to the emission line and measure the half - light radius @xmath115 in the hdf @xmath116 image .
we do not deconvolve the half - light radii with the wfpc2 psf @xmath117014 .
this will have only a small effect on the larger sources , but for the smaller sources our measurement will overestimate the true half - light radius and therefore the enclosed dynamical mass .
we did smooth the hdf images with a gaussian filter in an attempt to constrain ( and correct for ) the true seeing by matching the seeing - blurred spatial fwhm in the 2d spectra . however , significant uncertainties in the measured fwhms ( especially for low s / n sources ) and even variation within single 2d spectra made this approach impractical .
we deconvolve each observed linewidth with the measured 5 spectral resolution to derive intrinsic linewidths .
based on the discussion above of dynamical studies of local and high - redshift galaxies , we adopt the fwhm of the line profile as a reasonable but untested proxy ( modulo any outflow and extinction effects ) for circular velocity @xmath89 , we adopt @xmath115 for @xmath90 , and we calculate an estimate , again based on eq . [ eq : mdyn ] , of the dynamical mass within the half - light radius .
these measurements and the linewidth - based masses are also listed in table [ tab : mass ] , and all the derived masses are plotted vs. redshift in fig . [
fig : mz ] .
six of the eight ly@xmath23emission lines were at best marginally resolved , so we plot the derived dynamical masses for those sources as upper limits . here
we discuss the results for each target , listed in order of increasing ra , as in table [ tab : targs ] .
this relatively bright , elongated source has a disky isophote and captivated our interest as a potential high - redshift disk .
@xcite report for their object no .
2 a redshift @xmath118 , with low confidence ( @xmath119 , where 11 is the lowest confidence ) . our spectrum shows no sign of mg ii absorption at that redshift in either 2d or 1d , and no c iii ( although at 3634 , our spectrum has low sensitivity ) .
we conclude that the redshift assignment @xmath118 seems unlikely .
however , the final coadded spectrum shows weak emission lines at 4499 , 5869 , and 6045 , plausibly matching [ o ii]3727 , h@xmath120 , and [ o iii]5007 , respectively , at redshift @xmath121 .
we therefore conclude ( though with low confidence , given the lines weakness ) that the source is a low - redshift interloper that slipped through our color selection filter . this compact , high - surface brightness source is the brightest target in our survey .
@xcite were unable to measure a redshift , but found weak evidence for mg ii absorption at @xmath122 and @xmath123 .
@xcite report @xmath124 with low confidence ( @xmath119 , where 11 is the lowest confidence ) , and cite @xcite as the redshift source , but we are unable to find the target in that reference .
@xcite argued that the published spectrum does not support that redshift , and added that their photometric redshift technique favors @xmath125 .
meanwhile , @xcite report a photometric redshift using hst / wfpc2+nicmos @xmath126 .
we accumulated 24.5 ks of integration time observing the source with lris .
nevertheless , we find no strong emission or absorption lines , and we are unable to obtain an unambiguous redshift for the source , nor can we measure any kinematic signature .
no strong continuum break is visible in our spectrum , which extends down to 3320 with continuum detectable by eye in the 2d image down to 3920 .
since no strong ly@xmath23 emission or absorption is seen redward of that wavelength , we conclude that the emission redshift must be @xmath127 .
we again find weak evidence for mg ii absorption at @xmath122 and @xmath123 , as reported in @xcite .
unfortunately the mg ii lines for the lower redshift coincide with night sky lines , complicating their detection and measurement .
the lower redshift , which is consistent with the @xmath124 value of @xcite , is supported by possible fe ii absorption .
if real , either or both sets of redshifted absorption lines could be due either to intrinsic absorption from the emitting galaxy itself , or to intervening absorption if the emitting galaxy is at higher redshift .
the redshift of the galaxy goods j123640.85 + 621203.4 , with a position centroid 22 from hd4_1076_1847 , is reported by @xcite to be @xmath128 .
goods j123640.85 + 621203.4 has a highly linear morphology extending at least 15 in the hdf , aligned to within 10@xmath129 of the separation vector between the two sources .
the close agreement between its emission redshift and the tentative mg ii absorption redshift we find in the spectrum of hd4_1076_1847 suggests that it is either responsible for the absorption , or else associated with another as yet unidentified absorbing galaxy
perhaps hd4_1076_1847 itself .
we are therefore unable to confirm any of the previously published redshifts or derive a new one for this source , apart from the lower and upper limits mentioned above .
no kinematic information is available from our spectrum .
object c4 - 09 in the catalog of @xcite is a remarkable source consisting of four bright knots of emission with nearly identical colors , all within an area barely more than 1 across .
@xcite investigated and finally rejected the source as a possible gravitational lens system .
as reported by @xcite , the spectrum shows strong emission with two peaks at 5118.7 and 5132.6 .
continuum is clearly detected redward of the emission but not blueward , supporting the interpretation of the emission lines as ly@xmath23 at redshifts 3.211 and 3.222 , respectively , slightly lower than the @xmath130 reported by @xcite .
weak absorption lines are seen at 5314 , 5489 , 5624 , and 5633 , corresponding to si ii , o i , c ii , and si iv , respectively at an average absorption redshift @xmath131 , between the redshifts of the emission line peaks .
we observed c4 - 09 with two pas , one ( 98@xmath129 ) aligned with the longest axis of the parallelogram of knots , and the other ( 137@xmath129 ) closer to the short axis .
the four knots visible in the hdf image are so close together , with spacings between 057 and 117 , that the continuum and the red side of the emission line appear unresolved in both of our two - dimensional spectra .
the blue side of the ly@xmath23 emission line , however , appears to be slightly extended spatially in the long - axis spectrum , with fwhm=17 , compared with the seeing fwhm=11 , and offset towards the east with respect to the redder line and the continuum .
no such spatial extent or offset is visible in the short - axis spectrum .
no sign of ordered rotation is visible . the spatial extent and offset of the blue peak with respect to the red peak and the continuum
are consistent with a scenario in which some subset one , two , or three knots of the four emission knots is responsible for the blue peak , and another subset at a different velocity gives rise to the red peak .
the blue and red ly@xmath23 peaks are well fit in the spectral direction by gaussian profiles with fwhm 5.5 and 7.0 , respectively , i.e. , unresolved or marginally resolved .
after deconvolution with the 5 resolution , those widths correspond to @xmath132 and @xmath133 , respectively .
the 14 separation between the two peaks , however , corresponds to a velocity difference of 820 @xmath76 .
if the wavelength separation is indeed due to radial velocity differences among two or more of the four knots ( rather than , e.g. , an absorption line superposed on a broad emission line ) , then we can use it to estimate a system dynamical mass .
given the maximum spatial separation ( which will yield maximum mass ) of 117 , we derive a dynamical mass from eq .
[ eq : mdyn ] @xmath134 . alternately , the double - peaked emission profile is also well - fit by a single narrow absorption line with @xmath135 and @xmath136 superposed on a broad emission line with @xmath137 and @xmath138 .
such a broad emission line could be caused by strong gas outflows , and the absorption line could be caused by a foreground clump or screen of gas , even dust - free , that resonantly scatters ly@xmath23 photons out of the line center . in this scenario ,
the emission profile fhwm of 800 @xmath76 would reflect outflow velocity , perhaps coupled with resonant scattering , rather than circular velocity of a virialized system .
the redshifts from the ly@xmath23 emission and absorption lines would then be @xmath139 and 3.2170 , in excellent agreement with the redshift @xmath131 from the interstellar absorption lines but contrary to the commonly red - shifted ly@xmath23 emission reported by @xcite and @xcite , which they cite as evidence of outflows in lbgs .
thus an outflow model for c4 - 09 appears somewhat problematic .
we are not able with the current data to distinguish definitively between the two scenarios .
if outflow is indeed responsible for the line profile , then the true dynamical mass is most likely significantly lower than the value derived above .
this source is compact and relatively bright , but continues to elude efforts to obtain a definitive redshift .
@xcite list a redshift @xmath140 and @xcite report a corrected redshift @xmath141 with high confidence ( @xmath142 , where 1=highest confidence ) .
however , we see no al ii1671 absorption matching either redshift , only possible o i1302
/ si ii1304 and possible n v1243 in p - cygni profile matching @xmath141 ; no c iv1550 is detected .
@xcite give a photometric redshift @xmath143 .
continuum is clearly detected in our spectrum with a spatial profile matching an unresolved psf .
no strong emission or absorption lines are detected .
weak emission lines may be present at 5925 and 5976 . if these were [ o iii ] at @xmath144 , we would expect h@xmath120 at 5804 and / or [ o ii ] at 4450 , but neither is seen .
possible weak absorption lines appear at 4566 , 4678 , 5069 , and 5447 , but none is well - detected , nor were we able to discern a pattern among them indicating a source redshift .
we find no support for the tentative redshift @xmath145 reported by @xcite based on cross - correlation , apart from the possible line at 5069 , which would correpond to al ii1671 . without solid absorption or emission line features available
, we also find no kinematic indicators in our spectrum .
hd4_0298_0744 is another elusive target .
@xcite report @xmath146 with high confidence , while @xcite find @xmath147 .
continuum is clearly visible in the two - dimensional spectrum even before sky subtraction down to the blue limit of our spectrum , with no strong break visible .
no strong emission or absorption features are seen in the final two- or one - dimensional extracted spectrum .
we were therefore unable to confirm the redshift reported by @xcite , which we would have expected to produce a strong ly@xmath23 emission or absorption line or continuum break visible at 4621 . because no strong continuum break is visible down to 3900 in our spectrum ,
we conclude that the redshift is most likely @xmath148 .
we can not extract any dynamical information on the source .
c4 - 06 from the catalog of steidel et al .
( 1996 ) is one of the brightest , largest sources in our sample .
the hdf image shows a linear structure dubbed the `` hot dog galaxy '' @xcite
roughly @xmath149 in extent , with two major knots and several sub - knots .
the lris mask slit was aligned along that structure .
strong , spatially resolved continuum is clearly detected in the two - dimensional spectrum , although it appears as a single spatially continuous source , rather than the separate knots seen with hst resolution .
the spatial extent of the continuum emission , measured by compressing the two - dimensional spectrum into a one - dimensional spatial profile , is 32 , consistent with the hdf image .
no emission lines are visible , but several strong absorption lines are detected matching redshifted ly@xmath23 , si ii , o i , c ii , and si iv .
the ly@xmath23 line is broad ( 78 fwhm ) , while the other lines are unresolved .
additional marginal detections include si ii , , and si iii .
the redshift derived from averaging the strong narrow absorption line redshifts is @xmath150 , while cross - correlating the spectrum with that of a template average of 12 lbgs yields @xmath151 , similar to the redshift @xmath152 reported by @xcite .
the narrow absorption lines show no sign of any velocity shift across the extended continuum ( see figure [ fig:2da ] ) . to quantify any subtle velocity gradient
, we extracted three one - dimensional spectra at the middle and the two ends of the linear source and intercompared them .
the central extraction aperture was 12 pixels ( 26 ) wide and the apertures at the two ends were 6 pixels ( 13 ) wide , centered 8 pixels apart , so that the two end apertures had no overlap .
to constrain the velocity shift between the two end apertures , we selected regions of the spectra containing strong absorption lines and free from residual noise from strong sky emission lines and cross - correlated them .
the cross - correlation yielded a strong peak at @xmath153 . assuming a minimum relative wavelength accuracy of 1.0 , the velocity shift between the two apertures is then @xmath154 , consistent with zero within our measurement errors .
the dynamical mass limit implied by the observed constraint on line - of - sight velocity gradient @xmath154 over 322 is @xmath155 , the lowest dynamical mass estimate in our sample ( modulo , of course , the unknown inclination and three - dimensional morphology of the source ) .
we interpret the lack of observed velocity gradient over more than 20 kpc as evidence that we are observing not an edge - on disk with ordered rotation , but rather a truly linear , perhaps filamentary source .
the object may of course be collapsing or even expanding perpendicular to the line of sight ; collapse is more natural under most current galaxy formation scenarios .
this galaxy appears as a very close pair in the hdf image , with a separation of only @xmath156 .
@xcite report a redshift for the source @xmath157 .
our mask slitlet was aligned along the same pa , @xmath158 , as the two emission knots ; however , 1 ground - based seeing certainly blurs the two into a single unresolved source in our lris observations .
the spectrum shows very strong emission at 4823.2 and weaker emission at 4812.3 .
the stronger line appears extended both spatially and spectrally : the spatial fwhm of the line is 12 , comparable to the seeing , but the emission profile s faint wings extend to cover 33 .
the spectral fwhm is 7.0 , compared to the resolution of 5 , and the total detected emission spans 26.4 , including the two peaks and their wings .
the weaker line s spatial position centroid is offset 025 from that of the brighter line , closely matching the separation of the pair in the hdf image .
weak continuum is detected both redward and blueward of the emission lines , but appears to be stronger on the red side .
no other emission or absorption features are seen .
it is highly unlikely that this double emission line corresponds to an optical line emitted by a low - redshift ( @xmath4 ) system .
the [ o ii]doublet would , at @xmath159 , be separated by 3.5 , not the 11 we observe .
any [ o iii ] or balmer line should be accompanied by other optical lines as well .
we conclude that the most likely interpretation is ly@xmath23 at @xmath160 , confirming with only a slight revision the redshift reported by @xcite
. the double emission line profile may be due to a single emission line with strong absorption superposed by an intervening cloud or galaxy .
alternatively , as for c4 - 09 , it may be caused by radial velocity differences between the two emitting knots . given the close match between the spatial separations in the image and in the spectrum , we adopt the latter scenario .
the velocity difference between the emission peaks then corresponds to @xmath161 , while the entire 26.4 span of the emission line corresponds to @xmath162 , and the spectral fwhm of 7.0 corresponds , after deconvolving with the 5 instrumental resolution , to 4.9 , or @xmath163 . to derive a mass estimate for the system , we assume each knot in the pair emits one emission line , and we therefore adopt @xmath164 ( 2.3 kpc ) and @xmath165 to obtain @xmath166
. the large total range in observed emission velocity , @xmath167 , may be due to bulk gas outflow induced by merging of two or more sub - units and/or by the subsequent starburst . despite the relative brightness of this target ,
its redshift remains elusive .
@xcite reported a tentative redshift @xmath168 based on a few possible absorption lines . @xcite
estimated a photometric redshift @xmath169 , while @xcite give none .
the source is smooth , slightly extended , and elongated in the hdf image .
continuum is easily detected along the entire length of our spectrum .
however , no emission lines are visible , nor are any strong absorption features .
there are possible absorption lines at 4386 , 4528 , 4794 , 4816 , 5168 , and 5573 , but we were unable to discern any convincing pattern matching our redshifted template spectra or our previous tentative redshift @xmath168 .
cross - correlation with the template spectra likewise revealed no robust redshift .
the lack of a strong continuum break redward of 4000 implies @xmath170 , contrary to the photometric redshift @xmath169 calculated by @xcite . with no strong absorption or emission features , we were also unable to derive any kinematic information from the spectrum .
the hdf image shows the source hd4_0367_0266 to be an elongated object with two bright knots separated by 06 and an extended tail terminating in a fainter knot , all nearly aligned .
the redshift of @xmath171 reported by lowenthal et al .
( 1997 ) was based on several absorption lines .
the total extent of the source shown in the hdf is 23 ; the lris mask slitlet was aligned to cover all three knots as well as the tail .
our spectrum shows weak extended continuum emission with fwhm @xmath102 34 .
no emission lines are detected .
several possible weak absorption lines appear in the extracted one - dimensional spectrum at 4956 , 5117 , and 5478 .
these lines would correspond to si ii , o i , and si iv , respectively , at the previously reported redshift @xmath171 .
a possible broad absorption trough at 4757 could be a damped ly@xmath23absorber at @xmath172 .
overall our confidence in the redshift is somewhat lower than reported in lowenthal et al .
( 1997 ) , and we downgrade the redshift quality to @xmath173 . to search for velocity gradients or shifts across the galaxy , we extracted two spatially - independent one - dimensional spectra centered 24 apart and cross - correlated them , using only the clean regions of the spectra that were free from sky noise residuals and showed evidence of absorption features .
the cross - correlation function displays a weak peak consistent with @xmath174 , but unfortunately the signal - to - noise ratio was too low to allow any robust measurement of kinematics . the source hd2_1928_1041 ( c2 - 06 in the catalog of steidel et al 1996 ) is separated by only 20 spatially from c2 - 05 , and we consider them here as a close pair .
each of the galaxies is a compact source with one bright knot and a small cloud of extended emission as seen in the hdf .
@xcite reports @xmath175 for c2 - 05 , but this is revised to @xmath176 by @xcite . for hd2_1928_1041 , @xcite report @xmath177 , citing @xcite ; we here slightly revise that redshift .
both targets are covered in a single lris slitlet on our slit masks .
strong , unresolved continuum is detected from both objects .
no emission lines are visible , but numerous absorption lines corresponding to interstellar species such as si ii , o i , c ii , si iv , c iv , and al ii are well detected in the two spectra , corresponding to redshifts @xmath176 ( hd2_1928_1041 ) and @xmath178 ( c2 - 05 ) .
the velocity difference implied by the redshifts is @xmath179 .
cross - correlation of absorption line regions in the two one - dimensional extracted spectra produces a strong peak at @xmath180 ( where the error is estimated by dividing the fwhm of the cross - correlation peak by 10 ) , consistent with the simple redshift difference . the observed velocity difference and spatial separation provide a dynamical mass estimate @xmath181 for the two - source system . a highly elongated , multiple - knot source embedded in extended emission , extending over at least 26 along the long dimension , to which our lris slitlet was aligned .
spectroscopic redshifts reported in the literature include @xmath182 @xcite with low confidence and @xmath183 @xcite with high confidence ( @xmath184 , where 1=best ) . @xcite
cite @xmath185 , and @xcite cites @xmath186 , but we could not find the source of either measurement in the literature .
@xcite reported that a redshift was not measurable in their spectrum of the source .
photometric redshifts calculated for hd2_1739_1258 include @xmath187 @xcite , @xmath188 @xcite , and @xmath189 @xcite .
strong , extended continuum emission is detected in our 39 ks integration with a spatial @xmath190 9 pixels , or 19 and a total extent of 39 .
the continuum dies below about 4075 , implying @xmath191 if the dip is due to ly@xmath23 blanketing .
however , we also detect two possible weak absorption lines at 5305 , 5318 , which match mg ii at @xmath192
. these absorption lines could be due to intervening gas , but they could also be intrinsic to the target .
we detect no features that support either the redshift @xmath193 @xcite or @xmath194 @xcite . given the lines we do detect weakly , we tentatively conclude that hd2_1739_1258 in fact lies at @xmath192 rather than at @xmath11 .
we extracted two spatially independent 1-dimensional spectra from the extended two - dimensional spectrum and cross - correlated them in an attempt to measure velocity gradients across the source , but no statistically significant cross - correlation signal was detected .
this galaxy consists of a single compact core with two short wings of extended emission .
@xcite report @xmath195 .
our lris slitlet was aligned with the long axis of the system .
the spectrum shows a single weak emission line at 3934.2 with @xmath196 , i.e. barely resolved , with clear continuum redward but none blueward . interpreting the line as ly@xmath23 , which is supported by the continuum break
, we derive a redshift @xmath197 .
no other strong emission lines are detected , although possible weak c iv and al iiabsorption and he ii emission are visible .
the width of the line corresponds to a deconvolved velocity width of 240 @xmath76 , which we adopt as an upper limit .
neither the continuum nor the strong emission line shows any spatial extent in our two - dimensional spectrum , so we are unable to constrain kinematics for this system .
c3 - 02 is a small source with a single compact bright knot of emission and a diffuse tail extending barely 1 , along which our lris slit was roughly oriented .
@xcite reported a tentative redshift @xmath198 , @xcite report a photometric redshift @xmath199 , and @xcite report @xmath200 .
spatially unresolved continuum is easily detected with average s / n@xmath201 per resolution element over the entire length of our 25.5 ks spectrum , which extends from 3650 - 6285 .
no emission lines are visible .
several possible weak absorption lines appear at 4317 , 4663 , and 4608 , the first of those being the strongest ( @xmath202 ) . none of those wavelengths matches features expected from galaxies at @xmath198 .
both the photometric and spectroscopic redshifts previously reported therefore seem unlikely to be correct .
there is a possible match to o i1302 , si ii1304 , c ii1335/1336 , and si iv1402/1403 at @xmath203 .
no ly@xmath23emission is visible at the expected wavelength for that redshift , but the continuum shape is consistent with a moderate break there .
given the weakness of the lines , we adopt this redshift only tentatively ( @xmath204 ) .
kinematic measurements of this source are also impossible from our spectrum .
hd2_0698_1297 consists of two distinct knots embedded in an asymmetrical blob of diffuse emission .
the knots are 05 apart , and the diffuse emission visible in the hdf f814w image extends for 12 .
the system is 27 from hd2_0705_1366 , an unresolved point source ( measured @xmath117014 ) in the hdf image .
because of slit mask orientation constraints , we observed this pair at two different position angles : for 25.5 ks at 42@xmath129 , and for 10.2 ks at 58@xmath129 . due to the close proximity of the two sources ,
the 11-wide slitlets at both pas include emission from both targets .
both spectra show weak continuum with @xmath205 per resolution element and a single emission line from each source .
the emission line from hd2_0698_1297 is stronger in the spectrum with pa=42@xmath129 , while that from hd2_0705_1366 is stronger at pa=58@xmath129 , presumably because of slit placement with respect to each source .
we measure the wavelengths of the emission lines from the stronger spectrum in each case : 5396 ( hd2_0698_1297 ) and 5312 ( hd2_0705_1366 ) , each with @xmath136 , corresponding for redshifted ly@xmath23 to @xmath206 and 3.370 , respectively , within @xmath207 of the values reported in @xcite .
there is evidence for an absorption trough bluewards of the emission line in hd2_0698_1297 , supporting our identification of the line as ly@xmath23 .
we can not confidently rule out the possibility that each emission line is in fact spectrally unresolved , so we adopt the measured fwhm as an upper limit in each case . the continuum and the emission lines
are all spatially unresolved , except perhaps continuum from hd2_0698_1297 , which shows a slight spatial extent @xmath69 15 , although this may be due at least partly to small spatial registration errors in co - adding the 18 individual exposures . given the wavelength difference 84 and projected separation 27 between the two sources , and assuming the wavelength difference to reflect gravitationally - induced dynamics , we derive a system dynamical mass @xmath208 , the largest estimated dynamical mass in our sample .
see hd2_0698_1297 ( above ) .
see hd2_0624_1688 below .
this triple - knot compact source is part of a close pair with hd2_0529_1567 , 625 away .
both were covered by the slitlet at @xmath209 that targeted hd2_0725_1818 and hd2_0743_1844 ( see below ) .
hd2_0624_1688 , for which @xcite reported @xmath210 , was in our original kinematic sample target list , while hd2_0529_1567 was not : at @xmath211 , it offered little hope of providing a useful spectrum .
however , our 2d spectrum shows clear continuum and a single strong emission line from each source . the continuum emission from hd2_0624_1688 shows a significant drop blueward of the emission line .
continuum from the fainter hd2_0529_1567 is of course much weaker , and the proximity of the emission line to the blue end of the spectrum prevents us from assessing any continuum break across the line .
the single emission line from hd2_0624_1688 lies at 4162.7 , yielding redshift @xmath5 assuming that the emission line is ly@xmath23 , compared with @xmath210 reported by @xcite .
no other absorption or emission lines are visible , and ly@xmath23 is the only plausible interpretation .
the emission linewidth is @xmath212 , corresponding to 3.8 or @xmath213 after deconvolution with the 5 resolution , and we adopt that value as an upper limit .
the emission line from hd2_0529_1567 lies at 3973.3 , yielding @xmath140 , again under the assumption that the line is redshifted ly@xmath23 .
the deconvolved linewidth is @xmath214 or @xmath215 , which we adopt as an upper limit .
no other absorption or emission features are seen . together with the projected separation of 625 ( 50 kpc ) , the observed velocity difference of 1394 @xmath76 yields a dynamical mass estimate for the pair system @xmath216 .
this small ( @xmath217 026 ) source lies only 13 from hd2_0743_1844 .
@xcite reported a spectroscopic redshift @xmath218 for hd2_0725_1818 based on a ly@xmath23 emission line , but only a tentative redshift @xmath219 for hd2_0743_1844 .
the two sources have similar colors , with hd2_0743_1844 bluer by only 0.27 mag in @xmath220 and redder by only 0.05 mag in @xmath221 , less than the spread in the colors of lbgs reported by @xcite ; hd2_0743_1844 is fainter by 0.8 mag in @xmath116 ( all measurements from the catalog of williams et al 1996 ) .
our lris spectrum covers not only both of those sources , but also a spiral galaxy to the north with a redshift @xmath222 reported by @xcite and , to the south , hd2_0624_1688 and hd2_0529_1567 ( see above ) .
continuum from both galaxies is easily visible in the two - dimensional spectrum , but with some overlap of their spatial profiles , given their mere 13 separation .
the flux from hd2_0725_1818 is 2 - 3 times stronger than that hd2_0743_1844 , consistent with their photometric measurements .
several absorption features are apparent in the two - dimensional image ; while they seem to span spatially the continuum from both sources , the weakness of the fainter source makes this appearance difficult to confirm visually .
we extracted a one - dimensional combined spectrum of the pair . to assess their redshifts and velocities independently ,
we also extracted two separate one - dimensional spectra in the following way : first we extracted a one - dimensional spectrum of the brighter source , hd2_0725_1818 , using only those pixels in the two - dimensional spectrum extending spatially from the peak of the source away from hd2_0743_1844 , i.e. the side of the spatial profile that is less contaminated by flux from the fainter source .
we then used this 1d extraction to model the entire 2d spectrum and subtracted the scaled model from the 2d image .
this left a clean , isolated spectrum of hd2_0743_1844 , which we then extracted to 1d .
we repeated the process in reverse , modelling and subtracting the 2d spectrum of hd2_0743_1844 to produce a clean , isolated and uncontaminated spectrum of hd2_0725_1818 .
the absorption lines in the 1d spectrum of hd2_0725_1818 are well matched by interstellar siii , cii , and alii , and stellar siiv , civ at @xmath223 , thus confirming and slightly revising the redshift reported by @xcite .
we also find a tentative detection of c iii in emission , similar to their earlier result .
ly@xmath23 lies at the blue end of the spectrum , which is clean but low @xmath224 ; no emission or break is visible in either 2d or 1d .
the spectrum of the fainter source , hd2_0743_1844 , also shows absorption corresponding to si ii , c iv , and o i. no al ii is detected , though the @xmath224 is low in that part of the spectrum .
the implied redshift is @xmath225 , consistent with that of hd2_0725_1818 , and we adopt that redshift with quality @xmath226 rather than 4 given the source s extreme faintness . to measure any small velocity difference between the two sources , we cross - correlated the independent 1d spectra against each other , using only those parts of the spectra free of bright sky emission line residuals .
the result was a strong correlation peak ( correlation strength = 0.2 ) at @xmath227 , consistent with @xmath228 within our nominal velocity resolution of @xmath229 . for comparison
, we also cross - correlated each 1d spectrum with a 1d spectrum of blank sky extracted from the same 2d image .
no strong correlation peak is seen anywhere near @xmath230 , with a maximum of only 0.017 within @xmath231 of @xmath174 , and a maximum correlation strength 0.1 for the whole spectrum .
we thus conclude that the correlation between hd2_0725_1818 and hd2_0743_1844 is significant , and that they have identical redshifts within our uncertainties . adopting an upper limit @xmath154 and the projected separation 13 or 10.9 @xmath232 kpc ,
we then derive a dynamical mass upper limit @xmath233 , the lowest derived mass in our sample .
see hd2_0725_1818 above .
an elongated , wispy source consisting of two clumpy wings extending over 21 , with a 05 gap separating them , and aligned roughly east - west .
the source s total isophotal colors in the hdf vesion 1 catalog satisfy the @xmath234-dropout criteria of @xcite .
however , in the hdf version 2 catalog , the two sides colors are significantly different , with the western side bluer by 1.30 mag in @xmath235 and 0.52 mag in @xmath236 .
furthermore , in the version 2 catalog photometry , neither component individually satisfies either the @xmath237-dropout or the @xmath234-dropout criterion .
our lris slit covered both wings , as well as the source hd3_1824_1945 74 away and two other sources , one on either end of the slit .
@xcite report @xmath238 for the source .
only extremely very faint continuum is detected at the position of hd3_1633_1909 in the 2d spectrum , with no absorption lines discernible .
a single emission line appears , however , at 6145 . the emission line , though apparently spectrally resolved with @xmath239 , is not well fit by the [ o ii ] doublet at @xmath240 ; the peaks of the two components
would be separated by 4.1 , while the line in the spectrum is well fit by a single gaussian . if the line is ly@xmath23 , the source s redshift is @xmath3 , the highest redshift in our sample and consistent with the redshift reported by @xcite .
the resolution - deconvolved velocity width of the emission line is @xmath241 .
no spatially extended kinematic information is available from our spectrum .
this compact double - knot source was covered by the same slit that included hd3_1633_1909 ( see above ) , 74 away .
@xcite report @xmath242 ( object 145 in their online list ) .
our spectrum shows strong continuum with spatial @xmath11716 extending without any obvious break virtually to the blue limit of the spectrum at 3750 .
no strong emission lines are visible ; only a few weak possible absorption lines are present .
we see no compelling evidence to support the redshift @xmath243 reported by @xcite ; for example , al ii1671 , c iv , and si iv , which often appear in strong absorption in lbg spectra , are not seen .
the lack of a strong ly@xmath23 continuum break down to 3750 implies @xmath148 , in conflict with the photometric redshifts @xmath244 calculated by @xcite and @xmath245 by @xcite .
meanwhile , the lack of strong emission lines including [ o ii ] 3727 implies @xmath246 .
no kinematic information is available from our spectrum . of the 22 targets in our high - priority sample
, we confirm that 14 are lbgs with @xmath11 and @xmath173 or 4 , of which 13 provided some evidence of kinematics that we use to estimate dynamical mass or upper limits thereon .
two candidates are probably interlopers at @xmath4 , and the remaining six show insufficient features for confident redshift ( or kinematic ) measurement .
the failure rate , 6/22 = 27% ( or 8/22=36% if we include the two interlopers ) , is higher than the typical failure rate for lbgs with @xmath13 , @xmath247 .
this must be due to our color selection function , which is somewhat different from that of @xcite and may be less restrictive , as mentioned above ; our magnitude limit , which is 1 - 2 mags fainter than most previous spectroscopic surveys of lbgs ; our use of the 600 l / mm grating , which provides less spectral coverage than the lower - resolution gratings commonly used for lbg redshift surveys ; or a combination of those effects .
eleven of the 13 lbgs with kinematic signatures show spatially extended emission or absorption : eight are in close pairs , two contain three or four knots of emission each , and one is a clumpy elongated source extending over 3or @xmath102 25 kpc .
a different subset comprising eight of the 13 are sources with ly@xmath23 emission lines whose linewidths may provide some constraint on dynamical mass .
the redshifts of the combined subsets range from @xmath176 to @xmath3 with a median of @xmath5 .
the angular extent of emission ranges from 0.12 - 6(@xmath248 kpc ) , and the adopted circular velocities range from @xmath249 to 4700 @xmath76 , with a median of 330 @xmath76 .
we divide the mass estimates into two overlapping categories based on the source morphologies and angular extent : individual lbgs , comprising 11 sources , and lbg systems , comprising the four close pairs .
we derive dynamical masses for the lbgs ( using ly@xmath23linewidths and extended and multiple - knot sources ) ranging from @xmath250 to @xmath251 , with a median @xmath252 .
we see no evidence in any of our targets for ordered rotation such as measured by @xcite and @xcite for some of their more luminous star - forming galaxies at lower redshift ( @xmath57 ) .
dynamical mass estimates for the lbg systems range from @xmath253 to @xmath254 , with a median mass @xmath255 .
we emphasize that all these values are subject to the caveats discussed in [ sec : techniques ] , and that two of the four close pairs have probability @xmath256% of being chance projection rather than physical associations , according to our comparison with the models of @xcite .
the tentative dynamical masses we derive for individual lbgs , @xmath257 to @xmath7 , are very similar to those measured for brighter targets by @xcite based on h@xmath23kinematics measured with vlt / sinfoni , @xmath258 , implying that either ly@xmath23 and h@xmath23 emission linewidth measurements are both robust or else they both suffer from complications of comparable scale . @xcite also extrapolate the velocities beyond the half - light radii to estimate the virial mass of a typical halo , finding @xmath259 , similar to the halo masses inferred by @xcite based on clustering .
in contrast , the median dynamical mass @xmath260 that we measure for our four close - pair systems is an order of magnitude larger , and the maximum system dynamical mass estimate is @xmath261 , more than @xmath262 above the mean halo mass of @xcite .
we note , however , that in no case in our sample is the derived dynamical mass less than the reported stellar mass for the same galaxy ( [ sec : mstar ] ) , which is reassuring .
having measured redshifts and searched our spectra for dynamical signatures in 13 lbgs and lbg systems , we now compare those results to several other observational results and to theoretical expectations in order to explore implications for galaxy formation scenarios .
many lines of evidence point to the importance of merging in galaxy evolution over the entire cosmic history of galaxies . in this section
we explore the possibility that our sample of lbgs includes a significant merger fraction , and we compare the characteristics of our sample directly to the predictions of galaxy formation theories , including masses and close pair fractions .
cold accretion of gas and dark matter not associated with major majors is also increasingly appreciated as a significant source of fresh material for star formation and mass buildup @xcite . but major mergers , where the components have roughly equal mass , can erase disk morphologies and produce starbursts , agns , and elliptical galaxies , and even minor mergers can induce bursts of star formation or disturbed morphologies @xcite .
recent works have measured the evolution of the rate of galaxy mergers , the fraction of close pairs of galaxies , and the distribution of galaxy morphologies over cosmic time @xcite .
essentially all of those results are consistent with hierarchical galaxy formation models , in which galaxy mass is assembled largely through merging . as noted above , early studies of lbgs recognized the high rate of close pairs and multiple knots @xcite . of
the 14 confirmed lbgs in our sample , 12 are in close pairs or have multiple - knot or highly extended , clumpy morphologies .
all 12 are therefore excellent candidates for merging systems .
indeed , every close pair in our sample yielded potential kinematic information ; conversely , the targets that failed to provide such information were predominantly low - surface brightness , elongated , diffuse sources , which may themselves be interacting or merging systems . our entire sample , of course , is strongly biased away from isolated , symmetrical morphologies by virtue of our visual selection process , so it is to be expected that the sample would show a higher fraction of merging systems than do magnitude - limited lbg samples .
our sample contains only four close pair systems , so it is subject both to small number statistics and especially to the assumption that the pairwise velocities indicate dynamical mass ( see [ sec : techniques ] ) .
the components of the four pairs may be subclumps in various stages of infall and virialization , in which case their relative velocities could be dominated by random rather than circular motions .
they may , however , also reflect a true , large range in the masses of the lbg host halos .
the theoretical implications of observed lbg close pair incidence and velocity distributions for galaxy formation models have been explored in detail by several groups @xcite .
@xcite found that different models predicted very different halo occupation distributions , i.e. the number of lbgs per dark matter halo .
in most @xmath80cdm models , the number of lbgs predicted per halo increases with mass , with typically @xmath42 lbg halo@xmath68 for @xmath263 and @xmath13 .
the number of close pairs of galaxies defined as galaxies within a given angular separation within a redshift interval @xmath264 is an especially sensitive discriminator between models .
specifically , compared to the observed number of close lbg pairs , models postulating a one - to - one correspondence between lbgs and massive dark matter halos ( `` massive halo '' models ) underpredict close pairs by @xmath265 , while the `` colliding halo '' model overpredicts close pairs by @xmath266 within 15 .
the best agreement was provided by their `` collisional starburst models '' , in which lbgs host both ongoing quiescent star formation and rapid bursts of star formation triggered by minor and major galaxy mergers . to compare the observed incidence of close pairs in our sample with the models of @xcite
, we first note that only two of the four systems we call `` close pairs '' satisfy the @xcite criterion @xmath264 . fig .
6 in that paper shows the fraction @xmath267 as a function of pair separation for different models .
all four of the pairs in our sample have separation @xmath268 , so they all fall in the first bin in the figure .
if we assume that we have observed all the close pairs in the hdf survey volume in that separation bin , i.e. the total number of close pairs with @xmath264 is @xmath269 , and further that the total number of galaxies in the survey volume with comparable selection criteria is roughly the number of lbg candidates in the hdf times the success rate we observe here , or @xmath270 , then we find @xmath271 .
this value is already marginally inconsistent with ( higher than ) the values , all @xmath272 , predicted by the collisional starburst , constant efficiency quiescent , and accelerated quiescent models in fig . 6 of @xcite ; the massive halo model underpredicts close pairs by a factor of more than @xmath273 ( including only the error bars of wechsler et al 2001 ) . the colliding halo model , on the other hand , overpredicts it by a factor of less than @xmath274 .
obviously , with such small numbers of close pairs and large uncertainties on the selection function and completeness of our sample , the observed incidence ratios should be viewed as only tentative .
the close pairs and multiple knot systems we study here all have maximum angular separation @xmath83 , and therefore all fall in the first bin in the separation - pair fraction plot ( fig .
6 ) of @xcite , where indeed the discrimination among models is the strongest . because of the biased nature of our target selection , we do not attempt to compare quantitatively the observed pair fraction with the predicted one .
but we do note that close pairs are represented over the entire range of dynamical masses we derive , @xmath275 .
this distribution is likely due either to ( 1 ) misinterpretation of the kinematic signatures ( i.e. the velocities are not due primarily to orbital motion under gravity ) , resulting in erroneous mass estimates or ( 2 ) misidentification of individual knots of emission as separate galaxies , thus complicating the comparison with the models ; or else ( 3 ) it is inconsistent with the general prediction of @xcite that the number of lbgs per halo should increase with halo mass .
@xcite use observed sizes and sfrs of lbgs to conclude from their models that lbgs must be hosted by halos with circular velocities @xmath276 , although for some combinations of gas - to - total and dust - to - total mass ratios the distribution can reach 400 @xmath76 .
if lbg close - pair dynamics are indeed indicative of host halo mass , then the range of halo masses in our sample is significantly larger and the typical mass significantly higher than predicted by @xcite . @xcite further explore the possibility of using the observed pairwise velocity distribution ( pvd ) to discriminate between different galaxy formation models , but their investigation focuses on scales between 100 @xmath232 kpc and 30 @xmath232 mpc , much larger than the 10-kpc scale of multiple - knot and close pair targets studied here .
@xcite analyze dark matter halo @xmath44-body simulations and report that mass accretion occurs in a rapid phase , during which the halo s circular velocity is also predicted to rise rapidly , followed by a slow phase , during which the circular velocity remains almost constant .
however , they examine only isolated halos with total mass fixed at @xmath277 , for which the maximum circular velocity predicted is @xmath163 and the maximum halo radius is @xmath278 kpc .
these spatial scales are roughly comparable to those of three systems in our sample : two close pairs and the highly extended source c4 - 06 .
the velocity scale is comparable in only one close pair system .
most of the lbgs and lbg systems in our sample are therefore not apparently well described by the models of @xcite . however
, if we interpret our results within their framework , the large range in system dynamical masses we derive could imply that we are seeing the systems at different evolutionary stages , some before the rapid mass accretion phase is completed and the circular velocity plateaus , and some after .
typical group - sized halos today have circular velocities @xmath279 , and cluster halos have circular velocities @xmath280 .
only one of our lbg system velocities lies well above that range , with @xmath281 ; the rest have velocities consistent with those of normal galaxies to normal clusters of galaxies today .
as noted above in [ sec : kin ] ( fig .
[ fig : deltav ] ) , the distribution of pairwise velocities and projected separations of our four pairs is consistent with the predictions of model 3 of @xcite , supporting their scenario of quiescent disk star formation combined with merger - induced starbursts .
the dynamical masses we derive span the range of masses of dwarf galaxies to giant elliptical galaxies today .
our modest sample is certainly biased and is not representative of all lbgs at @xmath81 .
the range of estimated dynamical masses is similar to that derived for simulated lbgs in model 3 of @xcite ( [ sec : kin ] ) , but the kinematics of both simulated and real datasets may in fact not reflect true virial masses , especially given the lack of correlation with combined halo mass in the simulation ( fig .
[ fig : mhmv ] ) .
however , taken together , the sizes , sfrs , morphologies , velocities , and derived individual and system dynamical masses of the lbgs in our sample imply a large range of physical scenarios and dark matter halo masses .
they also further suggest that , rather than being identified one - to - one with the centers of future massive spheroids , bulges , or elliptical galaxies , lbgs are merging sub - clumps within halos that are destined to become galaxies with a large range of masses and potentially of morphological type as well .
this scenario is consistent with hybrid galaxy formation models such as the favored collisional starburst model of @xcite and @xcite , but not with their massive halo model .
here we compare our kinematic mass estimates to the stellar masses of the same galaxies calculated by sed fitting .
@xcite expanded on the sample of @xcite , fitting stellar population models to seds of uv - bright galaxies ( i.e. lbgs ) at @xmath282 in the hdf based on fluxes from hst / wfpc2 and nicmos .
the stellar masses for the 53 galaxies with @xmath283 in the nicmos f160w band range from @xmath284 , with a median around @xmath285 .
the stellar mass estimates for 765 of 1682 nicmos - detected sources with spectroscopic or photometric redshifts have since been supplemented by photometry from _ spitzer_/irac and were kindly provided to us by c. papovich ( 2007 , private communication ) .
we searched the _
spitzer_-supplemented nicmos+wfp2 catalog for positional matches to our high - priority lbg sample .
all 22 targets have at least one match well - detected by nicmos within 06 .
we visually examined and confirmed the position of each match in the hdf f814w images .
of the 22 lbg targets , 20 have stellar mass estimates in the papovich catalog , the other two lacking reliable redshifts .
neither of those two , however , is among the 13 lbgs in our sample showing kinematic signatures .
therefore we can directly compare all the individual and system ( i.e. , halo ) dynamical masses we estimated to the stellar masses estimated by @xcite . the hdf nicmos mosaic ( hnm ) catalog numbers , redshifts used , and stellar masses of all matched objects are listed in table [ tab : m * ] .
six targets have two matches in the papovich catalog within 13 . of those ,
three are in our list of 13 with dynamical mass estimates : c4 - 09 , c4 - 06 , and hd3_1633_1909 , each producing two matches .
both matches in each case correspond to emission knots we consider to be part of the same lbg . for c4 - 09 and c4 - 06 ,
the redshifts used by @xcite to calculate sed fits and stellar masses were for both matches essentially the same redshifts we report here .
for hd3_1633_1909 , however , @xcite used the same redshift we find ( @xmath3 ) for only one of the two matches , while for the other they used @xmath286 , i.e. what we consider to be a single lbg they consider to be two sources at widely disparate redshifts .
the resulting stellar mass estimate for the low - redshift source ( hnm 1499 ) is @xmath102 1/3 that of the high - redshift source ( hnm 1506 ) .
we consider below the stellar mass estimate of only the high - redshift source .
the low - redshift source certainly contributes flux to our spectrum .
however , as our dynamical mass estimate of hd3_1633_1909 is based only on the ly@xmath23 emission line width , the effect on our mass estimate should be negligible
. on the other hand , the redshift @xmath286 of the low - redshift source may be in error ; if the correct redshift is in fact close to @xmath287 , then the stellar mass estimate for hd3_1633_1909 would increase by roughly a factor of two .
we make the following three comparisons of the stellar mass estimates to our dynamical mass estimates : ( 1 ) for the eight ly@xmath23 linewidth - based dynamical mass estimates or upper limits , we compare those values directly with the single matching stellar mass estimates ( seven sources ) or the two components of a double match ( one source , c4 - 09 ) for a total of nine comparisons ; ( 2 ) for the three individual but extended or clumpy sources with single or multiple matches in the nicmos catalog , we combine the stellar mass estimates and compare the total to our individual lbg dynamical mass estimated from extended absorption or emission velocity or , if neither of those is available , from ly@xmath23 linewidth ; and ( 3 ) for the close pairs , we combine the two stellar mass estimates and compare the sum to the system dynamical ( halo ) mass from the pair s relative velocity .
the four - knot source c4 - 09 , which yields both ly@xmath23 linewidths and a double emission line peak , presumably from different emission knots , has two nicmos matches , so we apply both comparison types 1 and 2 ( i.e. we separately compare the smaller dynamical and stellar masses perhaps due to individual knots and the combined stellar and system dynamical masses ) .
these comparisons are shown in fig .
[ fig : m * ] .
the individual lbg stellar masses ( comparison types 1 and 2 above ) range from @xmath288 to @xmath289 with a median of @xmath290 .
two sources show stellar masses roughly equal to their dynamical masses . the stellar mass range and median , however , are both less than the range and median of dynamical masses ; the typical mass ratio @xmath291 , comparable to those of lcbgs at @xmath4 @xcite .
the stellar masses are also small compared to those measured for 181 lbgs in the egs by @xcite , @xmath292 with median @xmath293 .
no obvious correlation is seen in fig .
[ fig : m * ] between our dynamical mass estimates and the stellar masses from @xcite , although a trend could be masked by the six upper limits on dynamical mass from ly@xmath23 linewidths .
@xcite likewise compared stellar and dynamical masses for 68 galaxies at lower redshift @xmath25 , finding a weak correlation between the two , with a mean stellar mass @xmath294 more than 20 times the median stellar mass for our fainter , higher - redshift sample from @xcite and a typical mass ratio @xmath295 vs. 5 for our sample .
they also found that @xmath296 of their sample had large dynamical - to - stellar mass ratios @xmath297 ( vs. 10/12 = 83% for our sample ) , suggesting that this was likely due to young galaxy ages that had not yet built up significant stellar mass , or else faint older stellar populations that escaped detection .
the same scenarios may explain the large dynamical - to - stellar mass ratios and lack of correlation we observe .
our sample is more than one magnitude fainter than that of @xcite , so we are probing farther down the lbg luminosity function .
local galaxies and even starburst galaxies show a wide range in mass - to - light ratio , from @xmath298 in solar units for extreme starbursts such as h iigalaxies to @xmath299 for quiescently star - forming galaxies like the milky way @xcite .
the large range is presumably due to a combination of galaxy age ( older systems showing stronger nir flux from stars ) , burst age ( younger systems showing stronger uv flux from stars ) , merging history , and dust obscuration .
the large range we observe in dynamical - to - stellar mass ratio may in fact reflect a wide range of intrinsic lbg properties including dark matter halo or sub - clump mass , star formation history , merging history and status , and burst age .
we also compared the system masses derived from the four close pair relative velocities with the combined stellar mass estimates from papovich et al ( 2005 ; comparison type 3 above ) .
the right - hand panel of fig .
[ fig : m * ] shows the comparison . here the range in combined stellar mass is smaller than for the individual lbgs , while the range in system dynamical mass is larger .
the average total stellar mass is @xmath300 , and the average system dynamical - to - stellar mass ratio is @xmath301 , ranging from @xmath302 to @xmath303 .
this large mass ratio range suggests that one or both of two scenarios are at play : ( 1 ) the lbgs are embedded in dark matter halos with a wide range of masses that are not strongly correlated to the luminosity and star formation rates of the lbgs they host ; and ( 2 ) the relative velocities of the close pairs may reflect various stages of infall , i.e. not all the close pairs are necessarily in virialized systems ( see [ sec : techniques ] ) .
we have studied candidate and confirmed lbgs at @xmath52 in the hdf - n with deep two - dimensional spectroscopy , finding 3 tentative and 2 robust new redshifts .
we also find evidence for kinematics in 13 lbgs and four lbg close - pair systems . the observed morphologies and dynamics , including numerous cases of close pairs and multiple sources ,
appear to be most consistent with collisional starburst models of galaxy formation , rather than massive halo models with only one lbg per halo .
we derived tentative dynamical mass estimates from the observed velocity signatures , finding a range from @xmath304 to @xmath251 with a median @xmath305 for the individual lbgs and a range from @xmath253 to @xmath254 with a median mass @xmath255 for the close - pair systems .
all the mass estimates are subject to strong caveats including unknown inclination corrections , ly@xmath23 absorption , outflows , and non - virialized systems .
all our dynamical mass estimates are larger than the corresponding stellar mass measurements .
however , we find no evidence for a correlation between stellar mass and dynamical mass , in contrast to the weak correlation found by @xcite for their sample of brighter , lower - redshift sources .
we argue on the basis of comparison with local compact starburst galaxies , other studies of distant star - forming galaxies , and galaxy formation theories and simulations that the sizes , sfrs , morphologies , velocities , and dynamical mass estimates imply a wide range of physical scenarios and mass scales giving rise to lbgs at @xmath52 .
given the large time investment at keck needed to observe even this small sample , further progress in constraining the dynamical masses of lbgs will require technological advances including the deep diffraction - limited imaging and spectroscopy promised by _ james webb space telescope _ and increasingly available through adaptive optics and nir integral field spectroscopy at ground - based 10-meter class telescopes .
many thanks are due to the staff of the keck observatory for their help and expertise in carrying out the observations .
we appreciate assistance from sandra faber , drew phillips , and nicole vogt during the observations , and from dan kelson , gabriela mallen - ornelas , and nathan roche in reducing the data .
thanks to m. pettini for providing the comparison spectrum of cb-58 , to c. papovich for the electronic catalog of lbg stellar masses , and to d. hogg for help with the color image recipe . j.d.l .
acknowledges support from nsf grant ast-0206016 . this research has made use of nasa s astrophysics data system bibliographic services ; the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with nasa ; and iraf .
lowenthal , j. d. , durham , r. n. , lyons , b. j. , bershady , m. a. , gallego , j. , guzman , r. & koo , d. c. 2005 , _ starbursts : from 30 doradus to lyman break galaxies _ , eds .
r. de grijs and r.m .
gonzlez delgado , p. 17
tenorio - tagle , g. , terlevich , e. , terlevich , r. , mas - hesse , j. m. & leitherer , c. 2004 , in _ the formation and evolution of massive young star clusters _ , asp conf ser v 322 , eds . h.j.g.l.m .
lamers , l.j .
smith , and a. nota , p. 525
hd4_0259_1947 & 4 - 916 & 12:36:38.605 & 62:12:33.83 & 25.080 & 110 , 26.0 & 21.6 + hd4_1076_1847 & 4 - 878 & 12:36:40.963 & 62:12:05.30 & 24.009 & 26.0 , 137.0 & 24.5 + c4 - 09 & 4 - 858 & 12:36:41.245 & 62:12:03.07 & 24.904 & 98.0 , 136.9 & 21.6 + hd4_0818_1037 & 4 - 445 & 12:36:44.640 & 62:12:27.39 & 24.586 & 190.0 , 26.0 & 24.5 + hd4_0298_0744 & 4 - 316 & 12:36:45.087 & 62:12:50.81 & 24.983 & 145.0 & 24.5 + c4 - 06 & 4 - 555.1 & 12:36:45.409 & 62:11:53.18 & 24.190 & 43.0 & 25.5 + hd4_1006_0680 & 4 - 289 & 12:36:46.947 & 62:12:26.08 & 26.082 & 73.0 & 9.9 + hd4_1486_0880 & 4 - 382 & 12:36:46.951 & 62:12:05.34 & 24.957 & 13.0 , 26.0 & 25.5 + hd4_0367_0266 & 4 - 52 & 12:36:47.720 & 62:12:55.79 & 25.579 & 26.0 , 108.0 & 21.6 + hd2_1928_1041 & 2 - 454 & 12:36:48.266 & 62:14:18.42 & 25.081 & 165.0 & 25.5 + c2 - 05 & 2 - 449 & 12:36:48.338 & 62:14:16.63 & 24.249 & 165.0 & 25.5 + hd2_1739_1258 & 2 - 585 & 12:36:49.811 & 62:14:15.18 & 24.399 & 6.5 , 148.0 & 38.9 + hd2_1398_1164 & 2 - 525 & 12:36:50.123 & 62:14:01.03 & 25.596 & 83.0 & 9.9 + c3 - 02 & 3 - 550 & 12:36:51.335 & 62:12:27.51 & 25.706 & 174.0 & 25.5 + hd2_0698_1297 & 2 - 604 & 12:36:52.451 & 62:13:37.84 & 25.787 & 42.0 , 58.0 & 41.9 + hd2_0705_1366 & 2 - 637 & 12:36:52.760 & 62:13:39.09 & 25.916 & 42.0 , 58.0 & 41.9 + hd2_0529_1567 & 2 - 751 & 12:36:54.205 & 62:13:35.83 & 27.233 & 28.9 , 31.0 & 38.9 + hd2_0624_1688 & 2 - 824 & 12:36:54.627 & 62:13:41.37 & 26.139 & 28.9 , 31.0 & 38.9 + hd2_0725_1818 & 2 - 903 & 12:36:55.077 & 62:13:47.00 & 25.406 & 28.9 , 31.0 & 50.0 + hd2_0743_1844 & 2 - 916 & 12:36:55.184 & 62:13:48.07 & 26.214 & 28.9 , 31.0 & 50.0 + hd3_1633_1909 & 3 - 853 & 12:36:58.967 & 62:12:23.42 & 25.708 & 77.0 & 21.6 + hd3_1824_1945 & 3 - 875 & 12:37:00.127 & 62:12:25.22 & 24.671 & 77.0 & 21.6 + hd4_0460_1146 & 4 - 491 & 12:36:43.253 & 62:12:38.85 & 25.680 + hd4_1994_1406 & 4 - 631 & 12:36:45.310 & 62:11:38.47 & 26.269 + hd2_1949_0599 & 2 - 239 & 12:36:45.886 & 62:14:12.09 & 25.342 + hd4_1341_0299 & 4 - 85 & 12:36:49.515 & 62:12:20.11 & 26.103 + hd3_0408_0684 & 3 - 243 & 12:36:49.814 & 62:12:48.80 & 25.985 + hd2_0664_0879 & 2 - 373 & 12:36:50.302 & 62:13:29.73 & 25.951 + hd2_1410_1282 & 2 - 594 & 12:36:50.683 & 62:14:03.15 & 26.098 + hd3_0378_1536 & 4 - 677 & 12:36:51.620 & 62:12:17.31 & 25.979 + hd3_0457_2023 & 3 - 915 & 12:36:53.107 & 62:12:00.74 & 25.770 + hd2_0434_1377 & 2 - 643 & 12:36:53.422 & 62:13:29.53 & 25.457 + hd3_1455_0430 & 3 - 118 & 12:36:54.727 & 62:13:14.72 & 25.141 + lcccl hd4_0259_1947 & 0.207 & 2 & & new @xmath306 + hd4_1076_1847 & @xmath307 & 1 & & possible mg ii abs + c4 - 09 & 3.226 & 4 & ly@xmath23 & quadruple knot + hd4_0818_1037 & ( 2.500 ) & 1 & &
+ hd4_0298_0744 & @xmath308 & 1 & &
+ c4 - 06 & 2.794 & 4 & si ii , o i , c ii , si iv & `` hot dog '' + hd4_1006_0680 & 2.969 & 3 & ly@xmath23 & double knot and em .
peak + hd4_1486_0880 & @xmath309 & 1 & & + hd4_0367_0266 & 2.931 & 3 & & + hd2_1928_1041 & 2.005 & 4 & si ii , o i , c ii , si iv , c iv , al ii & c2 - 06 ; pair with c2 - 05 + c2 - 05 & 2.008 & 4 & si ii , o i , c ii , si iv , c iv , al ii & pair with hd2_1928_1041 + hd2_1739_1258 & 0.897 & 2 & & new @xmath306 + hd2_1398_1164 & 2.236 & 3 & ly@xmath23 & + c3 - 02 & 2.316 & 2 & & @xmath198 unlikely ; new @xmath306 + hd2_0698_1297 & 3.439 & 4 & ly@xmath23 & pair with hd2_0705_1366 + hd2_0705_1366 & 3.370 & 4 & ly@xmath23 & pair with hd2_0698_1297 + hd2_0529_1567 & 2.268 & 4 & ly@xmath23 & new @xmath306 ; pair with hd2_0624_1688 + hd2_0624_1688 & 2.424 & 4 & ly@xmath23 & pair with hd2_0529_1567 + hd2_0725_1818 & 2.232 & 4 & o i , si ii , al ii & pair with hd2_0743_1844 + hd2_0743_1844 & 2.230 & 3 & o i , si ii , al ii & new @xmath306 ; pair with hd2_0725_1818 + hd3_1633_1909 & 4.056 & 3 & ly@xmath23 & + hd3_1824_1945 & @xmath310 & 1 & & + lcccccccl hd4_0259_1947 & & & & & & & & + hd4_1076_1847 & & & & & & & &
+ c4 - 09 & @xmath311 & 0.14 & 1.0 & @xmath312 & 820 & @xmath313 & @xmath314 & @xmath315 + & @xmath316 & 0.14 & 1.0 & @xmath317 & & & & + hd4_0818_1037 & & & & & & & & + hd4_0298_0744 & & & & & & - & & +
c4 - 06 & & & & & @xmath104 & 3.2 & 24.8 & @xmath318 + hd4_1006_0680 & 300 & 0.22 & 1.7 & @xmath319 & 677 & 0.3 & 2.3 & @xmath320 + hd4_1486_0880 & & & & & & - & & + hd4_0367_0266 & & & & & & - & &
+ hd2_1928_1041 & & & & & 390 & 2.0 & 16.5 & @xmath321 + c2 - 05 & & & & & 390 & 2.0 & 16.5 & @xmath321 + hd2_1739_1258 & & & & & & & & + hd2_1398_1164 & @xmath322 & 0.26 & 2.2 & @xmath323 & & & & + c3 - 02 & & & & & & & &
+ hd2_0698_1297 & @xmath324 & 0.45 & 3.3 & @xmath325 & 4707 & 2.7 & 19.6 & @xmath326 + hd2_0705_1366 & @xmath324 & 0.14 & 1.0 & @xmath323 & 4707 & 2.7 & 19.8 & @xmath326 + hd2_0529_1567 & @xmath327 & 0.12 & 1.0 & @xmath328 & 1394 & 6.3 & 50.7 & @xmath329 + hd2_0624_1688 & @xmath330 & 0.31 & 2.5 & @xmath331 & 1394 & 6.3 & 50.1 & @xmath329 + hd2_0725_1818 & & & & & @xmath104 & 1.3 & 10.6 & @xmath332 + hd2_0743_1844 & & & & & @xmath104 & 1.3 & 10.6 & @xmath332 + hd3_1633_1909 & 330 & 0.64 & 4.5 & @xmath333 & & & & + hd3_1824_1945 & & & & & & & &
+ lcccl hd4_0259_1947 & 748 & 0.01 & -1.000 & 0.00 + hd4_1076_1847 & 1078 & 0.02 & 0.882 & 2.20e9 + c4 - 09 & 1114 & 0.57 & 3.216 & 7.84e8 + & 1115 & 0.13 & 3.216 & 3.56e9 + hd4_0818_1037 & 989 & 0.04 & 1.444 & 6.68e9 + hd4_0298_0744 & 782 & 0.54 & 1.761 & 4.21e9 + c4 - 06 & 1357 & 0.06 & 2.803 & 1.1e10 + & 1358 & 1.28 & 2.803 & 4.60e9 + hd4_1006_0680 & 1063 & 0.07 & 2.969 & 4.16e8 + hd4_1486_0880 & 1282 & 0.02 & -1.000 & 0.00 + hd4_0367_0266 & 813 & 0.04 & 2.931 & 7.34e9 + & 814 & 1.06 & 2.931 & 1.8e10 + hd2_1928_1041 & 109 & 0.23 & 2.009 & 2.81e9 + c2 - 05 & 110 & 0.27 & 2.005 & 9.70e9 + hd2_1739_1258 & 162 & 0.68 & 1.980 & 1.37e9 + & 163 & 0.41 & 1.980 & 2.3e10 + hd2_1398_1164 & 274 & 0.25 & 2.237 &
4.75e9 + c3 - 02 & 1179 & 0.23 & 2.218 & 1.25e9 + hd2_0698_1297 & 553 & 0.35 & 3.430 & 2.09e9 + hd2_0705_1366 & 516 & 0.13 & 3.368 & 1.77e9 + hd2_0529_1567 & 616 & 0.08 & 2.161 & 2.50e8 + hd2_0624_1688 & 561 & 0.10 & 2.419 & 8.61e8 + hd2_0725_1818 & 503 & 0.16 & 2.233 & 1.47e9 + hd2_0743_1844 & 503 & 1.30 & 2.233 & 1.47e9 + hd3_1633_1909 & 1499 & 0.48 & 2.175 & 4.88e8 + & 1506 & 0.99 & 4.050 & 1.36e9 + hd3_1824_1945 & 1513 & 0.38 & 2.050 & 4.15e9 + images ( left ) and closeups of extended kinematic features in 2d spectra ( right ) of the 11 lbgs in seven systems showing such features .
when a pair of targets is covered by a single slit , both targets are labeled in the image .
targets are shown in order of increasing ra , as in table [ tab : targs ] .
each image is 10 on a side and is oriented to show the slit vertical for comparison with the 2d spectrum .
the slit width is 11 @xmath334 , where @xmath74 is the relative position angle of the slitlet with respect to the slit mask .
each spectrum covers the same 10vertically as the image , and roughly 250 horizontally , the exact value depending on the projected pixel scale and therefore , again , the position angle of the particular slitlet used relative to its parent slit mask . the beginning and ending
wavelengths ( , lower left and right of spectrum ) of each section of spectrum and the position angle of the slitlet ( @xmath129 east of north , upper left of spectrum ) are labeled .
the features include spatially resolved or separated pairs of absorption lines of al ii , si ii , and o i ; and spatially or spectrally resolved or separated pairs of emission lines of ly@xmath23 .
the emission lines are usually the only relevant kinematic feature in their parent spectra ; the absorption lines however are chosen as representative from among several . ] | we present deep two - dimensional spectra of 22 candidate and confirmed lyman break galaxies ( lbgs ) at redshifts @xmath0 in the hubble deep field ( hdf ) obtained at the keck ii telescope .
the targets were preferentially selected with spatial extent and/or multiple knot morphologies , and we used slitmasks and individual slits tilted to optimize measurement of any spatially resolved kinematics . our sample is more than one magnitude fainter and is at higher redshift than the kinematic lbg targets previously studied by others .
the median target magnitude was @xmath1 , and total exposure times ranged from 10 to 50 ks .
we measure redshifts , some new , ranging from @xmath2 to @xmath3 , including two interlopers at @xmath4 , and resulting in a sample of 14 lbgs with a median redshift @xmath5 .
the morphologies and kinematics of the close pairs and multiple knot sources in our sample are generally inconsistent with galaxy formation scenarios postulating that lbgs occur only at the bottom of the potential wells of massive host halos ; rather , they support `` collisional starburst '' models with significant major merger rates and a broad halo occupation distribution . for 13 lbgs with possible kinematic signatures
, we estimate a simple dynamical mass , subject to numerous caveats and uncertainties , of the galaxies and/or their host dark matter halos .
dynamical mass estimates of individual galaxies range from @xmath6 to @xmath7 and mass estimates of halos , based on close lbg pairs , range from @xmath8 to @xmath9 with a median value @xmath10 .
comparison with a recent numerical galaxy formation model implies that indeed the pairwise velocities might not reflect true dynamical masses .
we compare our dynamical mass estimates directly to stellar masses estimated for the same galaxies from seds , and find no evidence for a strong correlation .
the diversity of morphologies and dynamics implies that lbgs represent a broad range of galaxy or proto - galaxy types in a variety of evolutionary or merger stages rather than a uniform class with a narrow range of mass . |
an _ rna sequence _ is a string composed of four types of nucleotides , namely @xmath0 , and @xmath1 .
given an rna sequence , the goal of the _ rna folding _
problem is to find a maximum cardinality set of crossing - free pairs of nucleotides , where all the pairs are either @xmath2 or @xmath3 .
the problem is central in bioinformatics and has found applications in many areas of molecular biology . for a more comprehensive exposition of the topic ,
the reader is referred to e.g. @xcite .
it is well - known that the problem can be solved in cubic time using a simple dynamic programming method @xcite . due to the importance of rna folding in practice ,
there has been a long line of research on improving the cubic time algorithm ( see e.g. @xcite ) .
currently the best upper bound is @xmath4 @xcite , and this can be obtained via four - russian method or fast min - plus multiplication ( based on ideas from valiant s cfg parser @xcite ) . whether the rna folding problem can be solved in @xmath6 time for some @xmath8 is still a major open problem .
other than attempting to improve the upper bound , we should also approach the problem in the opposite direction , i.e. showing a lower bound or arguing why the problem is hard .
a popular way to show hardness of a problem is to demonstrate a lower bound conditioned on some widely accepted hypothesis . [
strongly exponential time hypothesis ( seth ) ] [ c-1 ] there exists no @xmath9 such that @xmath7-sat with @xmath10 variables can be solved in time @xmath11 for all @xmath12 .
[ c-2 ] there exists no @xmath13 such that @xmath7-clique on graphs with @xmath10 nodes can be solved in time @xmath14 for all @xmath12 , where @xmath15 is the matrix multiplication exponent .
assuming that seth ( conjecture [ c-1 ] ) holds , the following bounds are unattainable for any @xmath8 : * an @xmath16 algorithm for @xmath7-dominating set problem @xcite , * an @xmath17 algorithm for dynamic time warping , longest common subsequence , and edit distance @xcite , * an @xmath18 algorithm for ( @xmath19)-approximating the diameter of a graph with @xmath20 edges @xcite . as remarked in @xcite , it is easy to reduce the longest common subsequence problem on binary strings to the rna folding problem as following : given two binary strings @xmath21 , we let @xmath22 be the string such that @xmath23 = a$ ] if @xmath24 = 0 $ ] , @xmath23 = c$ ] if @xmath24 = 1 $ ] , and we let @xmath25 be the string such that @xmath26 = u$ ] if @xmath27 = 0 $ ] , @xmath26 = g$ ] if @xmath27 = 1 $ ]
. then we have a 1 - 1 correspondence between rna foldings of @xmath28 ( i.e. concatenation of @xmath29 and the reversal of @xmath30 ) and common subsequences of @xmath31 and @xmath32 .
it has been shown in @xcite that there is no @xmath17 algorithm for longest common subsequence problem on binary strings conditioned on seth , and we immediately get the same conditional lower bound for rna folding from the simple reduction ! very recently , based on a conjectured hardness of @xmath7-clique problem ( conjecture [ c-2 ] ) , a higher conditional lower bound was proved for a generalized version of the rna folding problem ( which coincides with the rna folding problem when the alphabet size is 4 ) @xcite : [ @xcite ] [ thm-1 ] if the generalized rna folding problem on sequences of length @xmath10 with alphabet size 36 can be solved in @xmath33 time , then @xmath34-clique on graphs with @xmath35 can be solved in @xmath36 time . therefore , a @xmath37 time algorithm for the generalized rna folding with alphabet size at least 36 will disprove conjecture [ c-2 ] , yielding a breakthrough to the parameterized complexity of clique problem . however , the above theorem is irrelevant to the rna folding problem in real life ( which has alphabet size 4 ) .
it is unknown whether the generalized rna folding for alphabet size @xmath38 admits a faster algorithm than the case for alphabet size @xmath39 .
in fact , there are examples of string algorithms whose running time scales with alphabet size ( e.g. string matching with mismatched @xcite and jumbled indexing @xcite ) .
we also note that when the alphabet size is 2 , the generalized rna folding can be trivially solved in linear time . in this paper
, we improve upon theorem [ thm-1 ] by showing the same conditional lower bound for the rna folding problem : [ thm-2 ] if the rna folding problem on sequences in @xmath40 can be solved in @xmath33 time , then @xmath34-clique on graphs with @xmath35 can be solved in @xmath41 time .
note that we also get an @xmath42 factor improvement inside @xmath43 , though it does not affect the conditional lower bound .
the current state - of - art algorithm for @xmath44clique , which takes @xmath45 time , requires the use of fast matrix multiplication @xcite which does not perform very efficiently in practice . for combinatorial ,
non - algebraic algorithm for @xmath44clique , the current best one runs in @xmath46 time @xcite , which is only slightly better than the trivial approach . as a result , by theorem [ thm-2 ] , even a @xmath47 time combinatorial algorithm for rna folding will lead to an improvement for combinatorial algorithms for @xmath44clique ! in the proof of theorem [ thm-1 ] in @xcite , given a graph @xmath48 , a sequence of length @xmath49 is constructed in such a way that we can decide whether @xmath50 has a @xmath34-clique according to the number of pairs in an optimal generalized rna folding of @xmath51 .
such a construction requires many different types of letters in order to build various `` walls '' which prevent undesired pairings between different parts of the sequence .
hence extending their approach to handle the case where the alphabet size is 4 may not be easy without aid from other techniques and ideas .
* overview of our approach .
* at a high level , our reduction ( from @xmath34-clique problem to rna folding problem ) follows the approach in @xcite : we enumerate all @xmath7-cliques , and each of them is encoded as some gadgets .
all the gadgets are then put together to form an rna sequence .
the goal is to ensure that an optimal rna folding corresponds to choosing three @xmath7-cliques that form a @xmath34-clique , given that the underlying graph admits a @xmath34-clique . to achieve this goal without using extra types of letters that force the gadgets to match in a desired manner
, we construct the gadgets via a key lemma in @xcite , whose original purpose is to prove that longest common subsequence and other edit distance problems are seth - hard even on binary strings .
we will treat it as a black box and apply it multiple times during the construction .
this powerful tool will allow us to test whether two @xmath7-cliques form a @xmath52-clique by the longest common subsequence between the two strings representing the two @xmath7-cliques . in the final rna sequence ,
all clique gadgets are well - separated by some carefully designed sequences whose purpose is to `` trap '' all the clique gadgets except three of them . since we know that these three clique gadgets are guaranteed to match well if the graph has a @xmath34-clique , we can infer whether the graph has a @xmath34-clique from the optimal rna folding of the rna sequence .
* dyck edit distance .
* one other way to formulate the rna folding problem is as follows : deleting the minimum number of letters in a given string to transform the string into a string in the language defined by the grammar @xmath53 ( empty string ) .
dyck edit distance problem _ @xcite , which asks for the minimum number of edits to transform a given string to a well - balanced parentheses of @xmath54 different types , has a similar formulation . due to the similarity
, the same conditional lower bound as theorem [ thm-1 ] was also shown for the dyck edit distance problem ( with alphabet size @xmath55 ) in @xcite . in this paper , we improve and simplify their result by demonstrating a simple reduction from rna folding to dyck edit distance problem : [ thm-3 ] if dyck edit distance problem on sequences of length @xmath10 with alphabet size 10 can be solved in @xmath33 time , then the rna folding problem on sequences in @xmath40 can be solved in @xmath56 time .
combining theorem [ thm-2 ] , [ thm-3 ] , we get the following corollary : [ cor-1 ] if the dyck edit distance problem on sequences of length @xmath10 with alphabet size 10 can be solved in @xmath33 time , then @xmath34-clique on graphs with @xmath35 can be solved in @xmath41 time .
given a set of letters @xmath57 , the set @xmath58 is defined as @xmath59 .
we require that @xmath60 , and @xmath61 .
therefore , we have @xmath62 and @xmath63 . for any @xmath64
, we write @xmath65 to denote @xmath66 ( the letter @xmath67 stands for the prime symbol ) .
we denote the reversal of the sequence @xmath31 as @xmath68 .
the concatenation of two sequences @xmath21 is denoted as @xmath69 ( or simply @xmath70 ) .
we write _ substring _ to denote a contiguous subsequence .
two pairs of indices @xmath71 , @xmath72 , with @xmath73 and @xmath74 , form a _ crossing pair _ iff @xmath75 * generalized rna folding . * given @xmath76 , the goal of the generalized rna folding problem is to find a maximum cardinality set @xmath77 among all sets meeting the following conditions : * @xmath78 does not contain any crossing pair . * for any @xmath79 , either ( i ) @xmath80 \in \sigma$ ] and @xmath81 = s[i]'$ ] or ( ii ) @xmath81 \in \sigma$ ] and @xmath80 = s[j]'$ ] is true .
we write @xmath82 .
any set meeting the above conditions is called an _ rna folding _ of @xmath51 .
if its cardinality equals @xmath83 , then it is said to be _
optimal_. in the paper we will only focus on the generalized rna folding problem with four types of letters , i.e. @xmath84 , which coincides with the rna folding problem for alphabet @xmath85 . with a slight abuse of notation ,
sometimes we will write @xmath86 , s[j])$ ] to denote a pair @xmath79 .
the notation @xmath87 is used to indicate an unordered pair . * longest common subsequence ( lcs ) . *
given @xmath88 and @xmath89 , we define @xmath90 , where @xmath91 the length of the longest common subsequence of @xmath31 and @xmath32 .
it is easy to observe that @xmath92 equals the length of lcs @xmath93 . in this sense
, we can conceive of an lcs problem as an rna folding problem with some structural constraint on the sequence . in @xcite , a conditional lower bound for the lcs problem with @xmath94 based on seth
was presented .
a key technique in their approach is a function that transforms an instance of an alignment problem between two sets of sequences to an instance of the lcs problem .
* alignments of two sets of sequences .
* let @xmath95 and @xmath96 be two linearly ordered sets of sequences of alphabet @xmath57 .
we assume that @xmath97 .
an _ alignment _ is a set @xmath98 , @xmath99 with @xmath100 and @xmath101 .
an alignment @xmath78 is called _ structural _ iff @xmath102 and @xmath103 .
that is , all sequences in @xmath104 are matched , and the matched positions in @xmath105 are contiguous .
the set of all alignments is denoted as @xmath106 , and the set of all structural alignments is denoted as @xmath107 .
the _ cost _ of an alignment @xmath78 ( with respect to @xmath105 and @xmath104 ) is defined as : @xmath108 that is , unaligned parts of @xmath104 are penalized by @xmath109 .
given a sequence @xmath31 , the _ type _ of @xmath31 is defined as @xmath110)$ ] , where each letter is assumed to be a number . note that when the alphabet is simply @xmath111 , @xmath112 $ ] is simply the number of occurrences of @xmath113 in @xmath31 .
the following key lemma was proved in @xcite ( lemma 4.3 of @xcite ) : [ @xcite ] [ lem-1 ] let @xmath95 and @xmath96 be two linearly ordered sets of binary strings such that @xmath97 , all @xmath114 are of type @xmath115 , and all @xmath116 are of type @xmath117 .
there are two binary strings @xmath118 , @xmath119 and an integer @xmath120 meeting the following requirements : * @xmath121 . * the types of @xmath122 and the integer @xmath120 only depend on @xmath123 .
* @xmath122 , and @xmath120 can be calculated in time @xmath124 ( hence @xmath125 and @xmath126 are both @xmath127 ) .
note that the term @xmath128 comes from the word gadget .
intuitively , computing an optimal alignment ( or an optimal structural alignment ) of two sets of sequences is at least as hard as computing a longest common subsequence .
the above lemma gives a reduction from the computation of a number @xmath54 with @xmath129 ( which can be regarded as an approximation of optimal alignments ) to a single lcs instance .
we will use the above lemma as a black box to devise two encodings , the clique node gadget @xmath130 and the clique list gadget @xmath131 , for a @xmath7-clique @xmath132 in a graph in such a way that we can decide whether two @xmath7-cliques @xmath133 form a @xmath52-clique according the value of @xmath134 .
when invoking the lemma , @xmath105 , @xmath104 are designed in such a way that we can test whether a condition is met ( e.g. whether two given @xmath7-cliques form a @xmath52-clique ) by the value of @xmath135 . we will show that @xmath136 for the case we are interested in .
therefore , we can infer whether the condition we are interested in is met from the value of @xmath137 .
the goal of this section is to prove theorem [ thm-2 ] .
let @xmath138 be a graph , and let @xmath139 .
we write @xmath140 to denote the set of @xmath7-cliques in @xmath50 .
we fix @xmath141 . as in @xcite
, we will construct a sequence @xmath142 such that we can decide whether @xmath50 has a @xmath34-clique according to the value of @xmath143 . as our framework of
the construction of @xmath144 is similar to the one in @xcite , we will give the building blocks ( for constructing @xmath144 ) the same names as their analogues in @xcite , despite that they may have different lower - level implementations .
the high - level plan is described as following : in section [ ss-1 ] we describe two encodings @xmath145 for a @xmath7-clique @xmath132 based on the black box described in lemma [ lem-1 ] . in section [ ss-2 ] , adapting the encodings shown in the previous subsection as the building blocks , we present the definition of the binary sequence @xmath144
. we will give a lower bound on @xmath143 by demonstrating an rna folding of @xmath144 , and the bound will depend on whether @xmath50 has a @xmath34-clique .
the goal of the next two subsections is to show that the bound given in section [ ss-2 ] is actually the exact value of @xmath143 . in section [ ss-3 ] , we show that there exists an optimal rna folding of @xmath144 meeting several constraints
. these constraints will simplify the calculation of @xmath143 , and we will work out the exact calculation in section [ ss-4 ] .
we associate each vertex @xmath146 a distinct integer in @xmath147 .
let @xmath148 be the binary encoding of such integer with @xmath149 .
we define @xmath150 to be the binary string resulted by replacing each 0 in @xmath148 with 01 and replacing each 1 in @xmath148 with 10 .
it is clear that for each @xmath146 , @xmath150 is of type @xmath151 , and @xmath152 iff @xmath153 . in this subsection
we present two encodings @xmath145 for a @xmath7-clique @xmath132 such that we can infer whether two @xmath7-cliques @xmath133 form a @xmath52-clique from the value of @xmath134 . for each @xmath146 , the _ list gadget _ @xmath154 and the _ node gadget _
@xmath155 are defined as following : * @xmath156 , where @xmath157 @xmath158 , and the number of occurrences of @xmath159 is @xmath160 . *
@xmath161 .
[ lem-2 ] there is a constant @xmath162 , depending only on @xmath10 , such that for any @xmath163 , we have @xmath164 iff @xmath165 . we let @xmath166 .
let @xmath167 , where the number of occurrences of @xmath159 is @xmath168 , and let @xmath169 .
in view of lemma [ lem-1 ] , we have @xmath170 , @xmath171 , for some @xmath120 whose value depends on @xmath172 , and @xmath173 .
as these parameters depend solely on @xmath10 , the number @xmath120 only depends on @xmath10 .
since @xmath174 , any non - empty alignment between @xmath105 and @xmath104 is structural .
this implies that @xmath175 , @xmath176 .
when @xmath164 , since @xmath177 is contained in @xmath105 , clearly @xmath178 .
when @xmath179 , @xmath177 does not appear in @xmath105 , so @xmath180 .
note that @xmath181 , for any @xmath146 . as a result , @xmath164 iff @xmath175 , @xmath182 . hence setting @xmath183 suffices .
we let @xmath184 be the type of the list gadgets , and we let @xmath185 be the type of the node gadgets . for each @xmath7-clique @xmath186 , we define the _ clique node gadget _ @xmath187 and the _ clique list gadget _ @xmath188
as following : * @xmath189 , where the number of occurrences of each @xmath190 is @xmath7 . *
@xmath191 , @xmath192 , + @xmath193 , @xmath194 , where the number of occurrences of each @xmath195 is @xmath7 .
we are ready to prove the main lemma in the subsection : [ lem-3 ] there is a constant @xmath196 , depending only on @xmath197 , such that for any @xmath198 , @xmath199 is a @xmath52-clique iff @xmath200 .
let @xmath201 , and let @xmath202 .
let @xmath203 , where each @xmath190 appears @xmath7 times , and let @xmath204 , @xmath205 , @xmath206 , @xmath207 , @xmath206 , @xmath208 , @xmath205 , @xmath206 , @xmath209 , where each @xmath210 appears @xmath7 times .
in view of lemma [ lem-2 ] , we have @xmath211 , so we can lower bound @xmath212 by @xmath213 . if @xmath214 , any alignment has cost @xmath213 . when @xmath215 , it is easy to observe that in order to achieve @xmath216 , all sequences in @xmath104 must be aligned ( as the cost for any unaligned sequence in @xmath104 is now @xmath217 )
. therefore , any alignment @xmath78 with @xmath216 must be @xmath218 with @xmath219 , for all @xmath220 . in view of the above , @xmath221 iff @xmath219 for all @xmath220 . since @xmath218 is structural , @xmath221 iff @xmath222
. therefore , in view of lemma [ lem-1 ] , there exists a constant @xmath120 such that : * if @xmath221 , then @xmath223 , @xmath224 . *
if @xmath225 , then @xmath223 , @xmath226 .
moreover , the value of @xmath120 depends only on @xmath172 , @xmath227 .
as these parameters depend solely on @xmath197 , the number @xmath120 only depends on @xmath197 .
when @xmath199 is a @xmath52-clique , all vertices in @xmath228 are adjacent to all vertices in @xmath229 . in view of lemma [ lem-2 ] , @xmath230 .
hence @xmath221 , implying that @xmath223 , @xmath224 .
when @xmath199 is not a @xmath52-clique , there exist @xmath231 such that @xmath232 . according to our definition of @xmath105 and @xmath104
, we have @xmath233 , @xmath234 , and hence @xmath235 .
this implies that @xmath225 , which leads to @xmath223 , @xmath226 . as a result , @xmath199 is a @xmath52-clique iff @xmath236 , @xmath237 @xmath238
. setting @xmath239 suffices .
the following lemma is a simple consequence of lemma [ lem-1 ] : [ lem - length ] there exist four integers @xmath240 , @xmath241 , @xmath242 , @xmath243 , such that for any @xmath244 , * @xmath245 the number of occurrences of @xmath246 in @xmath187 , @xmath247 . * @xmath248 the number of occurrences of @xmath246 in @xmath188 , @xmath247 . as a consequence of lemma [ lem-1 ] ,
all @xmath187 have the same type , and all @xmath188 have the same type .
therefore , the existence of these four integers is guaranteed . in view of lemma [ lem-1 ] , for all @xmath146 ,
both @xmath154 and @xmath155 have length at most @xmath249 . applying lemma [ lem-1 ] again
, the length of both @xmath187 and @xmath188 for all @xmath244 is @xmath250 . as a result ,
the four integers can be bounded by @xmath251 .
based on the parameters in lemma [ lem - length ] , we define @xmath252 ; for @xmath253 , we set @xmath254 ; and @xmath255 .
the rna sequence @xmath144 is then defined as following : @xmath256 0^{\ell_4 } \left[{0'}^{\ell_3 } \underset{{t \in \mathcal{c}_k}}{\bigcirc } \left ( \textrm{cg}_\beta(t ) { 0'}^{\ell_3 } \right)\right ] 0^{\ell_4 } \left[{0'}^{\ell_3 } \underset{{t \in \mathcal{c}_k}}{\bigcirc } \left ( \textrm{cg}_\gamma(t ) { 0'}^{\ell_3 } \right)\right],\ ] ] where @xmath257 for any @xmath244 , @xmath258 , the string @xmath259 is called a _ clique gadget_. note that @xmath260 , @xmath261 , and @xmath262 .
it is obvious that @xmath263 . before proceeding further ,
we explain some intuitions behind the definition of @xmath144 and give a simple lower bound on @xmath143 by constructing an rna folding as following : * the pairings between letters in some @xmath264 and some @xmath265 sometimes make a clique gadget unable to participate in the rna folding with other clique gadgets . in this sense
, a clique gadget is said to be `` blocked '' if the letters within the clique gadget only pair up with other letters within the same clique gadget or some @xmath266 in a @xmath265 .
+ let s try linking all the @xmath267 in all @xmath264 to some @xmath266 in some @xmath265 in such a way that all clique gadgets are blocked except @xmath268 , @xmath269 , and @xmath270 .
this gives us @xmath271 amount of pairs .
see fig .
[ fig-1 ] . * for a clique gadget that is `` blocked '' , our design of @xmath144 ensures that the number of pairs involving letters in the clique gadget ( in certain optimal rna foldings ) is irrelevant to its corresponding @xmath7-clique ( we will prove it later ) : * * for a blocked @xmath272 , since @xmath273 is significantly larger than @xmath274 , an optimal way to pair up the letters is to match as many @xmath275 as possible .
this gives us @xmath276 pairs .
* * for a blocked @xmath277 , since we do not have any 1 here , the best we can do is to match all @xmath267 to some @xmath265 .
this gives us @xmath278 pairs .
* * for a blocked @xmath279 , no matching can be made .
+ therefore , the total amount of pairs involving blocked clique gadgets is @xmath280 .
see fig .
[ fig-2 ] for an illustration . * for the three clique gadgets that are not blocked , we will later see that ( in certain optimal rna foldings ) @xmath281 correspond to a @xmath34-clique if the graph has one .
it is a simple exercise to construct an rna folding for @xmath282 that uses up all the @xmath283 and has cardinality @xmath284 .
recall that @xmath285 @xmath286 is the length of the lcs between @xmath287 and @xmath288 .
see fig .
[ fig-3 ] for an illustration . and @xmath265 . ,
scaledwidth=100.0% ] .
, scaledwidth=100.0% ] in light of the above discussion , we define : * @xmath289 , * @xmath290 . the next lemma , which gives a lower bound on @xmath143 , is then implied instantly by the above discussion .
[ lem-5 ] @xmath291 .
ultimately we will show that @xmath292 , and clearly this offers enough information for us to decide whether @xmath50 has a @xmath34-clique .
the following lemma calculates @xmath293 of some sequences , and they will be useful in the subsequent discussion .
[ lem-4 ] the following statements are true for any @xmath294 : 1 .
@xmath295 2 .
@xmath296 3 .
@xmath297 4 .
@xmath298 5 .
@xmath299 6 .
@xmath300 the value of @xmath293 for each of the six sequences are calculated as following : 1 .
linking as many @xmath113 to @xmath301 gets a matching of size @xmath302 . to see that it is optimal , it suffices to show that both @xmath303 and @xmath304 can not appear in an optimal rna folding : * if the rna folding contains @xmath304 , then none of @xmath301 can participate in the rna folding .
as the total number of @xmath267 is @xmath305 , the size of rna folding is at most @xmath306 . *
if the rna folding contains @xmath303 , then at most @xmath307 number of letters within the middle @xmath308 ( the one between @xmath309 and @xmath310 ) can participate in the rna folding .
it implies that the number of @xmath311 pairs in the rna folding is at most @xmath312 .
hence the size of the rna folding can be upper bounded by @xmath313 .
since there is no @xmath113 , the equation follows from the fact that there are @xmath314 occurrences of @xmath267 , all of which can be matched to some @xmath266 without crossing .
no matching can be made since there is no @xmath315 .
the value of @xmath293 can be upper bounded by the number of @xmath113 and @xmath267 .
this is @xmath316 .
the value of @xmath293 can be upper bounded by the number of @xmath301 and @xmath267 .
this is @xmath317 .
we define @xmath318 , which is the result of removing the clique node gadgets and the clique list gadgets in @xmath319 .
it is clear that @xmath320 @xmath321 , as the total length of the removed substrings can be upper bounded by @xmath322 .
therefore , it suffices to show that @xmath323 .
+ let @xmath78 be any rna folding of @xmath51 : * case : there are some @xmath324 where the @xmath267 comes from the first @xmath309 in @xmath51 .
clearly , the first substring @xmath325 can not participate in any pairing .
therefore , @xmath326 .
* case : there are some @xmath327 where the @xmath267 comes from the first @xmath309 in @xmath51 .
in this situation , at most half of the @xmath328 can participate in the rna folding , since only the first @xmath329 in @xmath51 is reachable from @xmath328 without crossing a pair @xmath303 . therefore , @xmath330 is at most the total number of @xmath267 and @xmath113 in @xmath51 minus @xmath273 , i.e. @xmath331 . *
case : the first @xmath309 in @xmath51 does not participate in the rna folding .
then , @xmath332 @xmath333 .
note that ( 1 ) , ( 2 ) , ( 3 ) in lemma [ lem-4 ] imply that the rna folding for blocked clique gadgets described in fig .
[ fig-2 ] is optimal , and the optimal number of pairings is irrelevant to the corresponding @xmath7-clique . in the previous subsection
, we describe an rna folding of @xmath144 containing @xmath334 pairs .
the two key properties of this rna folding are : _ property 1 . _
all @xmath267 in all @xmath264 are paired up with some @xmath266 in some @xmath335 .
_ property 2 .
_ all clique gadgets are `` blocked '' by the pairings between @xmath264 and @xmath335 , except the three clique gadgets : @xmath281 , for some @xmath336 .
the goal in this section is to show that there is an optimal rna folding having the above two properties , which facilitates the calculation of @xmath143 in the next subsection .
[ lem-6 ] for any rna folding @xmath78 of @xmath144 , if there exists a pair linking a @xmath267 in a specific @xmath264 ( denoted as @xmath337 ) to a @xmath266 in a specific @xmath335 ( denoted as @xmath338 ) , then there exists another rna folding @xmath339 with @xmath340 where all letters in @xmath337 are linked to some letters in @xmath338 .
it immediately follows from the fact that @xmath341 is greater than the total number of @xmath267 in @xmath144 .
it makes rematching all the letters in @xmath337 to some letters in @xmath338 possible .
lemma [ lem-7 ] ensures that there is an optimal rna folding having property 1 : [ lem-7 ] there is an optimal rna folding @xmath78 of @xmath144 having property 1 .
let s choose any rna folding @xmath78 of @xmath144 with @xmath342 .
in view of lemma [ lem-6 ] , we can assume that for each @xmath264 in @xmath144 , either all its letters are matched to some @xmath266 in the same @xmath335 or none of its letters is matched to any @xmath266 in any @xmath335 .
let @xmath343 denote the number of @xmath264 such that none of its letters are matched to any @xmath266 in any @xmath335 . for some @xmath244 , and for some @xmath258 ,
@xmath259 is said to be `` trapped '' in @xmath78 if all letters within @xmath259 are either unmatched , matched to letters within @xmath259 , or matched to letters in some @xmath335 .
we note that a sufficient condition for @xmath259 to be trapped is that the letters in its two neighboring @xmath264 are all matched to the same @xmath335 .
the cases that the condition is violated is enumerated as follows : 1 .
the two neighboring @xmath264 of @xmath259 are matched to different @xmath335 , and this occurs at most @xmath344 times ( i.e. at most two times per @xmath258 ) .
a neighboring @xmath264 of @xmath259 is not matched to any @xmath335 , and this occurs at most @xmath345 times . therefore , the number of clique gadgets that are not trapped in @xmath78 is at most @xmath346 . using this information
, we can derive an upper bound of @xmath330 : @xmath347 in view of the calculation in lemma [ lem-4 ] , @xmath330 is at most @xmath348 since @xmath349 , and since the length of a clique gadget @xmath350 , we have : @xmath351 therefore , @xmath352 if @xmath353 . hence we must have @xmath354 , i.e. all @xmath267 in all @xmath264 are paired up with some @xmath266 in some @xmath335 .
to proceed further , some terminologies are needed to formally define the property 2 : let @xmath78 be an rna folding of a sequence where @xmath355 are two substrings ( subsequences of consecutive elements ) .
we write @xmath356 iff * there exists @xmath357 with @xmath358 .
* @xmath355 are disjoint substrings .
for example , `` @xmath359 is blocked in @xmath78 '' is equivalent to `` there exist no @xmath360 , @xmath361 such that @xmath362 '' .
@xmath363 is defined as the set of rna foldings of @xmath144 such that @xmath364 iff * @xmath78 has property 1 , and * there exist @xmath365 such that @xmath366 , and for any @xmath367 , @xmath368 , @xmath369 implies that @xmath370 .
@xmath371 and @xmath372 are defined analogously .
@xmath373 is defined as the set of rna foldings of @xmath144 such that @xmath374 iff * @xmath78 has property 1 , and * there exist @xmath375 such that for any @xmath367 , @xmath368 , @xmath369 implies that @xmath376 . using the above notions ,
it is clear that @xmath374 iff @xmath78 has both property 1 and property 2 . in the remainder of the subsection
, we will prove that there exists an optimal rna folding of @xmath144 that belongs to @xmath373 . to ease the notation , for each @xmath258 , we call @xmath259 an `` @xmath377 clique gadget '' , for all @xmath244 ; we write `` @xmath378 and @xmath379 are _ linked _ ( in @xmath78 ) '' to denote @xmath380 .
[ lem-8 ] let @xmath78 be an optimal rna folding of @xmath144 having property 1 .
for any @xmath258 , there does not exist two @xmath377 clique gadgets @xmath381 such that @xmath380 .
there is a substring @xmath264 located between @xmath378 and @xmath379 .
existence of a pair in @xmath78 linking a letter in @xmath378 and a letter in @xmath379 makes it impossible for any letter in the @xmath264 be matched to any letter in any @xmath335 , a contradiction .
[ lem-9 ] let @xmath78 be an optimal rna folding of @xmath144 having property 1 .
for any @xmath382 , @xmath383 , there does not exist two distinct @xmath377 clique gadgets @xmath381 and two ( not necessarily distinct ) @xmath384 clique gadgets @xmath385 such that @xmath386 and @xmath387 . clearly there must be a substring @xmath264 located between @xmath378 and @xmath379 .
however , since @xmath386 and @xmath387 , letters in the substring @xmath264 can only be matched to letters in @xmath388 , letters between @xmath381 , and letters between @xmath385 .
this contradicts property 1 .
[ lem-10 ] let @xmath78 be an optimal rna folding of @xmath144 having property 1 .
for any @xmath258 , suppose that there exist two distinct @xmath377 clique gadgets @xmath381 such that @xmath386 and @xmath387 for some clique gadgets @xmath385 .
then there does not exist any other pairs of clique gadgets that are linked in @xmath78 .
let @xmath389 such that @xmath390 is a @xmath384 clique gadget , and @xmath391 is a @xmath343 clique gadget .
by lemma [ lem-8 ] and lemma [ lem-9 ] , @xmath392 must be distinct .
suppose that there exist two clique gadgets @xmath393 that are linked in @xmath78 such that @xmath394 .
we show that this leads to a contradiction .
first of all , none of @xmath393 can be an @xmath377 clique gadget .
suppose that @xmath395 is an @xmath377 clique gadget .
then by lemma [ lem-8 ] , @xmath396 is either a @xmath384 clique gadget or a @xmath343 clique gadget . in any case , lemma [ lem-9 ] is violated .
therefore , we can ( without loss of generality ) assume that @xmath395 is a @xmath384 clique gadget , and @xmath396 is a @xmath343 clique gadget .
since @xmath381 are distinct , there must be a substring @xmath264 located between @xmath378 and @xmath379 . since @xmath378 is linked to a @xmath384 gadget , and since @xmath379 is linked to a @xmath343 gadget , letters in the substring @xmath264 can only be paired up with letters in the substring @xmath335 bordering both @xmath397 and @xmath398 ( imaging @xmath144 as a circular string ) .
however , the existence of a pair linking a letter in @xmath395 ( an @xmath384 clique gadget ) and a letter in @xmath396 ( a @xmath343 clique gadget ) implies that such @xmath335 can not be reached from the @xmath264 without a crossing , a contradiction .
the following lemma directly follows from lemma [ lem-7 ] and lemma [ lem-10 ] .
[ lem-11 ] there exists an optimal rna folding of @xmath144 that belongs to @xmath399 . by lemma [ lem-7 ]
, we can restrict our consideration to optimal rna foldings having property 1 .
let @xmath78 be any such optimal rna folding : * case 1 : * for each @xmath258 , there is at most one @xmath377 clique gadget that is linked to other clique gadgets
. then @xmath374 .
* case 2 : * for some @xmath258 , there are two distinct @xmath377 clique gadgets that are linked to other clique gadgets . by lemma [ lem-10 ] , @xmath400 .
we are now in a position to prove the main lemma in the subsection : [ lem-12 ] there exists an optimal rna folding of @xmath144 that belongs to @xmath373 . in view of lemma [ lem-11 ]
, it suffices to show that for any @xmath401 , we have @xmath402 .
let @xmath403 and @xmath404 be the ones in the definition of @xmath405 .
we let @xmath406 .
each pair in @xmath78 falls into one of the following categories : * the ones linking a @xmath267 in some @xmath264 to a @xmath266 in some @xmath335 .
there are exactly @xmath271 number of such pairs .
* the ones involving a letter in some @xmath407 , where @xmath408 .
as any letter in such @xmath407 can only be matched to the letters within @xmath407 or @xmath335 .
the number of such pairs can be upper bounded by @xmath409 . *
the ones involving a letter in some @xmath407 , where @xmath410 .
the number of such pairs can be upper bounded by @xmath411 @xmath412 .
therefore , using lemma [ lem-4 ] , we can upper bound @xmath330 as following : * when @xmath413 , @xmath414 . * when @xmath415 , @xmath416 . * when @xmath417 , @xmath418 . by lemma [ lem-5 ] , we always have @xmath419 ( recall that @xmath420 ) . in this subsection , we will prove that @xmath292 and finish the proof of theorem [ thm-2 ] . in view of lemma [ lem-12 ] , when calculating @xmath143
, we can restrict our attention to rna foldings of @xmath144 in @xmath373 .
based on the structural property of rna foldings in @xmath373 , we first reduce the calculation of @xmath143 to the calculation of optimal rna foldings of much simpler sequences . [ lem-13 ] @xmath421 . in view of lemma [ lem-12 ] ,
there is an optimal rna folding of @xmath144 in @xmath373 . for any @xmath374 ,
let @xmath375 be the ones in the definition of @xmath373 .
then , each pair in @xmath78 falls into one of the following categories : * the ones linking a @xmath267 in some @xmath264 to a @xmath266 in some @xmath335 .
there are exactly @xmath271 number of such pairs . *
the ones involving a letter in some @xmath407 , where @xmath422 .
as any letter in such @xmath407 can only be matched to the letters within @xmath407 or @xmath335 .
the number of such pairs can be upper bounded by @xmath423 . *
the ones involving a letter in some @xmath407 , where @xmath424 .
the number of such pairs can be upper bounded by @xmath425 . in view of lemma [ lem-4 ] , @xmath426 .
hence we conclude the proof .
the following auxiliary lemma is useful in the later discussion .
[ lem-14 ] let @xmath427 , where @xmath338 is either @xmath428 or @xmath429 .
then @xmath430 .
it suffices to show that there exists an optimal rna folding of @xmath51 such that the @xmath113 and the @xmath301 in @xmath338 are matched .
we first choose any optimal rna folding @xmath78 of @xmath51 , and then we show that we can modify @xmath78 in such a way that the @xmath113 and the @xmath301 in @xmath338 are matched without changing the number of matched pairs . *
case : the @xmath113 and the @xmath301 in @xmath338 are already matched . we are done .
* case : exactly one of the @xmath113 and the @xmath301 in @xmath338 is matched .
we first unmatch it , and then we pair up the @xmath113 and the @xmath301 .
doing so does not change the number of matched pairs . *
case : both of the @xmath113 and the @xmath301 in @xmath338 are matched to some other letters .
let the @xmath113 be matched with @xmath377 , and let the @xmath301 be matched with @xmath384 .
removing these two pairs from @xmath78 and adding @xmath431 and @xmath432 to @xmath78 does not change the number of matched pairs . for any choices of three @xmath7-cliques
@xmath375 , we define : @xmath433 where @xmath434 @xmath435 is simply a cyclic shift of the concatenation of @xmath268 , @xmath269 , and @xmath270 after removing the sequences of @xmath113s and @xmath301s at the beginning and the end of these clique gadgets . the next lemma ( together with lemma [ lem-13 ] ) reduces the calculation of @xmath143 to the calculation of @xmath436 .
[ lem-15 ] @xmath437 .
first of all , we state a few easy observations that will be applied in the proof : * by simply matching only the letters in @xmath438 , and @xmath270 ( as described in fig .
[ fig-3 ] ) , we can infer that @xmath439 . *
the total number of @xmath267 and @xmath113 in @xmath440 is at most @xmath441 . *
the difference between the number of @xmath113 and @xmath301 in @xmath440 is at most @xmath322 we claim that in any optimal rna folding @xmath78 of @xmath440 , all letters within all @xmath335 are not matched : * claim : there is no @xmath267 within @xmath269 matched to any @xmath266 in the two @xmath335 preceding and after @xmath269 .
recall that @xmath442 .
if there is such a pair , then at least @xmath273 amount of @xmath301 can not participate in the rna folding .
therefore , @xmath443 . *
claim : there is no @xmath267 within @xmath269 matched to any @xmath266 in the @xmath335 in the beginning of the sequence .
suppose that there is such a pair .
then the @xmath444 amount of @xmath301 within @xmath445 in @xmath268 and within the first @xmath329 in @xmath269 can only be matched to letters in @xmath268 .
however , the amount of @xmath113 in @xmath268 is at most @xmath446 , so at least @xmath447 amount of @xmath301 are not matched .
therefore , @xmath448 . *
claim : there is no @xmath267 within @xmath268 matched to any @xmath266 in any @xmath335 .
suppose that there is such a pair .
we can show that at least @xmath273 amount of @xmath301 can not participate in the rna folding , so @xmath449 . * * case : a @xmath267 within @xmath268 is matched to a @xmath266 in the first @xmath335 .
then the @xmath450 in the beginning of @xmath268 can not participate in the rna folding . * * case : a @xmath267 within @xmath268 is matched to a @xmath266 in the second @xmath335 .
then letters in the two @xmath308 in @xmath268 can only be matched to letters within @xmath451 .
hence at least @xmath452 amount of @xmath113 are unmatched .
since the difference between the number of @xmath113 and @xmath301 in @xmath440 is at most @xmath322 , at least @xmath453 amount of @xmath301 can not participate in the rna folding .
* * case : a @xmath267 within @xmath268 is matched to a @xmath266 in the third @xmath335 .
then all @xmath301 within @xmath269 can only be matched to @xmath113 within @xmath268 .
it is obvious that the number of @xmath301 within @xmath269 is at least @xmath273 more than the number of @xmath113 within @xmath268 , so at least @xmath454 amount of @xmath301 can not participate in the rna folding .
therefore , @xmath455 for the third equality , we just move @xmath456 from the end of the sequence to the beginning .
the fourth equality follows by applying lemma [ lem-14 ] iteratively ( which removes @xmath457 , @xmath458 , and @xmath459 ) . by calculating the exact value of @xmath436 , together with several previous lemmas
, the next lemma shows that @xmath292 .
[ lem-16 ] @xmath292 . in view of lemma [ lem-5 ] , [ lem-13 ] , [ lem-15 ]
, it suffices to show that @xmath460 @xmath461 .
first of all , it is easy to observe that @xmath462 , so for any optimal rna folding @xmath78 ( of @xmath435 ) , we must have @xmath463 .
we claim that in any optimal rna folding @xmath78 of @xmath435 , the following two statements are true : * for each of the two @xmath308 , there is a @xmath113 that is matched to a @xmath301 in the @xmath445 . * for each of the @xmath464 , there is a pair linking a @xmath267 in its @xmath309 and a @xmath266 in its @xmath310 . for the first statement , suppose that one @xmath308 does not have any letter matched to a @xmath301 in the @xmath445 .
it is easy to observe that the number of @xmath301 in @xmath435 that does not belong to @xmath445 is at most @xmath322 .
therefore , @xmath330 is at most the total number of @xmath267 plus the total number of @xmath113 minus @xmath465 . by a simple calculation , @xmath466 .
therefore , we conclude the first statement . for the second statement , suppose that there is an @xmath467 that has no pair linking a @xmath267 in its @xmath309 and a @xmath266 in its @xmath310 . due to the first statement , any pairing involving @xmath309 and @xmath310 are confined to be within @xmath51 .
therefore , the number of pairs involving letters in @xmath51 is at most @xmath468 .
this is certainly not optimal , since simply matching all @xmath267 in @xmath309 to all @xmath266 in @xmath310 gives us @xmath469 amount of pairs .
therefore , we conclude the second statement
. we can infer from the above two statements that for each @xmath467 , letters within @xmath51 are only matched to letters within @xmath51 in any optimal rna folding of @xmath435 . as a result ,
@xmath470 we are ready to prove the main theorem of the paper : * remainder of theorem [ thm-2 ] . *
_ if the rna folding problem on sequences in @xmath40 can be solved in @xmath33 time , then @xmath34-clique on graphs with @xmath35 can be solved in @xmath41 time .
_ given a graph @xmath50 , we construct the string @xmath144 . according to lemma [ lem-1 ] , [ lem - length ] ,
the length of @xmath144 is @xmath471 , and @xmath144 can be constructed in time @xmath472 .
we let @xmath375 be chosen such that @xmath473 is minimized . by lemma [ lem-3 ] , there exists a number @xmath196 such that : * the number @xmath196 depends only on @xmath474 , and @xmath475 . * if @xmath476 , then each of @xmath477 , @xmath478 , @xmath479 is a @xmath52-clique , which in turn is equivalent to `` @xmath480 is a @xmath34-clique '' . * if @xmath481 , then the graph has no @xmath34-clique . according to lemma [ lem-16 ] , @xmath292 . by its definition ,
@xmath482 only depends on @xmath474 , and @xmath483 .
hence we are able to decide whether @xmath50 has a @xmath34-clique from the value of @xmath143 , which can be calculated in time @xmath484 .
note that @xmath7 is treated as a constant instead of an input parameter .
in this section , we shift our focus to the dyck edit distance problem
. we will present a simple reduction from rna folding problem ( with alphabet size 4 ) to dyck edit distance problem ( with alphabet size 10 ) .
this leads to a much simplified and improved proof for a conditional lower bound of dyck edit distance based on the conjectured hardness @xmath7-clique ( the previous proof presented in @xcite requires 48 symbols ) .
* dyck edit distance . * given @xmath76 , the goal of the dyck edit distance problem is to find a minimum number of edit operations ( insertion , deletion , and substitution ) that transform @xmath51 into a string in the dyck context free language . given @xmath57 and its corresponding @xmath58 ,
the dyck context free language is defined by the grammar with following production rules : @xmath485 , @xmath486 , and @xmath487 ( empty string ) .
an alternative definition of the dyck edit distance problem is described as follows : given a sequence @xmath76 , find a minimum cost set @xmath77 satisfying the following conditions : * @xmath488 has no crossing pair .
* @xmath489 contains only pairs of the form @xmath490 , @xmath491 ( i.e. for all @xmath492 , we have @xmath80=x$ ] , @xmath81=x'$ ] , for some @xmath491 ) .
@xmath489 corresponds to the set of matched pairs .
* @xmath493 does not contain any pair of the form @xmath494 , @xmath495 ( i.e. for all @xmath496 we have either @xmath80 \in \sigma$ ] or @xmath81 \in \sigma'$ ] ) .
@xmath493 corresponds to the set of pairs that can be fixed by one substitution operation per each pair .
* let @xmath497 be the set of letters in @xmath51 that do not belong to any pair in @xmath78 .
each letter in @xmath497 requires one deletion / insertion operation to fix .
the cost of @xmath78 is then defined as @xmath498 , and the dyck edit distance of the string @xmath51 is the cost of a minimum cost set meeting the above conditions .
dyck edit distance problem can be thought of as an asymmetric version of the rna folding problem ( in rna folding , we allowed the aligned pair to be either @xmath490 or @xmath499 , @xmath491 ) that also handles substitution ( in addition to deletion and insertion ) .
though these two problems look similar , they can behave quite differently .
for example , in section [ sec.intro ] we describe a simple reduction from lcs to rna folding ; since lcs is edit distance problem without substitution , one may hope that the same reduction reduces edit distance problem to dyck edit distance problem . however , this is not true due to the following counterexample : both the two strings @xmath500 and @xmath501 require at least 4 edit operations to transform into the string @xmath502 ; but the dyck edit distance of @xmath503 is 4 ( by deleting all @xmath504 ) , while the dyck edit distance of @xmath505 is 3 ( by deleting all @xmath506 and substituting the second @xmath246 with @xmath507 ) .
intuitively , the substitution operation makes dyck edit distance more complicated than rna folding .
indeed , the same conditional lower bound as theorem [ thm-1 ] for dyck edit distance problem shown in @xcite requires a bigger alphabet size ( 48 instead of 36 ) and a longer proof . in the next , we prove theorem [ thm-3 ] by demonstrating a simple reduction from rna folding problem to dyck edit distance problem with alphabet size 10 .
this improves upon the hardness result in @xcite , and justifies the intuition that dyck edit distance is a harder problem than rna folding . *
proof of theorem [ thm-3 ] . *
_ if dyck edit distance problem on sequences of length @xmath10 with alphabet size 10 can be solved in @xmath33 time , then the rna folding problem on sequences in @xmath40 can be solved in @xmath56 time .
_ for notational simplicity , we let the alphabet for the rna folding problem be @xmath508 ( instead of @xmath85 ) .
let @xmath51 be any string in @xmath509 .
we define the string @xmath510 as the result of applying the following operations on @xmath51 : * replace each letter @xmath266 with the sequence @xmath511 .
* replace each letter @xmath267 with the sequence @xmath512 .
* replace each letter @xmath113 with the sequence @xmath513 .
* replace each letter @xmath301 with the sequence @xmath514 .
it is clear that @xmath510 is a sequence of length at most @xmath515 on the alphabet @xmath516 , though the letter @xmath517 is not used .
we claim that the dyck edit distance of @xmath510 is @xmath518 .
first , we show that the dyck edit distance of @xmath510 is at most @xmath518 .
given an optimal rna folding of @xmath51 , we construct a crossing - free matching @xmath78 with cost @xmath518 as follows : for matched pairs in the rna folding of @xmath51 : * for each matched pair @xmath304 in the rna folding of @xmath51 , we add two pairs @xmath519 to @xmath489 , and add three pairs @xmath520 to @xmath493 in its corresponding pair of substrings @xmath521 in @xmath510 . * for each matched pair @xmath303 in the rna folding of @xmath51 , we add two pairs @xmath522 to @xmath489 , and add three pairs @xmath523 to @xmath493 in its corresponding pair of substrings @xmath524 in @xmath510 . * similarly , for each matched pair @xmath525 in the rna folding of @xmath51 , we can add two pairs to @xmath489 and three pairs to @xmath493 . for unmatched letters in @xmath51 : * for each unmatched letter @xmath266 in @xmath51 , we add three pairs @xmath526 to @xmath493 in its corresponding substring @xmath527 .
similarly , for each unmatched letter
@xmath113 , we can add three pairs to @xmath493 .
* for each unmatched letter @xmath267 in @xmath51 , we add two pairs @xmath528 to @xmath493 in its corresponding substring @xmath529 . similarly , for each unmatched letter @xmath301 , we can add two pairs to @xmath493 .
the set @xmath489 has size @xmath530 , the set @xmath493 has size @xmath531 , and @xmath497 is an empty set .
therefore , the cost of @xmath78 is @xmath532 .
second , we show that the dyck edit distance of @xmath510 is at least @xmath518 . given a crossing - free matching @xmath78 ( on the string @xmath510 ) of cost @xmath120 , we recover an rna folding of @xmath51
that has @xmath533 number of matched pairs .
we build a multi - graph @xmath48 such that @xmath534 is the set of all substrings @xmath535 that constitute @xmath510 , and the number of edges between two substrings in @xmath534 is the number of pairs in @xmath489 linking letters between these two substrings .
note that @xmath536 .
it is clear that @xmath537 ( since @xmath538 ) .
therefore , we are done if we can recover an rna folding of size @xmath539 , since @xmath540 .
we observe the following : * @xmath50 has degree at most 2 ( due to our definition of @xmath535 , at most two letters in such a substring can participate in pairings of the form @xmath490 , @xmath541 , without crossing ) .
* in the graph @xmath50 , any edge must either links an @xmath542 with an @xmath543 or links an @xmath337 with an @xmath544 ( due to our definition of @xmath535 , any pairing of the form @xmath490 , @xmath541 , must be made between @xmath545 or between @xmath546 ) .
* @xmath50 does not contain any cycle of odd length ( due to the above observation ) . in view of the above ( second ) observation , a ( graph - theoretic ) matching
@xmath547 of @xmath50 naturally corresponds to a ( size @xmath548 ) rna folding of @xmath51 : for each edge ( a pair of substrings in @xmath510 ) in @xmath549 , we add its corresponding pair of letters in @xmath51 to the rna folding .
since a maximum matching has size @xmath539 in a graph of maximum degree 2 without odd cycles , we conclude the proof .
we note that for the case substitution is not allowed , the letter @xmath550 in the above proof is not needed , and this lowers the required alphabet size to 8 .
the reason that the letter @xmath550 is essential for the above proof to work is explained as follows : suppose that @xmath550 is removed . for each
matched pair @xmath304 in the rna folding of @xmath51 , after adding two pairs @xmath519 to @xmath489 , the letter @xmath507 between two @xmath551 in @xmath552 can not participate in any matching anymore . hence some letters will be in @xmath497 according to our construction of the crossing - free matching @xmath78 .
this indicates that our construction may not be optimal .
indeed , for the string @xmath553 ( after removing @xmath550 ) , if we insist on matching the two pairs @xmath519 in @xmath554 , then the cost will be at least @xmath555 ( three substitutions and two deletions are needed ) .
however , there is a solution that uses only 4 substitutions : @xmath556 .
in this paper we present a hardness result of rna folding problem with alphabet size 4 and demonstrate a reduction from rna folding problem to dyck edit distance problem .
a few open problems still remain : * there are still a few cases where the state - of - art conditional lower bound requires a certain alphabet size to work ( e.g. theorem [ thm-3 ] , corollary [ cor-1 ] , and the hardness result for dynamic time warping in @xcite ) .
is it possible to improve them using our technique or other ideas ? * is it possible to reduce dyck edit distance problem to rna folding problem ? * besides the classic rna folding problem , several problems in bioinformatics admit similar formulation ( see e.g. @xcite )
. it would be interesting to see whether the technique presented in this paper ( and @xcite ) can be adapted to give meaningful lower bounds for other problems .
89 tatsuya akutsu .
approximation and exact algorithms for rna secondary structure prediction and recognition of stochastic context - free languages .
journal of combinatorial optimization , 3(2 ) : 321 - 336 , 1999 .
mika amit , rolf backofen , steffen heyne , gad m. landau , mathias mohl , christina schmiedl , sebastian will .
local exact pattern matching for non - fixed rna structures . in combinatorial pattern matching ( cpm ) , 306320 , springer , 2012 .
amir abboud , arturs backurs , and virginia vassilevska williams .
if the current clique algorithms are optimal , so is valiant s parser . in proceedings of ieee
55th annual symposium on foundations of computer science ( focs ) , 2015 .
amir abboud , arturs backurs , and virginia vassilevska williams .
quadratic - time hardness of lcs and other sequence similarity measures . in proceedings of ieee
55th annual symposium on foundations of computer science ( focs ) , 2015 .
amihood amir , timothy m. chan , moshe lewenstein , and noa lewenstein . on hardness of jumbled indexing . in proceedings of the 41st international colloquium automata , languages , and programming ( icalp ) , 114125 , 2014 .
arturs backurs and piotr indyk .
edit distance can not be computed in strongly subquadratic time ( unless seth is false ) . in proceedings of the 47th annual acm symposium on theory of computing ( stoc ) , 5158 , 2015 .
karl bringmann and marvin knnemann .
quadratic conditional lower bounds for string problems and dynamic time warping . in proceedings of ieee
56th annual symposium on foundations of computer science ( focs ) , 7997 , 2015 .
liam roditty and virginia vassilevska williams .
fast approximation algorithms for the diameter and radius of sparse graphs . in proceedings of the 45th annual acm symposium on theory of computing ( stoc ) , 515524 , 2013 .
barna saha .
language edit distance & maximum likelihood parsing of stochastic grammars : faster algorithms & connection to fundamental graph problems . in proceedings of the ieee 56th annual symposium on foundations of computer science ( focs ) , 118135 , 2015 . | an rna sequence is a string composed of four types of nucleotides , @xmath0 , and @xmath1 .
given an rna sequence , the goal of the rna folding problem is to find a maximum cardinality set of crossing - free pairs of the form @xmath2 or @xmath3 .
the problem is central in bioinformatics and has received much attention over the years .
however , the current best algorithm for the problem still takes @xmath4 time , which is only a slight improvement over the classic @xmath5 dynamic programming algorithm . whether the rna folding problem can be solved in @xmath6 time
remains an open problem .
recently , abboud , backurs , and williams ( focs15 ) made the first progress by showing a conditional lower bound for a generalized version of the rna folding problem based on a conjectured hardness of the @xmath7-clique problem .
a drawback of their work is that they require the rna sequence to have at least 36 types of letters , making their result biologically irrelevant . in this paper , we show that by constructing the gadgets using a lemma of bringmann and knnemann ( focs15 ) and surrounding them with some carefully designed sequences , the framework of abboud et al . can be improved upon to work for the case where the alphabet size is 4 , yielding a conditional lower bound for the rna folding problem .
we also investigate the dyck edit distance problem .
we demonstrate a reduction from rna folding problem to dyck edit distance problem of alphabet size 10 , establishing a connection between the two fundamental string problems .
this leads to a much simpler proof of the conditional lower bound for dyck edit distance problem given by abboud et al . and lowers the required alphabet size for the lower bound to work .
keywords : rna folding , dyck edit distance , longest common subsequence , conditional lower bound , clique |
halting an epidemic outbreak in its early stages requires a detailed understanding of the progression of the number of new infections ( incidence ) .
current mathematical models predict that the incidence grows exponentially during the initial phase of an epidemic outbreak @xcite . within this exponential growth scenario
infectious diseases are characterized by the average reproductive number , giving the number of secondary infections generated by a primary case , and the average generation time , giving the average time elapse between the infection of a primary case and its secondary cases @xcite . in turn ,
vaccination strategies are designed in order to modify the reproductive number and the generation time @xcite .
i have recently shown , however , that this picture dramatically changes when the graph underlying the spreading dynamics is characterized by a power law degree distribution @xcite , where the degree of a node is defined as the number of its connections .
the significant abundance of high degree nodes ( hubs ) carry as a consequence that most nodes are infected in a time scale of the order of the disease generation time .
furthermore , the initial incidence growth is no longer exponential but it follows a power law growth @xmath0 , where @xmath1 is the characteristic distance between nodes on the graph . yet
, these predictions are limited to uncorrelated graphs and the susceptible - infected ( si ) model . in this work
i extend the theory of age - dependent branching processes @xcite to consider the topological properties of real networks .
first , i generalize my previous study @xcite to include degree correlations .
this is a fundamental advance since real networks are characterized by degree correlations @xcite that may significantly affect the system s behavior @xcite .
second , i consider the susceptible - infected - removed ( sir ) model that provides a more realistic description of real epidemic outbreaks @xcite , allowing us to obtain conclusions about the impact of patient isolation and immunization strategies on the final outbreak size
. finally , i survey our current knowledge about different networks underlying the spreading of infectious diseases and computer malwares and discuss the impact of their topology on the spreading dynamics .
consider a population of @xmath2 susceptible agents ( humans , computers , etc ) and an infectious disease ( human disease , computer malware , etc ) spreading among them .
the potential disease transmission channels are represented by an undirected graph , where nodes represent susceptible agents and edges represent disease transmission channels .
for example , when analyzing the spreading of sexually transmitted diseases the relevant graph is the web of sexual contacts @xcite , where nodes represent sexually active individuals and edges represent sexual relationships .
the degree of a node is the number of edges connecting this node to other nodes ( neighbors ) in the graph .
given the finite size of the population there is a maximum degree @xmath3 , where @xmath3 is at most @xmath4 .
i denote by @xmath5 the probability distribution that a node has degree @xmath6 .
the results obtained in this work are valid for arbitrary degree distributions .
nevertheless , recent studies have shown that several real networks are characterized by the power law degree distribution @xmath7 with @xmath8 @xcite .
therefore , i focus the discussion on this particular case .
real networks are characterized by degree correlations between connected nodes as well .
networks representing technological and biological systems exhibit disassortative ( negative ) correlations with a tendency to have connections between nodes with dissimilar degrees @xcite .
in contrast , social networks are characterized by assortative ( positive ) degree correlations with a tendency to have connections among nodes with similar degrees @xcite . to characterize the degree correlations i consider the probability distribution @xmath9 that a neighbor of a node with degree @xmath6 has degree @xmath10 .
it is important to note that the probability distributions @xmath5 and @xmath9 are related to each other by the detailed balance condition @xcite @xmath11 although @xmath9 contains all the information necessary to characterize the degree correlations it is difficult to analyze . a more intuitive measure which often appears in the analytical calculations @xcite is the average neighbors excess degree @xcite @xmath12
the empirical data indicates that @xcite @xmath13 where @xmath14 is obtained from the detailed balance condition ( [ db ] ) , resulting in @xmath15 when the degree correlations are disassortative the nearest neighbors of a low / high degree node tend to have larger / smaller degree . in this case
@xmath16 decreases with increasing @xmath6 .
in contrast , when the degree correlations are assortative the nearest neighbors of a low / high degree node tend to have proportional degrees . in this case @xmath16 increases with increasing @xmath6 . therefore ,
disassortative and assortative correlations are characterized by @xmath17 and @xmath18 , respectively .
real networks also exhibit the small - world property @xcite , meaning that the average distance @xmath1 between two nodes in the graph is small or it grows at most as @xmath19 .
for instance , social experiments such as the kevin bacon and erds numbers @xcite or the milgram experiment @xcite reveals that social actors are separated by a small number of acquaintances . this property is enhanced on graphs with a power law degree distribution ( [ pkpl ] ) with @xmath20 @xcite . in this case
the average distance between two nodes grows as @xmath21 , receiving the name of ultra small - world @xcite .
given the graph underlying the spreading of an infectious disease , let us consider an epidemic outbreak starting from a single node ( the root , patient zero , or index case ) . in the worst case scenario the disease propagates to all the nodes that could be reached from the root using the graph connections .
thus , the outbreak is represented by a spanning or causal tree from the root to all reachable nodes .
the generation of a node on the tree corresponds with the topological or hopping distance from the root .
this picture motivates the introduction of the following branching process : annealed spanning tree ( ast ) with degree correlations consider a graph with degree probability distribution @xmath5 and average degree @xmath22 , neighbors degree distribution @xmath9 given a node with degree @xmath6 , detailed balance condition ( [ db ] ) , and average distance between nodes @xmath1 .
the annealed spanning tree ( ast ) associated with this graph is the branching process satisfying the following properties : 1 .
the process starts from a single node , the root , at generation @xmath23 .
the root generates @xmath6 sons with probability distribution @xmath5 .
2 . each son at generation @xmath24 generates @xmath25 other sons with probability distribution @xmath9 , given its parent node has degree @xmath6 .
3 . a son at generation @xmath26 does not generate new sons .
[ acausaltree ] the term annealed means that we are not analyzing the true ( quenched ) spanning tree on the graph but a branching process with similar statistical properties . from the mathematical point of view
the ast is a generalization of the galton - watson branching process @xcite to the case where ( i ) the reproductive number of a node depends on the reproductive number of its ancestor and ( ii ) the process is truncated at generation @xmath1 .
this mathematical construction has been previously introduced to analyze the percolation properties of graphs with degree correlations @xcite .
the sharp truncation of the branching process at generation @xmath1 is an approximation . in the original graph
there are some nodes beyond the average distance between nodes @xmath1 and their average degree decreases with increasing generation .
therefore , a more realistic description is obtained defining @xmath9 generation dependent @xcite and truncating the branching process when the number of generations equals the graph diameter . yet
, an analytical treatment of this more realistic model is either unfeasible or results into equations that most be solved numerically , questioning its advantage with respect to direct simulations on the original graph . to allow for an analytical understanding
i truncate the branching process at generation @xmath26 , where @xmath1 represents the average distance between nodes @xmath1 in the original graph .
furthermore , i assume that @xmath9 is the same for all generations @xmath27 . at this point
it is worth noticing that all results derived below are exact for the ast but an approximation for the original graph .
the ast describes the case where all neighbors of an infected node are infected and at the same time .
more generally a node infects a fraction of its neighbors and these infections take place at variable times .
the susceptible @xmath28 infected @xmath28 removed ( sir ) model is an appropriate framework to consider the timing of the infection events @xcite .
the time scales for the transitions susceptible @xmath28 infected and infected @xmath28 removed are characterized by the distribution function of infection and removal times @xmath29 and @xmath30 , respectively .
for example , @xmath29 is the probability that the infection time is less or equal than @xmath31 .
consider an infected node @xmath32 and a susceptible neighbor @xmath33 .
the probability @xmath34 that @xmath33 is infected by time @xmath35 given @xmath32 was infected at time zero is the combination of two factors .
first , the infection time should be smaller that @xmath35 and , second , the removal time of the ancestor @xmath32 should be larger than the infection time .
more precisely @xmath36\ . \label{bt}\ ] ] from @xmath34 i obtain the probability that @xmath33 gets infected no matter when @xmath37 and the distribution function of the generation times @xmath38 in the original kermack - mckendrick formulation of the sir model @xcite the disease spreads at a rate @xmath39 from infected to susceptible nodes and infected nodes are removed at rate @xmath40 . in this case
the infection and removal rates @xmath39 and @xmath40 are exponentially distributed , @xmath41 and @xmath42 , resulting in @xmath43 @xmath44 some of the results obtained in this work are valid for any generation time distribution .
we focus , however , on the sir model with constant rate of infection and removal ( [ rsir])-([gtauexp ] ) . at this point
we can extend the ast definition to account for the variable infection times : age - dependent ast with degree correlations the age - dependent ast is an ast where nodes can be in two states , susceptible or infected , and 1 .
an infected node ( primary case ) infects each of its neighbors ( secondary cases ) with probability @xmath45 .
the generation times , the times elapse from the infection of a primary case to the infection of a secondary case , are independent random variables with probability distribution @xmath46 .
[ adast ] the age - dependent ast is a generalization of the bellman - harris @xcite and crum - mode - jagers @xcite age - dependent branching processes .
the key new elements are the degree correlations and the truncation at a maximum generation , allowing us to consider the topological properties of real networks .
let @xmath47 be the average number of nodes that are infected between time @xmath35 and @xmath48 given that patient zero ( the root ) was infected at time zero .
this magnitude is known in the epidemiology literature as the incidence @xcite .
consider an age - dependent ast and a constant infection and removal rate ( [ rsir])-([gtauexp ] ) .
making use of the tree structure i obtain ( appendix [ iterapproach ] ) @xmath49\ , \label{ntexp}\ ] ] where @xmath50 is the average number of nodes @xmath51 at generation @xmath52 , satisfying the normalization condition @xmath53 the interpretation of ( [ ntexp ] ) is the following .
if we count the time in units of one then on average @xmath51 new nodes are found at each generation .
since the infection times are variable , however , nodes at the same generation may be infected at different times .
this contribution is taken into account by the factor between @xmath54 $ ] in ( [ ntexp ] ) , giving the probability density of the sum of @xmath52 generation times .
independent of the particular @xmath52 dependence of @xmath51 , from ( [ ntexp ] ) it follows that the incidence decays exponentially for long times with a decay time @xmath55 .
this result is a consequence of the population finite size , i.e. sooner or later most of the nodes are infected and the number of new infections decays .
in contrast , the factor remaining after excluding the exponential decay is an increasing function of time and it dominates the spreading dynamics at short and intermediate times .
i obtain the following result determining the speed of the initial growth : consider the normalized incidence @xmath56 if there is some integer @xmath57 and real numbers @xmath58 and @xmath59 such that for all @xmath60 @xmath61 when @xmath62 then @xmath63 \ , \label{rhot}\ ] ] where @xmath64 [ nooo ] the symbol @xmath65 indicates that ( [ rhot ] ) is valid asymptotically when @xmath66 and it represents correction terms of the order of @xmath67 .
the demonstration of this result is straightforward . from ( [ cond ] )
it follows that for all @xmath60 the average number of nodes at generation @xmath52 ( [ zd ] ) is of the order of @xmath68 .
therefore , in the limit @xmath69 the sums in ( [ ntexp ] ) and ( [ no ] ) are dominated by the @xmath26 term and corrections are given by the ratio between the @xmath70 and @xmath26 terms .
plane showing the regions where theorem [ nooo ] is valid ( shadowed region ) for the case of a power law degree distribution ( [ pkpl ] ) and degree correlations ( [ qk ] ) .
the text inserts indicate the exponent @xmath71 in ( [ cond]).,width=384 ] the initial dynamics is characterized by a power law growth with an exponent determined by the average distance @xmath1 .
the characteristic time @xmath72 marks the time scale when this polynomial growth starts to be manifested .
this time is particularly small for graphs with a large maximum degree and satisfying the small world property , i.e. @xmath1 is small .
for instance , let us consider a power law degree distribution ( [ pkpl ] ) with @xmath8 and degree correlations ( [ qk ] ) .
the values of @xmath73 and @xmath74 for which the condition ( [ cond ] ) is satisfied are given in fig .
[ examples ] , together with the exponent @xmath71 .
disassortative degree correlations ( @xmath17 ) may invalidate the condition ( [ cond ] ) , indicating that strong disassortative correlations may lead to deviations from the theorem [ nooo ] prediction .
this observation is in agreement with a previous study focusing on the epidemic threshold @xcite .
in contrast , for assortative degree correlations ( @xmath18 ) the condition ( [ cond ] ) is satisfied for all @xmath8 . in other words ,
assortative correlations enhance the degree fluctuations , extending the validity of theorem [ nooo ] to the @xmath75 region .
focusing on the final size of the outbreak i obtain the following corollary : consider the average total number of infected nodes @xmath76 if the conditions of theorem [ nooo ] are satisfied then @xmath77\ . \label{ni}\ ] ] [ coro ] from this corollary it follows that increasing the rate of node removal , because of patient isolation or immunization , we just obtain a gradual decrease on the final outbreak size .
this implies that the concept of epidemic threshold loses sense since the outbreak size remains proportional to the population size for all removal rates .
this conclusion is in agreement with previous studies for the case ( @xmath78 ) @xcite and ( @xmath79 ) @xcite .
the above corollary extend these studies to the region @xmath75 , demonstrating that when @xmath18 there is not an epidemic thresholds for any value of @xmath73 .
theorem ( [ nooo ] ) proposes a new law of spreading dynamics characterized by an initial power law growth .
in essence the power low growth is a consequence of the small - world property and the divergence of the average reproductive number .
its origin is better understood analyzing the contribution of nodes at a distance @xmath52 from the root .
the distribution of infection times of nodes at generation @xmath52 is given by the distribution of the sum of @xmath52 generation times .
for the case of a constant infection rate this distribution is a gamma distribution , which is characterized by an initial power law with exponent @xmath80 .
this is the standard result for stochastic processes defined by a sequence of @xmath52 steps happening at a constant rate .
the total incidence is then obtained superimposing the contribution of each generation @xmath52 , weighted by the average number of nodes at that generation .
since most nodes are found at generation @xmath26 then the contribution from that generation dominates the incidence progression , resulting in a power law growth with exponent @xmath81 .
the small - world properties simply implies that @xmath1 is small and the resulting power law growth can be distinguished from an exponential growth .
the validity of this regime is restricted to time scales that are large enough such that an appreciable number of nodes at generation @xmath52 are infected , and it is followed by an exponential decay after most nodes at that generation are infected . to understand the relevance of this spreading law for real epidemic outbreaks , in the following
i analyze the validity of the conditions of theorem [ nooo ] for real networks underlying the spreading of human infectious diseases and computer malwares .
_ sexually transmitted diseases : _ sexual contacts are a dominant transmission mechanism of the hiv virus causing aids .
there are several reports indicating that the web of sexual contacts is characterized by a power degree distribution .
the current debate is if the exponent @xmath73 is smaller or larger than three @xcite . in either case
, it is known that social networks are characterized by assortative degree correlations @xcite , which extends the validity of theorem [ nooo ] to @xmath75 ( see fig .
[ examples ] ) .
there is also empirical evidence indicating that the number of aids infections grows as a power law in time for several populations @xcite .
when this empirical evidence is put together with that for sexual networks we obtain a strong indication that theorem [ nooo ] provides the right explanation for the observed power law growth . _ airborne diseases : _ airborne diseases require contact or proximity between two individuals for their transmission . in this case
the graph edges represent potential contact or proximity interactions among humans and the degree of an individual is given by the number of people with who he / she interacts within a certain period of time .
recent simulations of the portland urban dynamics @xcite shows that the number of people an individual contacts within a day follows a wide distribution up to about 10,000 contacts .
a report for the 2002 - 2003 sars epidemic shows a wide distribution as well , in this case for the number of secondary cases generated by a primary sars infection case .
although this data is not sufficient to make a definitive conclusion , it provides a clear indication that the number of proximity contacts a human undergo within a day is widely distributed .
this observation together with the high degree of assortativity of social networks opens the possibility that the spread of airborne diseases within a city is described by theorem [ nooo ] .
_ computer malwares : _ email worms and other computer malwares such as computer viruses and hoaxes spread through email communications .
the email network is actually directed , i.e. the observation that user a has user b on his / her address book does not imply the opposite .
this is an important distinction since the detailed balance condition ( [ db ] ) is valid for graphs with undirected edges .
there is , however , a significant reciprocity , meaning that if user a has user b on his / her address book then with high probability the opposite takes place as well .
thus , in a first approximation we can represent email connections by undirected links or edges and , in such a case , the detailed balance condition ( [ db ] ) holds .
recent studies of the email network structure within university environments indicate that they are characterized by a power law degree distribution with @xmath82 @xcite .
therefore , the spreading of computer malwares represents another scenario for the application of theorem [ nooo ] .
further research is required to determine the influence of the lack of reciprocity among some email users
. in conclusion , theorem [ nooo ] characterizes the spreading dynamics on complex networks with wide connectivity fluctuations .
its corollary [ coro ] determines the region of connectivity fluctuations and degree correlation where there is not an epidemic threshold .
the empirical data indicates that the theorem conditions are satisfied for several networks underlying the spreading of infectious diseases among humans and computer malwares among computers .
therefore , i predict that theorem [ nooo ] is a spreading law of modern epidemic outbreaks .
let @xmath83 be the probability distribution of the number of infected nodes at time @xmath35 ( including those that has been recovered ) , @xmath84 , on a branch of the ast [ acausaltree ] , given that branch is rooted at a node at generation @xmath52 and its has degree @xmath6 .
in particular @xmath85 is the probability distribution of the total number of infected nodes at time @xmath35 , given that patient zero ( the root of the tree ) became infected at time zero and its has degree @xmath6 .
based on the tree structure we can develop an iterative approach to compute @xmath83 recursively .
let @xmath32 be a node at generation @xmath52 with degree @xmath6 and let us denote by @xmath33 its neighbors on the next generation @xmath86 , where @xmath87 if @xmath23 , @xmath88 if @xmath89 , and @xmath90 if @xmath26 . then since node @xmath32 and its neighbors lie on a tree then @xmath95 are uncorrelated random variables .
furthermore , all nodes at generation @xmath52 has the same statistical properties , i.e. @xmath95 are identically distributed random variables .
let @xmath97 be the probability distribution of @xmath95 , given node @xmath33 is at generation @xmath86 and its ancestor has degree @xmath98 .
with probability @xmath99 the node @xmath33 is not infected at any time and with probability @xmath100 it is not yet infected at time @xmath35 given it will be infected at some later time . thus on the other hand , with probability @xmath45 node @xmath33 will be infected at some time @xmath31 , with distribution function @xmath46 , and continue the spreading dynamics to subsequent generations .
once node @xmath33 is infected , the number of infected nodes in the branch rooted at node @xmath33 is a random variable with probability distribution @xmath102 , given node @xmath33 has degree @xmath10 . | infectious diseases and computer malwares spread among humans and computers through the network of contacts among them .
these networks are characterized by wide connectivity fluctuations , connectivity correlations and the small - world property . in a previous work [ a. vazquez , phys .
rev .
lett .
96 , 038702 ( 2006 ) ] i have shown that the connectivity fluctuations together with the small - world property lead to a novel spreading law , characterized by an initial power law growth with an exponent determined by the average node distance on the network .
here i extend these results to consider the influence of connectivity correlations which are generally observed in real networks .
i show that assortative and disassortative connectivity correlations enhance and diminish , respectively , the range of validity of this spreading law . as a corollary
i obtain the region of connectivity fluctuations and degree correlations characterized by the absence of an epidemic threshold .
these results are relevant for the spreading of infectious diseases , rumors , and information among humans and the spreading of computer viruses , email worms and hoaxes among computer users . |
the einstein equations of general relativity have a hamiltonian formulation that arises as a consequence of the hilbert action principle in a 3 + 1 slicing @xcite .
the present article deals with numerical methods that respect the hamiltonian structure in the discretization . in various areas of scientific computing , such as the dynamics of particle accelerators , molecular dynamics , celestial mechanics , quantum dynamics and electrodynamics , _ symplectic integrators _ for hamiltonian systems have been essential for attaining favorable propagation properties ( see , e.g. , @xcite and references ) .
there have also been a few papers on symplectic integrators for free evolution in general relativity @xcite .
very recently , in @xcite the difficulties in devising constraint - preserving symplectic integrators have been addressed but not resolved . in this paper
we present symplectic integrators for free and constrained evolution in numerical relativity and illustrate and compare their numerical properties in numerical experiments with effectively 1 + 1-dimensional versions of einstein s equations .
in section 2 we describe the general framework of the hamiltonian formulation and its spatial discretization . for the free
, adm - like evolution we propose the _ strmer
verlet method _ in section 3 .
this standard symplectic integrator ( see @xcite ) here is an implicit method , but it is explicit in the computationally expensive terms containing the discretized ricci scalar .
moreover , the implicitness is point - wise in space and is readily resolved by simple fixed - point iteration .
however , in this free evolution scheme the hamiltonian and momentum constraints are not considered and may drift off . in section 4
we give a formulation of the ( spatially discretized ) einstein equations that enforces the momentum constraints in a holonomically constrained hamiltonian system .
this formulation interprets the shift vector as additional momentum variables and at the same time fixes it by gauge conditions . in the spatially continuous problem
, a result in @xcite implies that with the enforced momentum constraints , the hamiltonian constraint function satisfies a conservation law .
even if this property does not extend to the space discretization , it is an extra bonus for this momentum - constrained formulation .
the problem of constraint growth in numerical relativity has also been tackled , using various techniques , in other free and constrained evolution schemes ( see e.g. @xcite ) .
stable evolutions of the black hole binary problem are possible since 2005 @xcite . in the approach presented here
we avoid the growth of the momentum constraints and additionally respect the hamiltonian structure of the equations .
although no symplectic integrators are known for general constraints in hamiltonian systems , such integrators do exist for holonomically constrained systems .
the most basic of these methods is the _ rattle method _ , a constrained version of the strmer verlet method . in section 5
we discuss its application to the holonomically constrained formulation of section 4 .
section 6 presents the 1 + 1-dimensional examples with which we have done our numerical experiments : a perturbed minkowski problem and schwarzschild space - time .
we describe the spatial discretization , the dirac gauge condition , and the boundary treatment that we used in our numerical experiments .
section 7 compares a non - symplectic runge - kutta method and the symplectic strmer verlet method on a perturbed minkowski problem , showing very different dynamical behavior in the harmonic energies corresponding to different modes . these energies are almost - conserved quantities in the time - continuous problem and in the symplectic method , but completely fail to be preserved in the non - symplectic method .
section 8 compares the free and the constrained symplectic schemes on a different perturbed minkowski problem , illustrating that the unstable free evolution can be stabilized by enforcing the momentum constraints .
section 9 compares free and constrained symplectic integrators on the schwarzschild space - time .
it turns out that in the constrained scheme , boundary instabilities do not propagate into the interior of the domain , as opposed to the free scheme .
our numerical experiments thus illustrate remarkable properties of symplectic vs. non - symplectic and constrained vs. free integrators that apparently have not been addressed in the literature before .
for classical general relativity a variety of hamiltonian formulations is known ( see @xcite for a summary of the most popular ones ) .
our work is based on the famous _ super - hamiltonian _
@xcite ( see also @xcite ) .
it was discovered as a preliminary to the quantization of gravity and relies on a 3 + 1 splitting of space - time .
the geometry is described by the 3-metric of the spatial slices and their extrinsic curvature . in the super - hamiltonian ,
the position variables are provided by the 3-metric @xmath0 and the corresponding canonical momenta are denoted @xmath1 .
they are related to the extrinsic curvature @xmath2 by @xmath3 . here
@xmath4 is the determinant of the metric @xmath0 .
the super - hamiltonian takes the form @xmath5 with freely specifiable functions @xmath6 and @xmath7 .
the _ densitized lapse _
@xmath6 is related to the lapse function @xmath8 by @xmath9 , and the vector with components @xmath7 is the _ shift vector_. the functions @xmath10 and @xmath11 are given by @xmath12 where @xmath13 is the ricci scalar of the metric @xmath0 , and is a tensor density of weight @xmath14 . ]
@xmath15 solutions to the canonical hamiltonian equations of motion @xmath16 are solutions of einstein s equations if they satisfy the _
scalar constraint _
( or hamiltonian constraint ) and the _ vector constraints _ ( or momentum constraints ) @xmath17 if these constraints are satisfied for the initial data , then they remain satisfied in the course of the evolution . however , these functions do not satisfy a conservation law for general initial data . in this sense , @xmath10 and @xmath11 are _ weak invariants _ but not strong invariants . in section [
sect : constrained ] we will consider a formulation where the momentum constraints @xmath11 are enforced ( treating @xmath7 as dynamical variables ) and the hamilton constraint function @xmath10 becomes a strong invariant . the system is the adm system presented in @xcite .
it is weakly hyperbolic , but not strongly hyperbolic , and hence it has an ill - posed initial value problem . for practical applications one will therefore prefer to adopt another hamiltonian formulation of general relativity , like e.g. a generalized harmonic system @xcite .
however , we focus on the description of applications of symplectic integrators . adopting other formulations will not change our concept , but the details will be more complicated . for a reasonable comparison of the properties of the different time integration methods in the free evolution schemes we will consider an example where the unstable modes of the ill - posed system are not excited in the course of the simulation ( see section [ sect : exp1 ] ) .
concerning the constrained evolution scheme ( see sections [ sect : rattle ] ) we consider the modified system , where the equations as well as the vector constraints are explicitly imposed .
unfortunately nothing is known yet about the well - posedness of the initial value problem of this constrained system . to apply numerical methods
one will always approximate the continuous functions @xmath0 , @xmath1 , @xmath7 and @xmath6 through objects with finitely many degrees of freedom , and introduce discrete derivative operators that act on these finite dimensional spaces . with those ingredients a discrete hamiltonian
is derived by replacing the continuous functions and derivative operators in with the discrete ones .
a concrete example of such a finite - difference discretization is given in section [ sec : discretisation11 ] .
with some care in defining the discrete canonical momenta , also finite element or spectral discretizations yield finite - dimensional canonical hamiltonian equations of motion .
we collect the discrete degrees of freedom corresponding to the functions @xmath0 , @xmath1 , @xmath6 , @xmath7 in vectors @xmath18 , @xmath19 , @xmath20 , @xmath21 , respectively . the ordering in these vectors
is chosen such that components corresponding to the same spatial grid point are ordered consecutively .
since the super - hamiltonian consists of terms that are either quadratic in the momenta @xmath1 or linear in both @xmath1 and @xmath7 or independent of the momenta , any reasonable discretization of the super - hamiltonian @xmath22 will assume the form ( we ignore the dependence on the discrete densitized lapse @xmath20 in the notation ) @xmath23 where @xmath24 and @xmath25 are matrices of the appropriate dimensions .
the matrix @xmath24 is a square and symmetric matrix . in a finite - difference discretization ,
this matrix is block - diagonal with six - dimensional blocks corresponding to the six components of @xmath1 at a grid point , since the term in the super - hamiltonian that is quadratic in the momenta does not contain spatial derivatives .
we then have @xmath26 blockdiag@xmath27 , where @xmath28 contains the six components of the 3-metric at the grid point @xmath29 .
the discrete hamiltonian is thus quadratic in the momenta , but the functional dependence on the position variables is more complicated . the computationally expensive term with the discretized ricci scalar is subsumed in the potential @xmath30 .
the canonical equations of motion for this discrete hamiltonian then read @xmath31 \dot { { \mathbf p}}&= & -\frac12 { { \mathbf p}}^t { { \boldsymbol \nabla}}_{{\mathbf q}}{{\mathbf s}}({{\mathbf q}}){{\mathbf p}}-{{\boldsymbol \nabla}}_{{\mathbf q}}u({{\mathbf q } } ) -{{\boldsymbol \beta}}^t{{\boldsymbol \nabla}}_{{\mathbf q}}{{\mathbf d}}({{\mathbf q}}){{\mathbf p}}\ , .
\end{array}\ ] ] without further ado , the discretized momentum constraints @xmath32 , which read @xmath33 are _ not _ preserved under the evolution of the discretized system ( [ eq : discrete_evolution_equations ] ) , nor are the discretized hamilton constraints @xmath34 preserved
. there may be an exponential or even super - exponential drift away from these constraints along solutions of the discrete system ( [ eq : discrete_evolution_equations ] ) .
a standard symplectic integrator for hamiltonian systems is the strmer
verlet scheme , which is particularly prominent in the area of molecular dynamics and enjoys a number of remarkable properties ; see , e.g. , @xcite . when applied to ( [ eq : discrete_evolution_equations ] ) , a step from old values @xmath35 at time @xmath36 to values @xmath37 at time @xmath38 reads as follows : @xmath39 there is only one evaluation per step of the potential @xmath40 that contains the computationally expensive gradient of the discretized ricci scalar . the substeps ( [ sv - p1 ] ) and ( [ sv - q ] ) are implicit in @xmath41 and @xmath42 , respectively .
they are solved by fixed - point iteration , which is local at every grid point .
the choice @xmath43 for the shift vector further simplifies the formulas .
the last substep ( [ sv - p2 ] ) is explicit .
we will present numerical experiments with this method in sections [ sect : exp1][sect : exp3 ] .
the strmer verlet integrator is a second - order method .
we remark that higher - order symplectic methods are obtained by suitable compositions of steps with different step sizes ; see , e.g. , ( * ? ? ?
* section v.3 ) .
since the constraints are not taken care of in the free evolution ( [ eq : discrete_evolution_equations ] ) , there may be an uncontrollable drift in the discretized momentum and hamiltonian constraints . while there exist symplectic integrators for holonomically constrained hamiltonian systems , see (
* section vii.1 ) and ( * ? ? ?
* chapter 7 ) , there exist no general symplectic integrators for systems where the constraints depend on both position and momentum variables , as is the case with the momentum and hamilton constraints in general relativity .
we therefore look for a reformulation of the equations of motion for general relativity that enforces the momentum constraints via holonomic constraints and , as an extra benefit , turns the hamilton constraint function from a weak invariant into a strong invariant ( i.e. , satisfying a conservation law ) .
while this reformulation can equally be done on the continuous level , we here present it for the discrete equations of motion ( [ eq : discrete_evolution_equations ] ) to which we will apply a symplectic integrator in the next section .
we now consider @xmath21 as a dynamical variable and deal with it in two seemingly contradictory ways : 1 .
we fix @xmath21 by a _ gauge condition _
@xmath44 time differentiation of @xmath45 and using for @xmath46 gives @xmath47 which shows that indeed @xmath21 is determined by the gauge condition , when @xmath48 is invertible .
+ a candidate for the choice of the gauge function @xmath49 is a discretization of the _ dirac gauge _ , @xmath50 . with this choice , @xmath48 is a discretized second - order elliptic differential operator .
we consider @xmath21 as _ new momentum variables _ canonically conjugate to position variables @xmath51 that are not present in the hamiltonian ( [ h - disc ] ) , viz . , @xmath52 which still is quadratic in the momenta @xmath53 . in the free evolution ,
the equations of motion for this extended hamiltonian are ( [ eq : discrete_evolution_equations ] ) together with @xmath54 and @xmath55 . in order to enforce the discrete momentum constraints
@xmath56 , we therefore impose the _ holonomic constraints _ ( depending only on the position variables @xmath57 ) @xmath58 the corresponding _ hidden constraints _ obtained by time differentiation of these constraints and using the unchanged expressions for @xmath59 ( see ( [ eq : discrete_evolution_equations ] ) ) and @xmath60 then yields @xmath61 this way , the discrete momentum constraints appear as hidden constraints of a holonomically constrained hamiltonian system .
moreover , @xmath21 is completely determined by the first equation in ( [ hidden ] ) .
the full set of equations of motion with lagrange multipliers @xmath62 corresponding to the holonomic constraints is now @xmath63 \dot{{\mathbf p}}&= & -\frac12 { { \mathbf p}}^t { { \boldsymbol \nabla}}_{{\mathbf q}}{{\mathbf s}}({{\mathbf q}}){{\mathbf p}}-{{\boldsymbol \nabla}}_{{\mathbf q}}u({{\mathbf q } } ) -{{\boldsymbol \beta}}^t{{\boldsymbol \nabla}}_{{\mathbf q}}{{\mathbf d}}({{\mathbf q}}){{\mathbf p}}\ , - { { \boldsymbol \nabla}}_{{\mathbf q}}{{\mathbf g}}({{\mathbf q } } ) { { \boldsymbol \lambda}}\end{array}\ ] ] together with the constraints ( [ holo ] ) and ( [ hidden ] ) and formally also the equations @xmath64 with lagrange multipliers @xmath65 corresponding to the constraints @xmath66 .
it is to this formulation that we will apply a suitable numerical integrator in the next section .
this formulation , which can be similarly given also for the spatially continuous problem , does not enforce the hamilton constraint .
however , in the continuous case , a result by anderson & york @xcite shows that satisfying the momentum constraints @xmath32 implies that the hamilton constraint function satisfies a conservation law @xmath67 where @xmath68 is the covariant derivative operator in the direction of the shift vector @xmath69 .
the rattle method ( @xcite , ( * ? ? ? * section vii.1 ) , ( * ? ? ? * chapter 7 ) ) is an extension of the strmer verlet method to holonomically constrained systems .
it is symplectic and time - reversible , of second order accuracy , and enforces both the holonomic and the derived hidden constraints in the numerical solution . when applied to ( [ holo])([trivial - eom ] ) , a step of the rattle method consists of the following equations , which form a nonlinear system for @xmath42 and @xmath70 . 1
. first half - step for the momentum variables : @xmath71 2 .
full step for the position variables : @xmath72 3 .
position constraints : @xmath73 4 .
second half - step for the momentum variables : @xmath74 5 .
momentum constraints : @xmath75 \label{p - constr } { { \mathbf d}}({{\mathbf q}}^{n+1}){{\mathbf p}}^{n+1}&=&{{\mathbf 0}}\,.\end{aligned}\ ] ] equations ( [ p - half])([gamma - constr ] ) determine @xmath42 , and ( [ p - half-2])([p - constr ] ) determine @xmath70 .
the equations can be solved by an iterative procedure that requires only the solution of linear systems with the matrices @xmath76 and @xmath77 and their transposes .
* iterative solution of ( [ p - half])([gamma - constr ] ) : * we start with @xmath78 as initial iterate for @xmath42 , @xmath79 for @xmath41 , @xmath80 for @xmath81 , and @xmath82 for @xmath83 . with these values
we first update the iterates @xmath42 and @xmath41 through and respectively , inserting the initial iterates at the right - hand sides of these equations .
then , for given iterates of @xmath42 , @xmath41 and @xmath81 , equations ( [ gamma - constr ] ) , ( [ gamma - new ] ) and ( [ p - half ] ) yield an equation for @xmath83 : @xmath84 the matrix on the left - hand side is @xmath85-close to @xmath86 and therefore invertible under condition ( [ gauge ] ) . equation can be interpreted as the requirement to choose @xmath83 such that the momentum constraints are satisfied for the next iterate , @xmath87 .
instead of using to obtain @xmath87 from the current iterates we may also assume that it is approximately @xmath88 with a small correction @xmath89 to the lagrange multipliers .
this leads to the following equation , @xmath90 we solve this approximately for the increment @xmath89 , and replace the current iterate @xmath91 .
inserting ( [ q - new ] ) into ( [ q - constr ] ) gives an equation for @xmath81 . with the current iterates in the argument of @xmath49
, we solve for the increment @xmath92 in a step of a simplified newton iteration : @xmath93 and replace @xmath94 .
we then update the iterates of @xmath41 and @xmath42 using and ( [ q - new ] ) respectively with the current iterates on the right - hand side .
we thus have the following schematic iteration cycle : @xmath95 * solution of ( [ p - half-2])([p - constr ] ) : * inserting ( [ p - half-2 ] ) into ( [ p - constr ] ) yields a linear system for @xmath96 with the matrix @xmath97 . after solving this system
we compute @xmath70 from ( [ p - half-2 ] ) , and then @xmath98 is obtained from solving the linear system ( [ beta - constr ] ) with the matrix @xmath77 .
we remark that equations ( [ beta - half ] ) and ( [ beta - half-2 ] ) are ignored , since they only determine approximations to the lagrange multipliers @xmath99 that are without further interest .
* symplecticity . *
the rattle method is symplectic with respect to the canonical symplectic two - form @xmath100 for the _ extended _ phase space of @xmath101 restricted to the constraint manifold ; see ( * ? ? ?
* section vii.1 ) , ( * ? ? ?
* chapter 7 ) .
because of @xmath102 , the extended symplectic two - form here actually reduces to the original canonical symplectic two - form @xmath103 .
after the general discussion in the previous sections we now describe in detail the problems considered in our numerical experiments .
we derive a simplified hamiltonian in section [ sec : simple_hamiltonian ] , and in section [ sec : discretisation11 ] we describe the spatial discretization procedure that leads to a discrete hamiltonian . in section [ sec:11gauge ]
we introduce the gauge conditions that we use and discuss some problems that are related to that topic .
we describe a straightforward boundary treatment in section [ sec : boundary_conditions ] , before we turn to the particular test problems in section [ sec : free_test_scenarios ] . for a first test
we restrict to simple problems .
we therefore consider only those solutions of einstein s equations that satisfy the following requirements : @xmath104 where either @xmath105 or @xmath106 .
the latter case corresponds to spherically symmetric space - times , and the class of solutions with @xmath105 includes a perturbed minkowski geometry .
both cases will be considered later on .
it is easy to show that one obtains solutions in either of these classes if is satisfied at the initial hypersurface @xmath107 , if moreover at @xmath107 @xmath108 ( where @xmath109 or @xmath110 in the two cases , respectively ) and if the gauge is chosen such that everywhere @xmath111 to summarize , if is satisfied at every time slice and if , are satisfied at @xmath107 then , are satisfied for all times , as long as einstein s equations hold . because of the last equations in , it is natural to define @xmath112 it turns out that with these definitions @xmath113 is indeed the canonical momentum corresponding to @xmath114 . in the spherically symmetric case
we now consider the equatorial hypersurface , i.e. , @xmath115 and @xmath116 . using the new variables it can be shown that the following simplified hamiltonian provides the correct evolution equations for the functions @xmath117 , @xmath118 , @xmath114 and @xmath119 : @xmath120,\end{aligned}\ ] ] where @xmath121 in the spherically symmetric case and @xmath122 if @xmath105 .
this hamiltonian has a similar structure to the super - hamiltonian , its discretization will therefore also be of the form .
based on we now derive a discrete hamiltonian .
we introduce two uniform and @xmath123 . ] spatial grids @xmath124 and @xmath125 , , otherwise @xmath126 .
] staggered such that @xmath127 .
we approximate the functions @xmath128 , @xmath117 , @xmath118 , @xmath114 , @xmath119 by piecewise constant functions @xmath129 , such that @xmath130 and then replace the continuous functions in @xmath131 by the piecewise constant ones ( @xmath132 ) .
for the shift function @xmath133 the staggered grid @xmath125 is the basis of this discretization , i.e. , @xmath134 discretizing @xmath133 on the staggered grid leads to better results in the constrained evolution .
spatial derivatives are approximated by those piecewise constant functions whose functional values are obtained through centered finite differencing , i.e. , @xmath135 for @xmath136 .
an analogous formula is used also for the derivatives of @xmath133 on the grid @xmath137 .
these functions again replace the continuous functions @xmath138 and @xmath139 , respectively , in the continuous hamiltonian @xmath131 . to get the discretization of the integrand in one additionally needs to take a logarithm and to perform additions as well as multiplications .
these are pointwise operations where the continuous functions can be replaced by the piecewise constant ones easily .
hence , the integrand of the discrete hamiltonian @xmath140 is again a piecewise constant function . comparing the discrete hamiltonian for the 1 + 1 dimensional setting with its general form we identify the potential @xmath30 with the discretization of the second line in .
the matrix @xmath141 is derived from the kinetic term @xmath142 and the matrix @xmath143 comes from the shift term @xmath144 as we discussed in section [ sec : discretization_general ] , the matrix @xmath141 is blockdiagonal with each block corresponding to a single grid point , because does not contain spatial derivatives . here ,
in the 1 + 1-dimensional setting , the blocks are two dimensional corresponding to the functions @xmath118 and @xmath113 respectively , and the elements of the blocks can be obtained through @xmath145 concerning the matrix @xmath143 , the integrand in does contain derivatives . since we use a staggered grid for the discretization of @xmath133 ,
here @xmath143 becomes an @xmath146 matrix of the form @xmath147 where each star represents a row - vector @xmath148 for the constrained evolution scheme discussed in section [ sect : constrained ] we additionally need to choose a gauge constraint @xmath45 . in what follows
we derive this function from the dirac gauge @xcite , a condition that in the continuous problem fixes the spatial coordinate system .
this gauge condition is formulated in terms of a flat background metric , see @xcite for details . in the 1 + 1 dimensional context
the dirac gauge condition becomes @xmath149 where @xmath121 in the spherically symmetric case and @xmath122 for the @xmath105 class of problems .
the dependence on @xmath150 for spherically symmetric systems comes from the flat background metric , which is the minkowski metric in spherical coordinates there . from the continuous gauge condition we derive a discrete one as follows .
first we observe that we obtain equation when we require that the variation of the following integral with respect to @xmath151 vanishes @xmath152 this integral is again discretized by the procedure described in section [ sec : discretisation11 ] , that is , the functions @xmath117 , @xmath114 and @xmath150 are approximated by piecewise constant functions corresponding to the grid @xmath153 and @xmath151 becomes a piecewise constant function based on the staggered grid @xmath125 .
we thus obtain a discrete function @xmath154 of the grid variables @xmath18 , @xmath155 and @xmath62 . following the procedure in the continuous case
we take the derivative of @xmath154 with respect to @xmath62 as the discrete gauge constraints .
this defines functions @xmath156 , but as discussed in section [ sect : constrained ] we additionally need that the matrix @xmath48 is invertible .
it is not a priori clear that this condition is satisfied .
therefore we made several spot checks by calculating the singular values of @xmath48 .
it turns out that in the simulations where the computational domain possesses boundaries , in particular for spherical symmetry , we indeed find that the ratio of the largest to the smallest singular value was at most @xmath157 and the ratio of the smallest to the second smallest was always below @xmath158 . but with periodic boundary conditions it turns out that one singular value is more than ten orders of magnitude smaller than the others , and hence @xmath48 must be regarded as singular .
* constraints for periodic boundary conditions .
* the origin of the non - trivial kernel of @xmath48 can be found already in the continuous 1 + 1 dimensional formulation . there
the expression @xmath159 translates to @xmath160 \right).\end{aligned}\ ] ] it can be easily checked that with periodic boundary conditions the equations @xmath161 hold for any function @xmath162 .
that means the constant functions are in the kernel of the transpose operator @xmath163 , and functions of the form @xmath164 ( @xmath165 ) are in the kernel of @xmath166 . if the dirac gauge is satisfied then @xmath167 .
, so that the constant functions are also in the kernel of @xmath166 .
this argument can also be extended to other gauge conditions and probably to higher dimensions .
we do not want to elaborate on this here , we only mention that for periodic boundary conditions @xmath166 will have a non - trivial kernel for any pointwise gauge condition of the form @xmath168 .
now , the singularity of @xmath48 means that we can not solve equations and for @xmath169 and @xmath170 in the rattle scheme .
we therefore solve another system which enforces the discrete momentum constraints only up to a multiple of the spatially constant vector @xmath171 .
it turns out that for @xmath45 this vector is indeed in the kernels of @xmath48 and @xmath172 , and that the system @xmath173 has a unique solution for any right - hand side @xmath174 .
we then solve systems of this form instead of and , and take the obtained @xmath169 and @xmath170 as the right correction to the lagrangian multipliers and the shift variables respectively . however , with that procedure we do not have control of the mean value of the momentum constraints .
we see this immediately when we notice that @xmath175 is independent of @xmath62 ( since @xmath176 is in the kernel of @xmath48 ) .
hence , the mean value of the momentum constraints can not be brought to zero by this procedure , and instead of satisfying , we can only fulfill @xmath177 where @xmath178 is the projector to the subspace orthogonal to @xmath176 .
the full discrete momentum constraints @xmath179 could be enforced by adding an extra gauge condition as a discretization of an integral condition @xmath180 by @xmath181 where the function @xmath182 and the vector @xmath174 are chosen such that @xmath183 it turns out that this condition corresponds to @xmath184 in the continuous case .
we have , however , not included such an extension by an integral gauge condition in our numerical experiments .
we now describe the implementation of boundary conditions .
as we focus on the investigation of the time evolution , we tried to keep this aspect simple .
for the tests in the @xmath105 class of solutions we impose periodic boundary conditions and there is actually no boundary .
but in the spherically symmetric case it is not possible to impose periodic boundary conditions . for our spatial discretization
, the time derivatives of the variables at the grid point @xmath185 only depend on the variables at the neighboring grid points @xmath186 , @xmath187 , @xmath188 , @xmath189 , @xmath190 and @xmath191 , @xmath192 , @xmath193 , @xmath194 .
therefore we extend the spatial grids beyond the computational domain , that is , we introduce ghost zones to the left and to the right of the boundary ( see figure [ fig : spatial_grid ] ) .
the function values at the grid points in these ghost zones are then treated as non - dynamical variables . in general
these function values are fixed through the choice of boundary conditions , and here we simply set them to values that we read off from a reference solution .
the discrete hamiltonian still has the form , we only change the interpretation of the variables .
therefore the equations of motion for the variables in the computational domain have the same form as in the case of periodic boundary conditions and the variables in the ghost zones do not possess evolution equations .
( 0,0 ) # 1#2#3#4#5 @font ( 5159,915)(706,-3035 ) ( 2026,-2986)(0,0)[lb ] ( 3826,-2986)(0,0)[lb ] ( 1441,-2221)(0,0)[lb ] ( 4681,-2356)(0,0)[lb ] ( 4681,-2221)(0,0)[lb ] ( 2161,-2356)(0,0)[lb ] ( 721,-2356)(0,0)[lb ] in the general form of the hamiltonian we then have to distinguish between dynamical variables in the computational domain and non - dynamical variables in the ghost zones .
in particular we must subdivide the matrix @xmath143 such that the non - dynamical character of the ghost variables can be considered .
then , @xmath143 takes the form ( cf .
equation ) @xmath195 where @xmath196 corresponds to grid points in the computational domain and the remaining matrices to grid points in the ghost zones .
for @xmath141 the consideration of ghost variables leads to an analogous result , but @xmath141 is block diagonal and hence the only non vanishing `` ghost matrices '' are @xmath197 and @xmath198 .
these matrices only depend on variables in the ghost zones .
hence , they have no influence on the equations of motion and are therefore irrelevant .
we collect the ghost matrices in three matrices @xmath199 if we denote the variables in the ghost zones as @xmath200 , @xmath201 and @xmath202 , respectively , then the hamiltonian can be written as @xmath203 from this hamiltonian one obtains the equations of motion for @xmath18 and @xmath19 , as well as the one - step map in the strmer verlet scheme as described in sections [ sec : discretization_general ] and [ sect : sv ] , respectively .
we note that with the chosen space discretization , @xmath204 .
moreover , it turns out that @xmath205 does not depend on @xmath201 , and so the discrete momentum constraints become @xmath206 and thus do not depend on the variables in the ghost zones .
this structure is due to the discretization of the shift @xmath133 at the staggered grid @xmath137 .
if the shift is discretized at @xmath207 then @xmath143 does not have the bidiagonal structure , but there are more non - vanishing elements , and the momentum constraints depend on @xmath208 and @xmath209 as well .
the grid functions in the ghost zones are specified as @xmath210 for each grid function @xmath162 , each grid point in the ghost zone @xmath211 and all times @xmath212 , with given reference functions @xmath213 in the exterior .
this does not apply to the lagrange multipliers @xmath62 , which are defined only on the staggered grid in the interior domain .
in our example of the schwarzschild space - time given below , the analytically known stationary solution is chosen as initial data and exterior data , leaving only a numerical dynamics that reflects the stability properties of the various numerical schemes . in the following sections we investigate the properties of the strmer verlet and the rattle method in two different situations .
the first one , a perturbed minkowski space - time , is an example for the case @xmath105 , and the second one , the schwarzschild space - time , is a spherically symmetric solution .
[ [ perturbed - minkowski . ] ] perturbed minkowski .
+ + + + + + + + + + + + + + + + + + + + the minkowski metric describes a flat space - time , the analytical solution is ( with @xmath214 , @xmath215 , @xmath216 ) @xmath217 it is easy to check that for @xmath218const . slicing
this solution really is in the class with @xmath105 and thus the numerical schemes we described are applicable . here
we perturb the minkowski initial data .
we denote the perturbations @xmath219 and @xmath220 , and get that at a slice @xmath221 the independent components of the 3-metric are @xmath222 the canonical momenta become @xmath223 and without perturbations the slicing density is identical one and the shift vector vanishes : @xmath224 the spatial grid is chosen as the uniform grid @xmath225 and we apply periodic boundary conditions . [ [ schwarzschild - space - time - in - isotropic - coordinates . ] ] schwarzschild space - time in isotropic coordinates .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the schwarzschild space - time in isotropic coordinates is described by the metric ( with @xmath226 , @xmath227 , @xmath228 ) @xmath229 where @xmath230 . for @xmath218const .
slicing we thus obtain @xmath231 the extrinsic curvature and the canonical momenta vanish as well as the shift , @xmath232 and the densitized lapse is @xmath233 the spatial grid is chosen as @xmath234 to treat the boundaries we continue the uniform spatial grids beyond the boundaries , introducing @xmath235 grid points in ghost zones to the left and the right . the schwarzschild solution in isotropic
coordinates satisfies dirac gauge .
this permits us to obtain initial data easily and to compare the numerical results with the exact stationary schwarzschild solution . applying a numerical method with these initial data
yields a purely numerical dynamics that reflects the stability properties of the various numerical schemes .
[ [ symplectic-strmerverlet-vs.non-symplectic-icn-scheme . ] ] symplectic strmer verlet vs. non - symplectic icn scheme .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we compare the results of the strmer verlet scheme ( see section [ sect : sv ] ) with that of a free non - symplectic scheme , the iterated crank nicholson method ( icn ) . for a system of ordinary differential equations @xmath236 , icn reads @xmath237 in both schemes the densitized lapse as well as the shift are chosen not to change with time , @xmath238 and @xmath239 .
we apply the schemes to the perturbed minkowski example of section [ sec : free_test_scenarios ] , where the perturbation is chosen such that the functions denoted @xmath219 are gaussian functions with width @xmath240 , center @xmath241 and height @xmath242 .
the functions denoted @xmath220 are chosen to vanish .
that is , we perturb @xmath117 , @xmath119 , @xmath6 and @xmath133 , but we keep @xmath114 and @xmath118 from the unperturbed minkowski problem .
the case where all functions are perturbed by random noise , is discussed in the next section .
[ [ simulation - data . ] ] simulation data .
+ + + + + + + + + + + + + + + + the reason to choose @xmath243 here is the hyperbolicity of the considered system .
it turns out that the equations of motion that correspond to are weakly hyperbolic , but not strongly hyperbolic in the sense of @xcite .
however , if we have @xmath244 and @xmath245 at some time @xmath246 then the continuous as well as the discrete evolution equations lead to @xmath247 and @xmath248 .
we can hence omit these functions in the analysis of hyperbolicity .
the obtained system for @xmath117 and @xmath113 is then strongly hyperbolic . in the simulations we use grids with @xmath249 or @xmath250
grid points .
the size of the time step is the same as the spatial grid spacing , i.e. , @xmath251 .
the computational costs are comparable for both time - stepping methods .
[ [ simulation - results . ] ] simulation results .
+ + + + + + + + + + + + + + + + + + + it turns out that for both integrators the simulations are stable at least until @xmath252 .
we do not see any growing modes , and we expect that the simulations remain stable for a much longer time .
the discrete momentum constraint vanishes identically , because @xmath118 as well as @xmath253 vanish .
the discrete hamilton constraint does not vanish exactly , but it stays almost constant at about @xmath254 for both schemes . to see how the perturbations evolve , we calculate the _ harmonic energies _ corresponding to @xmath117 .
let @xmath255 be the vector composed of the grid values of @xmath117 , and let @xmath256 denote the @xmath257th component of its discrete fourier transform .
let further @xmath258 be the @xmath257th component of the fourier transform of @xmath259 .
then , the harmonic energy in the @xmath257th mode is @xmath260 .
as we see in figure [ fig : exp1:harmonic_energies_icn_sv ] , the @xmath261 are approximately preserved by the strmer verlet scheme over long times , whereas with the icn scheme they incorrectly decay exponentially until they reach the level of round - off error .
the exponential decay is much faster for high frequencies . [
cols="^,^ " , ] [ [ discussion - of - results.-2 ] ] discussion of results .
+ + + + + + + + + + + + + + + + + + + + + + we have seen that in the free evolution again high frequency errors occur , whereas they are absent in the constrained evolution scheme .
our simple choice of boundary conditions of section [ sec : boundary_conditions ] leads to boundary instabilities , and so the main source of errors are the boundaries . in the free evolution
the errors at the boundaries quickly propagate into the interior of the computational domain and amplify . in the constrained scheme ,
the propagation into the interior is suppressed .
we are grateful to sascha husa and bernd brgmann for their helpful advice .
this work was supported by dfg grant sfb / transregio 7 `` gravitational wave astronomy '' .
j. frauendiener . . ,
41 no 38:382005 , 2008 .
r. gambini and j. pullin .
consistent discrete space - time . in a.
ashtekar , editor , _ 100 years of relativity space - time structure : einstein and beyond_. world scientific , 2005 . | we consider symplectic time integrators in numerical general relativity and discuss both free and constrained evolution schemes . for free evolution of adm - like equations we propose the use of the strmer verlet method , a standard symplectic integrator which here is explicit in the computationally expensive curvature terms .
for the constrained evolution we give a formulation of the evolution equations that enforces the momentum constraints in a holonomically constrained hamiltonian system and turns the hamilton constraint function from a weak to a strong invariant of the system .
this formulation permits the use of the constraint - preserving symplectic rattle integrator , a constrained version of the strmer verlet method .
the behavior of the methods is illustrated on two effectively 1 + 1-dimensional versions of einstein s equations , that allow to investigate a perturbed minkowski problem and the schwarzschild space - time .
we compare symplectic and non - symplectic integrators for free evolution , showing very different numerical behavior for nearly - conserved quantities in the perturbed minkowski problem .
further we compare free and constrained evolution , demonstrating in our examples that enforcing the momentum constraints can turn an unstable free evolution into a stable constrained evolution .
this is demonstrated in the stabilization of a perturbed minkowski problem with dirac gauge , and in the suppression of the propagation of boundary instabilities into the interior of the domain in schwarzschild space - time .
pacs numbers : 04.25.d- , 04.20.fy |
the ` stingray nebula ' ( v839 ara , sao 244567 , cd -59@xmath0 6479 , cpd -59@xmath0 6926 , hen 3 - 1357 , and pn g331 - 12.1 ) is a unique case where an ordinary post asymptotic giant branch ( post - agb ) star suddenly changed its appearance to that of a young planetary nebula ( pn ) .
this ejection of the pn shell was roughly one millennium ago , but the nebula turned on suddenly around 1980 with some sharp increase in ionizing radiation .
the stingray represents our one opportunity to actually watch the turn - on of a pn . before 1980 ,
four spectra of the stingray have been published , all showing either very weak balmer emission lines or no emission lines ( see the timeline in table 1 ) . during this time , as best seen in the 1971 spectrum ( parthasarathy et al .
1995 ) , the spectrum was that of a normal b0 or b1 star , placed into luminosity class i or ii , with possible weak emission at @xmath1 only .
the star slowly began to attract attention , first for having just some @xmath2 emission ( henize 1976 ) , then as an _ iras _ far - infrared source selected out as a proto - planetary nebula ( volk & kwok 1989 ; parthasarathy & pottasch 1989 ) , then as a star that had a sudden appearance of a pn shell within the previous few years ( parthasarathy et al .
1993 ) . by 1990 and 1992 , the stingray optical spectrum was dominated by very bright and narrow [ oiii ] emission lines , plus other lines that are characteristic of a young pn ( parthasarathy et al . 1993 ; 1995 ) .
the stark difference between the 1971 spectrum ( a b1i star with weak emission only ) and the 1990 spectrum ( very bright pn emission lines ) impresses that the stingray is evolving fast and apparently the pn has just turned - on . from the start of the 1980 s event with the fast turn - on of the pn spectrum , the stingray has been fast evolving ( see table 1 ) . at the start , a b0 spectrum gives a surface temperature of around 30,000 k. from the time of the first _ international ultraviolet explorer _ ( _ iue _ ) spectrum , the effective temperature of the star has been heating up greatly ( parthasarathy et al .
1995 ) , going from 38,000 k in 1988 to 55,000 k in 2006 ( reindl et al . 2014 ) .
the surface gravity has increased from @xmath3 cm s@xmath4 in 1988 to @xmath5 cm s@xmath4 in 2006 ( reindl et al . 2014 ) .
the mass ejection rate has decreased from @xmath6 @xmath7 year@xmath8 in 1988 to @xmath9 @xmath7 year@xmath8 in 2006 ( reindl et al . 2014 ) .
the ultraviolet continuum brightness has fallen by a factor of three from 1988 to 1994 ( parthasarathy et al .
1995 ; feibelman 1995 ) , continued to fall by another factor of three to 1997 , but then brightened by a factor of two in 2002 and 2006 ( reindl et al .
we have a clear picture of a star in 1988 with a high stellar wind , that rapidly tapered off , while the star shrunk in size ( by a factor of four ) and and heated up .
the shrinking and heating of the central star is what would be expected for a simple picture of a stellar core following along the evolutionary track that leads to a white dwarf .
the first distance measure to the stingray was 5.64 kpc ( kozok 1985b ) , based on a presumed absolute magnitude appropriate for the star s color , and as such this estimate has a very large real uncertainty .
fresneau et al .
( 2007 ) measure the proper motion of the stingray , and from this derive a statistical parallax of [email protected] milli - arc - seconds ( for a distance of 830 pc and an uncertainty of 17% ) , while the possible deviation from their statistical model makes for a possibly large error in this distance .
the best distance to the stingray is based on the luminosity calculated from the measured temperature and surface gravity , with resulting distances of 1.6@xmath11 kpc ( reindl et al .
2014 ) and @xmath121.8 kpc ( arkhipova et al .
the interstellar extinction has @xmath13 equal to [email protected] mag ( as based on the 2200 feature in the ultraviolet ) and 0.14 mag ( based on the observed balmer decrement ) ( parthasarathy et al .
the measured extinction has not changed significantly from 1980 to 2011 ( reindl et al . 2014 ) .
the deduced mass for the star is variously given as @xmath140.55 @xmath7 ( reindl et al .
2014 ) or 0.2 + 0.59 @xmath7 for the ionized mass plus the core mass ( bobrowsky 1994 ) .
the expanding pn shell was resolved in 1992 by the _ hubble space telescope _ ( _ hst _ ) to be bipolar shaped with an embedded ` equatorial ring ' , where the largest radius was 0.8 arc - seconds ( bobrowsky 1994 ; bobrowsky et al .
the inclination of the ring is 56@xmath0 . just inside this ring ,
a [email protected] mag star is a possible wide companion star , making the stingray a binary with a separation of @xmath152200 au .
the presence of the companion star has been suggested to have some effect on the formation of the bipolar shape , but the very wide separation makes it hard for any such effects to be substantial .
importantly , the visible size of the pn requires an ejection many hundreds or thousands of years ago .
given an angular size of 1.15 " , a distance of 1.6 kpc , and an expansion velocity of 8.4 km s@xmath8 from the [ oiii ] line width , we get a kinematic age of the main pn shell to be 1013@xmath16 years ( reindl et al .
so we have a clear picture that the pn shell was ejected about a millennium ago , but was only ionized in 1980 .
the stingray has reasonable coverage from around 1920 to the present with ground - based spectroscopic monitoring , from 1988 to 1996 with _
spectra , and from 1992 to 2000 with _
hst _ ultraviolet spectroscopy and high - resolution imaging . but
no one has reported any photometry past a few isolated magnitudes , all with large problems , spread over various magnitude systems .
we realized that a full light curve was needed , and we had means for measuring full light curves all the way back to 1889 , so as to cover over eight decades of time before the 1980 s ionization event , the transition time of the pn turn - on , and three decades of the coasting phase after the turn - on . nor has anyone looked for an outer shell far away from the central star ( such as are seen around many ordinary pne ) , so we wondered whether the stingray has any far outer shells .
this paper reports on our results of the light curve of the stingray from 1889 to 2015 , plus our searches for any outer halo .
to get broad - band magnitudes for the stingray , we have pulled from a wide variety of sources ; the harvard photographic plate collection from 1889 - 1989 , the visual magnitude estimates of albert jones as archived by the _ american association of variable star observers _ ( _ aavso _ ) from 1994 to 2007 , the _ all sky automated survey _ ( _ asas _ ) from 2001 to 2009 , the _ aavso _ telescopes going into the _
aavso photometric all - sky survey _
( _ apass _ ) from 2011 to 2015 , plus our own photometry from ccd images with decam on the cerro tololo 4-m blanco telescope from 2014 . in all
, we have 1026 magnitudes , mainly in b and v , with good coverage from 1889 to 2015 .
we have added 15 magnitudes from the literature or derived by us from the literature , all on 6 nights from 1969 to 1996 .
all 1041 magnitudes are presented in table 2 .
a critical realization is that the magnitudes _ before _ 1980 are all measures of the continuum of the star , while all magnitudes _ after _ roughly 1989 are essentially measures of the [ oiii ] emission line strengths .
the reason is that before this cutoff date there were no significant emission lines seen in the spectrum so the brightness is of the observed stellar continuum , whereas after that date we see the optical fluxes dominated by emission lines . so before 1980 , we are measuring the brightness of the star alone , and this depends only on the effective size and temperature of the photosphere minus any extinction by dust .
after 1989 , the emission lines dominate , with the brightness depending on both emission line fluxes and the exact spectral sensitivity of the detector .
the nebular light in the v - band is entirely dominated by just three emission lines ; [ oiii ] at 5007 as the brightest with 77% of the v - band detected flux , [ oiii ] 4959 with 21% of the flux , and h@xmath17 at 4861 as the faintest with only 2% of the flux .
so jones , _ asas _ , and _ apass _ are essentially reporting the brightness of the [ oiii ] emission lines . a further realization from this is that there will likely be systematic differences in the post-1980 s magnitudes from observer to observer . for example , the exact width of the spectral sensitivity curve for albert jones eyes is somewhat different from the _ asas _ ccd / telescope / atmosphere combination , so jones will detect somewhat more [ oiii ] light relative to what he detects for the nearby comparison stars as compared to the _ asas _ measures .
these systematic observer - to - observer offset are real , and apparently at the quarter - of - a - magnitude level .
another critical realization is that the emission line spectrum for v839 ara makes its measured brightness respond slightly differently to atmospheric extinction than do the brightnesses of the nearby comparison stars .
the reason is that v839 ara has all of its light on the blue - side of the v - band , while the ordinary comparison stars have a continuum that fairly evenly covers the entire v bandpass .
ordinary atmospheric extinction dims the blue side of the v bandpass more than the red side .
extinction from the ground changes on many timescales due to differing airmass ( as the stingray is looked at with varying zenith distances ) and with ordinary changes in the atmosphere aerosol content .
thus , if atmospheric extinction increases for any reason , then the blueish emission light of v839 ara will be dimmed somewhat more than the light from the normal comparison stars will be dimmed .
all of the ground - based photometry reported here is straight differential photometry with respect to an ensemble of nearby comparison stars , so v839 ara will appear to slightly dim in our derived v magnitude as the atmospheric extinction increases .
this effect can not be taken out with the usual color terms derived from the comparison stars for the ccd / filter / telescope combination , and indeed , an exact correction is not possible without more information on the atmosphere and spectral sensitivities than is available .
so we are left with a relatively small atmospheric source of variability superposed on the real underlying light curve . from the _ apass
_ time series , we see that the effect can get up to around 0.05 mag in amplitude . for all other ground - based telescopes , we can not recover from this atmospheric variation , so we just have to acknowledge that the measured magnitudes have a non - astrophysical variability of roughly 0.05 mag superposed .
this problem arises from the nearly - unique nature of the stingray s emission line spectrum .
importantly , this mechanism does _ not _ apply to the pre-1980s harvard magnitudes , because the star dominated with no significant emission lines .
fortunately , for the post-1980s light curve , fast variations are impossible ( because the visible light is coming from emission lines spread out over a region one light - month in size ) , so we can use our many magnitudes to get time - averaged variations with atmospheric effects well - averaged out .
the harvard college observatory has a collection of over 500,000 archival photographic plates that cover the entire sky from 1889 to 1953 , plus a smaller collection from around 1969 to 1989 .
( the years 1953 - 1969 are the notorious menzel gap . )
these plates are mostly in the b band , with limiting magnitudes typically ranging from 14 to 18 mag .
any one position will have typically from 1000 to 4000 plates for coverage .
the _ digital access to a sky century @ harvard _ program ( _ dasch _ ) is currently @xmath1510% through digitizing all the harvard plates , but it will reach the stingray ( in the constellation ara at low galactic latitude ) only some years from now . we have constructed a light curve from 108 harvard patrol plates , all in the b band .
the plates were all from the b series as well as on patrol plates of the rb , am , ax , and dsb series .
( we could have measured many more magnitudes for such a bright star as the stingray , but the patrol plates taken have the complete coverage in time and already well - define the star s variability , so adding more plates would provide little new information . ) the comparison stars were chosen from nearby stars of similar magnitude , with their b magnitudes taken from the _ apass _ survey of the _ aavso_. the harvard plates formed part of the original definition of the b magnitude system , and they have been measured many times to have a near - zero color term for transformations from their native system to the johnson b magnitude system . with this
, we can be very confident that our resultant magnitudes are in the johnson b magnitude system .
the visual comparison of targets to the comparison stars is a long well - developed practice , now largely lost amongst living astronomers .
nevertheless , extensive experimentation over the last thirty years has proven that the visual estimation of magnitudes has an accuracy essentially equal to that obtained from digital scans and from iris diaphragm photometers .
importantly , an experienced eye will produce a real uncertainty that is more than a factor from 1-times to 3-times better than that produced by the _ dasch _
photometric pipeline ( e.g. , schaefer 2014a , 2014b ) .
given our very long and deep experience at visual estimation , plus its great speed , simplicity , and low - cost , the visual estimation method is to be preferred . for an average case with a good sequence and a target well above the plate limit ( as for the case of the stingray ) ,
the real one - sigma error bar is close to @xmath100.10 mag .
the harvard plates light curve from 1889 to 1989 is presented in table 2 and figure 1 .
the behavior of the stingray was startlingly unexpected . from 1889 to 1980 ,
the star slowly faded , with significant modulation .
the fading is consistent with a linear decline in magnitude , going from b=10.30 in 1889 to b=10.76 in 1980 , for a decline at a rate of 0.0051 mag / year .
superposed on this linear decline are apparent fluctuations on the decadal timescale with a total amplitude of near half a magnitude . then , starting suddenly around 1980 , the stingray central star faded fast at least until 1989 .
the rate of decline was approximately 0.20 mag / year . from early 1994 until the middle of 2007 ,
albert jones has reported 128 visual measures of the stingray .
these are all visual measures , through his 12.5-inch reflector telescope in nelson new zealand , made using a sequence of comparison stars that is accurate on the modern magnitude scale . thus , his measures are closely on the johnson v - magnitude system .
jones magnitudes were reported through the variable stars section of the _ royal astronomical society of new zealand _ , and are now available in the _ aavso _ database .
albert jones has been the world s best observer of variable stars for much of the last sixty years .
( he died in 2013 . )
some of his exploits include the discovery of the highly - important sn 1987a , the discovery of two recurrent novae eruptions ( t pyx in 1966 and v3890 sgr in 1990 ) , and the discovery of two comets . in the early 1990 s ,
janet mattei ( then _ aavso _ director ) pointed him out as being the all - time best variable star observer with incredibly accurate eyes , while daniel w. e. green ( now the director of the _ central bureau for astronomical telegrams _ ) pointed him out as being the world s best observer of comet magnitudes . and
he not only has an incredible accuracy , but he also has made over 500,000 variable star magnitude measures , nearly a factor of two times more observations than any other visual observer .
the authors of this paper have used jones observations in many prior studies for a wide range of stars , and we have always been deeply impressed by both his quality and quantity of observations , as well as by his characteristic of always looking at the fun , exciting , and useful variables .
results from our prior work , and from others , give jones as having a one - sigma photometric accuracy of 0.05 to 0.10 mag .
the reason for this recital is to point to jones reliability and accuracy .
the rms scatter of the differences between successive magnitudes in the light curve is 0.21 mag for pairs within 10 days .
jones is only measuring the [ oiii ] emission lines , and these can not vary substantially on such fast time scales .
( this is because the [ oiii ] light comes from all around the planetary nebula , with near - zero from the central star , and the nebula is about a light - month in radius . ) as 0.21 mag is the difference between two magnitudes , the one - sigma uncertainty for measuring one magnitude will be 0.15 mag .
this variation is significantly larger than jones photometric accuracy .
so we take the excess variance to be due to the effects of variations in the atmospheric extinction , where the emission lines from v839 ara are dimmed more than the continuum light from the comparison stars .
thus , the extinction effect is somewhat smaller than 0.15 mag in size .
jones visual light curve is plotted in figure 2 , with circles representing his measures .
his light curves shows a steady decline with superposed apparently - random fluctuations .
his observations are reasonably fit with a linear decline , from v=10.71 in 1994 to v=11.77 in 2007 , for a decline rate of 0.081 mag / year .
the rms scatter around this linear decline is 0.22 mag .
_ asas _ has been running an all - sky survey , measuring v magnitudes for approximately 10 million stars on a nightly basis ( pojmanski 2002 ) .
the telescopes all have lenses of aperture 200 millimeters and f - ratio of 2.8 , forming an 8.5@[email protected]@xmath0 field of view , while the ccd has a pixel size of 15 " .
with 180 second integrations , the limiting magnitude is roughly v=14 . for observations in v of the stingray ,
the telescope is at las campanas observatory in chile .
the publicly available light curve is from early 2001 to late 2009 , and we will have 425 v magnitudes . for the reported _ asas _ v magnitudes , we selected out only the grade a and b values .
these magnitudes show a slow secular decline from 2001 to 2009 , plus a large scatter superposed on top .
uncrowded neighboring stars of similar brightness show a good flat light curve with an rms scatter of 0.06 mag .
the publicly reported magnitudes are divided into ten blocks , where each block was made with a different field center and slightly different positions for the centers of the photometry apertures .
each field center produces a scattered light curve that parallels the light curves from the other centers , yet the field centers have offsets that vary by 1.6 mag ( with an rms scatter in the intercepts of 0.57 mag ) .
the cause for this variations is that a v=10.73 star is 35 `` from the stingray , so small changes in the center of the photometry aperture will make for a varying contribution from the nearby star . to pull out this effect ,
we have constructed a model with a presumed gaussian point spread function ( psf ) for the _ asas _ star profiles .
for the known aperture radii , the known star positions , and the known centers for the photometry aperture , this model has only one fit parameter , the gaussian width of the psf .
the _ asas _
literature gives typical values from 10 '' to 15 `` ( e.g. , pojmanski 2002 ) .
we have made a chi - square fit for the reported brightness for each of the ten field centers and the five aperture diameters ( 30 '' , 45 `` , 60 '' , 75 `` , and 90 '' ) .
our best fit is with a gaussian sigma value of 14 " . with this value plus our model ,
we have derived the v magnitude for the stingray .
the resultant light curve still shows substantial scatter . for pairs of magnitudes taken within 10 days of each other ( so that the nebular [ oiii ] light can not vary significantly ) , the rms scatter of the magnitude differences is 0.31 mag .
the one - sigma measurement uncertainty for one magnitude is 0.22 mag .
this large scatter is much larger than dictated by photon statistics .
the cause for this scatter can be due to the atmospheric extinction variations ( as discussed previously ) as well as the circumstance of having imperfect correction for a relatively bright star just near the edge of the photometry aperture .
small variations in the psf width and the center of the photometry aperture can make for substantial changes in the derived brightness for the stingray , and we have no way to recover this from the publicly available data . the _ asas _ light curve from 2001 to 2009 ( see figure 2 ) shows a steady decline , with this being highly significant despite the substantial measurement errors .
this decline is consistent with a linear change in the magnitude , going from 11.41 in 2001.0 to 12.53 in 2010.0 , at a rate of 0.124 mag / year .
the rms scatter around this trend line is 0.23 mag .
figure 2 shows us that the asas light curve has a similar slope from 2001 - 2009 as does jones from 1994 - 2007 , but that there is an offset by 0.2 to 0.4 mag in the overlap time interval .
it is possible that our model of the stingray produced a systematic underestimate of its contribution to the smallest photometry aperture .
however , we think that it is more likely due to jones and asas having different spectral sensitivities .
that is , the photochemicals in jones eye will produce a distinctly different relative sensitivity for the [ oiii ] lines ( when compared to the mean sensitivity for the continuum from the comparison stars ) than will the asas ccd - plus - filter . in this case , the observed offset is just the result of normal color terms between two different detectors , with this being exaggerated for an emission line source . in all cases , the photometric uncertainties are below the quarter - magnitude level , and we can easily see the overall behavior of the stingray .
_ apass _ is measuring b , v , g , r , and i magnitudes of all stars in the sky with approximately @xmath19 , with several telescopes in both the northern and southern hemisphere .
this has provided us with a reliable source of comparison stars for our differential photometry .
this has also provided us with magnitudes in the five filters on two nights in 2011 ( see table 2 ) . at our request ,
a. henden has put the stingray in the queue for time series photometry on the 0.61-meter optical craftsmen telescope at the mount john observatory in new zealand .
the 1-minute ccd integrations were through a johnson v filter on the nights of 23 , 26 , and 27 march 2015 .
the quoted magnitudes were based on differential aperture photometry with respect to 5 - 8 nearby comparison stars previously calibrated with _
apass_. on the first night , as the stingray rose from an altitude of 31@xmath0 to 54@xmath0 , the apparent brightness of the stingray suffered a dip in brightness by about 0.05 mag for one hour .
the stingray was constant , with an rms scatter of 0.008 mag , on the other two nights . during the time of the dip ,
the statistical error for the star increased from 0.012 mag up to 0.06 mag , indicating that some large atmospheric extinction had dimmed the target by much more than 0.05 mag . for the time through the dip and for all the nights , we found a tight correlation between the calculated statistical error and the differential magnitude of the target .
this has the easy interpretation that ordinary changes in the atmospheric aerosols substantially dimmed both the stingray and its comparison stars ( making for large statistical errors ) , while the stingray s [ oiii ] emission lines ( all on the blue side of the v bandpass ) were dimmed more than the continuum light from the comparison stars ( making for the small dip in the differential magnitude ) .
this effect is undoubtedly happening on all time scales to all ground - based observers , but we can see this effect only in a time series from one telescope . with this correlation , we can reduce all the _ apass _ magnitudes to those of some constant condition with minimal aerosols .
for the measures with statistical error greater than 0.020 mag , this correction could not be done with high accuracy , so we have chosen to delete these data .
the result is a time series on three nights with 352 v - band magnitudes ( see table 2 ) , with nightly averages of 12.62 , 12.57 , and 12.61 .
the _ apass _ magnitudes extends the v - band light curve past 2009 up until march 2015 ( see figure 2 ) . from the _ apass _
magnitudes alone , we see a significant decline from 2011 to 2015 .
this is just a smooth extension of the earlier decline .
the various observers are not expected to report the same v magnitudes to within a quarter of a magnitude or so , due to the emission line nature of the stingray . nevertheless , all observers show a consistent and highly significant steady fading from 1994 - 2015 .
this fading goes from v=10.71 in 1994.1 to v=12.61 in 2015.3 , for an average decline rate of 0.090 mag / year . with the cerro tololo 4-meter blanco telescope
, decam ccd images were taking on 30 june 2014 , the main goal being to search for any outlying faint optical shell . the large size ( 9x17 ) of one individual chip
allows for v839 ara , comparison stars , and for any far out shell to be contained on a single ccd ( out of the 64 ccd array ) .
observations were made using the the u , g , r , i ( which are the sloan digital sky survey , sdss , filters ) with pairs of exposure time of 20 and 90 seconds in each band .
standard procedures were used in the iraf data reduction package to extract magnitudes for v839 ara utilizing the phot package .
differential photometry was performed on v839 ara using multiple _ apass _ comparison stars , carefully avoiding saturated stars , and only with the short exposures .
these comparison stars were chosen such that they fell on the same chip as v839 ara , have similar brightness , and not crowded by surrounding stars . since _
apass _ only provides information for g , r , & i filters , a magnitude for the u observations could not be derived .
we measured the magnitudes of v839 ara to be [email protected] , [email protected] , & [email protected] for g , r , & i respectively .
the formal statistical error bars are smaller than 0.01 mag , while various systematic sources of error are likely around the 0.05 mag level .
hill et al .
( 1974 ) and kozok ( 1985a ) report on ubv photometry on four separate nights .
these magnitudes are of the central star , with any contributions from emission lines and the pn being negligibly small .
these four b magnitudes closely fit with the light curve from the harvard plates .
bobrowsky et al .
( 1998 ) used _ hst _ to resolve the central star and report v=15.4 as deduced from the flux measured in a continuum filter centered at 6193 .
this magnitude is of the central star alone ( with no emission lines and no pn shell ) was made in march 1996 . around this time
, albert jones was reporting the stingray to be v=10.8 mag .
the difference of 4.6 mag is because jones is including the shell light , almost entirely [ oiii ] emission lines .
this shows that the two [ oiii ] lines are contributing @xmath1570@xmath18 as much flux as the star over the entire broad v - band .
reindl et al .
( 2014 ) report on an _ hst _ spectrum from march 1996 with the _
fos_. we have taken the continuum flux for the center of the b- and v - bands , and converted these into magnitudes .
we get b=14.64 and v=14.96 , with this applying to the star alone .
the v magnitude is 0.44 brighter with the _ fos _ spectrum than with bobrowsky s continuum flux , despite there being only two days of separation in time .
it is unclear if the difference arises from variability or from the ordinary uncertainties in extracting magnitudes with two different non - standard methods .
other published magnitudes ( like the estimates from the _ cape durchmusterungen _ and _ hd _ catalogs ) have problems with their comparison stars that can be as large as one magnitude , and hence can not be used with any useful accuracy or confidence .
the presentation of the _ asas _ magnitudes by arkhipova et al .
( 2013 ) has not realized any of the problems or solutions caused by the varying centers and the nearby star , so their light curve is now to be replaced by our light curve described in section 2.3 . from the 1992 ph.d .
thesis of p. garcia - lario ( as quoted in parthasarathy et al .
1993 ) the j , h , and k magnitudes were 11.37 , 11.97 , and 11.38 .
these magnitudes presumably date from around 1991 .
like all the brightnesses in this subsection , these refer to the entire nebula , with near - zero contribution from the central star .
2mass _ magnitudes were [email protected] , [email protected] , [email protected] ( cutri et al .
the magnitudes are from may 2000 .
the _ iras _ fluxes were 0.65 jy at 12@xmath20 , 15.59 jy at 25@xmath20 , 8.05 jy at 60@xmath20 and 3.39 jy at 100@xmath20 ( parthasarathy & pottasch 1989 ) .
these observations were made from january 1983 to november 1983 .
this spectral energy distribution has the obvious interpretation as thermal emission from dust in the shell , with an average temperature of 125 k ( parthasarathy & pottasch 1989 ) .
the _ akari _ satellite all - sky survey gives the 9@xmath20 flux to be [email protected] jy and the 18@xmath20 flux to be [email protected] jy ( ishihara et al .
these brightnesses come from a survey running from may 2006 until august 2007
. the _ wise _ all - sky survey provides magnitudes ( on the vega magnitude system ) for the various wise bands to be [email protected] , [email protected] , [email protected] , [email protected] for the w1 ( 3.4@xmath20 ) , w2 ( 4.6@xmath20 ) , w3 ( 12@xmath20 ) , and w4 ( 22@xmath20 ) bands respectively .
the _ wise _ survey ran from january 2010 to january 2011 .
the _ atca _ radio light curve , from 1991 to 2002 , has the stingray declining steadily in flux density from [email protected] mjy ( parthasarathy et al .
1993 ) to [email protected] mjy at 4800 mhz ( umana et al .
roughly 50% of classical pn have halos , faint roughly - circular shells far outside the the bright well - known nebulosity ( chu et al .
for example , ngc 6720 ( the famous ring nebula ) has the obvious shell ( with outer dimensions 90``@xmath1865 '' ) surrounded by a faint shell extending 162``@xmath18147 '' filled with fairly uniform mottling and arcs , with this outer halo being bright in _ wise _ images .
corradi et al .
( 2003 ) and chu et al .
( 1987 ) have catalogs with pictures and intensity profiles .
the typical surface brightness is a thousandth of that of the inner classical pn .
the estimated ages for these outer shells are many tens of thousands of years as based on expansion velocities that default to 20 km s@xmath8 when not otherwise known .
corradi et al .
( 2003 ) call these faint shells structures outside the classic pn as ` agb halos ' , where the ordinary stellar wind of the star during its agb phase has become ionized .
they claim that the outer edge of the agb halo results from the last thermal pulse in the agb star .
it is possible that the stingray might have an agb halo outside its small classical pn shell
. such a shell might be too large to be visible in the _ hst _ images .
so we thought it worthwhile to search for any agb halo around the stingray .
we examined deep images from decam , shassa , 2mass , and _ wise _ , as reported below .
we have also examined the various available images from the digital sky survey and from _ hst _ , but our negative results for any shell structure or circular arcs centered on the stingray do not provide useful constraints .
no images were taken with _
galex _ or _
swift_. our decam images in u , g , r , and i were examined for any nebulosity surrounding the stingray . in our images ,
the classical planetary nebula is entirely in the near - saturated inner core of the normal stellar image .
so we are only sensitive to any outlying outer shell .
our examination was visual , mainly because there is no way to automate or quantify a shell search when we have no idea of the size or shape .
we have very long experience at searching for shells or light echoes around stars , and we know that the human eye / brain combination is very sensitive to detecting shells . indeed , for irregular shells , visual examination is greatly better at detecting any nebulosity by pushing down towards the background noise limit . with this , we detect no shell or nebulosity in any of our decam images .
the r image would record any h@xmath21 emission , while the g image would record any [ oiii ] emission .
the southern h - alpha sky survey atlas ( shassa ) is a robotic wide - angle ccd survey in the h@xmath21 line covering declinations south of + 15@xmath0 ( gaustad et al .
images have pixels that are 47.64 arc - seconds on a side , with a sensitivity down to about 0.5 rayleigh .
the stingray appears prominently about 4@xmath0 from the center of their field 38 .
the point - spread - function has a fwhm of around 1 arc - minute , so all the known shell appears as a point source , and any detectable shell would have to be larger than several arc - minutes in radius .
the continuum - subtracted and smoothed image shows no shells , circular arcs centered near the stingray , or any structure associated with the stingray out to a radius of over 1 degree .
there are also 2mass observations of v839 ara in j , h , and k bands taken in may of 2000 . the radius of emission in each of these three bands are less then 3 " .
no outer shell was seen in j , h , or k. the archival _ wise _ images covers v839 ara in four different bands centered on 3.3@xmath20@xmath22 ( w1 ) , 4.6@xmath20@xmath22 ( w2 ) , 12.1@xmath20@xmath22 ( w3 ) , and 22.2@xmath20@xmath22 ( w4 ) .
v839 ara is positioned nearby a star ( identified as tyc 8739 - 1088 - 1 with v=10.83 ) which is bright in both the w1 and w2 bands , however , it is noted that pne are expected to show strong emissions in the far infrared . in the w1 band , the emission is centered on the location of v839 ara with a radius of 11 `` .
going further out into the infrared , the emission centered on v839 ara extends out to a radius of 14 '' in w2 , 34 `` in w4 , and a finally a very large ` halo ' with a radius of 81 '' in w4 .
it should be noted that this ` halo ' seen out in 22@xmath20@xmath22 is consistent with psf rings around sources bright in the w4 band , and is not astrophysical .
this can be seen in figure 3 .
from 1889 to 1980 , the stingray s central star exhibits a linear decline in its light curve , with substantial variability superposed .
the best fit line goes from 10.30 mag in 1889 to 10.76 mag in 1980 , for a total drop of 0.46 mag .
this is the same as the apparent amplitude for the decade - long variations . over this time , with no substantial emission lines providing any significant fraction of the flux , the variability can only be caused by changes in the photosphere , either by variations in the effective temperature or radius .
( changes in the dust column seem unlikely given that the stellar wind must be weak , while even large winds in the 1980 s led to no change in the extinction . ) from _ c._1920 to 1980 , the spectral type remained largely unchanged , with everyone reporting spectral types from b0 to b3 , so the star s surface temperature was not changing by any large factor . for the b - band flux to change by 0.46 mag
, the radius would have to change by 20% .
this is a large change .
how can the stingray star have ` anticipated ' the 1980s ionization event ? apparently , some prelude to the sudden event had a surface manifestation for at least a century in advance .
the anticipation cooling and/or shrinking of the central star could not have been going on for many centuries . at its rate of fading by near half - a - magnitude per century , a millennium duration would imply that the central star was five magnitudes more luminous at the start of the anticipation .
such would imply a luminosity higher than supergiants .
so this anticipation fading can only have been going on from one century to a few centuries .
detailed calculations of the variations of the central star luminosity can have substantial changes in the century preceding thermal pulses ( blocker 1995 ; schnberner 1983 ) .
for a 0.553 m@xmath23 star , before a thermal pulse at 45,000 k , the luminosity will drop by a factor of 2.2 in the preceding century ( schnberner 1983 ) . for a 0.836 m@xmath23 star ,
the luminosity will drop by a factor of 6 in the century preceding a a thermal pulse at 200,000 k ( blocker 1995 ) .
handler ( 2003 ) has defined a new class of variable stars , called the ` zz leporis ' stars , which consist of the central stars of young pn with temperatures @xmath1450,000 k. the physical mechanism for these brightness changes is not known , but it is likely some combination of stellar pulsations and fluctuations in a stellar wind .
arkhipova et al .
( 2013 ) have already identified the central star of the stingray as being in this zz lep class . however , although the variations are fast and aperiodic in both cases , the zz lep stars have greatly different light curve properties than what we see for the stingray .
in particular , zz lep stars have time scales of 4 - 10 hours and amplitudes @xmath140.03 mag ( handler 2003 ; handler et al .
this is greatly different from the stingray s behavior in 1889 - 1980 , which has a time scale of around one decade and amplitude 0.5 mag . from 1994 - 2009
, our v - band light curve is not coming from the central star , while the central star may or may not have zz lep behavior . the fast decline for the light from the central star alone from 1980 ( b=10.8 ) to 1988.5 ( b=12.5 ) to 1996.3 ( b=14.64 ) is stark and unprecedented
. this can not be due to changing extinction , because the dimming , as measured by many methods , is essentially unchanged from 1980 to 2011 ( reindl et al .
so this fast fading star can only be due to some combination of the decrease of the stellar radius and a decrease in the temperature .
for the case of the stingray , reindl et al .
( 2014 ) show that the stellar temperature increased greatly while the star s radius decreased greatly from 1988 to 2006 .
the radius and temperature changes run against each other .
so it takes a detailed calculation as to whether these observed changes translate into the observed magnitude changes .
if we go by the usual luminosity equation , we have @xmath24 , where @xmath25 is the luminosity , @xmath26 is the stellar radius , and @xmath27 is the effective temperature of the photosphere .
we do not have direct measures of the radius , but the star s surface gravity , @xmath28 , will scale as @xmath29 . and
we should not be using the luminosity , but rather the blackbody flux in the b - band , @xmath30 , at @xmath31=4400 , with this being the usual planck function of @xmath31 and @xmath27 .
we can then get a b magnitude as @xmath32 , for some zero magnitude @xmath33 .
reindl et al .
( 2014 ) list measured values for @xmath34(g ) and @xmath27 for many years from 1988 to 2006 .
taking @xmath35 so that b=12.5 in 1988 , we get b=13.1 mag in 1996 , and b=14.9 in 2006 .
we see that in this case for a shrinking star that is getting hotter , it is the shrinking that is the dominant effect , making for the star dropping in b - band brightness .
however , the effect predicted from the measured @xmath34(g ) and @xmath27 is a much slower decline that is actually observed . that is , from 1988 to 1996
, the prediction is that the central star will dim by 0.6 mag , while the observed dimming is by 2.14 mag .
this is a large difference .
this points to the central star having additional light above the photosphere in 1988 that went away by 1996 , or to changing systematic errors in measures of the surface gravity ( perhaps associated with the changing stellar wind rate ) . from this analysis ,
we take the basic cause for the fading of the central star to be due to the shrinking of the stellar radius , although detailed calculations do not predict the central star to be fading as much as is observed .
the v - band light curve ( essentially the nebula s brightness ) from 1994 to 2015 has been steadily fading at the rate of 0.090 mag / year .
this is a highly significant result from three independent sources and has no prospect of being due to any artifact .
so we have the [ oiii ] emission line flux fading with a half - life of around 8 years . from 1994 to 2015 , the stingray emission lines have faded by near 2.0 mag , a factor of near 6@xmath18 .
this is the nebula fading , not the central star .
the fading of the nebula is also seen for wavelengths from the near infrared , the middle - infrared , and the radio : ( 1 ) from around 1991 to 2000.4 , the stingray faded by 0.73 , 0.28 , and 0.13 mag for the j , h , and k bands respectively .
these correspond to half - lives of 9.3 , 24 , and 54 years respectively .
( 2 ) we can use the _ iras _ , _ akari _ , and _ wise _ mid - ir fluxes to chart the changes in the 12@xmath20 flux . for this , the _ wise _ magnitude must be converted to jansky , and the _ akari _ flux must be interpolated to 12@xmath20 . with this
, we have 12@xmath20 fluxes of 0.65 jy in 1983.5 , 0.36 jy in 2007.0 , and 0.226 jy in 2010.6 .
this shows a steady decline over 27.1 years with a half - life of 20.4 years .
( 3 ) similarly , for 22@xmath20 , we have fluxes of 8.96 jy in 1983.5 and 3.20 jy in 2010.6 .
the extrapolation of the _ akari _ 18@xmath20 flux to 22@xmath20 is just where the spectral energy distribution is turning over , so the uncertainty is too large for this to be useful .
this shows a decline in flux with a half - life of 18 years .
( 4 ) from 1991.3 to 2002.7 , the atca radio flux at 4800 mhz ( 6 cm ) declined from 63.6 mjy to 48.8 mjy .
this corresponds to a half - life of 30 years .
we see a consistent picture for the light curves from 1994 - 2015 , where the optical , near infrared , middle - infrared , and radio fluxes have all been smoothly fading with a characteristic time scale of a decade or so .
the nebula had a sudden turn - on sometime between 1979.49 and 1988 , and the brightening of the nebula likely stopped in the early 1990 s with the turn - off of the fast stellar wind associated with the 1980s ionization event .
the fading of the nebula started after the end of the ionizing event and its fast stellar wind , in the early 1990s . at this time , the central star underwent fast fading ( see figure 1 ) , so its illumination of the nebula started to decline fast .
it is easy to ascribe the fading of the nebular light to the sharp fall off in its illumination by the central star .
even with the rise of surface temperature of the star , the fall in its surface area means that it is giving off much less ionizing radiation . with the turnoff of the illuminating source ,
the nebula light should start fading at some sort of a re - ionization time scale .
there is precedent for this fading and this interpretation for sakurai s object ( v4334 sgr , a very - late - thermal - pulse born - again star ) .
its emission line fluxes have also been seen to decline fairly rapidly ( with a half - life just under two years ) from 2001 to 2006 , with the fading attributed to cooling and recombination in the shell after the heating and ionization ended ( van hoof et al .
2007 ) . for the usual nebula conditions , the recombination time scale is around 80,000 years divided by the electron number density ( in units of @xmath36 ) .
( the nebula will have a range of densities , each recombining and fading on their own time scale , with continual - but - fading illumination from the central star , so the connection between the average electron density and the effective recombination times and the fading rate has substantial uncertainty and change over time . ) for the stingray , parthasarathy ( 2000 ) gives an electron density of 10,000 @xmath36 , so we expect a recombination time scale of around 8 years . the agreement between the half - life for the fading of the optical emission lines and the recombination time scale is an indication that the slow steady fading of the nebular light from 1994 - 2015 is caused by the ordinary recombination inside a nebula that had some sort of an ionization event between 1980 and 1994 .
the middle infrared emission is dominated by thermal light from warm dust in the nebula .
the dust grains are not likely to be changing in number or size , so the thermal flux can only be fading due to a cooling of the dust grains .
this is all consistent with a scenario where the dust was heated during the 1980s ionization event and has been cooling ever since .
the dust reaches thermal equilibrium on a fast time scale , so the fading of the middle infrared emission is a measure of the fading of the heating radiation field .
we suggest that the dust heating is dominated by the nebular emission lines , with this being _ in situ _ and the connection is made because the nebular lines are fading at the same rate that the dust is fading .
the radio emission is free - free light , with model derived electron densities of 1.23 - 2.5 @xmath18 10@xmath37 @xmath36 ( parthasarathy et al . 1993 ; umana et al .
2008 ) . like for the optical line emission , with a scenario of a short - duration ionization event in the 1980s , the radio flux should fade on some sort of a recombination time scale .
with the results from this paper , we can construct a detailed evolutionary path of the stingray through the hr diagram . for seven dates ,
we have calculated @xmath25 and @xmath27 for the central star ( see table 3 ) .
these can be directly plotted onto the hr diagram and compared to model predictions ( see figure 4 ) .
the first point , labeled for the year 1001 ad , is meant to illustrate the approximate position where the star ejected the pn shell , just as it was leaving the agb phase .
this is useful to show that the stingray took only around one millennium to cross from the agb to the thermal pulse region .
the two points for 1889 and 1980 have a modest uncertainty from the distance , so they can be moved up or down in the diagram , but they must be moved up or down together .
similarly , the last four points can be moved up and down together by modest amounts as the assumed stellar mass changes from the adopted 0.55 @xmath7 .
the end result is an observed evolutionary path from when the stingray leaves the agb until sometime during a thermal pulse .
the default or common idea is that the stingray central star has evolved off the agb with the usual nearly horizontal path across the top of the hr diagram ( e.g. , parthasarathy et al .
1993 ) , and is now undergoing a late thermal pulse ( e.g. , reindl et al . 2014 ) .
( see the next section for possible alternatives . )
late thermal pulses are when a shell of helium , just outside the carbon - oxygen core , is ignited , with the energy from this burning making the star temporarily increase back to giant size , with the result that the star forms a loop in the hr diagram .
detailed evolutionary paths through the hr diagram depend on the stellar mass , while the thermal pulses can occur anywhere along the path ( e.g. , schnberner 1983 ; blocker 1995 ; blocker 2001 ; schnberner 2008 ) .
for illustrative purposes , one such late thermal pulse evolutionary track ( from schnberner 1983 ) has been superposed on the stingray s observed evolutionary track in figure 4 .
this track was chosen because it is for a stellar mass that might be similar to that of the stingray , because the thermal pulse is at a similar temperature as for the fast evolution of the stingray , and because the track has conveniently labelled time tick marks to allow direct comparison .
the comparison between the observed and theoretical evolutionary paths in figure 4 shows fundamental problems : ( 1 ) the time scale for the evolution from a temperature of order 5000 k ( when the pn shell was ejected ) to a temperature of 50,000 k ( in 1996 ) is observed to be around one millennium , while theory dictates a time scale more like 13 millennia to cross from roughly 5000 k to 50,000 k. this discrepancy can be resolved if the star is greatly more massive than 0.55 @xmath7 , but this then raises other difficult discrepancies .
( 2 ) the harvard plates show the stingray was fading from 1889 to 1980 , while the spectra from 1920 to 1979.49 show a nearly - unchanging spectral type , so we have a distinct vertical segment on the hr diagram lasting about a century . from the theoretical paths ,
this could only match the initial turn in to a thermal pulse , like from the tick marks 13 " to 13.1 " in the theoretical path shown in figure 4 .
but if this match be made , then the further observed path to 1988 - 2006 is mystifying .
( 3 ) the harvard plates show a sharp drop in the central star s luminosity from 1980 to 1988 , all with the temperature not changing greatly , so we must have another connected nearly vertical segment which lasts only 8 years .
this segment does not match any theoretical segment , in particular because its 8 year duration ( 0.008 millennia for comparison with the theoretical tick marks ) is many orders - of - magnitude faster than any expected evolution , even for a massive star .
( 4 ) the 1988 - 1996 - 2002 - 2006 segment shows significant and complex evolution . in under two decades
, the central star has moved in the hr diagram by more than theoretical paths allow for within two centuries at its fastest . and
the total change in luminosity from 1889 to 2006 ( around 1.65 in log - units ) is greatly larger than allowed in theoretical models .
we can point to two major discrepancies between measures of fundamental parameters for the stingray : first , the mass of the star has contradictory evidence pointing to 0.55 m@xmath23 ( parthasarathy et al .
1995 ) , 0.59 m@xmath23 ( bobrowsky 1994 ) , 0.354@xmath38 m@xmath23 ( reindl et al . 2014 ) , or 0.87 m@xmath23 with an initial mass up to 6 m@xmath23 ( reindl et al .
second , the variations in the measured temperature and surface gravity for the central star _ qualitatively _ reproduce the sharp decline in brightness from 1980 to 2006 , but _
quantitatively _ , the predicted decline from 1988 to 1996 is 0.6 mag , in stark contrast to the observed 2.14 mag decline ( section 2.6 ) .
a fundamental problem posed by the stingray is the nature of the 1980s ionization event .
we have seen no paper that addresses the question of what is really going on during this event , likely because answers are not known to the fundamental questions .
( 1 ) why should the ionization event start so suddenly within a time scale of a few years ?
the evolution time scale for the increase in ionizing flux resulting from a simple transit across the top of the hr diagram is many millennia , while the evolution time scale for a thermal flash is centuries and longer .
( 2 ) what is the physical mechanism that makes for the high luminosity of ionizing flux ? presumably the pre - existing shell was ionized by ultraviolet or far - ultraviolet radiation , and such might come from the central star getting hotter . but
the surface temperature of the star only changed from around 30,000 k ( from 1920 - 1979.49 ) to 38,000 k ( in 1988 ) , and we need a detailed calculation to see whether such a modest temperature change can make the nebula evolve from only weak balmer emission lines to domination of the spectrum by very bright high - ionization emission lines . in any case , the simple use of the central star to provide the ionizing flux does not work because it heated to 60,000 k in the year 2002 , but the shell was already fading due to recombination .
( 3 ) why is the ionization event concurrent with the fast and heavy stellar wind as seen from 1988 to the early 1990s ?
the time coincidence between wind and ionization strongly implies a causal connection .
but did the wind make the ionizing radiation , or did the ionizing radiation drive the wind ? ( 4 ) why did the ionization event not produce any temporary brightening in the b - band light curve ?
a mechanism that suddenly produces a large luminosity of ultraviolet light will almost - certainly produce blue light , but there is no flare in the harvard light curve with good resolution throughout the entire time period when the flare should occur .
( 5 ) why should the duration of the ionizing event be less than one decade and why should the turn - off time be only a few years ?
all the evolution time scales are much longer than a decade .
a higher fundamental problem of the stingray is the overall nature of its evolution .
the general answer is that the star must be somehow traversing the hr diagram to go from the agb phase to the upper left where it will soon enough end up moving down the white dwarf cooling track .
but the simple right - to - left traverse of the hr diagram ( as depicted in textbooks ) can not explain the fast and complex evolution of the stingray s central star ( see figure 4 and table 3 ) .
so either some additional mechanism is superposing some complex evolution , or the stingray is not the post - agb star that it seems . here
, we will briefly discuss three possibilities : the first possibility to the nature of the stingray s evolution is that it is an ordinary post - agb star undergoing some sort of a thermal pulse , wherein a layer near the surface of the star suddenly ignites nuclear burning , and this influx of energy will puff up the star s outer envelope to giant portions .
thermal pulses will make a star appear to go through loops in the hr diagram .
perhaps the stingray is caught just at the time of a thermal pulse , already having completed part of the loop ?
thermal pulses come in two types , called ` late thermal pulses ' ( ltp ) and ` very late thermal pulses ' ( vltp ) .
the later starts only after hydrogen burning has become extinct , and a critical feature of distinction is that the stellar surface is virtually free of hydrogen ( schnberner 2008 ) .
the known vltp examples are v605 aql , sakurai s object ( v4334 sgr ) , and perhaps fg sge ( lawlor & macdonald 2003 ; schnberner 2008 ) .
but the stingray certainly can not be a vltp star because it has only evolved off the agb by roughly a millennium ( when it ejected the pn shell ) , and because the central star has high abundance of hydrogen ( as shown by the prominent balmer absorption lines visible before 1980 ) .
so for this solution , the stingray would have to be an ltp star .
the ltp can make the star trace out a wide variety of loops through the hr diagram ( schnberner 1983 ; blocker 1995 ; blocker & schnberner 1997 ; van winckel 2003 ) . in an extreme case ( see figure 2 of blocker & schnberner 1997 ) , the star s track has nine reversals of direction in the hr diagram . for all these published loops , none match the observed track of the stingray , as detailed in the previous section .
none of the published ltp tracks can account for the stingray , and this is the fundamental problem .
part of this fundamental mismatch between observation and model is that the stingray is moving in the hr diagram at much faster rates than is allowed by theory .
gesicki et al .
( 2014 ) has found similar problems in accounting for the white dwarf mass distribution , and they have had to postulate an acceleration by a factor of 3 in the evolutionary time scales of blocker ( 1995 ) . in his major review on post - agb stars , van winckel ( 2003 ) says `` the detailed description of the badly known external post - agb mass loss on top of the mass loss from the nucleosynthetic consumption is crucial for the transition time estimates '' for the star crossing the hr diagram in various phases .
it might yet be possible for some new model calculation to match the stingray s evolution , but until then , the ltp idea has a fundamental problem .
the second possibility is that the stingray is a product of a common envelope ejection of the outer envelope of a giant star .
that is , in ordinary binary evolution involving a red giant star , a binary can get entangled inside a common envelope , the stars will spiral close together , and the main sequence star can eject the outer envelope of the red giant , leaving a bare stellar core .
the ejected envelope will expand , and the hot red giant core will ionize the shell .
reindl et al .
( 2014 ) suggested this post - common - envelope ( post - ce ) state for the stingray , with the evidence being that the apparent luminosity was consistent with the models of hall et al .
hall et al .
make a variety of predictions that can be used to distinguish the post - ce systems .
( 1 ) the luminosity of the post - ce stars will always be less than about 3.16@xmath1810@xmath39 l@xmath23 while traversing across the hr diagram from right to left .
in sharp contrast , the stingray had a luminosity of 5.4@xmath1810@xmath39 l@xmath23 in 1889 .
it is only after the central star has faded by an order - of - magnitude does the observed luminosity come into agreement with the post - ce tracks .
it is possible to overcome this argument by merely having the stingray at a closer distance , with @xmath141.2 kpc being adequate and within the uncertainty for the distance .
( 2 ) the post - ce central star would be a close binary with an orbital period from a fraction of a day up to weeks .
we know of no observational tests for this prediction .
( 3 ) the evolution through the post - ce phase is greatly slower than for an ltp . hall et al .
calculate that the time scale from the ejection of the shell until the central star is 30,000 k is around 10,000 years , while the time scale for the small turn in the motion along the hr diagram ( with the central star hotter than 30,000 k ) is 0.1 to 10 million years .
this model prediction is greatly against the observed time scales of one millennium from ejection to 30,000 k and decades for the stingray to turn around in the hr diagram .
the stark failure of this post - ce model prediction makes for a strong argument against the second possibility .
the third possibility is that the stingray is not a pn at all , but is rather a ` pn mimic ' ( see frew & parker 2010 for a review ) .
this has been casually suggested by zijlstra ( 2015 ) in just one short sentence .
a pn mimic is a star with a surrounding ionized shell that does not come from the ejection of the outer envelope of an agb star .
pn mimics include wolf - rayet stars , ` young stellar objects ' , symbiotic stars , b[e ] stars , herbig - haro objects , reflection nebulae , diffuse hii regions , old novae , and supernova remnants .
with such a wide diversity of mimics , it is difficult to absolutely reject all mimic classes for the stingray .
nevertheless , the observed properties of the stingray have little in common with any of the classes of mimics . in detail ,
frew & parker point out how optical emission line ratios can discriminate between a typical pn and the various types of mimics . for this , taking the line fluxes from parthasarathy et al .
( 1993 ) , for the particular lines as defined by frew & parker , we have @xmath40})$]=1.83 , @xmath41})$]=0.48 , @xmath426584\aa}/f_{h\alpha})$]=-0.59 , and @xmath43}/f_{h\beta})$]=0.97 . with these , we can look in their figures 4 and 5 , seeing that the stingray is right in the middle of the pn region , and is far away from the various mimic class regions .
that is , the stingray s emission line spectrum is a classical pn spectrum , and greatly different from the mimics .
further , many properties of the stingray are characteristic for pne and are not found in many classes of mimics ; including the position of the central star on the hr diagram near b1 ii , a classic bipolar point - symmetric shaped shell , the @xmath1510 km s@xmath8 expansion velocity of the shell , dust with a temperature of 125 k , thermal free - free radio emission , a distance of around 340 pc from the galactic plane , and the lack of any nearby ism clouds or young stars . with all this
, we have strong confidence that the stingray is not a pn mimic . after considering all proposals for the nature of the stingray s evolution
, we have eliminated all but one , with the remaining idea being the common idea that the stingray is a post - agb star traversing the upper hr diagram while undergoing some sort of a thermal pulse .
the fundamental problem is that there is no published model for a thermal pulse that come anywhere near the fast time scale for evolution or the large vertical motion in the hr diagram .
we can propose four sets of observations that are currently ongoing or feasible in the near future .
first , it would be good to get optical and ultraviolet spectra of the shell and the central star , so as to measure the temperature while keeping up with the fast evolution of the system .
for this , a program with the _ hst _ in cycle 22 , with n. reindl as the principal investigator , will get near- and far - ultraviolet spectroscopy with the _ cos _ instrument .
second , further _
images in 2016 would provide a long enough time baseline so as to measure an accurate expansion age for the ordinary pn shell .
these same _ hst _ images might show the fast wind from the 1980s event .
this short duration wind has a velocity and total mass like that of the shell ejected by the recurrent nova t pyx , for which we see the impact of the wind onto a prior more - massive shell ( schaefer , pagnotta , & shara 2010 ) .
so we might expect to see this impact lighting up the inside of the old pn shell with bright emission lines .
hst _ images will also allow for resolving the central star and placing it onto the hr diagram , so we can see the evolution from 2006 - 2016 .
for this second set of observations , we have just been awarded time during cycle 23 of _ hst _ , with z. edwards as principal investigator , to use the _ wfc3 _ to take all these needed images .
third , the outer halo of the stingray can be sought in the radio regime ( c.f .
oettl et al .
fourth , time series spectra exposed to show photospheric absorption features of the central star might show a sinusoidal radial velocity curve that will discover a close - in companion star ( as needed to shape the bipolar nebula ) as well as define the separation and mass of the close - in companion .
this observing task could also discover a close binary remnant of a post - ce system .
this research has made use of the _ apass _ database , located at the _ aavso _ web site .
funding for _ apass _ has been provided by the robert martin ayers sciences fund .
we thank a. henden for help in getting the _ apass _ time series .
we thank h. bond and m. parthasarathy for comments on our manuscript .
arkhipova , v. p. , ikonnikova , n. p. , kniazev , a. yu . , & rajoelimanana , a. 2013 , astronomy letters , 39 , 201 blocker , t. 1995 , ap&ss , 275 , 1 blocker , t. 2001 , a&a , 299 , 755 blocker , t. & schnberner , d. 1997 , a&a , 324 , 991 bobrowsky , m. 1994 , apjlett , 426 , l47 bobrowsky , m. , sahu , k. c. , parthasarathy , m. , & garcia - lario , p. 1998 , nature , 392 , 469 chu , y .- h . ,
jacoby , g. h. , & arendt , r. 1987 , apjsupp , 64 , 529 corradi , r. l. m. , schnberner , d. , steffen , m. , & perinotto , m. 2003 , a&a , 340 , 417 cutri , r. m. , skrutskie , m. f. , van dyk , s. , et al .
2003 , vizier online data catalog , 2246 , 0 drilling , j. s. & bergeron , l. e. 1995 , pasp , 107 , 846 feibelman , w. a. 1995 , apj , 443 , 245 fresneau , a. , vaughan , a. e. , & argyle , r. w. 2007 , a&a , 469 , 1221 frew , d. j. & parker , q. a. 2010 , pasa , 27 , 129 gaustad , j. e. , mccullough , p. r. , rosing , w. , & van buren , d. 2001 , pasp , 113 , 1326 gesicki , k. , zijlstra , a. a. , hajduk , m. , & szyszka , c. 2014 , a&a , 566 , a48 hall , p. d. , tout , c. a. , izzard , r. g. , & keller , d. 2013 , mnras , 435 , 2048 handler , g. 2003 , in interplay of periodic , cyclic , and stochastic variability in selected areas of the h - r diagram , ed . c. sterken ( asp conference series , vol . 292 ) , pp .
183 - 190 handler , g. , prinja , r. k. , urbaneja , m. a. , antoci , v. , twicken , j. d. , & barclay , t. 2013 , mnras , 430 , 2923 henize , k. 1976 , apjsuppl , 30 , 491 hill , p. w. , kilkenny , d. , & van breda , i. g. 1974 , mnras , 168 , 451 ishihara , d. , onaka , t. , kataza , h. et al .
2010 , a&a , 514 , a1 kozok , j. r. 1985a , a&as , 61 , 387 kozok , j. r. 1985b , a&as , 62 , 7 oettl , s. , kimeswenger , s. , & zijlstra , a. a. 2015 , a&a , in press , see arxiv:1403.6715 parthasarathy , m. 2000 , bull .
india , 28 , 217 parthasarathy , m. & pottasch , s. r. 1989 , a&a , 225 , 521 parthasarathy , m. , garcia - lario , p. , pottasch , s. r. et al .
1993 , a&a , 267 , l19 parthasarathy , m. , garcia - lario , p. , de martino , d. et al .
1995 , a&a , 300 , l25 pojmanski , g. 2002 , acta astronomica , 52 , 397 reindl , n. , rauch , t. , parthasarthy , m. et al .
2014 , a&a , 565 , 40 schaefer , b. e. baas , 223 , 209.01 schaefer , b. e. baas , 224 , 110.01 schaefer , b. e. , pagnotta , a. , & shara , m. m. 2010 , apj , 708 , 381 schnberner , d. 1983 , apj , 272 , 708 schnberner , d. 2008 , in hydrogen - deficient stars , k. werner and t. rauch eds ( asp conference series , vol .
391 ) , pp .
139 - 149 umana , g. , triglio , c. , cerrigone , l. , buemi , c. s. , & lato , p. 2008 ,
mnras , 386 , 1404 van hoof , p. a. m. , hajduk , m. , zijlstra , a. a. et al .
2007 , a&a , 471 , l9 van winckel , h. 2003 , ara&a , 41 , 391 volk , k. m. & kwok , s. 1989 , apj , 342 , 345 zijlstra , a. 2015 , rev .
mexicana a. & a. , in press , see arxiv:1506.05508 \c .
1001 & reindl et al . ( 2014 ) & ejection of visible pn shell ( assuming 8.4 km / s velocity ) + 1889 & section 2.1 & b=10.1 , fading by 0.0051 mag / year for 1889 - 1981 + c. 1920 & hd catalog & b0 spectrum , no emission lines + c. 1950 & henize ( 1976 ) & h@xmath21 emission line only , no o[iii ] + 1971 & parthasarathy et al . ( 1995 ) & b1 or b2 , class i or ii , only emission is very weak h@xmath17 + 1979.49 & drilling & bergeron ( 1995 ) & o or b spectrum , no emission line + 1980 & section 2.1 & b=10.8 , starts fast fade at 0.20 mag / year + 1988.5 & section 2.1 & b=12.5 , central star is fading fast + 1988 & feibelman ( 1995 ) & bright uv continuum with p cygni profile + 1990 , 1992 & parthsaranthy et al .
( 1993 ; 1995 ) & very bright emission lines dominates optical spectrum + 1992 - 1994 & feibelman ( 1995 ) & bright uv emission lines with no p cygni profile + 1996.3 & section 2.5 & b=14.64 , v=14.96 for central star alone + 1994 - 2015 & sections 2.2 , 2.3 , & 2.4 & v - band ( i.e. , [ oiii ] lines ) fades steadily by 0.081 mag / year + 2411191.6463 & 1889.520 & b & 10.1 & 0.10 & hco ( b 3743 ) + 2411895.7797 & 1891.448 & b & 10.5 & 0.10 & hco ( b 6263 ) + 2411896.7927 & 1891.450 & b & 10.1 & 0.10 & hco ( b 6284 ) + 2412637.6167 & 1893.479 & b & 10.5 & 0.10 & hco ( b 9644 ) + 2412637.6247 & 1893.479 & b & 10.3 & 0.10 & hco ( b 9645 ) + 2412994.6657 & 1894.456 & b & 10.5 & 0.10 & hco ( b 11416 ) + 2413055.6143 & 1894.623 & b & 10.4 & 0.10 & hco ( b 11901 ) + 2413293.8362 & 1895.275 & b & 10.5 & 0.10 & hco ( b 12968 ) + 2413826.5242 & 1896.734 & b & 10.5 & 0.10 & hco ( b 17442 ) + 2414190.5163 & 1897.730 & b & 10.6 & 0.10 & hco ( b 20488 ) + 2414382.8107 & 1898.257 & b & 10.3 & 0.10 & hco ( b 21140 ) + 2414393.8365 & 1898.287 & b & 10.5 & 0.10 & hco ( b 21222 ) + 2414575.4898 & 1898.784 & b & 10.6 & 0.10 & hco ( b 22027 ) + 2414863.6364 & 1899.573 & b & 10.5 & 0.10 & hco ( am 22 ) + 2414925.5190 & 1899.743 & b & 10.4 & 0.10 & hco ( am 162 ) + 2414927.5695 & 1899.748 & b & 10.4 & 0.10 & hco ( am 173 ) + 2415274.5747 & 1900.698 & b & 10.4 & 0.10 & hco ( am 644 ) + 2415533.6520 & 1901.408 & b & 10.3 & 0.10 & hco ( am 816 ) + 2415646.5346 & 1901.717 & b & 10.2 & 0.10 & hco ( am 1061 ) + 2416011.5291 & 1902.716 & b & 10.2 & 0.10 & hco ( am 1545 ) + 2416022.5612 & 1902.746 & b & 10.4 & 0.10 & hco ( am 1584 ) + 2416297.6522 & 1903.499 & b & 10.2 & 0.10 & hco ( am 2084 ) + 2416302.6630 & 1903.513 & b & 10.3 & 0.10 & hco ( am 2101 ) + 2416311.5921 & 1903.538 & b & 10.2 & 0.10 & hco ( am 2128 ) + 2416563.7676 & 1904.228 & b & 10.3 & 0.10 & hco ( am 2492 ) + 2418102.5438 & 1908.441 & b & 10.4 & 0.10 & hco ( am 5549 ) + 2418533.5118 & 1909.621 & b & 10.5 & 0.10 & hco ( am 6461 ) + 2419239.5967 & 1911.554 & b & 10.6 & 0.10 & hco ( am 7589 ) + 2420654.7449 & 1915.428 & b & 10.3 & 0.10 & hco ( am 10631 ) + 2420720.5558 & 1915.609 & b & 10.3 & 0.10 & hco ( am 10972 ) + 2421022.6793 & 1916.436 & b & 10.2 & 0.10 & hco ( am 11914 ) + 2421081.5455 & 1916.597 & b & 10.3 & 0.10 & hco ( am 12308 ) + 2421449.5399 & 1917.604 & b & 10.1 & 0.10 & hco ( am 13484 ) + 2421757.6732 & 1918.448 & b & 10.2 & 0.10 & hco ( am 14060 ) + 2422175.5687 & 1919.592 & b & 10.5 & 0.10 & hco ( am 14890 ) + 2422542.5516 & 1920.597 & b & 10.6 & 0.10 & hco ( am 15379 ) + 2423902.8720 & 1924.321 & b & 10.6 & 0.10 & hco ( ax 484 ) + 2423918.7837 & 1924.365 & b & 10.6 & 0.10 & hco ( ax 527 ) + 2423997.6383 & 1924.581 & b & 10.6 & 0.10 & hco ( ax 732 ) + 2424074.5172 & 1924.791 & b & 10.4 & 0.10 & hco ( ax 862 ) + 2424357.5954 & 1925.566 & b & 10.3 & 0.10 & hco ( ax 1246 ) + 2424433.5119 & 1925.774 & b & 10.3 & 0.10 & hco ( ax 1364 ) + 2424648.7057 & 1926.363 & b & 10.2 & 0.10 & hco ( ax 1624 ) + 2424670.6955 & 1926.424 & b & 10.3 & 0.10 & hco ( ax 1675 ) + 2425795.4167 & 1929.503 & b & 10.6 & 0.10 & hco ( rb 426 ) + 2425886.2442 & 1929.752 & b & 10.3 & 0.10 & hco ( rb 569 ) + 2426189.4166 & 1930.582 & b & 10.5 & 0.10 & hco ( rb 1198 ) + 2426214.3066 & 1930.650 & b & 10.6 & 0.10 & hco ( rb 1241 ) + 2426505.4597 & 1931.447 & b & 10.6 & 0.10 & hco ( rb 1855 ) + 2426807.5668 & 1932.274 & b & 10.5 & 0.10 & hco ( rb 2601 ) + 2426812.5972 & 1932.288 & b & 10.5 & 0.10 & hco ( rb 2623 ) + 2426871.3811 & 1932.449 & b & 10.1 & 0.10 & hco ( rb 2850 ) + 2426939.2347 & 1932.634 & b & 10.3 & 0.10 & hco ( rb 3351 ) + 2427301.3199 & 1933.626 & b & 10.5 & 0.10 & hco ( rb 4536 ) + 2427601.4153 & 1934.447 & b & 10.5 & 0.10 & hco ( rb 5311 ) + 2427659.3741 & 1934.606 & b & 10.5 & 0.10 & hco ( rb 5537 ) + 2428049.2445 & 1935.673 & b & 10.8 & 0.10 & hco ( rb 6389 ) + 2428341.4863 & 1936.474 & b & 10.7 & 0.10 & hco ( rb 6862 ) + 2428448.2549 & 1936.766 & b & 10.7 & 0.10 & hco ( rb 7124 ) + 2428818.2554 & 1937.779 & b & 10.6 & 0.10 & hco ( rb 7621 ) + 2429130.3246 & 1938.633 & b & 10.8 & 0.10 & hco ( rb 8177 ) + 2429553.2404 & 1939.791 & b & 10.6 & 0.10 & hco ( rb 9185 ) + 2429746.4462 & 1940.320 & b & 10.7 & 0.10 & hco ( rb 9531 ) + 2429911.2340 & 1940.771 & b & 10.5 & 0.10 & hco ( rb 10034 ) + 2430150.3369 & 1941.426 & b & 10.6 & 0.10 & hco ( rb 10750 ) + 2430455.6135 & 1942.262 & b & 10.6 & 0.10 & hco ( rb 11406 ) + 2430963.2405 & 1943.652 & b & 10.6 & 0.10 & hco ( rb 12357 ) + 2431251.5632 & 1944.441 & b & 10.7 & 0.10 & hco ( rb 12897 ) + 2431262.5470 & 1944.471 & b & 10.6 & 0.10 & hco ( rb 12923 ) + 2431627.3593 & 1945.470 & b & 10.6 & 0.10 & hco ( rb 13738 ) + 2431993.5103 & 1946.472 & b & 10.5 & 0.10 & hco ( rb 14474 ) + 2432051.3238 & 1946.631 & b & 10.6 & 0.10 & hco ( rb 14677 ) + 2432353.4886 & 1947.458 & b & 10.7 & 0.10 & hco ( rb 15162 ) + 2432787.3021 & 1948.646 & b & 10.7 & 0.10 & hco ( rb 15853 ) + 2433512.3271 & 1950.631 & b & 10.7 & 0.10 & hco ( rb 16747 ) + 2433566.2594 & 1950.778 & b & 10.7 & 0.10 & hco ( rb 16799 ) + 2440200 & 1968.939 & u & 9.84 & 0.03 & hill et al .
( 1974 ) + 2440200 & 1968.939 & b & 10.73 & 0.03 & hill et al .
( 1974 ) + 2440200 & 1968.939 & v & 10.75 & 0.03 & hill et al .
( 1974 ) + 2440739.4498 & 1970.417 & b & 10.7 & 0.10 & hco ( dsb 38 ) + 2440741.4428 & 1970.423 & b & 10.8 & 0.10 & hco ( dsb 45 ) + 2440773.3530 & 1970.510 & b & 10.8 & 0.10 & hco ( dsb 73 ) + 2440805.2599 & 1970.598 & b & 10.6 & 0.10 & hco ( dsb 104 ) + 2441148.3256 & 1971.537 & b & 10.7 & 0.10 & hco ( dsb 257 ) + 2441160.3084 & 1971.570 & b & 10.6 & 0.10 & hco ( dsb 277 ) + 2441538.2577 & 1972.604 & b & 10.9 & 0.10 & hco ( dsb 377 ) + 2441539.2500 & 1972.607 & b & 10.8 & 0.10 & hco ( dsb 379 ) + 2441540.2471 & 1972.610 & b & 10.6 & 0.10 & hco ( dsb 381 ) + 2443612.7774 & 1978.284 & b & 11.1 & 0.10 & hco ( dsb 486 ) + 2443691.6038 & 1978.500 & b & 11.0 & 0.10 & hco ( dsb 524 ) + 2444018.6753 & 1979.395 & b & 11.0 & 0.10 & hco ( dsb 579 ) + 2444024.7914 & 1979.412 & b & 11.0 & 0.10 & hco ( dsb 618 ) + 2444163.5616 & 1979.792 & b & 11.0 & 0.10 & hco ( dsb 634 ) + 2444422 & 1980.498 & u & 9.95 & 0.02 & kozok ( 1985a ) + 2444422 & 1980.498 & b & 10.84 & 0.02 & kozok ( 1985a ) + 2444422 & 1980.498 & v & 10.88 & 0.01 & kozok ( 1985a ) + 2444431 & 1980.523 & u & 10.03 & 0.02 & kozok ( 1985a ) + 2444431 & 1980.523 & b & 10.92 & 0.02 & kozok ( 1985a ) + 2444431 & 1980.523 & v & 10.97 & 0.01 & kozok ( 1985a ) + 2444432 & 1980.526 & u & 10.08 & 0.02 & kozok ( 1985a ) + 2444432 & 1980.526 & b & 10.96 & 0.02 & kozok ( 1985a ) + 2444432 & 1980.526 & v & 10.99 & 0.01 & kozok ( 1985a ) + 2444697.1106 & 1981.253 & b & 10.8 & 0.10 & hco ( dsb 723 ) + 2444840.8754 & 1981.646 & b & 11.0 & 0.10 & hco ( dsb 774 ) + 2444843.8578 & 1981.655 & b & 11.0 & 0.10 & hco ( dsb 775 ) + 2445107.0202 & 1982.375 & b & 11.0 & 0.10 & hco ( dsb 885 ) + 2445134.9450 & 1982.452 & b & 10.9 & 0.10 & hco ( dsb 896 ) + 2445192.9166 & 1982.610 & b & 11.3 & 0.10 & hco ( dsb 937 ) + 2445520.9565 & 1983.508 & b & 11.6 & 0.10 & hco ( dsb 1076 ) + 2445819.9958 & 1984.327 & b & 11.7 & 0.10 & hco ( dsb 1289 ) + 2445960.8634 & 1984.713 & b & 11.8 & 0.10 & hco ( dsb 1462 ) + 2446145.1235 & 1985.217 & b & 12.1 & 0.10 & hco ( dsb 1582 ) + 2446319.8629 & 1985.696 & b & 12.3 & 0.10 & hco ( dsb 1769 ) + 2446348.8717 & 1985.775 & b & 12.3 & 0.10 & hco ( dsb 1812 ) + 2446695.8599 & 1986.725 & b & 12.3 & 0.10 & hco ( dsb 2060 ) + 2447019.8700 & 1987.612 & b & 12.2 & 0.10 & hco ( dsb 2331 ) + 2447356.9576 & 1988.535 & b & 12.5 & 0.10 & hco ( dsb 2634 ) + 2447413.9359 & 1988.691 & b & 12.3 & 0.10 & hco ( dsb 2694 ) + 2447682.1206 & 1989.425 & b & 12.3 & 0.10 & hco ( dsb 2832 ) + 2447760.9248 & 1989.641 & b & 12.3 & 0.10 & hco ( dsb 2866 ) + 2449418.1600 & 1994.178 & v & 11.3 & 0.15 & jones + 2449419.1800 & 1994.181 & v & 10.7 & 0.15 & jones + 2449420.1800 & 1994.184 & v & 10.5 & 0.15 & jones + 2449421.1400 & 1994.187 & v & 10.7 & 0.15 & jones + 2449422.2000 & 1994.189 & v & 10.7 & 0.15 & jones + 2449424.1600 & 1994.195 & v & 10.7 & 0.15 & jones + 2449428.1700 & 1994.206 & v & 10.7 & 0.15 & jones + 2449432.1800 & 1994.217 & v & 10.8 & 0.15 & jones + 2449433.2000 & 1994.220 & v & 10.7 & 0.15 & jones + 2449435.1900 & 1994.225 & v & 10.7 & 0.15 & jones + 2449438.1900 & 1994.233 & v & 10.8 & 0.15 & jones + 2449447.2000 & 1994.258 & v & 10.7 & 0.15 & jones + 2449452.9600 & 1994.274 & v & 10.5 & 0.15 & jones + 2449453.9500 & 1994.276 & v & 10.7 & 0.15 & jones + 2449472.8800 & 1994.328 & v & 11.1 & 0.15 & jones + 2449476.9400 & 1994.339 & v & 10.7 & 0.15 & jones + 2449482.9000 & 1994.356 & v & 10.8 & 0.15 & jones + 2449487.8600 & 1994.369 & v & 11.3 & 0.15 & jones + 2449488.9200 & 1994.372 & v & 11.3 & 0.15 & jones + 2449503.9300 & 1994.413 & v & 10.9 & 0.15 & jones + 2449505.9100 & 1994.419 & v & 11.1 & 0.15 & jones + 2449510.9100 & 1994.432 & v & 11.0 & 0.15 & jones + 2449522.8300 & 1994.465 & v & 10.8 & 0.15 & jones + 2449539.9100 & 1994.512 & v & 10.7 & 0.15 & jones + 2449576.9800 & 1994.613 & v & 10.6 & 0.15 & jones + 2449591.8700 & 1994.654 & v & 10.8 & 0.15 & jones + 2449651.8800 & 1994.818 & v & 11.3 & 0.15 & jones + 2449858.9000 & 1995.385 & v & 11.2 & 0.15 & jones + 2449953.9200 & 1995.645 & v & 11.0 & 0.15 & jones + 2450148.34 & 1996.176 & b & 14.64 & 0.05 & reindl et al .
( 2014 ) + 2450148.34 & 1996.176 & v & 14.96 & 0.05 & reindl et al .
( 2014 ) + 2450150.85 & 1996.183 & v & 15.4 & 0.05 & bobrowski ( 1998 ) + 2450191.1700 & 1996.295 & v & 10.8 & 0.15 & jones + 2450248.9200 & 1996.453 & v & 10.6 & 0.15 & jones + 2450318.8800 & 1996.644 & v & 10.8 & 0.15 & jones + 2451058.8600 & 1998.670 & v & 11.0 & 0.15 & jones + 2451064.8760 & 1998.687 & v & 10.9 & 0.15 & jones + 2451066.9500 & 1998.693 & v & 11.2 & 0.15 & jones + 2451068.8980 & 1998.698 & v & 11.1 & 0.15 & jones + 2451075.9480 & 1998.717 & v & 11.1 & 0.15 & jones + 2451078.9060 & 1998.725 & v & 11.1 & 0.15 & jones + 2451102.9260 & 1998.791 & v & 11.1 & 0.15 & jones + 2451110.9300 & 1998.813 & v & 11.1 & 0.15 & jones + 2451125.8900 & 1998.854 & v & 11.0 & 0.15 & jones + 2451132.9090 & 1998.873 & v & 10.9 & 0.15 & jones + 2451144.9000 & 1998.906 & v & 11.1 & 0.15 & jones + 2451198.1390 & 1999.052 & v & 11.1 & 0.15 & jones + 2451220.1590 & 1999.112 & v & 11.3 & 0.15 & jones + 2451238.1620 & 1999.161 & v & 10.9 & 0.15 & jones + 2451259.1950 & 1999.219 & v & 10.9 & 0.15 & jones + 2451278.9660 & 1999.273 & v & 11.1 & 0.15 & jones + 2451288.9310 & 1999.300 & v & 11.0 & 0.15 & jones + 2451305.9060 & 1999.347 & v & 11.1 & 0.15 & jones + 2451317.9130 & 1999.380 & v & 10.9 & 0.15 & jones + 2451325.2240 & 1999.400 & v & 11.1 & 0.15 & jones + 2451336.9530 & 1999.432 & v & 11.2 & 0.15 & jones + 2451348.9690 & 1999.465 & v & 11.1 & 0.15 & jones + 2451366.1980 & 1999.512 & v & 11.1 & 0.15 & jones + 2451369.2220 & 1999.520 & v & 11.1 & 0.15 & jones + 2451373.9570 & 1999.533 & v & 11.4 & 0.15 & jones + 2451379.9320 & 1999.549 & v & 11.1 & 0.15 & jones + 2451390.8930 & 1999.579 & v & 11.1 & 0.15 & jones + 2451403.9040 & 1999.615 & v & 11.2 & 0.15 & jones + 2451408.9580 & 1999.629 & v & 11.1 & 0.15 & jones + 2451428.9310 & 1999.684 & v & 11.1 & 0.15 & jones + 2451448.8670 & 1999.738 & v & 11.1 & 0.15 & jones + 2451463.9210 & 1999.779 & v & 11.1 & 0.15 & jones + 2451465.9040 & 1999.785 & v & 11.1 & 0.15 & jones + 2451470.8850 & 1999.798 & v & 11.1 & 0.15 & jones + 2451496.8830 & 1999.870 & v & 11.1 & 0.15 & jones + 2451557.1350 & 2000.035 & v & 11.1 & 0.15 & jones + 2451561.1470 & 2000.046 & v & 11.1 & 0.15 & jones + 2451571.1410 & 2000.073 & v & 11.2 & 0.15 & jones + 2451583.1150 & 2000.106 & v & 11.2 & 0.15 & jones + 2451603.1580 & 2000.161 & v & 11.1 & 0.15 & jones + 2451612.1080 & 2000.185 & v & 11.2 & 0.15 & jones + 2451618.1650 & 2000.202 & v & 11.1 & 0.15 & jones + 2451666.9090 & 2000.335 & v & 11.1 & 0.15 & jones + 2451689.2530 & 2000.396 & v & 11.1 & 0.15 & jones + 2451703.2210 & 2000.435 & v & 11.1 & 0.15 & jones + 2451730.8950 & 2000.510 & v & 11.1 & 0.15 & jones + 2451757.8490 & 2000.584 & v & 11.2 & 0.15 & jones + 2451762.8910 & 2000.598 & v & 11.5 & 0.15 & jones + 2451850.8910 & 2000.839 & v & 11.2 & 0.15 & jones + 2451869.9100 & 2000.891 & v & 11.1 & 0.15 & jones + 2451939.8804 & 2001.082 & v & 11.66 & 0.22 & asas + 2451946.8850 & 2001.102 & v & 11.38 & 0.22 & asas + 2451949.8658 & 2001.110 & v & 11.78 & 0.22 & asas + 2451953.8386 & 2001.121 & v & 11.40 & 0.22 & asas + 2451955.8488 & 2001.126 & v & 11.61 & 0.22 & asas + 2451976.1650 & 2001.182 & v & 11.2 & 0.15 & jones + 2451980.8003 & 2001.195 & v & 11.52 & 0.22 & asas + 2452025.7054 & 2001.317 & v & 11.67 & 0.22 & asas + 2452025.7094 & 2001.317 & v & 11.33 & 0.22 & asas + 2452026.7397 & 2001.320 & v & 11.30 & 0.22 & asas + 2452027.7596 & 2001.323 & v & 11.28 & 0.22 & asas + 2452030.7248 & 2001.331 & v & 11.58 & 0.22 & asas + 2452030.7511 & 2001.331 & v & 11.21 & 0.22 & asas + 2452031.7373 & 2001.334 & v & 11.38 & 0.22 & asas + 2452032.7199 & 2001.337 & v & 11.24 & 0.22 & asas + 2452033.7096 & 2001.339 & v & 11.36 & 0.22 & asas + 2452037.6818 & 2001.350 & v & 11.24 & 0.22 & asas + 2452037.7008 & 2001.350 & v & 11.41 & 0.22 & asas + 2452040.7062 & 2001.359 & v & 11.37 & 0.22 & asas + 2452052.6794 & 2001.391 & v & 11.36 & 0.22 & asas + 2452054.6889 & 2001.397 & v & 11.44 & 0.22 & asas + 2452055.6733 & 2001.400 & v & 11.34 & 0.22 & asas + 2452055.6939 & 2001.400 & v & 11.31 & 0.22 & asas + 2452055.9390 & 2001.400 & v & 10.9 & 0.15 & jones + 2452056.6745 & 2001.402 & v & 11.41 & 0.22 & asas + 2452056.6951 & 2001.402 & v & 11.42 & 0.22 & asas + 2452057.6703 & 2001.405 & v & 11.22 & 0.22 & asas + 2452058.6901 & 2001.408 & v & 11.38 & 0.22 & asas + 2452067.6349 & 2001.432 & v & 11.48 & 0.22 & asas + 2452068.6512 & 2001.435 & v & 11.50 & 0.22 & asas + 2452069.6673 & 2001.438 & v & 11.54 & 0.22 & asas + 2452070.6599 & 2001.441 & v & 11.28 & 0.22 & asas + 2452080.5344 & 2001.468 & v & 11.36 & 0.22 & asas + 2452081.6372 & 2001.471 & v & 11.61 & 0.22 & asas + 2452082.6507 & 2001.473 & v & 11.43 & 0.22 & asas + 2452083.6327 & 2001.476 & v & 11.36 & 0.22 & asas + 2452085.6257 & 2001.482 & v & 11.41 & 0.22 & asas + 2452086.6108 & 2001.484 & v & 11.55 & 0.22 & asas + 2452086.6316 & 2001.484 & v & 11.55 & 0.22 & asas + 2452086.8530 & 2001.485 & v & 10.9 & 0.15 & jones + 2452087.6062 & 2001.487 & v & 11.33 & 0.22 & asas + 2452088.6049 & 2001.490 & v & 11.42 & 0.22 & asas + 2452088.6235 & 2001.490 & v & 11.43 & 0.22 & asas + 2452093.5358 & 2001.503 & v & 11.35 & 0.22 & asas + 2452094.5729 & 2001.506 & v & 11.36 & 0.22 & asas + 2452094.5842 & 2001.506 & v & 11.49 & 0.22 & asas + 2452103.5795 & 2001.531 & v & 11.27 & 0.22 & asas + 2452103.5959 & 2001.531 & v & 11.49 & 0.22 & asas + 2452104.5816 & 2001.533 & v & 11.33 & 0.22 & asas + 2452105.5882 & 2001.536 & v & 11.36 & 0.22 & asas + 2452115.5335 & 2001.563 & v & 11.80 & 0.22 & asas + 2452122.1720 & 2001.582 & v & 11.1 & 0.15 & jones + 2452124.5178 & 2001.588 & v & 11.37 & 0.22 & asas + 2452125.5343 & 2001.591 & v & 11.35 & 0.22 & asas + 2452131.4906 & 2001.607 & v & 11.38 & 0.22 & asas + 2452135.5000 & 2001.618 & v & 11.36 & 0.22 & asas + 2452140.4924 & 2001.632 & v & 11.43 & 0.22 & asas + 2452142.5206 & 2001.637 & v & 11.31 & 0.22 & asas + 2452144.4891 & 2001.643 & v & 11.47 & 0.22 & asas + 2452151.4906 & 2001.662 & v & 11.52 & 0.22 & asas + 2452156.4760 & 2001.675 & v & 11.79 & 0.22 & asas + 2452164.4984 & 2001.697 & v & 11.53 & 0.22 & asas + 2452167.4762 & 2001.706 & v & 12.01 & 0.22 & asas + 2452168.5447 & 2001.709 & v & 11.65 & 0.22 & asas + 2452170.9569 & 2001.715 & v & 10.9 & 0.15 & jones + 2452172.4844 & 2001.719 & v & 11.73 & 0.22 & asas + 2452180.5243 & 2001.741 & v & 11.57 & 0.22 & asas + 2452184.5221 & 2001.752 & v & 11.81 & 0.22 & asas + 2452191.5163 & 2001.771 & v & 11.62 & 0.22 & asas + 2452195.4854 & 2001.782 & v & 11.11 & 0.22 & asas + 2452197.4924 & 2001.788 & v & 11.94 & 0.22 & asas + 2452198.5096 & 2001.791 & v & 11.61 & 0.22 & asas + 2452203.4956 & 2001.804 & v & 11.65 & 0.22 & asas + 2452203.5043 & 2001.804 & v & 11.77 & 0.22 & asas + 2452205.5173 & 2001.810 & v & 11.60 & 0.22 & asas + 2452302.1465 & 2002.074 & v & 11.2 & 0.15 & jones + 2452309.1167 & 2002.093 & v & 11.5 & 0.15 & jones + 2452320.1153 & 2002.124 & v & 11.2 & 0.15 & jones + 2452359.1382 & 2002.230 & v & 11.2 & 0.15 & jones + 2452368.9403 & 2002.257 & v & 11.2 & 0.15 & jones + 2452428.9417 & 2002.421 & v & 11.2 & 0.15 & jones + 2452505.9208 & 2002.632 & v & 11.1 & 0.15 & jones + 2452510.6398 & 2002.645 & v & 11.61 & 0.22 & asas + 2452526.5885 & 2002.689 & v & 11.79 & 0.22 & asas + 2452535.5610 & 2002.713 & v & 11.56 & 0.22 & asas + 2452540.6191 & 2002.727 & v & 12.15 & 0.22 & asas + 2452545.5895 & 2002.741 & v & 11.54 & 0.22 & asas + 2452549.5558 & 2002.752 & v & 11.48 & 0.22 & asas + 2452559.5464 & 2002.779 & v & 11.59 & 0.22 & asas + 2452560.5584 & 2002.782 & v & 11.64 & 0.22 & asas + 2452561.5578 & 2002.785 & v & 11.65 & 0.22 & asas + 2452563.5373 & 2002.790 & v & 11.62 & 0.22 & asas + 2452565.5198 & 2002.795 & v & 11.65 & 0.22 & asas + 2452566.5385 & 2002.798 & v & 11.89 & 0.22 & asas + 2452660.8594 & 2003.056 & v & 11.48 & 0.22 & asas + 2452663.8687 & 2003.065 & v & 11.53 & 0.22 & asas + 2452671.8520 & 2003.087 & v & 11.66 & 0.22 & asas + 2452671.8520 & 2003.087 & v & 11.95 & 0.22 & asas + 2452678.1431 & 2003.104 & v & 11.4 & 0.15 & jones + 2452678.8486 & 2003.106 & v & 11.82 & 0.22 & asas + 2452678.8716 & 2003.106 & v & 11.67 & 0.22 & asas + 2452685.8505 & 2003.125 & v & 11.66 & 0.22 & asas + 2452688.8499 & 2003.133 & v & 11.58 & 0.22 & asas + 2452703.8489 & 2003.174 & v & 11.22 & 0.22 & asas + 2452709.8384 & 2003.191 & v & 11.62 & 0.22 & asas + 2452711.8482 & 2003.196 & v & 11.66 & 0.22 & asas + 2452719.8236 & 2003.218 & v & 11.83 & 0.22 & asas + 2452719.8374 & 2003.218 & v & 11.38 & 0.22 & asas + 2452725.8025 & 2003.234 & v & 11.82 & 0.22 & asas + 2452727.8347 & 2003.240 & v & 11.21 & 0.22 & asas + 2452736.7614 & 2003.264 & v & 12.06 & 0.22 & asas + 2452747.8730 & 2003.295 & v & 11.31 & 0.22 & asas + 2452750.8662 & 2003.303 & v & 11.61 & 0.22 & asas + 2452751.7141 & 2003.305 & v & 11.84 & 0.22 & asas + 2452759.7983 & 2003.327 & v & 11.47 & 0.22 & asas + 2452770.8889 & 2003.358 & v & 11.3 & 0.15 & jones + 2452784.6988 & 2003.395 & v & 11.75 & 0.22 & asas + 2452787.7900 & 2003.404 & v & 11.60 & 0.22 & asas + 2452790.9107 & 2003.412 & v & 11.88 & 0.22 & asas + 2452790.9107 & 2003.412 & v & 11.69 & 0.22 & asas + 2452791.6407 & 2003.414 & v & 11.69 & 0.22 & asas + 2452802.8354 & 2003.445 & v & 11.40 & 0.22 & asas + 2452805.8355 & 2003.453 & v & 11.39 & 0.22 & asas + 2452812.2333 & 2003.471 & v & 11.3 & 0.15 & jones + 2452820.6991 & 2003.494 & v & 11.69 & 0.22 & asas + 2452844.8257 & 2003.560 & v & 11.3 & 0.15 & jones + 2452851.7313 & 2003.579 & v & 11.40 & 0.22 & asas + 2452865.7480 & 2003.617 & v & 11.73 & 0.22 & asas + 2452867.6999 & 2003.623 & v & 11.82 & 0.22 & asas + 2452876.6294 & 2003.647 & v & 11.64 & 0.22 & asas + 2452883.5493 & 2003.666 & v & 11.71 & 0.22 & asas + 2452885.5371 & 2003.672 & v & 11.64 & 0.22 & asas + 2452886.6790 & 2003.675 & v & 11.77 & 0.22 & asas + 2452893.5181 & 2003.693 & v & 11.88 & 0.22 & asas + 2452894.6070 & 2003.696 & v & 11.61 & 0.22 & asas + 2452906.6079 & 2003.729 & v & 11.57 & 0.22 & asas + 2452910.4864 & 2003.740 & v & 11.74 & 0.22 & asas + 2452912.5141 & 2003.745 & v & 11.62 & 0.22 & asas + 2452915.5404 & 2003.754 & v & 12.01 & 0.22 & asas + 2452917.5660 & 2003.759 & v & 11.67 & 0.22 & asas + 2452921.5653 & 2003.770 & v & 11.89 & 0.22 & asas + 2452923.5627 & 2003.776 & v & 11.94 & 0.22 & asas + 2452927.5127 & 2003.786 & v & 11.90 & 0.22 & asas + 2452929.5257 & 2003.792 & v & 11.71 & 0.22 & asas + 2452931.5299 & 2003.797 & v & 11.89 & 0.22 & asas + 2452933.5338 & 2003.803 & v & 11.84 & 0.22 & asas + 2452935.5351 & 2003.808 & v & 11.80 & 0.22 & asas + 2452954.5032 & 2003.860 & v & 11.70 & 0.22 & asas + 2453052.8736 & 2004.130 & v & 11.90 & 0.22 & asas + 2453056.8464 & 2004.141 & v & 11.72 & 0.22 & asas + 2453060.8541 & 2004.152 & v & 11.99 & 0.22 & asas + 2453061.8946 & 2004.154 & v & 11.80 & 0.22 & asas + 2453064.8438 & 2004.162 & v & 11.61 & 0.22 & asas + 2453069.8637 & 2004.176 & v & 11.78 & 0.22 & asas + 2453071.8658 & 2004.182 & v & 11.72 & 0.22 & asas + 2453077.8512 & 2004.198 & v & 11.86 & 0.22 & asas + 2453080.8684 & 2004.206 & v & 11.67 & 0.22 & asas + 2453080.8782 & 2004.206 & v & 11.90 & 0.22 & asas + 2453085.9064 & 2004.220 & v & 11.69 & 0.22 & asas + 2453091.8268 & 2004.236 & v & 11.72 & 0.22 & asas + 2453096.8798 & 2004.250 & v & 11.81 & 0.22 & asas + 2453098.8592 & 2004.256 & v & 12.02 & 0.22 & asas + 2453099.8203 & 2004.258 & v & 12.06 & 0.22 & asas + 2453109.7758 & 2004.285 & v & 11.82 & 0.22 & asas + 2453112.7640 & 2004.294 & v & 11.72 & 0.22 & asas + 2453114.8238 & 2004.299 & v & 11.83 & 0.22 & asas + 2453122.7609 & 2004.321 & v & 11.76 & 0.22 & asas + 2453126.7968 & 2004.332 & v & 11.69 & 0.22 & asas + 2453132.8082 & 2004.349 & v & 11.66 & 0.22 & asas + 2453160.7912 & 2004.425 & v & 11.90 & 0.22 & asas + 2453163.7740 & 2004.433 & v & 11.86 & 0.22 & asas + 2453165.8024 & 2004.439 & v & 11.99 & 0.22 & asas + 2453171.7415 & 2004.455 & v & 11.90 & 0.22 & asas + 2453175.7424 & 2004.466 & v & 11.95 & 0.22 & asas + 2453178.6073 & 2004.474 & v & 12.10 & 0.22 & asas + 2453180.7050 & 2004.480 & v & 11.95 & 0.22 & asas + 2453187.5360 & 2004.498 & v & 11.52 & 0.22 & asas + 2453187.7002 & 2004.499 & v & 11.80 & 0.22 & asas + 2453189.7613 & 2004.504 & v & 11.58 & 0.22 & asas + 2453294.5414 & 2004.791 & v & 11.73 & 0.22 & asas + 2453398.8613 & 2005.077 & v & 12.10 & 0.22 & asas + 2453404.8696 & 2005.093 & v & 11.98 & 0.22 & asas + 2453408.8702 & 2005.104 & v & 12.12 & 0.22 & asas + 2453416.8635 & 2005.126 & v & 12.02 & 0.22 & asas + 2453420.8560 & 2005.137 & v & 11.90 & 0.22 & asas + 2453430.8633 & 2005.165 & v & 12.01 & 0.22 & asas + 2453436.8921 & 2005.181 & v & 11.66 & 0.22 & asas + 2453443.8497 & 2005.200 & v & 11.69 & 0.22 & asas + 2453446.8599 & 2005.208 & v & 11.73 & 0.22 & asas + 2453448.8490 & 2005.214 & v & 11.70 & 0.22 & asas + 2453448.8490 & 2005.214 & v & 12.16 & 0.22 & asas + 2453449.8600 & 2005.217 & v & 12.12 & 0.22 & asas + 2453452.8075 & 2005.225 & v & 11.79 & 0.22 & asas + 2453453.7784 & 2005.227 & v & 12.18 & 0.22 & asas + 2453455.8246 & 2005.233 & v & 11.85 & 0.22 & asas + 2453457.8797 & 2005.239 & v & 12.20 & 0.22 & asas + 2453462.8115 & 2005.252 & v & 11.87 & 0.22 & asas + 2453465.8270 & 2005.260 & v & 11.90 & 0.22 & asas + 2453468.8339 & 2005.269 & v & 11.92 & 0.22 & asas + 2453474.8106 & 2005.285 & v & 12.08 & 0.22 & asas + 2453477.7987 & 2005.293 & v & 12.30 & 0.22 & asas + 2453481.8177 & 2005.304 & v & 11.94 & 0.22 & asas + 2453487.7945 & 2005.320 & v & 11.90 & 0.22 & asas + 2453492.8215 & 2005.334 & v & 11.44 & 0.22 & asas + 2453496.8055 & 2005.345 & v & 11.98 & 0.22 & asas + 2453499.9028 & 2005.354 & v & 11.7 & 0.15 & jones + 2453501.6721 & 2005.358 & v & 11.80 & 0.22 & asas + 2453502.8931 & 2005.362 & v & 11.8 & 0.15 & jones + 2453503.9014 & 2005.365 & v & 12.2 & 0.15 & jones + 2453510.2403 & 2005.382 & v & 12.0 & 0.15 & jones + 2453510.8345 & 2005.384 & v & 12.16 & 0.22 & asas + 2453512.7141 & 2005.389 & v & 11.95 & 0.22 & asas + 2453516.7668 & 2005.400 & v & 11.68 & 0.22 & asas + 2453520.7044 & 2005.411 & v & 12.30 & 0.22 & asas + 2453520.8889 & 2005.411 & v & 12.0 & 0.15 & jones + 2453522.7321 & 2005.416 & v & 11.81 & 0.22 & asas + 2453524.7617 & 2005.422 & v & 12.51 & 0.22 & asas + 2453525.2313 & 2005.423 & v & 11.9 & 0.15 & jones + 2453526.7830 & 2005.427 & v & 12.20 & 0.22 & asas + 2453527.8417 & 2005.430 & v & 12.2 & 0.15 & jones + 2453528.8239 & 2005.433 & v & 12.01 & 0.22 & asas + 2453533.8521 & 2005.447 & v & 12.0 & 0.15 & jones + 2453537.7329 & 2005.457 & v & 11.90 & 0.22 & asas + 2453544.7434 & 2005.476 & v & 12.32 & 0.22 & asas + 2453547.7901 & 2005.485 & v & 11.83 & 0.22 & asas + 2453548.8319 & 2005.488 & v & 12.1 & 0.15 & jones + 2453551.7742 & 2005.496 & v & 12.37 & 0.22 & asas + 2453554.7389 & 2005.504 & v & 12.23 & 0.22 & asas + 2453556.8377 & 2005.509 & v & 12.10 & 0.22 & asas + 2453561.8375 & 2005.523 & v & 11.8 & 0.15 & jones + 2453562.5703 & 2005.525 & v & 11.85 & 0.22 & asas + 2453564.6177 & 2005.531 & v & 12.02 & 0.22 & asas + 2453566.8306 & 2005.537 & v & 11.9 & 0.15 & jones + 2453570.6111 & 2005.547 & v & 12.12 & 0.22 & asas + 2453573.6604 & 2005.556 & v & 12.03 & 0.22 & asas + 2453575.6783 & 2005.561 & v & 11.87 & 0.22 & asas + 2453583.5856 & 2005.583 & v & 11.99 & 0.22 & asas + 2453585.6427 & 2005.588 & v & 11.97 & 0.22 & asas + 2453587.6631 & 2005.594 & v & 11.94 & 0.22 & asas + 2453589.7100 & 2005.599 & v & 12.31 & 0.22 & asas + 2453592.5372 & 2005.607 & v & 12.03 & 0.22 & asas + 2453592.9396 & 2005.608 & v & 11.8 & 0.15 & jones + 2453597.9118 & 2005.622 & v & 11.6 & 0.15 & jones + 2453601.6689 & 2005.632 & v & 12.21 & 0.22 & asas + 2453603.6683 & 2005.638 & v & 12.00 & 0.22 & asas + 2453615.9000 & 2005.671 & v & 11.8 & 0.15 & jones + 2453616.5523 & 2005.673 & v & 11.63 & 0.22 & asas + 2453620.6512 & 2005.684 & v & 11.90 & 0.22 & asas + 2453625.6121 & 2005.698 & v & 11.99 & 0.22 & asas + 2453631.5224 & 2005.714 & v & 12.30 & 0.22 & asas + 2453633.5429 & 2005.719 & v & 11.55 & 0.22 & asas + 2453637.9528 & 2005.732 & v & 11.5 & 0.15 & jones + 2453639.5684 & 2005.736 & v & 11.97 & 0.22 & asas + 2453644.5947 & 2005.750 & v & 11.88 & 0.22 & asas + 2453648.5583 & 2005.761 & v & 11.95 & 0.22 & asas + 2453650.5768 & 2005.766 & v & 11.89 & 0.22 & asas + 2453651.8847 & 2005.770 & v & 11.5 & 0.15 & jones + 2453653.5248 & 2005.774 & v & 11.96 & 0.22 & asas + 2453658.5283 & 2005.788 & v & 12.05 & 0.22 & asas + 2453660.5531 & 2005.793 & v & 12.08 & 0.22 & asas + 2453665.5400 & 2005.807 & v & 12.15 & 0.22 & asas + 2453669.9035 & 2005.819 & v & 11.5 & 0.15 & jones + 2453672.5139 & 2005.826 & v & 11.83 & 0.22 & asas + 2453675.5117 & 2005.834 & v & 11.30 & 0.22 & asas + 2453680.5014 & 2005.848 & v & 11.97 & 0.22 & asas + 2453693.8931 & 2005.885 & v & 11.5 & 0.15 & jones + 2453759.8684 & 2006.065 & v & 12.12 & 0.22 & asas + 2453764.1389 & 2006.077 & v & 11.8 & 0.15 & jones + 2453767.8726 & 2006.087 & v & 13.10 & 0.22 & asas + 2453774.8697 & 2006.106 & v & 12.12 & 0.22 & asas + 2453783.8886 & 2006.131 & v & 12.37 & 0.22 & asas + 2453790.8606 & 2006.150 & v & 12.08 & 0.22 & asas + 2453794.8718 & 2006.161 & v & 12.14 & 0.22 & asas + 2453797.8895 & 2006.169 & v & 12.07 & 0.22 & asas + 2453798.1688 & 2006.170 & v & 12.0 & 0.15 & jones + 2453802.8748 & 2006.183 & v & 11.96 & 0.22 & asas + 2453806.8387 & 2006.194 & v & 11.66 & 0.22 & asas + 2453809.8700 & 2006.202 & v & 12.02 & 0.22 & asas + 2453813.8289 & 2006.213 & v & 11.84 & 0.22 & asas + 2453820.8212 & 2006.232 & v & 11.93 & 0.22 & asas + 2453820.8581 & 2006.232 & v & 12.07 & 0.22 & asas + 2453823.8838 & 2006.241 & v & 12.73 & 0.22 & asas + 2453829.8389 & 2006.257 & v & 12.17 & 0.22 & asas + 2453832.8288 & 2006.265 & v & 12.21 & 0.22 & asas + 2453835.8145 & 2006.273 & v & 11.98 & 0.22 & asas + 2453849.8533 & 2006.312 & v & 12.19 & 0.22 & asas + 2453852.7762 & 2006.320 & v & 11.90 & 0.22 & asas + 2453858.7578 & 2006.336 & v & 12.34 & 0.22 & asas + 2453858.7578 & 2006.336 & v & 12.03 & 0.22 & asas + 2453860.8193 & 2006.342 & v & 11.94 & 0.22 & asas + 2453860.9424 & 2006.342 & v & 11.8 & 0.15 & jones + 2453864.7419 & 2006.352 & v & 12.37 & 0.22 & asas + 2453866.7336 & 2006.358 & v & 12.01 & 0.22 & asas + 2453871.7867 & 2006.372 & v & 11.89 & 0.22 & asas + 2453874.2285 & 2006.378 & v & 11.5 & 0.15 & jones + 2453874.8760 & 2006.380 & v & 11.89 & 0.22 & asas + 2453880.7593 & 2006.396 & v & 11.96 & 0.22 & asas + 2453887.7141 & 2006.415 & v & 12.00 & 0.22 & asas + 2453891.9069 & 2006.427 & v & 11.8 & 0.15 & jones + 2453893.7000 & 2006.432 & v & 11.75 & 0.22 & asas + 2453898.7406 & 2006.446 & v & 11.90 & 0.22 & asas + 2453900.8053 & 2006.451 & v & 11.97 & 0.22 & asas + 2453902.7901 & 2006.457 & v & 12.46 & 0.22 & asas + 2453904.8119 & 2006.462 & v & 12.09 & 0.22 & asas + 2453906.8617 & 2006.468 & v & 12.13 & 0.22 & asas + 2453909.6979 & 2006.476 & v & 11.49 & 0.22 & asas + 2453910.8438 & 2006.479 & v & 12.0 & 0.15 & jones + 2453913.7008 & 2006.487 & v & 12.60 & 0.22 & asas + 2453961.9479 & 2006.619 & v & 11.8 & 0.15 & jones + 2453969.9382 & 2006.640 & v & 11.5 & 0.15 & jones + 2453981.5070 & 2006.672 & v & 12.14 & 0.22 & asas + 2454006.8903 & 2006.742 & v & 11.7 & 0.15 & jones + 2454028.9028 & 2006.802 & v & 11.5 & 0.15 & jones + 2454133.8735 & 2007.089 & v & 12.12 & 0.22 & asas + 2454141.8739 & 2007.111 & v & 12.24 & 0.22 & asas + 2454145.8681 & 2007.122 & v & 12.31 & 0.22 & asas + 2454158.8794 & 2007.158 & v & 12.38 & 0.22 & asas + 2454162.8490 & 2007.169 & v & 12.33 & 0.22 & asas + 2454166.8473 & 2007.180 & v & 12.17 & 0.22 & asas + 2454176.8751 & 2007.207 & v & 12.56 & 0.22 & asas + 2454180.8332 & 2007.218 & v & 12.23 & 0.22 & asas + 2454187.8775 & 2007.237 & v & 12.09 & 0.22 & asas + 2454190.8382 & 2007.245 & v & 12.11 & 0.22 & asas + 2454193.8168 & 2007.253 & v & 11.93 & 0.22 & asas + 2454204.8726 & 2007.284 & v & 12.05 & 0.22 & asas + 2454208.8074 & 2007.294 & v & 12.14 & 0.22 & asas + 2454210.9472 & 2007.300 & v & 12.3 & 0.15 & jones + 2454212.8477 & 2007.306 & v & 12.37 & 0.22 & asas + 2454230.7997 & 2007.355 & v & 12.03 & 0.22 & asas + 2454232.7736 & 2007.360 & v & 12.20 & 0.22 & asas + 2454232.8009 & 2007.360 & v & 12.35 & 0.22 & asas + 2454234.8409 & 2007.366 & v & 12.24 & 0.22 & asas + 2454245.6876 & 2007.395 & v & 12.30 & 0.22 & asas + 2454247.6793 & 2007.401 & v & 11.93 & 0.22 & asas + 2454253.8798 & 2007.418 & v & 11.95 & 0.22 & asas + 2454256.6701 & 2007.426 & v & 11.96 & 0.22 & asas + 2454258.7451 & 2007.431 & v & 11.99 & 0.22 & asas + 2454266.7835 & 2007.453 & v & 11.74 & 0.22 & asas + 2454272.6893 & 2007.469 & v & 11.96 & 0.22 & asas + 2454274.6799 & 2007.475 & v & 12.01 & 0.22 & asas + 2454277.6464 & 2007.483 & v & 11.92 & 0.22 & asas + 2454283.6649 & 2007.499 & v & 12.27 & 0.22 & asas + 2454284.8117 & 2007.503 & v & 12.68 & 0.22 & asas + 2454300.7397 & 2007.546 & v & 12.52 & 0.22 & asas + 2454302.7716 & 2007.552 & v & 12.12 & 0.22 & asas + 2454303.8236 & 2007.555 & v & 11.8 & 0.15 & jones + 2454311.7078 & 2007.576 & v & 11.88 & 0.22 & asas + 2454319.5581 & 2007.598 & v & 12.47 & 0.22 & asas + 2454329.5505 & 2007.625 & v & 12.06 & 0.22 & asas + 2454331.5490 & 2007.631 & v & 11.99 & 0.22 & asas + 2454336.5705 & 2007.644 & v & 12.22 & 0.22 & asas + 2454338.7201 & 2007.650 & v & 12.15 & 0.22 & asas + 2454341.5459 & 2007.658 & v & 12.04 & 0.22 & asas + 2454343.5567 & 2007.663 & v & 12.24 & 0.22 & asas + 2454353.6701 & 2007.691 & v & 11.99 & 0.22 & asas + 2454358.5176 & 2007.704 & v & 12.12 & 0.22 & asas + 2454360.5034 & 2007.710 & v & 12.65 & 0.22 & asas + 2454363.5142 & 2007.718 & v & 12.08 & 0.22 & asas + 2454365.5575 & 2007.724 & v & 11.88 & 0.22 & asas + 2454368.6116 & 2007.732 & v & 12.33 & 0.22 & asas + 2454376.5456 & 2007.754 & v & 12.46 & 0.22 & asas + 2454378.5402 & 2007.759 & v & 12.65 & 0.22 & asas + 2454378.5689 & 2007.759 & v & 12.56 & 0.22 & asas + 2454383.5520 & 2007.773 & v & 12.19 & 0.22 & asas + 2454392.5262 & 2007.797 & v & 11.96 & 0.22 & asas + 2454400.5187 & 2007.819 & v & 11.90 & 0.22 & asas + 2454407.5123 & 2007.839 & v & 12.58 & 0.22 & asas + 2454411.5069 & 2007.849 & v & 12.65 & 0.22 & asas + 2454502.8725 & 2008.100 & v & 12.44 & 0.22 & asas + 2454510.8659 & 2008.121 & v & 12.34 & 0.22 & asas + 2454517.8930 & 2008.141 & v & 11.92 & 0.22 & asas + 2454521.8572 & 2008.152 & v & 12.48 & 0.22 & asas + 2454524.8782 & 2008.160 & v & 12.16 & 0.22 & asas + 2454527.8753 & 2008.168 & v & 12.07 & 0.22 & asas + 2454531.8385 & 2008.179 & v & 12.34 & 0.22 & asas + 2454535.9008 & 2008.190 & v & 12.57 & 0.22 & asas + 2454543.8230 & 2008.212 & v & 12.32 & 0.22 & asas + 2454546.8327 & 2008.220 & v & 12.11 & 0.22 & asas + 2454555.7964 & 2008.244 & v & 12.50 & 0.22 & asas + 2454561.8265 & 2008.261 & v & 12.07 & 0.22 & asas + 2454564.8139 & 2008.269 & v & 12.47 & 0.22 & asas + 2454567.8049 & 2008.277 & v & 11.80 & 0.22 & asas + 2454570.7971 & 2008.286 & v & 12.34 & 0.22 & asas + 2454572.8470 & 2008.291 & v & 12.08 & 0.22 & asas + 2454576.8270 & 2008.302 & v & 12.46 & 0.22 & asas + 2454583.8015 & 2008.321 & v & 12.09 & 0.22 & asas + 2454589.7645 & 2008.337 & v & 12.20 & 0.22 & asas + 2454592.7407 & 2008.346 & v & 12.22 & 0.22 & asas + 2454594.8225 & 2008.351 & v & 12.05 & 0.22 & asas + 2454602.7343 & 2008.373 & v & 12.65 & 0.22 & asas + 2454606.7516 & 2008.384 & v & 12.79 & 0.22 & asas + 2454609.7037 & 2008.392 & v & 12.60 & 0.22 & asas + 2454617.7273 & 2008.414 & v & 11.88 & 0.22 & asas + 2454623.7322 & 2008.430 & v & 12.47 & 0.22 & asas + 2454627.6957 & 2008.441 & v & 12.89 & 0.22 & asas + 2454629.6734 & 2008.447 & v & 12.15 & 0.22 & asas + 2454633.6629 & 2008.458 & v & 12.34 & 0.22 & asas + 2454638.7445 & 2008.472 & v & 12.33 & 0.22 & asas + 2454640.7113 & 2008.477 & v & 12.43 & 0.22 & asas + 2454642.6848 & 2008.482 & v & 12.21 & 0.22 & asas + 2454644.7043 & 2008.488 & v & 12.21 & 0.22 & asas + 2454646.7164 & 2008.493 & v & 12.52 & 0.22 & asas + 2454648.7597 & 2008.499 & v & 12.24 & 0.22 & asas + 2454650.8210 & 2008.505 & v & 12.44 & 0.22 & asas + 2454653.5883 & 2008.512 & v & 12.06 & 0.22 & asas + 2454655.6225 & 2008.518 & v & 12.62 & 0.22 & asas + 2454657.6614 & 2008.523 & v & 12.04 & 0.22 & asas + 2454660.6124 & 2008.531 & v & 12.36 & 0.22 & asas + 2454663.6202 & 2008.540 & v & 12.49 & 0.22 & asas + 2454664.7649 & 2008.543 & v & 12.44 & 0.22 & asas + 2454672.7114 & 2008.565 & v & 12.01 & 0.22 & asas + 2454681.6587 & 2008.589 & v & 12.27 & 0.22 & asas + 2454685.6391 & 2008.600 & v & 12.63 & 0.22 & asas + 2454687.5945 & 2008.605 & v & 12.12 & 0.22 & asas + 2454690.5458 & 2008.613 & v & 12.24 & 0.22 & asas + 2454701.6330 & 2008.644 & v & 12.92 & 0.22 & asas + 2454705.6663 & 2008.655 & v & 12.06 & 0.22 & asas + 2454710.6381 & 2008.668 & v & 12.10 & 0.22 & asas + 2454720.5595 & 2008.696 & v & 12.21 & 0.22 & asas + 2454728.5784 & 2008.718 & v & 12.07 & 0.22 & asas + 2454730.6178 & 2008.723 & v & 12.58 & 0.22 & asas + 2454733.6256 & 2008.731 & v & 13.12 & 0.22 & asas + 2454738.5167 & 2008.745 & v & 12.59 & 0.22 & asas + 2454741.5098 & 2008.753 & v & 12.06 & 0.22 & asas + 2454745.5096 & 2008.764 & v & 12.29 & 0.22 & asas + 2454747.5440 & 2008.769 & v & 12.48 & 0.22 & asas + 2454756.5339 & 2008.794 & v & 12.47 & 0.22 & asas + 2454759.5039 & 2008.802 & v & 12.72 & 0.22 & asas + 2454759.5257 & 2008.802 & v & 12.11 & 0.22 & asas + 2454762.5051 & 2008.810 & v & 12.31 & 0.22 & asas + 2454762.5246 & 2008.810 & v & 12.04 & 0.22 & asas + 2454765.5202 & 2008.819 & v & 12.41 & 0.22 & asas + 2454768.5156 & 2008.827 & v & 12.66 & 0.22 & asas + 2454772.4985 & 2008.838 & v & 12.52 & 0.22 & asas + 2454852.8693 & 2009.058 & v & 12.55 & 0.22 & asas + 2454869.8809 & 2009.104 & v & 13.09 & 0.22 & asas + 2454877.8747 & 2009.126 & v & 12.32 & 0.22 & asas + 2454882.8760 & 2009.140 & v & 12.67 & 0.22 & asas + 2454886.8588 & 2009.151 & v & 12.29 & 0.22 & asas + 2454890.8723 & 2009.162 & v & 12.76 & 0.22 & asas + 2454904.8806 & 2009.200 & v & 12.59 & 0.22 & asas + 2454916.8113 & 2009.233 & v & 12.55 & 0.22 & asas + 2454924.9081 & 2009.255 & v & 12.13 & 0.22 & asas + 2454928.8875 & 2009.266 & v & 13.05 & 0.22 & asas + 2454937.7961 & 2009.290 & v & 12.81 & 0.22 & asas + 2454939.8639 & 2009.296 & v & 13.11 & 0.22 & asas + 2454946.7752 & 2009.315 & v & 12.33 & 0.22 & asas + 2454951.7892 & 2009.329 & v & 11.64 & 0.22 & asas + 2454954.7473 & 2009.337 & v & 12.27 & 0.22 & asas + 2454969.8717 & 2009.378 & v & 12.69 & 0.22 & asas + 2454975.7602 & 2009.394 & v & 12.37 & 0.22 & asas + 2454984.7423 & 2009.419 & v & 12.45 & 0.22 & asas + 2454988.7129 & 2009.430 & v & 12.96 & 0.22 & asas + 2455002.6908 & 2009.468 & v & 12.09 & 0.22 & asas + 2455006.6930 & 2009.479 & v & 12.18 & 0.22 & asas + 2455012.6958 & 2009.495 & v & 13.07 & 0.22 & asas + 2455014.6438 & 2009.501 & v & 12.34 & 0.22 & asas + 2455019.7041 & 2009.515 & v & 12.44 & 0.22 & asas + 2455021.7344 & 2009.520 & v & 12.56 & 0.22 & asas + 2455023.7397 & 2009.526 & v & 12.44 & 0.22 & asas + 2455040.5943 & 2009.572 & v & 12.16 & 0.22 & asas + 2455042.5812 & 2009.577 & v & 12.32 & 0.22 & asas + 2455067.6128 & 2009.646 & v & 12.30 & 0.22 & asas + 2455069.6101 & 2009.651 & v & 12.05 & 0.22 & asas + 2455075.6351 & 2009.668 & v & 12.09 & 0.22 & asas + 2455087.6403 & 2009.701 & v & 12.89 & 0.22 & asas + 2455092.6229 & 2009.714 & v & 12.06 & 0.22 & asas + 2455097.5133 & 2009.728 & v & 12.16 & 0.22 & asas + 2455099.5400 & 2009.733 & v & 12.70 & 0.22 & asas + 2455106.4991 & 2009.752 & v & 12.29 & 0.22 & asas + 2455111.5380 & 2009.766 & v & 12.48 & 0.22 & asas + 2455649.8986 & 2011.240 & v & 12.27 & 0.01 & apass + 2455649.8992 & 2011.240 & b & 13.21 & 0.01 & apass + 2455649.9001 & 2011.240 & r & 12.22 & 0.01 & apass + 2455649.9013 & 2011.240 & i & 13.27 & 0.01 & apass + 2455649.9020 & 2011.240 & g & 12.31 & 0.01 & apass + 2455677.6330 & 2011.316 & v & 12.28 & 0.01 & apass + 2455677.6335 & 2011.316 & b & 13.21 & 0.01 & apass + 2455677.6343 & 2011.316 & r & 12.19 & 0.01 & apass + 2455677.6356 & 2011.316 & i & 13.22 & 0.01 & apass + 2455677.6361 & 2011.316 & g & 12.30 & 0.01 & apass + 2456839.4923 & 2014.496 & g & 12.52 & 0.01 & decam + 2456839.4975 & 2014.496 & r & 12.18 & 0.01 & decam + 2456839.5027 & 2014.496 & i & 13.35 & 0.01 & decam + 2457104.9560 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457104.9567 & 2015.224 & v & 12.65 & 0.01 & mount john + 2457104.9573 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9580 & 2015.224 & v & 12.65 & 0.01 & mount john + 2457104.9593 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9607 & 2015.224 & v & 12.59 & 0.01 & mount john + 2457104.9613 & 2015.224 & v & 12.58 & 0.01 & mount john + 2457104.9620 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457104.9707 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9714 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9720 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9727 & 2015.224 & v & 12.64 & 0.01 & mount john + 2457104.9734 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457104.9741 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457104.9747 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457104.9761 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9767 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457104.9774 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457104.9781 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457104.9787 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9794 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457104.9800 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9807 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9814 & 2015.224 & v & 12.59 & 0.01 & mount john + 2457104.9821 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457104.9827 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457104.9834 & 2015.224 & v & 12.59 & 0.01 & mount john + 2457104.9841 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457104.9847 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9854 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9860 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9867 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457104.9874 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9881 & 2015.224 & v & 12.58 & 0.01 & mount john + 2457104.9887 & 2015.224 & v & 12.59 & 0.01 & mount john + 2457104.9914 & 2015.224 & v & 12.59 & 0.01 & mount john + 2457104.9921 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457104.9928 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457104.9934 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457104.9941 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457104.9948 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457104.9955 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9961 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9968 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457104.9974 & 2015.224 & v & 12.64 & 0.01 & mount john + 2457104.9981 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9988 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457104.9994 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0001 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0008 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0014 & 2015.224 & v & 12.64 & 0.01 & mount john + 2457105.0021 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0028 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457105.0034 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0041 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0048 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0055 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0061 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0068 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0075 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0081 & 2015.224 & v & 12.65 & 0.01 & mount john + 2457105.0088 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0095 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0101 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0108 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0115 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0122 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0128 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0135 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457105.0141 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0148 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0155 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457105.0161 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0169 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0176 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0182 & 2015.224 & v & 12.66 & 0.01 & mount john + 2457105.0189 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0196 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457105.0202 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0209 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0216 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0222 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0229 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0236 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0242 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0249 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0256 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457105.0262 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0269 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0276 & 2015.224 & v & 12.65 & 0.01 & mount john + 2457105.0282 & 2015.224 & v & 12.59 & 0.01 & mount john + 2457105.0289 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0296 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0302 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0309 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0343 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0349 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0356 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0363 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0369 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0376 & 2015.224 & v & 12.65 & 0.01 & mount john + 2457105.0383 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0389 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0396 & 2015.224 & v & 12.64 & 0.01 & mount john + 2457105.0403 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0409 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0416 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0423 & 2015.224 & v & 12.64 & 0.01 & mount john + 2457105.0430 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0436 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0443 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0450 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0456 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0463 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0469 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0476 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0483 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0489 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457105.0496 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457105.0503 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0509 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0516 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0523 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0529 & 2015.224 & v & 12.64 & 0.01 & mount john + 2457105.0536 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0543 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0549 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0556 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0563 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0569 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0576 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0583 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0589 & 2015.224 & v & 12.59 & 0.01 & mount john + 2457105.0596 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457105.0603 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0609 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0616 & 2015.224 & v & 12.56 & 0.01 & mount john + 2457105.0623 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0629 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0636 & 2015.224 & v & 12.59 & 0.01 & mount john + 2457105.0643 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0650 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0656 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0663 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0670 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0676 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0683 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457105.0690 & 2015.224 & v & 12.66 & 0.01 & mount john + 2457105.0696 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0703 & 2015.224 & v & 12.58 & 0.01 & mount john + 2457105.0710 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0717 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0723 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0730 & 2015.224 & v & 12.65 & 0.01 & mount john + 2457105.0737 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0743 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0750 & 2015.224 & v & 12.65 & 0.01 & mount john + 2457105.0757 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0763 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457105.0770 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0777 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0784 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0790 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0797 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0803 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457105.0810 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0817 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0823 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0830 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0837 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457105.0843 & 2015.224 & v & 12.62 & 0.01 & mount john + 2457105.0850 & 2015.224 & v & 12.64 & 0.01 & mount john + 2457105.0857 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457105.0863 & 2015.224 & v & 12.61 & 0.01 & mount john + 2457105.0870 & 2015.224 & v & 12.60 & 0.01 & mount john + 2457105.0877 & 2015.224 & v & 12.63 & 0.01 & mount john + 2457107.9548 & 2015.232 & v & 12.57 & 0.01 & mount john + 2457107.9554 & 2015.232 & v & 12.56 & 0.01 & mount john + 2457107.9561 & 2015.232 & v & 12.58 & 0.01 & mount john + 2457107.9568 & 2015.232 & v & 12.58 & 0.01 & mount john + 2457107.9574 & 2015.232 & v & 12.56 & 0.01 & mount john + 2457109.1266 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1274 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.1281 & 2015.235 & v & 12.63 & 0.01 & mount john + 2457109.1287 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.1294 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1301 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1307 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1314 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1321 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1327 & 2015.235 & v & 12.63 & 0.01 & mount john + 2457109.1334 & 2015.235 & v & 12.56 & 0.01 & mount john + 2457109.1341 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1347 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.1354 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1361 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1368 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.1374 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.1381 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1387 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1394 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1401 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1407 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1414 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.1421 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.1427 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1434 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1441 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1447 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1454 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.1461 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.1467 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.1474 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.1481 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1487 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1494 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1501 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.1507 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1514 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.1521 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1527 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1534 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1541 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1548 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1554 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1561 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.1567 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.1574 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.1581 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1588 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1594 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.1601 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1608 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1621 & 2015.235 & v & 12.58 & 0.01 & mount john + 2457109.1628 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.1634 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1641 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1648 & 2015.235 & v & 12.58 & 0.01 & mount john + 2457109.1654 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1661 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1668 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1674 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1681 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1688 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.1694 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1701 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1708 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.1715 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1721 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1728 & 2015.235 & v & 12.58 & 0.01 & mount john + 2457109.1734 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1741 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1748 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1754 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1761 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.1768 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.1774 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1781 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1788 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.1794 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1801 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1808 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1814 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.1821 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1828 & 2015.235 & v & 12.57 & 0.01 & mount john + 2457109.1834 & 2015.235 & v & 12.58 & 0.01 & mount john + 2457109.1841 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1848 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1854 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1861 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1868 & 2015.235 & v & 12.63 & 0.01 & mount john + 2457109.1874 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1881 & 2015.235 & v & 12.63 & 0.01 & mount john + 2457109.1888 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.1894 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1901 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1908 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1914 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1921 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1928 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1934 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1941 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1948 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1954 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1961 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1968 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1975 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.1981 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1988 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.1994 & 2015.235 & v & 12.58 & 0.01 & mount john + 2457109.2001 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.2008 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2014 & 2015.235 & v & 12.58 & 0.01 & mount john + 2457109.2021 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2028 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.2035 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.2041 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2048 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.2054 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.2061 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2068 & 2015.235 & v & 12.63 & 0.01 & mount john + 2457109.2074 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.2081 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2088 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2094 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2101 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.2107 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.2114 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2121 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2127 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2134 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.2141 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2147 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.2154 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.2161 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.2167 & 2015.235 & v & 12.58 & 0.01 & mount john + 2457109.2174 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2180 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.2187 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.2201 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2207 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.2214 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.2221 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.2227 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2234 & 2015.235 & v & 12.63 & 0.01 & mount john + 2457109.2240 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2247 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2254 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2260 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.2267 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.2280 & 2015.235 & v & 12.57 & 0.01 & mount john + 2457109.2287 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.2294 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2300 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2307 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.2314 & 2015.235 & v & 12.58 & 0.01 & mount john + 2457109.2321 & 2015.235 & v & 12.63 & 0.01 & mount john + 2457109.2327 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.2334 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.2341 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.2347 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.2354 & 2015.235 & v & 12.59 & 0.01 & mount john + 2457109.2361 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.2367 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2374 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2381 & 2015.235 & v & 12.62 & 0.01 & mount john + 2457109.2388 & 2015.235 & v & 12.58 & 0.01 & mount john + 2457109.2394 & 2015.235 & v & 12.58 & 0.01 & mount john + 2457109.2401 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.2408 & 2015.235 & v & 12.61 & 0.01 & mount john + 2457109.2414 & 2015.235 & v & 12.60 & 0.01 & mount john + 2457109.2414 & 2015.235 & v & 12.62 & 0.01 & mount john + \c .
1001 & 3.70 @xmath10 0.20 & 3.74 @xmath10 0.5 + 1889 & 4.48 @xmath10 0.10 & 3.74 @xmath10 0.41 + 1980 & 4.45 @xmath10 0.11 & 3.38 @xmath10 0.41 + 1988 & 4.58 @xmath10 0.13 & 2.65 @xmath10 0.75 + 1996 & 4.70 @xmath10 0.10 & 2.73 @xmath10 0.67 + 2002 & 4.78 @xmath10 0.08 & 2.24 @xmath10 0.63 + 2006 & 4.74 @xmath10 0.09 & 2.09 @xmath10 0.65 + | up until around 1980 , the stingray was an ordinary b1 post - agb star , but then it suddenly sprouted bright emission lines like in a planetary nebula ( pn ) , and soon after this the _ hubble space telescope _ ( _ hst _ ) discovered a small pn around the star , so apparently we have caught a star in the act of ionizing a pn .
we report here on a well - sampled light curve from 1889 to 2015 , with unique coverage of the prior century plus the entire duration of the pn formation plus three decades of its aftermath . surprisingly , the star anticipated the 1980 s ionization event by declining from b=10.30 in 1889 to b=10.76 in 1980 .
starting in 1980 , the central star faded fast , at a rate of 0.20 mag / year , reaching b=14.64 in 1996 .
this fast fading is apparently caused by the central star shrinking in size . from 1994 - 2015 ,
the v - band light curve is almost entirely from the flux of two bright [ oiii ] emission lines from the unresolved nebula , and it shows a consistent decline at a rate of 0.090 mag / year .
this steady fading ( also seen in the radio and infrared ) has a time scale equal to that expected for ordinary recombination within the nebula , immediately after a short - duration ionizing event in the 1980s .
we are providing the first direct measure of the rapidly changing luminosity of the central star on both sides of a presumed thermal pulse in 1980 , with this providing a strong and critical set of constraints , and these are found to sharply disagree with theoretical models of pn evolution . |
although quasars were originally discovered as a result of their radio emission , it is now clear that a majority of quasars are radio - quiet. it is common to classify quasars as radio - loud or radio - quiet either based on their radio luminosity or on the ratio of their radio to optical luminosities .
the latter ratio is known to span a range of 5 orders of magnitude . to date
there is no generally accepted explanation for the difference between radio - loud and radio - quiet quasars .
one avenue of research has been to search for correlations between the fraction of radio - loud quasars ( f@xmath6 ) with other observables .
it is still an open question whether radio - loud and radio - quiet quasars are two distinct classes of objects or represent a continuous distribution of radio to optical power .
previous studies have attempted to define unbiased samples of quasars to address the nature of radio - loud quasars .
these samples have been drawn from optically and x - ray selected quasar surveys .
their main drawback has been the relatively small size of the samples .
in contrast , this study emphasizes quantity at the risk of introducing unrecognized biases . that is to say , in this work we will examine the radio emission from all known quasars over a large region of sky independent of how the individual quasars were discovered .
this has one key advantage , a very large sample , which optimistically may compensate for the diversity of the sample . in a recent study , hooper ( 1996 )
presented a sample of 61 vla radio detections among 359 optically selected quasars from the large bright quasar survey ( lbqs ) ( hewett 1995 ) .
they computed f@xmath6 as a function of redshift and absolute magnitude and found that it was only weakly dependent on either quantity . in this paper
we present a similar study for a much larger sample of quasars . using the national radio astronomy observatory very large array sky survey ( nvss ) ( condon 1995 ) , we have searched for radio emission from 4079 quasars ( 47 % of the quasars currently known ) in the latest edition of the veron catalog ( veron - cetty and veron 1996 , hereafter vv ) which lie within the area of sky covered by the nvss .
this exercise is possible because of new radio surveys which are becoming available . in particular ,
the vla is currently being used to survey 75 % of the sky at 1400 mhz .
the survey , formally known as the nvss , has an angular resolution of @xmath7 45 arcsec and a 5 sigma sensitivity of @xmath7 2.4 mjy .
to date , data for approximately one third of the eventual sky coverage has been released .
our methodology is straightfoward .
we compare the positions of quasars in vv to the positions of radio sources in the nvss catalog and calculate the separation between each quasar and the nearest radio source .
we estimate the chance coincidence rate by repeating the experiment with a false quasar catalog generated by shifting the positions in vv by 1 degree .
( note that the chance coincidence rate is in agreement with expectations for a random distribution . )
the spatial distribution of quasars that fall within the nvss survey as of may 1 , 1996 is shown in figure 1 .
the obvious nonuniform distribution of quasars on the sky is a direct result of inhomogeneous sky coverage of surveys . in figure 2
, we show a histogram of the angular separations between the known quasars and the nearest radio source both for vv as well as the false quasar catalog .
we conclude that matches within 10 arcsec are 98 % real while matches beyond 35 arcsec are almost totally chance coincidences .
( note that 10 arcsec is approximately equal to the radius of the 90 % confidence error circle for a 2.4mjy source ) .
the 149 matches between 10 - 35 arcsec are a mixture of real ( 60 % ) and chance ( 40 % ) coincidences and will not be included in this study even though they constitute 10 % of the real matches . in summary , we find 799 radio detections of quasars , including 168 new detections which are listed in table 1 , 3280 nondetections for which we assign an upper limit of 2.4 mjy , and 149 ambiguous matches which we discard . there has been no attempt to identify which of the quasars were originally discovered as a result of their radio emission .
in figure 3 , we show a histogram of the 1400 mhz flux densities up to 2 jy of the 799 quasars with counterparts in the nvss .
the shaded histogram depicts the same for the 168 quasars detected for the first time here .
it is apparent that there are very few new detections above 100 mjy . using the radio flux densities ( or upper limits ) in conjunction with the redshifts as given by vv
, we have calculated a radio luminosity ( or upper limit ) for each quasar using the equations from weedman ( 1986 ) , where we have assumed a radio spectral index of @xmath10.5 , a hubble constant of h@xmath8= 50 and q@xmath8= 0.5 .
furthermore , using the same spectral index we have adjusted the radio luminosity to 8.4 ghz to facilitate comparisons to previously published results(e.g .
hooper 1996 ) . in figures 4 and 5
, we show the distribution of l@xmath9 as a function of redshift z and m@xmath0 ( as given in vv with a correction to q@xmath8=0.5 ) . using 10@xmath10 ergs / sec / hz as the division between radio - loud and radio - quiet ,
it is clear that the nvss detects most of the radio - loud quasars out to a redshift of 1.5 while missing most of the radio - quiet quasars . of the 3280 upper limits , 1799 fall above 10@xmath10 ergs / sec / hz . it may be more informative to consider the ratio of radio to optical luminosity l@xmath9/l@xmath11 in defining radio loud quasars . in figures 6 and 7 , we plot log(l@xmath9/l@xmath11 ) against z and m@xmath0 . in this context ,
the divide between radio - loud and radio - quiet is often taken to be 1.0 .
there are only 139 upper limits which lie above this divide .
lastly , using 1.0 as the divide between radio - loud and radio - quiet , we calculate f@xmath6 as a function of m@xmath0 and z ( see figures 8 and 9 ) where we have not included upper limits above 1.0 in the calculation .
the dependence of f@xmath6 on m@xmath0 is minimal , remaining nearly constant at @xmath7 0.16 between m@xmath0 of @xmath123 to @xmath129 . at m@xmath0=@xmath130 ,
f@xmath6 may decline to below 0.10 .
the case is quite different for the dependence on redshift .
f@xmath6 rises dramatically between z=0 and z=0.5 from an initial value of 0.05 up to 0.25 . beyond z=1.0
, f@xmath6 appears to slowly decrease to 0.12 at z=2 and then remain approximately constant out to z=4 .
these results are similar to the z dependence reported by hooper ( 1995,1996 ) on a much smaller sample from the lbqs .
considerable effort has been expended over the past several decades to understand the underlying difference between radio - loud and radio - quiet quasars .
these studies have recently focussed on two specific questions ; namely , do the radio properties of quasars evolve with time and do they correlate with optical luminosity ? unfortunately the answers to these questions seem to differ from quasar sample to quasar sample .
for example , hooper ( 1996 ) found that f@xmath6 was nearly constant at 0.10 over a large range of absolute b magnitude .
in contrast they pointed out that a similar analysis of the palomar green sample ( schmidt and green 1983 ; kellermann 1989 ) revealed strong variations in f@xmath6 with absolute b magnitude .
hooper ( 1996 ) were at a loss to explain the difference .
our comparison between the vv catalog and the nvss has yielded radio data on a quasar sample which is an order of magnitude larger than previous samples . in general , the radio properties of this sample are very similar to those of the lbqs . in particular , f@xmath6 for this new large sample
is also largely independent of absolute b magnitude as shown in figure 8 .
likewise , in agreement with the lbqs sample , we see a peak in f@xmath6 between z of 0.5 to 1.0 .
however , with this larger sample , we can study the dependence on z with much higher resolution and as a result , find a significantly stronger dependence on z. in particular , we find that f@xmath6 is a factor of five lower today than it was at z=0.5 . while there has been considerable discussion in the literature on the differential evolution of radio - loud quasars at very high redshifts , our results suggest that differential evolution of radio - loud quasars has been most important over the last several billion years . as we noted at the outset of this paper , to obtain a larger sample we sacrificed the uniformity of the sample .
it is possible that the extreme differential evolution we observe between z of 0 and 0.5 is a consequence of various selection effects .
for example , the nearby quasar population has a higher proportion of low luminosity objects .
this in principle could bias the dependence of radio - loud fraction on redshift .
however we do not see a dependence of radio - loud fraction on optical luminosity .
while we can not rule out some unforseen bias , we are encouraged that the lbqs sample also shows a peak at z of 1 , albeit with less statistical significance and resolution .
this type of study and the conclusions drawn from it certainly should be used carefully in light of the heterogenous nature of the quasar sample .
this is particularly true in regions of parameter space where the sample is very sparse , such as at very high redshift or luminosity .
however , selection effects are probably less important at lower redshifts where quasars are well - sampled and easier to detect . in particular , it is hard to see how radio - loud quasars could be selected against at low redshift , so we think it very likely that this absence at low redshift is a real effect .
the authors thank chris impey for many useful comments .
this work was supported in part under the auspices of the department of energy by lawrence livermore national laboratory under contract w-7405-eng-48 and by the institute of geophysics and planetary physics .
condon , j.j . , cotton , w.d . ,
greisen , e.w .
, yin , q.f . , perley , r.a . , and broderick , j.j .
1994 preprint .
hewett , p.c .
, foltz , c.b . , and chaffee , f.h . , 1995 , aj , 109 , 1498 hooper , e.j . ,
impey , c.d . ,
foltz , c.b . , and hewett , p.c . , 1995 , apj , 445 , 62 hooper , e.j . , impey , c.d . , foltz , c.b . , and hewett , p.c . , 1996 , apj , 473 , 746 kellermann , k.i . , sramek , r. , schmidt , m. , shaffer , d.b . , and green , r. , 1989 , aj , 98 , 1195 .
schmidt , m. and green , r.f . , 1983 ,
apj , 269 , 352 veron - cetty , m .- p and veron , p. , ( vv ) , 1996 , eso scientific report no.x .
weedman , d.w .
, quasar astronomy , cambridge university press , 1986 , pages 62 - 64 lccccc 000041.9 & -022923 & 1.07 & 7.2 & -26.3 & 32.1 000150.0 & -015940 & 2.82 & 37.1 & -27.4 & 33.7 000507.0 & -013246 & 1.71 & 73.1 & -25.9 & 33.5 000727.1 & 024112 & 0.30 & 8.0 & -23.5 & 31.1 002717.3 & 003723 & 1.23 & 4.6 & -26.1 & 32.1 003136.5 & -013622 & 2.38 & 16.6 & -26.7 & 33.2 003942.4 & -392914 & 1.46 & 4.6 & -25.9 & 32.2 003942.5 & -352802 & 0.84 & 92.2 & -25.2 & 33.0 004150.6 & -382009 & 1.74 & 71.2 & -24.8 & 33.5 004159.9 & -392147 & 1.85 & 4.9 & -26.7 & 32.4 004216.5 & -025422 & 2.74 & 31.6 & -27.3 & 33.6 004240.7 & -333920 & 2.06 & 6.4 & -25.8 & 32.6 004259.9 & -300742 & 0.61 & 39.1 & -24.7 & 32.4 004322.4 & -263905 & 1.00 & 80.1 & -26.4 & 33.1 004351.3 & -282741 & 0.84 & 55.2 & -25.1 & 32.8 004810.3 & -400534 & 1.64 & 8.1 & -26.2 & 32.6 004940.8 & -263259 & 3.16 & 39.0 & -25.9 & 33.8 005040.9 & -254124 & 0.78 & 43.7 & -24.5 & 32.7 005130.4 & 004150 & 1.19 & 16.3 & -25.8 & 32.6 005153.6 & -275618 & 1.15 & 6.8 & -23.8 & 32.2 005205.5 & 003538 & 0.40 & 135.3 & -25.3 & 32.6 005245.4 & -262423 & 2.22 & 5.6 & -25.0 & 32.6 005252.4 & -282555 & 1.65 & 226.6 & -26.1 & 34.0 005404.0 & -271010 & 0.69 & 7.0 & -23.9 & 31.8 005424.9 & -252400 & 1.95 & 112.5 & -24.7 & 33.8 005441.2 & 000110 & 0.65 & 4.4 & -25.0 & 31.5 005445.4 & -384415 & 3.18 & 29.0 & -26.9 & 33.7 005519.8 & -295933 & 2.10 & 40.1 & -25.3 & 33.5 005535.1 & 014035 & 0.44 & 26.8 & -24.2 & 32.0 005542.8 & -272319 & 1.54 & 9.9 & -25.1 & 32.6 005604.3 & -265257 & 1.04 & 21.7 & -25.2 & 32.6 005801.8 & 155315 & 1.26 & 6.1 & -26.3 & 32.2 005841.2 & -390842 & 1.41 & 8.6 & -26.8 & 32.5 005917.6 & 014206 & 3.15 & 6.5 & -27.8 & 33.0 005958.1 & -393556 & 2.04 & 10.2 & -25.7 & 32.8 010012.3 & -270853 & 3.52 & 210.9 & -27.4 & 34.6 010059.8 & -285115 & 0.87 & 8.8 & -25.4 & 32.1 010200.6 & -301827 & 1.03 & 78.7 & -26.6 & 33.2 010259.8 & -272343 & 1.42 & 3.8 & -25.2 & 32.1 010329.4 & 004054 & 1.44 & 117.9 & -25.5 & 33.6 010403.5 & -294353 & 1.32 & 13.1 & -24.1 & 32.6 010928.8 & -340518 & 0.83 & 26.8 & -23.3 & 32.5 010952.0 & -352556 & 2.19 & 2.8 & -25.0 & 32.3 011108.1 & -353517 & 2.36 & 15.0 & -26.4 & 33.1 011119.5 & -330822 & 2.31 & 3.1 & -26.0 & 32.4 011547.7 & -330715 & 1.91 & 80.1 & -24.3 & 33.7 011636.9 & -345701 & 1.96 & 107.5 & -25.7 & 33.8 012204.4 & -331012 & 1.11 & 17.9 & -26.7 & 32.6 012655.8 & -320810 & 2.20 & 4.5 & -27.5 & 32.5 012744.0 & -345754 & 2.16 & 18.8 & -26.5 & 33.2 013405.1 & -393215 & 1.89 & 12.1 & -24.8 & 32.8 013438.6 & -193207 & 3.13 & 47.4 & -28.4 & 33.9 013505.7 & -072029 & 1.64 & 6.8 & -26.1 & 32.5 013535.7 & -385425 & 2.08 & 182.0 & -25.3 & 34.1 013555.5 & 003924 & 1.46 & 4.7 & -24.2 & 32.2 014254.7 & -302345 & 3.12 & 5.0 & -28.4 & 32.9 014254.8 & -385646 & 2.21 & 4.0 & -25.6 & 32.5 015227.3 & -200107 & 2.15 & 26.9 & -27.7 & 33.3 015259.1 & -012940 & 2.02 & 47.7 & -26.8 & 33.5 015708.6 & 013824 & 2.13 & 4.8 & -26.7 & 32.5 015921.0 & 190032 & 2.61 & 21.9 & -26.1 & 33.4 021612.3 & -010519 & 2.15 & 144.0 & -27.0 & 34.0 024205.2 & -234453 & 2.63 & 29.5 & -26.4 & 33.5 024240.3 & 005727 & 0.57 & 8.3 & -25.8 & 31.7 024450.5 & -221601 & 2.20 & 24.0 & -24.9 & 33.3 024547.8 & -313725 & 1.88 & 13.6 & -26.3 & 32.9 024739.1 & -005222 & 2.12 & 7.9 & -25.0 & 32.8 024806.6 & -293636 & 1.86 & 3.5 & -25.3 & 32.3 025048.6 & 000208 & 0.77 & 112.7 & -24.1 & 33.1 025146.8 & 155013 & 0.49 & 30.4 & -23.1 & 32.1 025156.2 & 005706 & 0.47 & 7.5 & -23.4 & 31.5 025658.5 & 005447 & 1.11 & 13.2 & -24.7 & 32.4 030029.8 & 024050 & 0.12 & 6.8 & -23.0 & 30.2 030138.2 & 013817 & 1.58 & 8.1 & -25.4 & 32.5 030548.9 & -024013 & 1.27 & 18.6 & -25.6 & 32.7 031002.0 & -372338 & 0.40 & 15.5 & -24.1 & 31.6 031139.8 & -400842 & 1.73 & 9.5 & -26.1 & 32.7 031631.3 & 013730 & 0.96 & 11.5 & -24.7 & 32.3 040728.8 & 175052 & 1.71 & 551.6 & -25.4 & 34.4 042241.8 & 003020 & 2.92 & 20.6 & -27.2 & 33.4 042935.5 & -224103 & 0.27 & 15.8 & -24.1 & 31.3 043848.1 & -363509 & 0.38 & 10.6 & -24.2 & 31.4 044122.6 & -270820 & 0.08 & 34.8 & -23.0 & 30.6 044314.3 & -364624 & 0.68 & 4.7 & -25.6 & 31.6 044415.2 & -300452 & 2.40 & 10.1 & -27.4 & 33.0 044921.1 & 072910 & 1.46 & 61.8 & -27.5 & 33.3 045230.1 & -295335 & 0.29 & 10.0 & -25.1 & 31.2 050455.3 & -272311 & 1.14 & 183.4 & -26.4 & 33.6 055813.5 & -361948 & 2.22 & 3.1 & -27.0 & 32.4 060607.2 & -344740 & 2.28 & 280.4 & -26.5 & 34.4 070713.2 & 643559 & 0.08 & 3.2 & -23.1 & 29.5 075100.7 & 032041 & 0.10 & 12.0 & -23.6 & 30.3 075819.8 & 421935 & 0.21 & 4.0 & -24.1 & 30.5 081520.6 & 273618 & 0.91 & 6.0 & -27.4 & 31.9 083658.9 & 442602 & 0.25 & 7.9 & -25.3 & 30.9 083840.5 & 473410 & 0.70 & 29.6 & -23.7 & 32.4 084118.0 & 354438 & 1.77 & 6.7 & -28.7 & 32.5 084215.2 & 452544 & 1.41 & 75.4 & -26.9 & 33.4 084756.4 & 314758 & 1.83 & 29.0 & -26.8 & 33.2 084847.8 & 142056 & 1.69 & 44.3 & -26.5 & 33.3 085516.1 & 561656 & 0.44 & 5.6 & -23.3 & 31.3 085658.8 & 512045 & 2.31 & 8.6 & -25.8 & 32.9 085706.3 & 190854 & 0.33 & 5.2 & -23.8 & 31.0 090603.6 & 194144 & 1.21 & 17.4 & -27.2 & 32.6 092837.9 & 602521 & 0.30 & 67.8 & -23.1 & 32.0 092958.4 & -260031 & 2.15 & 2.8 & -27.1 & 32.3 093235.0 & 215830 & 2.31 & 5.9 & -25.4 & 32.7 093857.2 & 424829 & 2.04 & 6.5 & -26.2 & 32.6 094128.8 & 444743 & 0.80 & 69.3 & -24.6 & 32.9 094704.5 & 472143 & 0.54 & 4.0 & -24.2 & 31.3 094756.0 & 535001 & 0.49 & 23.1 & -25.2 & 32.0 094835.9 & 432302 & 1.89 & 2.8 & -26.7 & 32.2 095227.2 & 504851 & 1.55 & 72.3 & -26.7 & 33.5 095538.0 & 333504 & 2.50 & 34.1 & -28.5 & 33.5 100254.6 & 324039 & 0.83 & 11.5 & -26.6 & 32.1 101703.4 & 592429 & 0.85 & 340.5 & -25.9 & 33.6 102809.6 & -264418 & 2.90 & 933.8 & -28.3 & 35.1 103808.1 & 473136 & 2.96 & 8.6 & -25.7 & 33.1 105333.4 & -273907 & 0.16 & 12.9 & -23.8 & 30.8 110325.3 & -264515 & 2.15 & 15.8 & -28.9 & 33.1 112241.5 & 303535 & 1.81 & 10.4 & -27.7 & 32.7 112736.4 & 265450 & 0.38 & 3.2 & -24.7 & 30.9 113004.8 & 411619 & 0.72 & 3.4 & -26.0 & 31.5 113314.8 & 281200 & 0.51 & 12.9 & -24.6 & 31.8 114116.1 & 215622 & 0.06 & 6.1 & -23.0 & 29.6 114800.0 & 263544 & 0.87 & 351.4 & -25.3 & 33.7 114818.8 & 315411 & 0.55 & 94.7 & -25.0 & 32.7 115227.5 & 320959 & 0.37 & 41.7 & -23.6 & 32.0 115734.7 & 731037 & 0.51 & 9.5 & -25.7 & 31.6 120243.3 & 373550 & 0.50 & 16.7 & -24.2 & 31.9 121209.4 & 462346 & 2.29 & 4.8 & -25.6 & 32.6 121721.3 & 305631 & 0.31 & 8.4 & -24.2 & 31.1 121736.6 & 515511 & 2.23 & 85.2 & -28.6 & 33.8 121943.6 & 540835 & 1.66 & 45.0 & -26.1 & 33.3 123322.9 & 473843 & 2.23 & 24.3 & -25.0 & 33.3 123714.4 & 261900 & 2.20 & 3.9 & -23.4 & 32.5 131021.1 & 473403 & 3.58 & 9.2 & -24.2 & 33.3 131104.3 & 292613 & 1.82 & 10.4 & -27.4 & 32.8 131211.0 & 480925 & 0.72 & 45.0 & -25.4 & 32.6 131801.9 & 470627 & 2.59 & 2.8 & -27.3 & 32.5 131850.9 & 264142 & 1.91 & 23.1 & -23.8 & 33.1 131913.8 & 264754 & 2.26 & 8.2 & -25.1 & 32.8 132250.7 & 473937 & 1.11 & 8.1 & -25.0 & 32.2 133602.8 & 272747 & 1.12 & 75.5 & -24.5 & 33.2 134208.3 & 270923 & 1.19 & 270.8 & -25.0 & 33.8 134425.6 & 261410 & 1.18 & 11.2 & -24.5 & 32.4 134543.8 & 262507 & 2.03 & 9.5 & -24.6 & 32.8 134544.5 & 262506 & 2.03 & 9.5 & -25.0 & 32.8 142314.3 & 505539 & 0.28 & 180.5 & -24.3 & 32.4 144300.6 & 520137 & 1.57 & 2550.8 & -24.5 & 35.0 144920.7 & 422102 & 0.18 & 160.1 & -23.1 & 31.9 145005.0 & 463519 & 0.29 & 4.6 & -23.1 & 30.8 151907.5 & 520605 & 0.14 & 11.1 & -23.3 & 30.6 152043.6 & 473249 & 2.81 & 87.5 & -26.2 & 34.0 152907.4 & 561607 & 0.10 & 6.0 & -23.0 & 30.0 152917.6 & 473838 & 3.07 & 55.5 & -26.2 & 33.9 153045.8 & 383953 & 2.02 & 3.0 & -28.0 & 32.3 153457.2 & 583924 & 1.90 & 129.5 & -25.9 & 33.9 154058.6 & 473828 & 2.56 & 230.0 & -26.0 & 34.4 155610.6 & 374040 & 2.66 & 9.2 & -26.9 & 33.0 162548.8 & 264659 & 2.52 & 6.9 & -29.6 & 32.8 163020.8 & 375656 & 0.39 & 23.8 & -24.8 & 31.8 163249.6 & 371631 & 2.94 & 3.2 & -27.6 & 32.6 164501.8 & 392548 & 2.15 & 3.0 & -26.0 & 32.3 165005.5 & 414033 & 0.59 & 234.7 & -25.2 & 33.1 165641.1 & 591539 & 0.50 & 13.0 & -23.7 & 31.8 175105.5 & 265902 & 0.14 & 146.4 & -24.5 & 31.7 183023.3 & 731310 & 0.12 & 9.1 & -23.8 & 30.4 | using the nvss radio catalog , we have searched for radio emission from 4079 quasars taken from the 1996 version of the veron - cetty and veron ( 1996 ) quasar catalog .
the comparison resulted in the positive detection of radio emission from 799 quasars .
of these , 168 are new radio detections .
examination of the radio luminosities shows a dramatic increase in the fraction of radio - loud quasars from the current epoch to z=0.5 and a gradual decline beyond z=1.0 .
inspection of the radio - loud fraction as a function of m@xmath0 shows little dependence fainter than m@[email protected] .
-0.4 in 0.0 in 0.0 in 9.0 in 6.5 in = cmsy7 = cmr7 = cmr5 = cmr5 = = = # 1#1 # 1 # 1 # 1to 0pt#1 3@xmath2 2 @xmath3 3mpch^-3mpc^3 0@xmath4 0@xmath5 # 1 # 110^#1 # 1to 0pt#1 version : 10 feb 1997 |
among all the processes manifested in stellar coronae , flares are the most energetic ones .
they are supposed to be the result of the energy release from magnetic field reconnection in the lower corona ( e.g. * ? ? ?
as a consequence , electrons and ions in the reconnection region are accelerated downwards , along the magnetic field lines , toward lower atmospheric layers . when they reach the upper chromosphere , the local gas is heated and evaporated into the new - formed magnetic loops .
thus , the density and temperature of these loops increase , causing intense emission of soft ( @xmath4 kev ) x - rays .
stellar flares have been observed in almost all the diagram ( see * ? ? ? * for a review ) .
however , they are more frequent in late - type stars , where they present a large variety of sizes and durations . in late - k and m dwarfs ( the so - called uv ceti - type stars ;
see * ? ? ?
* for a description of their main properties ) , moderate flares are frequently observed . during such events ,
the x - ray flux usually increases by a factor of ( e.g. * ? ? ?
giant flares , in which the x - ray flux increases from dozens to hundreds times the quiescent state value , have been detected in some m dwarfs such as @xcite , @xcite , and @xcite .
the duration of those flares are of the order of a few kiloseconds , even in the more energetic ones .
for instance , the giant flare observed in lasted @xmath5 ks .
long - duration flares are more common in pre - main sequence stars than in the uv ceti - type ones .
for example , during the 13 days observing run of the orion nebula complex by _ chandra _ , at least 19 flares with durations above half a day were detected @xcite .
long - duration flares were also observed in the taurus star - forming complex @xcite . the longer duration of these events is usually attributed to the presence of larger coronal structures in young stars .
the _ xmm - newton _ and _ chandra _ missions have contributed enormously to the understanding of the processes involved in the x - ray emission of late - type stars . in particular , the improved temporal and spectral resolution , together with the development of theoretical models ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , have provided us with powerful tools for investigating coronal flares .
detailed diagnostics of x - ray flares have been carried out by different authors .
@xcite determined general properties of flaring loops , in terms of temperature and semi - length , for young stellar objects in the orion nebula complex ( onc ) .
> from their analysis , the authors inferred loop semi - lengths comparable to the stellar radius in some cases . they speculated that these large structures are connected with the proto - planetary disk and are , in fact , the same structures that channel the plasma producing accretion . a similar work was done for the taurus star - forming region @xcite , where the authors remarked that some coronal loops extended up to a distance comparable with the stellar radius ( see also * ? ? ?
studies for more evolved stars were done by , e.g. , @xcite , @xcite , and @xcite , who found flaring loops with semi - lengths of @xmath6 @xmath7 ( i.e. , a relatively compact flaring corona ) . in the works presented in the previous paragraph
, the diagnostic of x - ray flares was done using the procedure described by @xcite to model the decay phase , which assumes the flare to be produced in a single loop where heating does not entirely drive the flare decay .
@xcite extended this method to the rise phase and compared the parameters determined in this way with those obtained by analyzing the decay phase for three stellar flaring loops ( see also * ? ? ?
* ) , finding a good agreement between them .
on the other hand , the solar two - ribbon model developed by @xcite , or its stellar version @xcite , should be used when heating totally drives the flare evolution @xcite . in this model ,
the reconnection energy is supposed to be dissipated immediately after being released .
@xcite assumed that only a fraction of the magnetic energy released by the reconnection process is used to supply the thermal energy of the newly formed flare loops . @xcite
assumed that a factor of @xmath8 of this thermal energy escapes into the x - ray regime , as suggested by detailed studies of solar flares @xcite .
we also refer the reader to @xcite , who included time - dependent conductive and radiative losses in the x - ray corona self - consistently . in this work
, we analyze an xmm - newton serendipitous observation of the young m - type star twa 11b in which the x - ray emission suffered a continuous increase during approximately 35ks .
the duration and statistics of the observed rise have allowed us : ( i ) to carry out a detailed spectral time - analysis ; and ( ii ) to derive properties of the star s magnetic configuration by using both the single - loop and the two - ribbon flare models described above .
the xmm - newton observation ( i d 0006220201 ) was perfomed in the revolution 197 , between 2001 january 4 and 5 , for a total duration of @xmath9ks .
the field is centred on the coordinates @xmath10 and @xmath11 ( the main target being the seyfert 2 galaxy ) .
the epic cameras were operated in full frame mode using the thick filter .
our target is situated 5.18arcmin at the north - east of the central source , in the field of view of both the mos and pn detectors .
the epic - xmm source coincides with the position of the optical counterpart of twa 11b .
the primary star , twa 11a ( an a0-type star ) , is situated at @xmath0 arcsec from twa 11b .
although none x - ray emission is expected from an a0 star , we investigated the possibility that the primary star produced some x - rays that could affect our results .
we studied a serendipitous detection of hetg chandra of the source ( observation i d 2150 ) . in this 140 ks exposure chandra observation ,
the x - ray source is clearly identified with twa 11b ( see fig .
[ fig0 ] ) , while none x - ray emission coming from twa 11a is detected .
therefore , we are confident that the x - ray emission detected in the xmm - newton observation comes only from the m star twa 11b .
the data reduction followed the standard operating procedure .
we used the version 7.1.0 of the xmm - newton science analysis system ( sas ) to derive a table of calibrated events in the energy range 0.3 10.0kev . to extract the events
, we chose a radius of 30arcsec , which is slightly larger than the 3@xmath12-level of the source s psf .
this assures us to lose less than 1% of the counts from the source .
different filters were applied to eliminate bad events and noise .
note that the observation was neither affected by pile - up nor by high flaring background periods .
the x - ray light curve ( fig .
[ fig1 ] ) showed a total increase in the star s count - rate of a factor of @xmath13 from the lower level ( @xmath14 counts / s ) to the maximum observed emission .
similar relative increases in flux were previously reported in flares from other stars ( e.g. * ? ? ? * ) . a first increase in the light curve
is observed @xmath15ks after the beginning of the observation , reaching a local maximum only @xmath15ks later . after a brief decrease ,
the count - rate continues increasing until reaching the global maximum in the light curve ( this happened 37ks after the beginning and only 3ks before the end of the observation ) .
the total duration of the enhancement is @xmath16ks ( @xmath17 hours kms@xmath18 , what leads to an upper limit in the rotational period of 2.7 days .
thus , the observation cover , at least , 17% of the rotational period of the star . nevertheless , no indication of occultation of the flaring region is observed . ] ) .
note that the observed maximum may or may not be the flare peak . in the latter case , the duration of the rise phase would be even longer . in fig .
[ fig2 ] we plot the evolution of the hardness ratio , which is a tracer of the temperature evolution , during the observation . here
, the soft energy band is defined as the range 0.3 0.8 kev and the hard energy band as 0.8 4.5 kev .
[ fig2 ] shows that the mean coronal temperature reached a maximum at @xmath19 ks , maintained high with significant fluctuations during approximately the next 27ks of exposure , and then it began to decrease gradually .
overall , the rise phase is unusually long compared with those observed in other stars ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
to date , so long rise phases have been observed only in some t tauri stars @xcite , some of them with accretion disks , and several rs cvn systems @xcite . in such stars ,
the relative increase in flux ranges from a factor 3 ( v410 tau ; * ? ? ?
* ) to a factor 10 ( v892 tau ; * ? ? ?
* ) with rise - phase times from 8 to 55 ks .
twa 11b is a weak - line t tauri m2.5 star with an age of @xmath0 myr ( estimated from its membership in the tw hya association ) . while it shows no signatures of an accretion disk ( near - infrared color excess or strong h@xmath20 emission ; * ? ? ?
* ) , it is still contracting .
thus , physical conditions in its atmosphere should be more similar to those of sub - giant stars than of main - sequence ones
. this may be the reason why long duration rise phase flares are observed in both t tauri stars and rs cvn systems .
a more detailed inspection of the light curve shows that the emission increased faster initially , it had a well - defined local maximum at @xmath21ks , and then it growed again , but more gradually this time , until the end of the observation .
at the very end , there is a hint that the emission was stopping to increase .
such a long rise phase suggests that we observed an uninterrupted sequence of flare events involving an extended coronal region .
however , the faster initial rise and the local emission to the earlier hardness ratio the evolution of a self - standing flaring episode possibly occurring ( at least during these initial stages ) in a single loop ( e.g. * ? ? ?
* ) , whose decay merged with the ignition of the later overlapping events .
[ fig2.1 ] demonstrates that this peak is indeed significative and can not be attributed to the effect of noise in the light curve .
this figure shows the cumulative distribution of counts for 200 simulated constantly increasing light curves ( for @xmath22ks ) with poissonian noise ( shadowed region ) together with the cumulative distribution of counts observed during the rise ( continuous line ) .
clearly , the simulations do not reproduce the observed distribution of counts for the first peak in the light curve .
the differences are high enough to justify the treatment of this event as independent . in the following , we work under this hypothesis .
ks ( continuous line ) .
the shadowed region contains cumulative curves of 200 simulated constantly increasing light curves ( also for @xmath22ks ) with poissonian noise.,width=302 ] the large number of counts collected from the source in the epic cameras allowed us to divide the observation in several time intervals with enough signal to perform a reliable spectral analysis ( @xmath23 counts in the quiescent and intervals 1 and 2 , and @xmath24 counts in the remaining intervals after background subtraction ) .
this permitted : 1 ) to investigate the nature of the processes taking place in our target ; and 2 ) to derive physical properties of it .
for the spectral analysis , the observation was split into 10 time - intervals ( vertical dashed lines in fig .
[ fig1 ] ) .
these intervals sample different features in the light curve .
we used the xspec spectral fitting package @xcite in the pn , mos1 , and mos2 detectors simultaneously .
we adopted the astrophysical plasma emission code ( apec , * ? ? ?
* ) included in the xspec software .
apec calculates spectral models for hot , optically thin plasmas using the astrophysical plasma emission database ( aped , * ? ? ?
* ) , that contains the relevant atomic data for calculating both the continuum and line emission .
interstellar absorption was taken into account using the interstellar photo - electric absorption cross - sections of @xcite , also available in xspec . the lowest ( constant ) count - rate level , found at the beginning of the observation , is assumed to be the quiescent state .
the x - ray luminosity of the star during this lapse of time ( assuming the same distance than the primary @xmath25 pc ) is @xmath26 } = 29.35 $ ] , corresponding to @xmath27 .
these values are typical of both pre - main sequence stars members of young stellar associations ( e.g. * ? ? ?
* ) and older field m dwarfs showing high x - ray activity @xcite .
the best fit to the x - ray quiescent spectrum ( fig .
[ fig3 ] ) is given by a plasma model with two equally - weighted thermal components ( the output parameters of this fit are summarized in table [ tab_quiescent ] ) . although high uncertainties are obtained for the hydrogen column density in our fit , the value of @xmath28 obtained by us is very similar to that determined for other members of the tw hya association ( e.g. * ? ? ?
* ; * ? ? ?
besides , the interstellar extinction was found to be negligible in the general direction of tw hya @xcite . lcl @xmath29 & = & @xmath30 @xmath31 + @xmath32 & = & @xmath33 @xmath34 + @xmath35 & = & @xmath36 kev + @xmath37 & = & @xmath38 @xmath39 + @xmath40 & = & @xmath41 kev + @xmath42 & = & @xmath43 @xmath39 + @xmath44 ( d.o.f . )
& = & @xmath45 + synthesized stellar - like spectra of solar flares ( previously subtracted by the quiescent `` background '' spectrum ) are generally with a single thermal component ( 1-@xmath46 model , see * ? ? ? * ) . consequently , to analyze physical properties of the flaring plasma in our observation , we first subtracted the spectrum of the quiescent state from the observed spectrum in each time - interval . then we fitted a @xmath47-model , leaving its temperature ( @xmath48 ) and emission measure ( @xmath49 ) as free parameters .
the values of the column density ( @xmath29 ) and abundance ( @xmath32 ) were fixed to those derived for the quiescent state . with this technique
, we interpret that the coronal spectrum during the flare results from adding to the quiescent spectrum a third thermal component , which is ascribed to heated material filling the flaring loops @xcite .
models with additional temperature components were checked , but did not improve the fit results significantly .
results from spectral fitting are given in table [ tab2 ] . as expected from the hardness ratio evolution ( fig .
[ fig2 ] ) , the temperature peaked in time - segment 1 .
then , it decreased gradually but almost continuously until the end of the observation . on the other hand ,
the emission measure increased continuosly from the beginning of the rise and starts to decrease only near the end of the observation .
c c c c c c c c time - segment & time - interval & central time & @xmath48 & @xmath49 & @xmath44 ( d.o.f . )
+ & ( ks ) & ( ks ) & ( kev ) & ( @xmath50 @xmath39 ) & + 1 & 5 8 & 6.5 & 8@xmath51 & 0.72@xmath52 & 1.05 ( 129 ) + 2 & 8 11 & 9.5 & 2.4@xmath53 & 4.0@xmath54 & 0.87 ( 215 ) + 3 & 11 17 & 14 & 2.5@xmath53 & 3.88@xmath55 & 1.08 ( 369 ) + 4 & 17 22 & 19.5 & 2.27@xmath56 & 5.3@xmath54 & 0.87 ( 355 ) + 5 & 22 27 & 24.5 & 1.88@xmath57 & 7.2@xmath54 & 0.89 ( 406 ) + 6 & 27 32 & 29.5 & 1.66@xmath58 & 8.3@xmath54 & 1.03 ( 430 ) + 7 & 32 36 & 34 & 1.31@xmath59 & 9.2@xmath60 & 0.97 ( 379 ) + 8 & 36 38 & 37 & 1.22@xmath61 & 9.9@xmath62 & 1.01 ( 243 ) + 9 & 38 41 & 39.5 & 0.99@xmath63 & 9.7@xmath60 & 0.86 ( 313 ) + in order to approximate physically more realistic continuous emission measure distributions ( emds ) of the plasma , we used a multi - temperature model as in @xcite . in that work , the authors used a @xmath64-model on a logarithmic , almost equidistant grid with temperatures fixed at 0.2 , 0.3 , 0.6 , 1.2 , 2.4 , and 4.8 kev ( which correspond to 2.3 , 3.5 , 7.0 , 14.0 , 28.0 and 56.0 mk ) , sampling those spectral regions where the xmm - newton detectors are more sensitive . to fit our spectra ,
the values of @xmath29 and @xmath32 were fixed to those previously determined for the quiescent state .
thus , the only variables are the emission measures .
note that in this case we are fitting each spectrum as a whole ( i.e. , the quiescent spectrum was not subtracted from the rest of the spectra ) .
results from applying this @xmath64-model to our observation are shown in table [ tab2.1 ] and fig .
[ fig4 ] , and are summarized in the following items : ccccccc time interval & @xmath65 & @xmath66 & @xmath67 & @xmath68 & @xmath69 & @xmath70 + quiescent & @xmath71 & @xmath72 & @xmath73 & @xmath74 & @xmath75 & @xmath76 + 1 & @xmath77 & @xmath78 & @xmath79 & @xmath80 & @xmath81 & @xmath82 + 2 & @xmath83 & @xmath84 & @xmath85 & @xmath86 & @xmath87 & @xmath88 + 3 & @xmath89 & @xmath90 & @xmath91 & @xmath92 & @xmath93 & @xmath94 + 4 & @xmath95 & @xmath96 & @xmath97 & @xmath98 & @xmath99 & @xmath100 + 5 & @xmath101 & @xmath102 & @xmath103 & @xmath104 & @xmath105 & @xmath106 + 6 & @xmath107 & @xmath108 & @xmath109 & @xmath110 & @xmath111 & @xmath112 + 7 & @xmath113 & @xmath114 & @xmath115 & @xmath116 & @xmath117 & @xmath118 + 8 & @xmath119 & @xmath120 & @xmath121 & @xmath122 & @xmath123 & @xmath124 + 9 & @xmath125 & @xmath126 & @xmath127 & @xmath128 & @xmath129 & @xmath130 + * the amount of plasma emitting at temperatures above 28mk is negligible in the quiescent state .
* plasma emitting at the highest temperatures ( @xmath131 56mk ) appears at the beginning of the first ( faster ) rise ( time - segment 1 ) , while the rest of the emd curve remains as in the quiescent state .
note that in this time - segment we obtained the maximum temperature for the flaring component when fitting it with the @xmath47-model ( see first part of @xmath132 [ sec : flaring ] and table [ tab2 ] ) . *
the only significant difference between the emd of the quiescent state and that derived for time - segment 2 ( where the light curve of the first flare peaks ) is an excess of plasma emitting at temperatures around 28mk .
we interpret it as the cooling of the material at higher temperatures that was detected in time - segment 1 together with additional plasma evaporated after the temperature peak was reached in time - segment 1 ( note that the total emission measure of the flaring component in time - segment 2 is higher than that measured in time - segment 1 , as it can also be seen in table [ tab2 ] ) . *
the excess of plasma emitting in the region around 28mk that was observed in time - segment 2 continues cooling towards lower temperatures during time - segment 3 . at the same time
, a high quantity of plasma appears again at the highest temperatures ( @xmath131 56mk ) .
this quantity is even larger than that measured in time - segment 1 and is approximately coincident with the beginning of the second ( more gradual ) rise . * during time - segments 4 7 , the amount of plasma emitting at the highest temperatures remains still high , while a large excess of material ( compared to the quiescent state ) appears also between 7 and 28mk ( reaching the maximum always at a temperature @xmath133 14mk ) .
the existence of intense , sustained heating is needed to explain the maintenance of plasma at temperatures around 56mk . the evolution of the rest of the emd curve can be interpreted as the cooling of the continuously new appearing very hot plasma together with heated material that evaporates from lower layers and fills the flaring loops . * during time - segments 8 and 9 the emission at temperatures @xmath134mk turns back to the level found at the quiescent state , which is an indication of the strong heating not being present any more .
however , the height of the maximum in the emd curve is even larger than in the previous time - segments , appearing also excess emission at even lower temperatures ( down to @xmath135mk ) .
the curves are shifted in the y - axis to better observe the flare evolution in time .
each one of the emds ( shown as continuous lines ) is labeled with the same number as its corresponding time - segment ( see table [ tab2 ] ) .
the quiescent is marked as q. for comparison , the emd of the quiescent state is also overplotted as a red , dashed line at the same zero level than each one of the rest emds.,width=309 ]
results from @xmath132 [ sec : observations ] and @xmath132 [ sec : specanalysis ] support the idea of the observed rise ( @xmath22ks ) being the consequence of two separate flare events : the first more impulsive one ( event a ) starting in time - segment 1 , followed by a more gradual energy release ( event b ) having a beginning that is merged with the decay phase of event a and becoming dominant at the end of time - segment 3 .
but , what kind of magnetic structures are producing each one of these events ? according to @xcite , multiple loop structures can be involved in a flare , but this frequently occurs only in its late phases .
the initial phases of an x - ray flare are usually quite localized and one can reasonably assume the presence of a single dominant loop ( e.g. * ? ? ?
* ; * ? ? ?
this assumption is realistic enough in most of the observed rise phases , where the impulsive heating typically involves a dominant loop ( while later residual heating may be released in other similar adjacent loops ) . the clear evidence of a delay between the temperature peak and the density peak is consistent with a single loop model , at least until the moment in which this maximum density is reached .
this delay is often observed both in solar flares , and in stellar flares , even in very long ones @xcite .
the presence of this delay is indeed a signature of a relatively short heat pulse ( the larger the delay , the shorter the heat pulse is ) and of the coherent hydrodynamic evolution of plasma confined in a loop .
arcades and two - ribbon flares are instead characterized by strong and/or lasting heating @xcite and/or irregular light curves @xcite .
thus , event a is likely characterized ( or dominated ) by a single flaring loop since its ignition occurs only at time - segment 1 ( as suggested by the plasma appearing at very high temperatures in its corresponding emd , see @xmath132 [ sec : emds ] ) and shows a delay between the temperature peak and the density peak reached by the flaring plasma in it ( see table [ tab2 ] ) .
event b is probably triggered by event a. the evolution of the whole emd during event b evidences continuous , but decreasing in efficiency , heating of material and is consistent with event b being a two - ribbon flare . in @xmath132 [ sec52 ]
we demonstrate that the light curve of event b is indeed well - reproduced by a two - ribbon flare model .
continual reconnections in an arcade may also be responsible for event b. however , in this case one would expect a more irregular pattern in the light curve than that observed in fig . [ fig1 ] , and the emd evolution would not necessarily point to a kind of heating that is decreasing in efficiency with time , as we have observed .
we used the results from fitting the spectra with the @xmath47-model ( see @xmath132 [ sec : specanalysis ] ) , given in table [ tab2 ] , to determine the length of the loop involved in event a. for the temperature in time - segment 1 , we used @xmath136 kev ( the lower limit from the fit ) since its upper limit is undetermined ( see table [ tab2 ] ) . for this event ,
the complete rise phase is observed and instead we miss satisfactory information from the decay phase .
a complete theoretical analysis of the rise phase of a flare occurring in a single loop is given by @xcite .
the author showed that the loop half - length ( @xmath137 ) can be determined from the maximum temperature reached by the flaring plasma ( @xmath138 ) and the temperature and time in which the density peaks ( @xmath139 and @xmath140 , respectively ) by : @xmath141 where all the parameters are given in c.g.s . units .
the maximum temperature ( @xmath138 ) is related to the temperature measured for the flaring component by spectral fitting ( @xmath142 ) . for the epic instrument : @xmath143 for event a , the maximum temperature and maximum emission measure were reached in the time - segments 1 and 2 , respectively . for a single loop description ,
the ( square root of the ) emission measure becomes a good proxy of the density , because the loop volume presumably does not change much during the event . in this case , from eq .
[ eq1 ] we obtain a loop half - length @xmath144 cm , i.e. @xmath145 @xmath7 ( assuming a stellar radius @xmath146 @xmath147 from the pre - main sequence models of * ? ? ?
* for an m2.5 star with 8 myr , such as twa 11b is supposed to be ) .
although this value is relatively large , similar loop sizes have already been derived for young stars in star - forming regions @xcite . as far as we are concerned
, this would be the first time that such a relatively long loop is detected in a star older than 3 4myr .
this loop would fill a volume @xmath148 @xmath149 ( see * ? ? ? * for further details on the relations used for determining this quantity ) . assuming a semi - circular geometry , its aspect ( @xmath150 , with @xmath151 being the loop cross - section ) would be of the order of 2 % , which is quite lower than that observed in the sun ( @xmath152 % ) , but compatible with the results in @xcite . in table
[ tab3 ] we summarize the main parameters of the loop involved in event a , i.e. those already mentioned and maximum density at the loop apex ( @xmath153 ) , average density in the loop when the maximum density is reached at the apex ( @xmath154 ) and cross - section area of the loop ( @xmath155 ) .
such parameters were determined with the relations given in @xcite .
comparing our results with those obtained by @xcite for events observed in algol , ab dor , and prox cen , we conclude that : 1 ) our densities seem to be similar to those determined in the algol and prox cen flares ; and 2 ) the loop volume and its subtended area are of the same order of magnitude as those found for prox cen , another m star .
note that the largeness of some of the errors shown in table [ tab3 ] is a consequence of the propagation of errors in the equations .
llc parameter & & value + @xmath153 & @xmath156 @xmath39 & @xmath157 + @xmath154 & @xmath156 @xmath39 & @xmath158 + @xmath159 & @xmath160 @xmath161 & @xmath162 + @xmath155 & @xmath163 @xmath164 & @xmath165 + @xmath151 & @xmath156 cm & @xmath166 + @xmath137 & @xmath167 cm & @xmath168 + for the sake of completeness , we made the exercise to repeat the same analysis to the entire flare event ( a@xmath169b ) as if it all occurred in a single loop . for a flare temperature peaking in time - segment 1 ( @xmath170 ks ) , and the density peaking in time - segment 8 ( @xmath171 ks ) , using eq .
[ eq1 ] , we obtained @xmath172 cm , i.e. @xmath173 .
obviously , such a long loop would be easily destroyed by the stellar rotation .
clearly , this result makes no sense . for the study of event
b , we used the two - ribbon flare model by @xcite extended to the stellar case @xcite .
this model supposes that a disruptive event opens a loop arcade , being the open field lines then driven toward a radial neutral sheet ( above the magnetic neutral line ) where they reconnect at progressively higher altitudes .
thus , the continuous heating provided by these reconnections is capable of reproducing the temporal profile of the energy rate released during both the rise and decay phases of a two - ribbon flare . by analogy to that observed on the sun , the model considers that the arcade of loops is extended along the east - west direction ( i.e. , axial symmetry around the polar axis is assumed ) .
it also assumes that the magnetic field is potential between the stellar surface and the location of the neutral line and extends radially outwards from there . the magnetic field in the meridional planes of the arcade
can therefore be expressed in terms of a single lobe of a legendre polynomial of degree @xmath174 . note that : 1 .
each lobe is latitudinally bounded by radial magnetic fields .
2 . the arcade corresponds to one lobe axisymmetrically continued over some longitude in the east - west direction .
3 . through an appropriate choice of @xmath174 , one can find a lobe placed in the range of latitudes covered by the active region .
however , as spatial information is available only for the sun , stellar flaring regions are generally assumed to be centered on the equator for odd @xmath174 and to end at the equator for even @xmath174 .
as stellar observations can not provide any information on the time - dependent rise of the neutral point , it is assumed to mimic the solar case .
thus , it follows an exponential law of the form given by eq .
[ eq : neutralpoint ] , where @xmath175 is the height of the neutral point ( in units of @xmath7 , measured from the star s center ) , @xmath176 is the time ( measured from the beginning of the two - ribbon flare , @xmath177 ) , @xmath178 is a time - constant , and @xmath179 is the maximum height reached by the reconnection point during its upward movement ( measured from the star s surface ) .
@xmath179 is typically chosen to be equal to the latitudinal extent of the arcade , which is in turn linked with @xmath174 ( see eqs .
[ eq : hmn1 ] and [ eq : hmn2 ] ) .
@xmath180 under all these assumptions , the rate of magnetic energy released by the reconnecting arcade per radian of longitude ( @xmath181 ) can be expressed as : @xmath182}{[n + ( n+1 ) y^{2n+1}]^3 } \left ( \frac{dy}{dt } \right)\end{aligned}\ ] ] where @xmath183 evaluated between the latitudinal borders of the lobe , @xmath184 is the legendre polynomial of degree @xmath174 , @xmath185 is the co - latitude , @xmath186 is the co - latitude of either boundary of the lobe , and @xmath187 is the maximum surface magnetic field in the active region .
the factor @xmath188 merely defines the normalization of the energy release light curve , while @xmath178 and @xmath174 determine its shape . as eq .
[ eq : dedt ] is given per radian of longitude , a length ( @xmath189 ) must be assumed for the arcade in order to calculate its total energy - release rate .
solar two - ribbon flares occur in loop arcades whose length is typically about 1.5 times their width ( @xmath190 ) . in our study , we adopted this ratio for any given @xmath174 .
since , in the view of many authors , the initiation process for the flare itself might be the result of a more rapid ( nearly explosive ) reconnection than the reconnection process about which @xcite and @xcite speak , they stressed that their model is applicable only after the initial flare trigger mechanism is terminated .
in fact , the model seems not to be able to describe the time of impulsive heating and steeply increasing temperatures .
however , the impulsive phase typically ceases before reaching the @xmath191 flare peak level in the soft x - ray bandpass commonly used for stellar observations . at this time in our observations , the temperature is gradually decreasing .
hence , the model is applicable from time - segment 4 of the light curve to the end of our observations ( time - segment 3 is rejected also for avoiding possible contamination from the decay of event a ) . in order to apply the described model , the photospheric magnetic field in the flaring region ( from which @xmath187 can be determinated ) , and its latitudinal location and size ( which dictate @xmath174 )
should be known . in the solar case , observations provide all these data , whereas in the stellar case , at best , they can only be inferred indirectly . since we are unaware of the location , size , and magnetic field strength of the active region that we are studying , we treated @xmath174 and @xmath187 as free parameters to be determined from the best fit of the model to the observations .
we created a grid of values for the free parameters @xmath174 , @xmath178 , and @xmath177 to fit event b with the two - ribbon flare model ( the lower limit of @xmath177 was fixed at the beginning of the whole enhancement because of obvious physical reasons ) .
for each set of these parameters , we determined the @xmath187 that best fit the data by minimizing the @xmath192 value , which is defined as : @xmath193 where @xmath194 is the expected luminosity from the model ( @xmath195 ) , @xmath196 is the observed luminosity , @xmath197 is the error in the observed luminosity , and @xmath198 the number of time - intervals with which we fitted the two - ribbon model ( from time - segment 4 to 9 ) . following the solar analogy ( * ? ? ? *
see @xmath132 [ sec : int ] ) , we assumed the measured radiative losses in the x - ray band to be @xmath199 of the thermal energy generated as consequence of magnetic reconnections , that is @xmath200 ( see * ? ? ? * for details ) .
actually , only a fraction ( @xmath201 ) of the liberated magnetic energy is indeed used to heat the plasma ( thermal energy that is subsequently lost via radiation and conduction ) , while the rest is transformed into mechanical energy , into fast particles ejected from the corona , etc . at this point , we want to notice that we did not find a unique solution corresponding to a single set of parameters , but a number of solutions producing a good fit ( @xmath202 ) . in table
[ tab4 ] , we show the two - ribbon flare parameters for some of the good fits we have obtained .
we plot these results together with the observations in fig .
other good fits were found also for larger values of @xmath174 ( i.e. , active regions with smaller width and , therefore , shorter loops ) , but note that a smaller active region needs higher surface magnetic fields for reproducing a given energy - release rate .
thus , loop systems that reach larger altitudes i.e. , small values of @xmath174 may be more realistic in our case , although other loop configurations can not be excluded ( see @xmath132 [ s : fluor ] ) .
( see table [ tab4 ] ) .
note that each set of parameters in table [ tab4 ] produces similar results in the fitted region while having different slopes in the previous part and in the decay .
the light curve segments used to fit the two - ribbon model ( time - segments 49 , see also fig . [ fig1 ] ) is shown with a solid line .
the dashed line is the model .
we have also plotted the average values of the model in each time - interval during event b ( dotted line ) for a clearer comparison between the observed light curve and the model . [ fig5],width=313 ] lcccc polynomial degree & 2 & 3 & 5 & 10 + region width [ deg ] & 90@xmath203 & 53@xmath203 & 33@xmath203 & 17@xmath203 + @xmath179 [ r@xmath204 & 1.57 & 0.90 & 0.57 & 0.30 + @xmath205 [ r@xmath204 & 2.46 & 1.41 & 0.89 & 0.47 + @xmath177 [ ks ] & 7.0 & 6.5 & 6.0 & 5.5 + @xmath178 [ ks ] & 247 & 192 & 183 & 174 + @xmath206 [ g ] & 440 & 730 & 1050 & 1830 + v@xmath207 [ km s@xmath18 ] & 2.8 & 2.1 & 1.4 & 0.8 + @xmath209 [ @xmath210 @xmath39 ] & 1.1 & 2.6 & 5.1 & 13.3 + @xmath192 & 1.02 & 1.02 & 1.04 & 1.04 + the values given in table [ tab4 ] for the maximum surface magnetic field were determined assuming that all the magnetic energy is used to heat the plasma that fills the flaring loops ( @xmath212 ) .
@xcite used @xmath213 for a solar flare . on the other hand , during the analysis of a stellar flare , @xcite found acceptable solutions with @xmath214 for @xmath215 and @xmath216 for @xmath217 . for event b we found that small values of @xmath218 ( @xmath219 ) imply very strong photospheric magnetic fields in the flaring region ( @xmath187 @xmath220 kg ) . using @xmath221 , we obtained @xmath187 @xmath222 kg for @xmath215 and @xmath187 @xmath223 kg for @xmath224 .
during the inspection of the x - ray spectrum in the different time - intervals , we observed a feature ( excess emission not reproduced by the plasma model ) close to the fe k@xmath225 line at redder wavelengths .
we identified this feature as the fe fluorescent line at 6.4 kev .
it was noticeable only during the time - segments 4 to 9 , while it was neither in time - segments 1 to 3 nor in the quiescent state ( see fig .
[ fig6 ] ) .
unfortunately , we could not monitor the time - variation of this emission line since the individual spectra of each time - interval have no counts enough for accurately fitting the line .
thus , to quantify the excess emission at 6.4 kev in our observations , we performed spectral fitting using an interstellar absorbed 3@xmath46-plasma model and an additional gaussian line component at 6.4 kev .
we fitted , on the one hand , the spectrum integrated from the quiescent to time - segment 3 ( inclusive ) and , on the other hand , the spectrum integrated from time - segment 4 to 9 ( inclusive ) . the gaussian flux obtained in this way for the former spectrum is zero .
the best fit results are shown in table [ tab5 ] .
note that we left free only the temperature and the emission measure of the third ( hottest ) thermal component and the gaussian , while the rest of the parameters were fixed to the values obtained for the quiescent state ( table [ tab_quiescent ] ) .
lcl @xmath226 ( kev ) & = & @xmath227 + @xmath228 ( @xmath50 @xmath39 ) & = & @xmath229 + gaussian central energy ( kev ) & = & @xmath230 + gaussian flux ( @xmath231 ph@xmath31s@xmath18 ) & = & @xmath232 + gaussian @xmath12 ( ev ) & = & @xmath233 + @xmath44 [ d.o.f . ] & = & @xmath234 $ ] + in the past , fluorescent fe emission were commonly observed in classical t tauri stars and protostars @xcite , where it has been attributed to the incidence of x - ray emission onto the proto - planetary disk or into the circumstellar gas surrounding very young objects .
however , the fe 6.4 kev line has been observed also in the giant star hr 9024 @xcite and the rs cvn system ii peg @xcite . in the case of hr 9024
, the authors attributed its presence to the incidence of hard x - rays onto the photosphere , in concordance with what is observed in the sun . in ii
peg , the excitation mechanism was ascribed to electron impact ionization of photospheric fe . in our observation
, we have not statistics enough to perform a robust analysis of the fluorescent line .
nevertheless , some constraints can be given .
on the one hand , the line is not observed during the first event .
this may be consistent with the scenario of the long loop for this event .
the efficiency of the fluorescence decreases with the distance of the x - ray source to the photosphere ( e.g. * ? ? ? * ) .
therefore , if the fluoresce line were produced by photoionization , it should not be observed during flares occurring in long loops . on the other hand , the presence of the line during the second event could be indicative of a not very high loop system , if it were produced by photoionization .
in contrast , if the line were produced by collisional ionization , there would be none constraint to the loop height .
a result in favor with the non - photoionization nature of the fluorescence line is shown in @xcite . in their study
, the authors modeled the illuminating input spectrum and obtained line fluxes below the observations . with our data ,
the collisional production of the fluorescence line can not be discarded .
in this paper , we analyzed the long rise phase of a flare observed in an xmm - newton archive data of the @xmath0 myr old star twa 11b .
the analysis of the light curve , of the hardness - ratio curve , and of the time - resolved spectra consistently indicates that probably the flare first involved mainly a single loop and then propagated to a loop arcade becoming a proper two - ribbon flare .
we split our analysis into three parts : the quiescent state , the single - loop flare ( event a ) , and the two - ribbon system ( event b ) .
event a was studied with the analysis described in @xcite . for event
b , we used the stellar version of the @xcite s solar two - ribbon flare model that is able to provide some limited information about the late flaring structures .
for the single loop , we obtained a semi - length of approximately @xmath235 cm ( @xmath145 r@xmath236 ) , with a volume @xmath237 @xmath149 and a cross - section @xmath238 .
these values are comparable to those found by @xcite in orion members .
this fact makes us suggest that large and thin loops are common in young active stars . for the two - ribbon system , different results consistent with the observed data were found .
good fits ( @xmath202 ) were obtained for both small and large values of @xmath174 ( i.e. for high and short loop arcades ) .
bearing the semi - length of the first loop in mind , the more realistic scenario for the two - ribbon system is that with long loops . in any case , the estimated values of the maximum surface magnetic field in the flaring region result to be quite strong , reaching 2 8 kg when a large fraction of the liberated magnetic energy ( @xmath221 ) is assumed to heat the plasma , and up to 18 kg when @xmath239 . during the inspection of the x - ray spectrum
, we observed the fe fluorescent line at 6.4 kev during the time in which the two - ribbon system evolved ( time - segments @xmath240 ) .
the absence of the line during the first event may be consistent with the long loop scenario for event a. in contrast , the detection of the fluorescence line during the second event could be indicative of a not very high loop system involved in the two ribbon flare , if the line were produced by photoionization .
otherwise , there would be none constraint to the loop system height .
jls acknowledges financial support by the pricit project s-0505/esp-0237 ( astrocam ) of the comunidad autnoma de madrid ( spain ) .
icc acknowledges support from the spanish _ ministerio de educacin y ciencia _ , under the grant f.p.u . ; and from the marie curie actions grant merg - ct-2007 - 046535 .
the madrid group acknowledges partial support by the programa nacional de astronoma y astrofsica of the spanish ministerio de educacin y ciencia ( mec ) , under grants aya2008 - 00695 and aya2008 - 06423-c03 - 03 .
fr acknowledges support from italian ministero delluniversit e della ricerca ( miur ) .
gm acknowledges support from the agenzia spaziale italiana ( asi ) and the istituto nazionale di astrofisica ( inaf ) under grant asi - inaf i/088/06/0 .
we would like to thank the referee and the editor for useful comments and discussion that helped us to improve this manuscript . | we analyzed a long duration flare observed in a serendipitous xmm - newton detection of the m star cd-39 7717b ( twa 11b ) , member of the young stellar association tw hya ( @xmath0 myr ) . only the rise phase ( with a duration of @xmath1 ks ) and possibly the flare peak were observed .
we took advantage of the high count - rate of the x - ray source to carry out a detailed analysis of its spectrum during the whole exposure .
after a careful analysis , we interpreted the rise phase as resulting from the ignition of a first group of loops ( event a ) which triggered a subsequent two - ribbon flare ( event b ) .
event a was analyzed using a single - loop model , while a two - ribbon model was applied for event b. loop semi - lengths of @xmath2 were obtained .
such large structures had been previously observed in very young stellar objects ( @xmath3 myr ) .
this is the first time that they have been inferred in a slightly more evolved star .
the fluorescent iron emission line at 6.4 kev was detected during event b. since twa 11b seems to have no disk , the most plausible explanation found for its presence in the x - ray spectrum of this star is collisional- or photo - ionization .
as far as we are concerned , this is only the third clear detection of fe photospheric fluorescence in stars other than the sun . |
longitudinal space charge ( lsc ) driven microbunching instability @xcite in electron linacs with bunch compressors ( used as drivers of short wavelength fels ) was a subject of intense theoretical and experimental studies during the past years @xcite .
such instability develops in infrared and visible wavelength ranges and can hamper electron beam diagnostics and free electron laser ( fel ) operation .
it was proposed in @xcite to use this effect for generation of vacuum ultraviolet ( vuv ) and x - ray radiation .
a concept of a longitudinal space charge amplifier ( lsca ) was introduced , scaling relations for an optimized lsca were obtained , and its possible applications were analyzed .
it was pointed out , in particular , that a broadband nature of lsca supports generation of attosecond pulses .
there are many different schemes for generation of attosecond x - ray pulses from free electron lasers ( fels ) @xcite .
most of the proposed schemes make use of a short intense laser pulse to manipulate electron energy in a small slice of an electron bunch and to select only this short slice with the help of fel mechanism . in ref .
@xcite a scheme was suggested that uses the longitudinal space charge amplifier in combination with laser manipulation of the electron beam to produce attosecond soft x - ray pulses . in this paper
we describe the operation of the lsca - based attosecond scheme in details , exemplify it with the parameters of the soft x - ray free electron laser facility flash @xcite , perform thorough numerical simulations , and discuss a possible implementation at flash making use of the existing infrastructure .
the scheme ( see fig . [ scheme ] ) of the longitudinal space charge amplifier ( lsca ) @xcite is simple both conceptually and technically .
an amplification cascade consists of a focusing channel and a dispersive element ( usually a chicane ) with an optimized momentum compaction @xmath0 . in a channel the energy modulations
are accumulated , that are proportional to density modulations and space charge impedance of the drift space . in the chicane
these energy modulations are converted into induced density modulations that are much larger than initial ones @xcite , the ratio defines a gain per cascade . in this paper
we will consider the case when the amplification starts up from the shot noise in the electron beam .
a number of cascades is defined by the condition that the total gain , given by the product of partial gains in each cascade , is sufficient for saturation ( density modulation on the order of unity ) after the start up from shot noise .
the amplified density modulation has a large relative bandwidth , typically in the range 50 - 100 % . behind the last cascade a radiator undulator
is installed , which produces a powerful radiation with a relatively narrow line ( inverse number of periods ) within the central cone .
this radiation is transversely coherent , and the longitudinal coherence length is given by the product of the number of undulator periods by the radiation wavelength .
when lsca saturates in the last cascade , a typical enhancement of the radiation power over that of spontaneous emission is given by a number of electrons per wavelength .
first few cascades of lsca ( as in fig .
1 ) operate as described above .
a broadband density modulation is amplified around the optimal wavelength @xmath1 where @xmath2 is a transverse size of the beam , and @xmath3 is relativistic factor .
such a choice of the wavelength is optimal because the lsc impedance reaches its maximum , and , at the same time , transverse correlations of the lsc field are on the order of the beam size @xcite , what guarantees a good transverse coherence of the radiation in the end .
length of a drift space , beta - function , momentum compaction @xmath0 of a chicane are optimized for the chosen wavelength .
the last cascade is modified as shown in fig . 2 .
a short two - period modulator undulator is installed in front of the last chicane . in this undulator
the electron beam is modulated in energy ( modulation wavelength is much longer than @xmath4 ) by a few - cycle powerful laser pulse in the same way as it is suggested for fel - based attosecond schemes @xcite , see fig .
[ atto - mod ] as an example . a short slice in the electron bunch ( between 0 and 0.3 @xmath5 m in fig . [ atto - mod ] )
gets the strongest energy chirp , and is being strongly compressed ( by a factor @xmath6 ) in the following chicane .
the @xmath0 of this chicane and the chirp are adjusted such that a required wavelength compression within that slice is achieved , and , at the same time the amplification of the microbunching within the slice through the last chicane is optimal ( parameters of the whole amplification chain are adjusted such that saturation is reached in the last chicane ) .
other parts of the electron bunch are either uncompressed or have much weaker compression than a slice with the strongest energy chirp . the chicane is followed by a short radiator undulator of which resonance wavelength , @xmath7 , corresponds to the ( broadband ) wavelength spectrum within the strongly compressed slice .
then only this short slice of the electron bunch produces undulator radiation within the central cone ( which is selected by a pinhole downstream ) , while modulation wavelengths in the rest of the electron bunch are much longer than the resonance wavelength of the undulator . in other words ,
the rest of the electron bunch produces only spontaneous emission within the central cone , and this emission would define the contrast of generated attosecond pulses .
note that the last part of the scheme ( modulator undulator , chicane , and radiator undulator ) operates in a similar way as the last part of the attosecond scheme proposed in ref .
the two schemes differ essentially by mechanisms of the generation of microbunching before wavelength compression .
the authors of @xcite considered echo - enabled harmonic generation @xcite , while our scheme is based on the longitudinal space charge amplifier , described above . finally , let us note that broadband nature of lsca makes wavelength compression especially attractive @xcite .
indeed , the compression factor is given by the formula : @xmath8 where @xmath9 is the linear energy chirp ( the derivative of relative energy deviation ) .
for a large @xmath10 a variation of the compression factor reads : @xmath11 after the compression the bands of density modulations and of the radiator must overlap .
this leads to the following requirement on the compression stability : @xmath12 where @xmath13 , and @xmath14 and @xmath15 are bandwidths of the density modulation and of the radiator , respectively .
thus , the stability of the chirp must satisfy the requirement : @xmath16 for coherent fel - type modulations and an undulator as a radiator @xmath17 what might set very tight tolerance for the chirp stability and limit practically achievable compression factors . for an lsca , however , @xmath18 , so that for a given chirp stability one can go for much stronger compression . alternatively , for a given compression factor one can significantly loosen the tolerances .
note also that nonlinearities of the longitudinal phase space do not play a significant role in the case of lsca .
we exemplify an operation of the attosecond scheme with the parameters of the soft x - ray fel facility flash @xcite . in the section 6
we will discuss a possible technical implementation of this scheme making use of the existing infrastructure at flash . concerning the choice of parameters of the electron beam , we rely on the results of beam dynamics simulations @xcite .
for the case of the bunch charge 100 pc we reduce compression by a factor of two ( with respect to the scenario considered in @xcite ) , thus ending up with the slice parameters presented in table 1 .
.electron beam parameters [ cols= " <
, < " , ] number of lsca cascades , their parameters , and the operating wavelength range are chosen with the help of the guidelines of ref .
@xcite . as a focusing structure
we choose fodo - lattice with the period 1.4 m and the average beta - function also 1.4 m ( even smaller beta - function would be preferable , but the technical feasibility and operational issues have to be taken into account ) . for electron beam parameters from table 1 ,
we calculate from ( [ opt - wl ] ) that the optimal wavelength for amplification in lsca is around 40 nm .
the length of a drift in each cascade is chosen to be 2.8 m , or two fodo periods .
this length should provide a sufficient gain per cascade , it is smaller than a reduced wavelength of plasma oscillations , and the longitudinal velocity spread due to emittance does not play any role @xcite .
the optimal @xmath0 of the chicane for a given energy spread and wavelength range is about 50 @xmath19 m .
the chicane in each cascade should fit in a space between two quadrupoles .
thus , the total length of a cascade is about 3.5 m ( or 2.5 fodo periods ) . according to our estimates ,
we need three regular cascades with the parameters , described above , and a last special cascade ( see fig .
[ atto - scheme ] ) , so that a total length of the system is about 14 m. in the last cascade we have a drift of the same length ( 2.8 m ) , followed by a two - period undulator and a chicane with reduced @xmath0 .
the undulator has a period length of 10 cm and a peak field of 1.4 t , so that a resonance of 1.2 gev beam with the laser beam ( wavelength is 800 nm ) is provided . in the last cascade a final amplification and , at the same time , a wavelength compression ( down to 4 - 5 nm ) in a short slice
take place . to estimate the required energy chirp and the @xmath0 of the last chicane
, we have to take the following considerations into account .
the amplitude of density modulation at the compressed wavelength should not be strongly suppressed by the uncorrelated energy spread during compression . for a gaussian energy distribution
an amplitude gain per cascade is given by @xcite : @xmath20 here @xmath21 is the modulation wavenumber before compression , @xmath22 is the impedance of a drift space ( per unit length ) , @xmath23 is the free - space impedance , @xmath24 is the length of the drift space , @xmath25 is the beam current , @xmath26 is the alfven current , @xmath3 is relativistic factor , and @xmath27 is rms uncorrelated energy spread ( in units of rest energy ) . the beam current and energy spread are taken before compression .
the optimal @xmath0 from ( [ gain - wake ] ) is given by @xmath28 , where @xmath29 is the modulation wavenumber after compression .
then the required energy chirp can be defined from ( [ comp - fac ] ) : @xmath30 if the electron beam is modulated by a laser with the wavelength @xmath31 , and the modulation amplitude is @xmath32 , then the chirp at zero crossing is @xmath33 .
thus , with the help of ( [ chirp - req ] ) we get an estimate of a required amplitude of energy modulation : @xmath34 if the compression factor is about 8 - 10 , @xmath35 nm , @xmath36 nm , then the required energy modulation amplitude is about 20 mev , and the corresponding @xmath0 is 6 - 7 @xmath5 m .
we should note that formula ( [ gain - wake ] ) is valid for a linear regime of amplification , while we deal with a saturation in the last cascade .
moreover , energy modulations due to a short laser pulse are not sinusoidal .
therefore , the above obtained estimate can only be used as a first guess . for our simulations
we choose the modulation amplitude of 20 mev , and the @xmath0 was optimized in the simulations at 7.1 @xmath5 m . in order to get the desired energy modulations , we use a laser pulse with the duration of 5 fs ( fwhm ) and pulse energy 3 mj .
the laser beam is focused into the modulator undulator , the spot size in the waist is @xmath37 m . finally , we should choose parameters of the radiator undulator that should be tuned to the desired wavelength of 5 nm for the given beam energy of 1.2 gev .
we take a planar undulator with the following parameters : undulator period is 2.5 cm , peak field is 0.67 t , number of periods is 5 .
we assume that the undulator has a tunable gap which would allow us to change wavelength by changing the gap and/or the beam energy .
the numerical simulations of the lsca operation were done as follows .
self - interaction of electrons in the beam by space charge fields in the focussing channels was simulated with 3-d version of the space charge tracking code astra @xcite .
all important aspects of the problem like betatron motion of particles in the fodo channel , three - dimensional calculation of the space charge field , start up from shot noise were included in the simulations . to properly treat the start up from noise
, we took only a short part of the bunch with the length about 2 @xmath5 m , and used a real number of particles , @xmath38 , distributing them randomly in 6-d phase space .
after the energy change of each particle in a drift was calculated , we simply applied @xmath0 to add the effect of a chicane ( later we checked that csr effects do not play any significant role , see below ) . in this way
the evolution of the particle distribution through a single cascade was simulated .
then the procedure was repeated for the next cascade , etc .
the results of simulations of the first three cascades of lsca are shown in figs .
[ fig : astra - time ] and [ fig : astra - spectr ] . in fig .
[ fig : astra - time ] we show the amplified noisy modulations of beam current , and in fig .
[ fig : astra - spectr ] the modulus of the bunch form factor ( fourier transform of the trace in fig .
[ fig : astra - time ] , left ) is presented .
note that a linear gain ( in amplitude ) per cascade is about 5 at 40 - 50 nm wavelength . in the third cascade
the amplification becomes slightly nonlinear , so that the gain is reduced by about 10% .
typical feature size in spectrum ( interval of spectral coherence ) can be estimated as @xmath39 @xcite , where @xmath40 m is the length of a part of the bunch in our simulations . converting this dependence to a wavelength interval
, we get @xmath41 , i.e. one should expect a quadratic dependence of the coherence interval on the wavelength .
also note that in case of simulations with astra we used only a short part of a 100 pc bunch , so that the fine structure in spectrum does not correspond to a real situation with much longer bunch .
alternatively , we have developed a simple and fast 1-d code losca ( * lo*ngitudinal * s*pace * c*harge * a*mplifier ) .
initially , particles are distributed randomly in the phase space time - energy .
beam is then cut into short slices ( much shorter than @xmath42 ) , and the lsc interaction between the slices in a drift space is simulated with the help of the lsc wake function , averaged over beam cross - section @xcite .
the drift is divided into pieces , and at the end of each piece the paricles change their positions depending on their energies .
the change of particles distribution through a chicane is simply performed by applying @xmath0 .
note that in our particular case the change of particles positions happens only in chicanes . in the drifts
it is negligibly small because of a very small spread of the longitudinal velocities due to both energy spread and emittance , and also because the length of a drift is much smaller that the reduced wavelength of plasma oscillations @xcite .
this means that for our parameter set it is sufficient to calculate only energy change of particles due to lsc interaction without dividing the drift into pieces .
in fact , due to a simple model , used in losca , we were able to relatively quickly calculate the evolution of particles distribution through lsca cascades with real number of particles in a bunch ( with the charge of 100 pc ) , namely with @xmath43 particles . to have a representative set of data
, we performed several simulation runs with different initial shot noise realizations .
the results of simulations of the first three cascades of lsca for one of realizations are presented in figs .
[ fig : losca-3rd - time ] and [ fig : losca-3rd - spectr ] .
it is interesting to observe ( see right plots in figs . [
fig : astra - time ] and [ fig : losca-3rd - time ] ) that losca gives the same amplification of shot noise modulations as astra , despite the fact that essentially different models are used in these two codes .
we propose the following explanation . for a considered parameter range
a typical transverse size of the lsc field is comparable to the electron bunch transverse size .
particles in astra execute betatron oscillations and sample different lsc fields , while in losca there is no transverse motion but the applied lsc wake is averaged over electron beam cross - section .
our conclusion is that both approaches lead to very similar results for the lsca gain , so that a simplified approach is justified for a typical parameter range of lsca @xcite : amplified wavelength is in the range described by eq .
( [ opt - wl ] ) , and beta - function is smaller than a length of a drift space in an amplification cascade . in the forth amplification cascade
the beam first moves in the 2.8 m long drift , accumulating energy modulation due to lsc field , and then it is modulated in energy by a short laser pulse .
modulation of the beam by a laser pulse in two - period undulator was simulated in the same way as it was done in @xcite , namely by a direct integration of equations of motion in the combined field of laser and undulator .
the resulting energy modulation is presented in fig .
[ atto - mod ] , it was imposed on the phase space distribution of the particles coming from the last drift .
then the effect of the last chicane was simulated by applying @xmath44 m . a typical current distribution for the beam , simulated with losca , is presented in fig .
[ fig : losca - comp ] . of course , amplitudes and positions of peaks ( right plot in fig .
[ fig : losca - comp ] ) are changing shot - to - shot .
a similar distribution is obtained from the simulation with astra , see fig .
[ fig : csr - nocsr ] .
the next problem is a possible influence of coherent synchrotron radiation ( csr ) field in the chicane on extremely short high - current spikes .
we have performed simulations with the code csrtrack @xcite .
we took the particle distribution at the exit of the forth drift ( simulated with astra ) , imposed energy modulation by a laser , reduced number of particles ( still having representative ensemble ) , and transferred this distribution into csrtrack .
finally , we did tracking through the chicane , taking into account csr wake with the help of the projected model @xcite .
the results can be seen in fig .
[ fig : csr - nocsr ] . as expected @xcite , no noticeable effects on longitudinal and transverse dynamics were observed ( except for a little time shift ) .
the reason for this is a practically complete smearing of microstructures inside the chicane due to the coupling of transverse and longitudinal dynamics described by @xmath45 and @xmath46 elements of transfer matrix @xcite .
ensemble - averaged spectrum of the beam current ( modulus of the bunch form factor ) after the last chicane is presented in fig .
[ fig : losca - comp - spectr ] . within the spectral window , shown in this figure , the spectral components are dominantly coming from the fine structure within the main peak in fig .
[ fig : losca - comp ] ( right ) .
the fine structures within two side peaks ( which are significantly weaker compressed ) , contribute to the spectral density for wavelengths longer than 10 nm . if a resonance wavelength of the radiator undulator is within the range of 2 - 8 nm , one would get quite powerful radiation generated by fine structure within the main peak , i.e. very short pulses - depending also on the number of undulator periods . some small contribution from side peaks is , in principle , possible due to a nonlinear harmonic generation mechanism during lsc amplification and compression . to illustrate the main properties of the radiation , we assume that the undulator is tuned to 5 nm , and the number of periods is 5 ( undulator parameters are presented in the previous section ) .
the radiation process was simulated with the code fast @xcite .
for this purpose , the particles distribution , simulated with the help of astra or losca , was transferred into fast input distribution .
we should note that fast , as well as other fel codes , uses resonance approximation , i.e. it deals essentially with narrow - band signals .
the question arises wether or not one can properly simulate a process if the input signal ( density modulation ) has a broad band . we can answer this question as follows ( see @xcite for more details ) : within the central cone of the undulator radiation the code produces correct results even if the incoming density modulation has a wide band ( in particular , if there is only shot noise having white spectrum ) .
the accuracy of simulations of radiation properties within the central cone is on the order of inverse number of undulator periods , i.e. about 20 % in our case .
main simulation results are presented in fig .
[ fig : power-4shots ] . several typical realizations of attosecond pulses
are shown , that were obtained from incoming particles distributions , simulated with astra ( solid line ) and losca .
a typical duration of these pulses is 50 - 70 attosecons ( fwhm ) , the peak power is at 100 mw level , and the bandwidth is about 20% . ensemble - averaged pulse energy is 5 nj with the pulse - to - pulse rms fluctuations about 35 % .
one can sometimes observe side peaks ( corresponding to those in density modulation in fig .
[ fig : losca - comp ] , left plot ) , but they are typically below 1% of the power in the main peak .
the rest of the bunch radiates only spontaneously ( amplified density modulations in the range 20 - 80 nm could have produced radiation into large angles , well beyond the central cone , but this radiation is totally suppressed due to a finite transverse size of the beam ) . as one
can see from fig .
[ fig : losca - comp ] , right plot , the half - angle of coherent radiation is about 50 @xmath5rad .
selecting ( with the help of a pinhole ) a cone with half - angle 100 - 120 @xmath5rad would mean that one does lose not any power of attosecond x - ray pulses , while spontaneous radiation background is reduced to the level of 50 pj for 100 pc bunch .
thus , we can state that a contrast ( defined as a ratio of radiation energy of attosecond pulses to the total pulse energy ) is high , above 98 % .
flash ( free electron laser in hamburg ) is the soft x - ray fel user facility @xcite operating in the wavelength range from 4 to 50 nm .
short vuv and soft x - ray pulses ( down to 10 fs @xcite ) are produced in a 27 m long undulator in self - amplified spontaneous emission ( sase ) mode @xcite .
recently , radiation at water window wavelengths down to 4 nm was generated .
flash layout is shown in fig .
[ flash ] .
the beam is produced by the laser - driven rf gun , compressed in two bunch compressors , and accelerated in superconducting rf modules up to 1.2 gev .
then the beam is guided through the dogleg ( where it is collimated ) , moves about 50 m towards sase undulator , radiates there , and goes to the beam dump ( alternatively , it can go to the dump via bypass ) . presently , there are two temporal experimental setups in the straight section between the dogleg and the undulator .
optical replica synthesizer ( ors ) @xcite is placed ( not shown in fig .
[ flash ] ) directly behind the dogleg / collimator .
the ors is followed by sflash @xcite , the experimental setup for seeding the fel with the radiation produced by the generation of higher harmonics of ti : sa laser in rare gases .
the laser beam together with high harmonics is sent from the laser laboratory to the tunnel , where it is in - coupled to the electron beam line through the dogleg . in the case of successful lasing of sflash
, the amplified vuv radiation is supposed to be outcoupled by a mirror and transported to the experimental hutch ( at the same time laser beam for pump - probe experiments is transported there from the laser laboratory ) .
the distance between the dogleg and the mirror is about 30 m. we note that this infrastructure is well suited for production of attosecond pulses in the lsca and using them for experiments in the hutch .
the total length of lsca is supposed to be about 14 m , so that it can substitute one of the two existing setups .
experiments can be started with the existing ti : sa laser ( that delivers 35 fs long pulses with a pulse energy of up to 50 mj @xcite ) for modulation of the electron beam , in this case one will produce a train of attosecond soft x - ray pulses .
we have simulated this situation with losca and fast , the results are presented in fig .
[ fig : long - pulse ] . to get required amplitude of energy modulations , we used only 15 mj out of available laser pulse energy .
later the laser can be upgraded in order to produce 5 fs long pulses with pulse energy about 3 mj what would allow to obtain an isolated attosecond x - ray pulse .
the attosecond pulses with the help of the outcoupling mirror can be transported to the existing experimental hutch where they can be used together with synchronized few femtosecond long powerful laser pulses . as a possible option one can consider seeding flash undulator with an attosecond pulse . in this case
the outcoupling mirror is moved out , and the chicane at that position is used to slightly delay the electron bunch so that the radiation pulse is parked on the fresh part of the bunch . despite the transverse size of the radiation at the entrance of flash undulator is much larger than that of the electron beam , an effective seed power is still much larger than an effective power of shot noise . during amplification the radiation pulse would be stretched , so that finally pulse duration would be comparable to fel coherence time ( about 1 fs ) . finally we note that the attosecond scheme , described in this paper , can also be used at flash ii @xcite , fermi@elettra @xcite and similar facilities , as well as at soft x - ray beamlines of large x - ray fel facilities like sase3 beamline of the european xfel @xcite . | a longitudinal space charge amplifier ( lsca ) , operating in soft x - ray regime , was recently proposed .
such an amplifier consists of a few amplification cascades ( focusing channel and chicane ) and a short radiator undulator in the end .
broadband nature of lsca supports generation of few - cycle pulses as well as wavelength compression . in this paper
we consider an application of these properties of lsca for generation of attosecond x - ray pulses .
it is shown that a compact and cheap addition to the soft x - ray free electron laser facility flash would allow to generate 60 attosecond ( fwhm ) long x - ray pulses with the peak power at 100 mw level and a contrast above 98% .
desy 11 - 040 march 2011 , |
the solvation force @xmath4 is the force per unit area between two surfaces , large colloidal particles or macromolecules due to the intervening fluid .
it arises in the thermodynamic description of confined fluids as an excess pressure over the bulk value ( fixed by the reservoir ) and is conjugate to the distance @xmath0 between the confining surfaces @xcite .
the study of this generalized force for simple and complex fluids is useful as it may play an important role in the field of colloid sciences , in self - assembling systems and in the biology of protein folding . from a fundamental viewpoint ,
solvation forces present an abundance of interesting problems to investigate : the behaviour at the phase transitions of a confined fluid , such as capillary condensation @xcite , bridging @xcite , localization - delocalization transition @xcite , or layering @xcite , the structure at small separations @xmath0 @xcite , or the dependence on the fluid - fluid interactions and the surface potentials @xcite .
an addition appeal lies in the fact that it may be measured experimentally by a number of different techniques @xcite ; e.g. , the surface force apparatus , the atomic force microscope , or the total internal reflection microscopy ( timr ) @xcite . by using the latter technique ,
the first direct determination of the universal contribution to @xmath4 between a single colloidal particle and a planar surface in the presence of a near - critical binary liquid mixture was provided @xcite .
the universal contribution , called the critical casimir force @xcite , is induced by the confinement of critical fluctuations of the order parameter of the second order phase transition , e.g. the density of the fluid or the concentration of one species for a binary liquid mixture . at the bulk critical point
this fluctuation - induced force has a leading power law decay @xmath5 as @xmath6 , where @xmath7 is the spatial dimension of the system and @xmath8 is the casimir amplitude . the finite - size scaling theory predicts ( see , e.g. , ref .
@xcite ) that in the vicinity of the bulk critical point @xmath4 is described by a scaling function that depends on the so - called universality class of the phase transition occurring in the bulk and on the geometry and surface universality classes of the confining surfaces @xcite .
the critical casimir force is a subject of recent considerable theoretical and experimental interest because at the submicron scale its strength and range is comparable with other interactions and because it can be easily controlled , including its sign , by changes of thermodynamic fields , such as the temperature , and by appropriate surface treatments .
as mentioned above , the solvation force is associated with fluid - fluid interactions and the surface potential . in the present paper
we want to study the relevant effects of long - ranged , i.e. , algebraically decaying forces on the near - critical behaviour of @xmath4 between two parallel walls and test predictions of the critical finite - size scaling theory .
examples of such long - ranged forces are dispersion or van der waals forces , dipolar forces , rkky interactions , forces in charged systems , and elastic forces in solids . here
we utilize the equivalence between a lattice gas model of a fluid and the ising model and consider ising spin films subject to identical boundary fields decaying in the orthogonal direction at a distance @xmath9 from the surface as @xmath10 , with @xmath11 and @xmath12 . using the density - matrix renormalization - group ( dmrg ) method
, we calculate the solvation force in two - dimensional ( @xmath13 ) systems along various thermodynamic paths .
the dmrg method is based on the transfer matrix approach and provides a numerically very efficient iterative truncation algorithm for constructing the effective transfer matrices for strips of fixed width and infinite length .
the advantage of this method is that it can treat arbitrary fields that are coupled to the single spin variable .
moreover , because the dmrg method gives quasi - exact results for the spectrum of the transfer matrix of the system , fluctuations are fully accounted for in this approach .
recall , that in @xmath13 systems fluctuation effects are particularly strong .
however , this method works only if the spin - spin interactions are nearest - neighbour and , at present , it is limited to the two - dimensional systems .
the studies carried out here complement and extend the earlier works ( see refs .
@xcite ) . in refs .
@xcite it was argued that the solvation ( casimir ) force in the vicinity of the critical point is strongly influenced by capillary condensation , i.e. , the shift of the bulk first - order transition which occurs below the critical temperature @xmath14 .
the residual condensation leads to the solvation force which is much more attractive at temperatures near @xmath14 and a reservoir densities slightly below the critical value ( or compositions slightly away from the critical composition in a binary mixture ) than the casimir value ( at the bulk critical point ) for the same @xmath0 .
these predictions were supported by explicit calculations for @xmath13 ising films with _ short - ranged _ ( contact ) boundary fields . by using the dmrg method it was possible to study the region between the capillary and bulk critical points , i.e. , for nonvanishing ordering field @xmath15 , which corresponds to the chemical potential difference @xmath16 . in particular , the scaling functions of @xmath4 along several isotherms were obtained showing that upon increasing the temperature towards @xmath14 a weakly rounded in @xmath13 discontinuous jump of @xmath4 on crossing the coexistence line @xcite transforms gradually into a minimum .
a location of this minimum follows roughly the continuation of the ( pseudo ) capillary condensation line towards @xmath14 @xcite .
the similar behaviour was found in the field - theoretic model solved numerically in the mean - field approximation @xcite , again for the short - ranged boundary fields . in ref .
@xcite , the case of short - ranged fluid - fluid interaction and long - ranged wall - fluid potentials decaying as @xmath17 for @xmath18 was considered for @xmath13 ising films and for the truncated lennard - jones fluid in a slit geometry .
the discrete model was treated within the dmrg method whereas the continuum one by the nonlocal density functional theory .
the study was focused on the asymptotic behaviour of the solvation force as @xmath6 . except for a high temperature ising system ,
results for both models agree with the predictions from the analysis based on the wall - particle ornstein - zernike equations that @xmath4 is repulsive and asymptotically decays with the same power law as the wall - fluid potential ; this prediction holds away from the critical temperature and from any phase transition . for ising films above @xmath14 the asymptotic behaviour
was found to be of the higher order then that of the boundary field , i.e. , @xmath19 .
this was explain by the specific symmetry of the order parameter with the spontaneous magnetization @xmath20 equal to zero above @xmath14 .
moreover , @xmath4 was calculated along the _ bulk _ two - phase coexistence line slightly on the liquid side of this line and at the critical density for @xmath21 ; in the ising system this path corresponds to the line @xmath22 .
the obtained results imply that for ordering field @xmath22 and @xmath6 @xmath23 where @xmath24 is the bulk density and @xmath25 is related to the strength of the wall - fluid potential . for ising systems @xmath26
is replaced by @xmath27 .
@xmath28 is a universal scaling function describing the contribution arising from critical fluctuations of a fluid . @xmath29 and @xmath30 is the critical exponent of the bulk correlation length @xmath31 .
@xmath32 is vanishingly small away from the critical region and is negative for identical walls . in the present study we consider the neighbourhood of the bulk critical point and
investigate the scaling behaviour of the solvation force along the critical and two subcritical isotherms and along the pseudo - coexistence ( capillary condensation ) line slightly on the `` liquid '' and on the `` gas '' sides .
a substantial progress in computer capacities allows to study sufficiently thick films to assure that the scaling limit is achieved .
here we consider strips of widths up to @xmath33 lattice constants . according to general scaling arguments @xcite the finite - size behaviour of the singular part of the solvation force
is modified by the presence of the long - ranged substrate - wall potentials . however ,
if @xmath34 the universal behavior is expected to hold . in this case
the long - ranged part of the boundary field is irrelevant in the rg sense with respect to a pure contact surface field @xcite . here
@xmath35 is the critical exponent governing the algebraic decay of the two - point correlation function in the bulk and at @xmath14 . in the present case of the @xmath36 ising model @xmath37 so that for @xmath38 we expect to observe the power law @xmath39 .
the scaling of the solvation force in films with long - ranged fluid - fluid and substrate - fluid potentials was analysed in ref .
@xcite by using the general scaling arguments and mean - field theory .
we will summarize the relevant conclusions of this analyses and relate our results to them in sec .
[ sec:3 ] .
our paper is organized as follows . in sec .
[ sec:2 ] we introduce the model and describe the determination of the phase diagrams for various values of the parameter @xmath1 .
[ sec:3 ] contains our results .
finally , we give conclusions in sec .
[ sec:4 ] .
we consider @xmath40 ising strips defined on a square lattice of size @xmath41 and subject to the same boundary fields on both sides .
the lattice consists of @xmath0 parallel rows at spacing @xmath42 , so that the width of the strip is @xmath43 ; in the following we set @xmath44 . at each site
there is an ising spin variable taking the value @xmath45 , where @xmath46 labels the site .
the boundary surfaces are located in the rows @xmath47 and @xmath48 and periodic boundary conditions ( pbcs ) are assumed in the lateral @xmath49 direction .
the hamiltonian of our model is given by @xmath50 where the first sum is over all nearest - neighbor pairs and the external potential is measured in units of @xmath51 .
@xmath52 is the total boundary field experienced by a spin in row @xmath9 ; it is the sum of the two independent wall contributions .
s_ingle - boundary field @xmath53 is taken to have the form @xmath54 with @xmath55 and @xmath56 .
@xmath57 is a bulk magnetic field . @xmath58 and @xmath57 are dimensionless ( see eq .
( [ eq:1 ] ) ) .
as already mentioned this model is equivalent to the @xmath13 lattice gas model of a two - dimensional one - component fluid with a short - ranged interaction potential between the fluid particles and either short - ranged or long - ranged substrate potentials ( see , e.g. , ref .
@xcite ) . in fig .
[ fig:1 ] we show the phase diagram for the present model calculated by using the dmrg method for a strip of width 600 and for three choices of the parameter @xmath1 describing the decay of the boundary field : @xmath2 , and 50 .
@xmath58 is chosen so that the scaling variable @xmath59 , which for the short - ranged boundary fields is sufficient to ensure each system corresponds to the infinite surface field scaling limit @xcite . the case @xmath60 is expected to resemble the behavior corresponding to short - ranged surface forces . in this figure we display the various thermodynamic paths along which we have calculated the solvation force . for surfaces which prefer the same bulk phase
, the phenomenon equivalent to capillary condensation takes place . the pseudo - phase coexistence between phases of spin up and
spin down occurs along the line @xmath61 , which is given approximately by the analogue of the kelvin equation @xcite . for positive surface fields capillary
condensation occurs at negative values @xmath61 of the bulk field @xmath57 .
the pseudo - coexistence lines have been identified as those positions @xmath62 in the phase diagram where the total magnetization of the strip vanishes , i.e. , @xmath63 with @xmath64 .
we have not attempted to localize the position of the pseudo - critical temperature @xmath65 which ends the pseudo - coexistence lines ; the unique determination of the pseudo - critical point is not possible because non - analytic behaviour is rounded in @xmath13 @xcite .
notice , that for @xmath66 and 50 and @xmath21 , the line defined by the zeros of the total magnetization moves to more negative values of @xmath57 upon increasing temperatures .
for @xmath67 we observe this trend already below @xmath14 . in ref .
@xcite the shift of pseudo - phase coexistence lines @xmath61 for @xmath67 and 3 relative to the short - ranged pseudo - phase coexistence line ( @xmath60 ) was analysed .
for @xmath60 and 3 also the scaling behaviour of the pseudo - coexistence line of capillary condensation was studied . ising strip subject to identical boundary fields @xmath68 ( see eq .
( [ eq : v_j ] ) ) obtained by using dmrg for a strip width @xmath69 and the amplitude @xmath58 for the boundary fields such that @xmath59 .
the thick solid line indicates the bulk coexistence line at @xmath22 ending at the bulk critical point ( black circle ) .
the lines interpolating the symbols represent the pseudo - phase coexistence lines @xmath61 for three different values of the exponent @xmath1 governing the algebraic decay of the boundary fields .
three isotherms along which the solvation force has been calculated are shown - paths @xmath70-@xmath71 .
the other thermodynamic path @xmath72 and @xmath73 run parallel to the pseudo - coexistence line slightly on the `` liquid '' and on the `` gas '' side , respectively . ]
the thick solid line in fig .
[ fig:1 ] indicates the bulk phase coexistence line @xmath74 terminating at the bulk critical point @xmath75 ( black circle ) .
the symbols ( triangles for @xmath67 , squares for @xmath66 , and circles for @xmath60 ) show the pseudo - phase coexistence .
it turns out that the pseudo - phase coexistence line for long - ranged boundary fields is located slightly further away from the bulk coexistence line , especially at lower temperatures , than the pseudo - phase coexistence line for the short - ranged boundary fields ( @xmath60 ) .
before presenting our results let us summarize predictions for the scaling behaviour of the solvation force based on general scaling arguments @xcite .
one expects that the finite - size behaviour of the singular part of the solvation force is modified due to the presence of long - ranged substrate - fluid potentials : @xmath76\ ] ] here @xmath77 is the bulk correlation length at bulk coexistence @xmath22 while @xmath78 is the bulk correlation length at the critical temperature @xmath79 .
@xmath80 is the correction to scaling exponents due to the long - ranged tail of the substrate - fluid interactions .
@xmath81 is the wegner s @xcite correction - to - scaling exponent for short - ranged systems , which is equal to @xmath82 for @xmath13 ising model .
@xmath83 is a dimensionless nonuniversal scaling field . in the case of @xmath60 that mimics the effect of the contact surface fields , the third argument of the scaling function in ( [ eq:3 ] ) should be replaced by @xmath84 , where the surface gap exponent @xmath85 for @xmath13 ising system . for large @xmath0 and @xmath86
, one can expand the scaling function @xmath32 in ( [ eq:3 ] ) @xmath87 + \left(l/\xi_0\right)^{-\omega_s}h_1\vartheta^{lr}_s\left[l/\xi_{\tau},l/\xi_{h}\right]\\ \nonumber & + & \left(l/\xi_0\right)^{-\omega}g_{\omega}\vartheta^{sr}_{\omega}\left[l/\xi_{\tau},l/\xi_{h}\right].\end{aligned}\ ] ] thus for @xmath88 , @xmath89 and @xmath90 represent corrections to the leading @xmath0-dependence provided by @xmath91 .
for @xmath92 the scaling function @xmath91 decays exponentially , therefore in this regime @xmath89 is the leading finite - size contribution in the singular behaviour of the force . as mentioned already in the introduction , in the present case of the @xmath13 ising model @xmath93 so that the expansion ( [ eq:4 ] ) should hold for @xmath38 , i.e. , for all cases considered in our work .
however , one can argue that for sufficiently thin films the contribution of the long - ranged substrate - fluid potential is always important , also at the bulk critical point . because the separation between two surfaces is finite and
the substrate - fluid potentials are long - ranged and identical , there will be always some nonzero field acting at the center of the film .
this effect can be interpreted as if the system experiences an external bulk field @xmath94^{-p}$ ] , although in fact it might be at the bulk coexistence curve @xmath22 .
the relevance of finite - size contributions due to this effective bulk field can be estimated by taking into account that @xmath57 scales as @xmath95 .
these finite - size contributions are negligible in the critical regime if @xmath96^{\delta/\nu - p}\ll 1 $ ] . from that inequality
one can identify the thickness of the film @xmath97 such that for @xmath98 the effect due to the long - ranged substrate - fluid potentials is relevant @xmath99 .
for @xmath13 ising model @xmath100 and @xmath101 therefore for @xmath67 one obtains @xmath102 , i.e. , the long - ranged tails practically always matter . for @xmath66 , @xmath103 which assuming @xmath104 gives @xmath105 .
the separation @xmath0 is in units of a lattice constant .
the total excess free energy per unit area for the case of identical surface fields @xmath106 and non - vanishing bulk magnetic field @xmath57 can be written as @xmath107 where @xmath108 is the free energy per site , @xmath109 is the bulk free energy , @xmath110 is the @xmath0-independent surface excess free energy contributed from each wall , and @xmath111 is the finite - size contribution to the free energy .
all energies are measured in units of @xmath112 and the temperature in units of @xmath113 .
@xmath111 , which vanishes for @xmath6 , gives rise to the generalized force , which is analogous to the solvation force between the walls in the case of confined fluids , @xmath114 in the transfer matrix approach the leading eigenvalue @xmath115 of the transfer matrix @xmath116 @xmath117 gives the free energy per spin of an ising strip as @xmath118 the components of the eigenvector @xmath119 related to the leading eigenvalue give the probabilities of various configurations . in order to calculate the size @xmath0 dependence of the solvation force at fixed values of parameters ( @xmath120 ) we calculate the excess free energy per unit area @xmath121 at @xmath122 and @xmath123 . having values @xmath124 and @xmath125 we approximate the derivative in eq .
( [ eq : solfordef ] ) by a finite difference @xmath126 .
the case with vanishing external field @xmath57 is relatively convenient to study because for the bulk free energy there exists an exact solution by onsager @xcite . for non - zero @xmath57 , in order to find the bulk free energy for each thermodynamic state one has to perform the calculation for finite systems and extrapolate the data to @xmath127 .
we have determined the largest eigenvalue of @xmath116 for the strips ( @xmath128 ) with free boundary conditions and widths @xmath0 ranging from @xmath129 to @xmath130 .
next the free energy was extrapolated to @xmath127 by means of the bulirsch and stoer method @xcite .
a new computational problem arises when one wants to determine the solvation force on the thermodynamic path along the pseudo - coexistence line . in order to calculate
the derivative of the excess free energy at certain @xmath131 the values of the free energy are necessary at @xmath122 and @xmath123 . according to the kelvin equation
, the pseudo - coexistence is shifted from the bulk position @xmath22 proportionally to @xmath132 .
therefore the path @xmath72 or @xmath73 can not be chosen to lie close to the pseudo - coexistence line corresponding to @xmath131 ( see fig . [ fig:2 ] ) . our way of defining the middle line between the @xmath122 and @xmath123 pseudo - coexistence lines and the shift @xmath133 quarantees that for both @xmath122 and @xmath123 one is on the same ( liquid or gas ) side of the above coexistence lines .
) boundary fields @xmath134 for strip widths @xmath135 and @xmath136 corresponding to @xmath137 .
the solid line denotes the middle line between both pseudo - coexistence lines defined as @xmath138 .
dashed lines indicate the pseudo - coexistence lines , whereas the dashed - dot lines present the thermodynamic paths shifted with respect to the @xmath139 line by @xmath140 . ]
the solvation force scaled by @xmath141 as a function of the scaling variable @xmath142 , evaluated along the pseudo - coexistence line slightly on its gas side ( path ( 4 ) ) and slightly on its liquid side ( path ( 5 ) ) is shown in fig .
[ fig:3 ] and fig .
[ fig:4 ] , respectively .
data correspond to the fixed @xmath143 and different values of @xmath0 .
note that in the case of @xmath60 which mimics a short - ranged contact surface field , @xmath49 is the relevant scaling variable .
panels ( a ) , ( b ) and ( c ) correspond to @xmath2 , and 50 , as depicted . on both sides of the pseudo - coexistence line and for all values of the parameter @xmath1 , the force is attractive in the whole range of the studied scaling variable @xmath142 . on the gas side , data corresponding to different separations @xmath0 do not collapse into a common curve , i.e. , there is no scaling except for @xmath66 and 50 near the minimum close to @xmath14 .
moreover , the solvation force @xmath144 is very weak - the weaker the larger value of the decay exponent @xmath1 . for @xmath66 and 50
the behaviour is qualitatively the same , we observe a slight linear variation of @xmath144 with a `` dip '' in the narrow interval below but close to @xmath14 . on the contrary , on the liquid side of the pseudo - coexistence line
an excellent scaling has been found for @xmath66 and @xmath3 . for @xmath67
one can see deviations from scaling , especially below @xmath14 .
the solvation force @xmath145 is much stronger than on the gas side - data shown in fig .
[ fig:4 ] correspond to the scaling function divided by 100 .
the presence of the pseudo - capillary critical point manifests itself as a crossover between a rapid linear increase of the solvation force at low temperatures and a saturation at small values for high temperatures . in ref . @xcite it was argued that at the transition slightly on the liquid side @xmath146 .
it follows that the amplitude of @xmath147 should increase in the same fashion as the interfacial tension as @xmath148 increases at fixed @xmath0 . for the @xmath13 ising model
, the surface tension is given exactly by @xmath149 , where @xmath150 , @xmath151 , and @xmath152 @xcite . since @xmath153 , the scaling function should vary linearly with @xmath154 at fixed @xmath0 , which agrees with our result for @xmath60 and 3 . the argument given in ref .
@xcite is based on the macroscopic approximation for the total free energy which ignores interactions between surfaces , i.e. , it is valid for @xmath6 . for @xmath67 such interactions
can not be ignored , even for large @xmath0 .
therefore it is not surprising that we can see some deviations from the linear variation of @xmath145 in that case . for the gas side of the pseudo - coexistence curve , the analysis in ref .
@xcite gives @xmath155 .
what we have found , is a nonzero but very weak solvation force .
it arises from the @xmath0-dependent part of the total free energy which is neglected in the previous argumentation .
as mentioned in the introduction , away from the bulk critical point the dominant contribution to the solvation force comes from the regular term which decays with @xmath0 in the same way as the substrate fluid potential , i.e. , @xmath156 or exponentially in @xmath0 for the case of short - range substrate potential - this explains the observed lack of the scaling behaviour .
finally , we note that on both sides of the pseudo - coexistence line @xmath4 is attractive for all values of @xmath1 . in contrast , in the same range of temperatures along the bulk coexistence line @xmath22 the solvation force is repulsive for @xmath67 and 3 and attractive for short - range boundary fields @xcite . calculated for different separations @xmath0 along the line of pseudo - coexistence slightly on its gas side - the path @xmath72 .
the amplitude of the substrate potential @xmath58 is chosen such that @xmath157 .
different panels show results for different values of the exponent @xmath1 describing the algebraic decay of the substrate fluid potential .
scaling does not hold . ] .
scaling is very well obeyed . ]
as mentioned in the introduction , the scaling behaviour of the solvation force along various isotherms was studied for @xmath13 ising model with short - range contact boundary fields in ref .
the conclusion was that in the range of the scaling variable @xmath142 in which the solvation force displays the features of ( weakly rounded ) capillary condensation , i.e. , the jump from values appropriate to a spin down @xmath158 ( gas ) phase @xmath159 to the negative values appropriate to a spin up @xmath160 ( liquid ) phase @xmath161 , scaling is not well obeyed . in the present study
we have been able to perform calculations for much bigger systems than that considered in @xcite .
we find that along the isotherms corresponding to @xmath162 and @xmath163 , data for different values of the separation @xmath0 collapse onto the common curve on the liquid side of the pseudo - coexistence line - see fig .
[ fig:5 ] .
this holds for all considered values of the parameter @xmath1 .
this is because on the liquid side of the pseudo - coexistence line the leading behaviour of the solvation force @xmath161 , so that @xmath164 is approximately a quadratic function of @xmath165 . on the gas side ,
the leading behaviour is the regular part of the solvation force ( see eq .
( [ eq:1 ] ) ) .
the maximum absolute value of the force is @xmath166 , which for a fixed temperature decreases as @xmath132 .
the location of the abrupt change in the value of the solvation force depends slightly on @xmath0 , as it should be for a weakly rounded transition .
what is also characteristic is that on the liquid side of the pseudo - coexistence curve the scaling function varies only very little from one isotherm to another .
as @xmath142 decreases the jump of @xmath4 gradually transforms into a minimum . for @xmath60 , scaling along the critical isotherm is excellent in the whole range of the scaling variable @xmath167 ( see fig .
[ fig:6 ] ) . for @xmath67 and 3
, data collapse only on the left shoulder of the minimum , which indicates some residual effect of the condensed phase . as one can expect the finite size effects increase with the range of the substrate potential .
notice also that for long - ranged boundary fields the maximal absolute value of the solvation force is strongly increased with respect to the one for the short - ranged one and that the force becomes repulsive for @xmath168 .
calculated for different separations @xmath0 along 3 various isotherms . for @xmath169 ( path ( 1)- bottom curve ) and 42 ( path ( 2)- middle curve ) , the sharp jumps to a strongly attractive solvation force as @xmath167
is reduced indicate weakly rounded capillary condensation of the @xmath160 ( liquid ) phase .
data corresponding to @xmath170 ( crosses ) , 501 ( squares ) , and 401 ( circles ) follow common curves to the left from the jumps . along the critical isotherm solvation force
exhibits a minimum . ]
calculated along the critical isotherm .
scaling is excellent for @xmath60 . for @xmath66 and 2 data collapse on the left shoulder of the minimum - to the right from the minimum we observe deviations from scaling . for @xmath67 and 3
, @xmath4 becomes repulsive for small values of @xmath167 .
symbols have the same meaning as in the fig . [ fig:5 ] . ]
we have presented dmrg results for the solvation force @xmath4 for @xmath13 ising systems between identical walls with surface potentials decaying algebraically with the exponent @xmath2 and 50 .
the range of the substrate potential does not influence the qualitative behaviour of the solvation force .
we found , in agreement with simple macroscopic arguments , that on the liquid side of the pseudo - coexistence line the behaviour of the solvation force is dominated by the bulk excess free energy associated with the @xmath160 phase being metastable in the bulk .
thus @xmath147 is almost @xmath0-independent along the isotherms ( for temperature sufficiently below @xmath14 ) which leads to the data collapse in plots of @xmath164 vs @xmath167 . along the pseudo - coexistence line ,
@xmath171 , which justifies scaling and the linear behaviour of the solvation force slightly on the liquid side of the pseudo - coexistence line .
contribution to the solvation force arising from the @xmath0-dependent regular terms in the free energy manifests itself strongly for @xmath67 and on the gas side of the pseudo - coexistence line for all values of the decay exponent @xmath1 .
there , except in the vicinity of the pseudo - capillary critical point for @xmath172 , scaling does not hold and the solvation force is weak .
close to the bulk critical temperature the solvation force changes its behaviour significantly , i.e. , the jump of @xmath4 transforms into a minimum , but the effects of ( pseudo ) capillary condensations remain strong .
this manifests itself at small values of @xmath167 where we observe data collapse for all values of @xmath1 and where the scaling function varies as @xmath173 , which implies @xmath174 for small @xmath57 and fixed @xmath0 corresponding to the residual metastable bulk phase
. for @xmath67 and 3 , along the pseudo - coexistence line on its gas side and along the isotherms for large values of @xmath167 , the regular contribution ( [ eq:1 ] ) dominates the behaviour of the solvation force . because @xmath175 , at these thermodynamic points @xmath4 is attractive . on the liquid side of the pseudo - coexistence line and far away from ( pseudo ) capillary condensation and bulk criticality , @xmath4 becomes repulsive in agreement with ( [ eq:1 ] ) ( see also ref .
@xcite ) . near the bulk critical point and close to the capillary condensation ( pseudo ) transition
the solvation force is attractive . in the case of short - range - like boundary potential @xmath60 , @xmath4 is always attractive .
99 see e.g. , r. evans , j. phys .
matter * 2 * , 8989 ( 1990 ) .
l. d. gelb , k. e. gubbins , r. radhakrishnan , and m. ' sliwi ' nska - bartkowiak , rep .
. phys . * 62 * , 1573 ( 1999 ) .
a. macioek , a. drzewi ' nski , and p. bryk , j. chem . phys .
* 120 * , 1921 ( 2004 ) .
d. dantchev , f. schlesener , and s. dietrich , phys .
e * 76 * , 011121 ( 2007 ) . c. bauer , t. bieker , and s. dietrich , phys .
e * 62 * , 5324 ( 2000 ) .
d. andrienko , p. patricio , and o. i. vinogradova , j. chem . phys .
* 121 * , 4414 ( 2004 ) .
h. shinto , k. uranishi , h. miyahara , and k. higashitani , j. chem .
phys . * 116 * , 9500 ( 2002 ) .
a. o. parry and r. evans , physica a * 181 * , 250 ( 1992 ) .
m. c. stewart , phd thesis , university of bristol .
j. israelachvili , _ intermolecular and surface forces _ , 2nd ed .
( academic press , london , 1991 ) . h. k. christenson and p. m. cleasson , adv .
colloid interface sci .
* 91 * , 391 ( 2001 ) .
p. attard , d. e. b ' erard , c. p. ursenbach , and g. n. patey , phys .
a * 44 * , 8224 ( 1991 ) .
g. volpe , t. brettschneider , l. helden , and c. bechinger , opt .
express * 17 * , 23975 ( 2009 ) . c. hertlein , l. helden , a. gambassi , s. dietrich , and c. bechinger , nature * 451 * , 172 ( 2008 ) .
a. gambassi , a. maciolek , c. hertlein , u. nellen , l. helden , c. bechinger , and s. dietrich phys . rev .
e * 80 * , 061143 ( 2009 ) .
m. e. fisher and p. g. de gennes , c. r. acad .
paris ser .
b * 287 * , 207 ( 1978 )
. m. krech , _ casimir effect in critical systems _
( world scientific , singapore , 1994 ) ; j. phys .
: condens .
matter * 11 * r391 ( 1999 ) . j. g. brankov , d. m. dantchev , and n. s. tonchev , _ theory of critical phenomena in finite - size systems _ ( world scientific , singapore , 2000 ) .
m. kardar and r. golestanian , rev .
mod . phys .
* 71 * ( 1999 ) 1233 .
m. n. barber , in _ phase transition and critical phenomena _
, edited by c. domb and j. l. lebowitz ( academic , london , 1983 ) , vol . 8 , p. 145 . h. w. diehl , in _ phase transition and critical phenomena _ , edited by c. domb and j. l. lebowitz ( academic , london , 1986 ) , vol .
10 , p. 75 .
a. drzewi ' nski , a. macioek , and r. evans , phys .
lett . * 85 * , 3079 ( 2000 ) .
a. macioek , a. drzewi ' nski , and r. evans , phys .
e * 64 * , 056137 ( 2001 ) .
r. evans , u. marini bettolo marconi , and p. tarazona , j. chem .
* 84 * , 2376 ( 1986 ) .
f. schlesener , a. hanke , and s. dietrich , j. stat . phys . * 110 * , 981 ( 2003 ) .
r. pandit , m. schick , and m. wortis , phys .
b * 26 * , 5112 ( 1982 ) .
a. drzewi ' nski , a. macioek , and a. ciach , phys .
e * 61 * , 5009 ( 2000 ) .
w. thomson , phil . mag . * 42 * , 448 ( 1871 ) .
a. drzewi ' nski , a. macioek , a. barasi ' nski , and s. dietrich , phys .
e * 79 * , 041145 ( 2009 ) .
f. j. wegner , phys .
b * 5 * , 4529 ( 1972 ) . l. onsager , phys .
rev . * 65 * , 117 ( 1944 ) .
r. bulirsch and j. stoer , numer . math .
* 6 * , 413 ( 1964 ) . | using the quasi - exact density - matrix renormalization - group method we calculate the solvation forces in two - dimensional ising films of thickness @xmath0 subject to identical algebraically decaying boundary fields with various decay exponents @xmath1 . at the bulk critical point
the solvation force acquires a universal contribution which is long - ranged in @xmath0 due to the critical fluctuations , a phenomenon known as the critical casimir effect . for @xmath2 and @xmath3 ,
we study the scaling behaviour of the solvation force along the pseudo - phase coexistence and along the critical and sub - critical isotherms . |
recently , the study of the field - induced effects in low - dimensional quantum spin systems has been attracting much interest from theoretical and experimental points of view @xcite,@xcite,@xcite,@xcite . for the system with an anisotropy of exchange interactions the magnetic properties essentially depend on the direction of the applied magnetic field . for example
, the behavior of the one - dimensional antiferromagnetic @xmath5 model in a transverse magnetic field is drastically different in comparison with the case of the longitudinal field . in particular
, the transverse field induces the staggered magnetization in the perpendicular direction and the continuous phase transition takes places at some critical field @xcite,@xcite,@xcite,@xcite .
this effect has been observed in quasi - one - dimensional antiferromagnet @xmath6 @xcite , where the magnetic field has both the transverse and longitudinal components .
therefore , it is important to study the properties of the antiferromagnetic @xmath7 @xmath5 model in mixed transverse and longitudinal magnetic fields @xmath8 we consider the most simple case of this model the antiferromagnetic ising chain given by the hamiltonian @xmath9 in spite of the simple form of this hamiltonian it can not be solved exactly . in @xcite
the ground state phase diagram of this model has been investigated using numerical diagonalization of finite systems and finite size scaling procedure .
it was found @xcite that the transition line between the ordered and disordered phases exists and it was assumed that the model belongs to the universality class of the two - dimensional ising model .
we note that the ferromagnetic ising chain in the mixed fields has been studied intensively @xcite,@xcite,@xcite .
though the ferromagnetic and the antiferromagnetic ising chains in the transverse magnetic field are equivalent , properties of these two models are very different at @xmath10 .
( in fact , the model ( [ h ] ) can be transformed to the ferromagnetic chain but in a staggered longitudinal field . ) for example , the ground state phase transition in the ferromagnetic model is smeared out by the longitudinal field in contrast to the antiferromagnetic model for which the phase transition remains at @xmath10 . in this paper
we study the model ( [ h ] ) using the density matrix renormalization group ( dmrg ) technique @xcite .
this method allows us to consider the systems up to a few hundred sites and to determine the transition line with high accuracy . using the finite - size estimation of the ground state energy and low - lying excitations
we will show that the model ( [ h ] ) on the transition line is described by the conformal field theory with the central charge @xmath11 .
besides , we consider the behavior of the model in the vicinity of the special points @xmath12 , @xmath13 and @xmath14 , @xmath15 and determine the form of the transition line near these points .
the paper is organized as follows . in sec.ii
we will provide a qualitative physical picture of the ground state phase diagram based on the classical approach . in sec.iii and
sec.iv the behavior of the system in the vicinity of the special points of the critical line will be considered . in sec.v
we will present the dmrg calculation of the critical line .
sec.vi is devoted to the connection of the model ( [ h ] ) with the statistical quasi - one - dimensional ising model in the external magnetic field . in summary
we discuss our results .
in order to provide a physical picture of the phase diagram of the model ( [ h ] ) we use the classical approximation , when spins are represented as three - dimensional vectors .
the classical ground state is given by a configuration in which all spin vectors lie in the @xmath16 plane with the spins on odd and even sites pointing respectively at angles @xmath17 and @xmath18 with respect to the @xmath19 axis .
the classical energy of this state is @xmath20 the angles @xmath17 and @xmath21 , minimizing this energy are solutions of the following equations @xmath22 the solution of these equations is simple in the particular cases @xmath13 @xmath23 and @xmath14@xmath24 but in general case , when @xmath25 and @xmath10 , the phase diagram is divided on two regions ( see fig . [ fig_1 ] ) . in the paramagnetic region ( pm )
the energy minimum is given by the configuration with @xmath26 , with @xmath27 determined by the equation @xmath28 the antiferromagnetic ( af ) region is characterized by non - zero staggered magnetizations in both @xmath19 and @xmath29 directions @xmath30 with @xmath31 .
thus , in the af region the magnetic fields induce a perpendicular antiferromagnetic long - range order ( lro ) . in the classical approach the value @xmath32 plays the role of the lro and it vanishes on the transition line determined by the equations @xmath33 where @xmath34 the solution of eqs.([eqtransclass ] ) gives the transition line in the explicit form @xmath35 the classical phase diagram
is shown on fig .
[ fig_1 ] .
the transition ( critical ) line separates the phase with non - zero afm lro from the phase with uniform magnetization ( the paramagnetic phase ) .
of course , the classical approach does not give the correct description of the phase transition .
at first , the quantum fluctuations shift the critical point @xmath36 to @xmath12 at @xmath13 .
secondly , the form of the critical line is certainly incorrect at @xmath37 . besides , the order parameter @xmath32 vanishes on the critical line with the critical exponent @xmath38 , which is not valid at least in the critical point @xmath39
nevertheless , the fact of the generation of the staggered magnetizations perpendicular to the field direction is qualitatively true .
before the numerical determination of the critical line we consider the behavior of the model ( [ h ] ) in the vicinity of two special points : @xmath39 and @xmath4
. this will give us the form of the critical line near these points .
in the case @xmath13 the model ( [ h ] ) reduces to the exactly solvable one - dimensional ising model in the transverse field ( the itf model ) .
this model is well studied @xcite and it belongs to the universality class of the two - dimensional ising model . at @xmath40
the itf model is gapped and there is non - zero staggered magnetization @xmath41 .
the ground state at @xmath40 is two - fold degenerated in the thermodynamic limit . at the point
@xmath12 the model becomes gapless and the staggered magnetization vanishes with the critical exponent @xmath42 .
this is the transition point from the antiferromagnetic state to the paramagnetic gapped state .
the gap vanishes linearly at @xmath12 .
now we consider the itf model with small longitudinal magnetic field . for small magnetic field @xmath1
we rewrite the hamiltonian ( [ h ] ) in the form @xmath43 after jordan - wigner transformation to the fermi operators @xmath44 , @xmath45 @xmath46 the itf hamiltonian @xmath47 takes a bilinear form @xmath48 the hamiltonian @xmath47 commutes with the parity operator @xmath49 because @xmath47 can change the number of @xmath50 excitation by an even number only .
therefore , the space of states of @xmath47 is divided on two sectors with odd ( @xmath51 ) and even ( @xmath52 ) number of the fermi particles @xmath50 .
the hamiltonian @xmath47 is diagonalized exactly @xcite @xmath53 with fermi particles @xmath54 and spectrum @xmath55 which gives a gap at momentum @xmath56 @xmath57 the gap vanishes at the critical point @xmath58 , where the spectrum becomes @xmath59 below in this section we will consider the perturbation theory in @xmath60 ( [ hv0 ] ) in the critical point @xmath58 . the transition operator @xmath61 in @xmath60 ( [ hv0 ] ) conserves the momentum and changes the parity , because as it follows from eq.([jw ] ) it changes the number of the fermi particles @xmath62 ( and also the number of @xmath54 particles ) by odd number .
therefore , the non - zero matrix elements with the transition operator @xmath63 have the states with equal momentum but different parity .
the last fact means that the perturbation theory in @xmath60 contains only even orders .
the second order correction to the ground state energy is : @xmath64 the ground state @xmath65 has the momentum @xmath66 and zero number of the @xmath54 particles ( @xmath52 ) .
therefore , the non - zero contribution to the sum in ( [ de2 ] ) is given by the intermediate states @xmath67 with zero momentum and the parity @xmath51 .
as follows from eq.([aspectrum ] ) , all states @xmath67 with momentum @xmath66 and odd number of the @xmath54 particles ( @xmath51 ) have ` high ' energies @xmath68 . on the contrary , among the states with @xmath66 and even number of the @xmath54 particles ( @xmath52 )
there are many states like @xmath69 with small @xmath70 having small excitation energies @xmath71 and they can lead to infrared divergencies .
hereinafter we consider large but finite systems of length @xmath72 .
we shall study the dependence of dominant contributions to the perturbation theory on @xmath72 , omitting numerical factors .
using the fact that @xmath73 one can rewrite eq.([de2 ] ) as @xmath74 in higher orders of the perturbation series for the ground state energy each second intermediate state @xmath67 has odd number of the fermi particles ( @xmath51 ) , and , therefore , high energy @xmath75 .
for example , let us consider the fourth order correction to the ground state energy @xmath76 all intermediate states @xmath77 have momentum @xmath66 .
the states @xmath78 have @xmath52 and some of them have small excitation energies @xmath79 , while the states @xmath80 and @xmath81 have @xmath51 and high energies @xmath82 .
therefore , one can sum over the intermediate states @xmath80 and @xmath81 , which reduces eq.([de4 ] ) to @xmath83 now we note , that this expression looks like the second order correction and eq.([de22 ] ) looks like the first order correction to the ground state energy with a perturbation @xmath84 . in a similar way one can sum over all intermediate states with @xmath66 and @xmath51 in all orders of perturbation series . as a result
, we arrive at the perturbation theory with the effective perturbation @xmath85 we note , that the perturbation theory with the perturbation @xmath86 coincides with the original perturbation theory ( [ hv0 ] ) in a sense , that both perturbation series have the same order of divergencies ( or power of @xmath72 ) at each order in @xmath1 . but numerical factors at each order in @xmath1 can be different . the perturbation @xmath86 commutes with the parity operator ( [ parity ] ) and conserves the momentum .
therefore , the perturbation series in @xmath86 contains the intermediate states with @xmath66 and @xmath52 only , and some of these states have small excitation energies @xmath87 .
these states give the most divergent contribution to the perturbation series in @xmath86 and further we shall take into account these states only .
now we need to estimate the matrix elements of the operator @xmath88 .
the behavior of the correlation function @xmath89 in the ground state and in the low - lying states with excitation energies @xmath87 on large distances is known @xcite @xmath90 and , therefore , due to oscillation of the correlator @xmath91 the sum over @xmath92 and @xmath93 can be estimated as @xmath94 where the constant was found from extrapolation of the exact results for finite chains with @xmath95 .
the non - diagonal matrix elements of the operator @xmath96 with two different low - lying states was also calculated numerically .
it was found that the only non - zero matrix elements ( in the thermodynamic limit ) are given by the states @xmath97 differing by two @xmath54 particles , like @xmath98 .
all such pairs of states give the same value for the matrix element @xmath99 while all other matrix elements exponentially drop with @xmath72 .
now we estimate all terms of perturbation series in @xmath1 . according to eq.([v1diag ] ) the second order correction ( [ de22 ] ) is proportional to @xmath100 .
the factor at @xmath100 was found numerically by exact diagonalization of finite systems and the following calculation of the sum ( [ de2 ] ) .
the data for @xmath95 are well extrapolated and give @xmath101 according to eq.([de23 ] ) the zero - field susceptibility is @xmath102 .
this result is in a perfect agreement with the value @xmath103 obtained analytically in @xcite . using eq.([v1nodiag ] ) we find that the non - zero contribution to the fourth order correction to the ground state energy ( [ de42 ] ) is given by the states @xmath104 with the excitation energy @xmath105 .
thus , because of small excitation energies at @xmath106 in denominator in ( [ de42 ] ) the fourth order correction to the ground state energy turns out to be proportional to @xmath72 ( we omit here logarithmic corrections ) @xmath107 the @xmath93-th order in @xmath86 is proportional to @xmath108 .
the denominator of the @xmath93-th order contains @xmath109 small excitation energies @xmath110 and all matrix elements in the numerator are of the order of unity .
therefore , the @xmath93-th order in @xmath86 diverges as @xmath111 .
but this is not valid for odd orders in @xmath86 .
the analysis shows that all odd orders in @xmath86 diverge as @xmath112 .
for example , the third order correction in @xmath86 ( the sixth order in @xmath1 ) for the ground state energy does not diverge : @xmath113 so , the odd orders in @xmath86 give the next order corrections to the ground state energy and we omit it below .
summarize all above , we arrive at the perturbation series in the form @xmath114 with @xmath115 ( eq.([de23 ] ) ) and unknown constants @xmath116 .
one can see that the first divergence appears only in the eight - th order in @xmath1 , and it is very difficult to observe it numerically . the exact numerical calculations up to @xmath117 of the second , the fourth and the sixth order corrections in @xmath1 confirm the form of the series ( [ de0ser ] ) .
the sum in eq.([de0ser ] ) forms the scaling function @xmath118 of the scaling parameter @xmath119 .
thus , the ground state energy takes the form @xmath120 since the ground state energy is proportional to @xmath72 , the scaling function @xmath118 has finite thermodynamic limit at @xmath121 .
thus , the leading term of the perturbation theory for the ground state energy is given by the second order and the divergent part of the perturbation theory gives the correction @xmath122 .
the perturbation series for low - lying states has the same form as in eq.([de0 ] ) , but each low - lying state has its own scaling function .
therefore , for the first excited state ( with momentum @xmath123 ) one has @xmath124 so , the mass gap @xmath125 appears as : @xmath126 where @xmath127 and the scaling function @xmath128 .
since the gap is finite , the scaling function @xmath129 in the thermodynamic limit at @xmath121 must tend to some finite limit @xmath130 from the last equation we see that the gap is proportional to @xmath131 , but the factor at @xmath131 is given not only by the second order correction @xmath132 but by all collected divergent orders of the perturbation series .
the numerical estimation of the second order correction to the gap gives @xmath133 in order to find the factor at @xmath131 in mass gap we performed dmrg calculations ( for details see sec.v ) of the model ( [ hv0 ] ) for @xmath134 and @xmath135 .
the dependence of the gap on the scaling parameter @xmath119 is shown on fig .
[ fig_2 ] .
one can see that the points with different @xmath72 and @xmath1 lie perfectly on one curve .
it definitely manifests that the scaling parameter is @xmath100 .
the dmrg calculation data give the mass gap @xmath136 from this equation follows that the second order correction to the mass gap ( [ m2 ] ) gives approximately one half of a total gap , and another half is collected by all other divergent orders contained in the scaling function @xmath137 .
we note , that in contrast to eq.([mvalue ] ) for the ferromagnetic itf model the mass gap at the critical point @xmath12 is proportional to @xmath138 @xcite,@xcite .
let us return to the estimation of the diagonal element of the operator @xmath86 ( [ v1diag ] ) .
using the fact of the oscillation of the correlator ( [ zzcorr ] ) in the ground state and in the low - lying states we can rewrite the sum in ( [ v1diag ] ) approximately as a one half of the first term : @xmath139 the last equation suggests that the perturbation @xmath86 can be reduced to the operator @xmath140 @xmath141 with some constants @xmath142 , which we will define later . in order to verify this assumption
we compared the matrix elements of the operator @xmath140 with those of @xmath86 .
we have found that the dependence of the matrix elements of the operator @xmath140 on @xmath72 is in a full accord with eqs.([v1diag ] ) and ( [ v1nodiag ] ) .
that is , the diagonal matrix elements of @xmath140 are proportional to @xmath72 and the only non - zero ( non - diagonal ) matrix elements have the pairs of states differing by two @xmath54 particles @xmath143 thus , we arrive at the effective hamiltonian @xmath144 again , the original model ( [ hv0 ] ) is equivalent to the effective model ( [ hmap ] ) in a sense , that the perturbation series for both models have the same order of divergencies ( or power of @xmath72 ) at each order of @xmath1 .
but numerical factors at each order of @xmath1 can be different .
the advantage of this mapping lies in the fact that the effective model ( [ hmap ] ) is of the itf type and , therefore , is exactly solvable one .
the spectrum of @xmath145 is @xmath146 with @xmath147 .
so , according to the spectrum of the effective model ( [ hmapspectrum ] ) the magnetic field @xmath1 produces the gap in the spectrum @xmath148 as was noted above , apart from numerical factors the effective model ( [ hmap ] ) has the same form of perturbation series in @xmath1 as the original model ( [ hv0 ] ) .
therefore , we can check the form of the perturbation theory ( [ de0ser ] ) by studying the perturbation series in @xmath1 of the exactly solvable model ( [ hmap ] ) . for large but finite systems the ground state energy ( with @xmath66 ) and the first excited state ( with @xmath123 ) of the hamiltonian ( [ hmap ] ) are known exactly @xcite : @xmath149 where @xmath150
. the perturbation series for the ground state energy and the gap can be easily found by formal expansion of these expressions in small parameter @xmath151 .
this results in the scaling form of perturbation series for the mass gap : @xmath152 where constants @xmath153 determine the scaling function @xmath154 of a scaling parameter @xmath155 @xmath156
the scaling function @xmath154 was found in @xcite . in the thermodynamic limit @xmath121 the function @xmath157 , resulting in the gap ( [ gapmap ] )
one can see that the scaling parameter and the form of the perturbation series ( [ hmapser ] ) coincides with eq.([mser ] ) .
now we define the constants @xmath158 and @xmath159 in eq.([hmap ] ) , so that the second order corrections to the ground state energy and the gap of the effective model ( [ hmap ] ) numerically coincide with eqs.([de23 ] ) and ( [ m2 ] ) .
this leads to the values @xmath160 the mass gap of the effective model ( [ hmap ] ) with constants @xmath158 and @xmath159 defined above is @xmath161 surprisingly , the value of the mass gap of the effective model turns out to be very close to one found numerically for the original model ( eq.([mvalue ] ) ) .
it means , that the gap scaling function @xmath154 of the effective model ( [ hmap ] ) has the same thermodynamic limit as that for the model ( [ hv0 ] ) .
moreover , we have calculated numerically the fourth order corrections in @xmath1 to the ground state energy and the gap for the models ( [ hv0 ] ) and ( [ hmap ] ) and found perfect numerical agreement ( the second order corrections in @xmath1 for both models coincide by the definition ( [ ab ] ) ) .
therefore , we expect that the mapping of the model ( [ hv0 ] ) at @xmath58 to the model ( [ hmap ] ) with constants @xmath158 and @xmath159 defined in ( [ ab ] ) is exact for low - lying excitations in the thermodynamic limit . the mapping of the model ( [ hv0 ] ) to the effective itf model can be extended to the case @xmath162 . for this case
the effective hamiltonian becomes @xmath163 where now the constants @xmath142 are the functions of @xmath0 .
the functions @xmath164 and @xmath165 are defined in such a way that the second order corrections to the ground state energy and the gap for the effective model ( [ hmap ] ) coincide with those for the original model ( [ hv0 ] ) . using the effective hamiltonian ( [ hmaphx ] )
one can calculate the susceptibility @xmath166 . in the vicinity of the critical point @xmath58 it has a logarithmic singularity @xmath167 with @xmath168 from eq.([de23 ] ) and @xmath169 ( eq.([ab ] ) ) .
another exactly solvable limit of the model ( [ h ] ) is the case @xmath14 .
the model at @xmath14 is the classical one . at @xmath170 its ground state
is the antiferromagnet . at @xmath15
the first order phase transition to the ferromagnetic ground state occurs .
this is so - called multicritical point .
the hamiltonian ( [ h ] ) at @xmath15 has a form @xmath171 where @xmath172 .
the ground state of @xmath47 is macroscopic degenerate : all spin configurations , excluding those with two neighbor spins pointing down , have the same energy @xmath173 .
the number of these states is @xmath174 @xcite .
the transverse field @xmath0 lifts the degeneracy .
the exact calculation of the first order correction in @xmath0 for @xmath175 is rather complicated because it involves the exponentially large number of degenerate states .
we carried out the approximate calculation of the perturbation theory in @xmath60 within degenerate manifold using a simple variational function in the form @xmath176 where @xmath177 are the sum ( with equal weights ) of all admissible states with @xmath93 spins down : @xmath178 here @xmath179 is a projector excluding states with two neighbor down spins , @xmath180 is the ferromagnetic state with all spins up and the normalization factors @xmath181 are @xmath182 matrix elements of @xmath60 with respect to @xmath177 are @xmath183 where @xmath184 coefficients @xmath185 in ( [ psi ] ) obey the equations @xmath186 the quantity @xmath187 has a sharp maximum for @xmath175 at @xmath188 and @xmath189 is @xmath190 the ground state energy in the thermodynamic limit is defined as a lowest eigenvalue of eqs.([cm ] ) , which is @xmath191 in the frame of variational function approach ( [ psi ] ) the ground state magnetizations @xmath192 and @xmath193 are @xmath194 the function ( [ psi ] ) has a momentum @xmath66 . to calculate the spectrum in the first order in @xmath0 it is necessary to choose the variational function of the type ( [ psi ] ) with momentum @xmath195 .
we omit here rather cumbersome calculations , which shows that the mass gap corresponds to @xmath123 and @xmath196 the numerical diagonalization of finite cyclic systems shows very rapid exponential convergence to the thermodynamic limit .
the extrapolated values for the ground state energy and the mass gap ( at @xmath123 ) for @xmath37 are @xmath197 comparison of these results with eq.([ehx0 ] ) shows that the variational ground state energy differs from the ` exact ' one within @xmath198 . as follows from eq.([emhx0 ] ) the mass gap in the multicritical point opens linearly with @xmath0 , which is correctly described by the variational approach ( [ mvar ] ) .
in general , the critical line at @xmath25 and @xmath10 can not be found exactly . to obtain it we used the dmrg technique @xcite .
we have performed dmrg calculations using both the infinite - size and the finite - size dmrg algorithms .
we calculated the ground state energy @xmath199 and two lowest excitations @xmath200 and @xmath201 . in order to check the accuracy of the dmrg method we compared the obtained results with the exact ones for the itf model .
we used the infinite - size algorithm and open boundary conditions .
the dependence of the results on the number of retained states @xmath80 in the dmrg computation and on a number @xmath202 of dmrg steps ( @xmath203 ) has been investigated .
we have found that the calculation with @xmath204 gives satisfactory accuracy up to @xmath205 .
for example , relative errors in the ground state energy and in the mass gap at @xmath12 are @xmath206 and @xmath207 for @xmath208 .
the accuracy becomes better when the value @xmath209 is increased .
the critical field @xmath210 at a fixed value @xmath1 ( @xmath211 ) is determined by vanishing of the gaps @xmath212 and @xmath213 . below the critical field the mass gap @xmath214 exponentially with @xmath72 .
this behavior confirms the fact that the ground state is doubly degenerate in the thermodynamic limit at @xmath215 .
the true mass gap in this region is defined by the value of @xmath216 .
the typical behavior of the gaps @xmath212and @xmath213 extrapolated to the thermodynamic limit is shown on fig .
[ fig_3 ] .
we note that at @xmath217 the gaps @xmath212and @xmath213 coincide at @xmath218 . the critical line obtained from the dmrg results extrapolated to the thermodynamic limit
is shown on fig .
[ fig_1 ] . in the plane
@xmath219 it connects two limiting critical points studied in previous sections : the critical point ( @xmath39 ) of the itf model and the multicritical point ( @xmath4 ) .
now we discuss the properties of the model ( [ h ] ) near the critical line . the itf model in the critical point
is described by a conformal field theory with a central charge @xmath11 .
we expect that the itf model is generic for the model ( [ h ] ) on the whole critical line ( except the multicritical point @xmath4 ) . to verify this suggestion we estimated the value of @xmath220 on the critical line . for this we used well known fact @xcite
that for conformal invariant model with periodic boundary conditions the central charge appears at @xmath221 correction to the ground state energy @xmath222 where @xmath223 is the ground state energy per site in the thermodynamic limit and @xmath224 is the sound velocity .
at first , we calculated the sound velocity @xmath224 as ( similar to the spectrum of the itf model ( [ aspectrum ] ) the sound velocity on the whole critical line is determined at @xmath225 ) @xmath226 \label{vcft}\ ] ] we carried out these calculations using the numerical diagonalization of the hamiltonian ( [ h ] ) with the periodic boundary condition for @xmath227 .
the size extrapolation is carried out by formula @xmath228 .
the example of the extrapolation procedure is shown on fig .
[ fig_4 ] .
the dependence of @xmath224 on @xmath1 along the critical line is shown on fig .
[ fig_5 ] .
after that , the central charge @xmath220 has been calculated using eq.([egscft ] ) .
the size extrapolation shows that the central charge is @xmath229 for all calculated critical points .
the critical exponents @xmath230 of the ground state correlation function @xmath231 is defined by the scaling dimension @xmath19 of the operator @xmath232 and @xmath233 @xcite .
the scaling dimension is related to the finite - size correction of the lowest excited eigenstate which can be reached from the ground state by the operator @xmath232 .
the finite - size correction to this excitation energy is @xmath234 for the itf model at @xmath12 the sound velocity is @xmath235 .
the ground state has a momentum @xmath66 and the lowest excitation energy in the sector with momentum @xmath66 is @xmath236 and in the sector with @xmath123 is @xmath237 thus , the scaling dimensions are @xmath238 and @xmath239 .
the corresponding associated operators are @xmath240 which is in accord with the well known results @xcite for the asymptotic of the correlation functions @xmath241 using eq.([decft ] ) we numerically checked that the scaling dimensions related to the finite - size corrections of the lowest excited eigenstates on the critical line remain as in itf model @xmath238 and @xmath242 .
therefore , we conclude that the model ( [ h ] ) on the critical line belongs to the universality class of the itf model .
this means that in accord with the prediction of the classical approach to the left of the critical line the staggered magnetizations @xmath243 and @xmath244 exist .
but in contrast to the classical approach they vanish on the critical line with the critical exponent @xmath42 .
the mass gap is closed on the critical line and the critical exponent for the gap is equal to unity , i.e. the scaling behavior of the gap near the critical line is linear . for the particular case @xmath245 ,
this is illustrated on fig .
[ fig_6 ] . where the scaling plot for the scaled mass @xmath246 with @xmath247 is shown .
the behavior of the critical line near the itf point @xmath3 can be found from the following consideration .
as it was established above in the vicinity of the critical line the gap is proportional to deviation from the line .
this is valid for any direction of deviation except the direction at a tangent to the critical line . in the vicinity of the itf critical point for fixed @xmath248
according to eq.([h0gap ] ) the gap is @xmath249 . on the other hand ,
the gap is given by eq.([mvalue ] ) .
combining these two expressions for the gap on the line @xmath58 , we obtain equation for the critical line in the vicinity of the point @xmath3 as @xmath250 this expression for the critical line is in a very good agreement with our numerical estimations up to @xmath251 . in the vicinity of the multicritical point the form of the critical line
can be found in a similar way . at @xmath15 and @xmath37
the mass gap is proportional to @xmath0 ( see eq.([emhx0 ] ) ) . on the other hand , for fixed @xmath37
this gap is proportional to a deviation from the critical line @xmath252 .
therefore , the critical line near the multicritical point behaves as @xmath253 .
the numerical coefficient @xmath254 has been found from the dmrg calculation of the critical line in the vicinity of the multicritical point . as a result ,
the form of the critical line at @xmath37 is @xmath255 this form differs from that given in @xcite , where @xmath256 .
it is well - known that the critical properties of the two - dimensional ising model are equivalent to those of the one - dimensional itf model @xcite . using the formalism of ref.@xcite
one can show that the quantum hamiltonian ( [ h ] ) is related to the transfer matrix of the strong anisotropic ( quasi - one - dimensional ) ising model in the uniform magnetic field .
this model describes ferromagnetic ising chains weakly antiferromagnetically coupled with each other .
the hamiltonian of the model is @xmath257 where @xmath258 , @xmath259 .
the relation between the model ( [ h ] ) and the model ( [ h2d ] ) are given by @xmath260 with @xmath261 .
the equivalence of these two models means that the model ( [ h2d ] ) undergoes the phase transition from the ordered ( at @xmath262 ) to the disordered ( at @xmath263 ) phase .
the critical line @xmath264 of the quasi - one dimensional model are related to the critical line @xmath210 of the model ( [ h ] ) . by use of mapping ( [ 1d-2d ] )
we obtain that the critical temperature @xmath264 is @xmath265 in particular , at @xmath266 @xmath267 and @xmath268 when @xmath269 as @xmath270 the order parameter @xmath271 vanishes at @xmath272 with the critical exponent @xmath273 .
we note that the phase transition does not occur if @xmath274 as well as in the ferromagnetic version of the model ( [ h ] ) .
the free energy of the model ( [ h2d ] ) is related to the ground state energy of the hamiltonian ( [ h ] ) by @xmath275 according to eqs.([chi ] ) , ( [ 1d-2d ] ) and ( [ free_en ] ) the zero - field susceptibility in the vicinity of the critical temperature @xmath276 can be obtained using exactly solvable effective model ( [ hmap ] ) . at @xmath277
the susceptibility is @xmath278 the last equation has a form coinciding with the results obtained for the two - dimensional antiferromagnetic ising model by the series expansion method @xcite .
we have studied the antiferromagnetic ising chain in the mixed transverse and longitudinal magnetic field .
it was shown that the quantum phase transition existing in the itf model remains in the presence of the uniform longitudinal field . using the dmrg simulations we have found the critical line in the ( @xmath2 ) plane where the mass gap is closed and
the staggered magnetizations along the @xmath19 and @xmath29 axes vanish .
it is found numerically that the model on the critical line is described by the conformal field theory with the central charge @xmath279 , i.e. it belongs to the universality class of the itf model . the scaling behavior in the vicinity of the itf critical point is studied in detail .
it is shown that the mass gap is proportional to @xmath131 and the contributions to it are given not only by the second order correction but also by all other divergent orders of the perturbation series in @xmath1 . besides , the analysis of the perturbation theory shows that the considered model at @xmath248 can be mapped to the effective itf model with renormalized parameters depending on @xmath1 . in a framework of the effective itf model the behavior of the susceptibility at @xmath280 is determined .
the behavior of the model in the vicinity of the multicritical point is investigated . using both the variational approach and the numerical diagonalization results we have found that the mass gap is proportional to @xmath0 . close to the multicritical point
the form of the critical line is linear .
of course , the considered model is the simplest case of the @xmath5 model ( [ hxxz ] ) .
it is interesting to extend the present analysis to this model as well as to study effects of interchain interactions in quasi - one - dimensional generalization of the model ( [ h ] ) .
the mapping of the quantum model ( [ h ] ) to the strongly anisotropic statistical two - dimensional ising model in the uniform magnetic field was considered .
the behavior of the susceptibility of this model near the critical temperature is found .
we expect also , that the main features of the considered model ( [ h ] ) are valid for the statistical two - dimensional antiferromagnetic ising model in the uniform magnetic field @xmath281 .
that is , the applied magnetic field does not smear the phase transition existing in the two - dimensional ising model , which is generic case for the whole transition line @xmath264 in the plane ( @xmath282 ) .
in particular , we believe that the analysis of the perturbation series in @xmath283 in the same manner as was done in sec.iii gives the scaling behavior for the correlation length and the form of the critical line near the point @xmath276 is @xmath284 . | we have studied the antiferromagnetic ising chain in a transverse magnetic field @xmath0 and uniform longitudinal field @xmath1 . using the density matrix renormalization group calculation combined with a finite - size scaling the ground state phase diagram in ( @xmath2 ) plane is determined .
it is shown that there is an order - disordered transition line in this plane and the critical properties belong to the universality class of the two - dimensional ising model .
based on the perturbation theory in @xmath1 the scaling behavior of the mass gap in the vicinity of the critical point ( @xmath3 ) is established .
it is found that the form of the transition line near the classical multicritical point ( @xmath4 ) is linear .
the connection of the considered quantum model with the quasi - one - dimensional classical ising model in the magnetic field is discussed . |
systems with triangular lattice structure have been a source of attention mostly due to the frustration effects resulting from their non - bipartite structure . as a result of the competition between the frustration and strong electron correlations
, these systems exhibit a wide range of exotic phases .
recent studies of the metal - insulator transition , superconductivity , and antiferromagnetism in the organic compounds @xmath4-(bedt - ttf)@xmath5x with x as an anion @xcite , discovery of superconductivity in na@[email protected]@xmath5o,@xcite and the recent discovery of the mott transitions in 0.33 monolayers of sn on ge(111 ) at @xmath7 k,@xcite as a few examples , underline the importance of these systems and their physics .
theoretical work has been dedicated to the discovery of the magnetically ordered phases in the ground state of the hubbard and @xmath8 models on triangular lattices as a function of the on - site electron - electron coulomb interaction @xmath2 at different fillings.@xcite unlike the square lattice at half - filling , which is a mott insulator with antiferromagnetic order at arbitrarily small values of @xmath9 , the ground state of a triangular lattice has a variety of magnetically ordered and disordered phases .
this is due to the lack of perfect nesting in the non - interacting fermi surface of a triangular lattice at half - filling . for triangular lattices , according to the hartree - fock calculations of krishnamurthy and co - workers , @xcite the mott transition occurs from a paramagnetic metal to a paramagnetic insulator at half - filling for values of @xmath2 larger than the band width @xmath10 .
a variety of the physical properties of the triangular lattices including tendencies towards superconductivity within the small to intermediate @xmath2 regime have been studied at finite temperature using correlated electron approaches which go beyond mean field theory , such as the fluctuation exchange approximation ( flex)@xcite or one - loop renormalization - group.@xcite however , when @xmath11 , a more powerful cluster solving technique such as the quantum monte carlo ( qmc ) is required to accurately describe the phase transition .
finite size lattice determinant quantum monte carlo ( dqmc ) has already been employed for triangular and kagom lattices.@xcite unfortunately , the finite size lattice qmc method incurs sign problems at both low temperatures and away from half - filling .
in addition , due to the finite size nature of the problem , the non - local correlations are always overestimated and the system consisting of a finite size cluster becomes an insulator as soon as the correlation length reaches the size of the lattice at low enough @xmath3 .
thus , the gaps or pseudogaps are also overestimated and are not guaranteed to persist as one reaches the thermodynamic limit.@xcite however , finite size effects can be reduced by size scaling and in general become less important as @xmath2 increases . in this work ,
we employ the dynamical mean - field theory ( dmft)@xcite to study the mott transition for the hubbard model on a triangular lattice at half - filling .
dmft describes systems in the thermodynamic limit and at the same time suppresses the physics due to the non - local correlations through coarse - graining over all momentum degrees of freedom .
therefore , susceptibilities can diverge , as in the thermodynamic limit but the mott gap in the strong - coupling regime is always smaller than its counterpart given by a finite size lattice approach .
the combination of the dmft and qmc also does not incur sign problems at either low temperatures or any fillings which makes the low temperature physics more accessible and reliable .
dmft is a more suitable approximation for the triangular lattice systems as opposed to those with square lattice structure .
for instance , dmft is incapable of describing the metal - insulator transition on the square lattice in the weak coupling regime .
this is a consequence of the suppression of the short - range antiferromagnetic correlations responsible for this transition at small @xmath2 by the single site nature of dmft.@xcite however , in triangular lattices , due to the frustration of the lattice structure , even in the presence of non - local correlations , the short - range antiferromagnetic correlations are greatly suppressed . as a result , heavy quasiparticles form at the fermi energy reminiscent of the dmft physics.@xcite dmft has already been employed for studying triangular lattices using the exact diagonalization and lanczos techniques.@xcite in this paper , we present for the first time , the combination of dmft and qmc for solving the hubbard model at half - filling on a triangular lattice . by computing the density of states at different values of interaction @xmath2 and temperature @xmath3
, we estimate the value of the critical interaction for the mott transition to be @xmath12 , less than the predicted value of @xmath13 in ref . , but , as expected , larger than the mean field result .
in addition , the mott transition features are also manifest in the evolution of the magnetic moment , total internal energy and the specific heat as functions of on - site interaction @xmath2 and temperature @xmath3 .
we consider the repulsive hubbard model on a triangular lattice , @xmath14 with @xmath1 the hopping amplitude , @xmath15 the chemical potential and @xmath2 the on - site coulomb interaction between the electrons of opposite spins residing on lattice site @xmath16 . note that unlike the square lattice , on a triangular lattice , due to the lack of particle - hole symmetry , @xmath17 does not yield half - filling .
also , this model exhibits quite different physics under the change of the sign of @xmath1 , especially away from half - filling .
the dispersion in the @xmath18 limit including the nearest neighbor hopping only is , @xmath19\,.\end{aligned}\ ] ] the resulting noninteracting density of states is markedly asymmetric about @xmath20 , extending from @xmath21 to @xmath22 , with a van hove singularity at @xmath23 . the dmft approach to tight binding hamiltonians starts by coarse - graining the green s function , @xmath24 where @xmath25 is the matsubara fermionic frequency and @xmath26 the local self - energy for spin @xmath27 .
note that the self - energy @xmath26 in dmft is local ( momentum independent ) . in the first dmft cycle iteration
, we chose @xmath28 .
the bath green s function is then computed , @xmath29 @xmath30 is used as an input to the qmc solver .
our dmft - qmc solver is based on the hirsch - fye algorithm @xcite in which the interacting two - body term in the hubbard model is decomposed into two one - body density terms coupled to a hubbard - stratonovich field . through the stochastic averaging over the most probable configurations of these ising like fields , one calculates the fully interacting green s function on a single site of a lattice @xmath31 .
the local self - energy @xmath26 is then calculated from the full and bath green s functions @xmath31 and @xmath30 respectively , @xmath32 this self - energy is again inserted in eq.[eq : green - cg ] , and the self - consistent process continues until convergence .
the final product of the dmft - qmc cycle is the full green s function in imaginary time @xmath33 which is used to compute the density of states ( dos ) @xmath34 , @xmath35 @xmath34 must be calculated by inverting eq.[eq : gtau ] , which is an ill - posed problem due to the statistical errors in @xmath33 from the qmc simulations .
this task in this paper is done by using the maximum entropy ( maxent ) analytical continuation technique developed by jarrell and gubernatis.@xcite this technique is based on bayesian inference in which an entropy function with respect to an original default model is defined as a function of @xmath34 .
the best @xmath34 is the one that maximizes this entropy function for a given @xmath33 .
we also present results for the double occupation , @xmath36 which is suppressed by the on - site repulsion .
the double occupation is related to the magnetic moment , @xmath37 the total internal energy can also be computed in this formalism in terms of the full green s function and double occupation : @xmath38e^{i\omega_n0^{+}}+ud\,,\ ] ] with @xmath39 defined as in eq.[eq : dbleoccp ] .
eq.[eq : int - eng ] is equivalent to the galitskii - migdal expression for the total internal energy.@xcite the specific heat at constant volume is @xmath40 as seen later in this article , the variation of the total internal energy and specific heat as a function of @xmath2 and @xmath3 will exhibit features reflecting the mott metal - insulator phase transition .
before presenting our data for the density of states , we comment on some of the details of the maximum entropy procedure , since careful characterization of the data is essential to obtain reliable results . in this work
, the dmft - qmc cycle as described in section [ sec : formalism ] converges at a desired tolerance which is about few percent for the time green s function @xmath33 .
the chemical potential is also adjusted simultaneously to keep the total occupation at half - filling . for
the qmc runs , we adopt the criterion @xmath41 with @xmath42 for finding @xmath43 , the number of the time slices .
@xmath43 can be small at high temperatures , however , even for temperatures as high as the bandwidth , we choose the lower limit of @xmath44 . the number of the lattice sweeps for the qmc runs varies between @xmath45 to @xmath46 depending on the values of @xmath2 and @xmath3 and the reliability of the maxent results .
the outputs of the cycle are the fully interacting time green s function , the magnetic moment on a single site and total internal energy for the temperature and interaction at which the qmc cycle has been running .
the full @xmath33 is employed to calculate the dos through eq.[eq : gtau ] using the maxent technique .
@xcite care must be taken when using maxent as the results can strongly depend on the quality of the qmc data .
maxent , as described in ref .
, requires statistically independent and gaussianly distributed measurements of @xmath33 for every time slice @xmath47 from qmc .
both these two properties are often lacking in the @xmath33 data initially obtained from the qmc output because of the presence of correlations between the measurements .
the solution is a careful rebinning of the @xmath33 data which eliminates these correlations and yields gaussianly distributed data .
correlations between different time slices are also eliminated in the maxent code by diagonalizing a covariance matrix coupling different time slices together.@xcite the other concern is a proper choice for the maxent default model . in general , for good qmc data , the choice of the default model should not change the qualitative features of the spectral function significantly . among the available default models ,
the gaussian and second order perturbation models have been most commonly utilized . in our work , we adopt the latter and perform the so called annealing method in which we start with the second - order perturbation theory default model at a high temperature close to the band width .
we then use the maxent output for that temperature as the default model for a slightly lower temperature and continue until we arrive at the lowest desired @xmath3 .
finally , as introduced in ref . , there are different methods of doing the maxent including the historic , classic and bryan .
we utilized both classic and bryan techniques in our calculations and found that both these two methods give qualitatively similar results . .
for clarity , the base lines have been shifted in steps of @xmath48 for different values of @xmath2 .
the @xmath49 line determines the location of the fermi energy .
three behaviors are visible : at weak coupling the density of states has a finite value at @xmath49 . at large coupling the density of states vanishes , signaling the opening of the mott insulating gap . at intermediate coupling the density of states
exhibits a kondo resonance at @xmath49 .
, width=307 ] in fig.[dos.evolve ] , we present the evolution of the dos as a function of @xmath2 at low temperatures . at @xmath18 , the dos shows a van hove singularity near @xmath50 .
as the interaction is gradually turned on , there is an overall broadening of the density of states , including a smearing out of the van hove singularity .
the dos at @xmath51 and @xmath52 exhibits the onset of the abrikosov - suhl ( kondo ) resonance in which there appears a sharp quasi - particle peak near the fermi energy with two shoulders , the upper and lower hubbard bands , around it . for @xmath53 and @xmath52 , the peak lines up with the fermi energy and shoulders
are approximately located at @xmath54 . at @xmath55 , and @xmath56
, we clearly observe a different trend in which the kondo resonance is absent and the dos has a pronounced minimum at the fermi surface , indicating that the system has been driven to the metal - insulator transition regime . for @xmath57 , and @xmath58 , the ( mott ) gap is well established and the system becomes an insulator . at half - filling and different temperatures . as the temperature
is lowered a sharp kondo resonance develops at @xmath49 .
the system is metallic .
panel ( b ) : the same results as for panel ( a ) for @xmath55 .
the system is on the verge of insulating behavior .
panel ( c ) : @xmath57 .
a mott gap has opened at @xmath58.,width=268 ] at half - filling .
@xmath59 drops down close to its minimum @xmath60 at large temperatures ( cut off in this figure ) . the comparison between the dqmc and dmft at @xmath61 and @xmath62
is also presented .
the inset shows how @xmath63 , the temperature for the maximum @xmath59 , drops as @xmath2 increases.,width=326 ] to better address the role played by temperature in the metal - insulator transition , in fig.[dos.compare ] the dos for three values of @xmath53 , @xmath64 and @xmath65 and different temperatures have been plotted . at @xmath53 ,
as @xmath3 decreases , the dos evolves towards a sharp abrikosov - suhl resonant quasi - particle peak .
however , @xmath55 exhibits a completely different trend .
as the temperature drops , the dos evolves towards the opening of a gap . at @xmath56
, we noticed that the maxent results are of less reliability due to the correlations in the qmc data which could not be totally removed up to @xmath66 qmc lattice sweeps by rebinning the data .
( it is , of course , typical of monte carlo simulations that statistical fluctuations are largest near phase transitions ) this is additional evidence that @xmath55 is close to the critical value for the opening of a gap but we were not able to determine conclusively whether the gap exists at @xmath55 if one can reach low enough temperatures . on the other hand , by the time @xmath57 ( at @xmath58 ) , the maxent results show a pronounced mott gap in the dos .
therefore , it is clear that the critical value of @xmath67 must lie very close to @xmath55 , leading to our estimate @xmath68 .
this value is smaller than the value @xmath69 obtained in ref . using the dmft with exact diagonalization and lanczos techniques as the impurity solvers , and comparable to @xmath70 in ref . with exact diagonalization on @xmath71 site lattices .
at both values of @xmath55 and @xmath65 , we also observe the appearance of a feature on the right hand side of @xmath49 .
this feature is likely a remnant of the peak in the non - interacting density of states it disappears as @xmath2 is increased to larger values and is completely absent for @xmath72 .
another signature of the metal - insulator ( mott ) transition is manifest in the behavior of the magnetic moment defined in eq.[eq : magmom ] as a function of temperature @xmath3 and interaction @xmath2 . as shown in fig.[mag - mom - t ] , for small to intermediate values of @xmath2
, @xmath59 reaches a maximum at a value of @xmath63 which is characteristic of a fermi - liquid dominated by spin fluctuations at low temperatures.@xcite at low temperatures with an entropy of @xmath73 per particle , a fermi liquid gains free energy upon heating by trying to localize the electrons in order to take advantage of a larger spin entropy.@xcite the inset in fig.[mag - mom - t ] shows how @xmath63 decreases with @xmath2 .
however , at the onset of the mott transition , @xmath59 becomes a monotonically decreasing function of @xmath3 as seen in fig .
[ mag - mom - t ] for @xmath55 and @xmath65 . in a localized mott phase ,
the spin entropy at @xmath74 is finite and the system can not gain free energy by further localizing the electrons upon heating .
comparison has also been made between dqmc on a @xmath75 finite size lattice and dmft in fig.[mag - mom - t ] for values of @xmath61 and @xmath62 .
the agreement between the two approaches is very close at @xmath61 down to @xmath76 .
however , for @xmath51 , at low temperatures , the dqmc approach exhibits a rather flat region for @xmath59 below @xmath77 in place of the downward curve in dmft .
this flat feature is similar to what is observed at larger values of @xmath2 in dmft as a precursor to the insulating phase .
this is typical of any finite size lattice approach such as dqmc in which at low enough temperatures , once the correlation length exceeds the size of the lattice , the system turns insulating . in the dmft , on the other hand , the system is in the thermodynamic limit and therefore , the insulating phase features do not appear except at large enough interaction @xmath2 values . nevertheless , dqmc has the advantage of including the non - local correlations which are ignored in dmft .
these correlations work in favor of the gap formation .
thus , dmft overestimates the value of @xmath2 for the formation of the gap in the mott transition .
the evolution of the total internal energy and specific heat defined in eq.[eq : int - eng ] and eq.[eq : spec - heat ] respectively also exhibits signatures of a metal - insulator phase transition at large interaction @xmath2 values and low temperatures @xmath3 . in fig.[thermodynamics ] , panel ( a ) , the total internal energy has been plotted as a function of temperature @xmath3 for different values of interaction @xmath2 .
we have employed a polynomial fitting procedure developed by duffy and moreo @xcite and obtained the curves presented in panel ( b ) . by analytically differentiating the polynomial fits with respect to temperature ,
we arrive at results for the specific heat plotted in panel ( c ) . for values of @xmath61 to @xmath78 , we observe a two - peak structure in the @xmath79 consistent with the results in ref . for the infinite dimensional and @xmath80 hypercubic hubbard models respectively .
the peak at low temperatures corresponds to the local spin fluctuations analogous to the formation of the kondo singlets in the kondo model .
it is also a signature for the validity of the fermi liquid picture as it indicates the existence of a conducting electron bath around the magnetic impurity.@xcite the upper peak at high temperatures , however , corresponds to on - site charge fluctuations .
separation between these two peaks occurs at @xmath61 and develops up to @xmath51 resulting in a sharp lower peak . as prescribed in ref .
, the lower and higher temperature segments of @xmath81 have been fit with two different polynomials which meet around @xmath82 .
this prescription explains the rather sharp feature in the @xmath79 curve at @xmath61 in fig.[thermodynamics ] , panel ( c ) , around @xmath83 .
however , the necessity of using two polynomials originates in the existence of two separate temperature regions , one dominated by spin and the other by charge fluctuations . the lower peak dwindles at @xmath53 as a sign of approaching the insulating phase.@xcite it is also striking that , similar to the result of ref .
, within the fermi liquid range up to @xmath53 , all @xmath79 curves more or less intersect at the same temperature close to @xmath84 .
the picture however is different at @xmath57 when the insulating phase sets in .
the lower peak completely vanishes and the @xmath79 curve goes to zero at a finite value of @xmath3 .
this is consistent with the behavior of @xmath79 in the presence of a gap @xmath85 going as @xmath86 at low temperatures , and having a broad peak at high temperatures resulting from charge fluctuations in the upper hubbard band.@xcite for @xmath55 , @xmath87 results were too noisy to be conclusive and therefore are not presented .
this was apparently due to @xmath55 being close to the critical @xmath67 which made it difficult to achieve convergence in the simulation .
lastly , fig.[eint - t - comp ] presents a comparison between dmft and dqmc for @xmath81 at @xmath61 and @xmath62 .
similar to the trend observed for @xmath59 in fig.[mag - mom - t ] , at high enough @xmath3 , the agreement betweeen the two approaches is clearly manifest in the results . for @xmath61 at low @xmath3 , dqmc exhibits an insulating phase feature in the rather flat tail of the @xmath88 curve near @xmath89 , typical of a finite size approach as discussed earlier . for @xmath51
, however , the effect of non - local correlations becomes more important and at the same time , finite size effects diminish .
hence , dqmc features may be more realistic in showing insulating phase signatures .
nevertheless , @xmath51 is still below the bandwidth @xmath10 and we do not expect to have entered the insulating phase as yet . unfortunately , mostly due to the sign problem in dqmc , attempts in going to lower temperatures and higher @xmath2 values have not yet been successful .
thus , it is not clear to us which approach more plausibly describes the physics for the intermediate value of @xmath51 . for different values of interaction @xmath2 including the error bars .
panel ( b ) : polynomial fits to data in panel ( a ) . panel ( c ) : the specific heat as a function of temperature @xmath3 for different values of interaction @xmath2 taken by analytical differentiation of polynomials in panel ( b ) .
, width=288 ] finite size lattice for the total internal energy versus temperature @xmath3 for interaction values of @xmath61 and @xmath62.,width=297 ] in order to bring the dqmc and dmft treatments into full agreement , it would be necessary to extrapolate both to a zero value of the discretization size of the inverse temperature @xmath90 and also to extrapolate dqmc and dmft to infinite spatial cluster size and full momentum resolution ( the dynamical cluster approximation ( dca ) ) , respectively .
in summary , we have performed the first implementation of the dmft using qmc as the solver for the half - filled triangular lattice hubbard model .
our approach confirms the previous predictions that for large enough @xmath91 , there is a metal - insulator ( mott ) phase transition from a paramagnetic metal to a paramagnetic insulator at finite temperature .
we demonstrate this phase transition by presenting the behavior of the dos , magnetic moment on a single site , total internal energy and specific heat as functions of the interaction @xmath2 and temperature @xmath3 .
our results suggest a critical value of interaction @xmath92 .
this is consistent with the value @xmath70 from exact diagonalization on @xmath71 site lattices @xcite and @xmath69 from dmft with exact diagonalization and lanczos impurity solvers.@xcite .
we have also investigated the intermediate interacting regime which is characterized by the presence of an abrikosov - suhl ( kondo ) resonance .
dmft is a more suitable approximation for studying the mott transition on a triangular lattice compared to a square lattice .
the reason is that in a triangular lattice , due to the frustration of the lattice structure , the short - range antiferromagnetic correlations are suppressed and do not play a major role in the metal - insulator transition at small @xmath2.@xcite hence , neglecting these correlations , as done by dmft , should still give us a qualitatively correct physical picture at least in the large @xmath2 limit .
the combination of the dmft and qmc also does not encounter the difficulties related to the sign problem at low temperatures and away from half - filling and also finite size effects associated with dqmc at low @xmath3 .
thus , the low temperature physics is better described by the dmft due to being in the thermodynamic limit . nevertheless , dqmc has the advantage of incorporating non - local correlations which act in favor of gap formation and dmft overestimates the value of @xmath67 because of lacking these correlations . due to the single site nature of the dmft , we were not able to address the physics of magnetically ordered phases for triangular lattices as one approaches the @xmath89 limit .
these studies have already been performed using the dqmc techniques.@xcite however , the combination of the dynamical cluster approximation ( dca ) and qmc @xcite can also be considered a promising future candidate for these studies as a complementary approach to dqmc .
work on different magnetically ordered phases at different coupling regimes , including ferromagnetism away from half - filling as discussed in ref .
is in progress .
we acknowledge useful conversations with m. jarrell , a. macridin , and s. savrasov .
this research was supported by doe de - fg03 - 03na00071 . and
de - fg01 - 06na26204 . v. m. galitskii and a. b. migdal , zh .
. fiz . * 34 * 139 ( 1958 ) [ sov . phys . *
jetp 7 * , 96 ( 1958 ) ] ; a. l. fetter and j. d. walecka , _ quantum theory of many - particle systems _ ( mcgraw - hill , new york , 1971 ) , sec . 7 . for the @xmath93 square lattice , a two - peak structure in @xmath94 persists for all interaction strengths .
see t. paiva , c. huscroft , a. k. mcmahan , and r. t. scalettar , phys .
b * 63 * , 125116 ( 2001 ) . | we employ dynamical mean field theory ( dmft ) with a quantum monte carlo ( qmc ) atomic solver to investigate the finite temperature mott transition in the hubbard model with the nearest neighbor hopping on a triangular lattice at half - filling .
we estimate the value of the critical interaction to be @xmath0 in units of the hopping amplitude @xmath1 through the evolution of the magnetic moment , spectral function , internal energy and specific heat as the interaction @xmath2 and temperature @xmath3 are varied .
this work also presents a comparison between dmft and finite size determinant quantum monte carlo ( dqmc ) and a discussion of the advantages and limitations of both methods .
[ theorem]acknowledgement [ theorem]algorithm [ theorem]axiom [ theorem]claim [ theorem]conclusion [ theorem]condition [ theorem]conjecture [ theorem]corollary [ theorem]criterion [ theorem]definition [ theorem]example [ theorem]exercise [ theorem]lemma [ theorem]notation [ theorem]problem [ theorem]proposition [ theorem]remark [ theorem]solution [ theorem]summary i*i * *p * |
binary neutron star mergers ( nsms ) are unique laboratories for the study of astrophysics .
the merger involves many elements of the theory of relativistic astrophysics , gravitational wave astronomy , and nuclear astrophysics .
furthermore , nsms are thought to be a possible source of detectable gravitational radiation , the _
r_-process elements , and gamma ray bursts . in order to develop accurate models of nsms one
must numerically solve the equations describing gas dynamics and the gravitational field arising from the matter .
however , the need for accurate numerical models of the inspiraling binary system presents some unique challenges that we address in this paper . in realistic astrophysical situations
the merger of binary neutron star systems is driven by gravitational radiation losses ( misner et al . 1973 ) .
this loss of energy will lead to the inspiral and eventual coalescence of the binary system .
the prediction of energy loss by gravitational radiation was confirmed by the observation of psr1913 + 16 , a binary neutron star system ( hulse & taylor 1975 ) . the observed rate of decrease of the orbital period of this system is in good agreement with predictions made by general relativity ( taylor & weisberg 1989 ) .
coalescing binary systems are expected to emit tremendous amounts of energy in the form of gravitational waves during the final stages of coalescence , and the gravitational waves produced in these events are expected to be observed by gravitational wave detectors currently under construction .
gravitational wave interferometers such as ligo ( abramovici et al .
1992 ) and virgo ( bradaschia et al . 1990 ) will soon be operational and will present the first opportunities to study nsms via gravitational waves .
theoretical templates of the expected signal are required to extract signal information from the noisy background ( cutler et al .
post - newtonian methods ( lincoln & will 1990 ) may be adequate for the prediction of waveforms for the early stage of the inspiral . however , the prediction of waveforms in the later stages of the merger , when tidal effects and neutron star structure become important , requires a full three - dimensional numerical solution of the equations describing the motion of matter and the gravitational field .
in addition to gravitational wave astronomy , nsms are of interest for nuclear astrophysical reasons .
nsms may yield information about the structure of neutron stars .
since the equation of state ( eos ) of neutron star matter is not well constrained , the observation of gravitational wave signals from nsms may provide constraints that could provide information about the dynamics of the merger and in turn the eos of dense matter .
additionally , the material ejected during the coalescence of binary neutron stars may be a site of _ r_-process nucleosynthesis ( lattimer et al . 1977 ) .
the _ r_-process , which is thought to be responsible for the production of about 50% of elements heavier than iron in the universe , occurs when the capture rate of neutrons bombarding nuclei exceeds the beta decay rate . in the material ejected during nsms there
are expected to be regions where the _ r_-process occurs robustly ( meyer 1989 ) .
simulations of nsms can allow us to study both the mass ejection and nucleosynthesis that occurs in the ejected material .
nsms are a suggested source for the mysterious gamma - ray bursts observed by cgro and other high - energy observational missions .
nsms are thought to release energy on the order of their gravitational binding energy @xmath0 erg , which may be larger than estimated gamma - ray burst energies @xmath1 ( quashnock 1996,rees 1997 ) to @xmath2 erg ( woods & loeb 1994 ) .
a popular model for bursts at cosmological distances is the relativistic fire - ball ( paczyski 1986 , goodman 1986 , shemi & piran 1990 , paczyski 1990 ) .
nsms are likely candidates for the source of relativistic fire - balls , but the mechanism by which the fire - ball develops has yet to be determined .
observations in the spring of 1997 of optical and x - ray counterparts to grb 970228 ( costa , et al .
1997 , guarnierni 1997 , piro et al . 1997 ) and grb 970508 ( bond 1997 , djorgovski et al .
1997 , metzger et al .
1997 ) , particularly the measurement of a mg ii absorption line at redshift @xmath3 ( metzger et al . 1997 ) , suggest that the bursts do indeed have a cosmological origin .
simulations of nsms can test the consistency of the energetics and time scales with the estimated energies and observed time scales of the observed bursts .
one of the major difficulties in carrying out numerical simulations of binary neutron star mergers is developing a numerical algorithm that does not introduce unphysical dynamical effects into the problem . in order to avoid spurious inspiral , both
the total energy and angular momentum must be conserved to sufficient accuracy . in the absence of physical instabilities or dissipative effects , such as gravitational radiation losses
, the numerical methods should be capable of maintaining a binary neutron star system in a stable orbit .
additionally , when physical instabilities capable of causing coalescence are present , the algorithms must continue to conserve all important physical quantities . without this capability
it is impossible to develop quantitatively accurate models in situations where radiation losses are present . in this paper
we consider several variations on the popular zeus hydrodynamic algorithm ( stone & norman 1992 ) as applied to nsms .
the comparisons that will appear later in this paper examine the numerical effects of the choice of rotating versus inertial frames as well as the choice of several possible schemes for the coupling of gravity to the hydrodynamics .
this paper is intended to lay the numerical groundwork for post - newtonian and relativistic studies that will follow in later papers . while realistic models of nsms are clearly relativistic or , at a bare minimum , post - newtonian ( pn ) in nature , the examination of newtonian models of orbiting binaries neutron stars is still of considerable value . many of the lessons learned from newtonian models will provide guidance for pn or gr modeling efforts . indeed ,
if a numerical self - gravitating hydrodynamics algorithm is incapable of maintaining stable orbits for binary star systems in the newtonian limit , then it is unlikely to be useful for more complex , and realistic , simulations of nsms . in this paper
we concentrate on purely newtonian models of orbiting binary neutron stars in both the stable and unstable regimes .
we will consider the evolution of initial configurations that are tidally unstable as well as initial configurations involving both spherical and `` relaxed '' neutron stars .
post - newtonian models that make use of our numerical techniques will be considered in a subsequent paper .
the newtonian and pn simulations that have been carried out to date can be placed into two categories : those that have employed eulerian hydrodynamic methods ( see bowers & wilson 1991 for a discussion of eulerian methods ) and those that have employed smoothed particle hydrodynamics ( sph ) methods ( see gingold & monaghan 1977 or hernquist & katz 1989 for a discussion of sph methods ) , an inherently lagrangean technique .
additionally , these simulations have been carried out in both inertial , i.e. laboratory , frames and in non - inertial , rotating frames .
these simulations have also utilized a range of techniques for calculating the gravitational potential .
finally , these simulations have employed both spherical stars and equilibrium binary configurations as initial data .
these choices can play a critical role in determining the outcome of the simulations .
for this reason , in this section we briefly describe existing work on newtonian and pn binary neutron star systems with a focus on the numerical techniques and the initial configurations that have been used .
we first consider the eulerian calculations followed by the sph models .
the earliest eulerian models of binary neutron star systems were carried out by oohara and nakamura ( 1989 ) .
this work and subsequent papers ( nakamura & oohara 1989 , oohara & nakamura 1990 , nakamura & oohara 1990 , nakamura & oohara 1991 , oohara & nakamura 1992 ) made use of purely newtonian hydrodynamics , while later work with shibata included pn effects ( oohara & nakamura 1992 , shibata , nakamura , & oohara 1992,1993 ) .
as with all eulerian , i.e. grid - based , hydrodynamics methods the underlying pdes are discretized onto a coordinate mesh .
the evolution of the mass distribution occurs as the material flows through the grid zones , and the equations governing the evolution are finite - differenced analogs of the euler equations .
there are many approaches to finite - differencing the euler equations , but most modern formulations are at least second order in time and space , and have methods of realistically modeling shocks .
the hydrodynamics method employed by the oohara / nakamura / shibata calculations utilizes leblanc s method for transport , making use of a tensor artificial viscosity .
a brief description is given in an appendix to oohara and nakamura ( 1989 ) .
the earlier calculations were carried out in the laboratory ( fixed ) frame , while later models utilized a rotating frame . in all of the calculations
the gravitational potential was found by a direct solution of the poisson equation .
however , none of the papers discuss the boundary conditions for this equation .
finally , the papers have considered both spherical stars and equilibrium configurations for initial data although no comparisons of the two types of initial data were offered .
ruffert et al .
( 1996,1997,1997 ) performed pn simulations of nsms with the prometheus code implementing the piecewise parabolic method ( ppm ) of colella and woodward ( 1984 ) .
the ppm is an extension of godunov s method that solves the riemann problem locally for the flow between zone interfaces , and , accordingly , it is well suited to addressing shocks .
the calculation of the gravitational potential was accomplished by means of the direct solution of poisson s equation using zero - padding boundary conditions ( hockney 1988 ) .
the initial conditions of the simulations were spherical neutron stars in both tidally locked and rotating configurations .
the stars were embedded in an atmosphere of @xmath4g/@xmath5 that covered the entire grid , and an artificial smoothing was performed on the surfaces of the stars to soften the edges .
the earlier models considered configurations with the realistic equation of state of lattimer and swesty ( 1991 ) while later studies considered models with a much simpler polytropic eos ( ruffert , rampp , & janka 1997 ) .
the most recent eulerian simulations of binary neutron star systems were performed by new and tohline ( 1997 ) .
their work focused on the evolution of equilibrium sequences of co - rotating , equal mass pairs of polytropes .
the equilibrium sequences were constructed with hachisu s self - consistent field ( scf ) technique ( hachisu 1986a , b ) .
the dynamical stability of these equilibrium sequences was tested by evolving them with a 2nd order accurate finite - differenced newtonian hydrodynamics code .
the gravitational potential was obtained by a direct solution of poisson s equation accomplished by means of the alternating direction implicit method .
no description was given of the boundary conditions that were applied to the poisson equation .
the calculation was carried out in a rotating frame that was initially co - rotating with the binary system in order to avoid problems with the advection of the stars across the grid . in a stability test ,
a comparison of two white dwarf binary system simulations starting from the same initial conditions , one carried out in the inertial reference frame and the other carried out in the initially co - rotating frame , revealed dramatic differences in the dynamics of the binary system .
this difference illustrates the need for very careful studies of purely numerical effects on these types of simulations . in the stability tests ,
new and tohline found no points of instability for polytropic models with fairly soft equations of state ( @xmath6 ) .
they did find , however , an instability for the stiffer @xmath7 case indicating that systems with stiffer equations of state are susceptible to tidal instabilities .
it is worthwhile to note that the authors state that they may have misidentified some stable systems as unstable had they performed their simulations in the inertial reference frame , and they therefore stress the importance of very careful studies of numerical effects and careful comparison of different numerical methods .
the earliest sph simulations of binary neutron star systems began with the work of rasio and shapiro ( 1992,1994,1995 ) .
in sph , a distribution of mass in a particular region of space is represented by discreet particles such that the mass density of the particles is proportional to the specified density of the fluid .
local calculations of fluid quantities are accomplished by smoothing over the local distribution of particles .
the method is inherently lagrangean .
this numerical work accompanied semi - analytical work with lai , with the goal of predicting the onset of instabilities in binary systems ( lai , rasio , & shapiro 1993a,1993b,1993c,1994a,1994b , 1994c ) .
the calculation of the gravitational potential was accomplished by mapping the particle distribution onto a density distribution on a mesh where the poisson equation was solved by means of fast fourier transform methods . in these calculations , the authors employed several different polytropic equations of state .
the equilibrium initial data for the models were obtained by allowing the spherical stars to relax to a steady state solution in a rotating frame of reference .
the work investigated sequences of binary star systems with a range of initial separations , and their construction of equilibrium initial configurations for evolutions was critical in determining if and when a dynamical instability forced the merger .
the group reported the presence of a dynamical instability at separations that increase with an increase of the polytropic exponent @xmath8 .
the drexel group performed newtonian sph simulations of nonaxisymmetric collisions of equal mass neutron stars ( centrella & mcmillan 1993 ) .
this effort was followed by calculations of newtonian nsms by zhuge et al .
( 1994,1996 ) they employed the treesph implementation of sph developed by hernquist and katz ( 1989 ) in which the gravitational forces are calculated via the hierarchical tree method of barnes and hut ( 1986 ) .
the calculations employed a polytropic eos with @xmath9 and @xmath10 .
these calculations were carried out in the laboratory frame and the initial data consisted of spherical stars in the non - rotating case or rotating stars that were produced using a self - consistent field method .
the work was aimed at studying the gravitational radiation emission from nsms and addressed effects due to the equation of state , spins , and mass ratio of the stars on the gravitational wave energy spectrum .
the coalescence in these calculations was driven by a frictional force term added to the hydrodynamic equations that models the effects of gravitational radiation loss .
davies , benz , piran , and thielemann ( 1994 ) performed sph simulations of nsms with a focus on the nuclear astrophysical and thermodynamic effects of coalescence .
the sph code used for these calculations was described in earlier work on stellar collisions ( benz & hills 1987 ) , and it makes use of a tree algorithm for calculating gravitational forces .
the calculations were carried out in the inertial frame .
the initial data for the neutron stars was modeled as equal mass @xmath11 polytropes , but a more realistic eos was employed for the dynamical calculation .
the driving force behind the coalescence was a frictional force model of gravitational radiation loss similar to that of the drexel group .
the rates of energy and angular momentum loss were determined by applying the quadrupole approximation to the equivalent point mass system , and the resulting acceleration of each sph particle was determined by expressions derived from these rates .
subsequent calculations by rosswog et al .
( 1998 ) have focused on _
r_-process nucleosynthesis and mass ejection .
more recently members of this group have developed a pn extension of the sph algorithm ( ayal et al .
1999 ) , which they have applied to nsms to study the dynamics and gravitational wave emission from the merger . in the remainder of this paper we will focus on comparisons of several eulerian numerical methods for modeling newtonian binary neutron star systems as well as comparisons of the effects of spherical versus equilibrium initial data . additionally , we will examine the stability of both the spherical and equilibrium initial data .
the subject of gravitational wave signals from nsms will not be considered here , but instead will be the subject of a subsequent paper . in [ sec : hydro ] we explain our numerical methods for evolving the equations of hydrodynamics and solving the poisson equation . in [ sec : eqd ] we delineate our method for obtaining equilibrium data with self - consistent boundary conditions . in [ sec : self ] we compare calculations carried out in both rotating and inertial frames with several different schemes for coupling gravity to matter via the gas momentum equations . in [ sec : stab ] we compare models with both equilibrium and non - equilibrium initial data . in [ sec : conc ] we offer conclusions about this work especially regarding its meaning for post - newtonian and fully general relativistic models of neutron star mergers .
as we have mentioned in the previous section , the accuracy of the numerical algorithms is of paramount importance if one is to obtain accurate hydrodynamic models for mergers .
previous work on such models have utilized a variety of hydrodynamic schemes in either a fixed ( inertial ) or rotating frame of reference .
the latter has been claimed to be more accurate by virtue of is obviation of the difficulties of advection , but no systematic comparison of the two has yet been published . in this section
we describe two numerical hydrodynamics schemes that we have employed to produce such a study .
we have not attempted an exhaustive study in which we compare the qualitative results of each hydrodynamic scheme that has been employed to date .
such studies have been conducted for most of these schemes on a number of problems involving shocks . however , the performance of the hydrodynamic algorithm on shocks is not the only metric by which one needs to measure the quality of the hydrodynamic algorithm .
for example , for simulations in which the linear momentum equation has been solved one should examine how well angular momentum is conserved . in the long timescale evolutions needed for multiple orbit simulations of orbiting stars , the addition or loss of angular momentum into the calculation could artificially enhance or delay inspiral during the mergers .
similar issues apply for linear momentum in simulations where the gas angular momentum equations are solved .
in the same vein , if the gas momentum equation is solved then one should monitor how well the total energy is conserved . or if the total energy equations is solved one should monitor how well the gas energy equation is solved .
the 3-d numerical hydrodynamics scheme we describe in the section is similar to the zeus scheme of stone & norman ( 1982 ) .
however , we have made some fundamental changes to the order of operations in order to improve the numerical accuracy of the scheme on self gravitating problems .
the flow of matter in the neutron stars can be taken to be inviscid . under these circumstances the newtonian description of the matter evolution
is described by the continuity equation together with the euler equations ( mihalas & mihalas 1984 , bowers & wilson 1991 ) of compressible inviscid hydrodynamics . in an inertial frame of reference
the equations are @xmath12 @xmath13 @xmath14 where the dependent variables are the mass density @xmath15 , the internal energy density @xmath16 , the fluid velocity @xmath17 , the fluid pressure @xmath18 , and the newtonian gravitational potential @xmath19 .
the gravitational potential is described by the the poisson equation @xmath20 in conjunction with boundary conditions that must be specified . in a frame rotating with angular frequency @xmath21 about the center of the fixed ( inertial ) grid the gas momentum equation is modified by the addition of the coriolis and centrifugal forces ( chandrasekhar 1969 ) @xmath22 where @xmath23 is the velocity of the rotating frame relative to the lab frame .
the notation @xmath24 indicates the @xmath25 component of the vector @xmath26 . in the limit of @xmath27
we recover the inertial frame momentum equation . the continuity , gas energy , and poisson equations are unchanged from the inertial frame case . for the remainder of this paper
we will consider the angular frequency vector @xmath28 to be co - aligned with the z axis so that @xmath29 we assume that the z axis passes through the center of the grid at coordinates @xmath30 . under this condition
the coriolis and centrifugal force terms in cartesian coordinates become : @xmath31 where @xmath32 and @xmath33 are the coordinates of the z - axis .
the set of hydrodynamic equations must be closed by specifying an equation of state expressing pressure as function of local thermodynamic quantities .
a standard choice for building the initial neutron star models is the polytropic equation of state , which has the form @xmath34 where @xmath8 is the polytropic exponent , which is related to the polytropic index , @xmath35 , by the relationship @xmath36 this particular type of eos is advantageous in that the gas energy equation becomes linear in @xmath16 rendering the solution trivial . in isentropic situations
this eos allows the pressure to be written purely as a function of density in the form @xmath37 where @xmath18 is the pressure , and @xmath38 is the polytropic constant .
this form of the eos is used for the construction of the initial models .
much of the numerical scheme we employ for the solution of the euler equations is derived from the zeus-2d hydrodynamics scheme invented by stone & norman ( 1992 ) ( hereafter sn ) .
in particular the finite - differencing stencils are identical to those of sn , with the exception of the coriolis and centrifugal forces , which are not included in the zeus-2d scheme .
however , the method we employ differs in one significant way : the order of solution of the various terms in the euler equations differs from that of the zeus-2d algorithm . as we will show in a subsequent section of this paper ,
the order of the solution of these equations is of fundamental importance to the accuracy of the algorithm in the case of self - gravitating hydrodynamics .
a final simplification for our algorithm , which we shall henceforth refer to as the v3d algorithm , employs only cartesian coordinates .
this simplification significantly increases the computational speed of the v3d code .
the finite - differencing algorithm we employ relies on a staggered grid in which the intensive variables @xmath16 , and @xmath15 are defined at cell centers , while the vector variables such as the velocity components @xmath17 are considered to be defined at their respective cell edges .
the gravitational potential @xmath19 is also defined at the cell center .
the centering of the variables on the grid is depicted in a 2-d plane of the 3-d grid in figure [ fig : grid ] . in our finite - difference notation
we employ superscripts to denote the time at which the variables are defined .
the timestep is taken to be @xmath39 and the numerical algorithm advances the solution of the pdes from @xmath35th time , @xmath40 , to the new @xmath41th time , @xmath42 .
the explicit finite - difference algorithm employed by sn for the solution of the euler equations decomposes the time integration into two steps by employing operator splitting among the various terms of the equations . in one step
the density , internal energy density , and velocities are updated by integrating the advective terms . in the nomenclature of sn
we refer to this as the _ transport _ step .
the remaining terms , i.e. the terms on the right hand sides of equations ( [ eq : cont ] ) , ( [ eq : gase ] ) , and ( [ eq : rgasm ] ) , are integrated forward in time .
following sn we refer to this step as the _ source _ step .
an additional consideration involves the solution of the poisson equation and how it relates to the solution of the euler equations . in the zeus-2d algorithm
the order of solution of the euler equations is described in the flow chart of figure [ fig : flow]a .
in contrast our algorithm , v3d , is described in figure [ fig : flow]b . in the transport
step the following equations are solved : @xmath43 @xmath44 @xmath45 the advection integration is carried out using norman s consistent advection scheme ( norman 1980 ) , which ties the advected internal energy density and advected velocity to the mass flux as described in sn .
this concept of tying the energy and momentum fluxes to the mass flux has been shown to posses superior angular momentum conservation properties ( norman 1980 ) .
the actual flux limiter employed is the van leer monotonic flux limiter ( van leer 1977 ) that is spatially second order . in the source
step the following equations are updated : @xmath46 @xmath47 the scalar viscous stress @xmath48 is added to the equations in order to allow for viscous dissipation by shocks in the fluid .
we employ the standard von neumann - richtmyer prescription for the viscous stress , as described in sn , with a length parameter of @xmath49 .
we monitor the viscous dissipation arising from this stress and have found that the total viscous energy generation is negligible for two merging polytropes .
the coupling to gravity enters through the gradient of the newtonian gravitational potential in equation ( [ eq : rgasm_s ] ) .
we wish to note that the continuity equation is not updated during the source step as it possesses no source term .
the lack of a source term for the continuity equation means that the density at a new time @xmath50 is known after the transport step is complete . as we will discuss in a subsequent section of this paper ,
this point is crucial to our preferred method of solution for these equations .
the explicit finite - differencing of the euler equations is briefly discussed in appendix a. for the remainder of this section we concentrate on the order of solution of the transport and source steps . in the method of sn the source terms ( equations [ eq : gase_s ] and [ eq : rgasm_s ] )
are integrated forward in time to arrive an intermediate solution for the new internal energy density @xmath16 and the velocity @xmath51 .
the intermediate energy density and velocity are used as initial values for the transport equations ( [ eq : cont_t])-([eq : rgasm_t ] ) , which are then integrated forward in time to find the values of the density , energy density , and velocity at @xmath50 .
sn have shown by means of convergence testing that this algorithm is spatially second order accurate .
there is no compelling reason to suggest that the order of source and transport updates as presented by sn is preferred .
one can easily reverse the order of updates so that the results of the transport step are utilized in the source update .
we henceforth will refer to this order of updates as the v3d algorithm while the opposite order will be referred to as the zeus algorithm . by comparing the two algorithms on a number of standard hydrodynamic test problems we have numerically verified that the reversal of these operations has no significant effect on the overall quality of solutions when self - gravity is not present .
however , as we will show in a later section the v3d algorithm offers significant advantages when modeling orbiting binary stars .
an example of the comparable performance of the zeus and v3d algorithms on a standard test problem is shown in figure [ fig : sod ] where the performance of the algorithm on a sod - like ( sod 1978 ) shock tube is shown .
the shock tube problem pictured employs a @xmath52 polytropic equation of state .
the grid is set up with 100 spatial zones over the range of @xmath53 cm with the initial contact interface at @xmath54 located at @xmath55 .
this initial configuration is that of a riemann problem , which results in a shock and a contact discontinuity propagating to the right and a rarefaction propagating to the left . because the exact solution is known ( chorin & marsden 1993 ) we can easily evaluate the numerical results from both algorithms . overall the character of the numerical solution is comparable between the two cases .
the values of the variables in both the contact discontinuity and the shock are slightly different as would be expected from different algorithms , but both methods resolve the shock and the contact discontinuity with the same number of zones .
the rarefaction is represented nearly identically by both methods .
the figure compares the two orders of update .
one can visually see that little difference exists between the two solutions .
we have verified this on a number of other non - self - gravitating test problems .
in contrast , for the case of self - gravitating hydrodynamics , we do find that the v3d algorithm is preferred as we will discuss in section [ sec : self ] . in order to describe self - gravitating phenomena , the gravitational field of the matter distribution
must be found by solving the poisson problem described by equation ( [ eq : poisson ] ) .
the poisson equation can be readily solved by a variety of techniques well suited to elliptic equations . in the simulations described in this paper
we have employed both w - cycle multigrid and fast - fourier - transform ( fft ) methods ( press et al .
both methods have been extensively tested on matter configurations where the solution is known .
because the poisson equation is linear one can easily generate test problems with known answers , but which also posses complex field geometries .
for example , by placing @xmath56 polytropes at random points within the computational domain we can create a complex gravitational field configuration .
since the gravitational potential for a @xmath56 polytrope is analytically known , the potential for the entire configuration at any point in the computational domain is readily found as a superposition of individual polytropic solutions . using this method
we have found that both methods give the correct answers to approximately @xmath57 for the grid resolutions employed in this work .
the numerical solution of equation ( [ eq : poisson ] ) requires the specification of boundary values along the edge of the computational domain . for the problem of merging neutron stars
these are _ a priori _ unknown .
the problem of determining the appropriate boundary conditions has been approached differently by a number of different groups .
ruffert et al .
have employed zero - padding boundary conditions in conjunction with their fft solution method .
the zero padding method has been shown by james ( 1977 ) to be algebraically equivalent to a direct summation by convolution of image charges ( defined on the edge of the grid ) over the green function for the poisson equation .
oohara and nakamura ( 1990 ) , rasio and shapiro ( 1992,1994,1995 ) , and new and tohline ( 1997 ) have not specified how boundary conditions on the potential were obtained for their hydrodynamic solutions . a number of groups ( davies et al .
1994 ; zhuge et al .
1994 ; zhuge et al .
1996 ) have carried out smoothed particle hydrodynamics ( sph ) simulations in which the field was computed by a tree - code summation obviating the need for the specification of boundary values .
the accurate specification of self - consistent boundary conditions for the poisson equation is challenging .
the expansion of the potential in terms of multi - poles may require the expansion to be carried out to very high order if one is to obtain accurate values for the potential .
this is especially true given that the initial configuration for a neutron star merger simulation consists of two widely separated fluid bodies , which has a very large quadrupole moment .
the zero - padding method necessitates a fairly large memory cost .
this memory cost can largely be eliminated by use of the james algorithm for the solution of the poisson problem for isolated systems ( james 1977 ) .
in fact we employ the james algorithm to obtain equilibrium initial data described in the next section .
however , we have found that the james algorithm does not scale well to large numbers of processors on shared - memory parallel computers . in order to obtain an accurate algorithm for the boundary conditions on the potential we have turned to direct integration over the green function for poisson s equation , i.e. @xmath58 because of the large number of grid zones present in the problem , it is computationally intractable to compute this sum directly .
if our computational domain is discretized into @xmath59 points in each of the three spatial dimensions then the summation is over @xmath60 zones in order to evaluate the potential at each of the @xmath61 points on the edge of the domain .
this implies the total algorithm is order @xmath62 .
since we typically employ @xmath63 for our simulations , this renders the direct summation over the entire grid computationally intractable for use in a hydrodynamic simulation .
this dilemma is further exacerbated by the need to calculate an inverse square root in order to evaluate the distance @xmath64 from the boundary point to each zone .
however , for the problems we are considering the mass is concentrated in only a relatively small region of the computational domain .
if we restrict the summation to only those zones in which a significant amount of mass is present , the summation becomes more tractable .
accordingly , we have adopted the following algorithm for obtaining the boundary values .
we evaluate the amount of mass in eight - zone cubic ( @xmath65 ) `` blocks '' of the grid .
if the mass of the zone is greater than a threshold value @xmath66 the total mass of the block and the blocks center of mass coordinates are stored in a list .
this operation is order @xmath60 but it is only required once per timestep . once the entire mesh has been scanned we obtain a complete list of all the block with a significant mass .
the size of the list is dependent on the mass threshold employed .
if the mass threshold is chosen too low , the list will become very large and the summation computationally intractable .
if the cutoff is chosen too large , the list will not include most of the mass in the domain .
we have experimentally found that a value of @xmath67 produces a list which fully represents the mass in the domain .
once the list of significant mass blocks has been produced the boundary values of the potential can be calculated by direct summation @xmath68 where @xmath69 is the mass in the @xmath70th block in the list , @xmath71 is the distance from the @xmath72th point on the edge of the grid to the center of mass coordinates of the @xmath70th block , and @xmath73 is the length of the list .
this operation is order @xmath74 .
however , @xmath75 , which renders the summation tractable . furthermore , this algorithm is readily parallelizable on a shared memory parallel computer thus allowing for a rapid solution .
we typically find that calculation of the boundary conditions never exceeds 20% of the overall computational effort .
finally , we wish to emphasize that this algorithm will not work efficiently in cases where the mass is more evenly distributed over the entire mesh
. the accuracy of this algorithm has been tested by two methods .
first , the algorithm was applied to the test problems we mentioned earlier in this section where the analytic answer was known .
secondly , the summation algorithm was also tested in more general situations by comparing the boundary values obtained by this method to those obtained by brute force direct summation . in all cases the boundary values agreed to better than @xmath76
; in most cases the agreement was better than @xmath77 .
additionally , we track the total mass in the list so that it can be compared to the total mass on the mesh . any significant difference between the two masses
will indicate a problem with the summation .
in order to accurately model two neutron stars in close orbits with one another , it is important to employ initial conditions that precisely reflect the true configurations of the two fluid bodies . in general , for close binary systems , these equilibrium configurations will not consist of two spherical stars .
instead the configuration will contain tidally distorted fluid bodies that are only approximately spherical . in numerical simulations of binary star systems
the stability of the orbits can be quite sensitive to the details of the initial configuration .
in a subsequent section we compare dynamical models which have employed equilibrium initial conditions with models that have utilized spherical stars .
the construction of initial data for these systems is a non - trivial task . in practice
each neutron star in a binary system will be non - synchronously rotating around it s own axis .
several calculations ( bildsten & cutler 1992 ; kochanek 1992 ) have shown that viscous dissipation at the causal limit is insufficient to tidally lock binary neutron star systems during their lifetime . accordingly ,
the most realistic configurations that one could model would be non - tidally locked . however , there is tremendous difficulty of obtaining equilibrium initial data for such cases .
finding initial conditions that correspond to the non - synchronous case would require the solution of the compressible darwin riemann problem that is well outside the scope of this paper . since the target of this paper is a study of the numerical methods and initial conditions needed for precise simulations of binary neutron star mergers , we restrict ourselves to the tidally locked case .
realistic binary systems will also contain unequal mass components . however , in this paper we consider only equal mass systems . our numerical algorithm for obtaining equilibrium initial data
is easily extended to the non - equal mass case which will be considered in a future paper .
a number of other research groups ( oohara & nakamura 1990 ; new & tohline 1997 ) have developed methods to obtain equilibrium data for the case of synchronous binary neutron star systems .
in both cases the equilibrium models must simultaneously satisfy both the bernoulli and poisson equations . in the case of oohara and nakamura
they have employed a method that in similar to ours in that it iteratively solves the bernoulli and poisson equations on a cartesian grid .
however , we have found some problems with this method for obtaining the initial conditions that we seek .
new & tohline have found initial conditions using the self - consistent field technique of hachisu ( 1986a,1986b ) , which iteratively solves the bernoulli equation and the integral form of the poisson equation on a spherical polar grid .
while this latter method avoids the problems we have found with the oohara method , it s use for our case would involve remapping of the data from the polar grid to the cartesian grid we employ for our dynamical simulations . this remapping would introduce small errors that would render the initial conditions on the cartesian grid slightly out of equilibrium . in turn , the deviation from equilibrium can cause spurious hydrodynamic motions away from the initial data once the evolution begins . in order to avoid this
we have combined techniques from both the oohara et al .
and new & tohline methods .
our objective is to develop equilibrium data on the same grid that the hydrodynamic simulation will employ .
we first need to identify the equations that describe the equilibrium configuration .
these equations result from taking the hydrostatic limit of the euler equations of hydrodynamics together with the poisson equation . in the hydrostatic limit the gas momentum equation ( [ eq : rgasm ] ) collapses to @xmath78 if we make the assumption that the equation of state is of the isentropic form given by equation ( [ eq : polyt ] ) , then equation ( [ eq : rhyd1 ] ) can be integrated by parts to find the bernoulli equation @xmath79 = c \label{eq : bern1}\ ] ] where @xmath80 is a constant and where we have assumed that the rotation is about the z - axis
. this equation must be satisfied in the interior of the fluid bodies . simultaneously
the equilibrium data must also satisfy the poisson equation @xmath81 however , two fundamental difficulties occur in the solution of these two equations on cartesian grids containing self - gravitating fluid bodies .
first , we do not _ a priori _ know where the boundaries of the fluid bodies lie , and hence we do not know where the bernoulli condition should apply .
second , we do not _ a priori _ know the boundary conditions on @xmath19 that must apply to the poisson equation .
the boundary conditions on @xmath19 must be determined in a self consistent fashion using the green function corresponding to the poisson equation .
this latter problem is the more difficult of the two problems to solve . while oohara et al .
have employed a direct solution of the poisson equation for equilibrium data , they have made no mention of what boundary conditions they have employed on equation ( [ eq : poisson2 ] ) .
we have found that the configuration resulting from the iterative solution of equations ( [ eq : bern1 ] ) and ( [ eq : poisson2 ] ) is quite sensitive to the use of non - self - consistent boundary conditions and we strongly recommend against employing such boundary conditions . in order to minimize the problems with deciding where the boundaries of the fluid bodies are , we have adopted a technique from the scf technique of hachisu et al .
we consider equilibrium binary systems in two different topological configurations as depicted by figure [ fig : binary ] . in the first case we consider non - contact binary systems . in the second case we consider contact binary systems . in the first case during our iterative solution of the combined bernoulli and poisson equations
we specify the extremal inner and outer points of the star as depicted in figure [ fig : binary]a .
we define the orientation of our grid so that the x - axis passes through the centers of mass of the two stars .
the z - axis passes through the barycenter of the system thus defining the origin of the grid . by specifying the inner and outer points we seek equilibrium solutions with a certain aspect ratio . by adjusting the locations of the extremal points we can find equilibrium configurations with varying separations between the components and or their centers of mass . in the case of contact binaries we specify the extremal outer point of the contact system and the extremal outer point of the neck connecting the high density portions of the two fluid bodies . as with the detached case , by varying these two points we can find a sequence of equilibrium configurations with varying separations between their centers of mass . in the contact binary case
we define the center of mass of each star by considering only the mass contained within each half of the computational domain as defined by a plane perpendicular to the line connecting the two highest density zones of the grid . once we have identified the configuration of the system we can then determine where the bernoulli equation can be applied during each iterative step .
assuming , for the moment , that we know @xmath82 , @xmath38 , and @xmath80 , and that we posses some iterative estimate of @xmath19 , when can then solve the bernoulli equation for a new estimate of the density @xmath15 : @xmath83\right ) \right]^{1/(\gamma-1)}. \label{eq : bern2}\ ] ] obviously , this equation only makes sense where the factor contained within the square brackets of equation ( [ eq : bern2 ] ) is positive .
we use this as a criterion to decide where to apply the bernoulli equation .
if the factor is positive the density is updated to the density as determined by equation ( [ eq : bern2 ] ) , otherwise the grid zone is considered to be vacuum and the density is set equal to zero .
the determination of self - consistent boundary conditions for the poisson equation is of paramount importance . in order to clarify what we mean by the use of
the term `` self - consistent '' we first clarify the problem .
we assume that the computational domain , @xmath84 , contains the self - gravitating fluid bodies which have compact support within the interior of the domain , i.e. the fluid density vanishes on the boundary of the domain @xmath85 .
more simply stated , we assume that the fluid bodies are contained inside the domain @xmath84 . this type of self - gravitating system has been termed an `` isolated '' system ( james 1977 ) . under these circumstances
we require that the boundary conditions on @xmath19 satisfy equation ( [ eq : poisson2 ] ) .
since the potential on the boundaries depends on the density distribution @xmath86 we can not _ a priori _ self - consistently know the boundary values prior to solving the problem .
this problem is not only relevant for initial data but is also relevant for the solution of the poisson problem during the course of a self - gravitating hydrodynamic simulation .
various groups modeling equilibrium binary configurations have attempted to avoid this problem .
the equilibrium sequence work of nt has utilized the scf method of hachisu , which avoids this problem by employing a multi - pole expansion of the potential in order to estimate the boundary conditions . in the case of isolated systems , james ( 1977 )
has shown that the boundary conditions can be obtained exactly by use of fft techniques .
we accordingly employ this method for use in our initial data algorithm .
furthermore , james has shown that this method is algebraicly equivalent to the `` zero - padding '' technique employed by ruffert et al .
the advantage of the james algorithm over the zero - padding technique is that it requires substantially less memory overhead which is a significant advantage in a 3-d simulation .
we could also employ this algorithm to compute the self - consistent potential during the course of our hydrodynamic simulations .
however , in the hydrodynamic simulations we have found it advantageous to employ equation ( [ eq : green1 ] ) directly to get the boundary conditions for @xmath19 followed by a straightforward poisson solve using either multigrid or fft techniques ( press et al .
we have found that this method is more amenable to implementation on the shared memory parallel computing architectures that we employ for our simulations .
the complete algorithm for finding the initial data is as follows : 1 .
fix inner and outer points of stars ( in the detached binary ) case and outer point and neck width ( in contact binary case ) .
denote the distance from the z - axis to these points as @xmath87 and @xmath88 .
2 . make initial guess at the density distribution throughout the computational domain .
also guess an initial value of @xmath38 .
3 . using density distribution solve for potential using the james algorithm to solve the poisson equation with self - consistent boundary conditions .
4 . using the bernoulli equation
evaluate @xmath82 by @xmath89 5 .
evaluate the bernoulli constant , @xmath80 , at @xmath87 using equation ( [ eq : bern1 ] ) 6 .
update @xmath38 by evaluating the bernoulli equation at some point @xmath90 which lies on the line through the centers of the stars @xmath91 7 .
calculate new value of density for every zone on the grid using the following algorithm : @xmath92 where @xmath93\right).\ ] ] 8 .
if the maximum relative density change in any zone of the grid is less than @xmath77 then consider the solution converged and stop
. otherwise go to step 3 .
one major difference between this algorithm and those utilized by others is step 6 , the update of @xmath38 . for a particular equation of state , e.g. @xmath56 , there may not be a solution for an equilibrium configuration with a given inner and outer point .
one can easily see this for the case of an isolated @xmath56 polytrope where the radius is determined by ( shapiro & teukolsky 1983 ) @xmath94^{1/2}.\ ] ] in this case only a specific value of @xmath38 will allow the star to `` fit '' into the specified number of grid points between @xmath87 and @xmath88 . if the value of @xmath38 is not allowed to change the iterative procedure described does not converge . in practice
the change in @xmath38 is small . when constructing an equilibrium sequence with fixed values of @xmath38 , we adjust the grid size slightly to get the desired value of @xmath38 .
we have found that this usually only requires changes of a few percent in the grid spacing @xmath95 in order to find an equilibrium solution for a specified value of @xmath38 .
the total mass of the converged equilibrium system is determined by the initial guess of the density distribution in step 2 of the algorithm . by multiplying the initial guess of the density distribution by some factor
we can converge to equilibrium systems of more or less mass .
for the case of @xmath56 both the bernoulli and poisson equations are linear in the variables @xmath19 and @xmath15 . in this case the total mass of the converged solution is affected only by the initial guess at the distribution while the value of @xmath38 is determined only by the grid spacing .
this renders the procedure of producing a sequence of equilibrium solutions for a given polytropic constant , @xmath38 , and total mass , @xmath96 , relatively easy . in the case
where @xmath97 the bernoulli equation becomes non - linear and changes in @xmath95 or the initial density guess affect both the resulting value of @xmath38 and @xmath96 . in this case
building an equilibrium sequence becomes much more difficult and time consuming .
for this reason we have constructed a @xmath98 equilibrium sequence only at @xmath99 resolution .
we have constructed a few specific @xmath98 equilibrium configurations at @xmath100 resolution for use in hydrodynamic studies of stability . using the method that we have described above we have constructed equilibrium sequences of data for @xmath56 ( @xmath101 ) and @xmath98 ( @xmath102 ) polytropes . for the @xmath56 case ,
the sequences were constructed with a total mass of @xmath103 and a value of @xmath104 erg @xmath5 g@xmath105 .
an isolated spherical @xmath56 polytrope with these parameters would have a radius of approximately @xmath106 km and a central density that is roughly 10 times the nuclear saturation density of @xmath107 g/@xmath5 .
such a configuration resembles a realistic neutron star .
the @xmath56 sequence is shown in figure [ fig : eq_seq_2 ] while the @xmath98 case is shown in figure [ fig : eq_seq_3 ] .
all separations shown are the center of mass separation @xmath108 which has been normalized to the spherical radius of a single undisturbed polytrope .
both the total energy , @xmath109 , and the angular momentum , @xmath110 , are plotted for each configuration . in the @xmath56 sequence
the models with a separation of less than about @xmath111 are contact binaries where the two stars are joined by a `` neck '' of matter passing through the barycenter of the system .
those systems with separations greater than @xmath112 are detached . for the @xmath98 case
the bifurcation point is at a separation of approximately @xmath113 . in order to obtain this number
more precisely we would have to construct models at substantially higher resolution .
because of the difficulty of constructing a large number of configurations with a specified value of @xmath38 and @xmath96 for the non - linear @xmath98 case , we have chosen not to do so .
our purpose was to construct initial data for hydrodynamic simulations using the same grid that we would employ for the simulation . for the models shown in figures [ fig : eq_seq_2 ] and [ fig : eq_seq_3 ] the grid resolution was approximately @xmath114 km for the @xmath115 models and @xmath116 km for the @xmath117 models . from figure
[ fig : eq_seq_2 ] it is easily observed , by comparing the @xmath115 and @xmath117 models , that the @xmath115 models do not have adequate spatial resolution at the wider separations . nevertheless both the @xmath115 and @xmath117 models show minima in both the total energy , @xmath109 , and angular momenta , @xmath110 , at approximately @xmath118 . the slight variation in the data near the bifurcation point between detached and contact binaries is due to the finite resolution of the grid .
the contact binaries in this case may have a neck consisting of only one or two zones , a situation which is likely to cause some fluctuation in both the energy and the angular momentum due to the discrete nature of the neck .
it is interesting to compare these results with the semi - analytic work of lai et al .
( 1993c ) ( hereafter lrs ) .
lrs constructed an equilibrium sequence of binary , compressible , darwin ellipsoids as an approximation to the equilibrium configurations of two synchronously orbiting polytropes . by identifying a turning point in the energy versus separation curves
lrs found a secular instability for @xmath56 polytropes at a separation of @xmath119 .
furthermore , lrs also found that these turning - points occured simultaneously in both the total energy , @xmath16 , and the angular momentum , @xmath110 .
there have also been a number of efforts to construct such equilibrium sequences numerically .
in addition to their semi - analytic work lrs also found equilibrium sequences obtained using a relaxation scheme which employed smooth particle hydrodynamics methods yielded a turning point in the energy and angular momenta for an equilibrium sequence at a separation of @xmath120 .
in contrast new and tohline ( 1997 ) , using the scf technique , found a turning point at @xmath121 .
our result , which can be readily seen from figure [ fig : eq_seq_2 ] , yields an approximate turning point of @xmath122 .
this result is somewhat closer to the compressible darwin ellipsoid value and much further from the recently obtained value of new & tohline . in agreement with both lrs and nt
, we find the turning point at a point where the equilibrium systems are still detached .
nevertheless , the turning point is quite close to the point at which attached systems would form .
the occurrence of a turning point on the detached binary branch of the curve seems to indicate that a binary system slowly spiraling inward by some energy and angular momentum loss mechanism will encounter an instability without ever becoming a contact binary .
the nature of this instability and its implication for the dynamical evolution of binary systems will be discussed in a subsequent section . in the @xmath98 case
our results indicate a turning point in the equilibrium sequence . in this case
the value of the polytropic constant was @xmath123 erg @xmath124 g@xmath125 and the total mass was @xmath103 .
note that in this case the turning point we seem to find is at approximately @xmath126 in comparison with the lrs semi - analytic value of @xmath127 .
in contrast nt find a value of @xmath128 .
unfortunately , hachisu ( 1986b ) does not present numerical values for the separation at this turning point so we are unable to compare to this work .
in contrast , although rasio & shapiro ( 1994 ) find a find an instability for the @xmath98 case of @xmath129 based on hydrodynamic simulations , the equilibrium sequence they obtain on the basis of relaxation methods using their sph code yields a turning point at @xmath130 .
this result can be contrasted with the results of nt and our own results which show a turning point in both energy and angular momentum at substantially larger separations . however , the semi - analytic results presented in both lrs and in lai et al .
( 1994b ) show a turning point at @xmath127 .
we will discuss the implications of this turning point for hydrodynamic evolution of a binary system in a subsequent section of this paper .
our numerical hydrodynamics method requires the density to be non - zero everywhere on the computational grid .
therefore , we include a low density ( @xmath131 1 g/@xmath5 ) `` atmosphere '' as a background in regions where stellar matter is not present .
we have varied the density between ( 1 - 10@xmath132g/@xmath5 ) in our hydrodynamic simulations and have found that this has no discernible effect on the dynamics of the simulations .
other models ( ruffert et al . 1995 , 1996 ) of neutron star mergers that have been carried out with eulerian codes have had to employ much higher densities ( @xmath4 g/@xmath5 ) for the surrounding material . however , adding matter in regions outside the stars presents two difficulties for hydrodynamic simulations .
first , such matter will not in general be in hydrodynamic equilibrium if it has the same entropy as the matter in the stars .
thus at the beginning of the simulation it will immediately infall towards the stars and form an accretion shock at the surface of the stars .
this accretion shock , while not physically troublesome because of the low density of the material in the atmosphere , will have the undesirable numerical effect of driving the timestep determined by the courant stability condition to a very small timestep because of the high infall velocities . in order to counter this effect
we make the atmosphere hot , i.e. we set the energy per baryon in the atmospheric material to approximately @xmath133 mev .
this has the effect of preventing the atmosphere from falling down onto the neutron star surfaces .
furthermore we decrease this energy slightly with distance from the stars so as to achieve a configuration that is slightly more hydrostatically stable .
a second problem originates if the atmosphere is put in place with a non - zero velocity with respect to the stars .
if the matter is placed on the grid with zero velocities in the lab frame , the motion of the stars quickly sweeps up the matter into a bow shock on the front sides of the orbiting stars . in a circumstance similar to the accretion shock mentioned in the previous paragraph
, the bow shock has the numerical effect of driving the courant timestep to zero . in order to avoid this problem ,
calculations that are carried out in the lab frame have an atmosphere with an initial velocity such that the material is rotating about the center of the grid at the same speed with which the stars are revolving . near the edge of the grid the velocity of the atmospheric material
is slowly tapered to zero so as to avoid a shock forming at the edge of the grid and to keep the velocity below the speed of light . in the case of rotating frame simulations the velocity of the atmospheric material
is set equal to zero in the rotating frame .
these two steps obviate the problem of having the matter accrete onto the stars .
we wish to note that this method requires no additional machinations to treat the material outside the stars ; the evolution of the material is described by the numerical solution of the euler equations .
one of the more difficult aspects of self - gravitating hydrodynamics is the need to self - consistently solve both the partial differential equations describing the dynamics of the fluid and the equation(s ) that describe the gravitational field arising from that matter .
one of the origins of this difficulty in the newtonian case is the different mathematical character of the two sets of equations : the euler equations are hyperbolic while the poisson equation is elliptic in nature .
while the euler equations can be numerically solved by explicit techniques , the poisson equation requires an implicit solution .
since the two sets of equations are coupled by the gravitational acceleration term in the gas momentum equation , one must take care that the numerical methods employed for these coupled equations adequately maintain all the desirable properties of the total system such as angular momentum conservation and total energy conservation . in this section
we compare several methods for these calculations that employ various methods for treating the gravitational acceleration term .
we wish to emphasize that no numerical scheme that solves the linear gas momentum equations in three dimensions will guarantee the numerical conservation of angular momentum .
the converse is also true : if one solves the angular momentum equations in three dimensions the solution will not in general numerically satisfy the linear momentum equations .
this discrepancy arises from the fact that the finite - differencing of the underlying partial differential equations reduces them to algebraic equations that must be solved for the new values of the density , internal energy , and velocity .
thus the five euler equations are sufficient to algebraicly determine the five variables .
the finite - differencing of the angular momentum conservation equations will be different from the linear momentum equations and thus give rise to five additional equations that must be algebraicly satisfied by the same five variables .
the problem is algebraicly over - constrained .
despite the fact that the linear gas momentum and linear angular momentum equations can be easily shown to be equivalent , i.e. that conservation of linear momentum guarantees the conservation of angular momentum and vice versa , there is no such equivalence between the finite - difference analogs to these two vector equations .
a similar statement can be made about the gas energy equation and the total energy equation .
the issue of importance to simulations of orbiting neutron stars is how badly does numerical conservation break down over the course of a simulation ?
that is , how badly conserved are the angular momentum and total energy over the course of a simulation ?
we will consider these issues for several different schemes for coupling the poisson and euler equations .
we will also show that a superior choice among these schemes emerges from these comparisons .
this is vital for quantitatively accurate models of binary neutron star mergers where it is necessary to conserve both angular momentum and energy in order to ensure that orbital decay is physical and not the spurious result of numerical non - conservation .
as we have previously mentioned in subsection [ subsec : numerical ] , there are two possible orders of update of the source and transport portions of the euler equations .
these two possibilities are illustrated algorithmicly in figure [ fig : flow ] .
the zeus algorithm of stone & norman employs the order of update shown on the left of figure [ fig : flow ] while our v3d code employs the method shown on the right .
as we have shown in figure [ fig : sod ] there is no substantive difference between these two method in the case where self - gravity is not relevant .
however , in the self - gravitating case these two approaches admit different possibilities for calculating the gravitational acceleration in the gas momentum equation . in the case of the zeus algorithm ,
the solution of the source step first requires the gravitational potential in order to calculate the gravitational acceleration in the gas momentum equation .
hence the need for first solving the poisson problem as described on the left side of figure [ fig : flow ] . since the density at the new time ( @xmath50 ) is not _ a priori _ known at the beginning timestep ( at time @xmath40 ) , the right hand side of the poisson equation
can only be constructed using the density that is known at time @xmath40 .
thus the newtonian gravitational potential is known only at time @xmath40 .
consequently the gravitational acceleration term which is calculated from the newtonian potential is not time - centered between times @xmath40 and @xmath50 .
our code , v3d , performs the advection step before the source step that updates the lagrangian terms ( the terms on the right hand side of the hydrodynamics equations ) .
this ordering , advection before the source update , allows the choice of computing the right - hand side of the poisson equation , @xmath134 , with the density at the old time step ( time lagged ) , the new time step ( time advanced ) , or the average of the two densities ( time centered ) .
the finite - difference expressions for these choices are : @xmath135 this choice can play a significant role in the dynamics of a simulation .
for example , in physical situations where the gravitational acceleration is always increasing in time , the use of the time - lagged centering will always underestimate the gravitational acceleration . over long time
scales this consistent underestimate can lead to significant deviations from the true physical behavior of the system . in the case of orbiting binary stars
this could lead to non - physical evolution of the orbits .
finally , the choice of time - centering for the gravitational acceleration can have a significant impact on the conservation of both angular momentum and total energy . both total energy and angular momentum conservation are vital in achieving good neutron star merger models .
a lack of conservation of either of these two quantities could lead to unphysical inspirals and both qualitatively and quantitatively incorrect outcomes of the simulations .
therefore , it is necessary for us to examine how well both of these quantities are maintained as various choices are made for the gravitational acceleration coupling method . in order to ascertain how the time centering affects the dynamics of binary orbits we have conducted a simple test .
we have placed two spherical @xmath56 , @xmath136 polytropes in a circular orbit with a separation of @xmath137 where @xmath138 km is the radius of the spherical polytrope . at such wide separations
the tidal distortion of the polytrope is minimal and the spherical approximation is valid .
this latter point will be confirmed in a subsequent section of this paper where we compare spherical and equilibrium initial data . in the physical situation
described in the previous paragraph , the two stars should remain in perfectly circular orbits with constant angular momentum . for this reason we have carried out six simulations in which we compare the effects of the three time centerings for both the rotating frame and inertial frame cases .
the results of these six simulations are shown in figure [ fig : sixpanels ] , which depicts the trajectories of the centers of masses of the stars , and the barycenter of the system , in the orbital plane .
note that the trajectories are terminated at the point where the stars merge or where the simulation was stopped ( if a merger did not occur ) .
the comparison among the matrix of plots reveals that the choice of centering has a major impact on the evolution of the orbits .
the results of the zeus algorithm , which employs a time lagged centering for the gravitational acceleration and is carried out in the inertial frame ( depicted in the bottom right panel ) , show a completely spurious inspiral of the two stars in the first orbit . in contrast , the middle right and the top right panels show inertial frame models with time centered and time advanced gravitational acceleration couplings . while the decay of the orbit is diminished with the time centered and time
advanced couplings , the overall evolution of the orbits is still unstable . in a forthcoming paper , ( calder et al .
2000 ) we shall show that this is a common feature of hydrodynamics simulations of this problem that employ inertial frames .
the decay of the orbits in the inertial frame case is due to the non - conservation of angular momentum .
this is shown directly in figure [ fig : jtot_comp ] where the angular momentum evolution for inertial frame models in the time centered and time
advanced cases are plotted over the first millisecond .
we have carried out additional models for each of these cases with a series of decreasing courant fractions .
we define the courant , or cfl , fraction as the ratio of our actual timestep to the maximal possible hydrodynamic timestep as determined by the timestep control for the zeus / v3d algorithm ( see sn for details ) . in most simulations
we employ a cfl fraction of @xmath139 .
the time - lagged models that we have carried out have revealed an even larger decrease in the angular momentum as a function of time than the time - centered models , which explains the rapid inspiral seen in the bottom right panel of figure [ fig : sixpanels ] .
as figure [ fig : jtot_converge ] shows , the lack of conservation of angular momentum is clearly related to the size of the timestep .
a perfect algorithm would show no evolution of the angular momentum . additionally , while the time
lagged case is much worse than the time
advanced case , both show a significant change in angular momentum over the first millisecond of evolution . in the time - lagged case
this loss results in the decreasing orbits seen in figure [ fig : sixpanels ] . in the time - advanced case the orbit outspirals and the system eventually acquires a small drift due to the slight interactions with the boundaries .
this drift becomes noticeable after many orbits .
the artificial loss of angular momentum in the calculation is due to the inability of the finite - difference scheme to maintain conservation of both angular momentum and linear momentum .
while this loss is mitigated through the use of time - advanced gravitational centering it is still sufficient to cause an unphysical inspiral of the system .
the use of a rotating frame helps to minimize the effects of angular momentum loss . with a rotating frame it is possible to choose the angular velocity of the frame so that the motion of the stars with respect to the frame is minimized .
the advantages of employing a rotating frame are clearly shown in figure [ fig : sixpanels ] . in the rotating frame the advection of the stars across the grid is minimized and the angular momentum is conserved to a much higher degree .
nevertheless , the time centering of the gravitational acceleration plays a role in determining the dynamics of the orbits .
the best combination of techniques is illustrated in the top left panel which shows the results from a simulation using both a rotating frame and the time - advanced coupling .
this particular scheme maintains stable orbits for the two stars for more than seven orbits at which point the simulation was terminated .
the simulation has shown no significant change in the orbits of the two stars over the course of the simulation .
an examination of the angular momentum evolution for this simulation , in figure [ fig : jtot_tarf ] , shows that the total angular momentum is well conserved .
the lines in this figure show the angular momentum contained in matter above various density thresholds .
note that nearly half of the angular momentum is contained in the high - density cores of the polytropes .
also note that the @xmath140 g/@xmath5 line shows that that there is initially a slight re - adjustment in the angular momentum distribution as the star relaxes on the grid .
nevertheless the total angular momentum is fairly well conserved over the course of the simulation .
in contrast the time - lagged inertial - frame case , shown in figure [ fig : jtot_tlff ] , reveals poor angular momentum conservation . this simulation shows a steady decline in the angular momentum at all densities . in particular
the high density core has lost most of the angular momentum .
the loss of angular momentum terminates approximately when the two stars have coalesced into a single central object . at this point
the object is fairly axisymmetric and could almost be though of as having achieved a steady state . under these circumstances
the time centering of the gravitational acceleration is not as critical as it was prior to coalescence and consequently the angular momentum is better conserved at late times .
the behavior of the total energy ( figure [ fig : eng_tlff ] ) in the the time - lagged inertial - frame case shows a slight decline which is not nearly so dramatic as the behavior of the angular momentum . again
, the decline ceases after coalescence .
note that figure [ fig : eng_tlff ] clearly shows the transfer of gravitational potential energy to kinetic energy during the inspiral and coalescence .
the total internal energy changes very little throughout the coalescence .
the time - advanced rotating frame case ( figure [ fig : eng_tarf ] ) shows a very steady behavior for all of the energies , with no substantial change throughout the length of the simulation .
finally , we wish to point out that virtually none of the loss of angular momentum or energy is due to dissipation by the artificial viscosity terms in the gas energy and gas momentum equations .
the total dissipation due to these terms is tracked throughout the simulation , and it is many orders of magnitude below the other energy and angular momentum scales involved .
this includes the case where the stars have coalesced .
we find no significant amount of shock generated dissipation as the stars merge in any of our models .
we have assumed that the angular velocity of the rotating frame with respect to the inertial lab frame is a constant .
thus in situations where the two stars inspiral due to physical processes , the stars will acquire a non - zero velocity with respect to the rotating frame .
in this situation one might suspect that the angular momentum conservation might begin to break down as the stars begin moving with respect to the grid .
however , in the next section we shall show that the angular momentum conservation is still well maintained even in the case where the stars inspiral and merge .
the significant amount of energy and angular momentum non - conservation in the fixed frame calculations clearly establish that there are significant problems associated with their use in modeling binary neutron stars .
we are currently surveying other hydrodynamic methods to see if the same difficulties are present in these other schemes . in order to avoid the problems associated with the inertial cartesian frames we have chosen to employ the rotating frame , time - advanced gravitational acceleration scheme as our preferred method for simulating orbiting and inspiral binary neutron stars . using this method we turn to the study of the stability of equilibrium models .
it has been known for some time that even in the purely newtonian case that tidal instabilities can drive coalescence in binary polytropic systems .
recent semi - analytic stability analyses have been performed by lai , rasio , and shapiro ( 1993a,1993c,1994a,1994c ) and lai and shapiro ( 1995 ) .
these models , which treat the binary polytropes as self - similar ellipsoidal figures of equilibrium , have found that close polytropic binary systems may be unstable to both dynamic and secular instabilities . in this context
we refer to a dynamical instability as one that takes place on the orbital timescale of the binary system while secular instabilities involve dissipative processes that may occur on much longer timescales .
the presence of these instabilities was confirmed numerically by rasio and shapiro ( 1992,1994,1995 ) using sph hydrodynamics methods .
more recently , new and tohline ( 1997 ) have performed similar calculations using eulerian hydrodynamics methods and have found results for the @xmath56 ( @xmath101 ) polytropic sequences that differ from those of rasio & shapiro . in this section
we discuss our investigations of these equilibrium sequences using the time - advanced rotating frame hydrodynamic scheme discussed in the previous section .
the initial data for these equilibrium sequences was constructed as described in section [ sec : eqd ] .
the models that we will discuss were all run at @xmath100 resolution with an approximate size of @xmath141 km in each dimension .
the grid used to construct the equilibrium data was the same grid that was used for the hydrodynamic simulation , thus obviating any introduction of error by remapping the data onto a new grid . since our primary interest is in neutron star mergers we have only carried out simulations for @xmath56 and @xmath98 equilibrium sequences .
the results of our simulations for the @xmath56 equilibrium sequence summarized in figure [ fig : binsep_2_eql ] , which shows the time evolution of the separation between the centers of mass between the two stars .
we have utilized the center - of - mass of the stars to define their separation in the same fashion as lrs .
in contrast with nt we have found pressure maxima to be ill suited for use as a separation diagnostic since extremely small changes in the values of the pressure in a given zone as the stars move can cause a discrete jump in the location of the maxima . because the center - of - mass is density - weighted , the location of these points changes smoothly . .
[ fig : binsep_2_eql ] note that the binary systems with initial separations greater than approximately @xmath142 km seem to be stable over may orbits while those with initial separations less than this radius do not .
normalized to the value of the unperturbed polytropic radius this cutoff corresponds to a separation of @xmath143 .
this is in close agreement with the minimal energy and angular momentum separation which was found for this @xmath56 sequence ( shown in figure [ fig : eq_seq_2 ] ) .
this also corresponds to the point at which the equilibrium sequence transitions from detached to connected binary systems .
this result is in close agreement with the predictions of lrs .
however , while we agree with the conclusion of rasio & shapiro ( 1992 ) that models with @xmath144 are unstable , we do not agree with their finding that the models inspiral on a timescale of 1 - 1.5 times the initial orbital period .
we find that the inspiral occurs on timescales of @xmath145 times the orbital period .
this timescale for evolution of these close systems seems to closely follow those of nt .
while the simulations of nt were not carried out for a time sufficient to show instabilities , the closest separation systems of nt showed an outward evolution comparable to ours over the first four orbital periods . while we seem to agree with the numerical results of nt , we disagree with the conclusion drawn by nt regarding stability of the @xmath56 polytropic sequence .
we see all systems interior to the minimum energy separation inspiral on the timescales of several , i.e. 3 - 5 , orbits .
the hydrodynamic simulations of nt have stopped at four orbital periods . however ,
most of our coalescing systems inspiral at precisely that time .
there is no reason to believe that the dynamical timescale for the inspiral must only be 1 - 2 orbital periods .
while the inspirals could be a result of a secular instability triggered by numerical inaccuracies within the code , it seems more likely that the dynamical process may take slightly longer than what is anticipated by nt .
an interesting feature emerges from figure [ fig : binsep_2_eql ] where we note that the systems with the smallest separations spiral out slightly towards the minimum energy point before undergoing tidal disruption .
similar behavior was seen by nt , who unfortunately terminated their calculations before the point where we see the inspiral occur .
this can be seen from figure 12 of nt , which shows the growth in the moment of inertia of their closest system .
this evolution can be interpreted as an instability that is driving the system towards a lower energy configuration at separations of @xmath146 .
the quality of the total energy and angular momentum conservation for the coalescing models is paramount . as the stars coalesce a significant amount of matter is rapidly advected about the grid even in the rotating frame calculations .
one might suspect that the quality of angular momentum and energy conservation might break down under such circumstances .
however we have found that this does not seem to happen .
this is illustrated by the results for the @xmath147 model which is typical of the coalescing cases .
as figures [ fig : jz_conserve ] and [ fig : eng_conserve ] indicate , both the angular momentum and the energy are well conserved .
the coalescence begins at a time of approximately 8 msec at which time there is a substantial transfer of angular momentum from the high density material to lower density material . as a result of this angular momentum transfer and the disruption of the stars
tidal `` arms '' are formed of material that is stripped from the stars .
these tidal arms contain a significant fraction of the total angular momentum .
some of the material in these arms is swept off of the grid , carrying with it angular momentum . in figure
[ fig : jz_conserve ] we separately track the total angular momentum on the grid at every instant in time along with the cumulative total of the angular momentum swept off of the grid .
the total angular momentum is the sum of these two curves .
the components of the angular momentum displayed in figure [ fig : jz_conserve ] are entirely composed of angular momentum about the z - axis ; the x and y components are effectively zero .
the angular momentum on the grid undergoes a sharp decline during the merger as matter flows off the grid .
this is also reflected in the rise of the cumulative total angular momentum that has been advected off the grid by the matter .
yet the total angular momentum remains quite well conserved .
we wish to emphasize that we have not adjusted the rotation speed of the frame as the stars have inspiraled ; the grid has maintained a constant rotation speed with respect to the laboratory inertial frame .
the introduction of a time - varying grid rotation speed would complicate the hydrodynamic equations and would at best yield only relatively small improvements in angular momentum conservation .
the evolution of the three components of the total energy is shown in [ fig : eng_conserve ] .
there is a slight rise of a few percent in the total energy over the course of the entire simulation but no significant jump during the coalescence .
a modest transfer of kinetic and potential energy occurs during the the merger but this does not have a pronounced effect on the conservation of total energy .
the hydrodynamic evolution of models from the @xmath98 equilibrium sequence shows an instability at a separation of approximately @xmath146 in good agreement with the minimum energy and angular momentum separation of @xmath148 .
the evolution of three models with @xmath100 resolution is shown in figure [ fig : binsep_3_eql ] .
[ fig : binsep_3_eql ] where the binary center - of - mass separation is shown . because of the difficulty of constructing high - resolution equilibrium models for the @xmath98 sequence , we have carried out only five simulations bracketing the predicted point of instability .
our results again agree with the location of the instability identified in rs94 ( see rs94 figure 3 ) .
rs94 found the instability occured at @xmath149 , a value within 10% of the semi - analytic prediction of lrs of a point of instability of @xmath150 .
in contrast , we do not agree with the results of nt who find that binaries at larger separations , e.g. @xmath151 models , are unstable to merger ( see figure 13 of nt ) .
our models with this initial separation exhibit no sign of instability .
a puzzling fact about the nt results for the @xmath98 sequence is that even the largest separation model with an initial value of @xmath152 show signs of a slow orbital decay .
neither we , nor rs , see such behavior .
many numerical investigations of the dynamics of binary neutron star coalescence have employed spherical stars as initial data . as we discussed in section
[ subsec : rot ] the tidal distortions for widely separated stars are small and one can often assume that spherical stars are a good approximation to the true equilibrium fluid bodies .
this assumption clearly breaks down as the separation between the two stars is reduced .
the critical question is as what point does this break down occur ? in order to clarify the realm of validity of the spherical initial data approximation we have carried out a series of simulations using @xmath56 and @xmath98 polytropes as initial data .
the initial separations were varied in the same fashion as the the series of runs for equilibrium data models .
the evolution of these separations for the @xmath56 case is shown in figure [ fig : binsep_2_sph ] . .
[ fig : binsep_2_sph ] at larger separations the systems are stable for long timescales ; the separations to do not significantly evolve .
the small oscillations present reflect the fact that the spherical stars are not in perfect equilibrium initially and consequently the evolving systems undergo small epicyclic oscillations .
similar behavior has been seen by rs94 . for binary separations
nearer the equilibrium sequence stability limit we can see substantial differences between the @xmath56 binaries shown in figure [ fig : binsep_2_sph ] and their equilibrium counterparts shown in figure [ fig : binsep_2_eql ] . for systems with
separations less than 30 km , the orbital separation is diminishing . in the equilibrium case these systems are stable as is seen in figure [ fig : binsep_2_eql ] .
similar behavior is seen in the @xmath98 case as is shown in figure [ fig : binsep_3_comp ] , which compares the equilibrium and spherical - star models . .
[ fig : binsep_3_comp ] the rate at which systems inspiral is clearly high for systems of smaller initial separation .
in fact , the closest systems are disrupted almost immediately .
however , the systems with initially wider separations show no sign of instability .
the results clearly indicate that for systems with initial separations well beyond the tidal instability limit that the spherical - star approximation is quite acceptable .
this point is very important for the case of the coalescence of two rotating neutron stars where one would have to solve the compressible darwin riemann problem in order to obtain equilibrium initial data . by starting sufficiently far beyond the tidal instability limit
one may be able to effectively employ static models of isolated rotating neutron stars as initial data for binary configurations .
furthermore , the isolated star approximation , using post - newtonian models for the isolated stars , could also greatly simplify the construction of initial data for post - newtonian simulations as well .
self - gravitating hydrodynamic models for binary neutron star phenomena pose some unique challenges for numerical modelers .
since the use of eulerian hydrodynamics techniques is prevalent in newtonian , post - newtonian , and relativistic models of binary neutron star coalescence , it is vital that we have a good understanding of the role that the numerical techniques play in determining the outcome of the models . to this end
, we have carried out a number of studies designed to compare rotating and inertial frame newtonian hydrodynamic models as well as to compare several choices that could be made for the coupling of gravity and matter .
the lessons that are learned from this efforts will be invaluable for post - newtonian and relativistic models as well .
we have been able to show that a combination of a rotating frame of reference and a time - advanced gravitational acceleration centering in the gas momentum equation yields adequate angular momentum conservation for the orbiting and merging binary problems . in contrast
, we have found that the use of a inertial laboratory frame together with a time - lagged gravitational coupling yields incorrect results .
the inertial frame methods produces a substantial angular momentum loss that leads to a spurious inspiral of what should be a stable newtonian binary system .
this result indicates that the use of inertial frames which involve stars advecting across the grid should be avoided where possible .
we have found a reliable method for constructing equilibrium initial data on the hydrodynamic grids for use in hydrodynamic simulations .
this method employs a method of solving the poisson problem , including the determination of consistent boundary conditions , for isolated self - gravitating systems without having to resort to multi - pole expansions of the mass distribution .
this iterative method is easily implemented and produces initial data that is consistent with the hydrodynamic grid . using this method
we have constructed equilibrium sequences that closely agree with the semi - analytic calculations of lai , rasio , and shapiro for @xmath56 and @xmath98 polytropic sequences .
while we see qualitative similarities with the results of new & tohline the locations of the minima differ somewhat from theirs and are closer to the lrs predictions . using the initial data from our equilibrium sequences we have investigated the stability of these models .
our results are in very close agreement with the numerical sph models of rasio and shapiro .
in contrast , we find that we disagree with the conclusions of new & tohline on the stability of the @xmath56 and @xmath98 equilibrium sequences .
finally , we investigate the effects of using the isolated star approximation for initial data .
we find that for separations modestly greater than the tidal instability limit that the use of isolated polytropes for initial data has little influence on the subsequent evolution of the binary system .
this point justifies the use of the isolated star approximation for the construction of equilibrium data for rotating , post - newtonian , and other complex binary neutron star systems .
we would like to express our thanks to our colleagues p. annios , g. daues , j. hayes , i. iben , e. seidel , p. saylor , m. norman , w .- m .
suen , i. foster , j. lattimer , p. leppik , m. prakash , p. marronetti , g. mathews , j. wilson , and d. mihalas for many helpful conversations regarding this work .
we would also like to thank john shalf , david bock , john bialek , and andy hall for extensive visualization support .
finally we wish to thank the nasa earth and space science high performance computing and communications program for funding for this work under nasa can s5 - 3099 .
acc would also like to thank the department of energy under grant no .
b341495 to the center for astrophysical thermonuclear flashes at the university of chicago .
computational resources were provided by the national center for supercomputing applications under metacenter allocation # mca97s011n .
in this appendix we briefly describe the details of our hydrodynamic algorithm .
the finite - differencing methodology is identical to that for the zeus algorithm as described by stone & norman ( 1992 ) .
an exception is the addition of the coriolis and centrifugal force terms , which do not appear in sn .
additionally , we have employed only cartesian coordinates , which significantly simplifies the equations as compared to the generalized coordinates of sn . for intimate details of the algorithm
we refer the reader to sn .
we wish to point out one important difference of our algorithm ( v3d ) from the zeus algorithm : although the finite - difference stencils of the equations are identical , the order of updates differs significantly as discussed in section [ subsec : numerical ] .
nevertheless , we still employ a multi - step ( operator split ) methodology as discussed in section [ sec : hydro ] . in the sn nomenclature
the euler equations are broken up into the _ transport _ and _ source _ terms .
the transport step results in the update of the hydrodynamic variables as due to the advective terms of the euler equations , while the source step results in the updates due to the remaining terms .
we describe each of these in turn .
in the transport step , equations ( [ eq : cont_t])-([eq : rgasm_t ] ) are updated .
these equations represent only the advective part of the hydrodynamic evolution . because of the 3-dimensional nature of these equations we employ the widely used dimensional operator splitting technique of strang ( 1968 ) which decomposes the 3-d update into a series of 1-d updates in each dimension . for simplicity we will describe our updates in only the x - direction .
the application to the y- and z - directions is obvious . in each dimension equations ( [ eq : cont_t])-([eq : rgasm_t ] ) are simple conservation laws of the form @xmath153 where @xmath154 is generic variable representing and advected quantity and @xmath155 is the flux of that quantity in the x - direction .
for the five equations ( [ eq : cont_t])-([eq : rgasm_t ] ) @xmath154 takes the respective forms of @xmath15 , @xmath16 , or @xmath156 , while @xmath155 takes the forms of @xmath157 , @xmath158 , and @xmath159 .
the fluxes are calculated for the x - faces of a cell centered around the point at which @xmath154 is defined .
note that these cells will differ for the five variables with the exception of @xmath15 and @xmath16 which are both defined at the same point . for
a given cell the update of @xmath154 will take the form @xmath160 where @xmath161 and @xmath162 are the fluxes on the right and left x - faces of the cell , @xmath163 is the cell width , and @xmath164 is the known value of @xmath154 at timestep @xmath40 . the notation @xmath165 is used to denote that this is only a partial update of @xmath154 only due to advection .
the advection scheme utilizes the consistent advection scheme of norman ( 1980 ) which ties the energy and momentum fluxes to the mass flux , i.e. we define @xmath166 where @xmath167 is the internal energy per gram and @xmath168 where @xmath169 where the @xmath170 indicates that values of variables are calculated using the monotonic advection scheme of van leer ( 1977 ) .
the implementation of this scheme is detailed in sn and we refer the reader to that paper for more information . the implementation of the source step in nearly identical to that of sn . in this step
we solve equations ( [ eq : gase_s ] ) and ( [ eq : rgasm_s ] ) .
note that there are no source terms for the continuity equation and thus the source step does not affect any change in the density .
thus the updates of the internal energy density , @xmath16 , and the momentum density components , @xmath156 , are of the form @xmath171 this update makes use of the intermediate result obtained from the previously undertaken transport step .
since the details of the finite - differencing for most of these source terms are given in sn we refer the reader to that work for further information .
however , in the gas momentum equation the the coriolis and centrifugal force terms are differenced as @xmath172 \\ \label{eq : candcx}\end{aligned}\ ] ] where @xmath173 is the x - coordinate of the grid center . in equation ( [ eq : candcx ] )
we have employed the difference notation as detailed in sn .
the update for the y - velocity component is similarly given by @xmath174 \label{eq : candcy}\end{aligned}\ ] ] where @xmath175 is the y - coordinate of the center of the grid .
note that since we are considering non - inertial frames that rotate around the z - axis , there are no coriolis or centrifugal contributions to the z - component of the momenta .
the coriolis and centrifugal updates are completed in the middle of the source step .
after the updates to the momenta due to pressure and gravitational accelerations have been completed , the velocities are calculated from the momenta .
equations ( [ eq : candcx ] ) and ( [ eq : candcy ] ) are then applied to get the non - inertial force updates to the velocities .
finally , the viscous stress updates to the velocities are completed .
as the last step the source terms for the gas energy equation are solved in order to obtain the new internal energy . | the numerical modeling of binary neutron star mergers has become a subject of much interest in recent years .
while a full and accurate model of this phenomenon would require the evolution of the equations of relativistic hydrodynamics along with the einstein field equations , a qualitative study of the early stages on inspiral can be accomplished by either newtonian or post - newtonian models , which are more tractable . however , even purely newtonian models present numerical challenges that must be overcome in order to have accurate models of the inspiral .
in particular , the simulations must maintain conservation of both energy and momenta , and otherwise exhibit good numerical behavior .
a spate of recent papers have detailed the results for newtonian and post - newtonian models of neutron star coalescence from a variety of groups who employ very different numerical schemes .
these include calculations that have been carried out in both inertial and rotating frames , as well as calculations that employ both equilibrium configurations and spherical stars as initial data .
however , scant attention has been given to the issue of the the accuracy of the models and the dependence of the results on the computational frame and the initial data .
in this paper we offer a comparison of results from both rotating and non - rotating ( inertial ) frame calculations .
we find that the rotating frame calculations offer significantly improved accuracy as compared with the inertial frame models .
furthermore , we show that inertial frame models exhibit significant and erroneous angular momentum loss during the simulations that leads to an unphysical inspiral of the two neutron stars .
we also examine the dependence of the models on initial conditions by considering initial configurations that consist of spherical neutron stars as well as stars that are in equilibrium and which are tidally distorted .
we compare our models those of rasio & shapiro ( 1992,1994a ) and new & tohline ( 1997 ) .
finally , we investigate the use of the isolated star approximation for the construction of initial data . |
the center of our galaxy harbors a @xmath2 bh coinciding with the strong radio source sgra*. x - ray / mev / gev transients in this region are observed on a regular basis .
various astrophysical sources are located close to the gc which may potentially be capable of accelerating particles to multi - tev energies , such as the supernova remnant sgra east or a pulsar wind nebula @xcite .
furthermore , the super - symmetric neutralinos @xmath3 are discussed as potential candidates of dark matter accumulating in the gc region and annihilating into @xmath0-rays @xcite .
the resulting spectrum would have a cut - off near the neutralino mass @xmath4 . assuming a certain dark matter density profile the expected @xmath0-ray flux along the line - of - sight integral can be calculated as a function of @xmath4 and the annihilation cross section @xcite and can in turn be compared to measured upper limits .
egret detected a mev / gev source 3egj1746 - 2851 coincident with the gc position @xcite and recently _
fermi_/lat resolved several sources in the gc region @xcite , see fig .
[ fig : skymap ] .
however , uncertainties in the diffuse galactic background models and limited angular resolution at mev / gev make it difficult to study the morphologies of these sources . at gev
/ tev energies a detection from the direction of the gc was first reported in 2001/02 by the cangarooii collaboration with a steep energy spectrum @xmath5 at the level of @xmath6 of the crab nebula flux @xcite . shortly after , evidence at the level of @xmath7 standard deviations ( s.d . )
was reported from the whipple @xmath8 collaboration @xcite .
the gc was finally confirmed as a gev / tev @xmath0-ray source by the h.e.s.s .
collaboration @xcite ( the position of the supernova remnant sgra east could be excluded as the source of the @xmath0-ray emission ) .
the energy spectrum is well described by a power - law @xmath9 with a cut - off at @xmath10@xmath11 .
the h.e.s.s .
observations revealed a diffuse gev / tev @xmath0-ray component ( dashed contour lines in fig .
[ fig : skymap ] ) which is aligned along the galactic plane and follows the structure of molecular clouds @xcite ; the emission was explained by an interaction of local cosmic rays ( crs ) with matter of the molecular clouds .
the magic collaboration detected the gc ( 7 s. d. ) in 2004/05 observations performed at lza @xcite , followed by a strong veritas lza detection in 2010 @xcite .
veritas sky map of the gc region ( smoothed excess significances , ring background , scale truncated ) .
the black contour lines indicate the gc and the supernova remnant g0.9 + 0.1 as seen by h.e.s.s .
the gray dashed lines indicate the h.e.s.s .
diffuse emission along the galactic plane and from hessj1745 - 303 @xcite .
the position of hessj1741 - 302 is indicated , as well ( circle ) .
the solid circles ( cyan color ) indicate the positions of the mev / gev sources taken from the second _ fermi_/lat catalog @xcite.,scaledwidth=99.0% ] [ [ gc - observations ] ] gc observations + + + + + + + + + + + + + + + due to its declination the gc can only be observed by veritas at lza ( @xmath12 ) strongly decreasing the angular resolution and sensitivity .
the use of the _ displacement _ parameter @xcite , between the center of gravity of the image and the shower position , has been used in the veritas event reconstruction which strongly improved the sensitivity for lza observations @xcite .
the performance and energy reconstruction have been confirmed on lza crab nebula data .
the column density of the atmosphere changes with @xmath13 . in a conservative estimate
, the systematic error in the energy / flux reconstruction can be expected to scale accordingly .
more detailed studied are needed for an accurate estimate ; for the gc observations we currently give a conservative value of a systematic error on the lza flux normalization of @xmath14 . the gc was observed by veritas in 20102012 for @xmath15 ( good quality data , dead - time corrected ) with an average energy threshold of @xmath16 .
[ [ gc - results ] ] gc results + + + + + + + + + + the veritas sky map of the gc region is shown in fig .
[ fig : skymap ] . an @xmath17 s.d .
excess is detected .
no evidence for variability was found in the 3-year data .
the energy spectrum is shown in fig . [ fig : sed ] and is found to be compatible with the spectra measured by whipple , h.e.s.s . , and magic . since the large lza
effective areas of the veritas observations compensate a shorter exposure of low - zenith observations , the statistical errors of the @xmath18 data points are found to be smaller as compared to the h.e.s.s
. measurements .
[ [ diffuse - flux - limit - and - dark - matter - annihilation ] ] diffuse flux limit and dark - matter annihilation + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + off - source observations were performed in a field located in the vicinity of the gc region ( similar zenith angles and sky brightness ) without a known tev @xmath0-ray source .
these observations are used to study the background acceptance throughout the field of view and will allow the estimate of a diffuse @xmath0-ray component surrounding the position of the gc .
an upper limit of the diffuse @xmath0-ray flux can in turn be compared with line - of - sight integrals along the density profile @xmath19 , in order to constrain the annihilation cross section for a particular dark matter model , dark matter particle mass and density profile @xmath20 . due to its likely astrophysical origin the excess at the gc itself , as well as a region along the galactic plane , will be excluded from this analysis ( work in progress ) . [
[ hadronic - models ] ] hadronic models + + + + + + + + + + + + + + + hadronic acceleration models @xcite are discussed involving : ( i ) hadrons being accelerated in the bh vicinity ( few tens of schwarzschild radii ) .
( ii ) the accelerated protons diffuse out into the interstellar medium where they ( iii ) produce neutral pions which decay into gev / tev @xmath0-rays .
linden et al .
( 2012 ) discuss the surrounding gas as proton target defining the morphology of the tev @xmath0-ray emission @xcite .
changes in @xmath0-ray flux in those models can be caused by changing conditions in the bh vicinity ( e.g. accretion ) .
the time scales of flux variations are @xmath10@xmath21 at mev / gev energies ( old flares ) and @xmath10@xmath22 at @xmath23 ( new flares caused by recently injected high - energy particles ) @xcite .
constraining the @xmath23 spectral variability would serve as an important test for this class of models .
[ [ leptonic - models ] ] leptonic models + + + + + + + + + + + + + + + atoyan et al .
( 2004 ) @xcite discuss a bh plerion model in which a termination shock of a leptonic wind accelerates leptons to relativistic energies which in turn produce tev @xmath0-rays via inverse compton scattering .
the flux variability time scale in this model is on the order of @xmath24@xmath25 .
the hadronic and the leptonic models are shown together with the veritas/_fermi _ data in fig .
[ fig : sed ] ( right ) .
the leptonic model clearly fails in explaining the flux in the mev / gev regime .
however , this emission may well originate from a spatially different region or mechanism other than the tev @xmath0-ray emission .
the hadronic models can explain the sed by the superposition of different flare stages .
fermi_/veritas flux correlation studies , as well as the measurement of the tev energy cut - off and limits on the @xmath26 variability will serve as crucial inputs for the modeling .
veritas is capable of detecting the gc within @xmath27 in zenith angle greater than @xmath28 observations .
the measured energy spectrum is found to be in agreement with earlier measurements by h.e.s.s . , magic , and whipple .
future observations to measure the cut - off energy in the spectrum and to determine limits on the flux variability at the highest energies will place constraints on emission models . the recently discovered giant molecular cloud heading towards the immediate vicinity of the gc bh @xcite represents further motivation for future tev @xmath0-ray monitoring of this region . an upper limit on diffuse @xmath0-ray emission and , in consequence , a limit on the photon flux initiated by the annihilation of dark matter particles is work in progress . this research is supported by grants from the u.s .
department of energy office of science , the u.s . national science foundation and the smithsonian institution , by nserc in canada , by science foundation ireland ( sfi 10/rfp / ast2748 ) and by stfc in the u.k .
we acknowledge the excellent work of the technical support staff at the fred lawrence whipple observatory and at the collaborating institutions in the construction and operation of the instrument . | the galactic center ( gc ) has long been a region of interest for high - energy and very - high - energy observations .
many potential sources of gev / tev @xmath0-ray emission are located in the gc region , e.g. the accretion of matter onto the central black hole ( bh ) , cosmic rays from a nearby shell - type super nova remnant , or the annihilation of dark matter .
the gc has been detected at mev / gev energies by egret and recently by _
fermi_/lat .
at tev energies , the gc was detected at the level of 4 standard deviations with the whipple @xmath1 telescope and with one order of magnitude better sensitivity by h.e.s.s . and magic .
we present the results from 3 years of veritas gc observations conducted at large zenith angles ( lza ) .
the results are compared to astrophysical models .
address = department of physics and mcss , washington university , st .
louis , mo , usa , [email protected] address = http://veritas.sao.arizona.edu/ |
the past few years have seen extensive investigations of the @xmath1 meson producing reactions at close to threshold energies .
the experiments are aimed either at directly searching @xcite for the possible existence of eta - mesic nuclei as a result of the strong attractive nature of the @xmath1-n interaction @xcite or studying the final state eta - nuclear interaction to eventually conclude on the existence of eta - mesic nuclear states @xcite .
a common feature of the data on @xmath1 production near threshold is the strong deviation of the cross sections from phase space .
it can be understood as a manifestation of the strong attractive @xmath1-n interaction ( arising basically due to the proximity of the @xmath1-n threshold to the @xmath9 resonance n@xmath10(1535 ) ) .
experiments on @xmath1 production have been performed in nucleon - nucleon collisions and have been extended to proton - deuteron and deuteron - deuteron collisions too @xcite . though the focus of reactions such as the p + d @xmath11 p + d + @xmath1 and @xmath12 is on investigating possibilities of bound states of eta mesons and 2 - 3 nucleon systems , theoretical studies of the reaction mechanism revealed interesting features too @xcite . in the p - d collisions ,
the production near threshold is found to be dominated by a two - step mechanism where the large momentum transfer in producing the @xmath1 meson is shared by three nucleons .
these findings naturally led to the curiosity of what happens when the @xmath1 interacts with more than three or four nucleons in the final nucleus . with this motivation ,
measurements of the @xmath0 reaction were carried out by the turin group in 1993 @xcite at an incident energy of 683 mev .
a theoretical study of this reaction along with others of the type , @xmath13 and @xmath14 was performed in @xcite .
part of the emphasis of this work was on obtaining the right form factors for @xmath2li ( and @xmath2be ) and the interactions of the mesons with the nuclei in the final states were neglected .
the interest in the @xmath0 reaction revived once again by the recent proposal of studying this reaction at cosy , jlich , at an incident energy of 673 mev @xcite .
having performed detailed theoretical studies of the p + d @xmath11 p + d + @xmath1 and @xmath12 reactions @xcite and the @xmath1 meson interactions with the deuteron , @xmath5he and @xmath15he nuclei @xcite , we now develop a model to study the interaction of @xmath1 mesons with a @xmath5he-@xmath15he cluster , namely , the @xmath2be nucleus .
the @xmath1-@xmath2be interaction is then incorporated in a theoretical calculation of cross sections for the @xmath0 reaction , with four possible low - lying states of the @xmath2be nucleus .
an analysis of its effects on the @xmath0 cross sections presented here , should be useful in motivating further experimental studies of this reaction .
based on literature which supports considering the @xmath2be nucleus as a bound state of an alpha ( @xmath15he ) and @xmath5he @xcite , we model the @xmath1-@xmath2be final state interaction in the form of a three body problem of the @xmath1-@xmath5he-@xmath15he interaction . regarding the @xmath3li too as a cluster of an alpha and a deuteron @xcite , the @xmath0 reaction is considered to proceed through the @xmath12 reaction with the @xmath4 remaining a spectator . besides , the present work focuses on the low energy region of @xmath1 production where ( a ) the @xmath12 production amplitude is large ( i.e. the p and d interact strongly to produce an @xmath1 ) and ( b ) the cluster picture of low lying levels of @xmath2be and @xmath3li is reasonably good .
there exists in principle , the possibility of considering the deuteron as a spectator .
however , a reaction of the type @xmath16 followed by the combination of @xmath17 and a deuteron to form the @xmath2be nucleus , does not agree with the cluster model approach here ( since @xmath2be is hardly known to be a cluster of d + @xmath18li , where @xmath18li is in fact unstable ) .
the possibility of an intermediate @xmath19 reaction , with further steps of the final state @xmath20 from this reaction combining with the spectator deuteron , i.e. , @xmath21 forming a @xmath5he which eventually combines with the final state @xmath4 ( from the above reaction ) to form @xmath2be still remains .
however , this is not a practical option with no information available on the @xmath22 reaction . as mentioned above
, we consider the @xmath2be nucleus as a two body system made up of a @xmath5he and @xmath15he nucleus and construct an elastic transition matrix for the three body problem of an @xmath1 meson , @xmath5he nucleus and an @xmath4 ( @xmath15he ) .
the individual scattering of the @xmath1 meson on @xmath5he and @xmath4 is evaluated using the @xmath23-matrices constructed earlier by the present authors @xcite .
these t - matrices are numerically evaluated using few body equations and include the off - shell rescattering of the @xmath1 on the nucleons inside @xmath5he and @xmath15he .
the @xmath1-@xmath5he t - matrix is tested to reproduce the @xmath12 cross section reasonably well @xcite . to formulate the three body problem of the @xmath1-@xmath5he-@xmath15he interaction ,
let us define @xmath24 and @xmath25 to be the coordinates of the @xmath5he and @xmath15he nuclei respectively , with respect to the mass-7 centre of mass system . defining the internal jacobi coordinate of the relative distance between the @xmath5he and @xmath15he nuclei as @xmath26
, one can see that @xmath27 and @xmath28 with @xmath29 and @xmath30 .
the @xmath1-@xmath2be t - matrix is then written as , @xmath31\ ] ] where @xmath32 and @xmath33 are the medium modified t - matrices for the off - shell @xmath1 scattering on the bound @xmath5he and @xmath15he respectively .
@xmath34 represents the cluster wave function of @xmath2be with angular momentum @xmath35 and @xmath36 , where , @xmath37 is the total @xmath1-nucleus energy in the centre of mass and @xmath38 is the energy required for the break up of @xmath2be @xmath39he+@xmath15he .
the in - medium @xmath1-@xmath5he and @xmath1-@xmath15he t - matrices are written using a faddeev type decomposition @xcite , namely , @xmath40 where @xmath41 and the indices @xmath42 and @xmath43 correspond to the @xmath5he and @xmath15he t - matrices respectively .
the @xmath44 s represent the single scattering terms and are the matrices for purely elastic @xmath1-@xmath5he and @xmath1-@xmath15he scattering .
they have the form , @xmath45 \ , , \ ] ] with @xmath46 and @xmath41 as mentioned above . at the low energies considered here
, the @xmath1-n interaction is dominated by the @xmath9 resonance n*(1535 ) and hence we perform a partial wave expansion and retain only @xmath47-waves , which reduces ( [ eta3he4he ] ) to the following two equations for @xmath1-@xmath5he ( @xmath48 ) and @xmath1-@xmath15he ( @xmath49 ) : @xmath50 @xmath51 where @xmath52 in eq .
( [ tetan ] ) has reduced to @xmath53 written in terms of the spherical bessel functions @xmath54 and @xmath55 . replacing the equation for @xmath49 in @xmath48
, we obtain a recursive relation for @xmath48 as follows , @xmath56 once @xmath48 is evaluated from ( [ eta3he ] ) , it can be replaced into the equation for @xmath49 and the two can be substituted in ( [ etabetmat ] ) to evaluate the @xmath1-@xmath2be t - matrix .
thus ( [ etabetmat ] ) is evaluated retaining only @xmath47-waves in @xmath48 and @xmath49 .
@xmath48 is evaluated numerically using the @xmath1-@xmath5he and @xmath1-@xmath15he t - matrices , @xmath57 and @xmath58 respectively , as inputs . @xmath57 and @xmath58 are themselves evaluated numerically using few body equations and an input coupled channel @xmath1-n t - matrix .
details of this formalism can be found in @xcite .
the two models of the elementary @xmath59 t - matrix used here will be discussed in the next subsection . the coupled channel @xmath1-n t - matrix which is required for the evaluation of the @xmath1-@xmath5he and @xmath1-@xmath15he t - matrices , @xmath57 and @xmath58 , is taken from two different models available in literature . in @xcite , a transition matrix including the @xmath60n and @xmath1n channels with the n*(1535 ) resonance playing a dominant role was constructed .
it consisted of the meson - n * vertices and the n * propagator as given below : @xmath61 with , @xmath62 , where @xmath63 @xmath64 are the self energy contributions from the @xmath65 and @xmath59 loops .
in @xcite elastic and inelastic @xmath1-deuteron scattering as well as @xmath1 photoproduction on the deuteron was studied using this @xmath59 model .
we shall use two parameter sets available , one which leads to a scattering length of @xmath66 = ( 0.75 , 0.27 ) fm and another which leads to @xmath66 = ( 0.88 , 0.41 ) fm .
we shall refer to this model henceforth as model a. model b used in the present work is taken from @xcite .
the t - matrix for @xmath67 is written in a separable form as , @xmath68 where the on - shell part , @xmath69 , is described in an effective range approximation as , @xmath70 the off - shell form factors have the form , @xmath71 , where @xmath72 is the length parameter in this model .
the parameter sets in this model are obtained from a fit to the @xmath73 , @xmath74 , @xmath75 and @xmath76 data .
the parameters required in eqs .
( [ green1 ] ) and ( [ green2 ] ) above can be found in @xcite .
we choose four parameter sets with @xmath59 scattering lengths of ( 0.88,0.25 ) fm , ( 0.77 , 0.25 ) fm , ( 0.51,0.26 ) fm and ( 0.4 , 0.3 ) fm . the choice of kinematic variables in a multiple scattering formalism where the many body scattering matrix is written in terms of a two body matrix is not unique in literature .
the two most commonly used approaches are ( i ) the fixed scatterer approximation ( fsa ) and ( ii ) the on - energy shell impulse approximation ( oei ) @xcite ( related to yet another approach , namely , the fixed impulse approximation ( fia ) @xcite ) . the difference between these approaches ( which becomes important at intermediate and high energies ) lies in the fact that in the fsa , the struck nucleon is assumed to recoil with the target as a whole , while in the oei , it recoils freely .
this means that in the fsa , the two - body operator does not follow from the equation for a free two - body t - matrix , but rather contains the mass of the nucleus ( in which the two - body system is embedded ) in the kinetic energy .
though this seems to be mathematically sound , such a two body t - matrix can have physically undesired features .
for example , in @xcite it was found from a phase shift study that such a two - body t - matrix may not display the resonant behaviour which it would be expected to . in the same work , in connection with @xmath60-nucleus scattering , it was found that considering a free two - body matrix simulated the contribution of continuum states ( otherwise neglected in that work ) and brought theory in closer agreement with data . though the differences arising from the particular approaches used above may not be crucial at the low energies considered in the present work , the many - body problem considered here is a bit more complicated and hence a small discussion is in order . in the works mentioned above ,
one studies the differences of the approaches involved , in a multiple scattering of a hadron on the individual nucleons in the nuclear target .
here however , the problem appears to be that of one multiple scattering problem embedded inside another .
the @xmath1 mesons scatter off the @xmath5he and @xmath15he nuclei as in a three body multiple scattering problem ( of the @xmath1-@xmath5he-@xmath15he system ) .
however , the individual @xmath1-@xmath5he and @xmath1-@xmath15he scatterings are represented by t - matrices for the multiple scatterings of the @xmath1 on the three and four nucleons inside @xmath5he and @xmath15he .
we choose then to work in a framework where we start by evaluating the momenta @xmath77 and @xmath78 in the @xmath1-@xmath2be centre of mass frame and then evaluate the individual @xmath1-@xmath5he and @xmath1-@xmath15he t - matrices at a lorentz boosted energy - momentum in the @xmath1-@xmath5he and @xmath1-@xmath15he centre of mass systems .
the energy in the in - medium propagators ( inside @xmath2be ) is however taken to be in the @xmath1-@xmath2be centre of mass system .
since the energy spacing between the first four low - lying levels of the @xmath79 nucleus is small , we include the contribution from these four levels .
we consider the angular momentum states with @xmath80 and @xmath81 corresponding to the @xmath82 and @xmath83 levels respectively .
the cluster wave functions for @xmath2be and those for @xmath3li required in the production amplitude of the @xmath0 reaction are ( 1 ) generated using a wood - saxon potential @xcite and ( 2 ) taken from a green s function monte carlo ( gfmc ) variational calculation with the urbana potential @xcite . in the gfmc case , the wave function for @xmath2li is used to represent the @xmath2be one .
the @xmath4 and deuteron cluster in @xmath3li is assumed to be in the @xmath84 state .
assuming that the beam proton interacts with a loosely bound deuteron in the @xmath3li nucleus to produce an @xmath1 meson and @xmath5he ( with the @xmath4 remaining a spectator ) the production amplitude for the @xmath0 reaction can be written in terms of that for the @xmath12 process .
the off shell @xmath1 meson thus produced then rescatters on @xmath2be ( i.e. the @xmath5he-@xmath15he cluster ) .
this production of the @xmath1 and its final state interaction ( fsi ) with @xmath2be is represented schematically in fig . 1 , and the corresponding transition matrix is written as , @xmath85 where @xmath86 and @xmath87 are the initial and final momenta in the centre of mass system .
@xmath88 , @xmath89 and @xmath90 are the spin projections of the proton , @xmath91 and @xmath79 respectively .
cluster model for @xmath1 production in the @xmath0 reaction .
diagram ( a ) corresponds to the direct on - shell @xmath1 production and ( b ) to an @xmath1 which is first produced off - shell and rescatters via the @xmath1 + @xmath2be @xmath92 + @xmath2be process to become on - shell .
, title="fig:",width=264,height=188 ] cluster model for @xmath1 production in the @xmath0 reaction .
diagram ( a ) corresponds to the direct on - shell @xmath1 production and ( b ) to an @xmath1 which is first produced off - shell and rescatters via the @xmath1 + @xmath2be @xmath92 + @xmath2be process to become on - shell . , title="fig:",width=264,height=188 ] + the production matrix for a relative angular momentum @xmath35 between the @xmath93 and @xmath4 is written as , @xmath94 where , @xmath95 is the momentum of the incoming proton and @xmath96\vec{k}_p + \vec{p}_1 $ ] of the deuteron in @xmath3li .
@xmath97 is the spin projection of @xmath5he .
the on - shell @xmath1 meson momentum is denoted as @xmath98 and the off - shell one as @xmath99 .
@xmath100\vec{k}_{\eta } + \vec{p}_2 $ ] or @xmath100\vec{q } + \vec{p}_2 $ ] is hence the momentum of on- or off - shell @xmath5he in @xmath2be .
since the @xmath4 particle remains a spectator , its momentum in @xmath3li and @xmath2be is required to be the same and @xmath101\vec{k}_p - \vec{p}_1 \ , = \
, -[4/7]\vec{k}_{\eta } ( \vec{q } ) - \vec{p}_2 $ ] ( for on - shell ( off - shell ) @xmath1 production ) .
thus @xmath102\vec{k}_p\,-\,[4/7]\vec{k}_{\eta } ( \vec{q})$ ] ( where @xmath103 and @xmath104 are the fermi momenta inside @xmath3li and @xmath2be respectively ) . the integration in ( [ prodtmat2 ] ) should in principle have been over both the fermi momenta , @xmath103 and @xmath104 , however , the above relation between them renders the integration over @xmath104 in ( [ prodtmat2 ] ) redundant .
further , when one evaluates the unpolarized cross sections , one sums over the spins in the final state and averages over those in the initial state . as a result , in such a calculation , some sums in ( [ prodtmat1 ] ) and ( [ prodtmat2 ] ) become redundant .
the @xmath105matrix for the process , @xmath106 , is written in a two - step model from our earlier work @xcite .
considering the complexity of the present calculations which include the off - shell rescattering as given by the second term in ( [ prodtmat1 ] ) with the @xmath1-@xmath2be fsi and the fact that the two - step model of the @xmath12 is itself quite involved , we neglect the effect of fermi motion on @xmath106 and hence take it out of the integral over @xmath107 in ( [ prodtmat2 ] ) .
the momentum space wave functions are expressed in terms of fourier transforms of their radial forms .
thus the integral in momentum space is transformed to that in coordinate space .
all this simplifies eq .
( [ prodtmat2 ] ) to a good extent and it can be written as , @xmath108 the @xmath3li-@xmath2be transition form factor for angular momentum states with @xmath80 and @xmath81 .
the solid line corresponds to the gfmc variational wave functions with the urbana potential for @xmath80 and the dashed , dot - dashed and double dot - dashed lines are those generated using a woods - saxon potential ( for @xmath80 and @xmath81).,width=264,height=264 ] is the transition form factor for @xmath3li @xmath109 @xmath2be , with momentum transfer @xmath110 for example in the off - shell case . though not written explicitly , the transition form factor depends on the total angular momentum @xmath111 since the radial wave function in @xmath2be depends , even if mildly , on @xmath111 . in fig .
2 we present the two form factors with @xmath112 ( @xmath113 ) and @xmath114 ( @xmath115 ) required in the present work , using two different prescriptions of the nuclear wave functions mentioned in the previous section . with the @xmath80 levels , @xmath116 and @xmath117 being very close to each other , the difference between the two @xmath80 form factors is not visible in the figure .
the form factor enters ( [ prodtmat1 ] ) via ( [ equation8 ] ) as ( [ prodtmat2 ] ) is simplified to ( [ equation8 ] ) due to the neglect of the fermi motion in the @xmath118 t - matrix .
it is evaluated at @xmath119 in the first term of ( [ prodtmat1 ] ) whereas over a range of momenta in the second term of ( [ prodtmat1 ] ) where it appears inside the integral .
the dotted vertical lines in fig .
2 indicate the relevant range corresponding to the beam energies of the present work .
we shall see later how this difference between the two types of form factors affects the cross sections .
in what follows , we shall present the cross section calculations for the @xmath0 reaction with an emphasis on the @xmath1-@xmath2be fsi .
the transition matrices for the elementary processes @xmath60 + n @xmath120 + n and @xmath1 + n @xmath120 + n which enter as inputs to the production and fsi matrices are chosen from coupled channels calculations . the calculations in figs 3 - 7 are done within model a with a choice of parameters which corresponds to a scattering length of @xmath121 fm . within the two - step model of the @xmath12 reaction we use here ,
the data on this reaction were reproduced well with this choice of the scattering length . in fig .
8 , a comparison of the total ( inclusive ) cross sections within models a and b of the @xmath67 t - matrix and using different sets of scattering lengths is made .
total cross sections as a function of the proton beam energy for different states of the @xmath2be nucleus .
the dashed lines are plots without the inclusion of the @xmath1-@xmath2be final state interaction .
the wave functions for @xmath3li and @xmath2be are generated using the woods - saxon potential.,width=377,height=377 ] same as fig .
3 except for the fact that the wave functions for @xmath3li and @xmath2be are from the gfmc variational method with the urbana potential.,width=377,height=188 ] .beam and excess energies for different levels of the @xmath2be nucleus [ cols= " < , < , < ,
< , < " , ] the smallest scattering length @xmath66 used here , leads to a positive real part of the @xmath1-@xmath2be scattering length , @xmath122 .
however , this choice of parameters ( model f in @xcite ) is mentioned as an unconventional solution obtained after dropping the photoproduction data from the fits . the first four entries in table ii corresponding to large @xmath66 , display @xmath122 to be large with negative real parts .
apart from the commonly known condition that the real part of the scattering length should be negative @xcite , in the third reference in @xcite , the authors found that the condition for the existence of a bound state is that @xmath123 for an eta - nucleus scattering length of ( @xmath124 ) .
with these two conditions it seems that the first few entries in table ii for the large @xmath59 scattering lengths support the possibility of eta - mesic states . in @xcite , while investigating the connection between the @xmath1-@xmath5he scattering lengths and the corresponding binding energies and widths , the authors also use the above conditions but mention that in reality none of the above can be taken as a sufficient condition .
the bottomline is then that it would indeed be premature to base the conclusions only on the signs or magnitudes of the scattering lengths .
one should rather perform a better analysis for the search of @xmath1-@xmath2be states using the present @xmath1-@xmath2be model and a time delay analysis as in @xcite or a k - matrix analysis as in @xcite before drawing definite conclusions .
the present work aimed at performing a thorough investigation of the effects of the @xmath1-@xmath2be interaction in the @xmath0 reaction near threshold .
the work was partly motivated by the recent revival of interest in this reaction by the cosy - gem collaboration @xcite after the first measurement in 1993 .
this work also comes as a sequel to our various earlier studies on @xmath1 meson production in light nuclei . a two - step model for the @xmath12 reaction including
the @xmath1-@xmath5he interaction was tested earlier with data by the present authors @xcite .
this model is used as an input to develop a cluster model approach for the @xmath0 reaction near threshold .
the @xmath1-@xmath2be interaction is included in a multiple scattering formalism for an @xmath1 scattering on a @xmath5he-@xmath15he cluster inside @xmath2be .
the @xmath1-@xmath5he and -@xmath15he scatterings are themselves included using few body equations .
the calculations are done for four low - lying levels of @xmath2be . to the best of our knowledge ,
this is the most detailed study of the @xmath1-@xmath2be interaction in the @xmath0 reaction performed so far .
the interesting two hump structure in the summed total cross section ( summed over @xmath111 ) hints toward a very strong near threshold effect of the @xmath1-@xmath2be interaction ( especially of the off - shell rescattering of the @xmath1 on @xmath2be ) which is worth verifying experimentally in future .
+ + * acknowledgements * + the authors gratefully acknowledge v. jha for the help in computing the cluster wave functions using the woods - saxon potential and k. p. khemchandani for her help with the @xmath125he programs and many useful discussions .
the authors are also thankful to the anonymous referee for his / her constructive criticism .
the work done by two of the authors , nju and bkj , was supported by the ramanna fellowship , awarded by the department of science and technology , government of india , to bkj .
+ 99 m. pfeiffer _
et al_. , phys .
lett . * 92 * , 252001 ( 2004 ) ; g. a. sokol _
et al . _ ,
. lett . * 102 * , 71 ( 2000 ) .
r. s. bhalerao and l. c. liu , phys .
* 54 * ( 1985 ) 865 ; l.c .
liu and q. haider , phys .
c * 34 * , 1845 ( 1986 ) ; q. haider and l.c .
liu , phys .
c * 66 * , 045208 ( 2002 ) .
n. g. kelkar , k. p. khemchandani and b. k. jain , j. phys .
g : nucl . part .
phys . * 32 * , l19 ( 2006 ) ; n. g. kelkar , k. p. khemchandani and b. k. jain , j. phys . *
g 32 * , 1157 ( 2006 ) ; n. g. kelkar , phys .
lett . * 99 * , 210403 ( 2007 ) .
s. wycech , anthony m. green and j.a .
niskanen , phys .
* c 52 * , 544 ( 1995 ) ; v.a .
tryasuchev , phys . atom .
60 * , 186 ( 1997 ) ( yad . fiz . * 60 * , 245 ( 1997 ) ) ; c.y .
song , x.h .
zhong and l. li , p.z .
ning , europhys .
* 81 * , 42002 ( 2008 ) ; d. jido , e.e .
kolomeitsev , h. nagahiro and s. hirenzaki , nucl . phys . * a 811 * , 158 ( 2008 ) ; h. nagahiro , d. jido and s. hirenzaki , arxiv:0811.4516 .
j. berger _
et al . _ ,
phys . rev .
lett . * 61 * , 919 ( 1988 ) ; b. mayer _ et al_. , phys .
c * 53 * , 2068 ( 1996 ) ; a. wroska _
et al . _ ,
a * 20 * , 640 ( 2005 ) ; h. , -h .
, adam _ et al .
c * 75 * , 014004 ( 2007 ) .
khemchandani , n.g .
kelkar and b.k .
jain , nucl .
phys . * a 708 * , 312 ( 2002 ) ; _ ibid _ , phys . rev .
c * 68 * , 064610 ( 2003 ) ; n.j .
upadhyay , k.p .
khemchandani , b.k . jain and n.g .
kelkar , phys .
c * 75 * , 054002 ( 2007 ) ; k. p. khemchandani , n. g. kelkar and b. k. jain , phys .
c * 76 * , 069801 ( 2007 ) .
scomparin e. _ et al .
_ , j. phys .
* 19 * l 51 , ( 1993 ) . j. s. al - khalili , m. b. barbaro and c. wilkin , j. phys .
g * 19 * , 403 ( 1993 ) .
m. ulicny _
et al . _ , aip conf
* 603 * , 543 ( 2001 ) ; h. machner , acta phys .
slov . * 56 * , 227 ( 2006 ) ; e - print : nucl - ex/0511034 .
t. kajino , t. matsuse and a. arima , nucl .
a * 413 * , 323 ( 1984 ) .
j. v. noble , phys .
c * 9 * , 1209 ( 1974 ) ; a. c. merchant and n. rowley , phys .
b * 150 * , 35 ( 1985 )
. s. a. rakityansky _
c * 53 * , 2043 ( 1996 ) ; s. a. rakityansky _
_ , few body syst .
* 9 * , 227 ( 1995 ) .
a. fix and h. arenhvel , eur .
j a * 9 * , 119 ( 2000 ) .
a. m. green and s. wycech , phys .
c * 71 * , 014001 ( 2005 ) .
e. kujawski and e. lambert , annals of physics * 81 * , 591 ( 1973 ) .
r. crespo , a. m. moro and i. j. thompson , nucl .
phys . a * 771 * , 26 ( 2006 ) .
v. b. belyaev , s. a. rakityansky and j. wrzecionko , nucl .
a * 368 * , 394 ( 1981 ) .
b. buck and a. c. merchant , j. phys .
g * 14 * l211 ( 1988 ) ; v. i. kukulin , v. g. neudatchin and yu .
f. smirnov , nucl .
phys . a * 245 * ( 1975 ) 429 .
j. l. forest _
et al . _ ,
phys . rev .
* c 54 * , 646 ( 1996 ) . c. j. joachain 1975
_ quantum collision theory _ ( north - holland ) , p.288 .
a. sibirtsev , j. haidenbauer , j. a. niskanen and ulf - g .
meissner , phys . rev . * c 70 * , 047001 ( 2004 ) ; see also j. a. niskanen , arxiv : nucl - th/0508021 .
n. g. kelkar , m. nowakowski , k. p. khemchandani and s. r. jain , nucl . phys .
* a730 * , 121 ( 2004 ) ; n. g. kelkar , m. nowakowski and k. p. khemchandani , nucl . phys . * a724 * , 357 ( 2003 ) ; _ ibid _ , mod .
a * 19 * , 2001 ( 2004 ) .
s. wycech and a. m. green , int .
phys . * a 20 * , 637 ( 2005 ) . | the @xmath0 reaction has been investigated with an emphasis on the @xmath1 meson and @xmath2be interaction in the final state .
considering the @xmath3li and @xmath2be nuclei to be @xmath4-d and @xmath4-@xmath5he clusters respectively , the reaction is modelled to proceed via the @xmath6 \
, \rightarrow \ , ^3{\rm he } [ \alpha ] \ , + \ , \eta$ ] reaction with the @xmath4 remaining a spectator .
the @xmath1 meson interacts with @xmath2be via multiple scatterings on the @xmath5he and @xmath4 clusters inside @xmath2be .
the individual @xmath1-@xmath5he and @xmath1-@xmath4 scatterings are evaluated using few body equations for the @xmath1 - 3n and @xmath1 - 4n systems with a coupled channel @xmath1-n interaction as an input .
calculations including four low - lying states of @xmath2be lead to a double hump structure in the total cross section corresponding to the @xmath7 and @xmath8 angular momentum states .
the humps arise due to the off - shell rescattering of the @xmath1 meson on the @xmath2be nucleus in the final state . |
at certain conditions , the critical state of a type - ii superconductor may become unstable @xcite , then thermomagnetic avalanche may develop . during the thermomagnetic avalanche , temperature of the superconducting sample increases rapidly and large amount of the magnetic flux enters into the sample volume .
thermomagnetic avalanches , called flux jumps , are problematic from a viewpoint of practical applications of superconductors , because they may drive a superconducting sample into a normal or , at least , into a resistive state .
the stability of critical state depends strongly on a relation between the thermal ( @xmath3 ) and magnetic ( @xmath4 ) diffusivity .
@xmath5 , where @xmath6 is the thermal conductivity , and @xmath7 is the specific heat at constant volume . @xmath8 . in the flux flow regime , @xmath9 , where @xmath10 is the flux flow resistivity
, @xmath11 is the normal state resistivity , @xmath12 is the second critical field and @xmath13 is the magnetic permeability of vacuum . in the case of hard conventional superconductors , usually @xmath14 .
such conditions are called `` locally adiabatic '' .
the most commonly studied parameter , which determines the stability of the critical state is a field of the first ( after cooling the sample in zero magnetic field ) flux jump , @xmath15 .
this parameter is closely correlated with the critical penetration depth , @xmath16 .
if the critical current density , @xmath17 , does not depend on the magnetic field , @xmath18 .
this parameter is very important from a viewpoint of the practical applications of the superconductors , because it is possible to avoid the thermomagnetic avalanches , if the width of the superconducting sample is smaller than @xmath19 .
we can also express the condition of the stability of the critical state in a form : @xmath20 , where @xmath21 is the field of full penetration , @xmath22 , and @xmath23 is half - width of the investigated sample . in the simplest approximations ,
the sample is usually assumed to be of a shape of an infinite slab , of width @xmath24 , and an external magnetic field is applied parallel to its surface . in locally adiabatic conditions ,
the field of the first flux jump can be expressed by a formula @xmath25 the thermal parameter , which determines the critical state stability in locally adiabatic conditions , is the specific heat @xmath7 . in locally adiabatic conditions ,
the stability of the critical state is not influenced by the cooling conditions at the surface of the superconducting sample .
the situation is different in a case when @xmath26 .
such conditions usually occur in superconducting composites and are called `` dynamic '' . in dynamic conditions , the thermal parameter which determines the stability of the critical state is the thermal conductivity , @xmath6 , or the thermal boundary conductivity , @xmath27 @xcite .
some problems connected with the critical state stability in high temperature superconductors ( htsc ) are not clarified yet .
the problems are connected with the complex structure of these materials .
the analysis of the critical state stability in htsc is usually performed within approximations developed for conventional superconductors @xcite , which sometimes lead to wrong results .
because of the increasing number of applications of htsc , it is very important to understand all aspects of the thermomagnetic avalanche development in these materials .
one of the specific features of htsc is their strong crystallographic anisotropy .
experimental data show that high temperature superconductors are usually more stable against flux jumping than conventional superconductors .
the larger stability of the high temperature superconductors against flux jumping can be correlated with the flux creep phenomenon , which is usually strong in these materials @xcite . in studies of the magnetization of superconductors ,
flux creep manifests itself as a relaxation of the magnetic moment .
the process of magnetic relaxation reflects strongly nonlinear current - voltage characteristics of superconducting sample .
these characteristics can be approximated by the formula @xmath28 where @xmath29 is the voltage criterion at which the critical current density , @xmath17 , is defined , and @xmath30 is a dimensionless parameter .
one usually assumes @xmath31 v / m . in this case @xmath32 .
the influence of nonlinear current - voltage characteristics on the critical state stability , in the case of a weakly cooled sample ( for @xmath33 , @xmath34 is so - called biot number ) , was analyzed in ref.@xcite .
it was shown that in such conditions the critical state stability depends on the external magnetic sweep rate,@xmath35 .
@xmath36 in previous works @xcite , we studied the influence of external magnetic sweep rate on the critical state stability in crystals of high temperature bi@xmath37sr@xmath37cacu@xmath37o@xmath38 superconductor .
it is extremely difficult to obtain thick crystals of bi@xmath37sr@xmath37cacu@xmath37o@xmath38 , and we have shown @xcite that the thickness of the crystal strongly influences the critical state stability . on the other hand ,
it is relatively easy to obtain large cubic samples of la@xmath0sr@xmath1cuo@xmath2 .
for this reason , we have chosen this system to study the influence of crystal anisotropy on the critical state stability .
a strong influence of cooling condition and a role of crystal anisotropy on the critical state stability in the case of textured ybacuo were considered in ref .
the effects of thermal insulation on the critical state stability in conventional nbti superconductor were investigated in ref .
@xcite .
we paid special attention to the dynamics of the flux jumps , which remains poorly understood up to date .
this dynamics is also important from a viewpoint of superconductor applications , because a large power of the heat released during short avalanches may , at certain conditions , lead to the destruction of superconducting devices . in the present work , we also studied the dynamics of the magnetic flux changes in a single crystal of la@xmath0sr@xmath1cuo@xmath2 during thermomagnetic avalanches . in particular
, we focused on the influence of external magnetic field on dynamics of the following stages of the thermomagnetic avalanche .
in our investigations , we used a large 3 x 3 x 3 @xmath39 single crystal of la@xmath0sr@xmath1cuo@xmath2 , obtained by the floating zone technique . in order to study the influence of the external magnetic sweep rate on the critical state stability
, we used the quantum design ppms system with the extraction magnetometer option and the maximal external magnetic field attainable of 9 tesla .
after cooling the sample in zero magnetic field , external magnetic field was swept with constant rate from 0 to 9 t. in the following experiments , we varied the sweep rate from the minimal ( @xmath40 t / s ) to the maximal ( @xmath41 t / s ) one attainable in our system .
the distance between the following data points taken in each magnetization curve was limited by the sweep rate and the time needed to take each data point . at the maximal sweep rate ( @xmath41 t / s ) , we were not able to obtain the distance smaller than about 0.1 t. this distance has limited an accuracy of @xmath15 , determined in our experiments . in order to determine the current - voltage characteristics of our sample , we performed studies of magnetic moment relaxation .
after cooling in zero external magnetic field , external magnetic field was first swept , with minimal sweep rate , up to 0.5 t , and next the relaxation of magnetic moment was registered at constant external magnetic field during approximately 1 hour .
the studies of @xmath15 , as well as of magnetic relaxation , were performed both for external magnetic field parallel to the @xmath42-axis and for external magnetic field parallel to the @xmath43-plane .
the dynamics of the flux jumps was studied in the cryogenics 12 tesla magnet system with variable temperature insert for external magnetic field parallel to the @xmath42-axis . a pick - up coil consisted of 6 turns of copper wire wound around the investigated sample and connected to data acquisition board ( daq ) in the computer .
because of small inductance of the applied coil and large internal resistance of daq , we were able to register the changes of the voltage with time resolution better that 10@xmath44 s. the time resolution of daq was about 10@xmath45 s. we registered time dependence of the coil voltage during the following flux jumps .
additionally , we used two miniature hall probes to register magnetic field dependence of the surface self - field , @xmath46 , of the investigated sample .
one of the probes was put in the centre of the surface of the investigated sample . the second probe measured external magnetic field .
the differential signal from the two probes was proportional to @xmath46 .
the cryogenic hall probes were made from tin doped insb films .
the probe was supplied with the current of 10 ma .
the sensitivity of the hall probe was about 5 mv / t .
the rate of external magnetic field sweep was approximately 1 t / min .
in the 12 tesla magnet system , we also performed magnetization measurements using vibrating sample magnetometer ( vsm ) .
it is important to note that the heat exchange conditions in both systems used in our experiments were very different . in the ppms system ,
the sample was put into a standard ppms holder , which was surrounded by helium gas with low pressure ( about 0.5 torr ) . in the 12 tesla magnet system , during the experiments at 2 k and 4.2 k , the sample was immersed in liquid helium . during experiments at higher temperatures ,
the sample was surrounded by flowing helium gas with approximately atmospheric pressure .
the sample was either glued to a quartz plate ( studies of the flux jumps dynamics ) or put into teflon holder ( vsm ) .
t / s at different temperatures for both sample orientations studied in our experiments . ] , with the sweep rate at different temperatures and for different sample orientations with respect to the external magnetic field , in a double - logarithmic scale , for the 3 x 3 x 3 @xmath39 la@xmath0sr@xmath1cuo@xmath2 single crystal surrounded by helium gas ( ppms experiments ) .
dashed lines show the predictions of eq .
3 , where @xmath15 is inversely proportional to the square root of the sweep rate . ]
the range of flux jumps occurrence , the field of the first flux jump as well as the distance between the following flux jumps depends on the sample orientation , temperature , sweep rate and the experimental system .
figure 1 presents some of the virgin magnetization curves taken by extraction magnetometer ( ppms ) with the sweep rate of @xmath41 t / s at different temperatures for the both sample orientations studied in our experiments .
magnetization curves on the left side in fig . 1 are cut - off for the signal amplitude above 20 emu , because of a limitation of the signal in extraction magnetometer .
nevertheless , in all our ppms experiments , we were able to determine the value of @xmath15 .
temperature range of the flux jumps occurence depends on the sample orientation .
for the external magnetic field parallel to the @xmath42-axis , flux jumps occur even at 9.5 k ( see fig .
for the external magnetic field parallel to the @xmath43-palne , flux jumps disappear already at 5 k ( see fig .
1f ) . with the increase of temperature
, we observed an increase of @xmath15 and of the distance between the following flux jumps .
we also observed a decrease of @xmath15 with an increase of external magnetic field sweep rate ( see .
2 ) hence , we observed the flux jumps in the largest temperature range , for the highest sweep rate ( @xmath41 t / s ) . at the lowest sweep rate , @xmath40 t / s
, we did not observe any flux jumps at the temperatures higher than 4 k , for the external magnetic field parallel to the @xmath42-axis , and higher than 2.2 k , for the external magnetic field parallel to the @xmath43-plane , respectively .
figure 2 shows the changes of @xmath15 with the sweep rate at different temperatures and for different sample orientations - the data were taken from ppms experiments .
the results are shown in a double - logarithmic scale .
if @xmath15 changes with the sweep rate according to the power law , we expect a straight line in such plot .
dashed lines show the predictions of eq .
3 , where @xmath15 is inversely proportional to the square root of the sweep rate . the experiments in the 12 tesla magnet ( with hall probes or vsm )
were all performed with the external magnetic field sweep rate approximately the same as the highest sweep rate in the ppms system .
however , we observed here the flux jumps only at the lowest temperature of about 2.2 k and for the external magnetic field parallel to the @xmath42-axis ( see fig .
3a - experiment with the hall probes ) .
no jumps were present neither at 4.2 k nor at higher temperatures . , measured by the hall probes in 12 tesla magnet system at 2.2 k and 4.2 k - sample immersed in liquid helium .
b , c ) time dependence of the voltage , per one turn of the coil , registered during the following flux jumps .
the numbers correspond to the numbers denoting successive flux jumps presented in part ( a ) .
initial voltages of the following flux jumps were shifted , in order to present all data in one figure . @xmath47 and @xmath48 describe duration of the two characteristic stages of the flux jump . ]
the relaxation of the magnetic moment , measured in external magnetic field @xmath49 t , was approximately logarithmic in time , which was consistent with the model of thermally activated motion of bundles of vortices @xcite . in figure 3a
, on can see the hysteresis loops of the surface self - field , @xmath46 , measured by the hall probes in 12 tesla magnet system at the temperature of about 2.2 k ( solid line ) and 4.2 k ( dotted line ) , for the external magnetic field parallel to the @xmath42-axis . at 2.2 k , one can see giant jumps of the surface self - field caused by the thermomagnetic avalanches .
no jumps were present at the temperature of 4.2 k or at higher temperatures . in figure 3b and 3c ,
one can see time dependence of the voltage per one turn of the coil wounded around the investigated sample , registered during the succeeding flux jumps .
the numbers at the curves correspond to the numbers denoting successive flux jumps presented in fig .
typical duration of the flux jumps was of several microseconds .
the dependence shown in fig .
3b corresponds to the lower ( ascending ) branch of the hysteresis loop .
the dependence shown in fig .
3c corresponds to the upper ( descending ) branch of the hysteresis loop , shown in fig .
before and after each flux jump , the voltage of the coil was close to zero .
however , in order to present all the data in one figure , the initial voltages of the following flux jumps were shifted .
the voltage of the coil is proportional to the time derivative of the magnetic flux in the investigated sample .
we can recognize two characteristic stages of the flux jump . during the first stage , the rate of the magnetic flux changes increases . during the second stage , the rate of the magnetic flux changes decreases .
we have denoted duration of these two stages as @xmath47 ( `` before maximum '' ) and @xmath48 ( `` after maximum '' ) , respectively ( see fig .
we have determined duration of the following stages of the flux jump at a noise level , which in our system was about 30 mv .
one can see that the voltage of the observed impulses for the ascending branch of the hysteresis loop ( fig .
3b ) has the sign opposite to the descending branch ( fig .
3c ) . this is because , for the ascending branch , the magnetic flux enters into the investigated sample during the succeeding flux jumps , while for the descending branch , we observe flux exit .
in order to compare the models of the critical state stability with our experimental results , it is necessary to know the current - voltage characteristics of the investigated sample @xcite .
these characteristics can be determined using the data of the magnetic moment relaxation . in order to obtain these characteristics
, we applied the following procedure .
we used the model of an infinite slab sample .
additionally , we assumed the screening current density to be the same in the whole sample volume penetrated by the magnetic flux . the time dependent screening current density can be in this case correlated with the magnetization
the electric field at the surface of the slab is proportional to the time derivative of the screening current density . treating the time as a parameter
, one can obtain the current - voltage dependence , @xmath50 .
the obtained by the described above procedure current - voltage characteristics , were fitted according to eq .
we applied this procedure for both sample orientations studied in our experiments .
our fitting procedure gave the critical current density of an order of @xmath51 a/@xmath52 and @xmath53 a/@xmath52 and the @xmath30-parameter ( see eq .
2 ) of about 40 and 30 for the external magnetic field parallel to the @xmath43-plane and parallel to the @xmath42-axis , respectively . in the range of the flux jumps occurrence ,
the @xmath30-parameter was only weakly dependent on temperature .
hence , we did not consider this dependence in our further analysis .
the temperature dependence of the critical current density was estimated using the widths of the magnetization hysteresis loops in the external magnetic field of 1 t , which was close to the field of the first flux jump , and the formulas from ref .
we used here the loops taken by vsm in 12 tesla magnet system . in vsm system , we did not observe any flux jumps neither at 4.2 k nor at higher temperatures , and there was no limitation of the signal above 20 emu like in ppms extraction magnetometer .
limitation of the signal as well as a large number of the flux jumps observed in ppms experiments made determination of the critical current density ( using ppms data ) impossible .
the experimentally obtained temperature dependence of the critical current density was fitted according to the exponential formula : @xmath54 , where @xmath55 , @xmath56 and @xmath5735.5 k. we have found the following fitting parameters . for the external magnetic field parallel to the @xmath42-axis , @xmath58 a/@xmath52 and @xmath59 = 7.4 k ( the critical current density in the direction parallel to the @xmath43-plane ) . for the external magnetic field parallel to the @xmath43-plane ,
@xmath60 a/@xmath52 and @xmath59 = 4.3 k ( the critical current density in the direction perpendicular to the @xmath43-plane ) .
our fitting procedure was performed only in the range 4 - 15 k , which was most interesting from the viewpoint of the flux jumps occurrence . in order to determine the conditions of the critical state stability and the thermomagnetic avalanche development , it is necessary to find the relation between the thermal and magnetic diffusivity or the parameter @xmath62 . in the case of the flux
creep model : @xmath63 . in the case of an infinite slab model , the electric field at the surface of the slab : @xmath64 for @xmath65 and @xmath66 for @xmath67 . the maximal value of the electric field which can be , at given conditions , induced in superconducting sample @xmath68 . using the maximal value of the electric field we can calculate the minimal value of @xmath61 , @xmath69 . in our analysis , we used the temperature dependence of the thermal parameters @xmath6 and @xmath7 from ref .
@xcite and ref .
@xcite , respectively . in all our calculations
, we used the value of in - plane thermal conductivity , which is approximately one order of magnitude higher than out - of - plane thermal conductivity @xcite . as a result ,
the ability of investigated sample to remove thermal fluctuations is limited by the in - plane thermal conductivity .
we have found that in the case of all our experiments @xmath70 .
hence , we can assume the dynamic conditions of the thermomagnetic avalanche development .
in dynamic conditions , we expect the critical state stability to be dependent on the cooling conditions .
this is consistent with the results of our experiments . in order to analyze this problem
, we calculated the biot number , @xmath71 .
the biot number depends on the thermal boundary conductivity , @xmath27 .
this parameter is very difficult to evaluate experimentally @xcite . for the sample immersed in liquid helium at normal pressure and in the absence of the boiling crisis , ( 12 tesla magnet experiments ) , we expect @xmath27 to be of an order of @xmath72 w/@xmath52k @xcite . in the case of ppms experiments
, we expect this parameter to be of several orders of magnitude lower .
the criteria of the critical state stability can be easily derived in two limiting cases .
1 ) @xmath73 - weakly cooled sample ( this case was considered in ref .
@xcite ) , 2 ) @xmath74 - so - called isothermal conditions .
we have found that the approximation of the weakly cooled sample can not be applied in the case of our sample immersed in liquid helium ( in this case @xmath75 10 ) .
however , it seems to be reasonable to use this approximation in the case of ppms experiments , where we expect @xmath34 to be several orders of magnitude lower .
we assumed the dynamic model for the weakly cooled sample ( @xmath73 ) to describe the critical state stability in the case of our experiments performed in ppms .
hence , we used eq . 3 , and our experimental data , to calculate the thermal boundary conductivity , @xmath27 .
if the boundary conductivity is governed by phonon processes , we expect this dependence to be described by a power function with the power of about 3 @xcite .
hence , we assumed @xmath76 .
for the highest sweep rate @xmath41 t / s , we have found , from the fit to experimental data , @xmath77 w/@xmath52k@xmath78 or @xmath79 w/@xmath52k@xmath78 for the external magnetic field parallel to the @xmath42-axis or to the @xmath43-plane , respectively . at 4.2 k the calculated , according to our procedure , thermal boundary conductivity was 2 - 3 orders of magnitude lower than that for the sample immersed in liquid helium at normal pressure .
thus our assumption about a low value of the biot number seems to be correct . however , it is necessary to notice that eq .
3 does not describe precisely the experimentally observed dependence of @xmath15 on the sweep rate ( see fig . 2 ) .
if @xmath15 changes with the sweep rate according to eq . 3 , we expect the dashed lines in fig . 2 to be parallel to the experimental curves .
we observed the larges disagreement for the external magnetic field parallel to the @xmath43-plane . in order to derive the critical state stability criterion in the isothermal conditions ( @xmath74 )
, we used the following approximations . let us assume the infinite slab to be penetrated by the magnetic flux to the depth @xmath80 . to simplify further calculations
, we assume the space coordinate @xmath81 at the sample border . according to the arguments presented in ref .
@xcite in dynamic conditions for an infinite slab model , we can correlate the fluctuation of the electric field @xmath82 with the fluctuation of the temperature @xmath83 according to the formula : @xmath84 the background electric field , @xmath85 , is induced by the sweep of the external magnetic field : @xmath86 . in the case of weakly cooled sample ( @xmath73 ) , one can assume the thermal fluctuation to be independent on the space coordinate .
the stability criterion ( eq .
3 ) can be in this case derived from the inequality @xmath87 , where the left side of the inequality is the heat generated by the fluctuation in the sample volume and the right side is the heat removed from the sample through its border @xcite . in isothermal conditions ( @xmath74 ) ,
because of strong cooling of the sample , we can assume the fluctuation of the temperature at the sample border to be equal to zero , @xmath88 . in our approximation , we assumed the space distribution of the thermal fluctuation to be described by a square function @xmath89 where @xmath90 is the maximal temperature fluctuation at the penetration depth .
the maximal heat power , generated in the sample by the fluctuation , which can be removed from the sample volume , is in this case limited by the thermal conductivity , and the critical state stability criterion can be described by the inequality @xmath91 .
hence , @xmath92 .
the left side of this inequality has a maximum for @xmath93 . as a result
, the critical state stability criterion is given by the condition @xmath94{\frac{\gamma
\mu_0 \kappa}{n \left| \frac{\partial j_\mathrm{c}}{\partial t } \right| \frac{\partial h_\mathrm{ext}}{\partial t}}}$ ] , or @xmath95{\frac{\gamma \mu_0 \kappa}{n \left| \frac{\partial j_\mathrm{c}}{\partial t } \right| \frac{\partial h_\mathrm{ext}}{\partial t } } } \label{eq : eleven}\end{aligned}\ ] ] where @xmath96 .
it is important to note that , also in the case of isothermal approximation , we expect the field of the first flux jump to decrease with an increase of the sweep rate , @xmath97 . at the sweep rate of @xmath41
t / s and in the external magnetic field parallel to the @xmath42-axis ( a ) and parallel to the @xmath43-plane ( b ) of the 3 x 3 x 3 @xmath39 single crystal of la@xmath0sr@xmath1cuo@xmath2 . the predictions of the approximated models of the critical state stability - see eqs . 1 , 3 and 5 are compared with the experimental data .
dashed line shows temperature dependence of the field of full penetration , @xmath98 . ] in figure 4 , we present the comparison of the predictions of the three theoretical approximations discussed in the present work , i.e. the local adiabatic approximation and two dynamic approximations ( of weakly cooled sample and isothermal ) with the experimentally obtained temperature dependence of @xmath15 .
the experimental data were taken for both sample orientations studied in our experiments : in the external magnetic field parallel to the @xmath42-axis ( fig .
4a ) or parallel to the @xmath43-plane ( fig .
4b ) , taken at the sweep rate of @xmath41 t / s . for the external magnetic field
parallel to the @xmath42-axis , we present the results both in the case when the sample is cooled by gaseous he ( the ppms experiment ) and in the case when the sample is cooled by liquid he ( the experiment in 12 tesla magnet with the hall probe - see fig .
3a , in this case @xmath99(2.2 k ) = 3.5 t ) . in the both figures , we also present temperature dependence of the field of full penetration : @xmath100 .
this parameter is very important , because in the infinite slab model of the critical state stability , we expect the flux jumps to vanish when @xmath15 exceeds @xmath21 . for the model of weak cooling , we assumed @xmath76 , where @xmath77 w/@xmath52k@xmath78 or @xmath79 w/@xmath52k@xmath78 for the external magnetic field parallel to the @xmath42-axis or to the @xmath43-plane , respectively .
it is clearly seen that our experimental results can not be explained by the local adiabatic approximation .
the best fit of the ppms results was found in the dynamic model of weak cooling , while the results obtained with the 12 tesla magnet are better described by the isothermal approximation .
it is worth to note that the isothermal curve in fig .
4a crosses the @xmath98 curve at the temperature of about 3 k. this result can explain the fact that , in the case of 12 tesla magnet experiment , we observed the flux jumps only at the temperature of 2 k and not at temperatures higher than 4.2 k. on the other hand , in the case of ppms experiments , we observed the flux jumps also above the @xmath98 curve .
it is worth to note that also in other experiments , one sometimes observes flux jumps in the external magnetic field higher than the field of full penetration @xcite .
such behavior of the flux jumps is not fully clarified yet . and
@xmath48 , before and after maximum of the coil voltage .
b ) the magnetic field dependence of the magnetic diffusivity , estimated form the exponential fit of the coil voltage at the begin and at the end of the flux jumps , respectively . ]
duration of the two characteristic stages of the flux jumps denoted in fig . 3b
as @xmath47 and @xmath48 decreases with an increase of the external magnetic field , and @xmath47 was approximately three times shorter than @xmath48 ( see fig .
the exponential character of the final stage of the flux jump suggests that it can be analyzed in terms of the magnetic diffusion .
the process of the magnetic diffusion depends on the initial magnetic field distribution in the sample .
it is possible to decompose this distribution into a series of functions characteristic for the sample geometry ( e.g. a series of cosines in the case of infinite slab geometry or a series of bessel functions in the case of a sample with cylindrical symmetry ) .
each harmonic of this decomposition has specific time dependence . from the sum of the time dependent harmonics ,
one can obtain the time dependent distribution of the magnetic flux in the investigated sample .
the voltage of the coil , wound up around the investigated sample , is proportional to the time derivative of the magnetic flux .
we assumed the sample to be of a shape of infinite slab . for sufficiently long times in comparison
to the parameter @xmath101 we can assume the voltage of the coil - @xmath102 .
the parameter @xmath103 is a characteristic length of the diffusion process . if the slab is fully penetrated by magnetic flux , @xmath104 .
hence , when one fits the time dependence of the coil voltage at the final stage of the flux jump , one can find the coefficient @xmath4 .
the magnetic field dependence of this parameter is shown in fig .
we have found an increase of this parameter with increasing external magnetic field .
the initial exponential increase of the voltage also depends on the magnetic diffusivity of the superconducting sample @xcite .
we have fitted this dependence by an exponential function @xmath105 ) . however , it is necessary to discuss the meaning of the length parameter ( the parameter @xmath103 in eq .
6 ) at the begin and at the end of the flux jump . during the flux jumps observed in our experiment ,
the sample was wholly penetrated by the magnetic flux ( see fig .
hence , as a characteristic length of the diffusion process ( in eq .
6 ) , at the end of the flux jump , we assumed the parameter @xmath23 , it means half of the sample width . on the other hand , at the begin of the flux jump , only a part of the sample is penetrated by the magnetic flux .
hence at this stage , one should assume the characteristic length parameter to be equal to the penetration depth , which is smaller than @xmath23 .
the magnetic field dependence of the parameter @xmath4 for the begin of the flux jump , we have found to be similar as that for the end of the jump .
these two parameters coincide very well , if we assume in eq . 6 @xmath106 ( see fig .
5b ) for the begin of the flux jump . the characteristic diffusivity found in our experiment was of an order of 1 @xmath52/s ( see fig .
such an order of diffusivity is characteristic for the magnetic diffusion in our system , providing the sample is in the normal state or in the flux flow regime .
the resistivity of la@xmath0sr@xmath1cuo@xmath2 is anisotropic .
however , we studied dynamics of the flux jumps only for the external magnetic field parallel to the @xmath42-axis of the investigated crystal . in such sample orientation , screening currents flow only in the @xmath43-plane .
if we assume the in - plane normal state resistivity @xmath107 @xcite , then @xmath108 . on the other hand , if one takes data of the thermal conductivity ( @xmath6 ) @xcite and of the specific heat ( @xmath7 ) @xcite , one can estimate the in - plane thermal diffusivity ( @xmath109 ) to be of an order of @xmath110 .
the out - of - plane thermal diffusivity is approximately one order of magnitude lower than the in - plane thermal diffusivity @xcite .
the critical state stability in a single crystalline la@xmath0sr@xmath1cuo@xmath2 is influenced by a large number of parameters .
all of them must be considered in order to determine the critical state stability in the framework of existing theoretical models . in la@xmath0sr@xmath1cuo@xmath2 crystal ,
most of these parameters are anisotropic .
hence , the stability of the critical state depends strongly on the sample orientation with respect to the external magnetic field . in our analysis
, we consider both anisotropy of critical current density and the anisotropy of the current - voltage characteristic ( the anisotropy of the @xmath30-parameter ) . in the present analysis
, we assumed the thermal conductivity of the investigated sample to be limited only by the in - plane thermal conductivity , which is approximately one order of magnitude higher that the out - of - plane thermal conductivity . according to our results in the case of la@xmath0sr@xmath1cuo@xmath2 crystal ,
the anisotropy of the @xmath30-parameter is relatively weak .
hence , the anisotropy of the critical state stability is connected mainly with the anisotropy of screening currents .
the difference between the in - plane and out - of - plane screening current , at given temperature , is about one order of magnitude . in both sample orientations
studied in our experiments , the thermomagnetic avalanche develops in dynamic conditions .
our sample is characterized by a strong flux creep phenomenon .
hence , the critical state stability is influenced both by the external magnetic field sweep rate and the cooling conditions .
these cooling conditions were very different in two cryostats used in our experiments . in ppms experiments , where the sample is surrounded by helium gas of low pressure
, the critical state stability can be analyzed in the approximation of weak cooling or low biot number . in such condition ,
the thermal parameter that influences the critical state stability is the thermal boundary conductivity .
this parameter is difficult to determine experimentally .
it depends e.g. on the roughness of the surface of the investigated sample . in the case of ppms experiments ,
the investigated sample is put additionally into a holder which makes an analysis of the heat exchange conditions even more difficult .
nevertheless , it seems to be slightly surprising that , according to our analysis , the thermal boundary conductivity changes by an order of magnitude after the change of the sample orientation .
it is possible that such results are , to some extent , connected with the approximations assumed by the models used in our analysis . in our analysis
we used models of an isotropic infinite slab sample . in order to determine ( in the framework of these models ) the critical state stability criterion ,
it is necessary to know a large number of parameters .
each of these parameters must be determined experimentally and all parameters have some experimental errors . as the result
, the uncertainty of the obtained critical state stability criterion can be relatively large .
the models used in our analysis can be applied only for some limiting cases ( @xmath73 or @xmath74 ) . in order to determine the critical state stability in an intermediate case , numerical calculations are necessary .
such calculations were performed e.g. in ref .
@xcite . in order to determine the current - voltage characteristics , we study the relaxation of the magnetic moment .
although eq .
2 can be applied to describe the the current - voltage characteristics in a wide range of the electric field , one must bear in mind that the electric field induced in the superconducting sample during magnetic relaxation experiments is significantly lower than the electric field induced by the external magnetic field sweep .
one should also bear in mind that the accuracy of determination of @xmath15 for the highest sweep rate ( @xmath41 t ) in our experiment was only about @xmath111 t. this experimental error has a special significance for the sample orientation with the external magnetic field parallel to the @xmath43-plane , because for this orientation all observed values of @xmath15 were in the range between 0.5 t and 1 t ( see fig .
2 ) . finally ,
more accurate analysis of the critical state stability in an anisotropic material should also take into account the value of the out - of - plane thermal conductivity . our experimental results , as well as the data for textured ybacuo @xcite ,
show that by the improvement of the cooling conditions , i.e. by immersing the sample in liquid helium , the critical state stability increases significantly .
both dynamic models used in our analysis predict the dependence of @xmath15 on the external magnetic sweep rate ( see eq . 3 and eq . 5 ) .
these equations were derived assuming the critical current density to be independent on the magnetic field and predict this dependence to be described by a power function with the power of @xmath112 and @xmath113 for the approximation of weak cooling and the isothermal approximation , respectively . however ,
if the critical current density is strongly dependent on the magnetic field , the dependence of @xmath15 on the sweep rate can be modified @xcite .
the comparison of the predictions of eq . 3 with the experimental results
, one can see in fig . 2 and in fig .
one can observe the largest inconsistency for the lowest temperatures and for the external magnetic field parallel to the @xmath43-plane . the influence of the heat exchange conditions on critical state stability as well as a strong dependence of @xmath15 on the sweep rate show that thermomagnetic instabilities are initiated in dynamic conditions , in which the magnetic diffusivity is smaller than the thermal one .
on the other hand , the effective diffusivity describing the dynamics of the flux jumps ( see fig .
5b ) is very high , it means comparable to the magnetic diffusivity in the normal state .
the magnetic field dependence of this effective diffusivity suggests that it can be correlated with the flux flow resistivity . in order to understand this discrepancy
, one should explain that the initial stage of the avalanche develops at the voltage level that is induced by the external magnetic field sweep .
we estimate that the sweep rates used in our experiments induced in the coil the voltage of an order of @xmath114 v , which is six or four orders of the magnitude lower than the noise level registered by the acquisition board . for this reason
, we were not able to register this very initial stage of the avalanche .
our experimental results show that after this initial stage of the avalanche the resistivity of the sample rapidly increases and the dynamics of the flux jumps is governed by the magnetic diffusivity , which depends of the flux flow resistivity .
the stability of the critical state in a large crystal of la@xmath0sr@xmath1cuo@xmath2 is influenced by its anisotropy , and it depends on the sample orientation with respect to the external magnetic field .
the most relevant is the anisotropy of the critical current . at relatively slow magnetic field sweep rates ,
thermomagnetic avalanches are initiated in the dynamic conditions .
hence , one observes a strong influence of the heat exchange conditions as well as of the sweep rate on the critical state stability . in the case of the experiments , where the investigated sample is surrounded by helium gas with low pressure ( ppms ) ,
the critical state stability criterion can be derived in the approximation of a weakly cooled sample . however , this approximation can not be applied in the case of large crystals immersed in liquid helium .
flux jumps in the single crystalline la@xmath0sr@xmath1cuo@xmath2 are very short , typical duration is of several microseconds .
the dynamics of the jumps is governed by the magnetic diffusivity , proportional to the flux flow resistivity .
this work was partly supported by polish national science centre under research project for years 2011 - 2012 ( grant n n202 1663 40 ) . | we studied the critical state stability in a large cubic sample of a single crystalline la@xmath0sr@xmath1cuo@xmath2 for different sample orientations with respect to the external magnetic field as well as the dynamics of the flux jumps .
it is shown that thermomagnetic avalanches develop in dynamic conditions characterized by significantly lower magnetic diffusivity than the thermal one . in this case
, critical state stability depends strongly on cooling conditions .
we compared predictions of the isothermal model and of the model for the weakly cooled sample with experimental results . in both models ,
the field of the first flux jump decreases with an increase of sweep rate of the external magnetic field .
we also investigated the influence of external magnetic field on the dynamics of the following stages of the thermomagnetic avalanche .
it is shown that the dynamics of the flux jumps is correlated with the magnetic diffusivity proportional to the flux flow resistivity . |
understanding the new physics behind small neutrino masses is a major current focus of particle theory research .
the simplest scenarios for this are the various types of seesaw mechanisms , which have also formed the basis for understanding the observed mixing pattern , which is so different from the quark sector .
two of them , type - i @xcite and the inverse seesaw @xcite postulate the existence of heavy sm singlet fermions , generally denoted by the symbol @xmath1 ( @xmath2 is a generation index ) in addition to the sm fermions .
the stability of these patterns can be guaranteed by extending the sm gauge group as we discuss below .
these singlet fermions couple to familiar lepton doublets via yukawa couplings of the form @xmath3 , which after electroweak symmetry breaking leads to the dirac mass of the neutrino , which is expected to be in the gev range . in the type - i seesaw ,
the tiny neutrino mass arises once we introduce an additional majorana mass term @xmath4 for the @xmath5 s .
being sm singlets , these majorana particles can take arbitrarily large mass values , which then guarantees that the neutrino mass given by the formula below is small : @xmath6 note that since the scale of neutrino masses is known to be in the sub - ev range , the scale @xmath7 is correlated with the magnitude of the yukawa couplings @xmath8 .
for instance , if the @xmath5 masses are in the 100 gev range , it implies @xmath9 or so .
however , for specific flavor pattern of both @xmath10 and @xmath8 , tiny neutrino masses can be realized without making @xmath11 necessarily tiny .
examples of such theories are given in @xcite .
on the other hand , in the case of inverse seesaw models , in addition to the set of singlets @xmath1 , one adds another set of singlet fermions @xmath12 which form a dirac mass @xmath7 with the @xmath1 fields , i.e. , @xmath13 .
the model so far has conserved lepton number and leads to zero neutrino mass @xcite .
one then allows the @xmath14 fields to have a majorana mass matrix @xmath15 whose overall scale is in the kev range @xcite .
the neutrino mass formula in this case is given by : @xmath16 this also leads to tiny majorana mass for neutrinos .
the small scale of the @xmath15 matrix can be explained in various ways @xcite and is a window to further new physics . in this model
since the small neutrino masses arise from the small values of @xmath15 matrix elements , the singlet neutrinos @xmath17 form a dirac pair with masses in the 100 gev range and yet have yukawa couplings of order @xmath18 without any fine - tuning .
a key question for neutrino mass physics is how to test the seesaw mechanism @xcite .
the obvious first step would be to determine the couplings @xmath8 and @xmath19 .
the vast literature on neutrino mass models is devoted to precisely this question where additional theoretical assumptions such as symmetries are used for this purpose .
the symmetries restrict the theory to a particular sub - space of the full parameter - space .
if the seesaw scale is high , as is the case in most type - i models , this appears to be the only realistic possibility .
we take a different approach in this paper .
since @xmath8 denotes the coupling of the @xmath5-fermions to the higgs boson , if @xmath7 is in the 100 gev range , experimental information on @xmath11 can be obtained from the lhc data on the higgs boson decays @xcite .
our goal in this paper is to focus on this question . in this paper
, we consider primarily the inverse seesaw case and we will comment on the specific type - i scenarios where our results will apply with slight changes .
the main result of the paper i.e. bounds on the the yukawa couplings @xmath8 , is derived from an analysis of the 7-tev lhc data .
the 8-tev data came out after this work was finished .
however , we have redone our analysis with this new data and found that the change in the bounds is very minimal .
we comment on it briefly at the end of the paper .
this paper is organized as follows : in section [ gaugemodel ] , we discuss the minimal gauge models with inverse seesaw ; in section [ generalpheno ] we give an overview of the higgs phenomenology of the seesaw models ; in section [ llnunubound ] we shall illustrate how to constrain the parameter - space in seesaw models using the results of the higgs searches in the final state @xmath20 ; in section [ postdiscovery ] we derive these constraints on the seesaw models by analyzing the rates of all the measured decay modes of the higgs boson ; the concluding remarks are in section [ conclusion ] . in appendix
a , we list the relevant decay widths of the heavy neutrino .
in this section , we briefly discuss two classes of model which can have @xmath18 dirac yukawa couplings and @xmath21 gev mass for the singlet fermions responsible for the seesaw mechanisms .
one class , as noted above , is the inverse seesaw model , where this situation can be naturally realized , and a second class , which belongs to type - i seesaw - based neutrino mass models where with specific texture for the dirac yukawa matrices as well as for the singlet fermion mass matrix , one can have lhc - accessible parameters . if we simply add two heavy sm singlet fermions @xmath22 to the sm , there are many gauge invariant terms in the potential and we will not get the simple inverse seesaw formula .
the simplest extension of sm where the inverse seesaw formula arises naturally is @xmath23 with gauge symmetry under which the singlet fermions @xmath5 transform as @xmath24 and @xmath14 fields are singlets .
the gauge symmetry is broken by a higgs field @xmath25 acquiring vacuum expectation value ( vev ) and the sm - doublet @xmath26 vev then breaks the gauge group down to electromagnetic @xmath27 .
the yukawa lagrangian for the lepton sector is given by : @xmath28 it is clear that once we substitute @xmath29 and @xmath30 , we get the inverse seesaw matrix @xcite @xmath31 with the mass matrix for the neutrinos given by eq .
( 2 ) with @xmath32 ( note that @xmath33 are in general @xmath34 matrices ) . at the renormalizable level , there are no other terms allowed in the yukawa lagrangian by the gauge symmetry and therefore the above inverse seesaw formula for neutrinos is stable under radiative corrections .
our goal is to find constraints on @xmath11 as a function of @xmath7 .
there exists an extensive analysis on the constraints for light singlet fermions @xcite .
although this analysis addresses only the constraints on majorana neutrinos , some of their results apply to our case as well
. however , the limits derived in @xcite for @xmath35 gev or so are very weak
. furthermore constraints from neutrino - less double beta decay @xcite derived on heavy sterile neutrinos do not apply to this case since in our model , the @xmath5 and @xmath14 form a pseudo - dirac pair and lepton number is almost exactly conserved . in order to use the lhc data to explore constraints on @xmath11 and @xmath7 in the 100 gev range
, we will assume that ( i ) @xmath36 and ( ii ) the mass of re(@xmath37 ) is heavy compared to the sm higgs boson so that neither the heavy gauge boson associated with @xmath38-symmetry nor the interactions of re(@xmath37 ) affect the higgs boson decay modes we consider .
it follows from the above lagrangian that if one of the singlet fermions has mass in the 100 gev range , it will affect the higgs branching ratios : for instance if @xmath39 , then this opens up a new mode for sm higgs decay , i.e. , @xmath40 , and the collider signal will arise from @xmath41 mixing diagram in fig . [
fig : hdecayiss ] where @xmath42 .
folding @xmath43 decays , one will get final states with @xmath44 where in the final state both charged leptons and anti - leptons will appear and the existing lhc data on these final states will provide constraints on @xmath11 .
clearly , which charged lepton appears will depend on the flavor structure of @xmath11 and @xmath45 . for @xmath45 we will go to a basis so that it is diagonal , i.e. a linear combination of @xmath46 and @xmath5 are mass eigenstates with @xmath14 field providing the chiral dirac partner .
mediated by the iss couplings.,title="fig:",width=302 ] mediated by the iss couplings.,title="fig:",width=302 ] turning to the type - i case , as noted earlier , in generic models , the dirac yukawa couplings are very small for the seesaw scale in the tev regime .
however , for specific textures for @xmath11 , it is possible to attain singlet fermion mass in the 100 gev range with dirac yukawa @xmath11 s of order @xmath18 while still satisfying the neutrino oscillation data . in this case
the singlet fermions could show up at the lhc .
two examples of this type of texture are @xcite : @xmath47 with the constraint that @xmath48 where @xmath19 are the singlet neutrino masses and the singlet fermion mass matrix is diagonal .
seesaw of this kind leads in the leading order to zero neutrino masses ; small @xmath49 can then generate the neutrino masses and mixings , and this will not affect the yukawa couplings or the masses @xmath19 .
a second example has @xmath8 of the following type @xcite : @xmath50 with right handed neutrino mass matrix of the form @xmath51 with @xmath52 . in the limit of @xmath53
the neutrino mass matrix is a null matrix regardless of the values of @xmath19 and @xmath54 and small neutrino masses arise from choosing @xmath55 s and @xmath56 s small . to the best of our knowledge , we are not aware of any symmetries that will guarantee these textures for the @xmath7 and @xmath11 matrices and it will be interesting to seek symmetry origin for them . therefore again in this case ,
lhc data should put constraints on @xmath54 vs @xmath19 . since in this case
, @xmath11 s are arbitrary , they could be of democratic type which will mean that there are multiple flavor final states .
two important points regarding these models are worth emphasizing . in both models , as the small parameters ( @xmath15 in the case of inverse seesaw , and @xmath56 and @xmath55 in the type - i model discussed above ) go to zero , the neutrinos become massless . as a result , the neutral higgs decays to two modes : @xmath57 . unlike the case where both @xmath46 and @xmath5 are majorana fermions ( as in generic type - i seesaw models )
, the higgs decay rate to @xmath58 final states becomes twice as large .
thus , the discussions of higgs decay in both classes of models are similar except for the flavor richness in case ( b ) compared to case ( a ) . in case ( b ) model ( ii ) , for @xmath59 , all flavors couple with equal strength to a single fermion flavor @xmath1 implying that final state signal becomes a combination of all three flavors .
similar situation also occurs for case ( b ) , model i if @xmath60 .
as discussed above , a new yukawa interaction in the leptonic sector can potentially alter the higgs phenomenology in a dramatic way .
in fact , in the models considered here , the yukawa coupling of the neutrino can be sizable in principle , much larger than the largest yukawa coupling involved in the decay of a light sm higgs boson , i.e. , the one with the bottom quark which is of order @xmath61 .
here we shall describe the observable consequences of the new yukawa coupling . focusing on the inverse seesaw case
, it is apparent from the neutrino mass formula in eq .
( [ massiss ] ) that neutrino mass fits will largely be dictated by the combination of the flavor structure of the @xmath15 matrix and the @xmath11-matrix . in quark - lepton unified theories ,
the natural expectation is that flavor mixings in @xmath11 are `` weak '' .
so we can assume @xmath11 to be a diagonal matrix .
the next question is the relative magnitude of different flavor yukawa couplings . below we will consider the cases when ( a ) @xmath62 , ( b ) @xmath63 , and ( c ) a democratic form where all the yukawa couplings are equal .
these different coupling structures imply distinctive flavor structure of the final states . from the interaction lagrangian ,
( [ lagrangianiss ] ) , it follows that @xmath8 mediates the decays illustrated in fig .
[ fig : hdecayiss ] : @xmath64 where charged and neutral leptons arise in all the flavor combinations allowed by the form of the @xmath11 matrix .
the gauge bosons @xmath65 and @xmath66 arise in the decay of @xmath5 through the @xmath67 mixing , that is @xmath68 .
the gauge bosons in turn can decay into leptons or hadrons , such that the new yukawa coupling contributes to the rates of both fully leptonic and semi - leptonic final states .
the heavy particles @xmath69 , @xmath66 and @xmath65 in the decay chain of eq .
( [ eq : genericdecays ] ) can be either on - shell or off - shell depending on the mass of @xmath69 .
in particular , when @xmath70 the higgs decay width in eq .
( [ eq : genericdecays ] ) scales as @xmath71 while for @xmath72 , it scales as @xmath73 .
the fully leptonic decay mode of the higgs @xmath20 is among the most sensitive search channels of the higgs and therefore is a very sensitive probe of the @xmath11-coupling , especially for @xmath74 not far above or below @xmath75 .
this allows us to derive a bound on the @xmath76 parameter space as described in detail in section [ llnunubound ] . for definiteness in what follows
, we shall focus on the case of the iss model with only one light flavor of heavy neutrino which we choose to be the electron flavor , @xmath77 , and the corresponding yukawa coupling @xmath78 to the higgs doublet .
we generically denote the mass of the heavy neutrino as @xmath79 , suppressing the explicit flavor index @xmath80 . in this scenario the decay chain of eq .
( [ eq : genericdecays ] ) becomes @xmath81 the decay of the @xmath65 boson to leptons produces pairs of charged leptons of all flavors , and hence , it contributes to the same flavor final states , namely @xmath82 the decay of the @xmath66 boson produces a single charged lepton . in this case , the @xmath66 is produced by a mixing effect of the heavy @xmath77 , and therefore , the @xmath66 is always produced in association with an electron or a positron .
hence , the decay process mediated by the @xmath66 only contributes to electron final states , namely @xmath83 altogether , due to the assumed coupling structure @xmath62 , the final state leptons of the higgs decays mediated by the iss are both of opposite flavor and of same flavor , with same flavor @xmath84 getting its only contribution from processes with @xmath65-boson intermediate state . for the other coupling patterns the resulting flavor structures are straightforward modifications of the one considered here . for the pattern ( b )
where @xmath63 , the main difference is that now it is the same flavor @xmath80 final state that gets its only contribution from the @xmath65-intermediate state .
the opposite flavor final states again get contribution from both @xmath65- and @xmath66-mediated processes , though with the kinematics of @xmath80 and @xmath84 exchanged .
as in our analysis described in the next sections there is going to be little difference between @xmath80 and @xmath84 , the bounds derived for the iss with light @xmath77 also apply for the case of light @xmath85 .
so far we considered a basis in which the @xmath45 matrix is diagonal .
there are however possible situations where one can derive bounds on yukawa couplings .
for instance , it is possible to have structure of the @xmath45 and @xmath11 matrices where democratic yukawa couplings can emerge i.e. by assuming an appropriate texture for the @xmath86 dirac mass matrix we can have the lightest heavy neutrino @xmath5 as a linear combination of all the three flavor states with equal probability , i.e. , @xmath87 . in this case , if @xmath11 is a unit matrix , then we expect all the possible flavor combinations to be equally populated as both the interactions of @xmath88 and the gauge bosons are flavor - universal . the analyses of the lhc experiments are mostly sensitive to @xmath80 and @xmath84 flavors only ,
therefore for such a democratic case , only 4 out the 9 possible final state flavor combinations are easily detectable may be used to recover some sensitivity .
however , they are only a fraction of the @xmath89 decay modes , and there are detection inefficiencies . as such , the gain might not be huge . at any rate ,
the use of the @xmath89 s requires careful experimental considerations that go beyond our competence ; hence we prefer not to consider them . ] .
furthermore the effect of the mixing of flavors in the lightest mass eigenstate @xmath5 in general reduces the rate of @xmath80 and @xmath84 leptons . because of these effects the bounds
should be recalculated carefully to account for the different flavor structure of the signal .
however , neglecting the sub - dominant processes mediated by the @xmath65 , one can obtain an estimate for the bound in the democratic case .
indeed , the rates of the processes mediated by the @xmath66 are just rescaled by mixing and multiplicity factors compared to those of the iss model with only a light @xmath77 . as such
, the bounds of the iss case should be relaxed by a constant factor . in particular , from the fact that the rate of eq .
( [ eq : issdecays ] ) scales as @xmath90 when @xmath91 and as @xmath92 when @xmath93 , we find that the bound gets relaxed by a factor @xmath94 or @xmath95 in the two cases . in the following , we shall illustrate in detail how to derive a bound on the yukawa coupling in the iss model . using the same procedure
, we also computed the bound for the type - i democratic case and we checked that this simple rescaling argument describes the actual bound to a very good accuracy .
hence , we give only the bounds for the iss case , and those for the democratic type - i case can be obtained by the simple rescaling described above . of course a more targeted analysis from the lhc experiment with flavor structure explicitly taken into account
could distinguish between these coupling structures .
for instance , an imbalance between the @xmath80 and @xmath84 same flavor final states can distinguish case ( a ) and case ( b ) , while case ( c ) would be favored if all the flavor combination appear to be equally populated .
irrespective of the flavor structure of the yukawa interactions , the new yukawa coupling mediates new decay modes of the higgs , which contribute to its total width , and hence , in these seesaw models , the higgs width is larger than in the sm : @xmath96 the seesaw contribution to the the total width is necessarily model - dependent . for the case of the iss with dominant coupling @xmath78 , the width and @xmath97 . ]
is given by @xmath98 where we have normalized the yukawa couplings such that the dirac mass terms of all fermions are @xmath99 for @xmath100 .
in the case of the democratic type - i seesaw , as in eq .
( [ eq : democratic ] ) , we can take @xmath101 ; hence the lagrangian essentially contains three interactions of equal strength for the higgs decay , and the higgs decay width is the same as in the iss model once @xmath102 is taken . in section [ postdiscovery ] we shall discuss in detail how to use the information on the measured rates of the several higgs decay final states to put a bound on the seesaw coupling .
the atlas and cms collaboration have both found evidence for a higgs - like particle at around 125 gev . the main evidence for the new particle comes from final states with resonant two photons or four leptons @xcite . in addition to the @xmath103 and @xmath104
searches the lhc and tevatron experiments searched for a sm higgs boson in several other final states , including in the final state @xmath105 @xcite . the experiments presented cut - based as well as multivariate analyses to put a bound on the mass of the sm higgs boson .
the bounds from the multivariate analysis are generically ( slightly ) more stringent than the one obtained from the cut - based analysis .
however the multivariate analysis can not be easily reproduced with our means , and hence , we shall only use the cut - based analysis to derive our bound . here
we shall reinterpret the results of @xcite to extract a bound on extra sources of @xmath105 events . to do this
, we shall repeat the cut - based analysis of @xcite on event samples generated by monte carlo tools
matrix elements computed with @xmath106 @xcite , showered and hadronized with @xmath107 @xcite and detector response parametrized by @xmath108 @xcite .
hadrons have been clustered into jet with the anti - kt algorithm as implemented in @xmath109 @xcite . in @xcite ,
the cms collaboration performed several analysis on the @xmath105 sample collected in the year 2011 with the lhc running at @xmath110 tev center - of - mass energy .
in particular they made a basic selection on the leptons , jets and missing energy of the events depending on the flavor of the final state leptons .
the cuts for the opposite flavor ( of ) and same flavor ( sf ) cases are reported in table [ tab : baseline - selection ] .
then the analysis is specialized for specific values of the sm higgs boson mass and further cuts are devised .
these cuts are collected in the tables [ tab : a120 ] and [ tab : a130 ] for the higgs mass hypothesis of @xmath111 gev and @xmath112 gev , respectively .
these analyses , which we call @xmath113 and @xmath114 , are the most sensitive to new physics connected to the higgs - like particle recently discovered , and hence , these are the only analyses we are going to repeat in order to extract our bound .
@xmath115 @xmath116 , @xmath117 & @xmath118 & @xmath119 @xmath120 , @xmath121 & @xmath121 & @xmath122 @xmath123 & @xmath124 & @xmath125 met @xmath126 20 gev & @xmath127 gev & @xmath128 gev & @xmath129 gev @xmath130 @xmath128 gev & @xmath129 gev to obtain a bound on the iss model we impose that the events yield of the iss after the cuts of both the analyses @xmath113 and @xmath114 is not larger than the total number of events allowed by each of the analysis .
this is done as follows : for each hypothesis for @xmath75 in the iss model we compute the yield of events after the cuts of cms .
this depends on the mass @xmath79 and coupling @xmath78 and it is given by @xmath131 \nonumber\\ \label{eq : niss}\end{aligned}\ ] ] where @xmath132 is the integrated luminosity used in the analysis , @xmath133 is the total higgs production cross - section taken from @xcite , @xmath134 and @xmath135 are flavor indexes @xmath136 , and @xmath137 and @xmath138 are the efficiencies of the cms selections for the decays mediated by the sm decay channel @xmath139 and by decays of the iss , respectively .
as we just want to illustrate here how to obtain an upper bound on the yukawa , we shall use only a few representative values of @xmath79 , namely 60 , 100 , 140 and 200 gev . for lighter @xmath79
the lhc searches tends to be rather ineffective .
in fact , in the iss signal both the charged leptons originate from @xmath140 and therefore the invariant mass of the two leptons can not exceed @xmath79 .
the cuts @xmath141 gev for of ( sf ) leptons , needed to reject leptons from qcd decays , would remove completely the iss contribution for @xmath142 gev . also , for light @xmath79 the bounds from other experiments are more stringent @xcite . the power of the lhc is the sensitivity to @xmath79 around and above the mass of the higgs , which improves significantly as compared to the reach of previous direct bounds .
we remark that our computation of the total events yield neglects the possible interference between the iss and the sm contributions , which in general is a small effect due to different flavor and lorentz structure of the decays .
the selection efficiencies @xmath137 and @xmath138 have been computed with showered events passed through @xmath108 and hadrons clustered into jets using the anti - kt algorithm with jet - cone radius parameter @xmath143 , as done in the cms analysis .
the obtained efficiencies have been rescaled such as to reproduce the sm higgs boson event yield in table 3 of @xcite for @xmath144 and @xmath112 gev for the analysis @xmath113 and @xmath114 respectively .
the results of the used chain of simulation codes are rather realistic , indeed the rescaling factor is almost flat w.r.t the cuts and the differences between the cms and @xmath108 efficiencies are within 20% . for the total width of the sm higgs boson @xmath145 and the partial width @xmath146
we take the reference values of @xcite .
the width @xmath147 due to the decays mediated by @xmath78 and the partial widths @xmath148 have been computed with @xmath106 .
these widths are the source of the dependence of @xmath149 on @xmath78 .
in particular , they scale as @xmath90 when @xmath91 , and as @xmath92 when @xmath93 , i.e. , the heavy neutrino that mediates the decays is off - shell . therefore ,
as @xmath5 becomes heavier than the higgs boson we expect the bound to quickly become less stringent . for completeness , we report our computation of the width @xmath147 and of the partial widths @xmath148 in appendix a. to derive the bound we compute the maximal @xmath78 such that @xmath150 where @xmath151 is the 95% cl limit on the number of events after the selection from the analysis under consideration . from @xcite we extracted @xmath152 and @xmath153 .
we take as the final bound , which we denote as @xmath154 , the most stringent one between the two bounds obtained for the analysis @xmath113 and @xmath114 .
it turns out that the bounds derived using the analysis @xmath113 are stronger for all the cases we have considered .
the obtained bound for a fixed @xmath155 gev as a function of @xmath79 is shown in figure [ fig : bound125 ] .
for the cases where @xmath91 , we exclude @xmath156 while for @xmath93 couplings @xmath157 are excluded . bound on @xmath78 for @xmath155 gev as a function of the mass of the heavy neutrino @xmath77 . ]
evidence of a new particle has been observed in the 2011 and 2012 lhc data @xcite .
the region of phase space where the excesses are concentrated suggests that they are originated by a higgs - like scalar particle with mass 125 gev . assuming that the new observed state is indeed the higgs boson which is also involved in the seesaw , further bounds can be obtained from a global study of the properties of the new particle instead of just using the bound on the rate of new phenomena in the @xmath20 channel . the presence of additional decay modes @xmath158 changes the properties of @xmath159 in several respects .
in fact the total width of @xmath159 is increased w.r.t to the sm value .
deviations from the sm value of the total width are potentially observable in a line - shape analysis ( when @xmath78 is large enough ) or in a global analysis of higgs decay data @xcite .
the change of the total width in turn affects all the branching fractions of @xmath159 . in particular the rates of modes that do not get contributions from the new decay mode ( such as @xmath104 , @xmath160 and @xmath103 )
are suppressed by a factor @xmath161 where @xmath147 is given in eq .
( [ gammaiss ] ) . altogether the iss model , compared to the sm , predicts a suppression by a factor @xmath164 in the observed rates of all channels but @xmath20 and @xmath165 which , for a light higgs boson , are less sensitive due to large backgrounds . ]
which instead is enhanced by the new decays .
the constraint that comes from the increase of the total width is far - reaching .
in fact , it also applies to seesaw models where , for any reason , the mode @xmath105 does not get enhanced .
this might happen , for instance , when the @xmath5 is very light , say @xmath166 , such that the leptons from the @xmath5 decay do not pass the selection cuts in table [ tab : baseline - selection ] . furthermore , the bound coming from the extra contribution to the width applies to other models as the case of iss models where the coupling @xmath167 dominates the decay . in this case
the searches into @xmath105 are much less effective , still the effect on the total width provides a bound on the seesaw coupling .
the constraints from the @xmath105 channel are more specific to each model
. also they depend on the details of each of the @xmath168 analysis .
in fact to make a correct use of the measured best - fit signal strengths of the @xmath168 channels one should consider each @xmath168 analysis separately and compute in detail the efficiencies as we have done for the analysis of @xcite discussed in section [ llnunubound ] . for these reasons we shall consider the bound from two sets of measurements . in both cases we take the iss model as reference . in the first case we put a bound on @xmath78 coming from all the measured signal strengths excluding the @xmath168 channels .
this is the safest possibility as we are making the least number of assumptions on the structure of the seesaw couplings .
additionally we compute the bound using all the available data on the higgs - like particle at 125 gev . to deal with the several @xmath168 analyses we make the simplifying assumption that all the efficiencies for the processes mediated by the seesaw coupling are the same as those computed for the analysis of section [ llnunubound ] . in eq .
( [ eq : niss ] ) . as we find in section [ llnunubound ] that the iss efficiencies are quite similar to those of the sm
, one can expect our simplifying assumption to be reliable for the atlas analysis as well . ] to put a bound we proceed as follows .
for each searched decay mode of the higgs the collider experiments give the best fit value of @xmath169 where @xmath170 runs on the higgs production modes and @xmath171 s are the branching fraction for the various higgs boson decay modes @xmath172 .
these measured best fits are expressed in units of the sm prediction and are referred to as best - fit signal strengths @xmath173 where by @xmath174 we mean the central value and by @xmath175 we mean the symmetrized 1@xmath176 error on the best - fit of the channel @xmath172 given by the experiments .
to place a bound we confront the iss prediction with the latest best - fit signal strengths given by the experiments at the lhc @xcite and the tevatron @xcite . for convenience of the reader
the best - fit signal strengths used in our analysis are reported in table [ strengths ] . .
[ strengths ] signal strength best - fits extracted from @xcite .
ap stands for associated production .
the best - fit for the ap and dijet analyses are extracted from @xcite .
we consistently take the best - fit signal strength for @xmath155 gev throughout . [
cols="^,^ " , ] 99 p. minkowski , phys . lett . *
b67 * , 421 ( 1977 ) ; t. yanagida in _ workshop on unified theories , kek report 79 - 18 _ , p. 95
( 1979 ) ; m. gell - mann , p. ramond and r. slansky , _ supergravity _ , p. 315
, amsterdam : north holland ( 1979 ) ; s. l. glashow , _ 1979 cargese summer institute on quarks and leptons _ , p. 687
, new york : plenum ( 1980 ) ; r. n. mohapatra and g. senjanovic , phys .
44 * , 912 ( 1980 ) .
a. pilaftsis , z. phys .
c * 55 * , 275 ( 1992 ) [ hep - ph/9901206 ] ; a. pilaftsis , phys .
* 95 * , 081602 ( 2005 ) [ hep - ph/0408103 ] ; a. pilaftsis and t. e. j. underwood , phys
. rev . * d72 * , 113001 ( 2005 ) [ hep- ph/0506107 ] ; a. de gouvea , arxiv:0706.1732 [ hep - ph ] ; j. h. chen , x. g. he , j. tandean , and l. h. tsai , phys .
d * 81 * , 113004 ( 2010 ) [ arxiv:1001.5215 [ hep - ph ] ] .
f. bazzocchi , phys .
* d83 * , 093009 ( 2011 ) [ arxiv:1011.6299 [ hep - ph ] ] ; s. c. park , k. wang and t. t. yanagida , phys .
* b685 * , 309 ( 2010 ) [ arxiv:0909.2937 [ hep - ph ] ] ; c. s. fong , r. n. mohapatra and i. sung , phys .
* b704 * , 171 ( 2011 ) [ arxiv:1107.4086 [ hep - ph ] ] .
for other recent works attempting to constrain neutrino mass physics using sm higgs decays at lhc , see l. m. carpenter and d. whiteson , arxiv:1107.2123 [ hep - ph ] ; s. m. davidson and h. e. logan , phys .
rev . * d82 * , 115031 ( 2010 ) [ arxiv:1009.4413 [ hep - ph ] ] . the considerations of these papers
however do not apply to the inverse seesaw and low scale type - i seesaw that we consider in this paper .
p. bamert , c. p. burgess and r. n. mohapatra , nucl .
b438 * , 3 ( 1995 ) [ hep - ph/9408367 ] ; p. benes , a. faessler , f. simkovic and s. kovalenko , phys .
rev . * d71 * , 077901 ( 2005 ) [ hep - ph/0501295 ] ; m. mitra , g. senjanovic and f. vissani , nucl .
* b856 * , 26 ( 2012 ) [ arxiv:1108.0004 [ hep - ph ] ] . the atlas collaboration , atlas - conf-2012 - 092 ( 2012 ) [ https://atlas.web.cern.ch / atlas / groups / physics / confnotes / atlas - conf-201% 2 - 092/[https://atlas.web.cern.ch / atlas / groups / physics / confnotes / atlas - conf-201% 2 - 092/ ] ] . the cms collaboration , cms - pas - hig-12 - 015 ( 2012 ) [ https://twiki.cern.ch/twiki/bin/view/cmspublic/hig12015twiki ] , cms - pas - hig-12 - 008 ( 2012 ) [ https://twiki.cern.ch/twiki/bin/view/cmspublic/hig12008twiki ] . y. zhang , h. an , x. -d .
ji and r. n. mohapatra , phys .
rev . * d78 * , 011302 ( 2008 ) [ arxiv:0804.0268 [ hep - ph ] ] ; m. hirsch , m. malinsky , w. porod , l. reichert and f. staub , jhep * 1202 * , 084 ( 2012 ) [ arxiv:1110.3037 [ hep - ph ] ] .
a. abada , c. biggio , f. bonnet , m. b. gavela and t. hambye , jhep * 0712 * , 061 ( 2007 ) [ arxiv:0707.4058 [ hep - ph ] ] ; m. malinsky , t. ohlsson and h. zhang , phys
. rev . * d79 * , 073009 ( 2009 ) [ arxiv:0903.1961 [ hep - ph ] ] ; m. malinsky , t. ohlsson , z .- z .
xing and h. zhang , phys .
b679 * , 242 ( 2009 ) [ arxiv:0905.2889 [ hep - ph ] ] ; p. s. bhupal dev and r. n. mohapatra , phys .
rev . * d81 * , 013001 ( 2010 ) [ arxiv:0910.3924 [ hep - ph ] ] .
a. ilakovac and a. pilaftsis , nucl . phys . *
b437 * , 491 ( 1995 ) [ hep - ph/9403398 ] ; f. deppisch and j. w. f. valle , phys .
d * 72 * , 036001 ( 2005 ) [ hep - ph/0406040 ] ; f. deppisch , t. s. kosmas and j. w. f. valle , nucl .
b752 * , 80 ( 2006 ) [ hep - ph/0512360 ] ; m. b. gavela , t. hambye , d. hernandez and p. hernandez , jhep * 0909 * , 038 ( 2009 ) [ arxiv:0906.1461 [ hep - ph ] ] ; m. hirsch , t. kernreiter , j. c. romao and a. villanova del moral , arxiv:0910.2435 [ hep - ph ] ; a. ibarra , e. molinaro and s. t. petcov , phys .
rev . * d84
* , 013005 ( 2011 ) [ arxiv:1103.6217 [ hep - ph ] ] ; a. abada , d. das , a. vicente and c. weiland , arxiv:1206.6497 [ hep - ph ] ; m. hirsch , f. staub and a. vicente , phys . rev .
d * 85 * , 113013 ( 2012 ) [ arxiv:1202.1825 [ hep - ph ] ] ; d. n. dinh , a. ibarra , e. molinaro and s. t. petcov , jhep * 1208 * , 125 ( 2012 ) [ arxiv:1205.4671 [ hep - ph ] ] . | we derive bounds on the dirac yukawa couplings of the neutrinos in seesaw models using the recent large hadron collider ( lhc ) data on higgs decays for the case where the standard model ( sm ) singlet heavy leptons needed for the seesaw mechanism have masses in the 100 gev range .
such scenarios with large yukawa couplings are natural in inverse seesaw ( iss ) models since the small neutrino mass owes its origin to a small majorana mass of a new set of singlet fermions .
large yukawas with sub - tev mass right - handed neutrinos are also possible for certain textures in type - i seesaw models , so that the above bounds also apply to them .
we find that the current higgs data from the lhc can put bounds on both electron- and muon - type yukawa couplings of order @xmath0 . |
a magnetic skyrmion is a nanometric magnetic domain wall structure of which the spin configuration is swirling in the planar space and would wrap a unit three - dimensional spherical surface with spins pointing in all directions in the compactification of the planar space @xcite .
recently , experiments have revealed the existence of both nel - type and bloch - type skyrmions in thin films as well as bulk materials @xcite .
the spin configuration of a skyrmion is topologically protected , which is stable as long as it does not overlap tilted spins on the sample edge or shrink to the lattice scale .
another merit of a skyrmion is that it can be displaced by a spin - polarized current in confined structures .
thus , the skyrmion is expected to be a key component of future device applications in the emerging field of skyrmionics @xcite .
recent experiments have demonstrated the creation , manipulation , and elimination of skyrmions in confined geometries using cutting - edge technologies @xcite . however , one significant obstacle to the essential transmission process of skyrmions in devices is the skyrmion hall effect ( skhe ) @xcite , where a skyrmion gains a transverse motion perpendicular to the imposed driving direction . as a consequence
, the skyrmion , which carries encoded information , might be destructed when it passes a narrow racetrack - type channel @xcite and/or when it travels at a high speed , by touching the edge of the device @xcite .
recently we have shown in ref . that a bilayer - skyrmion moves perfectly straight in the direction of the driving current in a bilayer synthetic antiferromagnetic ( saf ) racetrack , where the skhe is completely suppressed .
the two skyrmions in the top and bottom ferromagnetic ( fm ) layers of the bilayer saf racetrack have the opposite skyrmion numbers , and thus the potential transverse shift directions are opposite and canceled @xcite .
we note that the antiferromagnetic ( afm ) skyrmions in monolayer @xcite and bilayer @xcite afm racetracks are also free from the skhe , where the afm skyrmions have the skyrmion number of zero .
another challenging problem is about the effect of thermal fluctuations , which is detrimental to the skyrmion - based device applications .
a skyrmion is largely deformed and might be easily destroyed by the thermal fluctuations .
hence , the thermal effect is one of the important factors to realize the skyrmion - based device applications @xcite . in this paper
, we study the thermal fluctuation effect on the motion of skyrmions in multilayer saf racetracks as well as in monolayer fm racetracks based on numerical simulations and a multilayer thiele equation .
the skyrmions in all fm layers of a multilayer saf racetrack are bound to one single saf multilayer skyrmion by the interlayer afm exchange coupling .
we call it a saf @xmath2-layer skyrmion , where @xmath2 stands for the number of constituent fm layers .
first , we investigate the current - driven dynamics of a saf @xmath2-layer skyrmion at zero temperature ( @xmath3 k ) .
we find the odd - even effect of the number @xmath2 of constituent fm layers on the skhe , which is understood as the total skyrmion number of the skyrmions in each individual fm layer .
namely , the saf @xmath2-layer skyrmion shows the skhe only when @xmath2 is odd .
furthermore , we find that the velocity of the saf @xmath2-layer skyrmion at a given driving current density is inversely proportional to @xmath2 when the driving current is applied only in the bottom fm layer
. we also point out that when the driving current is applied in all fm layers , the velocity of the saf @xmath2-layer skyrmion is basically equal to that of the fm monolayer skyrmion at a given driving current density .
second , we investigate the current - driven dynamics of a saf @xmath2-layer skyrmion at finite temperature ( @xmath4 k ) .
it is shown that the fm monolayer skyrmion with typical material parameters is destroyed when @xmath1 k. on the other hand , the saf bilayer skyrmion is stable and goes straight along the bilayer saf racetrack even at room temperature ( @xmath0 k ) .
this is because that the two skyrmions consisting the saf bilayer skyrmion are tightly bound by the interlayer afm exchange coupling , where the skhes acting on the two skyrmions are opposite and canceled out . in general , for the system with a relatively large damping coefficient , the thermal fluctuation effect decreases rapidly as the layer number @xmath2 increases at a certain temperature .
our results provide a promising route to realizing the skyrmion - based device applications at room temperature .
figure [ fig1 ] illustrates the monolayer fm racetrack as well as the multilayer saf racetracks studied in this paper .
the monolayer fm racetrack contains one fm layer and a heavy - metal substrate underneath the fm layer .
the @xmath2-layer saf racetrack ( @xmath5 ) includes @xmath2 fm layers , which are separated by @xmath6 nonmagnetic spacer layers .
the fm layers are denoted from bottom to top as l1 , l2 , l3 , @xmath7 . in all models ,
the length along the @xmath8-axis , width along the @xmath9-axis , and thickness along the @xmath10-axis of each fm layer and nonmagnetic spacer are respectively equal to @xmath11 nm , @xmath12 nm , and @xmath13 nm , where the fm layer l1 is attached to the heavy - metal substrate . in the @xmath2-layer saf racetracks ( @xmath5 ) , the neighboring fm layers are antiferromagnetically exchange - coupled through their fm / spacer / fm interfaces .
the magnetization in each fm layer is perpendicular to the racetrack plane due to the high perpendicular magnetic anisotropy ( pma ) , while the magnetization in neighboring fm layers are antiparallel due to the interlayer afm exchange coupling .
the dzyaloshinskii - moriya interaction ( dmi ) in fm layers lead to the tilt of magnetization near the edges of the fm layers .
it is worth mentioning that the dmi in fm layers can be induced by both the heavy - metal substrate and spacer layers in real experiments @xcite .
recent theoretical @xcite and experimental @xcite studies have suggested and developed the methods to induce the dmi in multilayers .
more promisingly , it has been recently shown that by constructing the spacer layer made of two different heavy - metal materials , additive dmi can be achieved in multilayers @xcite . in our numerical simulations
we explicitly consider the saf @xmath2-layer skyrmions with @xmath14 .
we assume that the relaxed magnetization distributions of the fm layers l1 and l3 are almost pointing along the @xmath15-direction , while those of the fm layers l2 and l4 are almost pointing along the @xmath16-direction .
the background magnetization directions determine the skyrmion number of the skyrmion in each fm layer .
the skyrmion number equals @xmath13 in the fm layers l1 and l3 , while it equals @xmath17 in the fm layers l2 and l4 .
the total skyrmion number @xmath18 of the saf @xmath2-layer skyrmion is @xmath2 modulo @xmath19 .
namely , @xmath20 for @xmath14 , respectively ( see sec . [ se : skhe ] for details ) .
at the initial state , the skyrmions are first created and relaxed at the position of @xmath21 nm , @xmath22 nm . with regard to the injection scheme of the driving current , we consider a confined current - perpendicular - to - plane ( cpp ) geometry .
namely , an electron current flows through the heavy - metal substrate in the @xmath23-direction , which is converted into a spin current polarized along the @xmath24-direction and is perpendicularly injected into the fm layer l1 , due to the spin hall effect ( see ref . for a recent experimental example ) .
the skyrmion in the fm layer l1 is driven by the vertical spin current , while the other skyrmions in the fm layers l2 , l3 , and l4 move accordingly due to the interlayer afm exchange coupling between each adjacent fm layers .
it should be noted that , for comparison purpose , we also simulate the straightforward case of unconfined cpp geometry with the bilayer saf racetrack , where the spin current is injected into all fm layers , neglecting the spin current absorption in the model .
we investigate the multilayer saf racetrack comprised of @xmath2 fm layers , where the neighboring fm layers are antiferromagnetically exchange - coupled by the interlayer afm exchange interaction , as illustrated in fig .
the total hamiltonian @xmath25 is decomposed into the hamiltonian for each fm layer @xmath26 and the interlayer afm exchange coupling @xmath27 between neighboring fm layers , that is , @xmath28 the hamiltonian for each fm layer reads @xmath29+h_{\text{ddi } } , \label{eq : hamil - intralayer}\end{aligned}\ ] ] where @xmath30 is the fm layer index ( @xmath31 ) , @xmath32 represents the local magnetic moment orientation normalized as @xmath33 at the site @xmath34 , and @xmath35 runs over all the nearest - neighbor sites in each fm layer .
the first term represents the intralayer fm exchange interaction with the intralayer fm exchange stiffness @xmath36 .
the second term represents the dmi with the dmi coupling energy @xmath37 , where @xmath38 is the unit vector between sites @xmath34 and @xmath39 .
the third term represents the pma with the anisotropy constant @xmath40 .
@xmath41 represents the dipole - dipole interaction .
when @xmath42 , there exists an afm exchange coupling between the nearest - neighbor fm layers @xmath43 the sign of the interlayer exchange stiffness @xmath44 is negative for the interlayer afm exchange interaction .
we take the initial magnetization direction in the fm layer l1 to be pointing upward ( see fig .
[ fig1 ] ) . ) as a function of @xmath45 for the saf @xmath2-layer skyrmion at ( a ) @xmath46 and ( b ) @xmath47 .
the solid curve shows the case of the saf @xmath2-layer skyrmion with an odd @xmath2 , which has @xmath48 .
the dashed curve shows the case of the saf @xmath2-layer skyrmion with an even @xmath2 , which has @xmath49 .
here we assume @xmath50 in eq .
( [ eq : v - v - correlations ] ) . ,
scaledwidth=50.0% ] the dynamics of a skyrmion at a given finite temperature is described by introducing a gaussian stochastic magnetic field @xmath51 describing the thermal agitation of the magnetization @xcite , which satisfies @xmath52 where @xmath53 , and @xmath54 is the area of the lattice . in the cpp geometry
, we numerically solve the landau - lifshitz - gilbert ( llg ) equation including the spin - transfer torque ( stt ) term extended into the following form @xmath55 with the layer index @xmath30 suppressed .
here , @xmath56 is the effective magnetic field induced by the total hamiltonian @xmath57 , @xmath58 is the gyromagnetic ratio , @xmath59 is the gilbert damping coefficient originating from the spin relaxation , @xmath60 is the stt coefficient , and @xmath61 represents the unit spin polarization vector of the spin current .
we have @xmath62 with @xmath63 the vacuum magnetic permittivity , @xmath64 the thickness of the fm layer ,
@xmath65 the saturation magnetization , @xmath39 the applied current density , and @xmath66 the spin polarization rate .
the stt exerted on the fm layer l1 is induced by the spin hall effect .
it should be noted that we have @xmath67 in the fm layers above the fm layer l1 , thus there is no stt effect on the fm layer with the layer index number @xmath68 .
we employ the thiele equation @xcite with the inclusion of the stochastic force @xcite in order to interpret the numerical results .
we generalize it to the saf @xmath2-layer skyrmion system in multilayer saf racetracks driven by the spin current with the cpp geometry at finite temperature .
it would read in each fm layer as @xmath69 with @xmath30 the layer index , where @xmath70 , @xmath71 and @xmath72 represent the skyrmion velocity , the spin current , and the interlayer afm exchange force , respectively .
@xmath73 is the gyromagnetic coupling constant representing the magnus force with @xmath74 the skyrmion number , which is defined as @xmath75 and @xmath76 is the gaussian stochastic forces acting on the skyrmions representing the finite temperature effect , which satisfies @xmath77 where @xmath78 .
we have taken the same dissipation matrix @xmath79 and the same damping coefficient @xmath59 for all racetracks .
we now postulate that all skyrmions move together with the same velocity @xmath80 since they are tightly bound . summing all @xmath2 thiele eqs .
( [ eq : thieleeq - layer ] ) , we would phenomenologically obtain @xmath81 where the interlayer afm forces are assumed to be canceled out , that is , @xmath82 , and @xmath83 with @xmath84 and @xmath85 , @xmath86 . actually , @xmath87 , where @xmath88 is the spin current induced by the charge current in the heavy - metal substrate due to the spin hall effect ( see sec . [ se : monolayer - and - multilayer ] ) . the first term on the left hand side of eq .
( [ eq : thieleeq - total ] ) corresponds to the magnus force .
the total skyrmion number equals one , @xmath48 , when the number @xmath2 of the fm layers is odd , while it equals zero , @xmath49 , when the number @xmath2 of the fm layers is even since @xmath89 .
the variant of the sum of the gaussian noise is given by @xmath90 where @xmath53 .
as a result , the variance of the total noise becomes @xmath91 in the saf @xmath2-layer skyrmion .
therefore , the saf @xmath2-layer skyrmion is @xmath2 times more stable than the fm monolayer skyrmion at a certain temperature and a certain damping coefficient . the velocity is given by explicitly solving the thiele eq .
( [ eq : thieleeq - total ] ) as @xmath92 the mean velocity is thus given by @xmath93 when @xmath48 , which is the case for @xmath2 being odd , a skyrmion undergoes a transverse motion , where @xmath94 when @xmath49 , which is the case for @xmath2 being even , a skyrmion goes straight , where @xmath95 consequently the saf @xmath2-layer skyrmion experiences the skhe only when @xmath2 is odd .
we call it the odd - even effect on the skhe . the velocity of the skyrmion decreases with increasing @xmath2 at a given driving current density .
the velocity - velocity correlation functions are calculated as @xmath96 substituting eq .
( [ eq : eta - total ] ) , the correlation functions are obtained as @xmath97 which are functions of @xmath45 and @xmath59 for a given saf @xmath2-layer skyrmion .
since the scope of this paper is focused on the temperature effect on the saf @xmath2-layer skyrmion , we assume @xmath50 in eq .
( [ eq : v - v - correlations ] ) and show the correlation functions as a function of @xmath45 for the saf @xmath2-layer skyrmions under the assumptions of a large damping coefficient ( @xmath46 ) and a small damping coefficient ( @xmath47 ) in figs .
[ fig2](a ) and [ fig2](b ) , respectively .
it can be seen that , for the case of large damping coefficient ( @xmath46 ) [ fig .
[ fig2](a ) ] , the correlation functions of the @xmath2-layer skyrmion , where @xmath98 , increase with increasing @xmath45 . on the other hand ,
the correlation functions are inversely proportional to @xmath2 .
this indicates that the stability of the saf @xmath2-layer skyrmion with a large damping coefficient increases with @xmath2 but decreases with @xmath45 . in contrast , for the case of small damping coefficient ( @xmath47 ) [ fig .
[ fig2](b ) ] , it shows the correlation functions of the saf @xmath2-layer skyrmion , where @xmath98 , increase with increasing @xmath45 .
however , the correlation functions are nonmonotonic with respect to @xmath2 . at a certain @xmath45
, it can be seen that the correlation functions of the saf @xmath2-layer skyrmions with @xmath99 are larger than that of the saf @xmath2-layer skyrmion with @xmath100 .
this means that the stability of the saf @xmath2-layer skyrmion with a small damping coefficient decreases with @xmath45 , however , the fm monolayer skyrmion with @xmath48 is more stable than the saf bilayer skyrmion with @xmath49 at a certain @xmath45 .
it is noteworthy that this result is in good agreement with a recent study on the afm skyrmion ( see ref . ) , where the fluctuation of the afm skyrmion with @xmath49 is found to be inversely proportional to @xmath59 , and is more significant than that of the fm skyrmion with @xmath48 at a small damping coefficient @xmath101 .
it is also worth mentioning that the fluctuation of the fm skyrmion with @xmath48 is proportional to @xmath59 ( see refs . ) .
hence , in order to remain in a monotonic temperature dependence with respect to @xmath2 , we numerically study the saf @xmath2-layer skyrmion under the large damping coefficient assumption in the following sections .
the three - dimensional micromagnetic simulations are performed by using the 1.2 alpha 5 release of the object oriented micromagnetic framework ( oommf ) software developed at the national institute of standards and technology ( nist ) @xcite .
the simulations are handled by the oommf extensible solver ( oxs ) objects of the standard oommf distribution with the oxs extension modules for including the interface - induced dmi @xcite and the thermal fluctuation @xcite .
the magnetization dynamics at zero temperature ( @xmath3 k ) is controlled by the llg equation including the stt term @xcite , while a highly irregular fluctuating field representing the irregular influence of temperature is added into the llg equation including the stt term when the thermal effect is considered ( @xmath4 k ) .
the finite temperature simulations are performed with a fixed time step of @xmath102 fs , that is , @xmath103 s , while the time step in the zero temperature simulations is adaptive ( @xmath104 s ) .
each simulation with a certain finite temperature is performed @xmath102 times individually with different random seed values .
the models built in the micromagnetic simulations are discretized into regular cells with a constant cell size of @xmath19 nm @xmath105 @xmath19 nm @xmath105 @xmath13 nm , which allows for a trade - off between numerical accuracy and computational efficiency . in the cpp geometry ,
the spin current polarized along the @xmath24-direction flows upward in the bottom fm layer l1 , which is induced by the charge current flowing in the heavy - metal substrate due to the spin hall effect .
the current - induced oersted field is neglected in all simulations for simplicity since it makes only a minor contribution to the overall magnetization dynamics .
the typical material parameters used in the micromagnetic simulations are adopted from refs .
@xcite : gilbert damping coefficient @xmath106 which is large enough to satisfy the large damping coefficient assumption ; gyromagnetic ratio @xmath107 m a@xmath108 s@xmath108 ; saturation magnetization @xmath109 ka m@xmath108 ; intralayer fm exchange stiffness @xmath110 pj m@xmath108 ; interlayer afm exchange stiffness @xmath111 pj m@xmath108 ; interface afm exchange coefficient @xmath112 mj m@xmath113 ; dmi constant @xmath114 mj m@xmath113 ; pma constant @xmath115 mj m@xmath116 ; and spin - polarization rate @xmath117 .
we start with a numerical investigation of the current - velocity relation of skyrmions in @xmath2-layer saf racetracks , with @xmath14 , at zero temperature ( @xmath3 k ) ( see fig . [ fig3 ] ) .
let us recapitulate the current - induced motion of a fm monolayer skyrmion in a monolayer fm racetrack , which undergoes a transverse motion toward the upper edge of the racetrack because of the skhe ( see ref . for movie 1 ) .
we show the trajectory at a moderate driving current of @xmath118 ma cm@xmath113 in fig .
[ fig3](a ) .
the moving skyrmion reaches a stable velocity of @xmath120 m s@xmath108 and has a transverse shift of @xmath121 nm due to the skhe .
it does not touch the edge because of the repulsive force from the edge . nevertheless , when @xmath122 ma cm@xmath113 , it is destroyed by touching the edge shortly after the driving current is applied ( see ref .
for movie 1 ) .
let us also recapitulate the current - induced motion of a saf bilayer skyrmion in a bilayer saf racetrack ( see ref . for movie 2 ) .
it goes straight in the bilayer saf racetrack as a result of the suppression of the skhe ( @xmath49 ) . the trajectory at a moderate driving current of @xmath123 ma cm@xmath113 is shown in fig .
[ fig3](b ) , where it reaches a stable speed of @xmath124 m s@xmath108 .
the saf bilayer skyrmion strictly moves along the central line ( @xmath22 nm ) of the racetrack .
we go on to study the current - induced motion of a saf trilayer skyrmion with @xmath48 [ see fig .
[ fig3](c ) ] , and a saf quadrilayer skyrmion with @xmath49 [ see fig . [ fig3](d ) ] . the saf trilayer skyrmion experiences the skhe ( see ref . for movie 3 ) as in the case of the fm monolayer skyrmion . on the other hand ,
the saf quadrilayer skyrmion moves reliably in the quadrilayer saf racetrack even at a strong driving current ( see ref . for movie 4 ) , demonstrating the suppression of the skhe as in the case of a saf bilayer skyrmion .
the saf quadrilayer skyrmion moves along the central line ( @xmath22 nm ) of the racetrack .
we compare the skhes at @xmath100 and @xmath125 quantitatively .
we show the skyrmion hall angle @xmath126 as a function of time @xmath127 in fig .
[ fig4](a ) for @xmath128 .
the skyrmion hall angle is antiproportional to @xmath2 .
namely , the skhe for @xmath125 is three times smaller than that for @xmath100 .
the theoretical expectation eq .
( [ eq : hallangle ] ) explains the numerical data remarkably well with the choice of @xmath129 . in fig .
[ fig4](b ) , we show the velocity @xmath130 as a function of the applied driving current density @xmath39 for the motion of saf @xmath2-layer skyrmions , where @xmath14 . we have fitted the data successfully by theoretical expectation eq .
( [ eq : mean - velocity ] ) with the use of @xmath129 .
the velocity @xmath130 is almost antiproportional to @xmath2 as shown in fig .
[ fig4](c ) .
this is because the driving current is only applied to the bottom fm layer l1 . in order to further improve the @xmath39-@xmath131 relation of the saf @xmath2-layer skyrmions , here taking the saf bilayer skyrmion as an example
, we also investigate the @xmath39-@xmath131 curve when the driving current is applied in both constituent fm layers of the bilayer saf racetrack , which is plotted in fig .
[ fig4](b ) as a dashed curve .
it can be seen that , when both fm layers are driven by the current , the @xmath39-@xmath131 relation of the saf bilayer skyrmion matches well with that of a fm monolayer skyrmion moving in a monolayer fm racetrack at the small driving current regime . when @xmath118 ma cm@xmath113 , the velocity of the fm monolayer skyrmion is @xmath132 m s@xmath108 , while the velocity of the saf bilayer skyrmion is @xmath133 m s@xmath108 .
we proceed to investigate the effect of random thermal perturbations on the motion of saf @xmath2-layer skyrmions .
figure [ fig5 ] demonstrates the motion of a fm monolayer ( saf bilayer ) skyrmion driven by a moderate current of @xmath118 ( @xmath134 ) ma cm@xmath113 at temperatures ranging from @xmath3 k to @xmath0 k ( see ref . for movies 5 and 6 ) .
a fm monolayer skyrmion moves safely from the left to the right terminal of the racetrack at @xmath3 and @xmath12 k. however , when @xmath45 is increased and larger than @xmath135 k , it becomes unstable and is destroyed by touching the upper edge .
figures [ fig6](a)-(g ) show its trajectories at @xmath3 , @xmath12 , @xmath135 , @xmath136 , @xmath137 , @xmath138 , and @xmath139 k , respectively . at @xmath3 k , as we have already stated , the transverse shift is constant due to the balance between the skhe and the edge effect . at a finite temperature ,
the motion is affected by thermal effect , where the transverse shift is fluctuating .
it is destroyed when the fluctuated skyrmion touches the edge .
indeed , it is destroyed after having moved @xmath140 nm at @xmath1 k. when @xmath45 is raised to room temperature , that is , @xmath0 k , the fm monolayer skyrmion is destroyed very shortly after the driving current is applied . on the other hand , a saf bilayer skyrmion moves safely from the left to the right terminal of the racetrack even at @xmath0 k. the corresponding trajectories of the saf bilayer skyrmion at @xmath3 , @xmath12 , @xmath135 , @xmath136 , @xmath137 , @xmath138 , and @xmath139 k , are given in figs .
[ fig6](h)-(n ) , respectively .
since the saf bilayer skyrmion exhibits no transverse shift caused by the skhe , it is not destroyed by touching the edge as the racetrack width is wide enough here . in figs .
[ fig7](a)-(d ) , we show the distributions of the @xmath9-position of a saf @xmath2-layer skyrmion for @xmath141 nm of motion at different @xmath45 , which are fitted by using the gaussian distribution . in figs .
[ fig7](e)-(f ) , we show the mean value @xmath142 and standard deviation @xmath143 of the distribution of the @xmath9-position as functions of @xmath45 corresponding to figs .
[ fig7](a)-(d ) .
the mean @xmath142 is around the one half of the sample width for @xmath144 and @xmath145 since the saf @xmath2-layer skyrmion with an even @xmath2 goes straight reflecting the absence of the skhe . on the other hand ,
the mean @xmath142 is away from the one half of the sample width for @xmath100 and @xmath125 due to the skhe .
the deviation is larger for @xmath100 than @xmath125 , which indicates that the skhe exerted on a fm monolayer skyrmion is larger than that on a saf trilayer skyrmion . in fig .
[ fig6 ] we see that a fm monolayer skyrmion can go further at @xmath146 k than at @xmath1 k. this occurs accidentally due to the choice of the thermal random seed , which is a number employed by the random number generator to generate the thermal fluctuation field in thermal simulations @xcite . in order to evaluate and examine the effect of the thermal random seed on the simulation results , we perform each simulation at a given @xmath45 @xmath102 times individually with different thermal random seed values .
the trajectories of a fm monolayer skyrmion in a monolayer fm racetrack at @xmath147 and @xmath135 k with three selected random seed values are shown in figs .
[ fig8](a)-(b ) , respectively , where a moderate driving current of @xmath118 ma cm@xmath113 is applied .
it can be seen that the trajectories at a certain @xmath45 with different random seed values are modestly influenced by thermal perturbations . at @xmath147 k , although the motion of the fm monolayer skyrmion is fluctuating , the fm monolayer skyrmion reaches the right terminal of the racetrack in all @xmath102 simulations with different random seed values . at @xmath1
k , the fm monolayer skyrmion is destroyed by touching the upper edge of the racetrack in all @xmath102 simulations with different random seed values . on the other hand , the saf bilayer skyrmion is safely conveyed between the two terminals without touching the edge in all these simulations even at @xmath0 k , as shown in fig .
[ fig8](c ) .
we have studied the motion of skyrmions in multilayer saf racetracks in contrast to the motion of skyrmions in conventional monolayer fm racetracks . the thermal effect on the motion of skyrmions in monolayer fm racetracks as well as multilayer saf racetracks
have been investigated by including random thermal perturbations in the micromagnetic simulations .
we have found that a moving saf bilayer skyrmion is much more stable than a moving fm monolayer skyrmion , even when the temperature effect is taken into account , since the two skyrmions consisting the saf bilayer skyrmion are tightly bound by the interlayer afm exchange coupling , and thus the saf bilayer skyrmion is immune from the skhe . besides
, we have shown that the detrimental effect of the skhe on the moving fm monolayer skyrmion is enhanced as temperature increases , while the saf bilayer skyrmion can safely move along the racetrack even at room temperature .
in addition , the odd - even effect of the constituent fm layer number on the skhe in multilayer saf racetracks has also been demonstrated . due to the suppression of the skhe ,
the skyrmions have no transverse motion in multilayer saf racetracks with even constituent fm layers . in conclusion , we find that the bilayer saf racetrack is a preferred host for skyrmion transmission in racetrack - type device applications since it realizes a minimum system which does not show the skhe .
x.z . was supported by the ronpaku program of the japan society for the promotion of science . m.e .
acknowledges the support by the grants - in - aid for scientific research from jsps kakenhi ( grants no .
25400317 and no .
15h05854 ) .
y.z . acknowledges the support by national natural science foundation of china ( project no .
1157040329 ) , the seed funding program for basic research and seed funding program for applied research from the hku , itf tier 3 funding ( its/203/14 ) , the rgc - grf under grant hku 17210014 , and university grants committee of hong kong ( contract no .
aoe / p-04/08 ) .
thanks x.x .
liu for getting him involved in the jsps ronpaku program .
is very much grateful to n. nagaosa for many helpful discussions on the subject .
i. kzsmrki , s. bordcs , p. milde , e. neuber , l. m. eng , j. s. white , h. m. rnnow , c. d. dewhurst , m. mochizuki , k. yanai , h. nakamura , d. ehlers , v. tsurkan , and a. loidl , nat . mater . * 14 * , 1116 ( 2015 ) .
s. woo , k. litzius , b. krger , m .- y .
i m , l. caretta , k. richter , m. mann , a. krone , r. m. reeve , m. weigand , p. agrawal , i. lemesh , m .- a .
mawass , p. fischer , m. klui , and g. s. d. beach , nat . mater . * 15 * , 501 ( 2016 ) . c. moreau - luchaire , c. moutafis , n. reyren , j. sampaio , c. a. f. vaz , n. van horne , k. bouzehouane , k. garcia , c. deranlot , p. warnicke , p. wohlhter , j. m. george , m. weigand , j. raabe , v. cros , and a. fert , nat . nanotech . * 11 * , 444 ( 2016 ) .
o. boulle , j. vogel , h. yang , s. pizzini , d. de souza chaves , a. locatelli , t. onur mente , a. sala , l. d. buda - prejbeanu , o. klein , m. belmeguenai , y. roussign , a. stashkevich , s. m. chrif , l. aballe , m. foerster , m. chshiev , s. auffret , i. m. miron , and g. gaudin , nat . nanotech . * 11 * , 449 ( 2016 ) .
f. bttner , c. moutafis , m. schneider , b. krger , c. m. gnther , j. geilhufe , c. v. korff schmising , j. mohanty , b. pfau , s. schaffert , a. bisig , m. foerster , t. schulz , c. a. f. vaz , j. h. franken , h. j. m. swagten , m. klui , and s. eisebitt , nat . phys . * 11 * , 225 ( 2015 ) .
oommf oxs extension module of the spin - transfer torque and thermal fluctuation effect can be downloaded at http://kelvinxyfong.wordpress.com/research/research-interests/oommf-extensions/oommf-extension-xf_thermspinxferevolve .
see supplemental materials at [ url ] for movies on saf @xmath2-layer skyrmion motion ( @xmath14 ) at zero and finite temperatures . a movie on a saf bilayer skyrmion motion with @xmath148 pj m@xmath108 is also given .
the seed of the random number generator used to generate the thermal fluctuation field is set to @xmath135 in all thermal simulations . | a magnetic skyrmion is a topological magnetization structure with a nanometric size and a well - defined swirling spin distribution , which is anticipated to be an essential building block for novel skyrmion - based device applications .
we study the motion of magnetic skyrmions in multilayer synthetic antiferromagnetic ( saf ) racetracks as well as in conventional monolayer ferromagnetic ( fm ) racetracks at finite temperature .
there is an odd - even effect of the constituent fm layer number on the skyrmion hall effect ( skhe ) .
namely , due to the suppression of the skhe , the magnetic skyrmion has no transverse motion in multilayer saf racetracks packed with even fm layers .
it is shown that a moving magnetic skyrmion is stable even at room temperature ( @xmath0 k ) in a bilayer saf racetrack but it is destructed at @xmath1 k in a monolayer fm racetrack .
our results indicate that the saf structures are reliable and promising candidates for future applications in skyrmion - electronics and skyrmion - spintronics . |
raman spectroscopy and fundamental research on graphitic systems have been two inherently connected fields for many decades in solid - state physics . since the early works of tuinstra and koenig @xcite , raman spectroscopy has evolved as a very versatile tool to characterize graphite , graphene , and carbon nanotubes @xcite . especially after the introduction of the double - resonance raman concepts by thomsen and reich @xcite ,
many phenomena could be unraveled such as the anomalous dispersion of the @xmath3 and @xmath4 modes with laser excitation energy or the origin of many other , higher - order raman modes . however , most of the works focus on the double - resonant raman modes related to transverse and longitudinal optical phonons ( to and lo ) , such as the well - known @xmath3 , @xmath5 , and the @xmath4 mode , or the first - order @xmath6 band in graphene and the radial - breathing mode ( rbm ) in carbon nanotubes @xcite . recently , also
the investigation of small interlayer vibrational modes in few - layer graphene came into the focus of research , since these modes can directly probe the layer number and interlayer interaction @xcite . beside these modes , there are various other raman bands that result from double - resonant two - phonon scattering processes or from phonon - defect scattering @xcite . in this work ,
we focus on the defect - induced @xmath0 mode ( around 1100@xmath1 ) in graphene and carbon nanotubes that results from double - resonant scattering of longitudinal acoustic ( la ) phonons with defects .
low - energy acoustic phonons that stem from the brillouin - zone center strongly affect charge carrier mobilities and thermal transport properties and are thus very important for the performance of electronic devices @xcite .
however , first - order raman scattering does not allow to probe these phonons .
the la phonon was often observed in combination with other phonons in double - resonant raman scattering processes @xcite . especially in bi- and fewlayer graphene ,
combination modes containing the la phonon can be observed in the frequency range between the @xmath6 and the @xmath4 mode @xcite . however , none of the previous works analyzed the la phonon itself . here , we report the first measurements of the theoretically predicted @xmath0 mode in carbon nanotubes and graphene with intentionally created defects @xcite .
we investigate its dependence on laser excitation energy , the number of graphene layers and the carbon nanotube diameter .
we prove that this raman mode stems from so - called inner processes , again highlighting the importance of these contributions .
graphene samples were prepared by micro - mechanical exfoliation of natural graphite crystals onto silicon substrates with an 100 nm oxide layer .
the layer number of the prepared samples was unambiguously identified by their optical contrast and layer - number dependent raman modes @xcite .
the graphene samples were then transferred to a vacuum chamber and irradiated with swift heavy ions ( xe@xmath7 , 91 mev ) using a fluence of 65.000 ions/@xmath8 at normal angle to the graphene plane . under grazing - incidence irradiation
these projectiles cause extended modifications in graphene @xcite , while at normal incidence point - like defects are created @xcite . due to their high energy ,
the interaction of swift heavy ions with matter is exclusively by inelastic scattering .
as the penetration depth of 91mev xe ions is about 10@xmath9 m @xcite , defects are not introduced exclusively in single - layer but in bi- and tri - layer graphene as well . from the fluence of the xe ions , we deduce an average length between defects of approximately 4 nm . a determination of the defect length @xmath10 by the @xmath3/@xmath6-mode intensity ratio and formulas from ref .
is misleading in the present case .
the irradiation with 91mev xe ions creates significantly smaller defects compared to average defects sizes from previous studies with low - energy argon ions @xcite .
therefore , we observe a reduced @xmath3/@xmath6 mode ratio at the same @xmath10 compared to ref . .
approximate defect sizes in the present study are @xmath11 nm and @xmath12 nm ( using the notation from ref . ) . for the measurements on carbon nanotubes , we used buckypaper carbon nanotubes produced by the hipco process @xcite , having a diameter distribution of @xmath13 , which was verified by measurements of the rbm .
raman measurements were done with a horiba hr800 and a dilor xy spectrometer , equipped with solid - state lasers , as well as dye and gas lasers .
raman spectra were recorded in back - scattering geometry under ambient conditions using a 1800lines / mm grating and an @xmath14 objective , yielding a spectral resolution of approx .
1@xmath1 . since the @xmath0 mode and
the second - order raman modes of silicon are close in frequency , it was necessary to perform a background subtraction for graphene measured on silicon substrates . for this purpose
, we used the same experimental conditions as for the raman measurements on graphene , _
i.e. _ , the same laser power and integration time , and recorded the silicon background at a spot adjacent to the investigated graphene flake . during all measurements
the laser power was kept below 1mw in order to avoid sample heating , laser - induced doping or the creation of additional unwanted defects .
, @xmath6 , @xmath5 modes of graphene , as well as the second - order raman peak of the silicon substrate are labeled .
the inset shows an enlarged view of the @xmath0 mode at approximately 1100@xmath1 .
( b ) phonon dispersion of single - layer graphene obtained from _ ab - initio _ dft calculation .
the la phonon branch is marked in orange .
the hatched ellipse indicates the _
k_-space region where the resonant phonons stem from .
figure [ fig1](a ) shows an overview raman spectrum of defective single - layer graphene at 532 nm laser excitation wavelength .
the well - known @xmath3 , @xmath6 , and @xmath5 raman modes of single - layer graphene can be clearly identified . by enlarging the spectral range between the second - order raman peak of the silicon substrate and the @xmath3 mode
, we can identify another raman mode .
we explicitly verified that this raman mode can not be observed at arbitrary edges of exfoliated graphene , indicating that a larger number of defects is needed to result in a measurable raman signal . following the notation of venezuela _
, this mode is referred to as @xmath0 .
the observed @xmath0/@xmath3 mode intensity ratio in the spectrum of fig . [ fig1](a ) is approximately 1:100 .
the small intensity of the @xmath0 mode is attributed to the relatively weak electron - phonon coupling of the involved la phonons compared to the large electron - phonon coupling of the to - derived phonon modes near the @xmath15 point in graphene @xcite .
figure [ fig2](a ) presents raman spectra of the @xmath0 mode in single - layer graphene at four different laser excitation wavelengths .
the laser - energy dependent peak shift indicates that this raman mode results from a double - resonance process .
the observed downshift of approximately @xmath16@xmath1/ev with increasing laser energy is opposite to the well - known behavior of the @xmath3 and @xmath4 mode in graphene [ compare fig .
[ fig2](c ) ] .
this can be explained by the dispersion of the phonon branch that is involved in the scattering process .
the @xmath0 mode is assigned to a double - resonant intervalley scattering process that involves an la phonon and a defect , as theoretically predicted in ref . .
the phonon dispersion of single - layer graphene is shown in fig .
[ fig1](b ) , the la branch is highlighted .
supercell approach .
] since this raman mode results from an intervalley scattering process , the dominant contributions stem from phonons along the @xmath17 high - symmetry direction .
however , only between @xmath18 and @xmath15 the measured @xmath0-mode frequencies match the calculated la dispersion [ compare fig . [ fig2](b ) ] .
therefore , we can assign this scattering process to so - called inner processes , _
i.e. _ , the electronic transition must be between @xmath15 and @xmath19 .
the importance of inner processes was discussed extensively in recent literature on double - resonances in graphene @xcite , showing that the formerly proposed restriction to only outer processes is incorrect @xcite .
our results furthermore enable us to map the la phonon branch along the @xmath2 high - symmetry direction .
figure [ fig2](b ) compares the experimentally obtained @xmath0-mode frequencies with the calculated la phonon branch dispersion along @xmath17 .
the resonant phonon wave - vectors for each excitation energy were obtained from the resonance condition on the incoming and outgoing photon in the double - resonance process ; a _ gw_-corrected band structure of single - layer graphene was used .
the experimental values match the theoretical curve along @xmath2 within an error of less than 10@xmath1 . by tuning the laser excitation energy , one can now follow the la phonon branch along the high - symmetry line .
the experimentally obtained values , including the origin , can be fitted by a sine function ; the resulting fit shows a deviation of less than 10@xmath1 close to @xmath18 compared to our _ ab - initio _ calculations ( see si ) .
thus , our approach enables us to investigate acoustic phonons by an optical method , _
i.e. _ , double - resonant raman spectroscopy , not only at certain phonon wave - vectors .
in fact , the la phonon branch , obtained from fitting our experimental data , also showed good agreement close to @xmath18 , which is the @xmath20-space region that predominantly affects heat transfer .
we will now turn our discussion to the lineshape of the @xmath0 mode . from the raman spectra in fig .
[ fig2](a ) we can derive two statements : first , the @xmath0 mode shows a pronounced asymmetry towards higher frequencies and second , the full width at half maximum ( fwhm ) of @xmath0 mode increases with increasing laser excitation energy .
the asymmetric tail can be seen very clearly in all spectra .
we fitted all raman spectra with two lorentzian components ; the spectrum at 633 nm laser wavelength exemplarily shows their individual contributions .
the asymmetry of the @xmath0 mode can be directly explained with the two - dimensional phonon dispersion of the la branch around the @xmath15 point .
in contrast to the to branch , the la phonon branch shows a non - constant angular frequency dependence around the @xmath15 point , _ i.e. _ , the lowest frequencies can be found along @xmath21 and the highest along @xmath22 @xcite .
thus , we can assign the intense , lower - frequency lorentzian to contributions directly from the @xmath2 high - symmetry direction .
the high - frequency tail of the @xmath0 peak can be assigned to contributions next to the high - symmetry line in the brillouin zone , following the analysis of may _
the broadening of the lineshape at larger excitation energies can be easily explained by considering the dispersion of the la branch along @xmath2 , which has a steeper dispersion closer to @xmath18 [ see fig .
[ fig1](b ) ] . therefore , for shorter resonant phonon wave - vectors , _
i.e. _ , larger excitation energies , a larger frequency range can be accessed , leading to a broadening of this raman mode .
similar observations , both theoretically and experimentally , were made for the @xmath23 mode at approximately 2450@xmath1 in the raman spectrum of single - layer graphene @xcite .
the comparison of our measured @xmath0-mode spectra with calculations from venezuela _ et al .
_ ( compare ref . , fig .
11 ) shows very good agreement , both qualitatively and quantitatively .
the prediction of an asymmetric high - frequency tail was confirmed by our measurements .
also the absolute frequencies match the experimentally observed , supporting our assignment to the la - defect scattering process .
figure [ fig2](d ) shows @xmath0-mode spectra at 2.33ev laser excitation energy for different numbers of graphene layers .
as can be seen very clearly , the lineshape significantly broadens when going from single - layer to bilayer graphene .
this broadening can be directly attributed to the evolution of the electronic bandstructure around the @xmath15 point . since bilayer graphene has two valence and conduction bands , the number of resonant scattering processes is quadrupled compared to single - layer graphene .
the increased number of resonant phonon wave - vectors leads to an increased number of resonantly enhanced phonons , thus resulting in a broadening of the raman mode .
this effect can be also observed for the @xmath23 mode in bilayer graphene @xcite , as well as for the @xmath4 mode @xcite . by further increasing the layer number
, the lineshape does not show any noticeable changes in linewidth or by the appearance of additional peaks .
this can be again identified with the shape of the electronic bands around @xmath15 .
the bandstructure of trilayer graphene can be regarded as a superposition of the electronic bands of single and bilayer graphene @xcite .
thus , also the resonant phonon wave - vectors in the double - resonance process are very similar to the ones from single and bilayer graphene .
therefore , the @xmath0-mode lineshape in trilayer graphene does not differ from the peak observed in bilayer graphene . due to the degeneracy of the la phonon branch in bi- and tri - layer graphene @xcite
, the broadening of the @xmath0 mode can not result from the evolution of the phonon dispersion with increasing layer number , but rather from the evolution of the electronic bandstructure .
similar to graphene , only a few publications reported double - resonant raman modes in carbon nanotubes that involve la phonons @xcite . besides intravalley scattering with lo and la derived phonons , also intervalley scattering with to and la derived phonons was reported @xcite .
however , double - resonance processes combining a defect and an la phonon have not been reported so far in carbon nanotubes .
-mode region ; the @xmath0mode is marked with the red rectangle .
( b ) raman spectra of the @xmath0 mode in carbon nanotubes for various laser excitation energies .
spectra are vertically offset for clarity .
( c ) calculated @xmath0-mode frequencies for carbon nanotubes in the diameter range between 7 and 14 .
( d ) calculated frequencies as a function of the tube diameter for the semiconducting @xmath24 transition .
the solid lines in ( c ) and ( d ) represent linear fits to the calculated data points .
[ fig3 ] ] figure [ fig3](b ) shows raman spectra of the @xmath0 mode in carbon nanotubes for six different laser excitation energies . due to a broader @xmath3 mode in carbon nanotubes compared to graphene @xcite ,
the @xmath0 mode can only be observed as a low - frequency shoulder to the @xmath3 mode [ compare fig .
[ fig3](a ) ] . from the experimental spectra
we estimate a downshift of the @xmath0 mode with higher laser excitation energies of approximately @xmath25@xmath1/ev .
in contrast to graphene , we observe a complex peak structure with many different contributions .
this can be understood from the fact that the raman spectrum of carbon nanotubes results from many different tubes in or close to resonance with the excitation laser .
the analyzed hipco sample contains a large variety of carbon nanotubes with diameters around @xmath26 .
therefore , the @xmath0-mode lineshape is broadened by the different contributions of carbon nanotubes in or close to resonance .
however , an assignment of the different features in the @xmath0 band to distinct carbon nanotubes is not possible because of two reasons : the large number of different carbon nanotubes analyzed and the very low intensity of the @xmath0 band . in order to derive a systematic analysis , we present calculated @xmath0-mode frequencies for all carbon nanotubes in the diameter range between 7 and 14 . the calculation is based on a sixth - nearest neighbor tight - binding model with symmetry - imposed modifications for carbon nanotubes using the polsym code @xcite .
the calculated la phonon branch is scaled in frequency by 5.1% in order to fit the experimentally observed value at the @xmath15 point for graphite @xcite .
the resonant phonon wave vectors in the double - resonance process were obtained by assuming that scattering occurs between two equivalent extrema in the electronic band structure .
as indicated by other works before @xcite , we assumed that the dominant contribution to the double - resonance process results from the incoming resonance .
similar to graphene , the oscillator strength for optical transitions in carbon nanotubes is highest along the wavevectors derived from the @xmath22 direction in the graphene brillouin zone @xcite .
therefore , the raman spectrum is dominated by carbon nanotubes , where these transitions are probed , _ e.g. _ , the @xmath27 family for the semiconducting @xmath24 transition . of a carbon nanotube
is given by @xmath28 , where @xmath29 and @xmath30 are the chiral indices of the cnt .
] thus , we restrict our calculations to those transitions . the calculated @xmath0-mode frequencies in fig .
[ fig3](c ) reproduce the experimentally observed peak positions and peak shift very well .
the calculated shift rate of @xmath31@xmath1/ev is in reasonable agreement with the experiments . as can be seen in fig .
[ fig3](c ) , the calculated @xmath0-mode frequencies at each laser energy cover a large frequency range of approximately 40@xmath1 , in accordance with the experimentally observed broad lineshape of the @xmath0 mode in carbon nanotubes . fig .
[ fig3](d ) shows the calculated frequencies as a function of the tube diameter for the semiconducting @xmath24 transition ( all other transitions show the same behavior ) .
we observe an upshift of approximately 7@xmath1/ with increasing tube diameter .
since the la phonon branch itself shows nearly no dependence on the tube diameter @xcite , the observed diameter dependence basically reflects the diameter dependence of the optical transition energies .
in summary , we presented experimental raman spectra of the @xmath0 mode in graphene and carbon nanotubes at various laser excitation energies .
we showed that this mode results from double - resonant intervalley scattering of la phonons with defects and has a dispersion of @xmath32 in single - layer graphene .
we demonstrated that the @xmath0 must stem from so - called inner scattering processes with additional contributions from phonons next to the @xmath2 direction , explaining the observed high - frequency tail of this raman mode .
we further showed that the lineshape in graphene depends on the layer number , reflecting the evolution of the electronic bands around @xmath15 . in carbon nanotubes
, the lineshape of the @xmath0 mode is significantly broadened due to contributions from different tubes in or close to resonance with excitation laser .
our theoretical calculations of this raman mode in carbon nanotubes showed very good agreement with the experimental data .
our analysis presents a fast and simple method to investigate acoustic phonon branches by an optical method .
although the probed la phonons stem from a region close to the @xmath15 point , we demonstrated the possibility to derive the la dispersion along the complete @xmath2 direction .
furthermore , our results again highlight the importance of inner contributions to the double - resonance raman process in graphitic systems .
the authors acknowledge h. lebius and b. ban - detat from the irrsud beamline of the grand accelerateur national dions lourds ganil , caen , france , for performing the irradiation of the exfoliated graphene samples .
this work was supported by the european research council ( erc ) under grant no .
259286 and by the dfg under grant number ma 4079/7 - 2 and schl 384/15 - 1 within the spp1459 `` graphene '' .
57ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1063/1.1674108 [ * * , ( ) ] @noop _ _
( , ) link:\doibase 10.1098/rsta.2004.1454 [ * * , ( ) ] link:\doibase 10.1098/rsta.2004.1444 [ * * , ( ) ] link:\doibase 10.1038/nnano.2013.46 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.93.177401 [ * * , ( ) ] link:\doibase 10.1103/physrevb.72.205438 [ * * , ( ) ] link:\doibase 10.1021/nn2044356 [ * * , ( ) ] link:\doibase 10.1103/physrevb.61.14095 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1016/j.carbon.2009.12.057 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1021/nl302450s [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.121404 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1021/nl0731872 [ * * , ( ) ] link:\doibase 10.1103/physrevb.80.195423 [ * * , ( ) ] link:\doibase 10.1038/nnano.2008.58 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.075402 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.ssc.2011.05.017 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.035419 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevb.86.195434 [ * * , ( ) ] link:\doibase 10.1063/1.3559619 [ * * , ( ) ] link:\doibase 10.1063/1.4801973 [ * * , ( ) ] link:\doibase 10.1063/1.4808460 [ * * , ( ) ] link:\doibase 10.1016/j.nimb.2010.02.091 [ * * , ( ) ] link:\doibase 10.1021/nl201432 g [ * * , ( ) ] link:\doibase 10.1016/s0009 - 2614(99)01029 - 5 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.93.185503 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1073/pnas.1004595107 [ * * , ( ) ] link:\doibase 10.1103/physrevb.77.125401 [ * * , ( ) ] link:\doibase 10.1103/physrevb.66.245410 [ * * , ( ) ] link:\doibase 10.1103/physrevb.66.155418 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1016/j.carbon.2004.11.044 [ * * , ( ) ] link:\doibase 10.1021/jp803855z [ * * , ( ) ] link:\doibase 10.1063/1.2354081 [ * * , ( ) ] link:\doibase 10.1002/pssb.201200175 [ * * , ( ) ] in @noop _ _ , ( , , ) link:\doibase 10.1103/physrevb.80.085423 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.165423 [ * * , ( ) ] link:\doibase 10.1103/physrevb.74.195431 [ * * , ( ) ] link:\doibase 10.1103/physrevb.74.115415 [ * * , ( ) ] link:\doibase 10.1103/physrevb.75.035427 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/14/11/302 [ * * , ( ) ] @noop * * , ( ) | we present measurements of the @xmath0 raman mode in graphene and carbon nanotubes at different laser excitation energies .
the raman mode around 1050 - 1150@xmath1 originates from a double - resonant scattering process of longitudinal acoustic ( la ) phonons with defects .
we investigate its dependence on laser excitation energy , on the number of graphene layers and on the carbon nanotube diameter .
we assign this raman mode to so - called inner processes with resonant phonons mainly from the @xmath2 high - symmetry direction .
the asymmetry of the @xmath0 mode is explained by additional contributions from phonons next to the @xmath2 line .
our results demonstrate the importance of inner contributions in the double - resonance scattering process and add a fast method to investigate acoustic phonons in graphene and carbon nanotubes by optical spectroscopy . |
gravitational lensing has become one of important subjects in modern astronomy and cosmology ( e.g. , schneider 2006 , weinberg 2008 ) .
it has many applications as gravitational telescopes in various fields ranging from extra - solar planets to dark matter and dark energy at cosmological scales .
this paper focuses on gravitational lensing due to a n - point mass system .
indeed it is a challenging problem to express the image positions as functions of lens and source parameters .
there are several motivations .
one is that gravitational lensing offers a tool of discoveries and measurements of planetary systems ( schneider and weiss 1986 , mao and paczynski 1991 , gould and loeb 1992 , bond et al .
2004 , beaulieu et al .
2006 ) , compact stars , or a cluster of dark objects , which are difficult to probe with other methods .
gaudi et al . (
2008 ) have recently found an analogy of the sun - jupiter - saturn system by lensing .
another motivation is theoretically oriented .
one may be tempted to pursue a transit between a particle method and a fluid ( mean field ) one . for microlensing studies ,
particle methods are employed , because the systems consist of stars , planets or machos . in cosmological lensing , on the other hand , light propagation is considered for the gravitational field produced by inhomogeneities of cosmic fluids , say galaxies or large scale structures of the universe ( e.g. , refregier 2003 for a review ) .
it seems natural , though no explicit proof has been given , that observed quantities computed by continuum fluid methods will agree with those by discrete particle ones in the limit @xmath0 , at least on average , where @xmath1 is the number of particles .
related with the problems mentioned above , we should note an astronomically important effect caused by the finiteness of @xmath1 .
for most of cosmological gravitational lenses ( both of strong and weak ones ) , a continuum approximation can be safe and has worked well .
there exists an exceptional case , however , for which discreteness becomes important .
one example is a quasar microlensing due to a point - like lens object , which is possibly a star in a host galaxy ( for an extensive review , wambsganss 2006 ) .
a galaxy consists of very large number @xmath1 particles , and light rays from an object at cosmological distance may have a chance to pass very near one of the point masses . as a consequence of finite-@xmath1 effect in large @xmath1 point lenses , anomalous changes in the light curve
are observed .
for such a quasar microlensing , hybrid approaches are usually employed , where particles are located in a smooth gravitational field representing a host galaxy .
it is thus likely that @xmath1 point - mass approach will be useful also when we study such a finite-@xmath1 effect at a certain transit stage between a particles system and a smooth one .
along this course , an ( 2007 ) investigated a @xmath1 point lens model , which represents a very special configuration that every point masses are located on regular grid points . for a @xmath1 point - mass lens at a general configuration ,
very few things are known in spite of many efforts . among known ones
is the maximum number of images lensed by @xmath1 point masses . after direct calculations by witt ( 1990 ) and mao
, petters and witt ( 1997 ) , a careful study by rhie ( 2001 for n=4 , 2003 for general n ) revealed that it is possible to obtain the maximum number of images as @xmath2 .
this theorem for polynomials has been extended to a more general case including rational functions by khavinson and neumann ( 2006 ) .
( see khavinson and neumann 2008 for an elegant review on a connection between the gravitational lens theory and the algebra , especially the fundamental theorem of algebra , and its extension to rational functions ) . * theorem * ( khavinson and neumann 2006 ) : + let @xmath3 , where @xmath4 and @xmath5 are relatively prime polynomials in @xmath6 , and let @xmath7 be the degree of @xmath8 .
if @xmath9 , then the number of zeros for @xmath10 . here ,
@xmath6 and @xmath11 denote a complex number and its complex conjugate , respectively .
furthermore , bayer , dyer and giang ( 2006 ) showed that in a configuration of point masses , replacing one of the point deflectors by a spherically symmetric distributed mass only introduces one extra image .
hence they found that the maximum number of images due to @xmath1 distributed lensing objects located on a plane is @xmath12 .
global properties such as lower bounds on the number of images are also discussed in petters , levine and wambsganss ( 2001 ) and references therein . in spite of many efforts on @xmath1 lensing objects , functions for image positions
are still unknown even for @xmath1 point - mass lenses in a general configuration under the thin lens approximation .
hence it is a challenging problem to express the image positions as functions of lens and source locations .
once such an expression is known , one can immediately obtain magnifications via computing the jacobian of the lens mapping ( schneider et al .
1992 ) . only for a very few cases such as a single point mass and a singular isothermal ellipsoid ,
the lens equation can be solved by hand and image positions are known , because the lens equation becomes a quadratic or fourth - order one ( for a singular isothermal ellipsoid , asada et al .
for the binary lens system , the lens equation has the degree of five in a complex variable ( witt 1990 ) .
it has the same degree also in a real variable ( asada 2002a , asada et al .
this improvement is not trivial because a complex variable brings two degrees of freedom
. this single - real - variable polynomial has advantages .
for instance , the number of real roots ( with vanishing imaginary parts ) corresponds to that of lensed images .
the analytic expression of the caustic , where the number of images changes , is obtained by the fifth - order polynomial ( asada et al .
galois showed , however , that the fifth - order and higher polynomials can not be solved algebraically ( van der waerden 1966 ) .
hence , no formula for the quintic equation is known . for this reason
, some numerical implementation is required to find out image positions ( and magnifications ) for the binary gravitational lens for a general position of the source . only for special cases of the source at a symmetric location such as on - axis sources ,
the lens equation can be solved by hand and image positions are thus known ( schneider and weiss 1986 ) . for a weak field region ,
some perturbative solutions for the binary lens have been found ( bozza 1999 , asada 2002b ) , for instance in order to discuss astrometric lensing , which is caused by the image centroid shifts ( for a single mass , miyamoto and yoshii 1995 , walker 1995 ; for a binary lens , safizadeh et al .
1999 , jeong et al . 1999 ,
asada 2002b ) .
if the number of point masses @xmath1 is larger than two , the basic equation is much more highly non - linear so that the lens equation can be solved only by numerical methods . as a result
, observational properties such as magnifications and image separations have been investigated so far numerically for @xmath1 point - mass lenses .
this makes it difficult to investigate the dependence of observational quantities on lens parameters .
this paper is the first attempt to seek an analytic expression of image positions without assuming any special symmetry . for this purpose
, we shall present a method of taylor - series expansion to solve the lens equation for @xmath1 point - mass lens systems .
our method allows a systematic iterative analysis as shown later . under three assumptions of weak gravitational fields , thin lenses and small deflection angles ,
gravitational lensing is usually described as a mapping from the lens plane onto the source plane ( schneider et al .
bourassa and kantowski ( 1973 , 1975 ) introduced a complex notation to describe gravitational lensing .
their notation was exclusively used to describe lenses with elliptical or spheroidal symmetry ( borgeest 1983 , bray 1984 , schramm 1990 ) . for @xmath1 point lenses , witt ( 1990 ) succeeded in recasting the lens equation into a single - complex - variable polynomial .
this is in an elegant form and thus has been often used in investigations of point - mass lenses .
an advantage in the single - complex - variable formulation is that we can use some mathematical tools applicable to complex - analytic functions , especially polynomials ( witt 1993 , witt and petters 1993 , witt and mao 1995 ) .
one tool is the fundamental theorem of algebra : every non - constant single - variable polynomial with complex coefficients has at least one complex root .
this is also stated as : every non - zero single - variable polynomial , with complex coefficients , has exactly as many complex roots as its degree , if each root is counted as many times as its multiplicity . on the other hand , in the original form of the lens equation , one can hardly count up the number of images because of non - linearly coupled properties . this theorem , therefore , raises a problem in gravitational lensing .
the single - variable polynomial due to @xmath1 point lenses has the degree of @xmath13 , though the maximum number of images is @xmath2 .
this means that unphysical roots are included in the polynomial ( for detailed discussions on the disappearance and appearance of images near fold and cusp caustics for general lens systems , see also petters , levine and wambsganss ( 2001 ) and references therein ) .
first , we thus investigate explicitly behaviors of roots for the polynomial lens equation from the viewpoint of perturbations .
we shall identify unphysical roots .
secondly , we shall re - examine the lens equation , so that the appearance of unphysical roots can be avoided .
this paper is organised as follows . in section 2
, the complex description of gravitational lensing is briefly summarised .
the lens equation is embedded into a single - complex - variable polynomial in section 3 .
perturbative roots for the complex polynomial are presented for binary and triple systems in sections 4 and 5 , respectively .
they are extended to a case of @xmath1 point lenses in section 6 . in section 7
, we re - examine the lens equation in a dual - complex - variables formalism and its perturbation scheme for a binary lens for its simplicity .
the perturbation scheme is extended to a @xmath1 point lens system in section 8 .
section 9 is devoted to the conclusion .
we consider a lens system with n point masses .
the mass and two - dimensional location of each body is denoted as @xmath14 and the vector @xmath15 , respectively . for the later convenience ,
let us define the angular size of the einstein ring as @xmath16 where @xmath17 is the gravitational constant , @xmath18 is the light speed , @xmath19 is the total mass @xmath20 and @xmath21 , @xmath22 and @xmath23 denote distances between the observer and the lens , between the observer and the source , and between the lens and the source , respectively . in the unit normalised by the angular size of the einstein ring ,
the lens equation becomes @xmath24 where @xmath25 and @xmath26 denote the vectors for the position of the source and image , respectively and we defined the mass ratio and the angular separation vector as @xmath27 and @xmath28 . in a formalism based on complex variables , two - dimensional vectors for the source , lens and image positions
are denoted as @xmath29 , @xmath30 , and @xmath31 , respectively ( see also fig .
@xmath32 ) . by employing this formalism ,
the lens equation is rewritten as @xmath33 where the asterisk @xmath34 means the complex conjugate .
the lens equation is non - analytic because it contains both @xmath6 and @xmath11 .
( the circle ) and @xmath6 ( the filled disk ) , respectively .
locations of n point masses are denoted by @xmath35 ( filled triangles ) for @xmath36 . here , we assume the thin lens approximation . ,
the complex conjugate of eq .
( @xmath37 ) is expressed as @xmath38 this expression can be substituted into @xmath11 in eq .
( @xmath37 ) to eliminate the complex variable @xmath11 . as a result
, we obtain a @xmath39-th order analytic polynomial equation as ( witt 1990 ) @xmath40 equation ( a3 ) in witt ( 1990 ) takes a rather complicated form because of inclusion of nonzero shear @xmath41 due to surrounding matter .
bayer et al . ( 2006 ) uses a complex formalism in order to discuss the maximum number of images in a configuration of point masses , by replacing one of point deflectors by a spherically symmetric distributed mass .
their lens equation ( 3 ) agrees with eq .
( @xmath42 ) . in order to show this agreement
, one may use @xmath43 .
it is worthwhile to mention that eq .
@xmath44 contains not only all the solutions for the lens equation @xmath45 but also unphysical false roots which do not satisfy eq .
@xmath45 , in price of the manipulation for obtaining an analytic polynomial equation , as already pointed out by rhie ( 2001 , 2003 ) and bayer et al .
such an inclusion of unphysical solutions can be easily understood by remembering that we get unphysical roots as well as true ones if one takes a square of an equation including the square root .
in fact , an analogous thing happens in another example of gravitational lenses such as an isothermal ellipsoidal lens as a simple model of galaxies ( asada et al .
2003 ) .
in general , the mass ratio @xmath46 satisfies @xmath47 , so that it can be taken as an expansion parameter . without loss of generality , we can assume that the first lens object is the most massive , namely @xmath48 for @xmath49 . thus , formal solutions are expressed in taylor series as @xmath50 where the coefficients @xmath51 are independent of @xmath46 . up to this point
, the origin of the lens plane is arbitrary . in the following ,
the origin of the lens plane is chosen as the location of the mass @xmath52 , such that one can put @xmath53 .
this enables us to simplify some expressions and to easily understand their physical meanings , mostly because gravity is dominated by @xmath52 in most regions except for the vicinity of @xmath54 @xmath55 .
namely , it is natural to treat our problem as perturbations around a single lens by @xmath52 ( located at the origin of the coordinates ) . in numerical simulations or practical data
analysis , however , one may use the coordinates in which the origin is not the location of @xmath52 .
if one wishes to consider such a case of @xmath56 , one could make a translation by @xmath57 as @xmath58 , @xmath59 and @xmath60 in our perturbative solutions that are given below .
in this section , we investigate binary lenses explicitly up to the third order .
this simple example may help us to understand the structure of the perturbative solutions .
for an arbitrary @xmath1 case , expressions of iterative solutions are quite formal ( see below ) . in powers of @xmath65 ,
the polynomial equation is rewritten as @xmath66 where we defined @xmath67 ( w^ * z^2-w w^ * z - w ) , \nonumber\\ f_1(z)&=&(z - w ) \nonumber\\ & & \times % \left ( \bigl ( \epsilon ( z - w ) [ ( 2w^*-\epsilon^*)z+2 ] % \nonumber\\ % & & -\epsilon^ * z^2 ( z-\epsilon ) - \epsilon z \bigr ) , % \right ) , \nonumber\\ f_2(z)&=&\epsilon^2 ( z - w ) .
\label{f}\end{aligned}\ ] ] at @xmath69 , the lens equation becomes the fifth - order polynomial equation as @xmath70 .
zeroth order solutions are obtained by solving this .
all the solutions are @xmath64 ( doublet ) , @xmath71 and @xmath72 , where we defined @xmath73 one of the roots , @xmath71 , is unphysical , because it does not satisfy eq .
@xmath45 at @xmath69 . by using all the 0th order roots including unphysical ones , @xmath74 is factorised as @xmath75 next , we seek @xmath76st - order roots .
we put @xmath77 . at the linear order in @xmath65 , eq .
( @xmath42 ) becomes @xmath78 where the prime denotes the derivative with respect to @xmath6 .
thereby we obtain a @xmath76st - order root as @xmath79 the similar manner can not be applied to a case of @xmath64 , because it is a doublet root with @xmath80 , while @xmath81 . at @xmath82 ,
( @xmath42 ) can be factorised as @xmath83 + \epsilon \right ) \nonumber\\ & & \times \left(z_{(1 ) } [ ( \epsilon - w)(w^ * \epsilon + 1)-\epsilon ] + \epsilon ( \epsilon - w ) \right ) = 0 .
\label{2-lenseq-2nd}\end{aligned}\ ] ] hence , we obtain two roots as @xmath84 here , the latter root expressed by eq .
( @xmath85 ) is unphysical and thus abandoned , because it does nt satisfy the original lens equation ( @xmath86 ) . on the other hand , the former root by eq .
( @xmath87 ) satisfies the equation and thus expresses a physically correct image .
99 an j. h. , 2007 , mnras , 376 , 1814 asada h. , 2002a , a&a , 390 , l11 asada h. , 2002b , apj , 573 , 825 asada h. , kasai t. , kasai m. , 2002c , prog .
phys . , 108 , 1031 asada h. , hamana t. , kasai m. , 2003 , a&a , 397 , 825 asada h. , kasai t. , kasai m. , 2004 , prog .
phys . , 112 , 241 bayer j. , dyer c. c. , giang d. , 2006 , gen .
38 , 1379 beaulieu j. p. , et al . , 2006 , nature , 439 , 437 bond i. a. , et al . , 2004 , apj , 606 , l155 borgeest u. , 1983 , a&a , 128 , 162 bourassa r. r. , kantowski r. , norton t. d. , 1973 , apj , 185 , 747 bourassa r. r. , kantowski r. , 1975 , apj , 195 , 13 bozza v. , 1999 , a&a , 348 , 311 bray i. , 1984 , mnras , 208 , 511 gaudi b. s. , et al . , 2008 , science , 319 , 927 gould a. , loeb a. , 1992 , apj , 396 , 104 jeong y. , han c. , park s. , 1999 , apj , 511 , 569 khavinson d. , neumann g. , 2006 , proc .
, 1077 khavinson d. , neumann g. , 2008 , not . amer .
55 , 666 mao s. , paczynski b. , 1991 , apj , 374 , 37l mao s. , petters a. , witt h. j. , 1997 , in _ proceedings of the eighth marcel grossmann meeting on general relativity _
r. ruffini , ( singapore , world scientific ) astro - ph/9708111 miyamoto m. , yoshii y. , 1995 , aj , 110 , 1427 petters a. o. , levine h. , wambsganss j. , 2001 , _ singularity theory and gravitational lensing _
( boston , birkhuser ) refregier a. , 2003 , araa , 41 , 645 rhie s. h. , 2001 , arxiv : astro - ph/0103463 rhie s. h. , 2003 , arxiv : astro - ph/0305166 safizadeh n. , dalal n. , griest k. , 1999 , apj , 522 , 512 schneider p. , weiss a. , 1986 , a&a .
164 , 237 schneider p. , ehlers j. , falco e. e. , 1992 , _ gravitational lenses _
( heidelberg , springer - verlag ) schneider p. , 2006 , _ extragalactic astronomy and cosmology : an introduction _ ( heidelberg , springer - verlag ) schramm t. , 1990 , a&a , 231 , 19 van der waerden b. l. , 1966 , _ algebra i _
( heidelberg , springer - verlag ) walker m. a. , 1995 , apj , 453 , 37 wambsganss j. , 2006 , in _ proceedings of the 33rd saas - fee advanced course _ ( heidelberg , springer - verlag ) : arxiv : astro - ph/0604278 weinberg s. , 2008 , _ cosmology _ ( oxford , oxford univ . press ) witt h. j. , 1990 , a&a .
236 , 311 witt h. j. , 1993 , apj , 403 , 530 witt h. j. , petters a. , 1993 , j. math .
34 , 4093 witt h. j. , mao s. , 1995 , apj . 447 , l105 | this paper makes the first systematic attempt to determine using perturbation theory the positions of images by gravitational lensing due to arbitrary number of coplanar masses without any symmetry on a plane , as a function of lens and source parameters .
we present a method of taylor - series expansion to solve the lens equation under a small mass - ratio approximation .
first , we investigate perturbative structures of a single - complex - variable polynomial , which has been commonly used .
perturbative roots are found .
some roots represent positions of lensed images , while the others are unphysical because they do not satisfy the lens equation .
this is consistent with a fact that the degree of the polynomial , namely the number of zeros , exceeds the maximum number of lensed images if n=3 ( or more ) .
the theorem never tells which roots are physical ( or unphysical ) . in this paper ,
unphysical ones are identified . secondly , to avoid unphysical roots , we re - examine the lens equation .
the advantage of our method is that it allows a systematic iterative analysis .
we determine image positions for binary lens systems up to the third order in mass ratios and for arbitrary n point masses up to the second order .
this clarifies the dependence on parameters .
thirdly , the number of the images that admit a small mass - ratio limit is less than the maximum number .
it is suggested that positions of extra images could not be expressed as maclaurin series in mass ratios .
magnifications are finally discussed .
[ firstpage ] gravitational lensing
cosmology : theory stars : general methods : analytical . |
the identification of astrophysical mev and gev sources has been a long - standing question in gamma - ray astronomy , mainly due to the limited angular resolution of gamma - ray detectors .
the pioneer _
sas-2 _ and _ cos - b _ satellites reported detections of 26 sources with median location error of @xmath0 @xcite . however , only the emission from the and pulsars and the quasar could be firmly identified .
the deeper survey by egret reported 271 gamma - ray sources with a median location error of @xmath1 , but only 101 identifications were reported @xcite .
the large area telescope ( lat ) aboard the _ fermi gamma - ray space telescope _ started operations in 2008 .
the increased effective area , reduced dead time , and use of silicon tracker technology resulted in an order of magnitude improvement in source location compared to its predecessors . the second _
fermi_-lat source catalog ( 2fgl , * ? ? ?
* ) characterizes 1873 gamma - ray sources between 0.1 and 100gev with a median location error of @xmath2 .
a total of 1297 sources in the 2fgl are either identified through variability or morphology , or reliably associated with counterparts from catalogs of candidate gamma - ray - emitting source classes .
the remaining 576 sources for which no counterpart was identified are left unassociated .
[ b ] .list of source classes in the 2fgl catalog . [ cols="^,^,^,^",options="header " , ]
this work studied the sample of unassociated gamma - ray sources in the _
fermi_-lat 2fgl catalog , finding 231 agn candidates based on their gamma - ray properties .
two independent machine - learning algorithms ( random forest and neural networks ) were used to asses the likelihood of each source to be an agn , and intersected to add robustness to the classification method and reduce the number of false associations .
the study includes for the first time an estimate of the _ false - association rate _ that takes into account the differences between associated and unassociated gamma - ray sources .
by evaluating the performance using a test sample weighted in significance and galactic latitude , the obtained 11% _ false - association rate _ can be considered a realistic estimate of the fraction of spurious sources present in the agn - candidate list .
@xcite obtained a lower _ false - association rate _
directly evaluated on the test sample , which is likely an optimistic performance estimate as discussed in section [ sec : performance ] .
similarly , @xcite used cross - validation on the training sample , which is known to give an optimistic performance , as the same sources used to optimize the classification method are used to calculate the _ false - association rate_. the list of agn candidates ( table [ tab : sourcelist ] ) covers the whole sky , studying for the first time the strip that covers the galactic plane , where more than 50% of the unassociated 2fgl sources are located .
about 210 agn are expected at @xmath3 extrapolating from high - latitude observations @xcite , while only 104 are listed in the 2fgl catalog .
even though low - latitude sources are harder to classify , 72 agn candidates were found at @xmath3 ( see figure [ fig : lat ] ) , which could be a significant fraction of the missing agn close to the galactic plane . at @xmath4 ,
the list of agn candidates is in good agreement with previous work by @xcite .
their study found 60 additional agn candidates , which could be a combination of their method being more sensitive , as it was trained on a cleaner sample of high - latitude sources , and a lower confidence threshold to identify agn candidates .
close to the galactic plane , agn are difficult to identify due to optical extinction and the bright foreground in radio and soft x - rays .
counterpart catalogs are usually incomplete at low latitudes or skip the galactic plane altogether ( e.g. , * ? ? ?
* ; * ? ? ?
galactic absorption for gamma rays is negligible below 10tev @xcite , making low - latitude agn detectable in the gamma - ray band but difficult to catalog at lower frequencies .
in fact , numerous identifications of agn behind the galactic plane have been triggered by gamma - ray detections at gev @xcite and tev energies @xcite .
the list of 231 candidate agn presented here can not be considered source associations , but only objects likely to be associated with agn . in case of gamma - ray - emitting agn
, detectable levels of non - thermal emission in radio , optical , and x - ray frequencies are expected , and follow - up observations in those bands are needed to unambiguously identify the nature of the gamma - ray emission .
observations in the x - ray band ( 0.2 - 10kev ) have been successful in finding counterparts of unidentified gamma - ray sources ( e.g. , * ? ? ?
the angular resolution of x - ray telescopes ( @xmath5 for _ swift_-xrt ) can resolve individual sources inside the typical @xmath6 error box of unassociated 2fgl sources .
follow - up observations in radio and optical spectroscopy of candidate x - ray counterparts can then provide a solid agn identification and spectral class ( e.g. , * ? ? ?
the _ swift _ x - ray satellite has observed a good fraction of the 2fgl unassociated sources .
a complete summary of these observations can be found in http://www.swift.psu.edu/unassociated/ @xcite .
so far , 135 out of the 231 candidate agn have at least 2ks of _ swift_-xrt exposure .
a good fraction of these have also been analyzed in @xcite , finding 85 sources with at least one point - like x - ray counterpart .
infrared counterparts with agn - like spectra have been identified for 56 of the agn candidates in the wise all - sky survey @xcite .
recently , @xcite presented a multiwavelength study of seven unassociated sources where four agn candidates were investigated and confirmed to have agn - like properties in the radio and x - ray bands .
these counterparts are listed in table [ tab : sourcelist ] .
counterpart catalogs of gamma - ray - emitting agn candidates are mostly based on the population of agn detected by egret @xcite , that contained a large number of flat - spectrum radio quasars ( fsrq , 71% ) and fewer bl lac - type objects ( 27% ) .
similar fsrq / bl lac ratios are found in counterpart catalogs such as _ cgrabs _ ( 84% fsrq/10% bl lac , * ? ? ?
* ) or _ bzcat _
( 54%/39% , * ? ? ?
however , bl lacs are more numerous than fsrq in the 2fgl catalog ( 34%/40% , see table [ tab : classes ] ) .
lat - detected bl lacs have a median radio flux density of 86mjy @xcite , with a low - flux tail extending well below the completeness limit of the _ crates / cgrabs _ catalog ( 65mjy , * ? ? ?
the potential deficit of bl lacs in counterpart catalogs suggests that a fraction of unassociated 2fgl sources might indeed be bl lac - type blazars that have not yet been cataloged .
this could become relevant in searches for tev - emitting agn with present ground - based observatories ( e.g. , * ? ? ?
* ) , and prospects for future installations like cta @xcite , as the harder gamma - ray spectra of bl lacs favor their detection at tev energies over fsrq .
follow - up studies on the agn candidates presented here will reduce the number of unassociated gamma - ray sources and yield a more complete picture of the characteristics of gamma - ray - loud agn .
additionally , future observations could prove whether the population of gamma - ray - emitting bl lacs extends to sources with low radio flux density .
if confirmed , gamma - ray emission from bl lacs with luminosities @xmath7 will give additional information on the low end of bl lac luminosity function in the gamma - ray band @xcite , which is a key ingredient to estimate their contribution to the isotropic diffuse gamma - ray background @xcite .
this work was supported by dfg ( sfb823 , sfb876 ) , daad ( ppp usa ) and hap .
me acknowledges support from nasa grant nnx12aj30 g .
the authors thank reshmi mukherjee and daniel nieto for very useful discussions on the structure of the paper and the interpretation of the results , and sabrina einecke and ann - kristin overkemping for proof - reading the manuscript .
this study used rapidminer @xcite : an open - source data mining framework developed at technische universitt dortmund and maintained and distributed by rapid - i . | the second _ fermi_-lat source catalog ( 2fgl ) is the deepest all - sky survey available in the gamma - ray band .
it contains 1873 sources , of which 576 remain unassociated .
machine - learning algorithms can be trained on the gamma - ray properties of known active galactic nuclei ( agn ) to find objects with agn - like properties in the unassociated sample .
this analysis finds 231 high - confidence agn candidates , with increased robustness provided by intersecting two complementary algorithms .
a method to estimate the performance of the classification algorithm is also presented , that takes into account the differences between associated and unassociated gamma - ray sources .
follow - up observations targeting agn candidates , or studies of multiwavelength archival data , will reduce the number of unassociated gamma - ray sources and contribute to a more complete characterization of the population of gamma - ray emitting agn . |
in a gravitational - wave interferometer , the thermal noise of the mirrors has a similar effect as a gravitational wave since both change the optical path followed by the light in the interferometer arms . at every radial point @xmath0 of the mirror surface
, the field experiences a local phase - shift proportional to the longitudinal displacement @xmath1 of the surface ( the origin of the cylindrical coordinates is taken at the center of the mirror surface , see figure [ fig_model ] ) .
this leads to a global phase - shift for the reflected field which is actually related to the mirror displacement averaged over the beam profile@xcite .
the phase shift between the two interferometer arms thus contains information about the mirror displacement and the variable read out by this procedure corresponds to the averaged displacement @xmath2 where the brackets stand for the overlap integral in the mirror plane ( @xmath3 ) , @xmath4 and @xmath5 is the intensity profile of the light beam in the mirror plane . assuming that the beam is in a tem@xmath6 gaussian mode , this profile is related to the beam waist @xmath7 by @xmath8 any displacement can be decomposed onto the acoustic modes of the mirror.noting @xmath9 a basis of the internal acoustic modes , the displacement @xmath10 can be expressed as a linear combination of these modes @xmath11 where @xmath12 is the time - dependent amplitude of mode @xmath13 .
each acoustic mode corresponds to a harmonic oscillator characterized by a lorentzian mechanical susceptibility @xmath14 = \frac{1}{m_{n}\left ( \omega _ { n}^{2}-\omega ^{2}-i\omega _ { n}^{2}\phi \left [ \omega \right ] \right ) } , \label{eq_chin}\ ] ] where @xmath15 is the effective mass of mode @xmath13 , @xmath16 is its resonance frequency and @xmath17 $ ] is the loss angle assumed to be the same for all modes .
the mirror motion can be described by the fourier transforms @xmath18 $ ] of every amplitude coefficient .
assuming that the mirror is in thermal equilibrium at temperature @xmath19 , one gets @xmath20 = \chi _ { n}\left [ \omega \right ] f_{t , n}\left [ \omega \right ] , \label{eq_an}\ ] ] where @xmath21 is a langevin force describing the coupling of the mode @xmath13 with the thermal bath
. one can now determine the averaged displacement @xmath22 from eqs .
( [ eq_meanu ] ) , ( [ eq_decomp ] ) and ( [ eq_an ] ) . one gets@xcite @xmath23 = \chi _ { eff}\left [ \omega \right ] f_{t}% \left [ \omega \right ] , \label{eq_uchi}\ ] ] where @xmath24 appears as an effective susceptibility taking into account all the acoustic modes and their spatial overlap with the light beam , @xmath25 = \sum_{n}\left\langle u_{n}\left ( r , z=0\right ) , v\left ( r\right ) \right\rangle ^{2}\chi _ { n}\left [ \omega % \right ] . \label{eq_chieff}\ ] ] this effective susceptibility is then the sum over all modes of the susceptibilities @xmath26 weighted by the overlap with light .
the force @xmath27 in eq .
( [ eq_uchi ] ) is an effective langevin force related to the forces @xmath21 of each acoustic mode .
one finds@xcite that the noise spectrum @xmath28 $ ] of the force @xmath27 is related to @xmath24 by the fluctuation - dissipation theorem @xmath29 = -\frac{2k_{b}t}{\omega } % % tcimacro{\func{im}}% % beginexpansion \mathop{\rm im}% % endexpansion \left ( \frac{1}{\chi _ { eff}\left [ \omega \right ] } \right ) .
\label{eq_st}\ ] ] this relation means that the mirror is in thermodynamic equilibrium at temperature @xmath19 . in a gravitational - wave interferometer ,
the frequency of a gravitational wave is usually much smaller than the mechanical resonance frequencies of internal acoustic modes of the mirrors .
we are thus interested in the noise spectrum @xmath30 of the averaged displacement @xmath22 at low frequency compared to the resonance frequencies @xmath16 . from eqs .
( [ eq_uchi ] ) to ( [ eq_st ] ) , the background thermal noise can be approximated in this frequency domain to @xmath31 & = & \frac{2k_{b}t}{\omega } % % tcimacro{\func{im}}% % beginexpansion \mathop{\rm im}% % endexpansion \left ( \chi _ { eff}\left [ \omega \right ] \right ) \nonumber \\ & \approx & 2k_{b}t\frac{\phi \left [ \omega \right ] } { \omega } \chi _ { eff}\left [ 0\right ] .
\label{eq_su}\end{aligned}\ ] ] the effect of thermal noise on light is thus proportional to the effective susceptibility at zero frequency which is given by @xmath32 = \sum_{n}\frac{\left\langle u_{n}\left ( r , z=0\right ) , v\left ( r\right ) \right\rangle ^{2}}{m_{n}\omega _ { n}^{2}}. \label{eq_chieff0}\ ] ] in the next section we determine this susceptibility for a plano - convex mirror and we compare the values obtained to the ones of a cylindrical mirror .
we consider that the mirror is coated on the plane side of a segment of sphere of radius @xmath33 and of thickness @xmath34 ( see figure [ fig_model ] ) . for simplicity
we assume that the mirror has a sharp edge on its circumference so that its mass @xmath35 and its diameter @xmath36 are related to @xmath33 and @xmath34 by @xmath37 where @xmath38 is the density of the substrate ( 2200 kg / m@xmath39 for silica ) .
the total mass @xmath35 of the mirror is an important parameter for the suspension thermal noise@xcite .
we thus choose a mass of the same order as the one of the mirrors in gravitational - wave interferometers .
we will see however that the internal thermal noise is quite insensitive to the mirror mass so that we set the mass @xmath35 to 20 kg .
all the geometrical parameters of the mirrors ( curvature radius @xmath33 , diameter @xmath36 ) can then be expressed in terms of the thickness @xmath34 .
if the thickness is much smaller than the curvature radius , the acoustic propagation equation can be solved using a paraxial approximation and one gets analytical expressions for the acoustic modes corresponding to gaussian modes@xcite . each compression mode
is defined by three integers @xmath13 , @xmath40 , @xmath41 , corresponding to longitudinal , radial and angular indexes , respectively .
the longitudinal displacement @xmath42 at radial coordinate @xmath0 and axial coordinate @xmath43 is given by@xcite @xmath44 @xmath45 is composed of a transverse gaussian structure , a transverse laguerre polynomial @xmath46 and a cosine in the propagation direction ( @xmath47 is the mirror thickness at radial position @xmath0 , equal to @xmath48 for @xmath49 ) .
the acoustic waist @xmath50 and the eigenfrequency @xmath51 are given by @xmath52 , \label{eq_onpl}\end{aligned}\ ] ] where @xmath53 is the fundamental longitudinal frequency , @xmath54 being the longitudinal sound velocity ( 5960 m / s for silica ) .
from these equations one can derive an analytical expression for the effective susceptibility @xmath55 $ ] as an infinite sum over all modes @xmath56 ( eq . [ eq_chieff0 ] ) . in the case where the light beam is centered on the mirror , only modes that have a cylindrical symmetry
will contribute .
in particular the sum over @xmath41 is limited to @xmath57 .
the effective mass of each acoustic mode and the spatial overlap with light are then given by @xmath58 we have numerically computed the effective susceptibility for different thicknesses .
figure [ fig_thick ] shows the result obtained by computing 30 different values of the thickness . for each value
, the curvature radius @xmath33 and the diameter @xmath36 of the mirror are determined according to eqs .
( [ eq_m ] ) and ( [ eq_d ] ) .
for example one gets @xmath59 cm and @xmath60 cm for a thickness @xmath34 of 7 cm .
the two curves in figure [ fig_thick ] are obtained with different optical waists @xmath7 ( 2 cm for curve _ a _ and 5.5 cm for curve _ b _ ) .
these waists correspond to the beam waists on the front and end mirrors of the virgo interferometer@xcite .
one observes a decrease of the thermal noise for a thinner mirror .
this is partly due to the fact that the mechanical resonance frequencies are increased .
it would be however difficult to use a very thin mirror since its diameter would become very large ( @xmath36 evolves as @xmath61 for small @xmath34 ) .
if we consider a reasonable thickness of 7 cm , we obtain an effective susceptibility
@xmath55 $ ] equal to @xmath62 m / n for an optical waist of 2 cm , and @xmath63 m /
n for @xmath64 cm .
these results can be compared to the values obtained for cylindrical mirrors , that is @xmath65 m / n ( @xmath66 cm ) and @xmath67 m / n ( @xmath68 cm)@xcite
. the internal thermal noise of a plano - convex mirror is thus significantly smaller , by at least a factor 4 .
if the constraint on the diameter can be relaxed , even larger noise reduction may be obtained .
we have checked the validity of the numerical calculation by plotting the effective susceptibility as a function of the number of computed modes , for example in the case of a thickness of 7 cm and an optical waist of 2 cm ( figure [ fig_precis ] ) .
this curve shows that results become valid as soon as the number of computed modes is larger than 10@xmath69 .
since the numerical calculation only deals with simple analytical expressions , it can easily be processed with a very large number of modes , such as 10@xmath70 .
figure [ fig_waist ] shows the variation of the effective susceptibility as a function of the optical waist .
the thermal noise is reduced for a wider waist . the mirror displacement is actually averaged over the beam waist.since the maximum displacement is at the center of the mirror , one gets less noise for a wide waist .
it is possible to derive a simple approximation of the thermal noise in the case of a thickness @xmath34 much smaller than the curvature radius @xmath33 .
we can then assume that the transverse acoustic modes are degenerate and we can replace the resonance frequencies @xmath71 by the value for @xmath72 ( eq . [ eq_onpl ] ) . the sum over @xmath40 in the effective susceptibility ( eq . [ eq_chieff0 ] ) is then a geometric sum and one gets a simple estimate of the susceptibility @xmath55 $ ] in terms of an optical mass @xmath73 given by@xcite @xmath74 & \approx & 1/m_{opt}\omega _
{ m}^{2 } , \label{eq_chieffapprox } \\
m_{opt } & = & \frac{12}{\pi ^{2}}\left ( \frac{\pi } { 4}\rho h_{0}w_{0}^{2}\right ) .
\label{eq_mopt}\end{aligned}\ ] ] the effective susceptibility is thus equivalent to the one of a single harmonic oscillator of resonance frequency @xmath75 and of mass @xmath76 corresponding to the mass of the part of the mirror illuminated by the light beam ( term in brackets in eq .
[ eq_mopt ] ) .
the resulting susceptibility is shown as a dashed curve in figure [ fig_waist ] and appears to be a good approximation of the computed susceptibility .
it is actually an overestimation of the effective susceptibility since the resonance frequencies @xmath71 are always larger than @xmath77 .
this result shows that the internal thermal noise depends only on a few parameters , mainly the optical waist @xmath7 and the thickness @xmath34 .
the effective susceptibility evolves as @xmath78 and one obtains a lower noise for a thin mirror and a wide optical waist , as shown in figures [ fig_thick ] and [ fig_waist ] .
we have ckecked that the thermal noise is quite insensitive to other parameters such as the curvature radius @xmath33 or the mass @xmath35 of the mirror .
we have computed the effective susceptibility for different masses by varying the curvature radius @xmath33 while the thickness @xmath34 is kept constant .
the relative variation of the thermal noise is less than 5% for masses between 5 kg and 50 kg .
note that this relation with the optical mass seems to be a specific behavior of the plano - convex geometry .
one can define an effective mass for a cylindrical mirror by a relation similar to ( [ eq_chieffapprox ] ) but this mass is no longer related to the optical mass . from the point of view of thermal noise
, the comparison between cylindrical and plano - convex geometries shows that the effective mass of a cylindrical mirror is usually larger than the optical mass , whereas the fundamental resonance frequency @xmath79 is approximately 10 times larger for the plano - convex geometry ( typically 40 khz instead of 4 khz ) .
this drastic increase of the fundamental frequency explains why the plano - convex mirror is less noisy .
we now study the effect of a misalignment of the light beam .
since the maximum thermal displacement is at the center of the mirror , moving the light beam off center must improve the noise .
however , the number of modes ( @xmath80 ) which contributes to the thermal noise rapidly increases with the misalignment . as a consequence
there is only a slight effect for cylindrical mirrors@xcite . in the case of a plano - convex mirror ,
acoustic modes are more confined around the center of the mirror and one expects a larger effect of misalignment . to efficiently compute the effective susceptibility
, we replace the laguerre polynomial in eq .
( [ eq_unpl ] ) by two hermite polynomials and we use the recurrence relations between them in the numerical calculation . the resulting susceptibility is shown in figure [ fig_align ] , for a 7-cm thick mirror and for an optical waist of 2 cm ( curve _ a _ ) and of 5.5 cm ( curve _ b _ ) .
the reduction of thermal noise with misalignment is clearly visible on those curves since a reduction factor on the order of 2 can be reached .
we have studied the internal thermal noise at low frequency of a plano - convex mirror . we have shown that it can be related to the susceptibility of an equivalent pendulum of resonance frequency equal to the fundamental resonance frequency of the mirror and of mass equal to the optical mass . as a consequence
the thermal noise mainly depends on the thickness of the mirror and on the waist of the light beam .
we have found that a thin plano - convex mirror has a lower noise than a standard cylindrical mirror of same mass . for a 7-cm thickness ,
the thermal noise is approximately 10 times lower .
using such plano - convex mirrors in gravitational - wave interferometers would thus increase the sensitivity of the interferometer in the intermediate frequency domain .
the results derived in this paper are based on a paraxial approximation which allows to describe the acoustic modes of the mirror as gaussian modes .
this approximation is valid if the diameter and the curvature radius of the mirror are large as compared to its thickness and to the optical waist .
otherwise , high - order transverse modes may become sensitive to edge effects and they may be not adequately described as gaussian modes .
these modes however do not significantly contribute to the thermal noise since the overlap with light becomes small for high - order modes . anyway , considering the large noise reduction that would be obtained by a simple change of the mirror shape , we believe that it would be of great interest to experimentally compare the background thermal noises of cylindrical and plano - convex mirrors , for example by using a high - sensitivity displacement sensor such as a high - finesse cavity@xcite
. finally , let us note that we have studied a geometry with a sharp edge on the circumference . since the acoustic modes are confined near the center of the mirror , it is possible to truncate the borders of the mirror in order to have an edge with a finite size .
this would not degrade the thermal characteristics of the mirror , but it will change its diameter and its mass . a compromise must be found in practice between the thickness which determines the amplitude of internal noise , the mass which is important for suspension thermal noise , and the diameter of the mirror . | we study theoretically the internal thermal noise of a mirror coated on a plano - convex substrate .
the comparison with a cylindrical mirror of the same mass shows that the effect on a light beam can be reduced by a factor 10 , improving the sensitivity of high - precision optical experiments such as gravitational - wave interferometers . *
pacs :* 05.40.jc , 04.80.nn , 43.40.+s thermal noise is a basic limit in many precision measurements@xcite .
for example , the sensitivity in interferometric gravitational - wave detectors@xcite is limited by the brownian motion of the suspended mirrors .
this thermally excited motion can be decomposed into suspension and internal noises .
the former refers to the motion of the center of mass of the mirrors and it has been extensively studied both theoretically@xcite and experimentally@xcite . internal noise is due to thermally induced deformations of the mirror surface and constitutes the main limitation of gravitational - wave detectors in the intermediate frequency domain .
it has thus been studied theoretically for mirrors corresponding to the virgo and ligo gravitational - wave observatories , either by a decomposition of the motion over acoustic modes@xcite or by a direct approach based on the fluctuation - dissipation theorem@xcite .
the normal - mode expansion shows that the internal thermal noise may depend on the mirror shape and on the spatial matching between light and acoustic modes .
for example , a bar - shaped cylindrical mirror is less noisy than a gong - shaped one .
a slight noise reduction has also been obtained by moving the light beam off center .
these works however deal with standard cylindrical shapes although there is no evidence that such a geometry is the best candidate to reach low thermal noise . in this paper
, we study an alternative geometry in which the mirror is coated on the plane side of a plano - convex substrate .
such a plano - convex geometry has been extensively studied for mechanical resonators with high quality factors@xcite .
acoustic modes are confined near the central axis of the resonator and their spatial structure can be described by analytical expressions similar to gaussian optical modes of a fabry - perot cavity@xcite .
we show that this geometry leads to a drastic reduction of the thermal noise .
we first recall the effect of mirror deformations on the light reflected by the mirror .
we then determine the mirror motion induced by thermal excitation and show that the effect on light at low frequency can be expressed in terms of an effective susceptibility which takes into account every acoustic mode and its coupling to the light .
we compare the effect of thermal noise to the one obtained with a cylindrical mirror of the same mass .
we finally determine the noise reduction obtained by moving the light beam off center and we show that a global noise reduction of 10 can be reached . |
small size , complex structure and extreme openness or complete absence of the cavity are characteristic features of a number of `` exotic '' laser systems that have attracted the attention of physicists in recent years @xcite
. examples of lasers that fall into this category are chaotic microcavity @xcite and random @xcite lasers .
these systems are quite different from `` traditional '' cavity lasers composed of an amplifying medium in a high quality - factor cavity @xcite . from the theoretical point of view
, the very strong coupling to the external world requires a special treatment that is different from what can be found in standard laser textbooks @xcite .
a semiclassical model of lasing in open complex or random media was developed by treci _
this theory was successfully applied to understand lasing in random media @xcite . to develop the quantum theory of an open - cavity laser ,
one starts by facing the problem of quantization of the electromagnetic field in a space that can not be separated into `` system '' and `` bath '' parts unambiguously .
this problem was solved by hackenbroich _
@xcite who also put forward langevin and master equations to describe the dynamics of modes in open resonators @xcite .
hackenbroich also derived heisenberg - langevin equations for an open - cavity laser , which , however , he analyzed only in the semiclassical approximation @xcite . a related problem of light emission by an atom in a lossy cavity was also considered by di fidio _
@xcite who , however , considered a single - mode field and hence did nt discuss additional features that arise from the coupling between modes due to the openness of the system .
in the present paper we use a combination of the quantization procedure of refs .
@xcite with the standard density operator approach @xcite to develop the full quantum theory of an open - cavity laser .
we compute and analyze the lasing threshold , the photon statistics ( both below and above the threshold ) , as well as the emission linewidth of a laser that has no well - defined cavity , assuming that only two modes compete for gain .
this simple two - mode model allows us to capture some of the essential features of cooperative mode dynamics that seems to determine the behavior of the system .
we compare our results with those known from the standard laser theory @xcite and highlight common features as well as important differences .
it is worthwhile to note that a different master equation for a random laser was previously proposed by florescu and john @xcite .
these authors considered the random laser as a collection of low quality - factor cavities , coupled by random photon diffusion .
in contrast to this work , our approach has the advantage of not relying on any particular model of wave transport , as well as being based on a well - defined quantization procedure for the electromagnetic field and a fully quantum model for the atoms providing amplification .
let us start by considering an ensemble of two - level atoms interacting with the electromagnetic field .
we divide the modes of the electromagnetic field into those that belong to the system `` atoms + field '' ( a + f ) and those that constitute the `` bath '' . in the density operator approach ,
the system a + f is described by the density operator @xmath0 . the reduced density operator @xmath1 describing the electromagnetic field ( f ) is obtained by tracing over the atomic ( a ) degrees of freedom : @xmath2 .
the dynamics of the laser is due to the competition between gain ( due to the interaction of the field with atoms ) and loss ( due to the coupling of the system a + f to the bath ) , which can be expressed in the form of the following master equation : @xmath3 here the super - operators @xmath4 and @xmath5 describe the gain and the loss , respectively
. equation ( [ eq : me ] ) is quite formal and can be written for any quantum system interacting with environment .
let us now give expressions for the two terms on the r.h.s . of eq .
( [ eq : me ] ) in the case of an open - cavity laser . a way to deal with the second term @xmath6
was proposed in a series of papers by hackenbroich _
the idea is to separate the physical space @xmath7 into two subspaces and to quantize the field in terms of the modes @xmath8 and @xmath9 of these subspaces . in the context of the laser system considered here ,
it is natural to choose the first subspace such that it contains all the atoms and has a finite volume .
the discrete modes of the first subspace will constitute our sub - system f , whereas the modes of the second subspace will make up the bath .
an equation for the density matrix of the sub - system f is derived by tracing over the degrees of freedom corresponding to the modes that belong to the bath .
this yields @xcite @xmath10 here @xmath11 and @xmath12 are annihilation and creation operators corresponding to the modes of the sub - system f , and the coefficients @xmath13 depend on the precise geometry of the system .
these coefficients were calculated for a number of particular open cavities in ref .
@xcite but may be difficult to obtain in the general case . in a random laser system , they may be treated as random variables @xcite . the essential difference between eq .
( [ eq : losses rho ] ) and the analogous equation of the standard laser theory @xcite is that the damping matrix @xmath14 is not diagonal .
this shows that the openness of the system not only leads to losses described by the diagonal elements of @xmath14 , but also induces coupling between different modes .
the strength of the coupling is given by the off - diagonal elements of the damping matrix @xmath14 .
let us now turn to the first term in eq .
( [ eq : me ] ) .
it is not specific for the open - cavity laser , so that we will follow standard approaches to derive an explicit expression for it @xcite . as the first step , we consider the jaynes - cummings hamiltonian for an atom interacting with the electromagnetic field ( we set @xmath15 in the following ) @xcite : @xmath16 here @xmath17 is the frequency of the atomic transition , @xmath18 and @xmath19 are the atomic raising and inversion operators , respectively , the states @xmath20 and @xmath21 are the ground and excited states of the two - level atom , @xmath22 are the frequencies of the modes of the field , and the coefficients @xmath23 describe the coupling between the atom and the mode @xmath24 of the field .
it is convenient to introduce a reference frequency @xmath25 and the detuning parameters @xmath26 and @xmath27 to write @xmath28 where @xmath29 and @xmath30 . because @xmath31 and @xmath32 commute , @xmath33 = 0 $ ] , we will work in the interaction picture where the dynamics of the system is governed by @xmath32 . in this picture ,
the time evolution of the density operator is given by the evolution operator @xmath34 $ ] : @xmath35 where @xmath36 we will now restrict our consideration to the situations when the frequencies @xmath22 of the modes are close to @xmath25 , such that @xmath37 , @xmath38 , and will proceed by setting @xmath39 : @xmath40 here @xmath41 and @xmath42 .
the newly defined operator @xmath43 obeys the standard bosonic commutation relation @xmath44 = 1 $ ] .
after some algebra , with the help of the operators @xmath45^{1/2}$ ] and @xmath46^{1/2}$ ] , the evolution operator reads @xmath47 - i\left(\delta/2\right)\sin[\hat{\varphi}t]/\hat{\varphi } & -i \sin[\hat{\varphi}t]/\hat{\varphi } \hat{a}\\ -i\hat{a}^{\dagger}\sin[\hat{\varphi}t]/\hat{\varphi } & \cos[\hat{\phi}t ] + i\left(\delta/2\right)\sin[\hat{\phi}t]/\hat{\phi } \end{pmatrix } \label{eq : evol oper}\ ] ] we now assume that at the initial time @xmath48 , @xmath49 , with @xmath50 being the density operator of the atom , and that the atom is in its upper state : @xmath51 .
the density matrix of the full system `` atom + field '' at @xmath48 is then @xmath52 considering @xmath53 \pm i\left(\delta/2\right)\sin[\hat{\varphi}t]/\hat{\varphi}$ ] , equations ( [ eq : ro evol ] ) and ( [ eq : evol oper ] ) allow us to compute the density operator at arbitrary time as @xmath54/\hat{\varphi}\hat{a}\\ \ , & \,\\ -ig\hat{a}^{\dagger}\sin\left[\hat{\varphi}t\right]/ \hat{\varphi}\hat{\rho}_{f}(0)\hat{\phi}_{+}\;\ ; & g^{2}\hat{a}^{\dagger}\sin\left[\hat{\varphi}t\right]/\hat{\varphi}\hat{\rho}_{f}(0)\sin\left[\hat{\varphi}t\right]/\hat{\varphi}\hat{a}\end{pmatrix}.\ ] ] finally , the reduced density operator @xmath55 is @xmath56/\hat{\varphi}\hat{\rho}_{f}(0)\sin\left [ \hat{\varphi}t\right]/\hat{\varphi}\hat{a } \nonumber \\ & = & \hat{\lambda}(t ) \hat{\rho}_f(0 ) , \label{rhof}\end{aligned}\ ] ] where we defined a super - operator @xmath57
. equation ( [ rhof ] ) yields the evolution of the reduced density operator of the electromagnetic field interacting with a two - level atom which is initially in the excited state .
this is clearly insufficient to describe laser emission .
the first lacking ingredient stems from the fact that we want to describe an ensemble of many atoms , not just a single atom . to make use of eq .
( [ rhof ] ) in the case of many atoms , we assume that ( i ) the atoms interact with the field one after another , in sequence , and not all at a time , and ( ii ) the time of interaction of a given atom with the field @xmath58 is much shorter than the typical time @xmath59 at which the evolution of the field is calculated
. the density matrix of the field after a time @xmath60 during which the field interacted with @xmath61 atoms will be then equal to @xmath62 @xcite .
we will now introduce the second important ingredient of the laser system the pump . to model the pump in the framework of our two - level atom model
, we assume that atoms are transferred to the excited state at a rate @xmath63 by some external mechanism ( for example , via additional atomic levels that are not included in our model explicitly ) and that the probability for @xmath61 atoms to get excited during a time @xmath64 is @xmath65 , where @xmath66 , @xmath67 is the probability for a given atom to be in the excited state , and @xmath68 is the total number of atoms that can potentially participate in the lasing process ( i.e. , @xmath69 ) . the average number of excited atoms is @xmath70 .
the parameter @xmath67 describes statistics of pumping , with the limit @xmath71 ( that we will consider from here on ) corresponding to random pumping and the limit @xmath72 corresponding to a uniform ( regular ) pumping @xcite .
the density operator averaged over @xmath61 is @xcite @xmath73 \right\}^{k } \hat{\rho}_{f}(0 ) .
\label{eq : ro probab}\ ] ] to obtain a dynamic equation for @xmath1 , we take the derivative of eq .
( [ eq : ro probab ] ) with respect to time and expand the result in series in @xmath74 : @xmath75\right\}\hat{\rho}_{f}(t ) \nonumber \\ & \simeq & r [ \hat{\lambda}(\tau)-1]\hat{\rho}_{f}(t ) -\frac{r p}{2}[\hat{\lambda}(\tau)-1]^{2}\hat{\rho}_{f}(t ) .
\label{eq : deriv2 rho}\end{aligned}\ ] ] finally , we now take into account the fact that the time of interaction of a given atom with the field @xmath58 is , in fact , a random variable .
@xmath58 is finite due to the possible decay of the excited state without coupling to the modes of the electromagnetic field that make part of our sub - system f. this decay may be due , for example , to transitions involving additional atomic levels ( with or without emission of a photon ) , not included in our model . with @xmath76 being the rate of such transitions ,
the statistical distribution of @xmath58 is @xmath77 . by averaging eq .
( [ eq : deriv2 rho ] ) over this distribution , we obtain @xmath78-\frac{p}{2 } [ \hat{\lambda}(\tau)-1]^{2}\right\}\hat{\rho}_{f}(t ) .
\label{eq : rho gain fin}\ ] ] equations ( [ eq : losses rho ] ) and ( [ eq : rho gain fin ] ) provide explicit expressions for the two terms on the r.h.s . of eq .
( [ eq : me ] ) .
it is worthwhile to note that eq .
( [ eq : losses rho ] ) was derived in the schrdinger picture , whereas eq .
( [ eq : rho gain fin ] ) in the interaction picture . in the interaction picture , the general form of eq .
( [ eq : losses rho ] ) remains unchanged , except for the damping matrix that has to be transformed accordingly . in the present paper
we will not use any particular model for this matrix but will rather treat it as a free parameter , having in mind that in a random laser , for example , it is a random matrix .
under general conditions , many modes may coexist and compete for gain in an open - cavity laser .
the off - diagonal elements of the damping matrix @xmath14 couple the modes and make the single - mode regime hardly realizable .
it is worthwhile to note that this coupling is different from the coupling via interaction with the atomic subsystem which is always present and hardly depends on the type of the cavity under consideration . to analyze the interaction between modes ,
we consider a model in which only two modes are taken into account .
this simple situation often allows for important insights into the dynamics of laser systems , as , for example , it was the case for the quantum - beat or correlated - emission lasers @xcite .
our model differs from the previously considered two - mode models ( see , e.g. , ref .
@xcite but also refs .
@xcite and @xcite ) by the mode coupling through both the atomic subsystem ( i.e. , the coupling due to the fact that the modes interact with the same atomic transition ) and the _ common _ bath , whereas only the first type of coupling was considered in refs . @xcite .
thus , having a common bath for the two modes is essential in our model .
the strength of the additional coupling between the modes is given by the off - diagonal elements @xmath79 and @xmath80 of the damping matrix @xmath14 .
to put accent on this new element of the model , we will focus on the dependence of our results on these off - diagonal elements in what follows . to proceed ,
we rewrite the master equation ( [ eq : me ] ) assuming the limit of @xmath71 : @xmath81/\hat{\varphi}\hat{\rho}(t)\sin\left [ \hat{\varphi}\tau\right]/\hat{\varphi}\hat{\alpha}- \hat{\rho}(t)\right ] \nonumber \\ & + & \hat{l}^{\text{(loss)}}\hat{\rho } , \label{eq : me 2modes}\end{aligned}\ ] ] where @xmath82 and @xmath83^{1/2}$ ] . to lighten the notation
, we drop the subscript `` f '' of the reduced density operator and write @xmath84 .
the limit @xmath85 corresponds to the poissonian distribution of the number of excited atoms , @xmath86 , and hence to the realistic case of random pumping by an external source .
the loss term in eq .
( [ eq : me 2modes ] ) follows from eq .
( [ eq : losses rho ] ) : @xmath87 to solve the master equation ( [ eq : me 2modes ] ) , we will work with composite modes @xmath88 defined above and @xmath89 , with the commutation relations : @xmath90 = [ \hat{\alpha},\hat{\alpha}^{\dagger } ] = 1 $ ] and @xmath91 = [ \hat{\alpha},\hat{\beta } ] = 0 $ ] ( here we assume that @xmath92 are real numbers ) .
the composite modes @xmath93 and @xmath94 are not due to the phase locking phenomenon but are simply linear superpositions of the bare modes 1 and 2 . using the modes @xmath93 and @xmath94 instead of @xmath95 and @xmath96 simplifies further analysis because only the mode @xmath93 will be actually excited in the lasing process , as we will see from the following . to simplify analytical calculations
, we will assume @xmath97 from here on .
this corresponds , for example , to open cavities considered in ref .
the density operator can be represented in the basis of fock states @xmath98 as @xmath99 equations ( [ eq : me 2modes ] ) and ( [ eq : losses 2modes ] ) yield an equation for the density matrix @xmath100 : @xmath101\rho_{n_{\alpha},n_{\beta};\ , m_{\alpha},m_{\beta } } \nonumber \\ & + & c_{1}\sqrt{(n_{\alpha}+1)(m_{\alpha}+1)}\rho_{n_{\alpha}+1,n_{\beta};\ , m_{\alpha}+1,m_{\beta}}+c_{2}\sqrt{(n_{\beta}+1)(m_{\beta}+1)}\rho_{n_{\alpha},n_{\beta}+1;\ , m_{\alpha},m_{\beta}+1 } \nonumber \\ & + & 2c_{3}\sqrt{(n_{\alpha}+1)(m_{\beta}+1)}\rho_{n_{\alpha}+1,n_{\beta};\ , m_{\alpha},m_{\beta}+1}+2c_{3}\sqrt{(n_{\beta}+1)(m_{\alpha}+1)}\rho_{n_{\alpha},n_{\beta}+1;\ , m_{\alpha}+1,m_{\beta } } \nonumber \\ & -&c_{3}\sqrt{(m_{\alpha}+1)m_{\beta}}\rho_{n_{\alpha},n_{\beta};\ , m_{\alpha}+1,m_{\beta}-1}-c_{3}\sqrt{m_{\alpha}(m_{\beta}+1)}\rho_{n_{\alpha},n_{\beta};\ , m_{\alpha}-1,m_{\beta}+1 } \nonumber \\ & -&c_{3}\sqrt{(n_{\alpha}+1)n_{\beta}}\rho_{n_{\alpha}+1,n_{\beta}-1;\ , m_{\alpha},m_{\beta}}-c_{3}\sqrt{n_{\alpha}(n_{\beta}+1)}\rho_{n_{\alpha}-1,n_{\beta}+1;\ , m_{\alpha},m_{\beta } } , \label{eq : rho nondiag}\end{aligned}\ ] ] where we defined @xmath102 , @xmath103 , @xmath104 , @xmath105 , @xmath106 and @xmath107 $ ] . in order to facilitate the comparison with the standard laser theory , we defined the coefficients @xmath108 and @xmath109 in the same way as in the book @xcite ( p. 333 ) .
when @xmath110 and @xmath111 , the coefficients @xmath112 and @xmath113 reduce to the coefficient @xmath114 of this book ( p. 255 ) and @xmath115 vanishes .
the coefficient @xmath115 vanishes also for @xmath111 and @xmath116 .
this case is somewhat special because it corresponds to two modes with the same losses and the same coupling with atoms .
this leads to @xmath117 , but still @xmath118 for @xmath119 and hence the problem does not reduce to the case of a high quality - factor cavity .
high quality - factor cavities are described by eq .
( [ eq : rho nondiag ] ) with @xmath120 .
the coupling between the bare modes 1 and 2 then arises uniquely from their interaction with the same atomic transition ( and not with the common bath ) and the composite modes @xmath93 and @xmath94 still provide a useful basis for the description of lasing ( see , e.g. , ref .
@xcite ) . for the diagonal elements of the density matrix @xmath121 and up to the second order in @xmath122
we obtain @xmath123\rho_{n_{\alpha},n_{\beta}}+\frac{8c_{3}^{2}(n_{\alpha}+1)(n_{\beta}+1)}{k_{n_{\alpha}+1,\ , n_{\beta}+1}}\rho_{n_{\alpha}+1,n_{\beta}+1 } \nonumber \\ & + & \frac{2c_{3}^{2}(n_{\alpha}+1)n_{\beta}}{k_{n_{\alpha}+1,\ , n_{\beta}}}\rho_{n_{\alpha}+1,n_{\beta}-1}+\frac{2c_{3}^{2}n_{\alpha}(n_{\beta}+1)}{k_{n_{\alpha},\ , n_{\beta}+1}}\rho_{n_{\alpha}-1,n_{\beta}+1 } \nonumber \\ & + & \left[c_{1}(n_{\alpha}+1)-\frac{4c_{3}^{2}(n_{\alpha}+1)(n_{\beta}+1)}{k_{n_{\alpha}+1,\ , n_{\beta}+1}}-\frac{4c_{3}^{2}(n_{\alpha}+1)n_{\beta}}{k_{n_{\alpha}+1,\ , n_{\beta}}}\right]\rho_{n_{\alpha}+1,n_{\beta } } \nonumber \\ & + & \left[c_{2}(n_{\beta}+1)-\frac{4c_{3}^{2}(n_{\alpha}+1)(n_{\beta}+1)}{k_{n_{\alpha}+1,\ , n_{\beta}+1}}-\frac{4c_{3}^{2}n_{\alpha}(n_{\beta}+1)}{k_{n_{\alpha},\ , n_{\beta}+1}}\right]\rho_{n_{\alpha},n_{\beta}+1 } , \label{eq : rho diag}\end{aligned}\ ] ] where @xmath124^{-2 } m_{n_{\alpha},n_{\beta}}^{-1}$ ] and @xmath125 [ 1 + \bar{\delta}^{2 } + ( b / a)(n_{\alpha}+1/2)+(b/4a)^{2}]^{-1 } + c_{1}(n_{\alpha}-1/2)+c_{2}(n_{\beta}-1/2)$ ] . finally , the equations for the probability distribution of the number of photons in the modes @xmath93 and @xmath94 , @xmath126 and @xmath127 ,
follow : @xmath128 \nonumber \\ & + & 2c_{3}^{2}(n_{\alpha}+1)\sum_{n_{\beta}}n_{\beta}k_{n_{\alpha}+1,\ , n_{\beta}}^{-1}\left[2p(n_{\alpha}+1,n_{\beta})-p(n_{\alpha}+1,n_{\beta}-1)-p(n_{\alpha},n_{\beta})\right ] , \label{eq : rho alpha}\end{aligned}\ ] ] @xmath129 \nonumber \\ & + & 2c_{3}^{2}(n_{\beta}+1)\sum_{n_{\alpha}}n_{\alpha}k_{n_{\alpha},\ , n_{\beta}+1}^{-1}\left[2p(n_{\alpha},n_{\beta}+1)-p(n_{\alpha}-1,n_{\beta}+1)-p(n_{\alpha},n_{\beta})\right ] .
\label{eq : rho beta}\end{aligned}\ ] ] in the steady - state regime , @xmath130 and eqs .
( [ eq : rho alpha ] ) and ( [ eq : rho beta ] ) can be reduced to two - term recurrence relations by using the detailed balance condition and assuming that @xmath131 . here
@xmath132 and @xmath133 denote the average photon numbers in the modes @xmath93 and @xmath94 , respectively . for @xmath134 , @xmath135 ,
the resulting equations are @xmath136 \right\ } \nonumber \\ & - & p(n_{\alpha}-1 ) \left ( \frac{a}{1 + \bar{\delta}^{2}+(b / a ) n_{\alpha } } + 2 c_3 ^ 2 \bar{n}_{\beta } k_{n_{\alpha } , \bar{n}_{\beta}}^{-1 } \right ) = 0 , \label{eq : mode alpha } \\ & & p(n_{\beta } ) \left\ { c_2 - 2 c_3 ^ 2 \left [ ( \bar{n}_{\alpha } + 1 ) k_{\bar{n}_{\alpha}+1 , n_{\beta}}^{-1 } - 2 \bar{n}_{\alpha } k_{\bar{n}_{\alpha } , n_{\beta}}^{-1 } \right ] \right\ } \nonumber \\ & - & p(n_{\beta}-1 ) \times 2 c_3 ^ 2 \bar{n}_{\alpha } k_{\bar{n}_{\alpha } , n_{\beta}}^{-1 } = 0 .
\label{eq : mode beta}\end{aligned}\ ] ] the equations ( [ eq : rho nondiag])([eq : mode beta ] ) are the main result of this work . supplemented by the normalization condition @xmath137 , they will allow us to analyze the photon statistics , the threshold , the photon number fluctuations and the linewidth of the open - cavity laser in the steady - state regime . photon number distributions @xmath138 and @xmath139 can be readily obtained by solving eqs .
( [ eq : mode alpha ] ) and ( [ eq : mode beta ] ) numerically . but even without any numerical solution , it is easy to convince oneself that the only solution of eq .
( [ eq : mode beta ] ) is @xmath140 for @xmath135 [ and hence @xmath141 by normalization ] . equation ( [ eq : mode beta ] ) does not contain any gain , only damping terms are present .
besides , the terms proportional to @xmath142 cancel each other well above threshold , i.e. for @xmath143 , and therefore the steady - state solution vanishes , i.e. @xmath144 , with the exception that @xmath145 involving the normalization condition . in contrast , eq . ( [ eq : mode alpha ] ) does have a non - trivial solution @xmath146 for @xmath147 and this solution depends on the pump rate @xmath63 .
we therefore expect that if the laser effect occurs in our system , we should look for it signatures in the behavior of the composite mode @xmath93 . in the limit of weak pump @xmath148 , we may consider the linear approximation for the laser equations , i.e. @xmath149 , and the photon number distribution resulting from eq .
( [ eq : mode alpha ] ) approaches the thermal distribution as we see in fig .
1 . the analytical solution of eq . ( [ eq : mode alpha ] ) ( with @xmath150 ) can be approximated by @xmath151 with @xmath152 , which is similar to the standard result for the single - mode laser @xcite .
( dotted curve ) below ( curve 1 ) , at ( curve 2 ) , and above ( curve 3 ) threshold .
the dashed and solid lines show the thermal and the poisson distributions corresponding to the same average photon numbers @xmath153 as curves 1 and 3 , respectively . for this figure , we fixed @xmath154 , @xmath155 , @xmath156 , @xmath157 , @xmath158 , and @xmath159 .
[ photonstatistics],scaledwidth=90.0% ] in the limit of strong pump , @xmath160 , saturation effects become important and @xmath161 .
the analytical solution of eq .
( [ eq : mode alpha ] ) tends to @xmath162 where @xmath163 , and @xmath164 can be determined from the normalization of @xmath138 .
we thus observe that the distribution of @xmath134 changes qualitatively when the pump is increased and that its limiting forms ( [ weakpump ] ) and ( [ strongpump ] ) coincide with those for the single - mode laser @xcite .
this suggests that the laser transition occurs for the composite mode @xmath93 in our two - mode model .
this is illustrated by the distributions of the photon number @xmath134 found by solving eqs .
( [ eq : mode alpha ] ) and ( [ eq : mode beta ] ) numerically that we show in fig .
[ photonstatistics ] .
we see that at low pump ( below threshold ) , the distribution is close to the thermal one : @xmath165 \right ) \exp[-n_{\alpha } \hbar \bar{\omega } /k_{b}t]$ ] , where the effective temperature @xmath166 is determined by @xmath167=a/\widetilde{c}_{1}$ ] . at strong pump ( above threshold ) , @xmath138 approaches the poisson distribution : @xmath168 \bar{n}_{\alpha}^{n_{\alpha}}/(n_{\alpha}!)$ ] , where @xmath153 is given by eq .
( [ eq : phot no alpha ] ) below .
the average photon number @xmath153 can be found from eq .
( [ eq : mode alpha ] ) as @xmath169 . far above threshold ,
the distribution of @xmath134 is strongly peaked around @xmath153 and @xmath170 , as well as @xmath171 .
together with @xmath172 this yields @xmath173 hence , the threshold for the composite mode @xmath93 is given by the condition @xmath174 .
as a function of the pumping rate @xmath63 .
the off - diagonal elements of the symmetric matrix @xmath14 for the four curves are @xmath110 , @xmath175 , @xmath176 and @xmath177 , respectively .
other parameters are as in fig .
[ photonstatistics ] .
the inset shows the dependence of the threshold on @xmath178 .
[ averagen],scaledwidth=90.0% ] the dependence of the threshold pumping rate @xmath63 on the off - diagonal element @xmath79 of the damping matrix @xmath14 is shown in the inset of fig .
[ averagen ] .
in contrast , the mode @xmath94 does not have a threshold and the number of photons in it is always equal to zero .
the full dependence of @xmath153 on the parameters of the problem can be obtained by solving eq .
( [ eq : mode alpha ] ) numerically . in fig
. [ averagen ] we show the dependence of @xmath153 on the pumping rate @xmath63 for different values of the off - diagonal element @xmath79 of the damping matrix @xmath14 . figures [ photonstatistics ] and [ averagen ] show that the composite mode @xmath93 behaves as a lasing mode with a well - defined threshold .
this is also highlighted by the formal equivalence of eq .
( [ eq : phot no alpha ] ) and the standard expression for the average photon number in a single - mode laser @xcite . a common way to characterize fluctuations of the photon number in a mode of the electromagnetic field
is to compute the so - called mandel parameter , @xmath179 as a function of the pumping rate , for the same parameters as in fig .
[ averagen ] .
[ g2],scaledwidth=90.0% ] equations ( [ weakpump ] ) and ( [ strongpump ] ) readily allow us to compute this quantity for the composite mode @xmath93 analytically well below and far above threshold , respectively , as @xmath180 another quantity which is often used to characterize fluctuations of the photon number in experiments ( see , e.g. , ref .
@xcite ) is the second - order correlation function @xmath181 .
its value ranges from @xmath182 for thermal light to @xmath183 for coherent laser light .
hence , the dependence of @xmath184 on the pumping rate shown in fig .
[ g2 ] allows one to identify the laser transition quite clearly .
information about the frequency and the linewidth of light emitted by the two - mode open - cavity laser can be extracted from the off - diagonal elements of the density matrix , eq .
( [ eq : rho nondiag ] ) @xcite . defining @xmath185
, we follow ref . @xcite and use an ansatz @xmath186 .
when we insert this into eq .
( [ eq : rho nondiag ] ) , it follows that , up to the lowest non - vanishing order in @xmath187 ( valid for @xmath188 ) , @xmath189 the linewidth of the laser emission corresponding to the composite mode @xmath93 is given by the real part of @xmath190 , @xmath191 .
\label{eq : linewidth d}\ ] ] for @xmath192 and @xmath193 the linewidth reduces to @xmath194 .
this formally coincides with the result of the standard laser theory @xcite , except for the definition of @xmath112 , which includes an additional term , @xmath195 , in our case of the open - cavity laser . in fig .
[ linewidth ] , we plot the dependence of the linewidth of mode @xmath93 on the pumping rate @xmath63 for @xmath63 at least 25% above threshold .
the dependence of the linewidth on the off - diagonal element of the damping matrix @xmath14 is shown in the inset for @xmath196 , which is far above threshold for the range of @xmath79 shown in the figure . as a function of the pumping rate @xmath63 for the same parameters as in fig .
[ averagen ] .
the inset shows the dependence of the linewidth on @xmath178 for @xmath196 .
[ linewidth],scaledwidth=90.0% ] as in the usual single - mode laser with a cavity of high quality factor @xcite , the finite linewidth of our two - mode open - cavity laser is due to the spontaneous emission of the active atoms .
previous studies have shown that in an open cavity , the emission line is further broadened by a factor @xmath68 , called the petermann factor @xcite , due to the nonorthogonality of the cavity modes @xcite . however , our analysis here is based on the quantization procedure of refs .
@xcite that relies on the expansion of the electromagnetic field inside and outside of the cavity in terms of orthogonal modes .
it is therefore interesting to check if our model reproduces the large petermann factor expected from previous studies @xcite .
we define @xmath197 where the linewidth @xmath198 is considered as a function of @xmath79 that quantifies the openness of the system in our model . already from fig .
[ linewidth ] we see that @xmath198 increases with @xmath79 .
this is confirmed by fig .
[ kfactor ] where @xmath68 is shown as a function of @xmath79 far above threshold .
a simple analytical expression for @xmath68 can be obtained at @xmath199 and for @xmath200 : @xmath201 \left[a - c_1(0 ) \right]}{c_1(0 ) \left[a - c_1(\gamma_{12 } ) \right ] \left[a+c_1(0 ) \right]}. \label{eq : kasymp}\ ] ] because @xmath202 , we find that the growth of @xmath68 with @xmath79 is roughly linear , which is in agreement with fig .
[ kfactor ] .
thus , our model reproduces the increased petermann factor in open laser systems , despite the orthogonality of the basis in which we quantized the electromagnetic field . as a function of @xmath203 for the same values of parameters as in fig .
[ averagen ] and large pumping rate @xmath196 .
[ kfactor],scaledwidth=90.0% ] finally , the imaginary part of @xmath204 yields the shift of the laser frequency with respect to @xmath25 : @xmath205 this equation shows that the frequency shift depends on the average photon number @xmath206 which , in turn , is a function of the off - diagonal element of the damping matrix @xmath14 .
we developed the quantum theory of a laser with an open cavity .
the openness of the cavity is mathematically described by a nondiagonal damping matrix @xmath14 . assuming that only two modes of the `` cold '' cavity are allowed to participate in the competition for gain
, we have shown that the modes strongly interact with each other and that a composite mode ( denoted by @xmath93 here ) is built up .
this composite mode shows all the properties of a typical laser mode : threshold behavior , photon statistics evolving from thermal to poissonian as the pumping rate increases , augmented photon number fluctuations in the vicinity of the threshold , and linewidth narrowing . at the same time
, the precise behavior of the composite mode @xmath93 at given values of parameters depends explicitly on the off - diagonal element @xmath79 of the damping matrix @xmath14 .
more precisely , an increase of @xmath79 rises the lasing threshold and broadens the laser emission line .
an important result that can not be obtained from a semi - classical theory and needs the quantum theory developed in this paper to be understood is the broadening of the emission linewidth due to the coupling of the cavity modes through a common bath ( fig .
[ kfactor ] ) .
this broadening should be accessible experimentally : indeed , measurements of the laser linewidth in a laser with two coupled modes were already performed in , e.g. , ref .
it is interesting to note that in that work the mechanism of coupling was different and led to narrowing of the emission linewidth due to the so - called correlated spontaneous emission in a system of two excited levels that provide inversion for lasing on transitions sharing a common lower level ( see ref .
@xcite for the theoretical model ) .
in contrast to this , the two - mode open - cavity laser considered in this paper exhibits widening of the emission linewidth .
this would certainly limit applications of such a laser for high - precision measurements but might be beneficial for other applications where low - coherence light is required ( like , e.g. , optical coherence tomography @xcite ) .
one of the possible applications of our theory may lie in the field of random lasers . in this case , the damping matrix @xmath14 should be treated as a random matrix and our results should be averaged over the statistical distribution of its elements .
however , to obtain results that can be directly applied to random laser systems , one has to generalize our analysis to the multi - mode case because the number of active modes in random lasers is expected to be large @xcite . | we develop the quantum theory of an open - cavity laser assuming that only two modes compete for gain .
we show that the modes interact to build up a composite mode that becomes the lasing mode when pumping exceeds a threshold .
this composite mode exhibits all the features of a typical laser mode , whereas its precise behavior depends explicitly on the openness of the cavity .
we approach the problem by using the density - matrix formalism and derive the master equation for the light field .
our results are of particular interest in the context of random laser systems . |
astrophysical black holes are characterised by just mass and spin
. the measurement of spin requires observations which probe the immediate environment of the black hole event horizon .
two methods which have emerged using current x - ray observations of luminous accreting black holes rely on fitting either the shape of the reflection component of the spectrum including a broad iron line , or the shape and the flux of the quasi - blackbody continuum .
both assume that the accretion disc extends in to the innermost stable circular orbit ( isco ) , the radius of which is determined by the black hole spin .
the reflection fitting method essentially measures the largest gravitational redshift of the disc , which comes from the isco , and yields that radius in units of gravitational radii ( @xmath0 ) .
the continuum fitting method measures the area of the inner disc and obtains the radius of the isco in km . here
we apply the reflection method to suzaku x - ray data of the first stellar mass black hole , cyg x-1 .
early attempts to measure the spin parameter of cyg x-1 using the reflection method led to inconsistent results ( miller at al 2005 ; miller et al 2009 ) . adopting a newtonian emissivity profile ( surface flux on the disc varies with radius @xmath7 as @xmath8 with @xmath9 ) and using a special timing mode of xmm , a recent measurement by duro et al .
( 2011 ) gives a dimensionless spin parameter , where @xmath10 and @xmath11 are the angular momentum and mass of the black hole , respectively .
] of @xmath12 .
recent results using the continuum fitting method have been obtained by gou et al ( 2011 ) .
they find a near extreme black hole with @xmath13 the work relies on accurate optical determinations of the mass ( @xmath14 ) and inclination ( @xmath3 deg ) obtained by orosz et al ( 2011 ) . here
we study the reflection / iron line approach in detail , taking account of the relativistic effects on the emissivity profile expected if the black hole spins rapidly .
such effects demand that the emissivity be steep if the disc extends to the innermost region within @xmath15 ( wilkins & fabian 2011 ) .
we find @xmath16 , which is a higher than , but marginally consistent with , that reported by duro et al ( 2011 ) .
both reflection results are consistent with the continuum fitting result of gou et al ( 2011 ) .
we find the disc to be fairly highly ionized with an ionization parameter around 1800 erg cm s@xmath17 , which means that the reflection spectrum is dominated by a large edge due to ionized iron .
this blurred edge carries information about spin in the spectral fit .
early work emphasised how this edge can be important in fitting the spectrum of cyg x-1 ( ross , fabian , brandt 1996 ) .
the dataset we use is one of the 20 suzaku observations of cyg x-1 analyzed by miller et al ( 2012 ; the observation of 2009 apr 8) .
the focus of that work is not on measuring the black hole spin , but rather studying the detailed disc - jet coupling in cyg x-1 .
the spectra in that work have been fitted with a relativistically - blurred reflection model , but in order to make the work tractable , the fits kept the source inclination and iron abundance frozen at 27 deg and 1 , respectively .
the emissivity profile of the blurring function was also just a single power - law , fixed for the newtonian value of 3 .
we relax these assumptions here and apply them to an average dataset .
we later tested the model on several other datasets ( 2009 may 6 , 25 and dec 1 ) and obtained consistent results .
the best fit to the 1500 kev spectrum is shown in fig .
1 and values of key parameters are given in table 1 .
the model used , constant*tbabs*mtable(windabs.fits)(gaussian+gaussian + diskpbb+powerlaw+kdblur2f*atable(extendx.mod ) ) highecut , is shown in fig . 2 .
kdblur2f is a fast adaptation by jss of the broken power - law emissivity convolver kdblur2 .
extendx is a version of the self - consistent slab reflection model reflionx ( ross & fabian 2005 ) extended to allow the photon index to be less than 1.4 .
( we do not use the refhiden models of ross & fabian 2007 , which incorporate a blackbody component because they are only available for solar abundance iron and at the blackbody temperature found here for the low state of about 0.16 kev the differences are minimal . ) absorption features due to the stellar wind of the companion ( e.g. , hanke et al . 2009 , nowak et al . ,
2011 ) are modeled using windabs.fits ( miller et al . , 2012 ) and we refer to that paper for further details .
the absorption has little effect on our results .
we note that the highecut model used to fit the data above 100 kev is phenomenological and is not strictly consistent with the 300 kev exponential cutoff assumed by the extendx model , but has limited effect on the results presented here .
the quality of the fit is very good , and the emissivity index @xmath4 is found to be high @xmath18 .
the inclination of @xmath2 deg is consistent with the optical measurement of orosz et al ( 2011 ) .
it is clear that the spectrum is reflection - dominated below 10 kev ( fig .
2 ) , although the whole source , from an energy point of view integrating over the whole spectrum , is not reflection dominated .
this is a consequence of the hard irradiating continuum where much of the energy absorbed from compton recoil due to photons at @xmath19 is emitted by the surface layers if the disc below 6 kev .
the ratio of the 0.11000 kev flux from the reflection component is 0.70 times that from the power - law component , so the reflection fraction , calculated as the ratio of incident to emitted flux at the disc surface , @xmath20 .
the blackbody disc component is 0.4 times the power - law and if reflection is responsible for some of that emission then @xmath21 .
a narrow emission line is introduced at 6.4 kev to account for ( weak ) distant reflection ( e.g. from the stellar companion ) and a narrow absorption feature is included at 1.85 kev to compensate for problems in the model response near the detector silicon edge .
we show the model spectrum over the 265 kev band in the top panel of fig .
the inner radius has been stepped from 1.3 to 1.9 , and then to @xmath22 and the innermost emissivity index @xmath23 .
the effect of fixing @xmath4 at 3 is shown in the lower panel : the model is now less sensitive to the innermost radii .
.values of key model parameters used in the spectral fitting . [ cols="<,<,<,<",options="header " , ] [ par.tab ] a simple physical interpretation arises if the source of the steeper power - law is quasi - static and situated close to the centre of the disc , therefore producing the reflection .
the harder component could originate from further out within the moving jet ; both soft source photons from the disc and reflection produced by the continuum are reduced by beaming ( beloborodov 1999 ) .
time lags then occur if variations propagate up the jet from the inner , quasistatic and softer component to the outer , jetted harder component .
they need not propagate at jet speed , but could be due to changes in the magnetic structure or collimation of of the jet .
the power in these components and the jet is presumed to derive from the accretion disc and be transferred by magnetic fields coupled to a range of radii , so varying in a manner which is consistent with the lyubarskii model .
the lack of rapid reflection variations reported by revnivtsev et al 1999 ) could in part be due to light bending effects implicit in the proximity of the inner , softer , power - law component to the black hole ( miniutti & fabian 2004 ) , if some of the rapid variability is due to ( small ) changes in height of that inner component .
further development of such models is beyond the scope of this paper , which is focussed on the low state x - ray spectrum of cyg x-1 .
we have used the x - ray reflection fitting method to obtain a robust measurement of the dimensionless spin parameter of the black hole in cyg x-1 of @xmath16 .
our result agrees with the continuum fitting value of gou et al ( 2011 ) and yields an even higher spin than the reflection measurement of duro et al ( 2011 ) .
cygnus x-1 definitely appears to have high spin .
we argue that non - newtonian values must be used for the inner emissivity profile of the reflected emission of the innermost disc when the spin is high ( @xmath24 and @xmath25 ) .
the reflection fitting method relies on detection of the innermost radius of the disc which is dense enough for reflection to occur .
we assume that this is the isco , for within that radius the plunge orbits mean that the gas density drops rapidly so the reflection signal vanishes .
although magnetic fields may have some effect here , we consider it reasonable to assume that the dense parts of the disc that give the reflection spectrum are not seriously affected ( see discussion and simulations in reynolds & fabian 2008 and shafee et al 2008 ) .
there is nevertheless a systematic uncertainty here which simulations need to resolve . for the fit using the broken power - law is about 150 below that for the fit using a single power - law emissivity profile .
the results in table 1 were obtained with the outer index free . ]
the reflection fraction @xmath26 is low compared with the simple expectation of a source close to the black hole at @xmath27 .
light bending should give @xmath28 ( e.g. fabian et al 2011 ) . the observed radio emission
( miller et al 2012 ) means that a jet is operating , so it is plausible that the emission model is more complex , as discussed in section 3 .
we suggest there that irradiation of the disc by the quasi - static base of the jet at a @xmath29 gravitational radii is responsible for the reflected emission , whereas the observed power - law continuum is dominated by higher , faster parts of the jet which mildly beam emission in our direction .
systematic uncertainties in the models and ( relative ) spectral calibration can also affect the details of our fits and thus results , particularly of the innermost region .
the continuum fitting method is also subject to systematic uncertainties in both models and spectral calibration ( relative and absolute ) .
wide agreement on the high spin of cygnus x-1 boosts confidence in spin measurements from the reflection fitting method of other objects as well as agn , where it is the only method yet available .
the spectral fits strongly indicate that the innermost radius of the disc lies within @xmath6 .
relativistic effects are strong there and a simple newtonian assumption for the emissivity profile is inappropriate .
the effects are inescapable and involve both special relativity , since the velocity of the reflecting matter on the disc is high , and general relativity , through gravitational redshift and light bending ( wilkins & fabian 2010 , 2012 ) .
theoretical emissivity profiles due to point sources at different heights on the rotation axis above the disc plane are shown in fig .
12 of fabian et al ( 2011 ) .
when the source height exceeds @xmath5 , the emissivity profile breaks at radius @xmath30 to the newtonian value of 3 outside and flatter within .
the inner profile rapidly steepens to a high value within @xmath6 , if the disc extends in that far .
we recommend that @xmath9 be used initially .
should this lead to high spin being suspected then a power - law broken at least once , with a high inner index ( @xmath31 ) , should be used .
this is illustrated in fig .
10 where we have fitted our cyg x-1 data by a single power - law emissivity of free index .
it shows the best fit being of high index and relatively low spin ( @xmath32 ) , which is rejected by the physical sense argument of duro et al ( 2011 ) .
if instead we use a broken power - law emissivity fixing the outer index @xmath33 , then we find that the inner emissivity @xmath34 remains high and the spin rises to be very close to maximal ( @xmath35 ) .
this now makes physical sense .
allowing the outer index to be free gives the results of table 1 .
we thank the referee for comments leading to section 3 .
acf thanks the royal society for support .
rcr thanks the michigan society of fellows and nasa for support through the einstein fellowship program , grant number pf1 - 120087 .
arvalo p , uttley p. 2006 ,
mnras , 267 , 801 belloni t. , hasinger g. , 1990 , a&a , 227 , l33 beloborodov a.m. , 1999 , apj , 510 , 123 dauser t. , wilms j. , reynolds c.s .
, brenneman l.w . , 2010 , mnras , 409 , 1534 dauser t. , et al .
arxiv:1112.1796 duro r. , et al 2011 , a&a , 533 , l3 fabian a.c .
, et al 2011 , mnras , 419 , 116 gierliski m. et al 1997 , mnras , 288 , 958 gilfanov m. , churazov e. , revnivtsev m. , 2000 , mnras , 316 , 923 gou l. , et al 2011 , apj , 742 , 85 hanke m. , wilms j. , nowak m. a. , pottschmidt k. , schulz n. s. , lee j. c. 2009 .
apj , 690 , 330 kawabata r. , mineshige s. , 2010 , pasj , 62 , 621 lyubarskii y. , 1997 , mnras , 292 , 679 makishima k. , et al 2008 , pasj , 60 , 585 miller j.m . , wojdowski p. , schulz n.s . , marshall h.l .
, fabian a.c . ,
remillard r.a . , wijnands r. , lewin w.h.g . , 2005 ,
apj , 620 , 398 miller j.m .
, reynolds c.s . , fabian a.c . ,
miniutti g , gallo l , c .
, 2009 , apj , 697 , 900 miller j.m .
, pooley g.g .
, fabian a.c . ,
nowak m.a . ,
, cackett e.m .
, pottschmidt k. , wilms j. , 2012 , apj , submitted miyamoto s. , kitamoto s. , mitsuda k. , dotani t. , 1988 , nature , 450 miyamoto s. , kitamoto s. , 1989 , nature , 342 , 773 novikov i.d .
, thorne k.s . , 1973 , in black holes , ed .
c dewitt and b. dewitt , new york , gordon and breach nowak m. a. vaughan b. , wilms , j. , dove j.b .
, begeman m.c . , 1999 ,
apj , 510 , 874 nowak m. a. , et al .
apj , 728 , 13 orosz j.a .
, mcclintock j.e . , aufdenberg j.p . ,
remillard r.a . ,
, narayan r. , gou l. , 2011 , apj , 742 , 84 page d.n .
, thorne k.s . , 1974 , apj , 191 , 499 poutanen j. , 2001 , adv space res , 28 , 267 pottschmidt k. et al 2003 , a&a , 407 , 1039 shafee r. , et al 2008 , apj , 687 , l25 reis r.c .
, fabian a.c . , ross r.r . , miniutti g. , miller j.m . , reynolds c.s . , 2008 , mnras , 387 , 1489 revnivstev m. , gilfanov m. , churazov e. , 1999 , a&a , 347 , l23 reynolds c.s .
, fabian a.c . , 2008 , apj , 675 , 1048 ross r.r .
, fabian a.c . , 2005 , mnras , 358 , 211 ross r.r . , fabian a.c .
, brandt w.n . , 1996 , mnras , 278 , 1082 wilkins d.r . , fabian a.c . , 2011 , mnras , 414 , 1269 wilkins d.r . , fabian a.c .
, mnras submitted | the spin of cygnus x-1 is measured by fitting reflection models to suzaku data covering the energy band 0.9400 kev .
the inner radius of the accretion disc is found to lie within 2 gravitational radii ( @xmath0 ) and a value for the dimensionless black hole spin is obtained of @xmath1 .
this agrees with recent measurements using the continuum fitting method by gou et al . and of the broad iron line by duro et al .
the disc inclination is measured at @xmath2 deg , which is consistent with the recent optical measurement of the binary system inclination by orosz et al of @xmath3 deg .
we pay special attention to the emissivity profile caused by irradiation of the inner disc by the hard power - law source .
the x - ray observations and simulations show that the index @xmath4 of that profile deviates from the commonly used , newtonian , value of 3 within @xmath5 , steepening considerably within @xmath6 , as expected in the strong gravity regime .
ps . accretion , accretion discs black hole physics line : profiles x - rays : general |
for @xmath7 , the hahn polynomials are explicitly given by @xmath8 see @xcite and @xcite .
these polynomials are orthogonal on the discrete set @xmath9 with respect to the weight function @xmath10 more precisely , we have @xmath11 where @xmath12 see @xcite .
formula tells us that these polynomials are orthogonal on an unbounded interval as @xmath13 .
we now introduce a rescaling so that these polynomials become orthogonal on a bounded interval .
let @xmath14 denote the set defined by @xmath15 the @xmath16 s are called _ nodes _ and they all lie in the interval @xmath5 . also , let @xmath17 and @xmath18 it is readily seen that the polynomials @xmath19 are orthogonal on the nodes @xmath16 with respect to the weight @xmath20 ; that is , @xmath21 where @xmath22 .
for further properties and applications of the hahn polynomials , we refer to beals and wong @xcite and karlin and mcgregor @xcite . in 1989
, sharapudinov @xcite studied the asymptotic behavior of @xmath0 , when the degree @xmath1 is substantially smaller than @xmath23 .
his main result is an asymptotic formula for @xmath0 with @xmath24 as @xmath25 ; which involves the jacobi polynomial @xmath26 , where @xmath27 is related to @xmath28 via the formula @xmath29 .
recently , baik et al .
@xcite used the riemann - hilbert approach to investigate the asymptotics of discrete orthogonal polynomials with respect to a general weight function .
their results are very general , but it is difficult to use them to write out explicit formulas for specific polynomials . with regard to the hahn polynomials , they only considered the case of varying parameters ; that is , @xmath30 and @xmath31 , where @xmath32 and @xmath33 are fixed positive numbers .
moreover , their results are more local in nature ; that is , one needs more asymptotic formulas to describe the behavior of these polynomials in the complex plane .
the purpose of this paper is to study uniform asymptotic behavior of the hahn polynomials as @xmath34 , when the parameters @xmath2 and @xmath3 are fixed and the ratio @xmath35 is a constant @xmath36 .
our approach is based on a modified version of the riemann - hilbert method introduced by deift and zhou @xcite .
this version has been used successfully to obtain globally uniform asymptotic expansions of several discrete orthogonal polynomials ; see @xcite , @xcite and @xcite .
the presentation of this paper is arranged as follows : in section 2 , we review some preliminaries , including weak asymptotics of the zero distribution and the formulation of the basic interpolation problem @xmath37 . in section 3 , we transform this interpolation problem into an equivalent riemann - hilbert problem ( rhp ) . in section 4 ,
we give the matrix transformation @xmath38 that normalizes the rhp for @xmath39 presented in section 3 by using a function @xmath40 , which is related to the logarithmic potential ( g - function ) of the equilibrium measure .
the auxiliary functions @xmath41 and @xmath42 are also studied in section 4 . in section 5 ,
we factorize the jump matrix in the rhp for @xmath38 .
the final results are presented in sections 6 and 7 , where we divide the whole complex plane into three regions and in each region we derive a uniformly asymptotic formula . as a special case of our result , we give in section 8 the asymptotic formula of @xmath0 for fixed values of @xmath28 .
it is known that the zero distribution of discrete orthogonal polynomials is related to a constrained equilibrium problem for logarithmic potential with an external field @xmath43 , which is defined by the formula @xmath44 where @xmath45 and @xmath46 are real - analytic functions in a neighborhood of the closed interval @xmath47 $ ] , and @xmath46 is also a density function of the nodes ; see @xcite and @xcite . in our case , @xmath48 , @xmath49 and @xmath50=[0,1]$ ] . to see how the function @xmath45 is derived , we recall the nodes @xmath51 and the weights @xmath52 given in and , respectively , and write @xmath53 from
, an explicit formula can be given for @xmath54 .
indeed , since @xmath55 and @xmath56 we have @xmath57 since @xmath58 , by replacing @xmath51 and @xmath1 by @xmath28 and @xmath59 , respectively , in the above equation , we obtain @xmath60 for @xmath61 . using stirling s formula
, it can be shown that @xmath62 as @xmath63 . from
, we also have @xmath64 thus , it follows from that @xmath65 as @xmath66 .
this approximation formula suggests the definition of the function @xmath67 and we have @xmath68 observe that in our case , we also have @xmath69 .
the measure @xmath70 that minimizes the following energy functional @xmath71=-c \int_0 ^ 1 \int_0 ^ 1 \ln |x - y| \mu(x ) dx \mu(y ) dy + \int_0 ^ 1 \varphi(x ) \mu(x ) dx,\end{aligned}\ ] ] where @xmath72 satisfies the upper and lower constraints @xmath73 and the normalization condition @xmath74 is called the _ equilibrium measure_. it follows from ( [ mu_con ] ) that there is an upper bound for @xmath72 , which does not appear in the continuous case .
this fact is a crucial difference between discrete orthogonal polynomials and continuous orthogonal polynomials .
the equilibrium measure @xmath70 divides the interval of orthogonality into subintervals of three possible types : ( 1 ) achieving the lower constraint ; ( 2 ) attaining the upper constraint ; ( 3 ) not reaching the upper or the lower constraints .
we call the open intervals of type ( 1)-(3 ) , voids , saturated regions and bands , respectively . for reference
, we refer to @xcite . in the existing literature ,
the equilibrium measure is usually obtained by solving a minimization problem of a certain quadratic functional ( cf .
@xcite , @xcite ) . here , we prefer to use the method introduced by kuijlaars and van assche @xcite .
following @xcite , one can construct the equilibrium measure corresponding to the hahn polynomials .
recall @xmath75 , and let @xmath76 the density function is given by @xmath77 \cup [ b,1 ] .
\end{array}\right . \label{mu}\ ] ] observe that the equilibrium measure of the hahn polynomials corresponds to the saturated - band - saturated configuration defined in @xcite , and we also note that @xmath78 we begin with the basic interpolation problem for the hahn polynomials .
note that the leading coefficient of @xmath19 is @xmath79 hence , the monic hahn polynomials @xmath80 are given by @xmath81 clearly , they satisfy the new orthogonality relation @xmath82 where @xmath83 .
following @xcite , we construct the interpolation problem for a @xmath84 matrix - value function @xmath37 with the properties : 1 .
@xmath37 is analytic for @xmath85 ; 2 . at each @xmath86 , the second column of @xmath87
has a simple pole where the residue is @xmath88 3 .
as @xmath89 , @xmath90 by the well - known theorem of fokas , its and kitaev @xcite ( see also baik et al .
@xcite ) , we have the unique solution to the above interpolation problem is given by @xmath91
in this section , we will make a sequence of transformations : @xmath92 and transform the basic interpolation problem into an equivalent riemann - hilbert problem ( rhp ) .
first , we want to remove the saturated regions .
following baik et al .
@xcite , we introduce the first transformation @xmath93 it is readily verified that @xmath94 is a solution of the interpolation problem : 1 .
@xmath94 is analytic for @xmath85 ; 2 . at each @xmath95 , the second column of @xmath94 is analytic and the first column of @xmath94 has a simple pole where the residue is @xmath96 3 .
as @xmath89 , @xmath97 next , we shall employ an idea of ou and wong @xcite to remove the poles in @xmath94 , and transform this interpolation problem into an equivalent rhp ; see also @xcite .
let @xmath98 be a sufficiently small number , and we define @xmath99 for @xmath100 , and @xmath101 for @xmath102 , where the domains @xmath103 are shown in figure [ figure_r ] .
let @xmath104 be the boundary of @xmath105 in the upper half - plane , and @xmath106 be the mirror image of @xmath104 in the lower half - plane .
for the contour @xmath107 , see also figure [ figure_r ] . and
the contour @xmath108 . ] for each @xmath95 , the singularity of @xmath39 at @xmath16 is removable ; that is , @xmath109 .
[ r_lem ] for @xmath110 , it follows from ( [ r_h_ort ] ) that @xmath111 since @xmath112 is analytic by ( @xmath113 ) , the residue of @xmath114 at @xmath16 is zero . from
, we also have @xmath115 on the other hand , it is readily seen that @xmath116 thus , applying ( [ h_11 ] ) and ( [ h_12 ] ) to ( [ r_11 ] ) shows that the residue of @xmath117 at @xmath118 is zero .
the entries in the second row of the matrix @xmath39 can be studied similarly .
this completes the proof of the lemma .
note that this transformation makes @xmath119 and @xmath120 continuous on the interval ( 0,1 ) . as a consequence ,
we have created several jump discontinuities on the contour @xmath108 in the complex plane .
it is easily verified that @xmath39 is a solution of the following rhp : 1 .
@xmath39 is analytic for @xmath121 ; 2 .
the jump conditions on the contour @xmath108 : for @xmath122 , @xmath123 where @xmath124 for @xmath125 , @xmath126 where @xmath127 3 . as @xmath89 ,
to normalize the condition ( @xmath129 ) in the rhp for @xmath39 , we need some auxiliary functions . the @xmath130-function is the logarithmic potential defined by @xmath131 , \label{g_def}\ ] ] where @xmath72 is the density function given in , and the so - called @xmath132-function is defined by @xmath133 , \label{phi_def}\ ] ] where @xmath134 is called the lagrange multiplier , a being the constant given in .
now , we define @xmath135 where @xmath136 in and , we view @xmath137 as an analytic function of the variable @xmath138 with a branch cut along @xmath139 . since @xmath140
we obtain @xmath141 similarly , we have @xmath142 from , it follows that @xmath143 moreover , it can be easily verified that the @xmath144-function satisfies the jump conditions @xmath145 and @xmath146 on account of ( [ g_def ] ) , ( [ g_jum1 ] ) and , one readily sees that @xmath147 can be analytically extended to @xmath148 $ ] and @xmath149 \qquad \text{as\ } z\rightarrow \infty .
\label{g_z}\end{aligned}\ ] ] hence , by adopting the convention that @xmath150 denotes the pauli matrix @xmath151 we introduce the transformation @xmath152 to normalize the behavior of @xmath39 at @xmath153 , where @xmath154 is given in .
as we have mentioned before , the density function @xmath72 reaches its upper constraint near the endpoints of the interval of the orthogonality ; see ( [ mu ] ) .
furthermore , since @xmath72 is not differentiable at the point @xmath155 , the function @xmath156 is not analytic in the neighborhood of @xmath155 , and we can not construct our global parametrix ( such as airy parametrix ) by using @xmath157 in the region which includes the critical point @xmath158 . hence , for our future analysis , a few more auxiliary functions are needed . note by that @xmath159 is analytic for @xmath160 .
this together with ( 4.5 ) evokes us to consider a modified measure @xmath161 .
furthermore , we want to find a complex function @xmath162 which satisfies the requirement @xmath163 for @xmath164 .
let @xmath165- \frac1c \log(z - z^2 ) , \label{nu}\end{aligned}\ ] ] where @xmath166 is analytic in @xmath167 $ ] and behaves like @xmath6 as @xmath168 , and where @xmath169 is an analytic function with branch cuts along @xmath170\cup[1,+\infty)$ ] . in view of the identities
@xmath171 for @xmath172 $ ] , we have @xmath173 on the other hand , the measure @xmath174 can be rewritten as @xmath175 from and , it clearly follows that @xmath176 now , we are ready to introduce the auxiliary function @xmath177\cup[a,+\infty ) , \label{pt_def}\end{aligned}\ ] ] where the path of integration from @xmath158 to @xmath6 lies entirely in the region @xmath178\cup[a,+\infty)$ ] , except for the initial point @xmath158 .
similarly , we set @xmath179\cup[1,+\infty ) , \label{ps_def}\ ] ] where the path of integration from @xmath180 to @xmath6 lies entirely in the region @xmath181\cup[1,+\infty)$ ] , except for the initial point @xmath180 .
the functions @xmath41 and @xmath42 defined above play an important role in our argument , and the following are some of their properties .
the mapping properties of @xmath41 are given by : @xmath182 ; \label{pt_map_b } \\ \re \pt_\pm ( x )
< \re \pt_\pm(1 ) , \quad \im \pt_\pm ( x)=\pm ( 1-\frac1c x ) \pi , \quad x\in(1,+\infty ) . \label{pt_map_1}\end{aligned}\ ] ] [ pro_pt_map ] from the definition in ( [ nu ] ) , we note that @xmath162 is analytic in @xmath183\cup[a , b ] \cup[1,+\infty)$ ] .
when @xmath184 , it is readily seen that @xmath185 where @xmath186 -\frac1c \log(x^2-x ) .
\label{nu1}\end{aligned}\ ] ] from and ( [ nu_pm1 ] ) , we have @xmath187 it is easily verified that @xmath188 for @xmath189 , and hence ( [ pt_map_min ] ) holds by ( [ pt_p_min ] ) . for @xmath190
, @xmath191 is analytic and it can be shown from that @xmath191 is real and positive .
thus , from it follows that @xmath192 i.e. , holds . on the other hand
, @xmath193 follows immediately from the definition in ( [ pt_def ] ) .
note that @xmath194 for @xmath164 ; see ( [ nu_pm ] ) .
hence , ( [ pt_map_con ] ) holds by ( [ pt_def ] ) .
if @xmath164 , we also have from ( [ nu_pm ] ) @xmath195 note that @xmath196 for @xmath197 by ( [ mu ] ) , and hence ( [ t3 ] ) gives ( [ pt_map_ab ] ) . for @xmath198 $ ] , it follows from ( [ pt_def ] ) that @xmath199 by a change of variable @xmath200 , it can be verified that @xmath201 applying ( [ pt_map_con ] ) and ( [ t6 ] ) to ( [ t4 ] ) yields @xmath202 thus , ( [ pt_map_b ] ) holds by ( [ pt_map_0 ] ) and ( [ t5 ] ) . for @xmath203
, we have @xmath204 where @xmath205 -\frac1c \log(x^2-x ) . \label{nu2}\end{aligned}\ ] ] from ( [ pt_map_b ] ) and ( [ nu_pm2 ] ) , it follows that @xmath206 note that @xmath207 for @xmath208 .
thus , @xmath209 and ( [ pt_map_1 ] ) holds .
-plane under the mapping @xmath41.,title="fig : " ] + ( a ) the upper half @xmath6-plane under @xmath41 -plane under the mapping @xmath41.,title="fig : " ] + ( b ) the lower half @xmath6-plane under @xmath41 a geometric interpretation of the results established above is given in figure [ pt_fig ] .
furthermore , if we let @xmath210 and @xmath211 , then for @xmath212 we have from and @xmath213 the results in the following proposition can now be proved either directly from definition ( [ ps_def ] ) , or by using and proposition [ pro_pt_map ] .
the mapping properties of @xmath42 are given by : @xmath214 ; \label{ps_map_b } \\ \ps(b)=0 , \quad \ps_\pm(a)=\pm ( \frac1c-1)\pi i ; \label{ps_map_con } \\ \re \ps_\pm(x)=0 , \quad \arg \ps_\pm(x)=\pm \frac\pi 2 , \quad x\in(a , b ) ; \label{ps_map_ab } \\ \re \ps_\pm(x)<0 , \quad \im \ps_\pm(x)=\pm ( \frac1c-1)\pi , \quad x\in[0,a ) ; \label{ps_map_0 } \\
\re \ps_\pm ( x ) < \re
\ps_\pm(0 ) , \quad \im \ps_\pm ( x)=\pm ( \frac{1-c}{c}- \frac1c x ) \pi , \quad x\in(-\infty,0 ) .
\label{ps_map_min}\end{aligned}\ ] ] [ pro_ps_map ] the images of the upper ( lower ) half of @xmath6-plane under the mapping @xmath42 is depicted in figure [ ps_fig ] .
the next proposition shows the relationship between the @xmath132-function and the @xmath215-function ( @xmath216-function ) . with @xmath157
defined in ( [ phi_def ] ) , the following connection formulas between the @xmath132-function and the @xmath215-function ( @xmath216-function ) hold : @xmath217 and @xmath218 for @xmath212 .
[ pro_pt_phi ] -plane under the mapping @xmath42.,title="fig : " ] + ( a ) the upper half @xmath6-plane under @xmath42 -plane under the mapping @xmath42.,title="fig : " ] + ( b ) the lower half @xmath6-plane under @xmath42 to establish and , we need some properties of @xmath219 .
in fact , we shall derive explicit formulas for @xmath220 and @xmath219 . from ( [ mu ] )
, we have @xmath221 integration by parts gives @xmath222 ' ds \hh
\mu(s)\log(z - s ) s \big |_{a}^{b } - \int_a^b \mu(s ) ds -\int_a^b s \log(z - s ) \mu(s ) ' ds + z\int_a^b \frac{1}{z - s } \mu(s ) ds \label{g_cal1 } .\end{aligned}\ ] ] the last integral in ( [ g_cal1 ] ) is given by @xmath223 on account of ( [ mu ] ) and ( [ ab ] ) , a direct calculation gives @xmath224 and @xmath225 applying ( [ mu_der ] ) to ( [ g_cal2 ] ) , we obtain @xmath226 similarly , one can show that @xmath227 a combination of , ( [ mu_ab ] ) , ( [ g_cal3 ] ) and gives @xmath228 -(1-\frac{2a}{c } ) \notag \\ & \quad -\frac { z } { 2\pi } \int_a^b \frac{\log(z - s ) } { s \sqrt { ( s - a)(b - s ) } } ds + \frac { 1 } { 2\pi } ( z-1 ) \int_a^b
\frac{\log(z - s ) } { ( 1-s ) \sqrt { ( s - a)(b - s ) } } ds \notag \\ & \quad + \frac { 1}{\pi } \int_a^b \frac{\log(z - s ) } { \sqrt { ( s - a)(b - s ) } } ds . \label{g_cal_osc}\end{aligned}\ ] ] let the integrals on the right - hand side of the equality be denoted by @xmath229 , @xmath230 and @xmath231 , respectively . to evaluate @xmath229
, we note that from @xcite we have @xmath232 making the change of variable @xmath233 , we obtain from ( [ wangintegral ] ) @xmath234 and @xmath235 also , note that @xmath236 see ( * ? ? ?
* lemma 1 ) .
hence , @xmath237 a straightforward calculation shows that @xmath238 \notag \\ & \
+ \frac2c z \log\left [ z+ ( z - a)^{1/2}(z - b)^{1/2}+c/2 \right]+\frac2c ( 1-z ) \log \left [ z-1 + ( z - a)^{1/2}(z - b)^{1/2}-c/2 \right ] .
\label{g}\end{aligned}\ ] ] moreover , by using , and , we have @xmath239 + note that when @xmath164 , @xmath194 ; see .
hence , it follows that @xmath240 on the other hand , if @xmath241 , we have with the aid of ( [ nu_pm1 ] ) and ( [ nu_pm2 ] ) @xmath242 to establish ( [ pt_phi ] ) , we need one more function @xmath243 defined by @xmath244 from , one can show that for @xmath61 @xmath245 from ( [ g_der_plu ] ) , it is easily verified that @xmath243 is analytic for @xmath178\cup [ a , b]\cup[1,+\infty)$ ] and satisfies the jump conditions @xmath246 and @xmath247 comparing ( [ phat_jum1 ] ) and ( [ phat_jum2 ] ) with ( [ nu_jum1 ] ) and ( [ nu_jum2 ] ) , we find @xmath248 or , equivalently , @xmath249 for @xmath250 . on account of this , @xmath251 can be analytically extended to the whole complex plane .
note that @xmath252 , and that from it can be shown @xmath253 and so @xmath254 hence , from , and we have @xmath255 \hh 0 , \qquad x\in(a , b ) .
\label{nu : phih}\end{aligned}\ ] ] using analytic continuation , it follows from ( [ phat_def ] ) @xmath256 hence , it follows from ( [ pt_def ] ) that @xmath257 now , let @xmath258 ; on account of and the definition of @xmath259 , we have @xmath260 in a similar manner , one can show that for @xmath261 @xmath262 formula ( [ pt_phi ] ) now follows from ( [ pt_ps_pro1 ] ) , ( [ pt_ps_pro2 ] ) and ( [ phi_def ] ) . formula ( [ ps_phi ] ) is obtained by coupling ( [ pt_phi ] ) and ( [ ps_pt ] ) .
we conclude this section with the mapping properties of @xmath157 in ( [ phi_def ] ) . by using proposition [ pro_pt_map ] and ( [ pt_phi ] ) , one can prove the following proposition ; see also figure [ phi_fig ] .
( to establish the second equality in , we make use of instead of . )
the mapping properties of @xmath157 are given by : @xmath263 [ pro_phi_map ] -plane under the mapping @xmath157.,title="fig : " ] + ( a ) the upper half @xmath6-plane under @xmath157 -plane under the mapping @xmath157.,title="fig : " ] + ( b ) the lower half @xmath6-plane under @xmath157
recall from that @xmath264 where @xmath259 is the constant given in definition 4.1 and @xmath265 is shown in . by using , , and
, a straightforward calculation shows that @xmath38 is the unique solution of the following rhp : 1 .
@xmath38 is analytic for @xmath266 ; 2 . for @xmath267 , @xmath268 and for @xmath269 , @xmath270 where @xmath271 and @xmath272 3 .
@xmath38 behaves like the identity matrix at infinity : @xmath273 for the sake of simplicity , we introduce some auxiliary functions .
let @xmath274 for @xmath275 $ ] and @xmath276 for @xmath212 , where @xmath277 is defined in .
in view of the equation preceding ( 4.5 ) , it is easily verified that @xmath278 on account of this , it can be shown that @xmath279 for @xmath61 .
hence , @xmath280 can be analytically continued to the interval @xmath5 .
moreover , @xmath281 as @xmath282 , by applying stirlings formula to and evaluating the integral in explicitly , one can show that @xmath283 uniformly for @xmath6 bounded away from the interval @xmath284 $ ] .
moreover , from , we have for @xmath212 @xmath285 uniformly for @xmath6 bounded away from the real line . to compute the jump conditions of @xmath38 in ( @xmath286 ) above
, we first state the following lemma . for @xmath287
, we have @xmath288 and @xmath289 for @xmath290 , we have @xmath291 [ lem_gt ] for @xmath287 , it follows from and that @xmath292 the second equality is obtained by using , and . substituting into gives . for @xmath293 ,
the formula @xmath294 follows immediately from .
note that @xmath295 for @xmath164 ; hence , by and we obtain @xmath296 for @xmath297 , follows directly from , and . with the properties of @xmath41 and @xmath42 established in section 4 and lemma [ lem_gt ] , the jump matrices for @xmath298 in condition ( @xmath286 )
can be written as follows : @xmath299 for @xmath300 , @xmath301 for @xmath302 , @xmath303 for @xmath304 , and @xmath305 for @xmath125 .
note that for the jump matrix on @xmath306 , we have the following factorization : @xmath307 the first and third matrices on the right - hand side of have , respectively , the analytic continuations @xmath308 and @xmath309 on both sides of @xmath306 .
let @xmath310 , and let @xmath311 denote the contour shown in figure [ t_fig ] .
define @xmath312 and regions @xmath313 . ] where regions @xmath314 and @xmath315 are also depicted in figure [ t_fig ] .
it can be verified that the jump matrices of @xmath316 are given by @xmath317 and @xmath318 in proving , and , we note that @xmath319 is analytic in @xmath287 . for , we also make use of . in establishing ,
we have used and . for and
, we note that @xmath320 .
thus , @xmath316 is a solution of the following rhp : 1 .
@xmath316 is analytic for @xmath321 $ ] ; 2 .
for @xmath322 , @xmath323 3 . as @xmath89 , @xmath324 for condition ( @xmath325 )
, we note from that @xmath326 as @xmath327 . from the properties of @xmath42 and @xmath157
( see also figure [ pt_fig ] and figure [ phi_fig ] ) , the regions @xmath328 and @xmath329 can be chosen sufficiently small so that @xmath330 moreover , from ( [ e_asy ] ) it follows that @xmath331 as @xmath332 for @xmath6 bounded away from the real line .
these together with ( [ pt_map_0 ] ) and ( [ ps_map_b ] ) imply that the jump matrix @xmath333 tends exponentially to the identity matrix as @xmath34 for @xmath6 bounded away from @xmath334\cup \{0\}\cup \{1\}$ ] . on the other hand , the weight function @xmath335 in ( [ w_def ] )
can be rewritten as @xmath336 a direct calculation shows that @xmath337 therefore , it is natural to suggest that for large @xmath1 , the solution of the rhp for @xmath338 may behave asymptotically like the solution of the following rhp for @xmath339 : 1 .
@xmath339 is analytic for @xmath340 $ ] ; 2 . for @xmath341 ,
@xmath342 where @xmath343 3 . as @xmath89 ,
@xmath344 to solve the above rhp , we first consider the function @xmath345 where @xmath346 is analytic in @xmath167 $ ] and behaves like @xmath6 as @xmath327 . in view of
, it is easily shown that the function @xmath347 is indeed analytic in @xmath167 $ ] and @xmath348 its limit at infinity is given by @xmath349 therefore , if we set @xmath350 it follows from , and that @xmath351 is a solution of the rhp : 1 .
@xmath351 is analytic for @xmath340 $ ] ; 2 .
for @xmath341 , @xmath352 3 . as @xmath89 ,
@xmath353 this problem can be solved explicitly , and its solution is given by @xmath354 where @xmath355 with a branch cut along @xmath334 $ ] and @xmath356 as @xmath89 .
hence , @xmath357 and for @xmath358 , we expect that @xmath359 see the comment following equation . however , since @xmath360 for @xmath6 near 0 and 1 and since @xmath361 for @xmath6 near @xmath158 and @xmath180 , the jump matrix @xmath362 near these critical points ; see - .
that is , @xmath316 and @xmath339 are not uniformly close to each other near these points , and special attention must be paid to the neighborhoods of these points .
let @xmath363 be a neighborhood of the origin , where @xmath364 is a small positive number .
in view of , , and , we first wish to find a matrix - valued function @xmath365 that satisfies the following rhp : 1 .
@xmath365 is analytic for @xmath366 ; 2 . @xmath367
+ where the functions @xmath368 and @xmath369 denote the boundary values of @xmath365 taken from the left and right of the imaginary axis , respectively ; 3 . for @xmath370 , @xmath371 as @xmath34 .
next , we consider as in @xcite the following @xmath372-function @xmath373 .
\label{d}\end{aligned}\ ] ] by applying stirlings formula to , we have @xmath374 for @xmath375 .
we also introduce the function @xmath376 which is actually analytic in @xmath377 . from eulers reflection formula
, it follows that @xmath378 similarly , for @xmath379 , we have @xmath380 as @xmath34 .
set the following matrix - valued function @xmath381 it follows from , and that @xmath382 is a solution of the rhp : 1 .
@xmath383 is analytic for @xmath384 ; 2 . @xmath385
+ where the functions @xmath386 and @xmath387 denote the boundary values of @xmath383 taken from the left and right of the imaginary axis , respectively ; 3 . for @xmath6 bounded away from the origin , @xmath388 as @xmath389 . a comparison of conditions @xmath390 and @xmath391 with and shows that a solution of the rhp for @xmath392 is given by @xmath393 in a similar manner , for @xmath394 , we want to find a matrix - valued function @xmath395 that satisfies the rhp : 1 .
@xmath395 is analytic for @xmath396 ; 2 . @xmath397
+ where the functions @xmath398 and @xmath399 denote the boundary values of @xmath395 taken from the right and left of the line @xmath400 , respectively ; also see figure [ t_fig ] ; 3 . for @xmath401 , @xmath402 as @xmath34 . as in the case of @xmath392
, it can be verified that a solution of the rhp for @xmath403 is given by @xmath404 since @xmath365 satisfies the same jump conditions as @xmath316 on the contour @xmath405 , @xmath406 for @xmath407 .
similarly , @xmath408 for @xmath409 .
hence , as @xmath389 , we have from @xmath410 by - , it is clear that @xmath411 let @xmath412 be an arbitrary fixed point in the interval @xmath306 . from , and
, it follows that for @xmath413 , @xmath414 and by and we obtain @xmath415 for @xmath413 . from the definitions of @xmath280 and @xmath416 in ( [ et ] ) and ( [ ds_d ] ) , respectively
, it is easily shown that @xmath417 hence , it follows that for @xmath418 , @xmath419 similarly , for @xmath420 , we have by using , and @xmath421 where we have also made use of the identity @xmath422 obtained from and .
define @xmath423 on account of , and , one can work backwards to get @xmath424^{\s_3 } \label{r_asy}\end{aligned}\ ] ] for @xmath425 , where @xmath259 is given in definition 4.1 .
in this section , we will construct an approximation @xmath426 to the solution of the rhp for @xmath39 stated in section 3 for large @xmath1 . given @xmath427 , define the rectangle @xmath428 , and the line @xmath429 , where @xmath412 is given in section 5 .
we divide the complex plane into three regions i , ii , iii by using @xmath430 and @xmath431 ; see figure [ asy_fig ] . from , it is natural to set @xmath432^{\s_3 } \label{rt : i}\end{aligned}\ ] ] for @xmath433 i. for @xmath6 inside the rectangle @xmath431 , we will construct the parametrices by using airy functions as in @xcite . for @xmath434
, it follows from ( [ ds_d ] ) , and ( [ r_asy ] ) @xmath435^{\s_3 } , \qquad z\in \mathbb{c}_\pm , \label{r_asy_pt1}\end{aligned}\ ] ] and from ( [ pt_phi ] ) and ( [ d_asy ] ) we obtain @xmath436^{\s_3}. \label{r_asy_pt}\end{aligned}\ ] ] using and , one can derive an expansion for @xmath41 for @xmath437 .
indeed , we have @xmath438 for @xmath439 . by using ( 4.29 ) and the proposition of @xmath440 ( see figure [ pt_fig ] ) , it can be verified that the function defined by @xmath441 satisfies @xmath442 on account of this and , we conclude that this function is analytic in @xmath443\cup[b,+\infty)$ ] .
in particular , one can see that for @xmath444 @xmath445 and we obtain from @xmath446 note that @xmath447 by and , and @xmath351 has the factorization @xmath448 since @xmath449 , @xmath339 has fourth - root singularities at @xmath450 and @xmath451 . in order to remove the singularity at @xmath450
, we introduce the function @xmath452^{-\s_3/4 } .
\label{etn}\end{aligned}\ ] ] note that by using , we have from @xmath453 also , from we have @xmath454 thus , @xmath455 coupling and , we obtain @xmath456 since @xmath457 has a simple zero at @xmath450 and @xmath458 has a fourth - root singularity at @xmath158 , the last formula shows that @xmath459 has a removable singularity at @xmath450 .
furthermore , one can show that @xmath459 has no jump across the interval @xmath306 .
indeed , by @xmath460 from , and , it follows that for @xmath164 @xmath461^{-1 } ( \wt e_n ) _ + ( x ) \hh \frac12 [ n^{2/3}\ft_-(x)]^{\s_3/4 } \begin{pmatrix } 1 & -i \\
-i & 1 \end{pmatrix } \begin{pmatrix } 0 & -1 \\ 1 & 0 \end{pmatrix } \begin{pmatrix } 1 & i \\
i & 1 \end{pmatrix } [ n^{2/3}\ft_+(x)]^{-\s_3/4 } \hh [ n^{2/3}\ft_-(x)]^{\s_3/4 } \begin{pmatrix } -i & 0 \\ 0 & i \end{pmatrix } [ n^{2/3}\ft_+(x)]^{-\s_3/4 } = i. \label{e0t}\end{aligned}\ ] ] therefore , @xmath459 is analytic across @xmath306 . also , note that from , and , it is evident that @xmath339 , @xmath462 and @xmath457 are analytic for @xmath463 .
thus , the matrix - valued function @xmath464 is analytic in @xmath183\cup [ b,+\infty)$ ] . by virtue of ,
we can rewrite ( [ r_asy_pt ] ) as @xmath465^{\s_3 } \notag \\ & \sim \frac{(-1)^n}{2 } ( c_1 e^{nl } ) ^{\s_3/2 } \wt e_n(z ) [ n^{2/3 } \ft(z)]^{\s_3/4 } \begin{pmatrix } e^ { -n\pt(z ) } & -ie^ { n\pt(z ) } \\
-ie^ { -n\pt(z ) } & e^ { n\pt(z ) } \end{pmatrix } \left [ \frac { 2\cos(n\pi z ) } { ( c_1 h(z))^{\frac12 } \prod_{j=0}^{n-1}(z - x_{n , j } ) } \right]^{\s_3}. \label{r_asy_ft1}\end{aligned}\ ] ] to find an approximation to @xmath39 , we first look for a matrix which is asymptotic to @xmath466^{\s_3/4}\begin{pmatrix } e^ { -n\pt(z ) } & -ie^ { n\pt(z ) }
\\ -ie^ { -n\pt(z ) } & e^ { n\pt(z ) } \end{pmatrix}.\end{aligned}\ ] ] set @xmath467 from the asymptotic expansions of the airy function @xcite , we have @xmath468^{-1/4 } e^ { n \pt(z ) } , \\ & \ai ' ( \wt \xi ) \sim -\frac{\xt^{1/4 } } { 2\sqrt{\pi } } e^ { -\frac23 \wt \xi^{3/2 } } = -\frac{1 } { 2\sqrt{\pi } } [ n^{2/3}\ft(z)]^{1/4 } e^ { n \pt(z ) } , \\ & \ai ( \om \wt \xi ) \sim \frac { e^{-i\pi/6 } } { 2\sqrt{\pi } } \wt \xi^{-1/4 } e^ { \frac23 \wt\xi^{3/2 } } = \frac { e^{-i\pi/6 } } { 2\sqrt{\pi } } [ n^{2/3}\ft(z)]^{-1/4 } e^ { -n \pt(z ) } , \\ & \ai ' ( \om \wt \xi ) \sim -\frac { e^{i\pi/6 } } { 2\sqrt{\pi } } \wt \xi^{1/4 } e^ { \frac23 \wt\xi^{3/2 } } = -\frac { e^{i\pi/6 } } { 2\sqrt{\pi } } [ n^{2/3}\ft(z)]^{1/4 } e^ { -n \pt(z ) } , \end{aligned } \label{ai:1}\ ] ] where @xmath469 .
it is immediate that for @xmath444 , @xmath466^{\s_3/4}\begin{pmatrix } e^ { -n\pt(z ) } & -ie^ { n\pt(z ) } \\
-ie^ { -n\pt(z ) } & e^ { n\pt(z ) } \end{pmatrix } \sim 2\sqrt \pi \begin{pmatrix } -i\om^2 \ai ' ( \om\wt \xi ) & i\ai ' ( \wt \xi ) \\ -\om\ai ( \om \wt \xi ) & \ai ( \wt \xi )
\end{pmatrix}. \label{r_mat_2_plu}\end{aligned}\ ] ] similarly , for @xmath470 , we also have @xmath471 and @xmath466^{\s_3/4}\begin{pmatrix } e^ { -n\pt(z ) } & -ie^ { n\pt(z ) } \\ -ie^ { -n\pt(z ) } & e^ { n\pt(z ) } \end{pmatrix } \sim 2\sqrt \pi \begin{pmatrix } i\om \ai ' ( \om^2\wt \xi ) & i\ai ' ( \wt \xi ) \\
\om^2\ai ( \om^2\wt \xi ) & \ai ( \wt \xi ) \end{pmatrix}. \label{r_mat_2_min}\end{aligned}\ ] ] for @xmath472 , there is a result corresponding to ( [ r_asy_pt ] ) with @xmath41 replaced by @xmath42 .
indeed , it follows from ( [ dt ] ) , ( [ r_asy ] ) and ( [ ps_phi ] ) that @xmath473^{\s_3}. \label{r_asy_ps}\end{aligned}\ ] ] let @xmath474 which is analytic in @xmath475\cup [ 1,+\infty)$ ] . by proposition 4.3
, we see that for @xmath164 @xmath476 since @xmath339 has a fourth - root singularity at @xmath451 , we define @xmath477^{-\s_3/4 } . \label{esn}\end{aligned}\ ] ] as in the case of @xmath459 , it is readily seen that @xmath180 is a removable singularity of @xmath478 . by ( 6.11 ) and ( [ fs_jum ] )
, one also has @xmath479^{-1 } ( e_n^ * ) _ + ( x ) & = [ n^{2/3 } \fs_-(z)]^{\s_3/4 } \begin{pmatrix } i & 0 \\ 0 & -i \end{pmatrix } [ n^{2/3 } \fs_+(z)]^{-\s_3/4 } \hh i \label{eb}\end{aligned}\ ] ] for @xmath164 .
therefore , @xmath478 is analytic in @xmath480\cup[1,+\infty)$ ] . on account of this and ( [ r_asy_ps ] )
, we have for @xmath472 @xmath481^{\s_3 } \notag \\ & \sim \frac{(-1)^n}{2 } ( c_1 e^{nl } ) ^{\s_3/2 } e^*_n(z ) [ n^{2/3 } \fs(z)]^{\s_3/4 } \begin{pmatrix } e^ { -n\ps(z ) } & ie^ { n\ps(z ) } \\
ie^ { -n\ps(z ) } & e^ { n\ps(z ) } \end{pmatrix } \bigg [ \frac { 2\cos(n\pi z ) } { ( c_1 h(z))^{1/2 } \prod_{j=0}^{n-1}(z - x_{n , j } ) } \bigg]^{\s_3}. \label{r_asy_fs1}\end{aligned}\ ] ] hence , as before , it can be shown that the matrix @xmath482^{\s_3/4 } \begin{pmatrix } e^ { -n\ps(z ) } & ie^ { n\ps(z ) } \\
ie^ { -n\ps(z ) } & e^ { n\ps(z ) } \end{pmatrix}\end{aligned}\ ] ] is the leading term of the asymptotic expansion of the matrices @xmath483 for @xmath484 , and @xmath485 for @xmath486 , where @xmath487 .
define the matrix function @xmath488 an appeal to the formulas @xmath489 shows @xmath490 satisfies @xmath491 let @xmath459 and @xmath478 be defined as in and , respectively .
in view of the heuristic argument leading to , and , it is reasonable to suggest that @xmath39 is asymptotically approximated by @xmath492 where @xmath493^{\s_3 } \label{m1 } \end{aligned}\ ] ] for @xmath494 , and @xmath495^{\s_3 } \label{m2 } \end{aligned}\ ] ] for @xmath496 . from
, it can be shown that @xmath426 has the same large @xmath6 behavior as @xmath39 given in ( @xmath129 ) . to see this
, we only consider the case when @xmath494 and @xmath497 .
the discussion for the case @xmath498 and @xmath499 is very similar .
note that when @xmath494 and @xmath420 , @xmath500^{\s_3}. \end{aligned}\ ] ] from and , it is easily seen that @xmath501 for large @xmath6 .
furthermore , from and @xmath502 , we have @xmath503 this together with ( [ g_def ] ) and the asymptotic behavior of @xmath339 in ( @xmath504 ) gives @xmath505 ( c_1 e^{nl } ) ^{-\s_3/2 } z^{(n - n)\s_3 } \hh [ i+o(1/z ) ] \begin{pmatrix } z^{n - n } & 0 \\ 0 & z^{-n+n } \end{pmatrix } \qquad \qquad \mbox{as } z\ra\infty . \label{rt : inf}\end{aligned}\ ] ] let @xmath506 ; see figure [ js_fig ] .
since @xmath426 has the same large @xmath6-behavior as @xmath39 , it can be readily verified that the matrix - valued function @xmath507 is a solution of the rhp : 1 .
@xmath508 is analytic for @xmath509 ; 2 . for @xmath510 , @xmath511 where @xmath512 3 . for @xmath513
, @xmath514 where the contour associated with the matrix @xmath515 is @xmath516 ; see figure 1 and figure 6 . .
] to solve this problem , we first derive the asymptotic behavior of the jump matrix @xmath515 as @xmath34 .
for convenience , we consider only the cases when @xmath6 lies on the contour @xmath517 and @xmath518 , and @xmath6 in the interval @xmath5 , since the other cases can be discussed in a similar manner .
for @xmath519 , we have from , and @xmath520 similarly , @xmath521 since @xmath39 is analytic in @xmath522 , @xmath523 for @xmath524 ; see the rhp for @xmath39 in section 3 .
thus , from ( [ js ] ) it follows that @xmath525 on account of the well - known formula ( * ? ? ?
* , ( 9.2.9 ) ) ( ^2 ) ()-^2()(^2 ) = , [ ai_rel ] we obtain @xmath526 now , we set out to derive the asymptotic behavior of @xmath515 for @xmath519 . since by and @xmath527 , @xmath528 for @xmath519 .
we recall the asymptotic behavior of the airy function and its derivative in @xcite , and have as in ( 6.16 ) @xmath529 corresponding results can be given for @xmath530 , @xmath531 , @xmath532 and @xmath533 .
substituting , and the above asymptotic formulas into , and taking into account ( 4.36 ) , we readily see that @xmath534 \hh i+o(1/n ) .
\label{js:1:asy } \end{aligned}\ ] ] for @xmath535 , we again have @xmath536 and it follows from , ( [ p ] ) and that @xmath537 moreover , we note from - , and that @xmath538 and by and we obtain @xmath539 e^{-n\pt(z)\s_3}. \label{js:2:kk } \end{aligned}\ ] ] a combination of - , and yields @xmath540 e^{-n\pt(z)\s_3}e^{-n\pi iz\s_3 } d(z)^{-\s_3 } \hhh \times e^ { n\phi(z)\s_3 } h(z)^{-\s_3/2}n(z)^{-1 } \hh i+o(1/n ) . \label{js:2:asy } \end{aligned}\ ] ] let us now consider @xmath6 in the interval @xmath5 on the real line . for simplicity , we only consider the case @xmath541 $ ] .
for @xmath542 $ ] , we note by that @xmath543 , where @xmath544 is the point of intersection of the vertical line @xmath545 with the real axis ; see figure 7 .
it follows from - , and that @xmath546^{\s_3 } \begin{pmatrix } 1 & 0 \\
\dfrac{4\cos^2(n\pi x ) } { c_1 w(x ) \prod^{n-1}_{j=0}(x - x_{n , j } ) ^2 } & 1 \end{pmatrix } \bigskip \hhh \qquad \qquad \qquad \qquad \times \left [ \frac { d^*_+(x ) } { c_1^{1/2 } \prod^{n-1}_{j=0}(x - x_{n , j } ) } \right]^{-\s_3 } e^{n\phi_+(x)\s_3 } n(x)^{-1 } \hh n(x ) \begin{pmatrix } { d^*_- } /{d^*_+ }
e^{n(\phi_+-\phi_- ) } & 0 \\ \dfrac{4\cos^2(n\pi x ) } { w(x)d_+^*(x ) d_-^*(x ) } e^{n(\phi_++\phi_- ) } & { d^*_+ } / { d^*_- } e^{n(\phi_--\phi_+ ) } \end{pmatrix } n(x)^{-1}. \label{js:3:pre1 } \end{aligned}\ ] ] from , one can see that @xmath547 and @xmath548 since @xmath41 is analytic for @xmath463 , we obtain from @xmath549 in fact , it is easily verified that @xmath550 is bounded and @xmath551 in this case . since @xmath552 for @xmath553 $ ] ; see and figure [ pt_fig ] , @xmath554 as @xmath389 .
thus , @xmath555 finally , we consider the case @xmath556 $ ] . by , and , we have from @xmath557^{\s_3 } \begin{pmatrix } 1 & 0 \\
\dfrac{4\cos^2(n\pi x ) } { c_1 w(x ) \prod^{n-1}_{j=0}(x - x_{n , j } ) ^2 } & 1 \end{pmatrix } \hhh \times \left [ \frac { 2\cos(n\pi x ) } { ( c_1 h(x ) ) ^{1/2 } \prod^{n-1}_{j=0}(x - x_{n , j } ) } \right]^{-\s_3 } p_+(x)^{-1 } \et_n(x)^{-1 } \hh \et_n(x ) p_-(x ) \begin{pmatrix } 1 & 0 \\
h(x)w(x)^{-1 } & 1 \end{pmatrix } p_+(x)^{-1 } \et_n(x)^{-1 } .
\label{js:4:pre } \end{aligned}\ ] ] since by @xmath558 for @xmath559 $ ] , it follows from that @xmath560 in a similar manner , we can prove that for @xmath6 in other parts of the contour @xmath561 , the jump matrix @xmath515 also tends to the identity matrix .
hence , @xmath562 as @xmath563 on the whole contour @xmath561 . by theorem 3.8 in @xcite , the solution of the rhp for @xmath564 has the asymptotic behavior @xmath565 as @xmath34 uniformly for @xmath513 .
let @xmath566 denote the @xmath567 element in the matrix @xmath568 . for convenience ,
we define @xmath569 and @xmath570 . by and , a straightforward calculation gives @xmath571 and @xmath572 where @xmath158 and @xmath180 are given in and @xmath573 .
note that @xmath347 in is analytic in @xmath167 $ ] and hence @xmath574 is also analytic in @xmath167 $ ] .
however , this is not true with the function @xmath575 , which has additional cuts @xmath170 $ ] and @xmath576 .
we finally arrive at the following theorem : let i , ii , iii be the regions shown in figure [ asy_fig ] , and let @xmath259 denote the lagrange constant given in ( [ phi_def ] )
. with @xmath157 and @xmath577 defined in ( [ phi_def ] ) and ( [ dt ] ) , the asymptotic formula of the polynomial @xmath578 is given by @xmath579 + \delta(n ) \right\ } \label{p1}\end{aligned}\ ] ] for @xmath580 , where @xmath581 for @xmath582 and @xmath583 is exponentially small in comparison with its leading term for @xmath584 ; see figure 1 .
more precisely , @xmath585 , where @xmath586 is negative when @xmath587 is bounded away from the interval @xmath306 ; see figure [ phi_fig ] .
let @xmath588 be defined as in ( [ ft_def ] ) .
we have @xmath589 + \widetilde{\text{b}}(z , n)\left [ 1+o(\frac1n ) \right ] \right\ } \label{p2}\end{aligned}\ ] ] for @xmath590 , where @xmath591^{1/4 } [ m(z)+m^*(z ) ] \left\ { \sin(n\pi z ) \ai(n^{2/3}\widetilde{f}(z ) ) + \cos(n\pi z)\bi(n^{2/3}\widetilde{f}(z ) ) \right\},\end{aligned}\ ] ] and @xmath592^{-1/4 } [ m(z)-m^*(z ) ] \left\ { \sin(n\pi z ) \ai'(n^{2/3}\widetilde{f}(z ) ) + \cos(n\pi z)\bi'(n^{2/3}\widetilde{f}(z ) ) \right\}.\end{aligned}\ ] ] similarly , with @xmath593 defined in ( [ fs_def ] ) , @xmath594 + { \text{b}}^*(z , n)\left [ 1+o(\frac1n ) \right ] \right\ } \label{p3}\end{aligned}\ ] ] for @xmath595 , where @xmath596^{1/4 } [ m(z)-m^*(z ) ] \left\ { \cos(n\pi z)\bi(n^{2/3}{f}^*(z))-\sin(n\pi z ) \ai(n^{2/3}{f}^*(z ) ) \right\},\end{aligned}\ ] ] and @xmath597^{-1/4 } [ m(z)+m^*(z ) ] \left\ { \cos(n\pi z)\bi'(n^{2/3}{f}^*(z))-\sin(n\pi z ) \ai'(n^{2/3}{f}^*(z ) ) \right\}.\end{aligned}\ ] ] since @xmath598 by , we have @xmath599 and @xmath600 from , it follows that @xmath601 first , let us consider @xmath494 .
recalling the definition of @xmath426 in , one obtains @xmath602 \label{r11:p1 } \end{aligned}\ ] ] and @xmath603 .
\label{r12:p1 } \end{aligned}\ ] ] note by theorem 2.1 and ( 3.1 ) that @xmath604 from ( 3.3)-(3.4 ) , we know that @xmath605 has different expressions in different parts of the region i. from , and , it follows that @xmath606 \label{p1:nom } \end{aligned}\ ] ] for @xmath607 , since @xmath569 .
recalling the notation @xmath608 in , we have from ( 3.3 ) and - @xmath609 \mp e^{n\phi(z ) } \frac { 2 e^{\pm n\pi i z } \cos(n\pi z ) } { \wt d(z)w(z ) } n_{12}(z ) \left [ 1+o(1/n ) \right ] \right\ } \label{p1:ome : pre } \end{aligned}\ ] ] for @xmath610 . for simplicity ,
we only consider @xmath499 .
it follows from and that @xmath611 \mp e^{n\phi(z ) } { d(z)^{-1}w(z)^{-1 } } n_{12}(z ) \left [ 1+o(1/n ) \right ] \right\ } .
\label{p1:ome } \end{aligned}\ ] ] note that @xmath612 for @xmath494 ; see figure [ phi_fig ] .
moreover , it is readily seen from that @xmath613 is bounded as @xmath389 for @xmath614 . from .
we also have @xmath335 bounded as @xmath615 for @xmath616 and @xmath617 . near @xmath618 , @xmath619 .
thus , in terms of the notation @xmath620 , @xmath621 + o ( n^{\a } e^{n \re \phi(z ) } ) \right\ } \end{aligned}\ ] ] for @xmath610 and @xmath418 .
similarly , we can show @xmath621 + o ( n^{\b } e^{n \re \phi(z ) } ) \right\ } \end{aligned}\ ] ] for @xmath610 and @xmath420 . coupling the above two formulas with gives .
next , we consider the case @xmath622 ; see figure 6 .
let us first restrict @xmath623 . from ( 7.6 ) and ( 7.7 ) , we obtain by a combination of ( [ etn ] ) , ( [ p ] ) and ( [ rt ] ) @xmath624 n_{11}(z ) [ 1 + o(1/n ) ] \hhh \quad \qquad \qquad \qquad \qquad \qquad + [ \xt^{-\frac14 } \om^2 \ai'(\om\xt ) -\xt^{\frac14 } \om \ai(\om \xt ) ] n_{12}(z ) h(z)^{-1 } [ 1 + o(1/n ) ] \bigg\ } \label{r11:p2}\end{aligned}\ ] ] and @xmath625 n_{11}(z ) h(z ) [ 1 + o(1/n ) ] \hhh \qquad \qquad \qquad \qquad \qquad \qquad \qquad \qquad + [ \xt^{\frac14}\ai(\xt)- \xt^{-\frac14 } \ai'(\xt ) ] n_{12}(z ) [ 1+o(1/n ) ] \bigg\ } \label{r12:p2 } .\end{aligned}\ ] ] here , use has also been made of ( 7.8 ) .
since @xmath626 , by and we have @xmath627 see the first equality in ( 7.13 ) . a combination of this with ( 7.17 ) and ( 7.18 ) gives @xmath628 n_{11}(z ) [ 1 + o(1/n ) ] \hhh \qquad \qquad \ \qquad \qquad \qquad \qquad \ + [ \xt^{-\frac14 } \om^2 \ai'(\om\xt ) -\xt^{\frac14 } \om \ai(\om \xt ) ] n_{12}(z ) h(z)^{-1 } [ 1 + o(1/n ) ] \bigg ] \hhh \qquad \qquad \qquad \quad \ - e^{n\pi i z } w(z)^{-1 } \bigg [ [ i \xt^{-\frac14 } \ai'(\xt ) + i \xt^{\frac14}\ai(\xt ) ] n_{11}(z ) h(z ) [ 1 + o(1/n ) ] \hhh \qquad \qquad \qquad \qquad \qquad \qquad \quad \quad \quad + [ \xt^{\frac14}\ai(\xt)- \xt^{-\frac14 }
\ai'(\xt ) ] n_{12}(z ) [ 1+o(1/n ) ] \bigg ] \bigg\ } \label{p2:pre } \end{aligned}\ ] ] for @xmath629 .
recall the well - known formula of the airy functions ( * ? ? ?
* ( 9.2.11 ) ) @xmath630 . \label{bi}\end{aligned}\ ] ] also , note that @xmath631 $ ] as @xmath389 ; see and .
this together with gives @xmath632 + \widetilde{\text{b}}(z , n)\left [ 1+o(\frac1n ) \right ] \right\ } \label{p2:p } \end{aligned}\ ] ] for @xmath629 . in a similar manner
, one can see that @xmath633\cdot n_{11}(z ) [ 1 + o(1/n ) ] \hhh \qquad \qquad \
\qquad \qquad \qquad \qquad \
+ [ -\xt^{-\frac14 } \om \ai'(\om^2 \xt ) + \xt^{\frac14 } \om^2 \ai(\om^2 \xt ) ] \cdot n_{12}(z ) h(z)^{-1 } [ 1 + o(1/n ) ] \bigg ] \hhh \qquad \qquad \qquad \quad \ + e^{-n\pi i z } w(z)^{-1 } \bigg [ [ i \xt^{-\frac14 } \ai'(\xt ) + i \xt^{\frac14}\ai(\xt ) ] n_{11}(z ) h(z ) [ 1 + o(1/n ) ] \hhh \qquad \qquad \qquad \qquad \qquad \qquad \quad \quad \quad + [ \xt^{\frac14}\ai(\xt)- \xt^{-\frac14 } \ai'(\xt ) ] n_{12}(z ) [ 1+o(1/n ) ] \bigg ] \bigg\ } \label{p2:pre1 } \end{aligned}\ ] ] for @xmath634 .
again by , we obtain the exactly same formula given in , thus proving .
following the same argument as given above , one can establish for @xmath635 .
this completes the proof of theorem 7.1 .
in this section , we want to investigate the asymptotic behavior of the hahn polynomials @xmath636 for fixed values of @xmath28 .
first , we introduce the notation @xmath637 then , it follows from ( 1.6 ) and ( 2.14 ) that @xmath638 here , we have made use of the fact that @xmath643 .
substituting into gives @xmath644 moreover , it is readily verified that @xmath645 as @xmath389 , and by @xmath646 as @xmath647 .
we now derive an explicit formula for @xmath648 . from ( [ pt_phi ] ) , and , one has @xmath649 for @xmath463 .
note that since @xmath28 is fixed , by @xmath650 as @xmath389 . on account of and
( 4.58 ) , by letting @xmath389 , we obtain @xmath651 + z \left [ \frac1c \log(z ) -\frac2c \log c -\frac1c \right]+o(z^2 ) , \label{pt : x : asy}\end{aligned}\ ] ] and so @xmath652 ( the last two equations have also been used in [ 10 , ( 6.27 ) & ( 6.28 ) ] . ) a combination of , , and gives @xmath653 for @xmath640 .
next , we consider the case when @xmath654 , i.e. , @xmath655 . from and , it can be shown that @xmath656 is analytic in the interval @xmath657 .
hence , we obtain from and @xmath658 note that @xmath659 . on account of
, we have @xmath660 + z \left [ \frac1c \log(-z ) -\frac2c \log c -\frac1c \right]+o(z^2 ) \label{pp}\end{aligned}\ ] ] as @xmath661 . letting @xmath34
, we obtain @xmath662 equations and are also given in [ 10 , ( 6.33 ) & ( 6.34 ) ] .
inserting in ( [ ds ] ) yields @xmath663 hence , it follows from , , ( [ mx ] ) and -([dsx ] ) that @xmath664 for @xmath654 . to conclude this chapter , we make a comparison of our results with those presented in @xcite for the discrete chebyshev polynomials @xmath665
first , note that by taking @xmath666 , the weight function @xmath667 in becomes 1 .
furthermore , we have @xmath668 see ( * ? ? ?
* & p.176 ) and @xcite . for convenience ,
we consider only the case when @xmath6 is real and @xmath418 . by and , we have @xmath669 if @xmath670 denotes the monic polynomials associated with the discrete chebyshev polynomials , then we have the relationship @xmath671 in view of , and ( * ? ?
* ( 2.1 ) ) .
since the function @xmath220 in does not depend on @xmath2 and @xmath3 , it is the same @xmath130-function defined in ( * ? ? ? * ( 3.4 ) ) ; also see and ( * ? ? ?
* ( 6.25 ) ) . for @xmath622 ( that is , @xmath672 )
, in the notations of this paper we can rewrite ( 5.1 ) in @xcite as @xmath673 \frac{(z - b)^{1/4 } } { ( z - a)^{1/4 } } [ n^{2/3}\widetilde{f}(z)]^{1/4 } \nonumber \\ & \qquad \quad+ \left [ \sin(n\pi z ) \text{\upshape ai}'(n^{2/3}\widetilde{f}(z))+\cos(n\pi z)\text{\upshape bi}'(n^{2/3}\widetilde{f}(z ) ) \right ] \frac { ( z - a)^{1/4 } } { ( z - b)^{1/4 } } [ n^{2/3}\widetilde{f}(z)]^{-1/4 } \bigg\ } , \label{pi_left}\end{aligned}\ ] ] on account of ( 3.37 ) in @xcite . in view of
, it can be readily seen that agrees with .
moreover , we obtain from ( * ? ? ?
* ( 6.2 ) ) , ( * ? ? ? * ( 3.34 ) ) , and @xmath674 for @xmath655 . similarly , by ( * ? ? ?
* ( 6.1 ) ) , ( * ? ? ?
* ( 3.33 ) ) , and we have @xmath675 for @xmath676 . a comparison of and with and implies that our results agree with those in @xcite .
in fact , one can easily show that when @xmath666 , we do not need to construct the parametrix in region i. however , @xmath426 defined in ( 6.31 ) for @xmath434 can not be extended to the region i when @xmath677 .
this is because the function @xmath462 involved in @xmath426 is not analytic in the neighborhoods of @xmath678 and @xmath679 .
j. baik , t. kriecherbauer , k. t .-
mclaughlin and p. d. miller , _ discrete orthogonal polynomials : asymptotics and applications _ , annals of mathematics studies , vol
princeton university press , princeton , nj , 2007 . | in this paper , we study the asymptotics of the hahn polynomials @xmath0 as the degree @xmath1 grows to infinity , when the parameters @xmath2 and @xmath3 are fixed and the ratio of @xmath4 is a constant in the interval @xmath5 .
uniform asymptotic formulas in terms of airy functions and elementary functions are obtained for @xmath6 in three overlapping regions , which together cover the whole complex plane .
our method is based on a modified version of the riemann - hilbert approach introduced by deift and zhou .
+ + * keywords : * global asymptotics ; hahn polynomials ; riemann - hilbert problems ; airy function . + * mathematics subject classification 2010 * : 41a60 , 33c45 |
low mass stars form via the gravitational collapse of `` dense cores '' with densities @xmath1 @xmath2 which are built up either via the diffusion of neutrals past ions tied to magnetic field lines or the dissipation of turbulent support ( evans 1999 ) .
the lynds catalog ( lynds 1962 ) of optically selected dark cores provides an excellent starting point for the observational study of the processes involved .
lynds cores are generally compact , dense and emit strongly in a wide range of molecular lines and continuum at millimeter wavelengths .
this allows for detailed investigations of their dynamics and comparison with theoretical models ( e.g. gueth et al .
1997 ) .
the first stage in star formation , the collapse of a starless core , can be diagnosed through the analysis of self - absorbed spectral features ( leung & brown 1977 ) . for sufficiently high optical depth and a decreasing excitation temperature with radius ,
the outer part of a core absorbs the emission from the center .
the resulting spectral profile will show a dip near the velocity of the absorbing layer .
if the core is collapsing , this dip will appear at higher velocities than the bulk of the core emission .
such red - shifted self - absorption is the principal signature of inward motions . probably because self - absorption requires a fairly well - tuned combination of optical depth and excitation conditions , however , the window of observability of infall in starless cores is fairly short ( e.g. gregersen & evans 2000 ) .
l1544 is the prototypical example of a collapsing starless core .
it shows strong red - shifted self - absorption in a number of molecular species ( e.g. , tafalla et al .
1998 ; caselli et al .
these have been interpreted to show both rotation and collapse ( ohashi et al . 1999 ) and the infall motions have been compared in detail with a variety of collapse models , ranging from gravitational free - fall ( whitworth & ward - thompson 2001 ) , ambipolar diffusion ( ciolek & basu 2000 ) , external compression ( hennebelle et al .
2003 ) , and turbulent cooling ( myers & lazarian 1998 ) .
l1544 has received so much attention as it shows the most clearly defined infall signatures of any starless core .
it is important , nevertheless , to study other cores to constrain the models further .
l694 - 2 was identified as a collapsing , starless core by lee and collaborators ; lee & myers ( 1999 ) used the digital sky survey to define the optical properties of the most opaque lynds cores in detail and followed up with a molecular line survey of those that were determined to be starless ( lee , myers , & tafalla 1999 ) .
l694 - 2 came to prominence in this work as a `` strong infall candidate '' on the basis of pronounced red - shifted self - absorption in the cs(21 ) line . subsequently
lee , myers , & plume ( 2004 ) showed similar infall profiiles in cs(32 ) and dco@xmath3(21 ) .
their spectral fits indicated that the infall speed increased with optical depth and , presumably therefore , toward the center of the core .
the starless nature of l694 - 2 was deduced from the lack of an iras point source and confirmed by a vla non - detection at 3.6 cm ( harvey et al .
2002 ) . together
these limit the luminosity of an embedded source to @xmath4 .
subsequent millimeter continuum observations by harvey et al .
( 2003a ) showed that there was no compact , disk - like source to a sensitivity @xmath5 .
the core is included in the `` c2d '' spitzer legacy program which will show whether a very low luminosity source exists ( evans et al .
2003 ) l694 - 2 also has similar structural and chemical properties as l1544 ( crapsi et al .
the low temperatures and high densities of starless cores result in substantial molecular depletion onto grain surfaces ( tafalla et al .
these chemical properties imply that to follow the dynamics toward the core centers requires a relatively inert molecular probe that remains in the gas phase .
to measure relative ( infall ) motions , in particular , requires high optical depth and therefore high abundance .
only a handful of molecules meet these conditions .
2hp is well suited as a relatively abundant species in starless cores that suffers little depletion except at the very highest column densities ( bergin et al .
tafalla et al .
( 1998 ) found that the @xmath6 line was double - peaked due to self - absorption at the center of l1544 .
williams et al .
( 1999 ; hereafter paper i ) subsequently mapped this line at @xmath7 resolution to examine the velocity field at scales approaching that of a protostellar disk ( 1000 au ) . in that work
, we found a centrally condensed , slightly elongated core , and an increasing infall velocity toward the center .
lee , myers & tafalla ( 2001 ) mapped 2hp(10 ) emission in l694 - 2 at @xmath8 resolution and found a compact , centrally condensed core . here , we present combined interferometric plus singledish observations of 2hp in l694 - 2 at @xmath7 resolution .
although the 2hp spectra do not split into two distinct peaks as in l1544 , they do have identifiable `` red shoulders '' due to self - absorption that are equally indicative of collapse .
we compare the two datasets and model the spectra in both cores using new analytic tools made available by de vries & myers ( 2005 ) .
the observations are presented in 2 , and the results and modeling in 3 .
we show that l1544 is more centrally condensed and has a higher infall speed and gradient than l694 - 2 .
this general trend is discussed in the context of gravitational collapse models in 4 . throughout the paper ,
we adopt a distance to l694 - 2 of 230 pc based on optical star counts by kawamura et al .
( 2001 ) and 140 pc for l1544 based on its association with the taurus cloud ( elias 1978 ) .
l694 - 2 was observed with the 10 element berkeley - illinois - maryland array in its compact c configuration on 1999 , june 7 , 8 and 11 .
the phase center of the observations was @xmath9 , @xmath10 and the total ( on - source ) integration time was 14 hours .
the amplitude and phase calibration was performed from 3 minute integrations of the quasar 1925 + 211 interleaved with each 22 minute observation on source .
the passband and flux calibration was carried out from a 30 minute observation of uranus toward the end of each track .
the derived flux for 1925 + 211 was 1.7 jy .
the digital correlator was configured with 512 channel windows at a bandwidth of 6.25 mhz ( @xmath11 velocity resolution per channel ) centered on the 7 hyperfine components of 2hp(10 ) in the lower sideband and 34s(21 ) in the upper sideband .
eight 32 channel windows at a bandwidth of 100 mhz . were used to measure continuum radiation .
data were reduced with the miriad package using standard procedures .
the data sets from each day were calibrated separately and then transformed together with natural weighting .
the resulting beamsize was @xmath12 .
the 2hp lines were detected with a high signal - to - noise ratio in the lower sideband , but no emission was detected in the upper sideband , implying a @xmath13 upper limit of 1.5 k for the 34s emission integrated over velocities @xmath14 .
no continuum emission was detected at a @xmath13 level of 2.1 mjy beam .
the projected baselines of the interferometer data ranged from 1.9 to 27 k@xmath15 . to provide information at lower spatial frequencies , we observed l694 - 2 using the five college radio astronomy observatory 14 m telescope ( fcrao ) on 1999 december 20 .
we used the 16 element , focal plane array sequoia and a system of autocorrelation spectrometers at a spectral resolution of 20 khz .
standard position - switching procedures were used and resulted in a nyquist sampled map with an rms noise of 0.05 k per 0.06 channel .
antenna temperatures were multiplied by a gain of 43.7 jy kto convert to flux units and the singledish data were then added to the interferometer data using the combination method described in stanimirovic et al . ( 1999 ) .
the analysis in this paper was performed on the combined bima+fcrao dataset convolved to a spatial resolution of @xmath7 and interpolated to velocity channels of width 0.05 .
l1544 had been previously observed in a similar way with the bima array and iram 30 m antenna ( paper i ) . to make an unbiased comparison between the two cores , we re - reduced the data and combined the interferometer and singledish datasets using the same linear combination technique described above
the @xmath6 transition of 2hp has 7 hyperfine components , two groups of three and one isolated ( caselli , myers & thaddeus 1995 ) .
the linewidths of l694 - 2 and l1544 are similar and sufficiently small that all 7 components can be distinguished .
figure 1 shows three different hyperfine components , ordered from high to low optical depth , toward the two cores .
the l1544 data have been smoothed to @xmath16 to achieve the same linear resolution , 2300 au , as the l694 data . in each case , the spectra are most asymmetric in the top panel and become more symmetric as the optical depth of the hyperfine component decreases in the middle and bottom panels
. the asymmetries must be due , therefore , to radiative transfer effects rather than kinematics .
we conclude that both cores exhibit red - shifted self - absorption .
this is strong evidence for core collapse .
the self - absorption features are less pronounced in the l694 - 2 spectra , which show shoulders , than in the l1544 spectra , which show dips .
this is indicative of a lower opacity and is in agreement with the multi - transition analysis by crapsi et al .
( 2005 ) who found that the 2hp column density is @xmath17 lower in l694 - 2 than in l1544 .
the isolated component , @xmath18 , has the lowest optical depth and therefore is the weakest and most symmetric .
although we were able to analyze the dynamics using this line in l1544 in paper i , the lower optical depth and signal - to - noise ratio precludes a similar study of the l694 - 2 dataset .
we therefore focus the remainder of the analysis and comparison of the two cores on the central @xmath19 component at 93.1737767 ghz , as it is the brightest and has the highest optical depth , making it the most sensitive for tracing relative motions in the core .
maps of the first two moments of the @xmath19 hyperfine component are plotted in figures 2 ( l694 - 2 ) and 3 ( l1544 ) .
the contours of zeroth moment , or integrated intensity , are similar to dust continuum images ( visser et al . 2002 ; ward - thompson , motte , & andr 1999 ) suggesting that there is negligible 2hp depletion in each core and that this transition is a reliable tracer of the dynamics toward the core centers at @xmath20 au resolution scales . the first moment , or mean velocity , of the line is overlaid in color scale over the integrated intensity contours . in l694 - 2 , there is a central low velocity ( blue ) hole surrounded , almost in a complete ring , by higher velocities .
in l1544 , there is a clear trend of low to high ( red ) velocities across the core , but the lines of constant velocity are curved in a ` c'-shape . in each case , the velocity pattern can not be explained solely by rotation or shear across the core .
rather , following the earlier work by tafalla et al .
( 1998 ) in l1544 and lee et al . ( 1999 ) in l692 - 2 , we suggest that the 2hp velocity pattern can be explained by a combination of rotation and core contraction .
the spectroscopic signature of core collapse is red - shifted self - absorption .
the mean velocity of such spectra is lower than in the absence of self - absorption or infall and therefore appears in a first moment map as a dip toward low , or blue , velocities .
core rotation manifests itself as an one - dimensional gradient in the first moment map , however , and the two effects can therefore be distinguished ( adelson & leung 1988 ) .
walker , narayanan , & boss ( 1994 ) modeled the radiative transfer of collapsing , rotating cores and coined the phrase `` blue - bulge '' to describe the feature in the mean velocity field where infall motions dominate over rotation . for a centrally condensed core with uniform rotation and increasing infall speed toward the center , the contours of constant velocity form a c-shape near the peak of emission , dependent on the inclination of the rotation axis to the line of sight .
they point out that this feature is a more robust signature of infall than the classic line asymmetry and subsequently demonstrate its applicability in two star - forming systems ( narayan , walker , & buckley 1998 ; narayan et al . 2002 ) .
the mean velocity maps of both l694 - 2 and l1544 also show the `` blue - bulge '' signature of collapse superimposed on rotation .
a central blue dip can be clearly seen toward the center of l694 - 2 .
the velocity gradient is higher in l1544 so the lines of constant velocity do not wrap onto themselves but nevertheless are clearly curved . using these maps , we are able to follow the core dyanamics beyond the center where the spectral asymmetries in figure 1 are most clear . to isolate and analyze the infall motions , we removed the rotation in each core by subtracting a linear velocity gradient from the maps .
this is an approximation as the 2hp line is optically thick and self - absorbed .
we define the best fit , however , by minimizing the dispersion in mean velocity within the central @xmath21 , and it is weighted ( simply by the number of pixels ) to the outer parts of the core where the optical depth is low .
therefore , the error in measuring the rotation is likely to be small .
the subtracted mean velocity image are shown for l694 - 2 and l1544 in the right panels of figures 2 and 3 respectively .
the magnitude and direction of the rotation vector is shown in the bottom right corner .
the velocity gradient in l1544 , 4.1 , is similar to the value determined in paper i from a different modeling technique .
the gradient in l694 - 2 is much smaller , 0.5 . without knowledge of the core geometries and inclination of the rotation axes , however , it is not possible to compare their angular momenta directly . in both cores ,
the residual ( mean minus rotation ) velocity field shows a marked bipolar deviation from circular symmetry .
the signal - to - noise ratios of the data are unfortunately too low to study this quantitatively but it is perhaps noteworthy that the bipolarity is not aligned with the direction of core elongation .
this suggests that the infall speed is not spherically symmetric , as would be expected for a purely gravitational flow resisted by isotropic pressure .
the high velocity gradient in l1544 stretches the color scale and masks the central blue - bulge feature in the residual map .
the increase in the residual velocity field away from the core center is shown for both cores in figure 4 and demonstrates that the bluebulge dip is steeper in l1544 than l694 - 2 .
this could be due to either or both a steeper radial variation of opacity and infall speed . to examine this
further we return to an examination of the spectral profiles . by removing the velocity field due to rotation
, we can isolate the infall dynamics in each core .
away from the center , however , the signal - to - noise ratios are too low to allow modeling of individual spectra and we binned the data into concentric radial annuli with @xmath7 width .
this allows us to measure the radial variations of optical depth and infall speed .
the radial averaging is justified in l694 - 2 by the approximate circular symmetry of the core and the analysis by harvey et al .
( 2003a , 2003b ) who showed that the core density profile can be fit either by a spherically symmetric model or a cylinder that is nearly pole on .
harvey et al . ( 2003b ) also noted that a cylindrical configuration would be prone to collapse .
l1544 is more elongated but we found an approximately circular distribution of infall speed in paper i. the average spectra for the two cores are shown as histograms in figure 5 .
the systemic motion of each core ( defined from the fits described below ) is shown as a dotted vertical line . moving out from the core centers ( top panels ) to the outer parts ( lower panels ) ,
the spectra in both cores become weaker , more symmetric , and their mean velocity shifts to higher values .
this behavior can be explained by an absorbing layer of gas falling onto the core . at small radii ,
the optical depths are high and the profiles show relatively strong self - absorption . because the absorbing layer is falling onto the core , the absorption is redshifted with respect to our line of sight and the profiles therefore appear asymmetric .
as the optical depth decreases at larger radii , there is less self - absorption and the spectra become more symmetric .
the values of the optical depth and infall speed can be determined via detailed modeling .
we were able to produce very good fits to the radially averaged spectra using the two layer infall model described in myers et al .
( 1995 ) and the parameter search technique of de vries & myers ( 2005 ) . following the recommendation in that paper
, we used the hill5 de vries & myers model .
the hill5 analytic models are characterized by 5 parameters : the systemic velocity , velocity dispersion , total optical depth , peak temperature , and the infall speed ; @xmath22 , and @xmath23 respectively .
the infall speed is defined such that the front layer moves at velocity @xmath24 with respect to the observer and the rear layer at velocity @xmath25 .
measurement of the infall speed requires a sufficiently high opacity to create significant self - absorption in the spectra , and was not possible in the outer parts of each core .
the fits are only shown , therefore , for the spectra where the modeled optical depth is greater than unity ; @xmath26 for l694 - 2 and @xmath27 for l1544 .
the model spectra are overlaid as solid lines over the histograms in figure 5 and the parameters that define them are listed in table 1 .
the parameter uncertainties were estimated by taking the best fit models in table 1 , adding random noise with the same rms as the observed noise , and re - fitting using the same procedure .
the error in each parameter was set equal to the dispersion in the fits over 50 repeats of this process .
we found that the errors in @xmath28 , @xmath29 , and @xmath30 were negligible , but the errors in @xmath31 and @xmath23 , whose value depends on relatively subtle features in the spectral profiles , were larger and ranged from @xmath32 .
the radial variations of modeled optical depth and infall speed are shown in figure 6 . within the @xmath33 uncertainties ,
the analysis here indicates similar , but slightly higher , infall velocities in l1544 than those measured in paper i using a different fitting method to the isolated hyperfine component
. we can be more confident of the relative differences between l1544 and l694 - 2 , however , because we have applied an identical analysis to observations of the same hyperfine component .
we find that both the optical depth and infall speed are higher in the central 3500 au in l1544 than in l694 - 2 and both decline more steeply with radius .
the higher opacity was immediately apparent from the stronger self - absorption features in the spectra in figure 1 .
the rapid decline toward larger radii reflects the steeper blue - bulge signature seen in figure 4 .
we have compared similar maps of the @xmath6 transition of 2hpin two collapsing , starless cores , l694 - 2 and l1544 . toward the core centers ,
the hyperfine components show an increasing asymmetry with optical depth .
this is strong evidence for inward motions and complements earlier studies of collapse at larger scales by lee et al .
( 2001 ) in l694 - 2 and tafalla et al .
( 1998 ) in l1544 , using lower resolution observations of cs .
we then analyzed the dynamical structure of the cores in more detail by examining the moment maps of the most optically thick , @xmath34 hyperfine component .
the mean velocity pattern showed a systematic variation across the cores from rotation and an additional `` bluebulge '' toward the center from infall .
this follows earlier modeling work and observations of other star forming regions by walker et al .
( 1994 ) and narayanan ( 1997 ) . by removing the velocity gradient from the data
, we were able to isolate the blueward shift due to infall asymmetry .
the depth of the bluebulge is similar , @xmath35 , in both cores but the size of the dip is larger in l694 - 2 reflecting a shallower opacity gradient .
these velocity - shifted spectra were then radially averaged and modeled using an automated parameter search technique by de vries & myers ( 2005 ) applied to the two - layer infall model of myers et al .
( 1995 ) .
the spectral fits show that both cores have low velocity dispersions , @xmath36 , for a kinetic temperature @xmath37 k. the non - thermal motions are therefore very small , @xmath38 .
in fact , much of the non - thermal motion may be due to infall itself . whether this is the case or not , the pressure support in the cores , which is proportional to the square of the velocity dispersion , is predominantly thermal .
the turbulence that is seen at larger scales has almost completely dissipated and the infall motions that we have measured here can constrain theoretical models of the gravitational flows in a core supported by a thermal , or thermal plus magnetic , pressure ( e.g. shu 1977 ; mouschovias 1976 ) . indeed
the variation of infall speed with radius in l1544 has been successfully reproduced using an ambipolar diffusion model by ciolek & basu ( 2000 ) and the free - fall collapse of a sphere with plummer - like density profile by whitworth & ward - thompson ( 2001 ) .
these two models differ in that the total star formation timescale is very different , @xmath39 myr for ambipolar diffusion , compared to @xmath40 myr for the free - fall collapse timescale of the central flat region of the plummer core .
they agree , however , in how far l1544 is from the end of this process .
ambipolar diffusion lengthens the build - up of a core to a critical state where magnetically diluted collapse takes place and observable infall motions only occur in the last few percent of @xmath41 .
the best fit ciolek & basu model places the age of l1544 at @xmath42 , or only @xmath43 yr from forming a star .
whitworth & ward - thompson ( 2001 ) place the age of l1544 of about 50% of the central free - fall tiemscale so only @xmath44 yr from star formation . myers ( 2005 ) also predicts a similar late stage of collapse for l1544 in a purely gravitational collapse model applied to different geometries . a general feature of both models ( and indeed any model of gravitational collapse )
is that the infall speed increases with time _ at all radii_.
the central density also increases , although the density profile in the outer parts , radii @xmath45 au , is independent of time . at the resolution of these observations , the density profiles of the two cores
should therefore appear very similar but the central column density should be higher in the more evolved core and the infall speeds should be higher throughout . the main results of the spectral fitting are the radial variations of infall speed and opacity in the two cores ( table 1 and figure 6 ) . in the inner regions , @xmath46 au , l1544 is @xmath47 more opaque and collapsing about @xmath48 faster .
this may indicate that l694 - 2 is somewhat less evolved toward star formation than l1544 , but it may also be due to its lower mass .
the precise masses of the cores depend on how their outer boundaries are defined , but within the @xmath49 au size of our 2hp maps , we estimate a mass @xmath50 for l694 - 2 from visser et al .
( 2002 ) and @xmath51 for l1544 from ward - thompson et al .
whether evolution or mass explains the differences in the model fits , the centrally concentrated blue - bulge signatures of collapse are remarkably similar and indicate that point mass formation at the centers of _ both _ cores will occur in a few @xmath0 yr .
we caution that the observed values of the radii in the plot of infall speed ( figure 6 ) do not directly translate to physical core radii because the spectral features essentially convolve the emission from several distinct regions along the line of sight .
a quantitative comparison with theory requires calculating the radiative transfer of dynamical models ( e.g. choi et al .
this is also necessary to tie the 2hp observations presented here to measurements of infall at large scales from cs observations ( lee et al .
2001 ; tafalla et al .
this adds the significant additional complexity of core chemistry .
the relevant chemical timescales for molecule formation and depletion are similar to the dynamical timescales of collapse .
consequently , the comparison of infall motions measured in two different molecular species requires combined modeling of structure , dynamics , chemistry , and radiative transfer ( e.g. doty , schier , & van dishoeck 2004 ) . due to the relatively low signal - to - noise ratio of these data and the subtle self - absorption features in the spectra , we have only been able to examine the radial variation of core properties .
the residual moment maps , after subtraction of a a velocity gradient , show additional structure .
these approximately bipolar velocity features are not aligned with the elongation of the cores , measured from the central intensity contours of l694 - 2 and l1544 .
optical depth effects alone , therefore , are not the sole explanation and it appears that the infall speed is not spherically symmetric .
both core elongation and a preferred direction for collapse are expected from a static magnetic field geometry .
a quantitative investigation of the two - dimensional density and velocity structure requires more sophisticated radiative transfer models ( e.g. hogerheijde , & van der tak 2000 ) and higher quality data , such as will soon be available with the carma array ( woody et al .
2004 ) .
several other starless collapsing core candidates are known .
it will be very worthwhile to study them at high resolution in molecular species that probe the dynamics at their centers .
the comparison of such data will give further insight into the processes that collapse the dense ism down to planetary disk scales .
we thank chris de vries for advice concerning his code to fit the spectra .
we acknowledge support from nsf grant ast-0324328 ( jpw ) and nasa origins of solar systems grant nag5 - 13050 ( pcm ) .
= 1adelson , l. m. , & leung , c. m. 1988 , , 235 , 349 4.6 in * figure 1 : * comparison of 2hp(10 ) spectra toward the centers of l694 - 2 and l1544 at the same linear resolution , 2300 au .
three hyperfine components are shown , ordered from high to low optical depth , and the dotted line shows the systemic velocity for each core .
the increasing asymmetry with opacity in each core is strong evidence for collapse . 2.9 in * figure 2 : * 2hp @xmath52 intensity and velocity structure in l694 - 2 .
the left panel shows the mean velocity in colors overlaid on contours of integrated emission .
the right panel shows the residual velocity field after subtracting a linear gradient due to core rotation .
the direction and magnitude of the velocity gradient are shown in the lower right corner .
the @xmath7 resolution of the data and scale bar are shown in the lower left corner .
the color scale ranges from 9.49 to 9.62 and contours have starting point and increments of 0.12 k in both panels .
2.9 in * figure 3 : * similar plot as figure 1 for l1544 .
the color scale ranges from 7.08 to 7.28 and contours have starting point and increments of 0.16 k in both panels .
these plots were made with the data smoothed to the same linear resolution , 2300 au , as the l694 - 2 data .
note that the velocity gradient is much higher in l1544 and the color scale is therefore stretched over a wider range .
this has the effect of reducing the contrast of the central blue - bulge feature .
3.5 in * figure 4 : * comparison of the blue - bulge signature in l694 - 2 and l1544 .
each point is a radial average in @xmath7 annuli of the residual velocity field in the right panels of figures 2 and 3 .
the velocity scale is set to zero at the outermost radii to show the blueward shift of mean velocity toward the center of each core .
the errors are the standard deviation of the mean .
the bluebulge is larger in extent but shallower in l694 - 2 ( squares connected by a dashed line ) compared to l1544 ( circles connected by a solid line ) .
3.3 in * figure 5 : * radially averaged 2hp @xmath52 spectra ( histogram ) and two - layer model fits ( line ) for the two cores .
l694 - 2 is on the left hand side , l1544 on the right .
the averaging interval , in arcseconds , is shown in the top left hand corner of the left panel and is the same in the right panel .
the vertical dotted line is the systemic velocity of each core .
the shoulder like profile toward the center of l694 - 2 contrasts with the two distinct peaks toward the center of the more opaque l1544 . 3.7 in * figure 6 : * optical depth and infall speed variation with radius , derived from the two - layer spectral fits .
the errors were determined from multiple fits to fake spectra created by adding noise to the best fit models .
l694 - 2 ( squares connected by a dashed line ) has a shallower opacity gradient and lower central values of opacity and infall speed than l1544 ( circles connected by a solid line ) .
this may reflect an earlier evolutionary state or a slightly lower mass .
cccccc + annulus & @xmath28 & @xmath29 & @xmath30 & @xmath31 & @xmath23 + ( @xmath53 ) & ( ) & ( ) & ( k ) & & ( ) + + + @xmath54 & 9.61 & 0.078 & 9.3 & @xmath55 & @xmath56 + @xmath57 & 9.61 & 0.081 & 8.9 & @xmath58 & @xmath59 + @xmath60 & 9.61 & 0.081 & 8.0 & @xmath61 & @xmath62 + + + @xmath54 & 7.20 & 0.084 & 7.1 & @xmath63 & @xmath64 + @xmath57 & 7.21 & 0.083 & 6.1 & @xmath65 & @xmath66 + @xmath60 & 7.22 & 0.086 & 5.8 & @xmath67 & @xmath68 + @xmath69 & 7.22 & 0.086 & 5.4 & @xmath70 & @xmath71 + | we present interferometric observations of 2hp(10 ) in the starless , dense core l694 - 2 and compare them to previously published maps of l1544 .
both cores are starless , centrally condensed , and show spectral signatures of rotation and collapse .
we fit radially averaged spectra using a two - layer infall model and measure the variation of opacity and infall speed in each core .
both functions increase toward the center of each core but the radial gradients are shallower , and the central values lower , in l694 - 2 .
this general behavior is predicted in models of gravitational collapse with thermal plus magnetic support and the lower values in l694 - 2 may be due to its lower mass or a slightly earlier evolutionary state . in either case
, it appears that both cores will form stars within a few @xmath0 yr . |
quantum processor stems from analogy between quantum computer and classical computer .
similarly as in classical computer a fixed classical gate array , which takes the program and data as input , performs operations on the data depending on the program , quantum processor consists of a fixed quantum gate array , a quantum system as data register carrying data state and another quantum system as program register containing information about operation to be performed on the data state .
when the data state and program state go through the fixed quantum gate array together , an unitary operation is executed on the data state depending on the program state .
but it is shown by nielsen and chuang @xcite that in quantum computer no such fixed quantum gate array can be realized in deterministic fashion .
suppose that the action of a fixed quantum gate array is operator @xmath0 , @xmath1 is a data state and @xmath2 is a program state , the total dynamics of the quantum processor can be described as @xcite : @xmath3 where @xmath4 is a unitary operation corresponding to the state @xmath5 . from eq .
( 1 ) it follows that distinct ( up to a global phase ) unitary operation @xmath4 requires different orthogonal state of program system .
it means that the dimension of the program system has to be infinite because even for a single qubit the set of operations is infinite .
therefore no universal quantum gate array can be constructed in deterministic fashion .
fortunately there is possibility to build a universal quantum gate array in a probabilistic fashion .
an example based on teleportation is given in ref.@xcite .
another explanation in terms of control - u operation is given to realize a one - qubit arbitrary rotation around the @xmath6 axis of a spin @xmath7 particle , in which the probability of success increases exponentially with the number of qubits of the program register @xcite .
on the other hand it is possible to construct a programmable quantum gate to implement an interesting subclass of unitary operations perfectly , accordingly it is possible to implement an arbitrary operation probabilisticly which can be decomposed into linear combinations of the subclass of unitary operations .
the problem is recently examined by hillery et .
generalized to the case of qudits of dimension @xmath8 @xcite .
also a mathematical description of programmable quantum processor is developed and its classes are analyzed by the same group@xcite .
we note that the processors in the previous works @xcite are from an input to an output . in quantum network one - input - to - multiple - output processor
is more desirable .
so it is worthwhile to extend the one - to - one processor to the case of multiple outputs . in this paper
we combine telecloning @xcite and programmable quantum gate array based on teleportation @xcite to discuss the possibility of multi - output quantum processor for restricted sets of operations commuting and anti - commuting with pauli operator @xmath9 . here
we encode operation into a particular multiparticle - entangled state instead of maximally two - qubit entangled state .
in addition , we slightly modify the procedure of the quantum gate array and redefine the location of the data register output .
similarly , when joint bell - state measurement gives some appropriate values , several identical states come out in the data register outputs .
the output data states are approximately - transformed versions of input data state . then these several approximate data outputs could be distributed to next different quantum processors for different tasks as described in @xcite .
though , compared with one - to - one processors in refs.@xcite , our multi - output processor has several disadvantages , it has extensive applications in remote control and distributed computation . for simplicity , we focus on single qubit operation and two data outputs in the following discussion .
we elaborate the details in sec.ii , and give our discussion in sec.iii and conclusion in sec.iv .
in this section , we will discuss a two - output quantum processor which produces two data register outputs for sets of operations on single qubit commuting or anti - commuting with pauli operator @xmath10 . in the processor the operation is encoded into a four - qubit entangled state and the quantum gate array is similar with one proposed by nielsen and chuang @xcite . after the bell - state measurement , two approximate data register outputs appear at the terminal with probability of @xmath7 , and the fidelity of each data register output is @xmath11 .
first we consider how the quantum gate array implement the operation commuting with operator @xmath10with two data outputs .
the set of operations has the form of @xmath12 where an arbitrary angle @xmath13 @xmath14 .
the gate array is illustrated in fig.1 .
the four - qubit entangled state @xmath15 to be codified into program state is formulated as : @xmath16 where qubit @xmath17 is a port qubit subjected to bell - state measurement m with the qubit as data input register , qubits a , b and c serve as data output register together , qubit a is an ancillary which is discarded eventually . the sates @xmath18 and @xmath19 are given , respectively as : @xmath20 @xmath21 where the subscript is neglected for concision , the qubits are arranged in accordance with the order a , b , c from the left to the right . in order to prepare a program state , the state @xmath15 is transformed through @xmath22 , that is , @xmath23 where only the port qubit p is altered by @xmath4 . for an arbitrary input data state @xmath24 , where @xmath25 and @xmath26 are
complex numbers and satisfy @xmath27 , the total state @xmath28 feeding into the gate array can be rewritten as : @xmath29 \nonumber \\ & = & \frac{1}{2}[\left| \phi ^{+}\right\rangle ( ae^{i\frac{\alpha } { 2}}\left| \phi _
{ 0}\right\rangle + be^{-i\frac{\alpha } { 2}}\left| \phi _ { 1}\right\rangle ) + \left| \phi ^{-}\right\rangle ( ae^{i\frac{\alpha } { 2}% } \left| \phi _ { 0}\right\rangle -be^{-i\frac{\alpha } { 2}}\left| \phi _ { 1}\right\rangle ) \nonumber \\ & & + \left| \psi ^{+}\right\rangle ( be^{i\frac{\alpha } { 2}}\left| \phi _ { 0}\right\rangle + ae^{-i\frac{\alpha } { 2}}\left| \phi _ { 1}\right\rangle ) + \left| \psi ^{-}\right\rangle ( be^{i\frac{\alpha } { 2}}\left| \phi _ { 0}\right\rangle -ae^{-i\frac{\alpha } { 2}}\left| \phi _ { 1}\right\rangle ) , \nonumber\end{aligned}\ ] ] where @xmath30 and @xmath31 are bell states .
when the measurement result from m gives an eigenvalue corresponding to @xmath32 , the three data qubits are projected onto a state @xmath33 dropping the ancillary qubit a and calculating the reduced density matrices on the qubits b and c respectively , we get @xmath34 however the correctly transformed version of the input data state should be @xmath35 .
we estimate the fidelity between the correct version and the resulting state as @xmath36 hence we obtain two data register outputs with fidelity of @xmath11 .
when the measurement result from m gives an eigenvalue corresponding to @xmath37 , the three data qubits are projected onto a state @xmath38 which can be converted into the state @xmath39 by applying the operation @xmath10 on all three data qubits .
again we can have the resulting state in the fidelity of @xmath11 with the correctly transformed version of data input state .
for the two possible results of the measurement m corresponding to @xmath40 and @xmath41 , the quantum gate array produces ruined state from which the state @xmath39 can not be retrieved effectively and the process of quantum gate array on the input data state fails .
consequently , from the results of measurement m we know whether the process is successful .
the four results of the measurement m are uniformly distributed at random , containing no information about @xmath4 and @xmath42 .
so our processor is no deterministic and succeeds with certain probability of @xmath7 .
because in the process there is only one data input but there are two data outputs , it implies a cloning transformation .
in fact , it involves a 1@xmath432 telecloning procedure .
each output data state has a fidelity of @xmath11 with the correctly transformed version of input data state .
thus our processor consisting of a fixed gate array illustrated in fig.1 and a particular program register and a data register can implement arbitrary rotation around @xmath6 axis with two data outputs with fidelity of @xmath11 and successful probability of @xmath7 .
next , we adjust our processor to implement operations of anti - commuting with pauli operator @xmath10 with two data outputs .
the set of operations has the form of @xmath44 where an arbitrary angle @xmath45 @xmath14 .
the quantum gate array modified is illustrated in fig.2 .
the initial state of the program register is reset as @xmath46 where the notion and the states @xmath47 and @xmath48 are defined as before . in order to create a program state corresponding to the operation @xmath49 , the state @xmath50 is subjected to a transformation @xmath51 : @xmath52 where only the port qubit is transformed by the operation @xmath49 .
the quantum gate array firstly flips the data qubit by pauli operator @xmath53 , then makes joint bell - state measurement m on data qubit d and port qubit p , and outputs four results . in detail , here is how it works : @xmath54 \nonumber \\ & = & \frac{1}{2}[\left| \phi ^{+}\right\rangle ( be^{i\frac{\beta } { 2}}\left| \phi _
{ 0}\right\rangle + ae^{-i\frac{\beta } { 2}}\left| \phi _ { 1}\right\rangle ) + \left| \phi ^{-}\right\rangle ( be^{i\frac{\beta } { 2}}\left| \phi _ { 0}\right\rangle -ae^{-i\frac{\beta } { 2}}\left| \phi _ { 1}\right\rangle ) \nonumber \\ & & + \left| \psi ^{+}\right\rangle ( be^{-i\frac{\beta } { 2}}\left| \phi _
{ 1}\right\rangle + ae^{i\frac{\beta } { 2}}\left| \phi _ { 0}\right\rangle ) + \left| \psi ^{-}\right\rangle ( be^{-i\frac{\beta } { 2}}\left| \phi _ { 1}\right\rangle -ae^{i\frac{\beta } { 2}}\left| \phi _ { 0}\right\rangle ) ] .
\nonumber\end{aligned}\ ] ] as has been discussed above , when the measurement result from m gives eigenvalues corresponding to @xmath32 or @xmath55 , the three data qubits are projected onto or can be effectively converted by pauli operator @xmath10 into a state @xmath56 similarly , dropping the ancillary qubit and tracing out qubit c or b , we have @xmath57 the correctly transformed version of input data state we desired is @xmath58 .
according to the eq.(9 ) we get the fidelities of the states @xmath59 and @xmath60 as @xmath61 for other two results of measurement m , the quantum gate array yields ruined state from which the state @xmath62 we desired can not be retrieved .
therefore for the operation anti - commuting with pauli operator @xmath10 the modified quantum processor yields two approximate transforms with successful probability of @xmath7 .
the fidelity of each successful data output is @xmath11 . in the same way
, the results of measurement give the information about whether the process is successful .
we have designed a two - output programmable processor respectively for sets of operations on single qubit commuting and anti - commuting with pauli operator @xmath10 .
the programmable processor implements the restricted operations on arbitrary data input state with successful probability of @xmath63 and provide two outputs with fidelity of @xmath11 .
it is noticed that the states @xmath64 and @xmath65 just are mixed states since @xmath66 .
that is , we apparently do not have complete knowledge about @xmath67 ( or @xmath62 ) .
we have presented how to construct a two - output programmable processor .
it is necessary to point out that the programmable quantum processor is approximate and probabilistic .
the set of operations which it applies on the input data state is limited .
it comes from three restrictions .
one is that in itself no universal quantum gate array exists in deterministic fashion .
the other is that it involves a telecloning procedure complying with no - cloning theorem @xcite .
the third lies in , during the creation of program state , executing the operation on the port qubit which later is subjected to the joint measurement with the data qubit in the bell basis .
simple algebra shows that @xmath68 , where @xmath69 denotes the transpose of @xmath4 with respect to the computational basis . unlike the quantum gate array proposed by nielsen and chuang @xcite
, we has to concern transforming @xmath70 to @xmath71 .
it is easily proved that for an unknown operation @xmath4 and an arbitrary state @xmath42 , the transformation can not be completed effectively .
nielsen - chuang processor @xcite produces one data output for arbitrary operation on the input data with successful probability of @xmath72 and the perfect fidelity of 1 , and vidal s @xcite one - to - one processor can be programmed to implement restricted set of operations on input data with the successful probability of @xmath73 and the same perfect fidelity .
compared with them our processor creates two data outputs for restricted set of operations on input in price of decreasing of the fidelity and probability of success .
but our processor is more powerful in remote control and distributed computation .
for example , if we expect to rotate two spin-@xmath7 particles in different states simultaneously , we can use our two - output programmable quantum processor as a program encoder and match one particle and one data output of the processor respectively , then feed them into another quantum gate arrays such as the nielsen - chuang @xcite and vidal s @xcite . on the other hand , a sequential scheme , in which a nielsen - chuang processor on teleportation @xcite is followed by a 1@xmath432 telecloning procedure ,
may achieve the same purpose as our two - output processor .
however , the sequential scheme accomplishes arbitrary transformation on input data state with less probability of success and the same fidelities compared with the two - output processor .
furthermore , the sequential scheme is more complex and more consumptive .
it not only consumes more1 ebit of entanglement and 2 bits of classical information communication than the two - output processor , but also involves more physical qubits and quantum operations than the two - output processor . in a sense ,
our two - output processor reduces the resource consumption and complexity scales of the information process .
if a general approximate 1@xmath432 cloning procedure follows the nielsen - chuang processor instead of 1@xmath432 telecloning procedure in the sequential scheme , besides the above disadvantages , whether the two data output registers of the sequential scheme can be separate from each other is to be queried . since our two - output processor takes advantage of the nonlocal cloning of telecloning procedure and has two distant outputs , it is suitable for remote control and distributed computation of multiple spatial location . at this point
our two - output processor is more useful than the sequential scheme
. the two - output programmable quantum processor can be straightforward generalized to the case of multi - output and multi - qubit operation .
it is easy to understand that the fidelity decreases with the increase of the numbers of data output and the probability decreases exponentially with the number of qubits the operation involved . whether there is more effective scheme of multi - output quantum processor is still an open question
in this paper , we have proposed a multi - output programmable quantum processor . as a result of three restrictions , our processor is approximate and probabilistic , and the set of operations which it can implement is restricted . as an example , we give a two - output programmable quantum processor for sets of operations commuting and anti - commuting with pauli operator @xmath10 , which is based on telecloning procedure . with probability of @xmath7 , it yields two approximately transformed data outputs in the fidelity of @xmath11 , which can be distributed to next processors for different task , respectively . and due to the property of entanglement ,
the data registers in terminal can separately be located , even far from the processor if it allows locc ( local operation and classical communication ) .
therefore the multi - output processor can be extended to distributed computation and remote control of multi - user . compared with a sequential scheme in which a nielsen - chuang processor is followed by an approximate cloning procedure ,
our processor reduces the resource consumption and complexity scales of the information process and has two distant outputs . but the probability of success and the fidelity of data output of the multi - output processor decrease with the numbers of qubits involved in the computation and of data outputs required .
it would be extremely intriguing to know if it is possible to build a more effective multi - output processor . | by combining telecloning and programmable quantum gate array presented by nielsen and chuang [ phys.rev.lett . 79 : 321(1997 ) ] , we propose a programmable quantum processor which can be programmed to implement restricted set of operations with several identical data outputs .
the outputs are approximately - transformed versions of input data .
the processor successes with certain probability .
pacs number(s ) : 03.67.lx |
as a distinctive type of the neutrino source , the core - collapse supernova ( sn ) provides a rich physical content that is lacking in the terrestrial environment . with its unique production and detection processes , the neutrino burst from a sn has long been considered as one of the promising tools for probing the unknown neutrino intrinsic parameters @xcite , in particular , the neutrino mass hierarchy and the tiny mixing angle @xmath0 .
however , difficulties also arise from the complexity caused by the unavoidable astrophysical uncertainties , which could lead to ambiguous interpretations of the observed events .
the existing paradigm for the neutrino msw flavor conversion in a sn has been established with the assumption of small neutrino self interaction .
however , a possible new paradigm has been shaped in recent years .
the argument is that , in addition to the standard msw effect , the neutrinos may encounter certain non - msw effects in the sn environment when the neutrino number density is extremely high .
these not so well - known effects may be independent of the msw effect and may introduce additional factors that alter the efficiency of the neutrino flavor transition .
the current consensus is that both types of effects could contribute to the flavor transition and should all be included in the more complete analysis of the sn neutrino signals .
even the standard treatment of the msw flavor conversion alone is not immune from ambiguity .
the uncertain local density profile of the sn matter and the undetermined neutrino spectral parameters , such as the average energy and the luminosity of each neutrino flavor , are the crucial factors involved in the msw effect of the sn neutrinos .
however , an analysis based on the simple global density model : @xmath1 , which is widely adopted in the literature , could lose the generality if the variation of the local density shape near the resonance is not taken into account . with these uncertain factors ,
it is worth while to investigate how the contribution from the standard msw effect should be modified in analyzing the sn neutrino signals , unless the overall contribution from the non - msw effects is much greater than that from the msw effect .
the resonant neutrino flavor transition in a sn and in earth would in principle give rise to observable signatures that reflect the properties of neutrino parameters .
the unknown local density profile of sn matter near the resonance takes part in the adiabaticity parameter of the level crossing , and plays a role in the determination of transition probabilities .
in fact , the adiabaticity of the level crossing could vary abruptly with the local density profile in certain neutrino parameter space .
in addition , the knowledge to the primary spectrum for each neutrino flavor is essential in assessing the efficiency of neutrino flavor conversion .
various mixing scenarios , which arise from the uncertain sn physics , the possible neutrino mass hierarchies , and the uncertain magnitude of @xmath0 lead to distinct neutrino survival probabilities .
an analysis of the observed neutrino signals should , in principle , be able to single out the working scenario .
the promising features of the multi - detector experiments have been well recognized @xcite . with the uncertain local density profile and the spectral parameters
, this work investigates the potential signatures that may be related to the contribution from the msw effect in a multi - detector analysis of the sn neutrinos . as suggested by the previous work @xcite
, the consequences due to the uncertain local density profile near the resonance can be accounted for by the adoption of independent and variable power - law density functions @xmath2 . however , the calculations in ref . 13 and 14
are performed under the assumption of nearly constant @xmath3 , which is similar to that adopted in the usual analysis based on @xmath4 . in the present work ,
the key improvement in dealing with the variable density profile is the establishment of variable @xmath3 , which reproduces reasonable densities that agree with the numerical simulations at a wide range of the radial location . with the more general parametrization of the density function ,
this work is further devoted to analyzing the expected neutrino events at two water cherenkov detectors .
focus is aimed at the modified contribution from the msw effect .
certain physical observables derived from the expected event rates at the two detectors are proposed as the discriminators of various transition scenarios for the msw resonance . in searching for the model - independent properties of the observables , three existing sn
neutrino spectral models proposed by the garching group @xcite and the lawrence livermore group@xcite are adopted in the calculation .
there is no definite way of modeling the practically unknown local variation of density profile . considering the narrow thickness of the resonant layer ,
as compared to that of the whole scope of the sn matter distribution , the density profile with a form of variable power - law in each resonant layer would be a reasonable simplification that accounts for the possible dynamical consequences .
the essence of this approximation is that the power characterizing the density profile in each resonant layer is allowed to vary independently : @xmath5 for the lower resonance layer , and @xmath6 for the higher resonance layer , where the factor @xmath3 denotes the magnitude of the profile @xmath7 , with @xmath8 or @xmath9 . in this present work ,
the density profile in the resonance layer is parameterized as @xmath10[\frac{r}{r_{0}}]^{n_{k}},\ ] ] where @xmath11 is a mass scale , and @xmath12 is a distance scale , which is conveniently taken as the solar radius , @xmath13 cm . with the variable @xmath14 ,
an important criterion for a reliable parametrization of the density profile is that the predicted density at a specific radial location agrees with that given by the numerical simulation , at least in order of magnitude .
this suggests that in eq.(1 ) the variation of @xmath14 is accompanied by the variation of @xmath11 at a different location .
one may thus write the density profile in a more convenient form as @xmath15[\frac{c}{c_{0}}][\frac{r}{r_{0}}]^{n_{k}},\ ] ] where @xmath16 is a constant mass scale , and @xmath17 is a reference density scale . in the typical model with a fixed power , @xmath18 ,
the magnitude of @xmath11 varies weakly@xcite : @xmath19 in the density range @xmath20 . for variable @xmath14 near a given location ,
however , the mass scale @xmath11 may vary by several orders of magnitude so as to reproduce reasonable density at this specific radial location .
note that with this variable parametrization of the density profile , the choice of @xmath16 does not alter the results since the variation of the mass scale is represented by @xmath11 . in this present analysis
we shall adopt the value @xmath21 , which is simply the mean value of @xmath11 in the typical density model .
the variables @xmath11 and @xmath22 may be rewritten as @xmath23 and @xmath24 , respectively , and the density profile becomes @xmath15 x z^{n_{k}}.\ ] ] by fitting the predicted densities with the numerical results@xcite at different radial locations , this parametrization leads to an approximate relation that regulates the variation of @xmath11 , @xmath22 , and @xmath14 : @xmath25 as summarized in the following , the variation of @xmath14 can lead to non - trivial effects on the crossing probabilities at the resonance .
one may first write the electron number density @xmath26 for a typical core collapse sn as @xmath27,\ ] ] where the electron number per baryon @xmath28 is adopted , and @xmath29 is the baryon mass .
the adiabaticity parameter for the resonant transition , defined as@xcite @xmath30 becomes @xmath31 [ \frac{\sin^2 2\theta_{ij}}{\cos 2\theta_{ij } } ] [ ( \delta m^2_{ij}/e)\frac{\cos 2\theta_{ij } } { \sqrt{2 } g_{f}(\frac{y_{e}}{m_{n}})(\frac{c_{0}}{r_{0}^{3}})}]^{\frac{1}{n_{k } } } r_{0 } z_{0}^{(1+\frac{3}{n_{k}})},\ ] ] where @xmath32 , @xmath33 is the mixing angle between the eigenstates @xmath34 and @xmath35 , @xmath36 is the fermi constant , @xmath37 is the neutrino energy , and @xmath38 indicates the location of resonance .
note that @xmath39 is evaluated at the resonance , where @xmath40 the adiabaticity parameter @xmath41 appears in the level crossing probability @xmath42 ( with @xmath8 or @xmath9 ) as : @xmath43- \exp[-\frac{\pi}{2 } \gamma_{k } \frac{f_{k}}{\sin^{2 } \theta_{ij } } ] } { 1-\exp[-\frac{\pi}{2 } \gamma_{k } \frac{f_{k}}{\sin^{2 } \theta_{ij}}]},\ ] ] where @xmath44 is the correction factor to a non - linear profile .
the origin of eq .
( 9 ) and the correction function @xmath44 can be found in , @xmath45 , ref .
. we shall adopt @xmath46 ev@xmath47 , @xmath48 ev@xmath47 , and @xmath49 in the calculation . with a given radial location for the resonance
, it can be verified that the adiabaticity parameter @xmath41 , and thus the neutrino survival probability , vary weakly with the energy @xmath37 , as compared to the influences from the variations of @xmath14 and @xmath0 .
this result suggests that one may simply adopt the average neutrino energies , @xmath45 , @xmath50 mev and @xmath51 mev , in the following calculation for the neutrino transition probability in sn .
the energy dependence would only appear in calculating the earth effect .
when the neutrinos arrive at earth , the survival probabilities for @xmath52 and @xmath53 are given respectively by @xmath54 @xmath55 for the normal hierarchy , and @xmath56 @xmath57 for the inverted hierarchy , where @xmath58 and @xmath59 represent the higher and the lower level crossing probabilities for @xmath60 , respectively , and @xmath61 is the element of the mixing matrix . in calculating the above level crossing probabilities , it should be emphasized that both the higher and the lower level crossings occur in the @xmath62 sector for the normal hierarchy , while the higher crossing occurs in the @xmath63 sector and the lower crossing occurs in the @xmath62 sector if the mass hierarchy is inverted .
one thus needs to calculate @xmath64 and @xmath65 in eq.(10 ) for the normal hierarchy , and @xmath66 in eq .
( 13 ) and @xmath65 in eq.(12 ) for the inverted hierarchy at the individual resonance .
note that since the antineutrinos do not encounter the resonance at the location where the neutrinos cross the lower resonance , the crossing probability @xmath67 for both hierarchies in eq.(11 ) and eq.(13 ) is vanishing : @xmath68 .
in addition , it can be verified that the adiabaticity parameter @xmath69 at the lower resonance is very large , @xmath70 .
this suggests that the lower resonance is adiabatic for both the mass hierarchies : @xmath71 , which leads to @xmath72 . on the other hand , to show that the variation of @xmath73 and @xmath0 leads to non - trivial structures of @xmath64 in the @xmath74 parameter space , we first estimate the typical density near the location of the higher resonance . using eqs.(5 ) and
( 8) , the density at resonance is given by @xmath75[y_{e}/m_{n}]},\ ] ] which leads to @xmath76 with small @xmath0 .
this density corresponds to an approximate location of resonance @xmath77 according to the numerical results @xcite .
it follows from eqs.(7 ) and ( 9 ) that the crossing probability @xmath64 is in general non - adiabatic ( @xmath78 ) for @xmath79 .
this leads to @xmath80 , as can be checked using eqs .
( 10 ) and ( 12 ) .
in addition , it can be verified that @xmath81 remains constant through out the parameter space with @xmath82 .
however , the probability @xmath83 varies with both @xmath73 and @xmath0 near @xmath84 since @xmath64 drops rapidly from @xmath78 ( non - adiabatic ) to @xmath85 ( adiabatic ) within a narrow region of the parameter space .
it follows that @xmath83 drops from @xmath86 to @xmath87 .
this property distinguishes @xmath83 from @xmath81 . as an illustration
, we show @xmath64 as a function of @xmath88 for @xmath89 and @xmath90 in fig . 1 .
.the survival probabilities for @xmath52 and @xmath53 , the possible neutrino mass hierarchy , and the function @xmath91 as predicted by the specific scenarios .
the uncertainty in the power of the local density profile can alter the predicted bound of @xmath0 through the function @xmath91 . [
cols="<,<,<,<,<,<,<,<,<",options="header " , ] [ ta2 ]
in addition to the brief pulse of the neutronization @xmath52 burst , all three flavors of neutrinos and antineutrinos are emitted from the sn through pair production processes during a typical time scale @xmath92s .
the primary sn neutrino spectrum , which is not pure thermal , is usually modeled as a pinched fermi - dirac distribution with several spectral parameters : @xmath93}+1},\ ] ] where @xmath94 is the luminosity of the neutrino flavor @xmath34 , @xmath95 is the effective temperature of @xmath34 inside the respective neutrinosphere , @xmath37 is the energy , @xmath96 is the normalization factor , and @xmath97 is the pinching parameter for @xmath98 .
notice that @xmath99 is assumed here , where @xmath100 , @xmath101 , @xmath102 , and @xmath103 .
the most significant difference between the garching models ( g1 , g2 ) and the lawrence livermore model ( ll ) is that the garching models exhibit a less hierarchical structure in the luminosity and in the average energy among different neutrino flavors , as shown in table [ ta2 ] .
we assume a core collapse sn at a distance 10 kpc away .
in a typical water cherenkov detector , the isotropical inverse @xmath104-decay , @xmath105 , is the most dominant process . in this work ,
other isotropical cc processes , @xmath106 and @xmath107 , are also included in the calculation . as for the directional events ,
we include the scattering events induced by @xmath34 and @xmath53 : @xmath108 , where @xmath109 .
the cross sections for @xmath110 and @xmath111 are adopted from ref.@xcite , and that for @xmath112 are adopted from ref.@xcite . as shown in fig .
3 , the neutrino incident angle at a detector , denoted as @xmath113 and @xmath114 for detector @xmath115 and @xmath116 , respectively , may be defined as the angle measured from the local zenith . for a given @xmath113 , the allowed range for @xmath114
is limited . if the angle formed by @xmath115 , the earth center , and @xmath116 is given by @xmath117 , then @xmath118 for @xmath119 , and @xmath120 for @xmath121 .
the density of earth matter is of the order @xmath122 , and its variation is much smaller than that of the sn matter .
one may adopt a simple step function@xcite as the approximation : @xmath123 @xmath124 for @xmath125 ( mantle ) , and @xmath126 @xmath124 for @xmath127 ( core ) , where @xmath128 is the earth radius .
for the purpose of illustration , one may analyze the expected results at two future detectors : the hyperkamiokande@xcite ( detector @xmath115 ) and the memphys@xcite ( detector @xmath116 ) .
the locations of the two detectors lead to @xmath129 .
since the neutrino flux penetrates significant amount of earth matter only when the incident angle at a detector is greater than @xmath130 , no observable difference is expected from the two detectors if both @xmath113 and @xmath114 are less than @xmath130 . in this case , both @xmath131 and @xmath132 reduce to @xmath133 , and all the observed differences between @xmath115 and @xmath116 vanish . given the variety of spectral models , the present analysis is aiming at seeking for model - independent properties that would help identify the working scenario .
simultaneous observation of the sn neutrinos at two terrestrial detectors would yield the most useful results when pronounced earth regeneration effect is observed at only one of the detectors .
given the two terrestrial detectors @xmath115 and @xmath116 , a proper analysis of the expected neutrino event rates at the two detectors would in principle reveal the signatures representing distinct scenarios of the neutrino properties and mixing schemes . to pave the way for the analysis , we first examine the energy dependence of the relative flux difference at the detectors @xmath115 and @xmath116 : @xmath134 for the @xmath52 flux , and @xmath135 for the @xmath53 flux . with @xmath136 and @xmath137 denoting the primary flux of @xmath34 and @xmath138 , respectively , one may relate the observed and the primary fluxes as @xmath139 @xmath140 the three scenarios then lead to several distinct relative flux differences : @xmath141 [ p_{2e}-\sin^{2}\theta_{12 } ] } { f^{0}_{\nu_{e}}[\sin^{2}\theta_{13}\sin^{2}\theta_{12}]+ f^{0}_{\nu_{x}}[1-\sin^{2}\theta_{13}\sin^{2}\theta_{12}]},\ ] ] @xmath142 [ p_{2e}-\sin^{2}\theta_{12 } ] } { f^{0}_{\nu_{e}}\sin^{2}\theta_{12}+ f^{0}_{\nu_{x}}\cos^{2}\theta_{12}},\ ] ] @xmath143 [ p_{2e}-\sin^{2}\theta_{12 } ] } { f^{0}_{\nu_{e}}\sin^{2}\theta_{12}+ f^{0}_{\nu_{x}}\cos^{2}\theta_{12}},\ ] ] @xmath144 [ \bar{p}_{2e}-\sin^{2}\theta_{12 } ] } { f^{0}_{\bar{\nu}_{e}}\cos^{2}\theta_{12}+ f^{0}_{\bar{\nu}_{x}}\sin^{2}\theta_{12}},\ ] ] @xmath145[\bar{p}_{2e}-\sin^{2}\theta_{12 } ] } { f^{0}_{\bar{\nu}_{e}}[\sin^{2}\theta_{13}\cos^{2}\theta_{12}]+f^{0}_{\bar{\nu}_{x } } [ 1-\sin^{2}\theta_{13}\sin^{2}\theta_{12}]},\ ] ] @xmath146 [ \bar{p}_{2e}-\sin^{2}\theta_{12 } ] } { f^{0}_{\bar{\nu}_{e}}\cos^{2}\theta_{12}+ f^{0}_{\bar{\nu}_{x}}\sin^{2}\theta_{12}}.\ ] ] it is seen that the ratios also depend on the chosen model for the primary @xmath34 and @xmath138 fluxes , as well as on @xmath131 and @xmath132 , which introduce the non - trivial energy dependence into @xmath147 and @xmath148 . as an illustration
, we show the variation of @xmath149 as a function of energy in fig .
4 , with @xmath150 ( detector @xmath115 is unshadowed ) and @xmath151 ( detector @xmath116 is shadowed by mantle and core ) .
the primary neutrino spectra are adopted from the g1 , g2 , and the ll models .
note that with the finite resolution power of the detector , not all the details of the curve can be observed in practice .
it is clear from eq.(30 ) that the energy dependence of @xmath149 originates from the factor @xmath152 .
the energy dependence of @xmath147 and @xmath148 , as shown in eqs.(27 - 32 ) , can be used to illustrate the properties of the observables that shall be proposed in this work . in terms of the event rates
, one may first examine the relative difference of the observed event rates per unit target mass for detectors @xmath115 and @xmath116 : @xmath153 the total event rates per unit target mass , @xmath154 and @xmath155 , consist of all the isotropical and directional events mentioned in the previous section , with @xmath156 where @xmath157 denotes the total target mass of a detector .
the expected event rates @xmath158 induced by the flavor @xmath34 ( or @xmath138 ) at the detector is given by @xmath159de,\ ] ] where @xmath160 is the target number at the detector , @xmath161 is the distance to the supernova , @xmath162 is the cross section for @xmath34 ( or @xmath138 ) in a particular reaction channel @xmath163 .
the detection efficiency is assumed to be one . in the following
, one may denote the resultant values of @xmath164 for the six scenarios as @xmath165 , where @xmath166 , @xmath167 . as an illustration
, we show in fig .
5 the expected values of @xmath164 for scenarios @xmath168 , @xmath169 , and @xmath167 . the reasonable estimation of the energy bin size is roughly a few mev for the detection of typical sn neutrinos@xcite .
we adopt an energy bin size of 5 mev in the figure .
note that for scenario @xmath169 , the deviation of @xmath164 from zero is @xmath170 or smaller , and the sign of @xmath164 is undetermined with the error bar .
the estimated width for @xmath164 is given by the different results from the three spectral models and the statistical errors from each .
as an example , one notes that the expected values of @xmath171 under ll , g1 , and g2 models in 25 mev @xmath17230 mev are estimated to be @xmath173 , @xmath174 , and @xmath175 , respectively . the resultant @xmath171
is then shown in fig .
5 with the combined width @xmath176 for 25 mev @xmath17230 mev .
note that there is no apparent energy dependence of the width .
even though some details of the energy - dependence in fig . 5 are lost , certain useful qualitative properties are still available .
it is seen that the earth effect could suppress the event rates to an observable level ( of order 10% or more ) for @xmath177 mev if the working scenario is @xmath168 or @xmath167 .
the sign and the magnitude of @xmath164 in @xmath177 mev can be used to separate the three scenarios into two groups : @xmath178 $ ] and @xmath169 .
it should be pointed out that for @xmath179 mev , the uncertainties wipe out the small deviation from zero and the sign of @xmath164 is undetermined for all scenarios .
more hints may be available from analyzing the directional events alone , which are predominantly the @xmath52-induced elastic scattering events .
however , the directional event rate is approximately @xmath180 orders of magnitude less than the isotropical event rate .
extracting the directional events effectively from the dominant background of the isotropical events may be quite challenging in practice . in the following analysis
, one assumes that the directional event rates in the forward cone with a narrow solid angle can be identified .
we may define another ratio based on the directional events per unit mass of target at the two detectors , @xmath181 where @xmath182 and @xmath183 denote the shadowed and the unshadowed directional events per unit target mass , respectively .
the properties of @xmath184 , as shown in fig .
6 , lead to new information that is unseen from @xmath164 .
one notes that two groups of scenarios , @xmath168 and @xmath185 $ ] , could be identified by the magnitude of @xmath184 at the high energy end since the magnitude could differ by a factor of three or more .
thus , one may conclude that the degeneracy between @xmath171 and @xmath186 could be removed by using the observable @xmath184 , and that the three scenarios could be reasonably distinguished by analyzing both @xmath164 and @xmath184 at the high energy end .
it should be pointed out that even though @xmath164 and @xmath184 represent energy - integrated quantities , they still exhibit general energy dependence with the chosen finite bin size , as can be seen from figs . 5 and 6 .
to qualitatively illustrate this general properties , one may approximate @xmath164 , for simplicity , as the event ratios that consist only of the predominant @xmath187 events . for scenario @xmath168 , eqs.(25 - 26 ) and
( 33 - 35 ) lead to @xmath188 \int ( \bar{p}_{2e}-\sin^{2 } \theta_{12 } ) \sigma ( e ) de } { \int ( f^{0}_{\bar{\nu}_{e } } \cos^{2 } \theta_{12 } + f^{0}_{\bar{\nu}_{x } } \sin^{2 } \theta_{12 } ) \sigma ( e ) de},\ ] ] where @xmath189 is the cross section . with the input parameters and a given model for the fluxes , @xmath190
if one calculates @xmath171 in a relative small step in energy , @xmath191 a few mev , the cross section @xmath192 , which is roughly @xmath193 , would be relatively smooth as compared to the function @xmath132 . as a consequence
, one may expect @xmath194 .
it is seen from eq.(30 ) that the qualitative behavior of the factor , @xmath195 , which dictates the general trend of the energy dependence of @xmath149 in fig .
4 , propagates to @xmath171 in fig .
the energy dependence of @xmath184 , and that of @xmath196 , which is introduced in the next subsection , can be realized in a similar way .
the observables , @xmath164 and @xmath184 , tend to flip signs near the peak of the spectrum ( @xmath197 mev ) , and become less useful for analyzing the abundant events induced by the neutrinos with energies near the peak
. it would be worth while to further investigate whether there is more to learn from the middle range of the spectrum .
theoretically , the earth regeneration effect is signaled by the oscillatory modulations of the observed spectra
. a quantitative analysis of the resultant event rates , however , may be quite subtle .
the resolution power of the detector is limited and the modulation may not alter the total event rates significantly enough to render an observable difference between the shadowed and the unshadowed events . to optimize the results for an analysis , neutrino event rates from a specific range of energy cut may be needed .
one may construct a double ratio with both the isotropical and the directional events : @xmath198/(n_{a}-n_{a})}{(n_{b}-n_{a})/n_{a}},\ ] ] where @xmath199 and @xmath200 denote the isotropical and the directional events , respectively .
depending on the time of a day , a given incident angle @xmath113 for detector @xmath115 corresponds to a definite incident angle @xmath114 for detector @xmath116 .
notice that the possible values of @xmath114 are regulated by eqs .
( 21 ) and ( 22 ) . if @xmath201 or @xmath202 , the angle @xmath114 is limited to a narrow range : @xmath203 . on the other hand
if @xmath204 , the allowed range for @xmath114 becomes maximized : @xmath205 , and the observables are expected to vary with a wide range of @xmath114 .
each of the two detectors may be unshadowed ( if @xmath206 ) , shadowed by the mantle ( if @xmath207 ) , or shadowed by the mantle and the core ( if @xmath208 ) . in any case , more useful hints from a simultaneous observation would be available if one of the incident angles is greater than @xmath130 and the other is less that @xmath130 . the observable @xmath196 may be used as a supplementary tool to the other two observables , @xmath164 and @xmath184 . as an illustration ,
the ratio @xmath196 is evaluated here for @xmath209 mev and analyzed as a function of @xmath114 , with @xmath210 and a conservatively assumed directional resolution of @xmath211 for detector @xmath116 , as shown in fig .
notice that the pointing accuracy of the incident neutrinos at the sk detector can reach the level to within a few degrees at 10 kpc @xcite .
the results for each scenario are expected to show no @xmath114-dependence for @xmath212 .
in addition , calculations show that scenarios @xmath169 leads to small deviations from zero : @xmath213 , which can be distinguished from the other two scenarios .
furthermore , @xmath214 and @xmath215 can also be distinguished by the difference in magnitude .
one would expect that a combined analysis of @xmath164 , @xmath184 , and @xmath196 should be able to provide certain useful constraint or prediction about the neutrino parameters .
the sn neutrino burst has long been considered as a promising tool for probing the neutrino mass hierarchy and the mixing angle @xmath0 .
the current consensus is that the contributions from both the msw and the non - msw effects should be included in analyzing the flavor transition of the sn neutrino signals .
however , the uncertain local density profile of the sn matter and the neutrino spectral parameters may lead to diverse interpretations to the neutrino signals from the contribution of the msw effect . to investigate the consequences resulting from these uncertainties ,
this work suggests a more general parametrization for the uncertain local density profile , and outlines a scheme using ratios of experimental observables as the discriminators for the possible scenarios that are related to the msw effect . as an illustration
, we analyze the expected event rates at two terrestrial cherenkov detectors . the consequences due to choices of the existing spectral models are included in the analysis .
the possible scenarios arising from the uncertainties are then examined through the observables @xmath164 , @xmath184 , and @xmath196 .
it is seen that the uncertain local density profile of the sn matter can alter the predictions for @xmath0 and the mass hierarchy from the contribution of the msw effect alone .
the predicted bound for @xmath0 is regulated by the proposed function @xmath91 .
the results in this work are derived from the expected event rates at two megaton - scale cherenkov detectors that would be available in the future .
analyses of the future sn neutrinos based on other types of detectors , such as the liquid argon and kton - scale scintillation detectors that are sensitive to distinct channels of the neutrino interactions , should also provide valuable information toward a better understanding of the neutrino parameters and the sn physics . a detailed analysis based on all the three types of detector
is performed in , @xmath45 , ref .
6 , in which seven sn and neutrino parameters are taken into consideration . to extend the present work
, it would be intriguing to investigate how the uncertain density profile would impact the outcome of the analysis based on the detectors other than the cherenkov ones . in shaping a possible new paradigm for the neutrino flavor conversion in a sn
, recent studies and simulations suggest that the shock wave propagation and certain types of non - msw flavor conversion may occur in the complex environment of a sn at different space and time scales if the proper physical conditions are met .
these effects , such as the neutrino self coupling in dense media@xcite and the neutrino flavor de - polarization associated with the after shock turbulence@xcite , may have different origins from that of the msw effect and would impact the efficiency of neutrino flavor transition . it is suggested that the neutrino self interactions induce collective flavor transitions , which may occur before or in the same region as the resonant flavor transition .
the size of modification to the original neutrino flux depends on the mass hierarchy , as well as on the mixing angle @xmath0 . as an example , one considers the simplest case when the self - interaction effects can be factored out from the ordinary msw effects . for the inverted hierarchy
, it is suggested that if @xmath0 is nonzero , the collective pair - conversion of the type @xmath216 can be triggered before the ordinary msw effect occurs@xcite .
thus , even in the simplest case , the altered primary spectra could impact the validity and the usefulness of the observables proposed in this work . a better understanding of the modified neutrino spectra resulting from the collective effects would help in establishing more convenient and useful observables since the knowledge of the neutrino spectra before and after the msw effect is essential for the present analysis . note that the treatment of the variable density profile for the msw effect , as that proposed in this work , is unaffected by the non - msw effects . the shock wave propagation , on the other hand , may alter the density profile of the sn matter and induce certain non - linear effects that modify the neutrino survival probabilities and the neutrino spectrum .
in particular , there is a possibility that more than one msw - resonance related to the same scale of mass - squared difference could arise due to the shock wave propagation .
the effects of the shock wave on certain physical observables , with the emphasis on the time - dependent properties , are analyzed@xcite .
although the proposed scheme in the present work is limited to the discussion without considering specific scenario due to the shock effects , it takes into account the possible varying density profile , and is not limited to whether the profile is actually primary .
in addition , the quantities @xmath164 , @xmath184 , and @xmath196 are proposed in this work to examine the time - integrated behaviors of the event rates . as an extension to this present analysis
, it would be intriguing to further investigate weather the time structure of the shock waves effects and the possible multiple msw - resonance due to the shock effect can alter the time - integrated properties of the proposed quantities to an observable level . even with the unknown neutrino parameters , the uncertain model for the sn neutrino spectrum , and the possible statistical errors , the quantities @xmath164 , @xmath184 , and @xmath196 might still be able to provide another potential means for probing the sn physics if they reveal time - independent signatures that are related to the shock wave propagation . on the other hand , if the shock signatures are unobservable with these time - integrated quantities , then these quantities might be able to provide insights to the neutrino parameters without being affected by the shock waves .
it should be emphasized that the present work is devoted to effects that are related to the msw resonant flavor transition in a sn .
the focus is aimed at the possible modification to the results of msw effect due to the uncertain density profile .
this modification is expected to have certain impact to the complete analysis based on both the msw effect and the non - msw effects .
the whole details of the non - msw effects , as well as the interplay between the msw and the non - msw consequences , are still not fully understood .
the neutrino flavor conversion in a sn remains as a complicated problem that awaits for more complete answers .
whether the flavor transition arising from one origin would dominate over that from the others depends on the uncertain spectral parameters , the undetermined neutrino intrinsic properties , and the physical conditions that the neutrinos encounter at different stages of the sn environment .
in addition , the reliability of the analysis relies on whether the msw and non - msw effects can be analyzed separately .
however , these unrelated types of effects decouple only when specific physical conditions are met . in any case ,
analysis of the non - trivial msw flavor conversion due to the possible drastic variation of the local density profile , as discussed in this work , should provide certain insight in probing the neutrino parameters and in shaping a more convincing paradigm for the neutrino flavor conversion in the sn . | the unknown neutrino parameters may leave detectable signatures in the supernova ( sn ) neutrino flux . however , even the contribution from the msw flavor transition alone could cause ambiguity in the interpretation to the neutrino signals because of the uncertain local density profile of the sn matter and the model - dependent sn neutrino spectral parameters .
a specific parametrization to the unknown local density profile is proposed in this work , and the contribution from the standard msw effect is investigated through a multi - detector analysis of the sn neutrinos . in establishing the model - independent scheme ,
results based on the existing spectral models are included .
the limitation of the analysis is also discussed . |
exploring the properties of nuclei with unusual proton to neutron ratios is nowadays one of the major tasks and most active topics in nuclear structure physics , both theoretical and experimentally @xcite . in this context , the beta - decay properties of these exotic objects are of extreme importance in two different aspects .
first , most nuclei of astrophysical interest @xcite are those far from stability and their beta - decay rates have to be estimated theoretically .
second , beta - decay properties and nuclear structure are intimately related .
it is clear that a precise and reliable description of the ground state of the parent nucleus and of the states populated in the daughter nucleus is necessary to obtain a good description of the @xmath3strength distribution , and vice versa , failures to describe such distributions would indicate that an improvement of the theoretical formalism is needed .
microscopic models to describe the @xmath4-strength based on spherical single - particle wave functions and energies with pairing and residual interactions treated in random phase approximation ( rpa ) or quasiparticle random phase approximation ( qrpa ) were first studied in ref .
extensions of these models to deal with deformed nuclei were done in ref .
@xcite , where a nilsson potential was used to generate single - particle orbitals .
subsequent extensions including woods - saxon type potentials @xcite , residual interactions in the particle - particle channel @xcite , consistent hartree - fock ( hf ) mean field with residual interactions treated in tamm dancoff approximation @xcite , selfconsistent approaches in spherical neutron - rich nuclei @xcite and based on an energy - density functional @xcite , can be also found in the literature . in a previous work @xcite we studied ground state and @xmath4-decay properties of exotic nuclei on the basis of a deformed selfconsistent hf+bcs+qrpa calculation with density dependent effective interactions of skyrme type .
this is a well founded approach that has been very successful in the description of spherical and deformed nuclei within the valley of stability @xcite . in this method
once the parameters of the effective skyrme interaction are determined , basically by fits to global properties in spherical nuclei over the nuclear chart , and the gap parameters of the pairing force are specified , there are no free parameters left . both the residual interaction and the mean field
are consistently obtained from the same 2-body force .
this is therefore a reliable method , suitable for extrapolations into the unstable regions approaching the drip lines .
it is worth investigating whether these powerful tools designed to account for the properties of stable nuclei are still valid when approaching the drip lines .
the residual interaction introduced in refs .
@xcite , consistent with the mean field , acts in the particle - hole channel ( @xmath5 ) .
however , it has been pointed out ( see for instance ref .
@xcite and refs .
therein ) that for a complete description of the @xmath6 and @xmath7 strengths , the inclusion of the particle - particle ( @xmath8 ) residual interaction is required .
we gave in @xcite an example of the sensitivity of the results to the inclusion of such a force but a complete calculation within our formalism including this force is still missing .
one of the purposes of this work is to extend our previous calculations by including a residual separable force in the @xmath8 channel and to study its effect in a systematic way to distinguish what is general and what is particular in the behavior of the nuclear response . to this end
we include in the qrpa calculation a separable neutron - proton pairing force as part of the residual interaction .
following the same criteria as in our previous work @xcite , we apply this formalism to the study of a series of proton rich isotopes approaching @xmath9 in the mass region @xmath10 .
there are several reasons why this is a region of special interest to study @xmath4-decay .
one is that the @xmath11-value of the decay ( @xmath12 ) is quite large in proton rich nuclei @xcite .
this means that a large fraction of the gamow - teller ( gt ) strength is contained within the @xmath12 window .
therefore , one can investigate directly the structure of the gt strength distribution by @xmath4-decay measurements without dealing with other indirect methods to extract the gt strength such as @xmath13 or @xmath14 charge exchange reactions .
another aspect to stress is that the mass region @xmath15 is at the border or beyond the scope of the full shell model calculations .
predictions for the strength distributions , half - lives , and other decay properties obtained from the present formalism are of special relevance in this mass region since they may be at the moment the most reliable calculations .
the present results can then be used as a guidance to further experimental searches .
it is also worth mentioning that this mass region is known to present particular structure effects , such as shape coexistence , which are the result of a delicate balance of competing nuclear shapes @xcite .
it is then important to explore whether the dependence of the beta - decay properties on the deformation of the parent nucleus could be used as an additional piece of information to elucidate the shape of the nucleus . as a last point
, we mention that , by approaching systematically the @xmath9 isotopes in various isotopic chains , we are placed in optimum conditions to observe whether agreement with experiment breaks down at some point as we approach @xmath9 .
these isotopic chains are a laboratory to test the validity of our formalism and to look for failures related to the appearance of new physics such as neutron - proton pairing correlations not taken into account explicitly in the present mean field calculations , that may be important in @xmath9 nuclei . in this work
we extend the study of refs .
@xcite by including a residual neutron - proton pairing interaction in the @xmath16 channel , by studying the influence of usual bcs pairing correlations in the @xmath9 isotopes , and by discussing the similarities between the gt strength distributions and the spin @xmath2 strength distributions , which are the @xmath17 isospin counterparts of the @xmath18 gt transitions . in section 2
, we present a brief summary containing the basic points in our theoretical description .
section 3 contains the results obtained for the gt strength distributions with a discussion of the dependence on the residual interaction , pairing correlations , and deformation .
we also compare our results to the experimental available information .
spin @xmath2 strength distributions are studied in section 4 discussing their analogies with the gt strength distributions .
the conclusions are given in section v.
in this section we summarize briefly the theory involved in the microscopic calculations presented in the next sections . more details can be found in refs .
our method consists in a selfconsistent formalism based on a deformed hartree - fock mean field obtained with a skyrme interaction including pairing correlations in the bcs approximation .
we consider in this paper the force sg2 @xcite of van giai and sagawa , that has been successfully tested against spin and isospin excitations in spherical @xcite and deformed nuclei @xcite .
comparison to calculations obtained with other skyrme forces have been made in refs .
@xcite , showing that the results do not differ in a significant way .
the single particle energies , wave functions , and occupations probabilities are generated from this mean field .
for the solution of the hf equations we follow the mcmaster procedure that is based in the formalism developed in ref.@xcite as described in ref.@xcite .
time reversal and axial symmetry are assumed .
the single - particle wave functions are expanded in terms of the eigenstates of an axially symmetric harmonic oscillator in cylindrical coordinates .
we use eleven major shells .
the method also includes pairing between like nucleons in the bcs approximation with fixed gap parameters for protons @xmath19 and neutrons @xmath20 , which are determined phenomenologically from the odd - even mass differences through a symmetric five term formula involving the experimental binding energies @xcite .
the values used in this work are the same as those given in ref .
@xcite . in a previous work
@xcite we analyzed the energy surfaces as a function of deformation for all the isotopes under study here .
for that purpose , we performed constrained hf calculations with a quadrupole constraint @xcite and we minimized the hf energy under the constraint of keeping fixed the nuclear deformation .
calculations in this paper are performed for the equilibrium shapes of each nucleus , that is , for the solutions , in general deformed , for which we obtained minima in the energy surfaces .
most of these nuclei present oblate and prolate equilibrium shapes @xcite that are very close in energy .
we add to the mean field a spin - isospin residual interaction , which is expected to be the most important residual interaction to describe gt transitions .
this interaction contains two parts .
the particle - hole part is responsible for the position and structure of the gt resonance @xcite and is derived selfconsistently from the same energy density functional ( and skyrme interaction ) as the hf equation , in terms of the second derivatives of the energy density functional with respect to the one - body densities @xcite .
the residual interaction is finally written in a separable form by averaging the landau - migdal resulting force over the nuclear volume , @xmath21 where @xmath22 the coupling strength is given by @xcite @xmath23 and it is completely determined by the skyrme parameters @xmath24 , the nuclear radius @xmath25 , and the fermi momentum @xmath26 , obtained with the same skyrme force for the given nucleus . the particle - particle part is a neutron - proton pairing force in the @xmath16 coupling channel .
we introduce this interaction in the usual way @xcite , that is , in terms of a separable force with a coupling constant @xmath27 , which is fitted to the phenomenology , @xmath28 where @xmath29 the two forces @xmath5 and @xmath8 in eqs .
( [ vph ] ) and ( [ vpp ] ) are defined with a positive and a negative sign , respectively , according to their repulsive and attractive character , so that the coupling strengths @xmath30 and @xmath31 take positive values .
the peak of the gt resonance is almost insensitive to the @xmath8 force and @xmath27 is usually adjusted to reproduce the half - lives @xcite .
however , one should be careful with the choice of this coupling constant . since the @xmath8 force is introduced independently of the mean field
, if @xmath27 is strong enough it may happen that the qrpa collapses , because the condition that the ground state be stable against the corresponding mode is not fulfilled .
this happens because the @xmath8 force , being an attractive force , makes the gt strength to be pushed down to lower energies with increasing values of @xmath27 .
a careful search of the optimal strength can certainly be done for each particular case , but this is not our purpose here .
instead , we have chosen the same coupling constant for all nuclei considered here .
this value has been obtained under the requirements of improving in general the agreement with the experimental half - lives while being still far from the values leading to the collapse .
this will be discussed in more detail in the next section . the proton - neutron quasiparticle random phase approximation ( qrpa ) phonon operator for gt excitations in even - even nuclei is written as @xmath32\ , , \label{phon}\ ] ] where @xmath33 are quasiparticle creation ( annihilation ) operators , @xmath34 are the excitation energies , and @xmath35 the forward and backward amplitudes , respectively .
the equations of motion are solved in the proton - neutron qrpa . from the qrpa equations
the forward and backward amplitudes are obtained as @xcite @xmath36 \ , , \label{xwa}\ ] ] @xmath37 \ , , \label{ywa}\ ] ] with @xmath38 the two - quasiparticle excitation energy in terms of the quasiparticle energies @xmath39 .
@xmath40 are given by @xmath41 @xmath42 @xmath43 @xmath44 with @xmath45 where @xmath46s are occupation amplitudes ( @xmath47 ) and @xmath48 spin matrix elements connecting neutron and proton states with spin operators @xmath49 the solution of the qrpa equations can be found solving first a dispersion relation , which is now of fourth order in the excitation energies @xmath50 .
then , for each value of the energy the amplitudes @xmath51 are determined by using the normalization condition of the phonon amplitudes @xmath52 = 1 \ , .\ ] ] the technical procedure to solve these qrpa equations is well described in ref .
@xcite . in the intrinsic frame
the gt transition amplitudes connecting the qrpa ground state @xmath53 to one phonon states @xmath54 , are given by @xmath55 the ikeda sum rule is always fulfilled in our calculations @xmath56 = 3(n - z ) \ , .\ ] ]
the bulk properties of the isotope chains considered here were already studied in ref . @xcite .
binding energies , charge radii , quadrupole moments , and moments of inertia were discussed and compared successfully to available experimental data . here
we will concentrate on the decay properties . in this section
we discuss the results obtained for the gt strength distributions , half - lives , and summed strengths in the proton rich isotopes belonging to the @xmath15 mass region .
the results correspond to qrpa calculations with the skyrme force sg2 and they have been performed for the nuclear shapes that minimize the hf energy . figures showing the gt strength distributions are plotted versus the excitation energy of daughter nucleus .
the distributions of the gt strength have been folded with @xmath57 mev width gaussians to facilitate the comparison among the various calculations , so that the original discrete spectrum is transformed into a continuous profile .
these distributions are given in units of @xmath58 and one should keep in mind that a quenching of the @xmath59 factor , typically @xmath60 is expected on the basis of the observed quenching in charge exchange reactions and spin @xmath2 transitions in stable nuclei , where @xmath61 is also known to be approximately @xmath62 . therefore , a reduction factor of about two is expected in these strength distributions in order to compare with experiment .
this factor is of course taken into account when comparing to experimental half - lives .
figure 1 illustrates the effect of the residual interactions on the uncorrelated 2-quasiparticle calculation .
the calculations are done for the oblate and prolate shapes of @xmath63kr
. the coupling strength of the @xmath5 residual interaction @xmath64 is obtained from eq .
( [ chiph ] ) , and its value for @xmath65 and skyrme force sg2 is @xmath66 mev .
the coupling strength of the @xmath8 residual interaction is varied from @xmath67 to @xmath68 mev .
if we compare first the calculation with only the @xmath5 residual interaction ( dotted line ) to the uncorrelated one ( thin solid line ) , we can see that the repulsive @xmath5 force introduces two types of effects .
one is a shift of the gt strength to higher excitation energies with the corresponding displacement of the position of the gt resonance .
the other is a reduction of the total gt strength
. obviously these effects are more pronounced as we increase the value of the coupling strength @xmath64 .
if we now introduce a @xmath8 residual interaction ( dashed and solid lines ) , we can see that its effect , being an attractive force , is to shift the strength to lower excitation energies , reducing the total gt strength as well .
the shift and reduction of strength is more pronounced at high excitation energies and the position of the gt resonance is hardly affected by this interaction .
the gt strength is also pushed a bit by the @xmath8 interaction to lower energies in the low energy excitation region .
this effect , although small , is of great relevance in the calculation of the @xmath69 half - lives , which are only sensitive to the distribution of the strength contained in the energy region below the @xmath12 window . by comparing the curves obtained from @xmath70 mev ( dashed line ) and @xmath68 mev ( solid line )
, we can also see how these effects are more pronounced as we increase the value of the coupling strength .
this is the reason why the usual procedure to fit these two coupling strengths @xmath64 and @xmath27 is as follows : first one chooses @xmath64 to reproduce the position of the gt resonance , usually determined from charge exchange reactions @xmath13 and @xmath14 , and then one chooses @xmath27 to reproduce the half - life . in our work , since the value of @xmath64 is determined selfconsistently and the known gt resonances are reasonably well described @xcite , we do not carry out this case by case fitting procedure .
however , to determine the value of @xmath27 , we calculate first gt strength distributions and half - lives and compare the latter with experiment to extract a value that producing a reasonable agreement is still within the range of values compatible with the correct treatment of the qrpa .
in figure 2 we can see the result of the calculation of the half - lives in @xmath71sr as a function of the coupling strength @xmath27 .
the half - lives are obtained from the familiar expression , @xmath72 where @xmath73 are the fermi integrals .
note that in our previous work @xcite , we used for these integrals the values tabulated in @xcite . in the present calculations we compute the fermi integrals numerically for each @xmath74 values . by this procedure
we get more accurate results .
we use @xmath75 s and include standard effective factors @xmath76 ^{2}=\left [ 0.77\left ( g_{a}/g_{v}\right ) _ { free}\right ]
^{2}=0.90 \ , . \label{quen}\ ] ] the half - life decreases with increasing values of @xmath27 .
this is clear because as @xmath27 increases the strength becomes more concentrated at low excitation energies below @xmath12 and therefore the half - lives are smaller .
this is true up to values around @xmath77 mev , where the collapse of the qrpa takes place . in the case of @xmath71sr , we get an optimum value of @xmath78 mev and @xmath79 mev for the oblate and prolate shapes , respectively
. this value will of course depend , among other factors , on the nucleus , shape , and skyrme interaction and a case by case fitting procedure could be carried out .
nevertheless , we have done calculations for other cases and found that , in general , values around @xmath80 mev improve the agreement with experiment in most cases and since this is a valid value far from collapse , we have chosen @xmath80 mev as the value of the coupling constant of the @xmath8 residual interaction . in ref .
@xcite homma and collaborators studied @xmath3decay properties using nilsson+bcs+qrpa with @xmath5 and @xmath8 separable residual interactions .
they considered a wide range of nuclei to extract phenomenologically the coupling strength of such residual interactions by fits to experimental half - lives .
they obtained the following dependence on the mass number : @xmath81 mev , @xmath82 mev . for @xmath10
this gives a @xmath5 strength @xmath83 mev and a @xmath8 strength @xmath78 mev .
the @xmath5 strength is somewhat smaller than the one obtained from eq .
( [ chiph ] ) , derived selfconsistently from our mean field .
the @xmath8 strength is also smaller than the value adopted here .
this is consistent with the trend observed in fig . 1 of homma _ et al .
_ @xcite that shows that for decreasing values of @xmath64 one needs smaller values of @xmath27 . on the other hand ,
a discussion of the adequacy of our @xmath5 strength is also demonstrated in ref .
@xcite , where we compared our results on the position of the gt resonance with experimental data from @xmath13 and @xmath14 reactions in this mass region .
the fact that our @xmath5 and @xmath8 strengths are somewhat different from those in ref .
@xcite is not surprising since the mean fields are also different .
as already mentioned in the introduction , our theoretical treatment does not explicitly include neutron - proton pairing correlations in the mean field .
thus , our quasiparticle basis only includes neutron - neutron and proton - proton pairing correlations in the bcs approach . in principle
, one could extend the bcs treatment to include also neutron - proton pairing correlations in the mean field .
this may be important particularly for n = z nuclei . in ref .
@xcite we studied bulk properties ( binding energies , charge radii , quadrupole moments , moments of inertia , ... ) of the nuclei considered here and found that the agreement between theory and experiment is as good for the @xmath9 as for the @xmath84 isotopes .
we concluded that the effect of neutron - proton pairing correlations on these bulk properties is roughly taken into account by the use of the phenomenological gap parameters @xmath85 the main effect of taking into account neutron - proton pairing in the hf+bcs calculation would be to increase the diffuseness of the fermi surface @xcite .
this diffuseness is proportional to the gap parameters .
it is therefore interesting to study the sensitivity of the gt strength to the gap parameters in the @xmath9 nuclei @xmath86ge , @xmath87se , @xmath88kr , and @xmath71sr . to this end
we compare in fig .
3 the qrpa results obtained with the sg2 force in those nuclei for various values of the gap parameters differing by @xmath89 mev from the values extracted from the phenomenology . to make the discussion easier we did not include the @xmath8 residual interaction in those figures . the main effect of the bcs correlations is to create new transitions that were forbidden in the absence of such correlations . since the occupation probabilities are now different from 0 or 1 , the already existing peaks at @xmath90 will decrease when @xmath91 , while new strength will appear at other energies and will increase with increasing gap parameters .
this new strength is in general located at high excitation energy , while the strength already present at @xmath90 is mainly concentrated at lower energy . as a consequence , the main effect of increasing the fermi diffuseness is to smooth out the profile of the gt strength distribution , increasing the strength at high energies and decreasing the strength at low energies .
it is clear from fig .
3 that each particular case has its own peculiarities but the general trend described above plays the major role .
it is also interesting to remark that in the high energy tail beyond the gt resonance , the effects of increasing @xmath92 and increasing @xmath27 are opposite . as we mentioned before ,
increasing @xmath27 tends to deplete those tails .
the role of deformation on the gt strength distributions can be summarized , in general , in two types of effects .
first , deformation breaks down the degeneracy of the spherical shells and this implies that the gt strength distributions corresponding to deformed nuclear shapes will be much more fragmented than the corresponding spherical ones .
second , the energy levels of deformed orbitals coming from different spherical shells cross each other in a way that depends on the magnitude of the quadrupole deformation as well as on the oblate or prolate character .
this level crossing may lead in some instances to similar profiles in the gt strength distributions of the various coexisting nuclear shapes but in other cases it may lead to sizeable differences between the gt strength distributions corresponding to different shapes of the same parent nucleus .
this fact can be exploited to gain information on what is the nuclear shape of a nucleus by just looking at the structure of its @xmath4-decay .
4 summarizes the main results of this work .
they correspond to the gt strength distributions obtained from our hf+bcs+qrpa with sg2 in the isotope chains of ge , se , kr , and sr .
we can see in this figure the results obtained for the possible nuclear shapes of the isotopes ranging from @xmath9 to @xmath93 . as usual the strengths are in units of @xmath58 and are plotted versus the excitation energy of the daughter nucleus .
we can see that in general the @xmath9 isotopes of each chain contain the maximum strength as it corresponds to the most unstable nuclei .
this strength becomes smaller and smaller as we increase the number of neutrons approaching the stable isotopes .
the excitation energy of the gt resonance also decreases with the number of neutrons . if we look in fig .
3 the case of ge and se isotopes , we can see that there are no signatures of the actual character , oblate or prolate , of the deformation of the parent nucleus : both equilibrium shapes produce similar gt profiles independently of whether it is oblate or prolate .
then , it can be concluded that in these isotopes one can not use gt distributions to distinguish between the two coexistent shapes predicted by the theory .
on the other hand , the figures corresponding to kr and sr isotopes , show differences in the gt profiles of the various shapes for each isotope that can be easily identified even within the @xmath12 window .
a case by case analysis allows one to conclude that the most favorable candidates to look for deformation effects based on @xmath4-decay measurements are @xmath63kr and @xmath94sr . in these isotopes
the gt strengths within the @xmath12 window are different enough to distinguish between different equilibrium shapes . on the contrary , there are cases like @xmath88kr where though the large @xmath12 value ( @xmath95 mev ) make it worth to investigate experimentally , the measured strength may not be conclusive on the shape because oblate and prolate shapes produce similar profiles .
other isotopes like @xmath96kr have small @xmath12 values that will not allow a clear conclusion either , while @xmath63kr seems to be the best kr candidate .
it has a @xmath12 window ( 3.1 mev ) big enough to distinguish between a continuously increasing profile , as the prolate shape predicts , or a completely developed bump structure peaked at around 1.5 mev and vanishing at about the @xmath12 value , as the oblate shape predicts .
a similar situation happens in the case of @xmath71sr ( @xmath97 mev ) .
the oblate shape produces a peak centered at around 1 mev and a second one centered at 6 mev , while the prolate shape produces a gt distribution that increases almost continuously up to 6 mev .
the case of @xmath98sr ( @xmath99 mev ) is again a clear case where one can distinguish between the bump structure generated by the spherical shape or the continuously increasing pattern of the prolate case .
the same is true for the @xmath100sr isotope , although in this case @xmath12 is not so big ( @xmath101 mev ) . in this subsection
we compare to experiment our calculated @xmath12 and total @xmath102 half - lives @xmath103 , as well as a few cases where the gt strength distribution have been partially measured by @xmath104-decay .
we also discuss the gt strengths contained within the @xmath12 windows .
the difference with respect to the values quoted in our previous work @xcite is that now the @xmath8 interaction is included .
there is also a minor difference in the calculation of @xmath12 , @xmath105 the @xmath12 values in tables 1 - 4 of ref .
@xcite were calculated approximating the lowest quasiparticle energies @xmath106 and @xmath107 by the gap parameters @xmath108 and @xmath109 , respectively , while @xmath12 values in table 1 here are calculated using their actual values @xmath110 .
we can see in table 1 these results obtained from qrpa calculations with the skyrme force sg2 and for the different shapes , oblate , prolate or spherical , where the minima occur for each isotope .
we did not include the stable nuclei , all having @xmath12 less than zero and infinite half - lives .
it should also be mentioned that the total gt strength below @xmath12 has been reduced by a quenching factor 0.6 to be consistent with the same quenching used in the calculation of the half - lives in eq .
( [ t12 ] ) .
we get a very good agreement with the measured @xmath12 values in practically all cases and the values obtained with the various shapes are quite similar .
the half - lives are also well reproduced in most cases . only in the most stable isotopes , where the half - lives are very large , we find some noticeable discrepancies but this is not so relevant because in these cases the @xmath12 values are very small and therefore , the half - lives are only sensitive to a tiny region of the gt tail at low energies . in other cases
the agreement is very reasonable .
the sums of the gt strength up to the @xmath12 value do not differ much from one shape to another although this does not mean that the structure of the profiles are equivalent . as we have seen in the last subsection , there are cases , @xmath63kr and @xmath94sr , whose profiles can be easily distinguished although the final summed strength is very similar
. table 2 contains experimental information on gt summed strengths in the @xmath9 nuclei @xmath88kr @xcite and @xmath71sr @xcite up to different values of the excitation energy , always below @xmath12 .
they are compared to our theoretical calculations for the two equilibrium shapes using the same quenching factor as in eq .
( [ quen ] ) . from this comparison
an oblate shape for @xmath88kr and a prolate shape for @xmath71sr seems to be favored , but this is not yet sufficient for a conclusive answer . in any case
it would be worth to extend the measurements in @xmath71sr to higher excitation energies because it is a very good example where deformation effects are visible ( see fig . 4 for @xmath71sr ) .
on the other hand , the extension of these measurements to the case of @xmath88kr may not be so relevant since one could not distinguish one shape from another ( see fig . 4 for @xmath88kr ) .
the spin @xmath2 transitions are the @xmath111 isospin counterparts of the @xmath112 gt transitions .
the study carried out in @xcite for several isotope chains showed that the @xmath2 strength depends on deformation , not only for orbital but also for spin excitations .
thus , on general grounds , one may argue that the deformation dependence of the gt strength discussed earlier should be comparable to that of the spin @xmath2 strength considered in @xcite .
this is expected to be particularly the case when @xmath9 . in a similar way to fig . 4 for the gt distributions , we show in fig .
5 the profiles of the spin @xmath2 strength distributions for the same isotope chains .
the results correspond to the selfconsistent hf+bcs+qrpa calculations with the sg2 interaction .
the details of these calculations are described in @xcite and closely follow those of the gt strength except for the @xmath111 character of the @xmath2 operator .
similarly to the gt strength distributions in the ge and se isotope chains in fig .
4 , we can see that the @xmath113 strength distributions for these two isotopes in fig .
5 have similar structures for the oblate and prolate shapes .
they have a big resonance located at the same energy around 10 mev and a small bump at about 5 mev .
however , the strength contained in the prolate peak is always larger than the oblate one , contrary to what happened with the gt strengths that were comparable .
the spherical shape in @xmath63se produces much less strength than the deformed shapes .
we can also see in fig . 5 for kr and sr isotopes that the profiles of the @xmath2 strengths corresponding to the prolate and oblate distributions can be clearly distinguished , similarly to the case of the gt distributions in fig .
the strength corresponding to the prolate shape is again the largest .
therefore , clear similarities between the gt and @xmath2 strength distributions are observed , but some differences can be seen as well . in particular , for spin @xmath2 strength distributions the position and strength of the resonance are practically the same in all nuclei in a given isotope chain .
this is different to what happened with the gt strength distributions , where the @xmath114-gt strength decreases very fast with increasing @xmath115 due to pauli blocking , while @xmath2 strengths are not affected by this .
we studied @xmath0-decay in various isotopic chains of medium - mass proton - rich nuclei within the framework of the selfconsistent deformed hf+bcs+qrpa with skyrme interactions .
our spin - isospin residual interaction contains a particle - hole part , which is derived selfconsistently from the skyrme force , and a particle - particle part , which is a separable force representing a neutron - proton pairing force .
compared to previous calculations , where the latter residual interaction was not considered , we obtain in general better agreement with experiment for half - lives .
it is worth mentioning that for the n = z isotopes this agreement is comparable to that achieved for the other isotopes with an excess of neutrons .
this indicates that using the phenomenological gap values for neutrons and protons , as well as the neutron - proton pairing correlations as a residual force in qrpa , is sufficient to account at least for this experimental information . from our study of the dependence on the shape of the gt strength distributions
we conclude that there are some interesting cases that are worth to explore experimentally such as @xmath63kr and @xmath94sr . in these examples
the measured @xmath0 strengths below @xmath12 could be used to identify their equilibrium shapes . in particular ,
we have observed in @xmath71sr various compatible indications pointing out towards the same conclusion : a strong prolate component in the ground state that is suggested from the comparison to experiment of half - lives and gt strength measured at low excitation energies .
nevertheless , more experimental information is still needed to reach a conclusive answer .
we also studied the energy distributions of the spin @xmath2 strength and the similarities and differences with their gt counterparts .
we conclude that the main features of gt and spin @xmath2 strength distributions are similar .
this suggests that we may learn about properties that are observable in highly unstable nuclei ( like @xmath4 decay ) from properties that are observable in stable nuclei ( like @xmath2 s ) , and vice versa .
information on spin @xmath2 excitations can be extracted from both leptonic and hadronic probes . in stable nuclei ,
the combined analysis of electron , photon , and proton scattering data has provided reliable information on orbital and spin @xmath2 strength distributions @xcite . in principle
, the same type of experiments could be performed on the proton rich nuclei considered here .
inelastic @xmath116 scattering experiments might be particularly suitable to explore the spin part of the @xmath2 strength and to test the predictions of the present paper .
c. dtraz and d.j .
vieira , annu .
nucl . part .
39 ( 1989 ) 407 ; a. mueller and b.m .
sherrill , annu .
nucl . part .
43 ( 1993 ) 529 ; nupecc report on _ nuclear structure under extreme conditions of isospin , mass , spin and temperature _ ,
mueller et al . (
1997 ) .
m. hirsch , a. staudt , k. muto and h.v .
klapdor - kleingrothaus , nucl .
a535 ( 1991 ) 62 ; k. muto , e. bender , and h.v .
klapdor , z. phys . a 333 ( 1989 ) 125 ; h. homma , e. bender , m. hirsch , k. muto , h.v .
klapdor - kleingrothaus and t. oda , phys .
c54 ( 1996 ) 2972 .
h. flocard , p. quentin and d. vautherin , phys .
b 46 ( 1973 ) 304 ; m. beiner , h. flocard , n. van giai , and p. quentin , nucl . phys . a 238 ( 1975 )
29 ; p. quentin and h. flocard , annu . rev .
nucl . part .
28 ( 1978 ) 253 ; p. bonche , h. flocard , p.h . heenen , s.j .
krieger , and m.s .
weiss , nucl .
phys . a 443 ( 1985 ) 39 .
* table 1 . * results for @xmath12 values , half - lives @xmath117 , and gt strength summed up to @xmath12 energies @xmath118 .
the results correspond to qrpa calculations performed with the skyrme force sg2 in four isotopic chains and are calculated for the various equilibrium shapes in each nucleus .
experimental values for @xmath12 and @xmath103 are from @xcite . 0.5 cm | a selfconsistent approach based on a deformed hf+bcs+qrpa method with density - dependent skyrme forces is used to describe @xmath0-decay properties in even - even deformed proton rich nuclei .
residual spin - isospin forces are included in the particle - hole and particle - particle channels .
the quasiparticle basis contains neutron - neutron and proton - proton pairing correlations in the bcs approach , while neutron - proton pairing interaction is treated as a residual force in qrpa .
we discuss the sensitivity of gamow - teller strength distributions and @xmath1 half - lives to deformation , pairing and the strength of the particle - particle interaction .
the dependence on deformation is also compared to that of spin @xmath2 strength distributions . |
recent high-@xmath2 rhic experimantal results on pion production in central @xmath1 collision at mid - rapidity at @xmath4 and @xmath5 agev have shown a strong suppression compared to binary scaled @xmath6 data @xcite .
analysing this effect in different impact parameter ranges , the suppression vanishes with increasing centrality . in @xmath3
the lack of the suppression @xcite was found also . a quantitative analysis of the suppression pattern , the so called _ `` jet - tomography '' _ @xcite , yields information about the physical aspects of the produced dense matter and the impact parameter dependence of the formation of deconfined matter in @xmath1 collisions .
latest data on @xmath3 collision confirmed that initial state effects can not be responsible for suppression in @xmath1 collision .
thus medium induced non - abelian jet - energy loss in the final state becomes a strong candidate to explain the missing pion yield .
jet energy loss can be calculated in a perturbative quantum chromodynamics ( pqcd ) frame @xcite .
our aim is to calculate @xmath0 production in nucleus - nucleus @xmath7 collision in a pqcd improved parton model , taking into account a phenomenological intrinsic transverse momentum distribution ( intrinsic-@xmath8 ) for the colliding partons , which is necessary to get better agreement between data and calculations in @xmath6 collisions @xcite . for @xmath7 collision we included initial state effects into our calculations : nuclear multiple scattering , saturation in the number of semihard collisions , and shadowing effect inside nucleus .
we used glauber - like collision geometry . here
we investigate jet energy loss . as high - energy quark and
gluon jets are travelling throught dense colored matter , they can lost their energy described by induced gluon radiation in thin non - abelian matter .
the glv - description @xcite of energy loss can be included into the pqcd improved parton model as a strong final state effect . comparing theory to data we can extract the opacity values ( @xmath9 ) of the produced colored matter at different centralities .
applying the above sketched model for reproducing @xmath0 yields in @xmath1 , and @xmath3 collisions we can accomplish a tomographic analysis of the produced dense and colored matter in reactions . in fig .
[ fig:1 ] we display the nuclear modification factor : @xmath10 as a function of @xmath2 for pion production from central to peripheral @xmath1 ( _ lower curves _ ) and in @xmath3 ( _ the upper curve _ ) . in the most central @xmath1 collision the measured suppression
is reproduced with opacity @xmath11 3.5 . as fig .
[ fig:1 ] displays , in @xmath1 collisions the suppression of the high - energy pions goes away with decreasing centrality , which is followed by the descreasing values of the opacity . in @xmath3 collision
we do not need to include jet quenching to reproduce pion production , this final state effect does not appear there , only multiscattering and shadowing .
g. david _ et al .
_ ( phenix coll . ) , nucl .
a698 ( 2002 ) 227 ; d. denterria _ et al . _
( phenix coll . ) , nucl .
a715 ( 2003 ) 749 ; s.s .
( phenix coll . ) , phys .
91 ( 2003 ) 072301 , _ ibid .
_ 072303 ; nucl - ex/0308006 .
j.c . dunlop _
( star coll . ) , nucl .
a698 ( 2002 ) 515 ; j.l .
( star coll . ) , nucl .
a715 ( 2003 ) 733 ; j. adams _ et al . _
( star coll . ) , phys .
91 ( 2003 ) 172302 , _ ibid . _ 072304 ; nucl - ex/0308023 .
e. wang , x.n .
wang , phys .
( 2002 ) 162301 ; m. gyulassy , p. lvai , i. vitev , phys .
b538 ( 2002 ) 282 .
m. gyulassy , p. lvai and i. vitev , phys .
85 ( 2000 ) 5535 ; nucl .
b571 ( 2000 ) 197 , _ ibid .
_ b594 ( 2001 ) 371 . g. papp et al .
hep - ph/0212249 ; y. zhang _ et al .
c65 ( 2002 ) 034903 .
et al . _ , aph ns heavy ion phys . 18 ( 2003 ) 79 , nucl - th/0206006 .
_ et al . _ , nucl - th/0306019 . | a strong suppression of @xmath0 production in @xmath1 collision was discovered at rhic energies in the high-@xmath2 region .
the lack of this suppression in @xmath3 collision proved , this effect can be connected to final state effects , namely medium induced jet energy loss caused by dense color matter . here
we display our quantitative results on jet quenching in @xmath1 collision using our pqcd - based parton model calculations . in parallel
we present results on @xmath3 collision . |
a qcd - based space - time model that allows to simulate nucleus - nucleus collisions ( among other particle collisions ) , is now available as a computer simulation program called vni @xcite .
it is a pqcd parton cascade description @xcite , supplemented by a phenomenological hadronization model @xcite , with dynamically changing proportions of partons and hadrons , in which the evolution of a nuclear collision is traced from the first instant of overlap , via qcd parton - cascade development at the early stage , parton conversion into pre - hadronic color - singlet clusters and hadron production through the decays of the clusters , as well as the fragmentation of the beam remnants at late times .
( i will in the following refer to the parton - cascade / cluster - hadronization model as pcm , and to its monte carlo implementation as vni . ) the pcm description has been used @xcite to provide very useful insight into the dynamics of the evolution of the matter at energies likely to be reached at bnl rhic and cern lhc .
the experimental data for these will , however , come only in the next millennium .
recently , d. srivastava and myself have taken steps to investigate collisions at the cern sps and compare the model calculations with the increasing body of experimental data , thereby addressing the the question : _ can one really use the pcm picture already at the sps energies _
( @xmath2 1720 gev / a ) ? recall that the parameters of the models were fixed by the experimental data for @xmath3 ( @xmath4 ) cross - sections over @xmath5 = 101800 gev , and @xmath6 annihilation @xcite .
the nucleon - nucleon energy reached at sps is well within this range . indeed , as shown in @xcite , this approach does remarkably well in comparison to the gross particle production properties observed at cern sps .
this came as a surprise , since no attempt was made to fine - tune the model to the data .
in figs . 1 and 2 , i have plotted examples of the results of @xcite , for the relatively ` light ' system @xmath7 and for the most ` heavy ' system @xmath8 .
the first plot shows the rapidity distribution of negative hadrons in central collisions of sulfur nuclei , whereas the second plot shows the transverse mass spectra of negative hadrons in central lead - lead colisions . in comparison with the corresponding experimental data ,
it is evident that a decent description is obtained , and this without any adjustments of parameters .
the analysis of @xcite concluded : \(i ) the simulations of heavy - ion collsions give a good overall description of the cern sps data on rapidity and transverse energy distributions , the multiplicity distributions , as well as a reasonable description of the transverse momentum distribution of hadrons .
this success may be taken as a ` post - hum ' justification of the applicability of a pqcd parton decription even at cern sps energies .
\(ii ) contrary to wide - spread belief , at cern sps energy the ` hard ' ( perturbative ) parton production and parton cascading is an important element for particle production at central rapidities - at least in @xmath8 collisions . in effect
it provides almost 50 @xmath9 of final particles around mid - rapidity .
` soft ' ( non - perturbative ) parton interactions on the other hand are found to be insignificant in their effect on final - state particle distibutions .
let me now turn to the problem of the observed strong suppression of @xmath0 production in pb + pb collisions , following a recent work @xcite by b. mller and myself .
in particular , i would like to discuss briefly the @xmath11-dependence of the absorption of the @xmath0 by comoving produced hadronic matter .
the key element in the following arguments is the fact that , for most nuclear collision systems at the relatively low center - of - mass energy of the cern - sps experiments , perturbative mini - jet production is largely due to quark - quark scattering , because the typical value of bjorken-@xmath12 probed is @xmath13 , where the gluon density is small .
this implies that shadowing effects are negligible as explained in @xcite and the critical momentum @xmath14 where these effects turn on , lies below the perturbatively accessible range of momenta @xmath15 gev at the sps energy , and may just begin to reach into it for the pb + pb system .
if this is correct , then _ all _ mini - jet production involving momenta @xmath15 gev lies safely above @xmath16 . as immediate consequence , the density of comovers which can effectively interact and absorb the @xmath0 at the sps grows like @xmath17 , or @xmath18 in asymmetric collisions .
this is a _ much stronger @xmath11-dependence _ than naively expected and embodied in most comover suppression models @xcite .
it also implies a _ stronger impact parameter or @xmath19-dependence _ of comover suppression than predicted by existing models .
quantitative support for these speculations comes from recent calculations of secondary particle production in the parton cascade model , usisng vni @xcite .
the calculations predict that the energy density @xmath20 produced by scattering partons ( mini - jets ) at central rapidity grows by a factor of more than 2 between @xmath21 and @xmath22 , compatible with the scaling law @xmath23 , but much faster than expected in the usual glauber model approach which predicts @xmath24 , in which case @xmath20 increases only by 30 @xmath9 . in @xmath22 collisions the initial partonic energy density @xmath20
is predicted to reach 5 gev/@xmath25 at times @xmath26 fm/@xmath27 in the comoving reference frame ( fig .
3 ) . for a thermalized gas of free gluons and three flavors of light
quarks this corresponds to an initial temperature @xmath28 mev , clearly above the critical temperature @xmath29 mev predicted by lattice - qcd calculations @xcite shown in fig .
4 . on the other hand , the parton density @xmath30 2 gev/@xmath25 predicted for @xmath21 collisions ( fig .
3 ) , just lies at the upper end of the transition region in the equation of state from lattice - qcd .
the temperature @xmath31 required for dissociation of the @xmath0 bound state due to color screening is known @xcite to be higher than @xmath32 , namely @xmath33 mev .
the parton cascade model results are therefore compatible with the experimental finding that there appears to exist no significant comover - induced suppression of @xmath0 in @xmath21 collisions , but a large and strongly impact parameter dependent effect is observed in pb + pb collisions .
it must be stressed that the model results do not contradict the experimental data collected at the sps , as one might suspect , since it is usually claimed that those are fully consistent with the glauber picture , in particular with the transverse energy production ( fig .
5 ) and the linear @xmath34 relation ( fig .
6 ) , both wich are in good agreement with the corresponding measured data @xcite . 9 k. geiger , comp .
comm . * 104 * , 70 ( 1997 ) , and references therein .
k. geiger and b. mller , nucl .
phys . b*369 * ( 1992 ) 600 . k. geiger , phys . rep . *
258 * , 376 ( 1995 ) .
j. ellis and k. geiger , phys .
d * 54 * , 949 ( 1996 ) ; phys .
d * 54 * , 1755 ( 1996 ) .
k. geiger and d. srivastava , phys . rev .
* c56 * , 2718 ( 1997 ) .
d. srivastava and k. geiger , @xmath35 nuc - th/9708025 @xmath36 .
k. geiger and b. mller , @xmath35 nuc - th/9707048 @xmath36 .
s. gavin and r. vogt , _ nucl . phys . * a610 * _ , 442c ( 1996 ) . c. bernard et al . , _ phys
. rev . * d54 * _ , 4585 ( 1996 ) . f. karsch and h. satz , _ z. phys . * c51 * _ , 209 ( 1991 ) .
l. ramello , these proceedings . | there seems to be a general consensus now that a first glimpse of a qgp - like effect has become visible in the beautiful na50 data on @xmath0 production and the ` anomalous supression ' phenomenon . on the other hand ,
it is still widely believed that the dynamics of heavy - ion collisions at cern sps energy is predominantly governed by soft , non - perturbative physics .
this is ironic : after all , it is unlikely that a qgp could be formed if the underlying dynamics were essentyially soft , rather than that it requires intense quark - gluon production with sufficient cascade - like reinteractions which drive the matter to large density and equilibrium .
therefore , i advocate in this contribution that for ultra - relativistic nucleus - nucleus collisions a description based on the pqcd interactions and cascade evolution of involved partons can and should be used , owing to the claim that short - range parton interactions play an important role at sufficiently high beam energies , including cern energy @xmath1 a gev . here
mini - jet production which liberates of quarks and gluons can not be considered as an isolated rare phenomenon , but can occur quite copiously and may lead to complex multiple cascade - type processes . |
a @xmath18 bound in a nucleus decays via either mesonic weak decay , @xmath18 @xmath19 , or nonmesonic weak decay ( nmwd ) , @xmath20 .
mesonic decay is essentially similar to free @xmath18 decay and has been studied in detail . in a nucleus
a @xmath18 can decay via an interaction with a neighbor nucleon , either a proton ( @xmath21 , proton - induced nmwd ) or a neutron ( @xmath22 , neutron - induced nmwd ) ; this is referred to as a one - nucleon - induced ( 1n ) nmwd process . the two - nucleon induced ( 2n ) process ( @xmath18@xmath23 ; @xmath24 ) , which is another nmwd mode , has been predicted theoretically ; however , it has not yet been verified experimentally .
the nmwd process has attracted considerable attention since it provides the only practical means of studying the strangeness - changing baryonic weak interaction at present @xcite . since the discovery of hypernuclei , one of the primary concerns in the study of nmwd of @xmath18 hypernuclei has been the relative strength of the two channels of 1n nmwd - the ratio of the decay widths ( / ) .
experimental ratios over the broad mass range of @xmath18 hypernuclei are consistently greater than unity , thereby indicating the dominance of the neutron channel ; on the other hand , the theoretical ratios are only one tenth of unity .
the predominance of the proton channel in the theoretical prediction is due to the fact that the tensor term of one - pion exchange ( ope ) contributing only to the proton channel is very high , and therefore , the contributions of other meson exchange terms become minor corrections .
the ope contribution is included in all model calculations for the long range contribution ; hence , the @xmath7 ratios have remained at around 0.1 .
this intriguing problem is referred to as the @xmath7 puzzle . in order to reduce the predominance of the proton - induced nmwd of the ope model ,
many theoretical models for short range contribution , such as the heavy meson exchange ( hme ) models and the direct quark ( dq ) model have been studied .
all these models have been unsuccessful in increasing the @xmath7 ratio significantly until recently .
however , in recent important development , the incorrect sign in kaon exchange amplitudes was identified ; the correction significantly increased the @xmath7 values of these models @xcite .
thus , the current theoretical values of the @xmath8/@xmath10 ratio have increased to 0.34 - 0.70 for @xmath25he and 0.29 - 0.53 for @xmath0c @xcite . until recently experimental @xmath7 values were extracted by fitting intranuclear cascade ( inc ) calculation yields to the experimental proton yields
almost all the experimental @xmath7 ratios have been close to or greater than unity @xcite .
however , recent high - quality neutron yield spectra obtained by our group have made it possible to derive the @xmath7 ratio directly by comparing the neutron yields to the corresponding proton yields ; the ratios obtained by this method are @xmath260.5 for @xmath0c @xcite and @xmath260.6 for @xmath15he @xcite agreeing well with the recent theoretical values .
these ratios are significantly smaller than unity .
they are the first experimental indications of the proton channel dominance ( @xmath27 ) in the nmwd .
it appears that the discrepancy between the experimental and theoretical values of the ratios of nmwd has been removed .
however , the @xmath7 ratios of 5lam and derived by comparing the nucleon yields still contain the ambiguity due to the possible contributions of 2n nmwd and the final state interaction ( fsi ) effects on the emitted nucleons . the current fsi model calculation which takes into account the 2n nmwd and fsi can not reproduce the yield spectra @xcite .
therefore , the uncertainty of the ratios due to the ambiguity could not be estimated in such single nucleon ( singles ) measurements . in order to remove such ambiguity experimentally
, the decay channel has to be explicitly identified for each event . in order to determine each decay channel of 1n - induced nmwd exclusively
, we have performed a coincidence measurement of both n+p ( @xmath28 pair ) and n+n ( @xmath29 pair ) pair nucleons from the nmwd of the hypernuclei , 5lam and .
since @xmath18n @xmath30 nn is a two - body process , the outgoing nucleon pairs will exhibit a clear back - to - back ( @xmath31 ) correlation with regards to their opening angle , and their energy sum distribution would exhibit a peak at around the decay q value , @xmath26155 mev , although they are broadened due to nuclear medium effects . on the other hand , in 2n nmwd ,
three nucleons are produced in the final state and a pair of nucleons among them would exhibit neither the @xmath32 feature in the opening angle nor the energy sum peak at the q value .
the nucleon from 1n nmwd that suffered intense fsi on the way out of the residual nucleus would be deflected from the original momentum direction and lose kinetic energy , thereby degrading the kinematics characteristic of 1n nmwd .
therefore , by applying the event selection criteria of kinematics , we can exclusively select 1n nmwd events suppressing possible contributions from the 2n process and the events that suffered serious fsi .
recently we have reported the @xmath7 value of @xmath16-shell @xmath25he that was determined in the two - nucleon coincidence measurement of the nmwd @xcite . by applying the kinematical condition for the opening angle
, we obtained the pair number ratio of @xmath29 to @xmath28 pairs of 1n nmwd , @xmath33 , to be [email protected]@xmath340.03 .
this ratio becomes essentially the ratio when fsi is weak as in 5lam .
the @xmath16-shell hypernucleus , @xmath35he , was selected for the first exclusive coincidence measurement , since the fsi effect for such a light hypernucleus must be small . however , the unambiguous determination of for a @xmath14-shell hypernucleus also is essential in the study of baryonic weak interaction since only the @xmath14-shell hypernuclei can provide the @xmath14-wave initial state for the interaction while the @xmath16-shell 5lam does only that of @xmath16-wave .
the importance of the @xmath14-wave contribution in the nmwd in the finite nucleus has been discussed in the reference @xcite . in such respect , has been extensively studied together with 5lam .
exclusive determination is even more important for the @xmath7 ratio of heavier @xmath14-shell hypernuclei such as since the nucleons propagating there will suffer more intense fsi .
although @xmath0c is the most extensively studied @xmath14-shell @xmath18 hypernucleus , no unambiguous @xmath7 ratio for it is available yet .
the values of the asymmetry parameter @xmath36 of protons from @xmath37 nmwd of the polarized hypernuclei were measured and reported for both @xmath15he and formed via the ( @xmath1,@xmath2 ) reaction . although small positive values of [email protected] and [email protected]@xmath340.04 @xcite were observed for @xmath15he , a large negative value ( @xmath38 )
was reported for the @xmath14-shell hypernuclei @xcite .
authors indicated a possible explanation that the interaction mechanism of nmwd for short and long range could be different and therefore the initial @xmath14-state contribution in made the difference .
the theoretical predictions for the asymmetry parameter for 5lam are about -0.7 @xcite from both the recent heavy meson exchange ( hme ) and the direct quark ( dq ) models .
they significantly overestimate the magnitude of the parameter .
the recent hme model prediction for @xmath36 of is 0.36 that is again far off from the experimental value .
the current difficulty in understanding the decay asymmetries of @xmath16- and @xmath14-shell hypernuclei implies that an accurate and unambiguous determination of @xmath7 ratios for both @xmath16- and @xmath14-shell hypernuclei is essential for nmwd studies .
in this paper we report the results of the two - nucleon observables measured in the coincidence experiment of the decay of @xmath0c especially focusing on its @xmath7 ratio .
captypefigure the experiment was carried out at the k6 beam line of 12-gev proton synchrotron ( ps ) in high energy accelerator research organization ( kek ) .
@xmath18 hypernuclei @xmath0c were produced via the ( @xmath1,@xmath2 ) reaction at the beam momentum of 1.05 gev / c on the @xmath39c target . in order to improve the vertex resolution , we used thin segmented plastic scintillators for an active @xmath39c target .
[ fig : e508setup ] shows the schematic view of coincidence counters for detecting the decay product particles from the weak decay of @xmath0c .
it comprises three sets of coincidence counter - two located at the top and bottom of the target , optimized for @xmath32 coincidence events and one at the side to allow the detection of non - back - to - back events .
each of top and bottom counter sets comprise a fast timing counter ( t2 ) , a drift chamber ( pdc ) , veto or stop timing counter ( t3 ) and neutron counter arrays ( t4 ) .
the side set is similar ; however , the pdc is absent .
the inclusive @xmath0c excitation energy spectrum derived from the momenta of incoming pion and the outgoing kaon is shown in fig .
2(a ) . the two single particle @xmath18 states ( 1@xmath16 and 1@xmath14 )
are clearly identified .
the 1@xmath14 @xmath18 state is located just above proton emission threshold ; it decays into @xmath40b by emitting a proton .
2(b ) and 2(c ) show the excitation energy spectra in coincidence with the emitted pair nucleons @xmath28 and @xmath29 , respectively .
the vertical lines show the applied gates of the decay measurement for ground state of @xmath0c .
the enhancement of the yield in the quasi - free @xmath18 region of the @xmath29 coincidence excitation energy spectrum as compared with that of the @xmath28 pairs occurs due to the emission of two neutron via the absorption of @xmath41 from the quasi - free @xmath18 mesonic decay .
captypefigure neutral particles , @xmath42 and neutrons , were measured using six layers of 5 cm thick plastic scintillators ( t4 ) vetoed with t3 . the particle identification ( pid ) of the neutrons and @xmath42
was achieved by using the flight time from the decay vertex to the fired t4 counter .
the neutron and @xmath42 were well separated , and the accidental background within the neutron gate was negligible .
the charged particles were identified by combining @xmath43 , @xmath44 and @xmath45 .
@xmath43 denotes the energy loss per unit length measured by @xmath46 ; @xmath44 , flight time between @xmath46 and @xmath47 ; and @xmath45 , the total energy deposited in the sequentially fired counters , @xmath46 , @xmath47 , and @xmath48 .
protons are well separated from pions and deutrons .
the kinetic energies of neutrons and protons were calculated from the flight times and the ranges , respectively .
proton energy was corrected for the energy loss inside the target material thanks to the good vertex resolution .
further experimental details are provided in the previous report @xcite .
[ fig : raw](a ) and [ fig : raw](b ) show the raw distribution of @xmath28 and @xmath29 pair yields @xmath49(cos@xmath50 ) and @xmath51(cos@xmath50 ) in the opening angle @xmath50 , respectively .
only the pair events of which each nucleon has an energy greater than 30 mev are counted .
the dominance of the peaks shown in the raw spectra is enhanced by the maximized acceptance for the coincidence back - to - back kinematic events .
a total of 116 and 43 events were observed in the @xmath32 angular region of cos@xmath52 for the @xmath28 and @xmath29 pair yields , respectively . the shoulder bump observed in the region of the pair
can be understood as a combined result of a statistical fluctuation due to the limited statistics of pair events ( total 43 counts only in the region ) and the binning effect of neutron z - position analysis ( refer to fig .
[ fig : e508setup ] ) .
it was revealed that the background nucleon pairs produced through the absorption of @xmath53 from the mesonic decay of @xmath0c in the materials around the target were insignificant unlike the situation in the case of @xmath25he @xcite .
the angular resolutions for the and pairs were estimated to be @xmath54 and @xmath55 at cos@xmath56 , respectively .
[ sec : discussion ] captypefigure fig .
[ fig : raw](c ) and [ fig : raw](d ) show the angular correlation of pair nucleons , namely the normalized pair yields per nmwd , @xmath57(cos@xmath50 ) .
@xmath57(cos@xmath50 ) can be expressed as @xmath58(cos@xmath50)@xmath59(cos@xmath50)@xmath60 , where @xmath61 and @xmath62 are the total number of nmwd observed and the overall efficiency for detecting two nucleons from the and pairs in coincidence including the detector acceptance .
@xmath61 is obtained by multiplying the number of hypernuclei produced in the ground state of @xmath63 with the nmwd branching ratio @xmath64 , which is defined as @xmath65 , where @xmath66 and @xmath67 denote the @xmath41 and @xmath68 branching ratios , respectively .
the accuracy of the @xmath64 value of [email protected]@xmath340.005 has been improved drastically due to the accurate measurement of @xmath67 in the present experiment @xcite in comparison with the previous one @xcite .
this accurate @xmath64 value made it possible to normalize the pair yields per nmwd without introducing significant systematic errors .
the measured @xmath57 angular correlation of @xmath29 and @xmath28 pairs could now be directly compared with those obtained from the fsi model calculation @xcite .
@xmath69(cos@xmath50 ) was calculated with energy dependent event - by - event simulations , and it was tested to be in good agreement with the data @xcite .
the neutron detection efficiency was calculated by the monte - carlo simulation code - the modified demons code - which is applicable to a multi - element neutron detector and has been tested for producing various experimental data well @xcite .
the angular correlation of @xmath28 pair clearly exhibits a peak at cos@xmath50 @xmath70 , thereby indicating the @xmath32 emission of two nucleons , which is the signature of two - body decay final state .
the kinematic regions defined as cos@xmath52 essentially include all the events in the kinematics peak .
however , although dominance of kinematics events is degraded to a certain extent in the pairs , the events continue to be a major contribution .
the dominance of the kinematic events observed in both and pair confirms that 1n nmwd , @xmath71 and @xmath72 , is the main nmwd mode of . in non - region , we observe pair events more or less uniformly distributed over all angle in cos@xmath50 whose total number reaches @xmath2640 % of the total pair numbers as shown in table 1 .
table [ tab : pair_number ] shows the pair numbers in each opening angle region .
we consider the uniform distribution event extended even in region .
therefore , we will subtract the corresponding component of the uniform distribution events in region from the pair numbers @xmath57 for the later discussion of 1n nmwd .
.the yields of coincidence events @xmath49 , @xmath51 and @xmath73 and the normalized pair numbers per nmwd @xmath6 , @xmath5 and @xmath74 are shown in the opening angle regions @xmath32 ( back - to - back , cos@xmath75 ) and @xmath76 ( cos@xmath77 ) .
@xmath57 values listed are the numbers simply integrated over the corresponding angular regions .
the pair numbers after the subtraction of the uniform background in the @xmath32 region are also listed .
errors are statistical .
[ cols="^,^,^,^,^,^,^",options="header " , ] in fig .
[ fig : raw](e ) and [ fig : raw](f ) , the spectra of the @xmath28 and @xmath29 pair yields in the nmwd of @xmath0c in the @xmath32 region are shown in the energy sum of the two emitted nucleons .
the rms energy - sum resolutions , @xmath78 , for @xmath29 and @xmath28 pairs were estimated to be 11 and 9 mev for typical cases , such as 75 mev nucleons , respectively .
the energy - sum spectra show that the pair yields are distributed over a broad energy region from 70@xmath2680 mev to the q value at around 155 mev ; on the other hand , a few pair yields exist above the q values . those present above the q values are considered to be due to the rapid deterioration in the neutron - energy resolution with an increase in the neutron energy .
since the energy sum spectrum of the two nucleons emitted in the two - body decay process , @xmath79 , would exhibit a peak at around the q value , @xmath26155 mev , the broad energy distribution indicates that a large number of the kinematics nucleons suffered an energy loss on the way out of the residual nucleus with a mass of @xmath80 .
on the other hand , a dominant narrow peak around q value was clearly visible in the pair energy sum spectrum of @xmath15he @xcite .
therefore even if we consider the pair numbers confined in the @xmath32 region only , they include a considerable number of events that suffered fsi .
we observe a similar or even more energy degraded energy - sum spectrum in @xmath29 pairs ; however , the statistics are considerably lower .
captypefigure fig .
[ fig : compare ] compares the angular correlations @xmath6(cos@xmath50 ) and @xmath5(cos@xmath50 ) of 5lam(open circle ) @xcite and ( filled circle ) .
we observe dominant @xmath32 peaks in both @xmath25he and @xmath0c ; this indicates that the 1n nmwd is the major mechanism of nmwd . by comparing the angular correlations of @xmath25he and @xmath0c ,
we observe two clear features - broadening of the @xmath32 peak in angular correlations and quenching of the pair numbers of @xmath0c over those of @xmath25he in both @xmath28 and @xmath29 pairs .
the total @xmath32 pair numbers of @xmath28 and @xmath29 in @xmath0c are reduced in comparison with those of 5lam by @xmath2640% and @xmath2620% , respectively .
the features can be understood in terms of the same physics responsible for the energy degradation in the @xmath81(@xmath0c ) spectra ( fig .
[ fig : raw](e ) ) in comparison with those of @xmath25he . by considering that 2n theoretical branching ratio does not vary considerably with the mass number , and therefore
, the relative importance of 2n with respect to 1n would be more or less the same for @xmath25he and @xmath0c @xcite , we infer that fsi is the major cause of broadening and quenching in the angular correlation and the degradation in the energy - sum spectrum .
this implies that the effect of fsi on the pair numbers , even in the @xmath32 region , would not be negligible , and we must take it into account in order to extract @xmath82 ratio from the pair number ratio . a significant portion ( @xmath2640% ) of the pair number
is observed in the non - back - to - back region as shown in table [ tab : pair_number ] .
the possible contributions in the non - back - to - back region are such as 2n(or multinucleon)-induced nmwd events , those that suffer strong fsi , @xmath41 absorption contamination and random coincidence events etc .
however , in @xmath0c , the contributions from the last two cases are found to be almost negligible , and the first two are considered to be the major contributions in non - back - to - back region .
these non - events , whose strength reaches @xmath2640 percent of nmwd , are the sources of the ambiguity in the previous results of since they can not be distinguished from 1n nmwd in the singles measurement .
however , we can reject them by considering only the coincidence pair numbers in the region in the derivation of , thereby resulting the ratio which is almost free from the ambiguity inherent in the previous results .
next , we would like to derive the @xmath82 ratio from the pair number ratio , @xmath33 , of the @xmath32 region .
if neither @xmath24 contribution nor fsi existed , the ratio of @xmath33 would become @xmath82 .
however , as discussed , the effects of fsi still remain in the pair numbers of @xmath0c .
the most harmful effect of fsi is the channel crossover such as that from to or from to @xmath29 .
it directly contaminates the pair number ratio .
we observed 8 @xmath83 pair coincidence events in the @xmath32 region , which are possible only via the fsi process .
a neutron in the @xmath28 pair emitted at the vertex is converted into a proton via the nuclear interaction on the way out of the nucleus , thereby registering as a @xmath83 coincidence event .
therefore , they represent the channel crossover information of fsi on the nucleons emitted .
table [ tab : pair_number ] shows @xmath74 yields - the normalized @xmath83 pair numbers per nmwd .
we will utilize @xmath74 for the correction of the fsi crossover effect on @xmath7 ratio .
the pair numbers in the @xmath32 region @xmath5(@xmath32 ) , @xmath6(@xmath32 ) and @xmath74(@xmath32 ) can be expressed as @xmath84 where @xmath85 and @xmath86 are the fractions of neutron and proton induced channels out of the nmwd .
@xmath87 is the loss factor of neutrons ( protons ) from the energy region of interest due to fsi and @xmath88 is the resulting crossover influx factor of neutrons ( protons ) from protons ( neutrons ) channel .
the first terms of @xmath5 and @xmath6 represent the pair numbers emitted out of the nucleus that maintain the channel starting at the initial weak vertex point .
the second terms represent the pair numbers that cross the channel and end in a channel different from the starting one .
@xmath89 and @xmath90 are the loss factors from the angular region of the original and channel crossing pairs ; however , their actual values are not required in the following derivation .
instead , they are replaced with the observed @xmath83 pair numbers .
if there were no fsi effects , @xmath87 and @xmath88 would be one and zero , respectively .
we put @xmath91 and @xmath92 by considering the charge symmetry of a strong interaction .
a similar shape of proton and neutron spectrum obtained from the present experiment @xcite supports this approximation .
a simple simulation reveals that the @xmath93 term in eq .
[ eq : pairs_bb ] is less than 1 percent of the noncrossing terms and can therefore be neglected .
then , the second terms of @xmath94 and @xmath95 of eq . [ eq : pairs_bb ] are replaced by @xmath96 and the @xmath97(@xmath98 ) ratio can be expressed in terms of pair numbers only as follows : @xmath99 the @xmath100 numbers are the integrated pair numbers in the region after the subtraction of the background , which is assumed to be uniformly distributed in all cos@xmath50 as we pointed out in the discussion of fig .
3(c ) and 3(d ) .
the value of the constant background was determined by @xmath101 fitting .
the ratio of the raw pair numbers simply integrated over the region [email protected] becomes [email protected] after the subtraction of the uniform background .
the obtained ratio @xmath102 in the eq .
[ eq : gngp ] is only 4 percent reduced from the pair number ratio @xmath103 .
this is surprising considering that we observed the strong fsi effect both in the energy - sum spectra and in the angular correlations .
however , the small correction can be understood considering that @xmath104 and hence the crossover pair numbers from to and from to are almost balanced , and therefore , the crossover - influx and -outflux due to fsi cancel each other .
the systematic errors mainly originate from the uncertainties of the neutron detection efficiency ( 6% ) and the @xmath83 pair acceptance ( 4% ) .
many sources of uncertainty such as @xmath105 and detection efficiencies are canceled out by taking the ratio @xcite .
the present result determined with the pair nucleon number ratio @xmath33 is in good agreement with the recent theoretical results based on heavy meson exchange ( hme ) model calculations @xcite .
it finally resolved the longstanding ratio puzzle without ambiguity by removing the effect of fsi self - consistently and by rejecting non - events applying kinematic condition .
it is quite accidental that the present result is close to that of the previous singles measurement which were determined only by assuming the nonexistence of 2n ( or multi - nucleon induced ) nmwd @xcite .
however , we have pointed out that there exist a significant amount ( @xmath106% of the total ) of non - events , which probably are the candidates of them
. the present @xmath82 ratio of @xmath0c is in good agreement with that of the previous exclusive measurement of @xmath25he ( @xmath107 ) , which indicates that the nmwd mechanism of @xmath16-shell ( @xmath25he ) and @xmath14-shell ( @xmath0c ) hypernuclei are rather similar .
though the ratio determined from the ratio of pair numbers @xmath33 agrees well with those of the theoretical predictions @xcite , there remain discrepancies between the pair numbers @xmath57 of our measurement and those of the theoretical calculation @xcite : the pair numbers for events and non - events , and the significant quenching of pair numbers of compared with those of 5lam .
for example , the observed pair number of events is about half of the prediction while that of non - events is only a fraction .
we have measured , for the first time , the angular correlation @xmath57(cos@xmath50 ) of the pair nucleons and from the nmwd of in coincidence measurement .
we have exclusively identified the dominant 1n nmwd process , @xmath108 and @xmath4 , by requiring the angular correlation of the two emitted pair nucleons , which is the characteristic of two - body decay .
then we have obtained the @xmath7 ratio from the ratio of pair numbers @xmath33 of 1n nmwd in which most of the non - events are excluded and fsi effects have been corrected using the simultaneously measured @xmath83 pair events .
therefore , the present @xmath7 result is almost free from the ambiguities due to 2n(or multinucleon induced ) nmwd contribution and the fsi effects which were inherent in the previous results obtained with singles spectra .
the obtained @xmath82 ratio of @xmath102 is in good agreement with the recent theoretical results calculated based on hme models .
this has finally resolved the long standing @xmath82 ratio puzzle unambiguously .
this ratio is very close to that of the previous exclusive measurement of @xmath25he , thereby indicating the decay mechanisms of @xmath16-shell ( @xmath25he ) and @xmath14-shell ( @xmath0c ) hypernuclei are rather similar .
moreover , we have obtained the pair number yields in the non - back - to - back kinematics region , which would provide information on the possible 2n nmwd contribution .
we are grateful to prof . k. nakamura and kek - ps staff for their support during our experiment and for ensuring the stable operation of kek - ps .
the authors , mjk and hb , acknowledge the support provided by kosef(r01 - 2005 - 000 - 10050 - 0 ) and krf(2003 - 070-c00015 ) . | we have measured the angular correlation of the pair nucleons and emitted from the nonmesonic weak decay ( nmwd ) of @xmath0c produced via the ( @xmath1,@xmath2 ) reaction in coincidence measurement .
the @xmath3 and @xmath4 modes were clearly identified by measuring the back - to - back correlation of the emitted nucleon pairs which is the characteristic of two - body kinematics . from the measured nucleon pair numbers @xmath5 and @xmath6 , the ratio @xmath7 of the partial decay widths @xmath8(@xmath9 ) and @xmath10(@xmath11 ) of @xmath0c
was extracted to be @xmath12(stat)@xmath13(syst ) ; this result is almost free from the ambiguity due to the nuclear final state interaction and 3-body decay process , which were inherent in the previous results .
the obtained @xmath7 ratio of @xmath0c ( @xmath14-shell ) is close to that of @xmath15he ( @xmath16-shell ) .
the results are consistent with those of recent theoretical calculations . , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , nonmesonic weak decay , @xmath0c , @xmath17 21.80.+a , 13.30.eg , 13.75.ev |
galaxy clusters promise great statistical power as a low redshift cosmological probe , if their masses are estimated accurately @xcite .
a bias in the mass estimation poses the most serious challenge to the success of cluster cosmology .
for example , the recently reported tension in cosmological inferences from the cosmic microwave background ( cmb ) and galaxy clusters measured by the planck satellite , the so - called ` planck cmb - cluster tension ' @xcite , may point to new physics , but it may also reflect the urgent need for better understanding and correction for systematic biases ( see e.g. * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the most prominent astrophysical bias has to do with the assumption of hydrostatic equilibrium ( hse ) of the intracluster medium ( icm ) inside the gravitational potential .
hse no longer holds in the presence of non - thermal sources of pressure .
hydrodynamical simulations show significant pressure associated with bulk and turbulent motions of the gas , which leads to a ` hse mass bias ' of order ten percent or more in the mass estimates of individual clusters ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
as the amplitude of non - thermal pressure can hardly be measured directly by observations , one has to resort to theoretical estimations to correct for this bias .
to this end , we have developed a physically motivated analytical model for non - thermal pressure in the icm ( * ? ? ?
* sk14 hereafter ) . given the accretion history of the cluster , the model predicts the amplitude of non - thermal pressure and its radial , mass , and redshift dependencies .
the model successfully reproduces the average non - thermal pressure profiles seen in cosmological hydrodynamical simulations @xcite , the average thermal pressure profile measured by planck @xcite , as well as non - thermal pressure profiles of _ individual _ clusters in hydrodynamical simulations ( * ? ? ?
* sknn15 hereafter ) . in this paper
, we study the ability of the sk14 model in correcting the hse mass estimation by testing it on simulated clusters in a set of high - resolution cosmological hydrodynamical simulations ( @xcite ) , with special attention paid to the dynamical state of galaxy clusters .
the dynamical state of a cluster influences its mass estimation in two ways .
first , the more dynamically relaxed a cluster is , the smaller its non - thermal pressure ( e.g. sk14 ; @xcite ) , and thus the smaller its hse mass bias .
second , the cluster profiles are smoother when a cluster is dynamically more relaxed , which gives a better precision in the mass estimation . in the early days , hse mass estimates were applied to clusters irrespective of their dynamical state ( e.g. * ? ? ?
these studies use x - ray surface brightness data and make the assumption that the icm is isothermal . after the recognition that gas temperature decreases systematically with increasing radius outside the cluster core region ,
precise hse mass estimates require also spatially - resolved , high signal - to - noise gas temperature measurements , which are harder to obtain especially out to high redshifts .
the current cluster mass estimates for a large sample of clusters are mostly performed statistically ( rather than individually ) using scaling relations between the mass and spatially - averaged observables , such as the mean x - ray temperature , luminosity , or their combination , and the sunyaev - zeldovich ( sz ) effect @xcite averaged over certain cluster radii ( e.g. @xcite , see also @xcite for a review ) . in this context , hse mass estimates are usually performed only on a small sample of dynamically relaxed clusters where the bias from non - thermal pressure is minimized , and are used for observationally calibrating the scaling relations with the hope that the adopted scaling relation is robust against varying dynamical states ( which holds true for some scaling relations , see @xcite ) .
however , even for relaxed clusters , non - thermal pressure is expected to be non - negligible because non - thermal motions are continuously sourced by the mass accretion and are never fully thermalized , especially at the cluster outskirts ( sk14 ; @xcite ) .
therefore , it is important to study the correction of hse mass bias induced by non - thermal pressure for dynamically relaxed clusters .
this is one of the primary goals of this paper . in recent years ,
the advances in sz observations make it possible to estimate hse masses using x - ray surface brightness combined with the gas pressure profiles probed by the sz effect , without the expensive x - ray temperature measurements .
this will make individual cluster mass estimates for a large number of clusters possible out to high redshifts , and this time with high precision to large radii thanks to the relative insensitivity of the sz effect to gas density . in light of these observational advances ,
another goal of this paper is to study the possibility of correcting hse mass bias for a population of clusters in a wide range of redshifts and dynamical states .
the rest of the paper is organized as follows : we introduce our method of correcting non - thermal pressure in cluster mass estimation in sect.[sec : mass_method ] , where we discover the necessity of regulating the input cluster profiles by fitting or smoothing . then we present four methods for regularizing the profiles in sect.[sec : smoothing ] and [ sec : hsemethods ] and apply them to our simulated clusters . after introducing the way we select subsample of clusters based on their dynamical state in sect.[sec : dynstate ] , we present our results in sect.[sec : results ] and conclude in sect.[sec : conclusion ] . in the two appendices we study the effect of velocity anisotropy ( appendix[sec : anisotropy ] ) and how an error on the pressure profile propagates to an error on the mass ( appendix[app : massbias ] ) .
the assumption of hse relates the thermal pressure @xmath0 that is observable from both x - ray and sz observations to the hydrostatic mass @xmath1 as @xmath2 where @xmath3 is the gas density .
since the total pressure in the icm , @xmath4 where @xmath5 is non - thermal pressure support , is greater than @xmath0 by definition , @xmath1 is an underestimate of the true mass .
we thus define the ` corrected mass ' @xmath6 as , random motions producing @xmath5 need not be isotropic . while we have implicitly assumed the random motions to be isotropic here , we investigate the consequence of velocity anisotropy in appendix[sec : anisotropy ] . ]
@xmath7 this @xmath6 is expected to be close to an unbiased estimate of the total mass when the non - thermal pressure support is the dominant source of the hse mass bias , as is the case found in hydrodynamical simulations at least for regions within @xmath8 ( e.g. * ? ? ?
* ; * ? ? ?
solving for @xmath13 using equation ( [ eq : mhse ] ) or ( [ eq : mcorr ] ) is equivalent to finding the corresponding radius @xmath9 from a spherically - averaged ` overdensity ' profile defined by @xmath14 with the implicit relation @xmath15 . here , @xmath16 is either @xmath0 or @xmath17 .
an error on the pressure gradient due to neglecting @xmath5 will lead to an error on @xmath13 through an incorrect estimation of @xmath9 . ) with respect to the critical density , @xmath18 , of a simulated cluster .
the horizontal dotted lines show @xmath19 and 200 . the true overdensity profile ( black solid line )
is calculated from the mass profile in the simulation , whereas the estimated profiles are calculated from spherically - averaged gas density and pressure profiles in the simulation using equations ( [ eq : mhse ] ) and ( [ eq : mcorr ] ) ( red and blue lines , respectively ) . structures and noises in the gas density and pressure profiles are greatly amplified by the derivative and division procedures required in computing the overdensity profile ( dashed lines ) , suggesting the need for fitting ( solid lines ) or smoothing.,scaledwidth=48.0% ] we apply the hse and corrected mass estimates to a mass - limited sample of 65 galaxy clusters with @xmath20 at redshift @xmath21 and their most massive progenitors at @xmath22 in the omega500 simulation at @xmath21 , with @xmath23 being the mass enclosed in a radius within which the average density is 200 times the mean matter density of the universe . ]
@xcite , an eulerian hydrodynamic cosmological simulation .
radiative cooling , star formation and feedback are not included in the current runs of the omega500 simulation .
these additional physics are crucial for reproducing the observed behaviors of cluster cores .
they also affect the detailed shapes of the thermo - dynamical profiles in the bulk of icm , especially for low mass systems . for low mass clusters and galaxy groups , testing non - thermal
pressure and cluster mass estimations would be better performed with simulations with cooling and feedback .
however , these additional physics do not change the overall shape of the icm profiles outside the cluster core ( e.g. * ? ? ?
in particular , the velocity dispersion of the gas and fraction of non - thermal pressure at around @xmath24 are consistent to typically @xmath2510% between simulations with and without cooling and feedback @xcite even for clusters with relatively low masses @xmath26(corresponding to @xmath27 ) @xcite .
therefore , non - radiative simulations are adequate for our purpose .
we refer to sect.2 of sknn15 and references therein for further descriptions of the simulation and the cluster sample used in this paper .
small clumps have been removed when constructing the one - dimensional cluster profiles by excluding high - density tail in the probability distribution of gas densities using the @xcite method .
we first compute the overdensity profile @xmath18 from simulated profiles with equation ( [ eq : deltaprof ] ) , taking @xmath28 for @xmath1 and @xmath29 for @xmath6 .
we then use equation ( [ eq : mdelta ] ) to compute @xmath30 .
here , @xmath31 is the ` non - thermal fraction ' defined by @xmath32 .
we compute @xmath33 of each cluster using the sk14 model and the mass accretion history of the cluster measured from the simulation ( see sknn15 for details ) .
numerically calculating a cluster mass using equation ( [ eq : deltaprof ] ) is not trivial , as taking derivative of pressure and dividing by density would amplify small - scale structures and noise . if we apply no regulation of the simulated profiles other than a cubic - spline interpolation when computing the derivative of the pressure profile , the resulting overdensity profile will be far from smooth ( dashed lines in fig.[fig : r500c_est ] ) , despite that the gas density and pressure profiles are rather smooth after clumps are removed . in the previous x - ray studies ,
the observed gas density and temperature profiles are often fit to parametric models to reduce this noise ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
non - parametric smoothing methods have also been adopted in simulation studies @xcite .
here we use and compare four approaches : one based on cubic spline smoothing , and the others based on fitting pressure and density to different parametrized models . for the parametrized models , we further compare two fitting ranges .
first , we compute the overdensity profile ( equation [ eq : deltaprof ] ) from simulated gas density and pressure profiles with cubic spline interpolation applied to pressure when taking a derivative , @xmath34 .
we then smooth the computed @xmath35 profile with a cubic - spline smoothing algorithm described in @xcite .
the algorithm fits the data points with a piece - wise cubic spline with an increasing number of nodes , until the input constraint on the value of @xmath36 statistic is satisfied .
we use the constraint that @xmath36 divided by the number of degrees of freedom is less than 0.04 .
this constraint is chosen to minimize the variance of @xmath37 while not biasing the sample mean . when estimating @xmath30 with the smoothing method , a large radial range of data around the true @xmath24 is needed to guarantee that the estimated @xmath38 profile covers @xmath19 .
therefore , we apply this method only to rather relaxed clusters using data on the radial range of [ 0.1 , 1.5 ] @xmath39 .
next , we fit parametrized models to the pressure and gas density profiles and estimate masses .
we fit pressure rather than temperature because pressure is the quantity that directly enters the hse and corrected mass estimation . also , thermal pressure is the quantity that the thermal sunyaev - zeldovich ( sz ) effect @xcite directly probes , so fitting gas density and pressure may be a better option when applying to combined sz and x - ray data .
the data points used in the fit are equally spaced logarithmic bins with uniform weighting on the radial range of either [ 0.1 , 1 ] @xmath40 or [ 0.1 , 1.5 ] @xmath39 .
these are the typical radial ranges of the currently available data on well - resolved clusters used for the mass calibration .
we first use the model presented in @xcite , which is based on an isothermal @xmath41 model @xcite with modifications capturing the cool cores and the observed steepening of the gas density profile at large radii . here
we ignore the modification for the cool cores , both because cool cores are non - existent in our simulated clusters , and that our focus is on the outer region of the clusters , especially around @xmath24 where we determine their masses .
the resulting simplified model is : @xmath42 and @xmath43 following @xcite , we fix @xmath44 . after this
, this model still has 10 parameters ( including 2 for normalization ) , and is highly nonlinear , suggesting the possible existence of local minima in the @xmath36 function . for all the parametrized models
, we fit the simulated gas density and pressure profiles to the models by searching for the global minimum of the @xmath36 function in a wide parameter range . here
, we set @xmath45 , @xmath46 , @xmath47 , @xmath48 , @xmath49 , and @xmath50 and @xmath51 being less than 1 mpc / h .
the vikhlinin model was designed to fit gas density and temperature of the icm .
it is then not surprising that the fitting form for the pressure ( equation[eq : vikp ] ) is rather redundant .
here , instead , we use a generalized navarro - frenk - white ( gnfw ) profile @xcite for pressure , as suggested by @xcite : @xmath52}^{\alpha ( \beta-\gamma)}}\ , .
\label{eq : pressure - fit}\ ] ] this simpler form already has enough flexibility to describe the shape of the observed thermal pressure profiles ( e.g. * ? ? ?
* ) , as well as our simulated @xmath53 and @xmath17 .
here we further fix @xmath54 which provides an adequate fit to all our simulated clusters . for the gnfw parameters , we search for the best fit values within @xmath55 , @xmath56 and @xmath57 between 10
kpc / h and 1 mpc / h .
an ideal parametrized model fits the data with enough flexibility with a minimum number of fitting parameters .
the way to achieve this is to utilize existing knowledge or well - founded assumptions on the underlying physical system that the data represent .
here we try to do so by assuming that the total mass density profile follows an nfw profile @xcite .
the gravitational acceleration , @xmath58 , is then given by @xmath59 where @xmath60 is the scale radius of nfw profile , whose value we search for between 10 kpc / h and 1 mpc / h .
for @xmath0 and @xmath17 we continue to use gnfw ( equation [ eq : pressure - fit ] ) .
the gas density then follows from @xmath61 and @xmath58 as @xmath62 } } { { \left [ ( 1+r / r_s)\ln(1+r / r_s ) - r / r_s \right ] } r^{\gamma-1 } { \left [ 1+(r / r_{p})^{1/\alpha } \right]}^{\alpha(\beta-\gamma)+1 } } \ , .
\label{eq : density - fit } \end{split}\ ] ] after setting @xmath54 , the model has 6 parameters .
the overdensity profiles constructed with this model for one representative cluster in the sample are shown as the colored solid lines in fig.[fig : r500c_est ] .
for this cluster , the fits to the gas density and _ total _ pressure profiles ( i.e. , corrected for non - thermal pressure ) give an overdensity profile ( blue solid line ) that agrees very well with the true one ( black solid line ) .
this suggests that at least for this cluster , the nfw assumption works and this model provides a good fit to the simulated profile .
recently , @xcite developed an automatic method for classifying the relaxation state of galaxy clusters .
their classification is based on the symmetry ( ` s ' ) , peakiness ( ` p ' ) , and alignment ( ` a ' ) of their x - ray morphologies , similar to the factors we consider for our classification ( sect.[sec : dynstate ] ) .
here we compare the two classification methods on our simulated mass - limited cluster sample at @xmath21 using mock x - ray images with 100 kilo - seconds exposure time ( fig.[fig : comparison ] ) . in @xcite classification , a cluster is categorized as relaxed when a s - p - a criterion of @xmath63 , @xmath64 , and @xmath65 ( dashed lines in fig.[fig : comparison ] ) is satisfied in more than 50% of the cases of its bootstrap analysis .
this results in a relaxed cluster fraction of about 16% , slightly smaller than the fraction ( 21% ) of our ` rather relaxed ' clusters ( blue circles and the green star in @xcite ) . in general , more relaxed clusters in our simulated sample
according to our classification also have higher @xmath66-@xmath67-@xmath68 values , showing overall agreement of the two classification methods .
the specific distribution of @xmath66-@xmath67-@xmath68 values of our simulated cluster sample is slightly shifted compared to the distribution of morphological values of the _ chandra _ sample ( fig.8 , @xcite ) : our simulated clusters have systematically higher alignment values , and lower peakiness values .
this is likely due to the inability of non - radiative simulations in reproducing the observed properties of the cluster core .
after taking the corresponding shifted @xmath66-@xmath67-@xmath68 criterion into consideration , the distribution of @xmath66-@xmath67-@xmath68 values of our ` rather relaxed ' clusters aligns rather well with that of the relaxed clusters as classified with @xcite method . | non - thermal pressure in galaxy clusters leads to underestimation of the mass of galaxy clusters based on hydrostatic equilibrium with thermal gas pressure .
this occurs even for dynamically relaxed clusters that are used for calibrating the mass - observable scaling relations .
we show that the analytical model for non - thermal pressure developed in @xcite can correct for this so - called ` hydrostatic mass bias ' , if most of the non - thermal pressure comes from bulk and turbulent motions of gas in the intracluster medium .
our correction works for the sample average irrespective of the mass estimation method , or the dynamical state of the clusters .
this makes it possible to correct for the bias in the hydrostatic mass estimates from x - ray surface brightness and the sunyaev - zeldovich observations that will be available for clusters in a wide range of redshifts and dynamical states .
[ firstpage ] galaxies : clusters : general galaxies : clusters : intracluster medium cosmology : observations methods : analytical methods : numerical |
from ginzburg - landau ( gl ) theory @xcite landau s thesis student alexei abrikosov predicted that superconductors with a gl parameter @xmath2 may contain a lattice of vortices of supercurrent , or flux lines ( fluxons ) with quantized magnetic flux @xmath3 tm@xmath4 .
abrikosov had linearized the gl equations with respect to a small order parameter @xmath5 and discovered a solution @xmath6 possessing a regular lattice of zero lines .
this lattice solution appears when the applied magnetic field @xmath7 ( along @xmath8 ) is decreased below the upper critical field @xmath9 , where @xmath10 is the gl coherence length . at @xmath11 one
has the average induction @xmath12 . with decreasing @xmath7 ,
the induction decreases and reaches @xmath13 at the lower critical field @xmath14 with @xmath15 for @xmath16 ( see below ) . at the same time
the vortex lattice spacing @xmath17 increases and diverges when @xmath18 .
abrikosov tells that he had obtained this vortex solution in 1953 but landau did nt like it , stating that there are no line - like singularities in electrodynamics .
only when feynman @xcite had published his paper on vortices in superfluid helium , did landau agree and abrikosov could publish his solution in 1957 @xcite . for his prediction of the vortex lattice
abrikosov 50 years later in 2003 received the nobel prize in physics together with vitalii ginzburg and anthony leggett .
after first evidence of the triangular vortex lattice in superconducting niobium by small - angle neutron scattering in saclay @xcite , truble and essmann in stuttgart succeeded @xcite to observe the vortex lattice directly by decorating the surface of a superconductor with iron microcrystallites ( `` magnetic smoke '' ) .
at that time i joined this research group headed by a. seeger and wrote my thesis on the theory of defects in the vortex lattice @xcite .
parts 1 and 2 deal with low inductions @xmath19 when london theory may be used and the vortices interact with each other pairwise , similar to 2d atomic lattices .
parts 3 and 4 consider high inductions @xmath20 , where the shape of the gl solutions @xmath6 and @xmath21 may be obtained from linearized gl theory , while the nonlinear gl terms determine the amplitudes of this @xmath22 and @xmath23 .
my thesis extended abrikosov s theory of periodic vortex lattices to non - periodic vortex arrangements , see below .
such distorted - lattice solutions are required to calculate the elastic energy of the vortex lattice and the energy of lattice defects like vacancies and dislocations .
they are also helpful to visualize where the solutions of the linearized gl theory apply and how they have to be modified at lower inductions .
in the usual reduced units ( length @xmath25 , induction @xmath26 , energy density @xmath27 , where @xmath28 is the thermodynamic critical field ) the spatially averaged free energy density @xmath29 of the gl theory referred to the meissner state ( @xmath30 , @xmath31 ) within the superconductor reads @xmath32 here @xmath33 is the complex gl function , @xmath34 the magnetic induction , @xmath35 the vector potential , and @xmath36 means spatial averaging over the superconductor with volume @xmath37 . introducing the super velocity @xmath38 and the magnitude @xmath39 one may write @xmath29 as a functional of the real and gauge - invariant functions @xmath40 or @xmath41 and @xmath42 , @xmath43 in the presence of vortices @xmath44 has to be chosen such that @xmath45 has the appropriate singularities along the vortex cores , where @xmath40 vanishes . by minimizing this @xmath29 with respect to @xmath22 , @xmath46 or @xmath40 , @xmath42 ,
one obtains the gl equations together with the appropriate boundary conditions . for the superconducting film considered in sec . 5 , one has to add the energy of the magnetic stray field outside the film , which makes the perpendicular component @xmath47 of @xmath48 continuous at the film surface , see below .
the two gl equations are obtained by minimization of @xmath29 ( 1 ) with respect to @xmath22 and @xmath46 , @xmath49 and @xmath50 , yielding @xmath51 = |\psi|^2 { \bf q } \ , .
\end{aligned}\ ] ] with @xmath7 and @xmath0 chosen along the @xmath8 axis and in the gauge @xmath52 , @xmath53 ( @xmath54 , @xmath55 are terms of higher order ) the linearized first gl equation , obtained by omitting the term @xmath5 in ( 3 ) , has the general solution @xmath56 this means @xmath57 , @xmath58 , can be any analytical function . for a periodic solution satisfying @xmath59 , @xmath60 , with real and reciprocal lattice vectors @xmath61 ( @xmath62 ; triangular lattice : @xmath63 , @xmath64 , @xmath65 ; square lattice : @xmath66 , @xmath67 ) and with a zero at @xmath68 , one obtains for @xmath69 the function @xmath70 defined as @xcite @xmath71 \sin(2n+1)z \,.\ ] ] thus , the periodic abrikosov solution with zeros at the @xmath72 may be written as @xmath73 this solution has the mean induction @xmath74 , normalized order parameter @xmath75 , and the fourier series @xmath76 , @xmath77 from the zero @xmath78 follows that the sum over all fourier coefficients in ( 11 ) is zero for all lattice symmetries .
( abrikosov @xcite chose a different position for @xmath79 and thus obtained the function @xmath80 @xcite ) .
the gl solution for a distorted vortex lattice near @xmath24 is obtained as follows .
assume that each of the straight and parallel vortex lines is displaced from its ideal lattice positions @xmath81 by displacements @xmath82 , @xmath83 , such that the displacement field itself is periodic with a super lattice @xmath84 times larger than the vortex lattice , but with same symmetry , @xmath85 . where needed we use a continuous displacement field @xmath86 defined such that it has the same fourier transform as the discrete @xmath87 . a distorted triangular vortex lattice with spacing @xmath88 then has the solution , eq . (
5 ) , @xmath89 \over \vartheta_1 [ ( \pi / na ) ( z - z_\nu ) , \tau ] } \ ] ] with @xmath58 , @xmath90 , @xmath91 , @xmath92 , the product is over one super cell , and @xmath93 is a normalization constant .
when all @xmath94 , the product in ( 12 ) is unity , @xmath95 , thus the first two factors in ( 12 ) are the ideal lattice solution with @xmath96 , cf .
each factor of the product shifts one zero from @xmath97 to @xmath98 .
the absolute value of @xmath99 of this gl function may also be expressed in terms of the fourier series @xmath100 ( 11 ) , @xmath101 \over \omega_a [ ( { \bf r - r}_\nu ) /n ] } \,.\ ] ] in the limit of infinite super cell , @xmath102 , one may use @xmath103 for @xmath104 , thus one may replace the function @xmath70 by its argument since all the constant factors cancel or combine to a normalization factor that follows from numerics .
one then obtains simply @xmath105 = .71.0 = .71.0
removing the central vortex at @xmath106 adds a factor @xmath107 to the linearized solution @xmath108 .
obviously , if the other vortices are not allowed to relax , this solution at large distances vanishes as @xmath107 ; it can not be normalized and its energy is infinite .
however , if the relaxation of the other vortices is chosen appropriately it will minimize the defect energy and make it finite .
this can be seen from the solution @xmath109 the constant factor @xmath110 was added to force convergence of the infinite product .
the solution for a super lattice of vacancies positioned at the @xmath111 , is given by expression ( 13 ) divided by @xmath112 that removes the zeros at positions @xmath113 .. the energy of both the ideally periodic and the distorted vortex lattices is calculated via the abrikosov parameter @xmath114 this @xmath115 enters the free energy of the linearized gl theory ( referred to the normal state ) , that has to be minimized when @xmath0 is held constant , @xmath116}\ ] ] and the free enthalpy that has to be minimized when @xmath7 is held constant , @xmath117 the elastic energy of the distorted vortex lattice is the product of the derivatives @xmath118 or @xmath119 times the change of @xmath120 times the volume , with the limit of infinite volume taken .
this means that all elastic energies and energies of structural defects near @xmath24 vanish as @xmath121 .
this is true also for the shear modulus @xmath122 of the vortex lattice , which can be obtained using abrikosov s periodic lattice solution @xcite . in the case of the vortex vacancy ,
the resulting defect energy is finite only if the vortices relax ( shift towards the removed vortex ) such that at large distances ( and after numerical minimization practically at all distances ) the vortex displacements are @xmath123 with @xmath124 the vortex density .
if the radial displacements were chosen smaller ( larger ) than in ( 19 ) , the order parameter ( 15 ) would vanish ( diverge ) at large distances @xmath125 . but with the correct displacements that minimize @xmath115 and thus the defect energy , the amplitude of the oscillating order parameter stays almost constant , even near the vacancy .
this can be seen in fig . 1 , where the profiles of @xmath126 along @xmath127 are plotted for ideal triangular vortex lattice and for the lattice with a central vacancy with the simple relaxation ( 19 ) and with an improved relaxation field @xmath128 $ ] .
figure 2 shows the contour lines of the fully relaxed order parameter @xmath126 , which has a maximum at the origin ( vacancy position ) and minima ( zeros ) at the vortex positions .
the displacement ( 19 ) means that the area of the `` relaxing ring '' , @xmath129 , exactly equals the area of one lattice cell . in other words ,
the removal of one vortex at @xmath130 is compensated by the relaxation of the surrounding vortices such that the average vortex density in any contour ( containing or not containing the vacancy ) stays constant and equals the density @xmath131 that was there before the vacancy was introduced .
note that the field ( 19 ) satisfies @xmath132 , and thus describes a pure shear deformation .
more precisely , one has @xmath133
( @xmath134 is the 2d delta function ) , i.e. , the displacement field ( 19 ) `` remembers '' that one vortex cell area was removed .
further interesting properties of structural defects in vortex lattices and other two- or three - dimensional soft lattices are discussed in [ 10 ] .
as shown with the vacancy example , the distorted - lattice solution ( 14 ) of the linearized gl equations yields finite energies of lattice defects only if most of the vortex displacements are allowed to relax appropriately .
but without this relaxation , the elastic energy is infinite .
for example , if only the one vortex at the origin is displaced , @xmath135 ( @xmath136 if @xmath137 , else @xmath138 ) one has from ( 14 ) @xmath139 i.e. , the periodic order parameter is modulated by a slowly decreasing function .
the abrikosov @xmath115 of this defect times the volume , @xmath140 , diverges and so the defect energy diverges .
this unphysical divergence of defect energies of the vortex lattice is removed when the influence of the nonlinear gl terms on the solutions @xmath126 and @xmath21 is accounted for .
this calculation was performed in a series of 4 papers @xcite : parts 1 and 2 deal with the linear elastic energy of the vortex lattice at low and high inductions @xmath0 .
parts 3 and 4 derive the gl solutions for the distorted vortex lattice when the vortex lines are straight and parallel or arbitrarily curved .
the essential result is that the long - ranging modulation factors like @xmath141 in the linearized solution ( 14 ) become exponentially damped over a new length @xmath142 with @xmath143 .
as @xmath144 , this screening length becomes infinite and the linearized solution ( 14 ) is recovered . at @xmath145 , the distorted - lattice solution ( 14 ) should be replaced by @xmath146 ^ 2 + o(s^2 ) \
, , \end{aligned}\ ] ] where @xmath147 is a modified bessel function with the limits @xmath148 ( @xmath149 ) , @xmath150 ( @xmath151 ) . this generalized expression up to terms linear in the vortex shifts @xmath152 reproduces the linearized solution ( 14 ) , ( 20 ) when @xmath153 , but it does not possess the correct zeros at @xmath154 . this may be corrected by replacing in ( 21 ) the periodic order parameter @xmath100 by the `` phase modulated '' @xmath155 $ ] and cutting the infinity of @xmath156 off .
the resulting solution is still exact up to linear terms in @xmath152 since @xmath157 vanishes at the @xmath158 and thus the expansion of @xmath155 $ ] contains no linear term .
the screening length @xmath159 may be derived by considering only one fourier component of the displacement field , @xmath160 with @xmath161 and re = real part .
one may then write the linearized solution as @xmath162 ^ 2
+ o(s^2 ) \ , , \\ \eta({\bf r } ) = 2 \sum_\nu { \bf s}_\nu { { \bf r - r}_\nu \over ( { \bf r - r}_\nu ) ^2 } % \nonumber \\ = { 2b\over \xi^2 } { \rm re } \bigg\ { { \bf s}_0 \sum_{\bf k } { i({\bf k+k } ) \over ( { \bf k+k})^2 } \exp [ i{\bf ( k+k)r } ] \bigg \ } \ , .
\end{aligned}\ ] ] in @xmath163 ( 24 ) the terms with reciprocal lattice vectors @xmath164 shift the zeros of @xmath108 ( `` phase modulation '' ) , while the term @xmath165 yields an `` amplitude modulation '' of @xmath108 .
this term diverges as @xmath166 , i.e. , it yields a diverging amplitude modulation when the wavelength of the displacement field is large . from physical reasons
it is clear that this term @xmath167 has to be cut off , e.g. , replaced by @xmath168 .
accounting for all the gl terms nonlinear in @xmath169 ( @xmath170 , terms like @xmath171 , @xmath172 , @xmath173 ) indeed yields such a cut off , with @xmath174 .
the resulting solution for periodic @xmath86 may be written as @xmath175 \bigg [ 1 + { 2b\over \xi^2 } { \nabla { \bf s(r ) } \over k^2 + k_\psi^2 } \bigg ] + o(s^2 ) \ , .
\end{aligned}\ ] ] in a similar way , the solution for the induction @xmath21 of the linearized gl theory , @xmath176 is modified by the nonlinear terms to give for periodic @xmath86 @xmath177 - { \bar b~ \nabla { \bf s(r ) } \over 1 + k^2 /k_h^2 } + o(s^2 ) \end{aligned}\ ] ] with @xmath178 and @xmath179 the ideal periodic solution for @xmath180 . in deriving ( 27 ) all terms containing @xmath181
have cancelled . from the solutions ( 25 ) and
( 27 ) for periodic @xmath86 , the generalization to arbitrary displacement fields is obtained by fourier transform .
the above method can be extended to 3d displacement fields @xmath182 $ ] describing distorted lattices of curved vortices , @xmath183 where now @xmath184 , @xmath185 , @xmath186 , and the @xmath187 integration extends over the first brillouin zone of the ideal vortex lattice [ since @xmath188 and over @xmath189 .
the coordinate @xmath8 plays here the role of a line parameter .
the order parameter which solves the gl equations near @xmath24 and has zeros at the vortex positions @xmath190 is @xmath191 + o(s^2 ) \ , .
\end{aligned}\ ] ] the 3d solution for the induction @xmath192 for periodic @xmath86 after averaging over a vortex cell may be written as @xmath193 these expressions coincide with the first - order expansion terms ( in @xmath194 ) of a linear superposition of spherical `` source fields '' centered at each vortex element : @xmath195 where @xmath196^{1/2}$ ] has an inner cut - off @xmath197 , half the vortex spacing in our derivation from the vortex lattice .
the expression ( 32 ) is also the solution of london theory for arbitrarily arranged curved or straight vortices if one puts @xmath198 ( i.e. @xmath199 ) and the vortex core radius @xmath200 for the inner cutoff .
the line element of the path integral in ( 32 ) may be parameterized with @xmath8 as line parameter and integration variable , @xmath201
the distorted - lattice solution up to terms linear in the displacements @xmath202 can be used to calculate the linear elastic energy of the vortex lattice , @xmath203 , referred to the perfect lattice ( the equilibrium state ) .
the most general expression quadratic in the 2d displacements @xmath204 , or in their fourier transforms @xmath205 , ( 28 ) , is @xmath206 where the sum over the indices @xmath207 , @xmath115 = ( @xmath208 ) is taken .
the @xmath209 matrix @xmath210 is the elastic matrix .
this expression applies for both an elastic continuum and for a lattice . for a lattice @xmath210
is periodic , @xmath211 , and thus the integral should be restricted to the first brillouin zone ( bz ) .
the bz for the triangular lattice is a hexagon , and for the square lattice a square . where required , the bz may be approximated by a circle with radius @xmath212 , @xmath213 , and area @xmath214 , @xmath215 . for a uniaxial elastic continuum
the elastic matrix @xmath216 has the form @xmath217\ , .
\end{aligned}\ ] ] in it the coefficients are the elastic moduli : @xmath218 the isotropic compression modulus , @xmath219 the uniaxial compression modulus , @xmath122 the shear modulus , and @xmath220 the tilt modulus .
the elastic moduli of the vortex lattice are obtained by deriving the elastic energy , e.g. , from gl theory and comparing it at @xmath221 with the continuum limit ( 35 ) .
this yields @xmath222 ^ 2 } ( 1 - 0.3b ) \\
c_{44}(k ) & = & { \bar b^2 \over \mu_0}\ , { 1 \over 1 + k^2/k_h^2 } + { \bar b ( b_a -\bar b ) \over \mu_0 } \ , . \end{aligned}\ ] ] these expressions are exact at large reduced induction @xmath223 and for all @xmath1 , but they are written such that they reduce to the correct values also in the limit of small induction @xmath19 . in @xmath122 , @xmath224 is the abrikosov parameter of the triangular lattice ( the square lattice is unstable and thus has negative @xmath122 ) ; the third factor reduces to 1 for @xmath225 and to @xmath226 for @xmath227 , which means the shear stiffness of the vortex lattice is zero in superconductors with @xmath228 ; the factor @xmath229 interpolates between the the correct limits at @xmath144 and @xmath199 .
in particular , for @xmath230 and @xmath225 , ( 37 ) reproduces the london result @xmath231 .
an interesting result is the dependence of @xmath219 ( 36 ) and @xmath220 ( 38 ) on @xmath232 , which means the elasticity of the vortex lattice is * non - local*. in the limit of uniform stress , @xmath233 , these expressions reproduce the known values of the compression and tilt moduli obtained by thermodynamics , @xmath234 , @xmath235 .
however , when the wavelength of the periodic compression or tilt decreases , i.e. , the wave vector @xmath236 increases , these moduli decrease .
this means , the vortex lattice is softer for short - wavelengths compression and tilt than it is for long wavelengths .
the two characteristic lengths or wave vectors were already introduced above , @xmath237 and @xmath238 .
this dispersion or elastic non - locality means , e.g. , that a point force exerted by a small pinning center on the vortex lattice , deforms the vortex on which it acts not like plugging a string but more , causing a sharp cusp since a local deformation costs little energy .
if the interaction of the vortices with the pinning center is via the order parameter @xmath5 or via the gradient term in the gl functional , then this interaction itself is nonlocal , smeared over the length @xmath239 . in the expressions for the elastic force and the elastic energy there is thus a factor @xmath240 in the numerator that compensates the same factor in the denominator originating from @xmath241 , ( 36 ) .
therefore , the factor @xmath242 in @xmath219 has no physical meaning in pinning problems since near @xmath24 where @xmath243 can be larger than the vortex spacing @xmath244 , it is not possible to exert a pinning force on one single zero of the order parameter but only on an area with radius @xmath243 containing several such zeros .
the nonlocality factor @xmath245 in @xmath220 , however , is important in pinning theories since it strongly enhances the elastic deformations caused by small pins acting on the vortex cores . in (
not very realistic ) models where the pinning force acts only the magnetic field of the vortex but not on the vortex cores , this enhancement of the elastic displacement may vanish , cancelled by the non - locality of this model force .
the correct , non - local elasticity thus effectively softens the vortex lattice and leads to large , pinning - caused distortions and disorder of the vortex lattice .
furthermore , the thermal fluctuations of the vortex lattice are strongly enhanced by this non - local elasticity . in both cases
the lattice softening is caused mainly by the dispersion of @xmath246 , while the dispersion and reduction of @xmath241 is not so important since the shear modulus @xmath122 is typically much smaller than @xmath241 and the shear modes of the elastic deformation thus dominate over the compressional modes .
= .71.0 = .71.0 = .71.0 = .71.0 = .71.0 = .71.0 = .71.0 = .71.0 = .71.0 = .71.0
at low inductions @xmath247 and not too small @xmath248 , the gl theory for arbitrary 3d arrangements of vortices reduces to the london theory , which may be expressed by the energy functional @xmath249 \ , .
\end{aligned}\ ] ] here @xmath25 is the london depth equal to the gl magnetic penetration depth .
minimizing @xmath250 with respect to the induction @xmath251 using @xmath252 , and adding appropriate singularities along the positions @xmath253 of the vortex cores , one obtains the modified london equation @xcite , @xmath254 with @xmath255 the 3d delta function . from this one obtains the energy of an arbitrary arrangement of straight or curved vortices , @xmath256 in this double sum the terms @xmath257 describe the pairwise interaction of the vortex line elements @xmath258 , @xmath259 over the distance @xmath260 . the term @xmath261 is the self - energy of the @xmath262th vortex line , which depends on the shape of this vortex . in it
an inner cut off is needed , obtained , e.g. , by putting @xmath263 with @xmath200 the vortex core radius , to avoid divergence when in the integral the parameters equal , @xmath264 . from the gl nonlocal elastic energy ( 34)-(38 ) one may construct an effective interaction potential between vortex line elements such that the full nonlocal linear elastic energy is reproduced at small displacements @xcite . at the same time
, this interaction at low @xmath265 reproduces the london interaction for arbitrary vortex arrangements , and an approximate gl interaction valid at all @xmath266 and @xmath1 , @xmath267 \end{aligned}\ ] ] with @xmath260 .
for @xmath265 the first term in ( 42 ) reproduces the magnetic repulsion of london vortices with @xmath268 ; this magnetic interaction is vectorial due to the product @xmath269 containing the cosine of the angle between two line elements .
the second term of shorter range @xmath270 may be interpreted as an attraction caused by the overlap of the vortex cores , the regions where the order parameter is reduced : two overlapping cores require less ( positive ) condensation energy than two separated cores , thus the cores attract .
this attraction has scalar character , hence the product @xmath271 . the attractive second term in ( 42 )
removes the logarithmic divergence of the magnetic repulsion at zero distance , since both terms have the same singularity but of opposite sign , such that they cancel .
the total potential ( 42 ) is thus a smooth function that at @xmath272 starts with a finite value and then decreases monotonically to zero with increasing distance @xmath273 . for straight parallel vortex lines the general 3d energy expression ( 42 )
simplifies to the sum of the vortex self energies , @xmath274 per unit length and per vortex , and the interaction energy @xmath275 of all vortices per unit length , @xmath276 \ , .
\end{aligned}\ ] ] here @xmath147 is a modified bessel function , see eq . (
the effective 2d interaction potential in ( 43 ) is a smooth , monotonically decreasing function with a finite value at @xmath272 since the two logarithmic singularities of the @xmath156 functions cancel each other . as in the 3d expression
( 42 ) , the first term in ( 43 ) is the magnetic repulsion of the straight vortices , and the second term is an attraction due to gain in condensation energy during the overlap of vortex cores .
abrikosov s solution method for the periodic vortex lattice starts from the linearized gl theory and is thus valid only at large inductions @xmath0 near the upper critical field @xmath24 .
first numerical solutions for all @xmath0 and @xmath1 were obtained by the `` circular cell method '' @xcite that approximates the hexagonal wigner - seitz cell of the triangular vortex lattice by a circle and solves a cylindrically symmetric problem , see also ref .
the periodic solution in the entire ranges of reduced induction @xmath277 and gl parameter @xmath278 may be obtained for bulk superconductors by the following numerical method @xcite .
we start from the free energy functional @xmath29 , eq .
( 2 ) , and minimize it with respect to the real and periodic functions @xmath279 ( order parameter ) and @xmath280 ( negative super velocity ) or @xmath281 ( induction ) .
we consider periodic lattices with one flux quantum per vortex . in the sense of a ritz variational method
we use fourier series for the periodic trial functions with a finite number of fourier coefficients @xmath282 and @xmath283 , @xmath284 where @xmath285 and @xmath286 are the reciprocal lattice vectors ( 8) of the vortex lattice with positions ( 7 ) . in all sums here and below the term @xmath287
is excluded . in ( 46 ) @xmath288 is the super velocity of the abrikosov @xmath24 solution , which satisfies @xmath289 { \bf\hat z}\ , , \end{aligned}\ ] ] where @xmath290 is the 2d delta function .
this relation shows that @xmath291 is the velocity field of a lattice of ideal vortex lines but with zero average rotation .
close to each vortex center one has @xmath292 and @xmath293 with @xmath294 . in principle @xmath295
may be expressed as a slowly converging fourier series by integrating ( 47 ) using @xmath296 as in ref .
but it is more convenient to take @xmath291 from the exact relation @xmath297 where @xmath298 is the abrikosov @xmath24 solution given by the rapidly converging series ( 11 ) .
with ( 48 ) the numerical method becomes highly accurate .
note that the ansatz ( 46 ) assumes that @xmath299 .
this assumption can be shown to be exact at high and low inductions , but i did not find a proof that it is true in the general case , though it is satisfied numerically with high precision for the periodic vortex lattice at all @xmath0 and @xmath1 . the solutions @xmath108 and @xmath300 may be computed by using a finite number of fourier coefficients @xmath301 and @xmath283 and minimizing the free energy @xmath302 with respect to these coefficients as done in @xcite . however , a much faster and more accurate solution method @xcite is to iterate the two gl equations @xmath303 and @xmath304 written in appropriate form .
the iteration is stable and converges rapidly if one isolates a term @xmath305 on the l.h.s . and puts the remaining terms to the r.h.s . as an `` inhomogeneity '' of such london - like equations , e.g. , @xmath306 with the abbreviations @xmath307 , @xmath308 , @xmath309 , and @xmath310 .
equations ( 49 ) , ( 50 ) introduce some `` penetration depths '' @xmath311 and @xmath312 ( in real units ) , which stabilize the convergence of the iteration .
acting on the fourier series @xmath313 ( 44 ) and @xmath314 ( 46 ) the laplacian operator @xmath315 yields a factor @xmath316 , which facilitates the inversion of ( 49 ) and ( 50 ) . using the orthonormality @xmath317 valid for @xmath164 , one obtains from ( 44 ) , ( 45 ) @xmath318 and @xmath319 .
the convergence of the iteration is considerably improved by adding a third equation which minimizes @xmath29 ( 2 ) with respect to the amplitude of @xmath313 , i.e. , @xmath320 .
this step gives the largest decrease of @xmath29 .
the resulting three iteration equations for the parameters @xmath282 and @xmath283 then read @xmath321 \cos{\bf k r } \rangle \over
k^2 + \bar \omega } \
, , \end{aligned}\ ] ] with @xmath322 and @xmath323 as above .
the solutions @xmath108 , @xmath251 , and @xmath324 are then obtained by starting , e.g. , with @xmath325 [ the abrikosov solution ( 11 ) ] and @xmath326 and then iterating the three equations ( 52)-(54 ) by turns until the coefficients do not change any more . after typically 25 such triple steps , the solution stays constant to all 15 digits and the gl equations are exactly satisfied .
since all terms in ( 52 ) - ( 54 ) are smooth periodic functions of @xmath327 , high accuracy is achieved by using a regular spatial 2d grid , e.g. , @xmath328 ( @xmath329 ) and @xmath330 ( @xmath331 , @xmath332 ) with constant weights @xmath333 and @xmath334 .
these @xmath335 = 100 to 5000 grid points fill the rectangular basic area @xmath336 , @xmath337 , which is valid for any unit cell with the shape of a parallelogram .
spatial averaging @xmath338 then just means summing @xmath84 terms and dividing by @xmath84 .
best accuracy is achieved by considering all @xmath339 vectors within a half circle @xmath340 , with @xmath341 chosen such that the number of the @xmath339 is slightly less than the number @xmath84 of grid points .
the high precision of this method may be checked with the identity @xmath342 , which is valid at @xmath343 for all @xmath266 .
this relation is confirmed with an error @xmath344 .
the reversible magnetization @xmath345 and the equilibrium field @xmath346 ( the applied field ) are easily computed from doria s virial theorem @xcite , which in our reduced units reads @xmath347 in this way we find the lower critical field @xcite , @xmath348 , @xmath349 \ , , ~~ \nonumber h_{c1 } = { b_{c1 } \over b_{c2 } } = { \ln \kappa + \alpha(\kappa ) \over 2\kappa^2 } \ , , \\
\alpha(\kappa ) & = & \alpha_\infty + \exp [ - c_0 - c_1 \ln\kappa - c_2 ( \ln\kappa)^2 ] \pm \epsilon % \alpha(\kappa ) = 0.49693 + \exp[-0.775\ln\kappa % \\ \nonumber -0.1303(\ln\kappa)^2 - .41577 ] \pm 0.00076 \end{aligned}\ ] ] with @xmath350 , @xmath351 , @xmath352 , @xmath353 , and @xmath354 .
this expression yields at @xmath343 the correct value @xmath355 and for @xmath16 it has the limit @xmath356 .
a simpler expression for @xmath357 , yielding an @xmath358 with error still less than 1% and with the correct limits at @xmath359 and @xmath16 , is @xmath360 the resulting magnetization curves @xmath361 are shown in fig .
they are well fitted by @xmath362 with @xmath358 from eq .
( 56 ) and @xmath363 $ ] ( @xmath364 ) , @xmath365 ( 1.18034 ) for the triangular ( square ) vortex lattice .
this @xmath366 satisfies the exact relations : @xmath367 , @xmath368 , @xmath369 , @xmath370 .
the @xmath371 from the fit ( 58 ) applies for not too large @xmath372 . for larger @xmath1
, better fits are given in ref .
@xcite , where it is also shown that the often used `` logarithmic law at @xmath373 '' for the magnetization @xmath374 has very limited range of validity .
figure 4 shows the profiles @xmath375 and @xmath376 for an isolated vortex line @xmath18 from gl theory for @xmath377 2 , 5 , and 20 .
profiles for the vortex lattice are shown in fig . 5 for @xmath378 and for two inductions at which the vortex spacing is @xmath379 and @xmath380 .
contour lines for @xmath126 and @xmath21 are plotted in fig . 6 for @xmath381 at @xmath382 . at @xmath383
the contours of @xmath126 ( see right hand part in fig
. 2 ) and @xmath21 practically coincide , and at @xmath384 the contours are nearly circular , around well separated vortex cores and field peaks .
the variance of the magnetic field @xmath385 ^ 2 \rangle = \sum_{{\bf k}\ne 0 } b_{\bf k}^2 \end{aligned}\ ] ] is shown in fig .
7 . in the low - field range @xmath386
one has for the triangular lattice the london limit @xmath387 ( upper frame in fig .
7 ) , at very small @xmath388 one has @xmath389 ( dash - dotted straight lines in in fig .
7 ) , and near @xmath390 one has the abrikosov limit @xmath391 ^ 2 $ ] .
this field variance is needed , e.g. , for the interpretation of muon spin rotation ( @xmath262sr ) experiments @xcite . in fig .
8 the shear modulus of the bulk triangular vortex lattice is plotted versus the reduced induction for various gl parameters @xmath1 .
note that for @xmath392 where @xmath393 , one has @xmath394 .
one can show that in the particular case @xmath395 all possible vortex configurations have the same free energy @xmath396 , e.g. , triangular and square lattice , lattices with two flux quanta per vortex , or all vortices merged into one giant vortex .
the 2d fourier method for the bulk vortex lattice in sec .
9 can be generalized to the 3d problem of vortex lattices in infinite films of arbitrary thickness @xmath397 put into a uniform magnetic field @xmath398 , since then the functions @xmath399 and @xmath400 are still periodic in the @xmath401 plane .
i consider here the case when @xmath402 is perpendicular to the film plane @xcite , though in principle the fourier method applies also to tilted applied field .
the total free energy @xmath403 per unit volume of the infinite film is the free energy , eq .
( 2 ) , plus the stray - field energy @xmath404 , i.e. , the energy of the magnetic field variations outside the film , @xmath405 one has @xmath406 since all field lines have to cross the infinite film .
the factor of 2 in ( 60 ) comes from the two half spaces above and below the film , which contribute equally to @xmath404 .
the stray field @xmath407 with constant planar average @xmath408 is determined by the laplace equation @xmath409 ( since @xmath410 and @xmath411 in vacuum ) and by its perpendicular component at the film surface @xmath412 , since @xmath47 has to be continuous across this surface .
the trial functions for @xmath108 , @xmath413 , and @xmath414 inside the film ( @xmath415 ) are now 3d fourier series @xcite , @xmath416 here @xmath184 , @xmath417 , @xmath418 with @xmath419 from eq .
( 8) , and @xmath420 , @xmath421 . in all sums here and below the term @xmath422 is excluded . minimizing @xmath403 with respect to the coefficients @xmath282 and @xmath283 and using the appropriate orthonormality relations one arrives at iteration equations for the @xmath282 and @xmath283 similar to eqs .
( 52)-(54 ) .
the solution is then obtained by first finding the 2d bulk solution as in sec .
10 by considering only the terms with @xmath423 .
the magnetic field lines then still have an unphysical sharp bend at the surface , see left part of fig .
next we allow for the terms with @xmath424 .
this yields a `` mushrooming '' of the field lines of each vortex when it approaches the film surface such that these lines smoothly cross the surface with no bend , see right part of fig .
9 . profiles @xmath425 and @xmath426 are shown in fig . 10 for @xmath427 ( middle plane of the film ) and @xmath428 ( film surface ) .
one can see that the spatial variation of @xmath23 at the surface is reduced from its bulk value by nearly @xmath429 . outside the film , the transverse field components @xmath430 , @xmath431 rapidly decrease as @xmath432 where @xmath433 is the distance from the surface .
interestingly , the profile of the order parameter @xmath313 in films is almost independent of @xmath8 .
the shear modulus @xmath122 of the triangular vortex lattice in films can be positive even when @xmath434 , provided the film thickness @xmath397 is smaller than the coherence length @xmath435 , see fig .
this means that a stable vortex lattice may exist in thin type - i superconductor films .
our numerical result confirms the @xmath122 of films that was calculated analytically near @xmath24 by albert schmid @xcite .
in such infinitely extended films one has @xmath406 since all field lines have to pass the film .
therefore , the magnetization @xmath436 of the film can not be calculated as a difference of fields , but one has to take the derivative of the total free energy , @xmath437 .
a more elegant method calculates @xmath436 by doria s virial theorem @xcite , which for bulk superconductors yields eq .
indeed , it was shown recently @xcite that this virial idea can be generalized to films of arbitrary thickness and @xmath436 may be calculated directly from the gl solution for the film , with no need to take an energy derivative .
the resulting magnetization of the film is plotted versus the applied field in fig . 12 for various film thicknesses @xmath397 and various gl parameters @xmath1 .
the curves for various @xmath397 cut each other at @xmath438 .
for @xmath439 the thick film limit ( @xmath440 but still film width @xmath441 ) yields a straight line , @xmath442 .
this result is valid for large demagnetization factor @xmath84 , @xmath443 ; it differs from the bulk result for @xmath436 shown in fig . 3 , which is valid for the demagnetization - free limit of @xmath444 .
one may ask why the magnetization @xmath445 is not calculated via the general definition of the magnetic moment @xmath446 @xcite from the current density @xmath447 , which is easily calculated as a periodic function by our fourier method , both for bulk and film superconductors .
however , into this definition enters not only the periodic part , i.e. , the vortex currents circulating inside each vortex cell ; this contribution even would give the wrong sign of @xmath448 . the main contribution to
the magnetic moment of a superconductor of any shape comes from the screening currents that flow near the surface of the specimen .
the magnetization in superconductors is thus not a volume property as it is in magnets .
for example , for a long cylinder in parallel field @xmath7 the magnetization @xmath449 is composed of the positive contribution @xmath0 of the vortex currents and the negative ( and larger ) contribution @xmath7 of the surface currents that screen the cylinder from the applied field @xmath7 before vortices are allowed to penetrate . near @xmath24 , both terms nearly compensate and @xmath450 is a small difference of two big terms . in the film geometry , and for most other shapes of the superconductor ,
the screening current is not easily known but has to be computed ; such computations for thin and thick strips , disks , and plates of macroscopic size @xmath451 are presented in @xcite .
landau knew this problem , since he worked on demagnetization factors and on the intermediate state in type - i superconductors @xcite , and he used the thermodynamic definition of the magnetization as an energy derivative . in my view , the simple formula ( 55 ) derived by doria , gubernatis and rainer @xcite by scaling the coordinates in gl theory and finding a novel virial relationship between kinetic and potential energies from which the equilibrium field @xmath452 follows , was a fundamental discovery that occurred long after the publication of gl theory @xcite . i. l. landau and h. keller , phys .
c * 466 * , 131 ( 2007 ) .
e. h. brandt , j. low temp .
phys . * 73 * , 355 ( 1988 ) .
e. h. brandt and a. seeger , adv .
physics * 35 * , 189 ( 1986 ) .
e. h. brandt , to be published in physica b : phys .
cond . mat .
, special issue @xmath262sr2008 . | abrikosov s solution of the linearized ginzburg - landau theory describing a periodic lattice of vortex lines in type - ii superconductors at large inductions , is generalized to non - periodic vortex arrangements , e.g. , to lattices with a vacancy surrounded by relaxing vortices and to periodically distorted lattices that are needed in the nonlocal theory of elasticity of the vortex lattice .
generalizations to lower magnetic inductions and to three - dimensional arrangements of curved vortex lines are also given .
finally , it is shown how the periodic vortex lattice can be computed for bulk superconductors and for thick and thin films in a perpendicular field for all inductions @xmath0 and ginzburg - landau parameters @xmath1 . for electronic journal of theoretical physics ,
landau issue 2008 . 05/06/2008 + * the vortex lattice in ginzburg - landau superconductors * + ernst helmut brandt + max planck institute for metals research , d-70506 stuttgart , germany |
this paper concerns with the numerical solution of the finite horizon optimal investment problem with transaction costs under potential utility .
let us consider an investor whose wealth can be inverted in a risky stock and in a riskless bank account .
we suppose that the investor is risk averse with constant relative risk aversion ( crra ) . in @xcite
, merton showed that , in absence of transaction costs , the problem can be explicitly solved .
the optimal strategy consists in keeping a fixed proportion between the money invested in the risky asset and the bank account . when transaction cost are considered , the merton strategy is unfeasible because it requires a continuous portfolio rebalancing with unbounded costs .
proportional transaction costs were first introduced in @xcite .
more recently , in @xcite , the problem was reformulated as a non - linear parabolic double obstacle problem posed in one spatial variable , defined in an unbounded domain .
several explicit properties and formulae were obtained in @xcite , although explicit formulae for the solution are not available .
the problem was numerically solved in @xcite , where the authors employ a characteristics method with a projected relaxation scheme .
the scheme proportionates satisfactory results with good agrement with the results in @xcite .
when we want to solve financial problems , like finding investment strategies or pricing derivative contracts , in general , there is no known closed form solution of the different problems and several numerical methods have been employed . without the aim to be exhaustive , monte - carlo based methods ( @xcite , @xcite ) , piecewise linear interpolations ( @xcite ) , lattice methods ( @xcite , @xcite ) , finite elements ( @xcite ) or spectral methods ( @xcite ) are some of them .
a general review of financial problems or models , numerical techniques and software tools can be found in @xcite .
the objective of this paper is to construct a spectral method specifically adapted to the optimal investment problem with potential utility when proportional transaction costs are present .
as it is well known , spectral methods @xcite are a class of spatial discretizations for partial differential equations which offer fast convergence in the case of smooth solutions .
they are not widely used yet in numerical finances because it is usually believed that the lack of smoothness present in most interesting problems makes spectral methods uncompetitive . however
, several papers have used spectral methods for problems in finance with good results .
for instance , in @xcite a fourier - hermite procedure to the valuation of american options has been presented . in @xcite
a spectral method based on laguerre polynomials has been employed to numerical valuation of bonds with embedded options . a fourier spectral method to compute fast and accurate prices of american options written on asset following
garch models has been presented in @xcite . in @xcite
the authors use an adaptive method with chebyshev polynomials coupled with a dynamic programming procedure for contracts with early exercise features . in @xcite
a very efficient procedure for asian options defined on arithmetic averages has been proposed . in all cases ,
the spectral - based methods have been proved to be competitive with other alternatives in terms of precision versus computing time needed to compute the numerical solution . in the present paper , we restate the problem using polar coordinates .
this allows to consider a double parabolic obstacle problem in one spatial - like variable defined in a bounded domain .
furthermore , this formulation avoids the emergence of nonlinear terms simplifying the numerical treatment .
we present a chebyshev spectral approach based on adaptive meshes to locate the optimal frontiers .
although some of the numerical difficulties that appear with the parabolic double obstacle problem are avoided with our approach , we still have to deal with the so - called gibbs effect , which comes from the fact that the objective function is continuous but not differentiable at maturity .
we show that this issue can be circumvented by using a time - adapted spatial mesh .
we show that our approach is efficient by comparing it with a standard finite difference scheme .
the outline of the paper is as follows . in section [ ch3toiprecon ] ,
a description of the optimal investment problem as it can be found in @xcite or @xcite is presented . in section [ ch3dpr ] ,
the problem is reformulated as a parabolic double obstacle problem as it was done in @xcite .
afterwards , we propose an equivalent formulation of the problem employing polar coordinates .
ch3 nm ] is devoted to a mesh - adapted chebyshev - collocation method which solves the problem of section [ ch3dpr ] . in section [ ch3numresults ]
we perform the numerical analysis of the method .
section [ ch3conclus ] presents some conclusions and future research .
we consider an optimal investment problem with transaction costs , @xcite , @xcite .
let @xmath0 be a filtered probability space .
let us consider an investor who holds amounts @xmath1 and @xmath2 in a bank and a stock account respectively .
the dynamics of the processes is @xmath3 where @xmath4 denotes the constant risk - free rate , @xmath5 is the constant expected rate of return of the stock , @xmath6 is the constant volatility of the stock and @xmath7 is a standard brownian motion such that @xmath8 where @xmath9 is the natural filtration induced by @xmath7 .
we suppose that @xmath10 and @xmath11 are adapted , right - continuous , nonnegative and nondecreasing processes representing the cumulative monetary values of the stock purchased or sold respectively and @xmath12 and @xmath13 , represent the constant proportional transaction costs incurred on the purchase or sale of the stock . in this paper
we assume that @xmath14 .
the finance meaning of equation ( [ ch3ecudinacc ] ) is natural . along time , the rate of change of the amount of money invested in the risky asset , represented by the stochastic process @xmath2 , evolves according to a standard geometric brownian motion modified by the difference between the amount of money invested in buying stock , @xmath15 , and the amount of money obtained selling stock , @xmath16 . at the same time
, the value of the bank account , @xmath1 , is instantaneously increased by the difference @xmath17 , that represents the net flow of money resulting from stock negotiations , including the transaction costs .
processes @xmath10 and @xmath11 can be financially understood as an historical record of the total purchases and sales of stock of the investor .
the net wealth is the money the investor would have if he closes his positions .
it can be written as @xmath18 if the investor is long in the stock or @xmath19 in case the investor is short in the stock .
let @xmath20 be a utility function , that is , a continuous , strictly increasing , concave function .
the optimal value function is given by : @xmath21,\ ] ] for all @xmath22 $ ] , where @xmath23 is the set of admissible strategies , defined as the set of processes @xmath24 such that if @xmath25 then @xmath26 for @xmath27 and where @xmath28 is the solvency region , @xmath29 in this paper , we assume that @xmath20 is a potential function ( constant relative risk aversion utility function ) of the form @xmath30 for some constant @xmath31 , @xmath32 .
the optimal value function ( [ ch3defvarpphi ] ) , see @xcite , is the viscosity solution in @xmath33 $ ] of @xmath34 subject to : @xmath35 where @xmath36 the existence and uniqueness of a viscosity solution of ( [ chp3hjbecu1])-([chp3hjbecu2 ] ) has been proved in @xcite .
there , it is proved that at any time @xmath37 , the spatial domain is divided in three regions , namely , in financial terms , the buying region @xmath38 , the selling region @xmath39 and the no transactions region @xmath40 .
the selling and buying regions do not intersect .
for simplicity in the exposition , we suppose that @xmath41 . with this hypothesis ,
short - selling is always a suboptimal strategy @xcite , @xcite , @xcite .
this means that the optimal trading strategy is always to have a nonnegative amount of money invested in the stock .
as remarked in @xcite , the choice of the potential utility function is interesting since it leads to the homothetic property in the optimal value function , @xmath42 this property is used in @xcite to reduce the dimensionality of the problem . setting @xmath43 , a new function @xmath44 is introduced in @xcite , so that : @xmath45 in @xcite , the authors prove that @xmath46 is the solution of an one dimensional parabolic double obstacle problem with two free boundaries equivalent to ( [ chp3hjbecu1 ] ) .
furthermore , it is also proved in @xcite , that there exist two continuous monotonically increasing functions @xmath47\rightarrow(-(1-\mu),+\infty],\ ] ] such that @xmath48 .
the buying and selling regions are characterized by @xmath49 \
\mid z\leq\text{sr}^c_f(t ) , t\in[0 , \ t ] \right\ } , \\
\text{\textbf{br } } & = \left\{(z , t)\in\omega\times[0,t ] \
\mid z\geq\text{br}^c_f(t ) , t\in[0 , \ t ] \right\}. \\ \end{aligned}\ ] ] although other properties and explicit formulas are obtained in @xcite , a complete analytical solution is still missing and numerical procedures have to be used , see , for example , @xcite .
here , inspired by @xcite , we take advantage of ( [ homothetic ] ) by working in polar coordinates @xmath50 , @xmath51 .
it is not difficult to show that ( [ chp3hjbecu1])-([chp3hjbecu3 ] ) are equivalent to @xmath52 subject to : @xmath53 where @xmath54 and @xmath55 based on ( [ homothetic ] ) , we conjecture a solution to ( [ ch3ecuinicpol ] ) of the form : @xmath56 taking into account that @xmath57 and substituting in ( [ ch3ecuinicpol ] ) , ( [ finalcondition1 ] ) , we see that @xmath58 satisfies , @xmath59 subject to : @xmath60 the functions @xmath61 , @xmath62 , are given by @xmath63 the solvency region in the new coordinates is given by : @xmath64 where @xmath65 this formulation has several advantages over the formulation of @xcite .
as in @xcite , the problem is one dimensional ( ( [ ch3enunprobpolar])-([ch3enunprobpolar2 ] ) do not depend of @xmath66 ) , but in our case the domain is bounded ( @xmath67 ) .
furthermore , the operators involved in ( [ ch3enunprobpolar ] ) are linear in @xmath68 , whereas in @xcite , the equations contains a nonlinear term .
next , we characterize the buying and selling regions in terms of the polar coordinates .
first , let us observe that @xmath69 where @xmath46 is the function defined in ( [ checamborig ] ) .
let us define the functions @xmath70 and @xmath71 by @xmath72,\ ] ] where @xmath73 and @xmath74 are the boundaries of the buying and selling regions in cartesian coordinates defined in ( [ ch3funfronc ] ) .
the following proposition is an immediate consequence of the results in @xcite .
[ ch3tradpropertapolar ] functions @xmath71 , @xmath70 are monotonically decreasing functions .
it holds that @xmath75 and that @xmath76,\quad \hat{t}_0=t-\frac{1}{\alpha - r}\log\frac{1+\lambda}{1-\mu}.\ ] ] if @xmath77 , then @xmath78 , with @xmath79 it holds that @xmath80 , where @xmath81 is the merton line .
if @xmath82 , there exist two values @xmath83 , such that @xmath84 the limit values @xmath85 and @xmath86 are defined by @xmath87 where @xmath88 and @xmath89 are the constants defined in ( * ? ?
* theorem 6.1 ) .
the functions @xmath71 , @xmath70 satisfy ( see also ( * ? ? ?
* proposition 3.4.2 ) ) , @xmath90.\ ] ] it is now easy to see that , for @xmath91 $ ] , the buying , selling and no transaction region can be described by as follows : * 1 . * the buying region
is defined by @xmath92 $ ] . in @xmath93
the value function satisfies @xmath94 * 2 . * the selling region
is defined by @xmath95 . in @xmath96
the value function satisfies @xmath97 * 3 . * the no transaction region
is defined by @xmath98 . in @xmath99 ,
@xmath68 satisfies of the following partial differential equations @xmath100 we remark that if the buying ( @xmath101 ) and selling ( @xmath102 ) frontiers are known , we can compute the value function @xmath58 in @xmath93 and @xmath96 explicitly by a simple integration of equations ( [ chp3brexpformula ] ) and ( [ chp3srexpformula ] ) respectively . for @xmath103 , we have @xmath104 and for @xmath105 , @xmath106 value function ( numerical solution ) for @xmath107\times[0,30]$ ] .
the colour code is blue if @xmath108 is in the buying region , green in the no transactions region and red in the selling region.,width=453,height=188 ] figure [ ch3stationarycn2 ] represents the value function in perspective ( left ) and from above ( right ) in @xmath109 $ ] , for a maturity of @xmath110 years .
the figure shows the numerical values obtained for the function @xmath58 with the method described in section [ ch3 nm ] .
we have coloured the function depending in whether @xmath108 is in the buying , selling or no transactions region .
we can visually check the expected monotonicity of the buying and selling frontiers studying from above ( right ) the two curves which divide the different colours ( red - green and green - blue ) .
the buying frontier remains constant ( @xmath111 ) for a certain period near maturity and the stationarity value of both frontiers as we move away from maturity is also observable .
the numerical method described in this section is constructed upon the following strategy .
let @xmath112 denote an admissible trading strategy where @xmath113 and @xmath114 are the amount of money in the bank and stock accounts at @xmath115 .
let @xmath116 and @xmath117 and define @xmath118 where @xmath119 , @xmath120 are the amounts in the bank and stock accounts if strategy @xmath121 is followed .
proposition [ ch3tradpropertapolar ] implies that @xmath122 where @xmath123 denotes the optimal trading strategy solving ( [ ch3defvarpphi ] ) .
therefore , the optimal value function can be computed as the solution of ( [ ch3enunprobpolar])-([ch3enunprobpolar2 ] ) in @xmath124 $ ] subject to the boundary conditions : @xmath125 these conditions are equivalent to a mandatory buying or selling the stock if @xmath126 reaches @xmath127 or @xmath128 respectively ( see formulas ( [ chp3brexpformula])-([chp3srexpformula ] ) ) .
the solution can be extended to @xmath129 $ ] taking into account that for @xmath130 $ ] : @xmath131 and @xmath132 , so that we can compute @xmath58 with ( [ ch3exfsrbr1 ] ) in @xmath93 and ( [ ch3exfsrbr2 ] ) in @xmath96 .
let @xmath133 be a nonnegative integer and let us define the time mesh @xmath134 by @xmath135 the spatial mesh will depend on the time step .
the main idea is to adapt the mesh in such a way that it evolves through time following the approximate location of the buying and selling frontiers ( i.e. evolving as the green zone in figure [ ch3stationarycn2 ] ) . to this end , let @xmath136 be a control parameter .
we define @xmath137 where @xmath138 $ ] is the stationary state of the selling frontier ( see proposition [ ch3tradpropertapolar ] ) . for @xmath139 ,
let us consider the @xmath140 chebyshev nodes in @xmath141 $ ] , @xmath142 we define the integer @xmath143 as the unique integer such that @xmath144 note that @xmath145 is well defined because @xmath146 . from the definition of the chebyshev nodes
, it is easy to check that it exists @xmath147 such that for all @xmath148 , @xmath149 .
let us suppose that at time @xmath150 , we know the locations of the buying and selling frontiers , @xmath151 and @xmath152 . given @xmath153 , @xmath154
we define an interval @xmath155 by : @xmath156,\ ] ] if @xmath157 , or , in case @xmath158 , @xmath159.\ ] ] here @xmath160 .
we remark that , with this definition , the interval @xmath155 always contains the no transaction region @xmath161 $ ] and it is contained in the solvency region @xmath162 $ ] .
furthermore , @xmath163 and @xmath164 are always one of the @xmath165 chebyshev nodes in the interval @xmath155 , while restriction @xmath154 implies that @xmath164 is an interior point of @xmath155 .
the control parameter @xmath166 guaranties that a maximum of @xmath167% of interval @xmath155 is contained in the selling region , another maximum of @xmath167% of @xmath155 in the buying region , whereas a minimum @xmath168% of @xmath155 is contained in the no transactions region . note also that for @xmath169 , the values @xmath170 and @xmath171 are , of course , known data ( see proposition [ ch3tradpropertapolar ] ) , whereas for @xmath172 , we have to substitute @xmath151 and @xmath152 by some approximation that we will denote @xmath173 and @xmath174 where @xmath175 .
we will describe in subsection [ ch3maccmcc ] how to compute them prior to the construction of the interval @xmath155 .
next proposition proves that , for @xmath133 big enough , @xmath176 , so that we can compute recursively the intervals @xmath177 for @xmath178 .
[ ch3propoinclusnteninter ] let @xmath179 $ ] where @xmath180 and @xmath181 are the exact location of the buying and selling frontiers .
for any @xmath154 , where @xmath147 is the restriction which guarantees that @xmath102 will be in the interior of @xmath182 , compute @xmath182 with ( [ ch3forint1 ] ) or ( [ ch3forint2 ] ) .
it exists @xmath183 such that for any time mesh @xmath184 given by ( [ ch3timediscret ] ) , it holds @xmath185 for any @xmath186 . from @xcite , we know that @xmath187
. therefore , @xmath188 is @xmath189 .
let @xmath89 from ( [ ch3paramk ] ) be fixed . since @xmath102 is in the interior of @xmath182 , it will exist @xmath190 such that for all @xmath191 : @xmath192 this guarantees that for any equally spaced time mesh @xmath193 , with @xmath194 , @xmath195 . to finish the proof , note that from proposition [ ch3tradpropertapolar ] , @xmath196 and that @xmath197 , so the result follows directly from the definition of @xmath155 .
let us suppose that we know an approximation of the function value @xmath198 , @xmath199 and approximate values of @xmath200 and @xmath201 at time @xmath150 .
for @xmath202 big enough ( proposition [ ch3propoinclusnteninter ] ) , we can compute @xmath203 $ ] defined as in ( [ ch3forint1 ] ) if @xmath204 , or with ( [ ch3forint2 ] ) otherwise . for @xmath205 $ ]
, we define the function @xmath206 as the function value which gives the expected terminal value when the trading strategy is to perform no transactions if @xmath207 , to buy the stock if @xmath208 and to sell the stock if @xmath209 , subject to @xmath210 .
therefore , @xmath206 is the solution of the equation @xmath211 subject to @xmath212 let us consider the @xmath165 chebyshev nodes in @xmath155 @xmath213 where @xmath214 are the chebyshev points ( [ ch3chebynodes ] ) .
the numerical approximation @xmath215 , @xmath216 to the function @xmath206 is the collocation polynomial @xcite of degree @xmath202 defined for @xmath217 by : @xmath218 subject to @xmath219 with ( neumann ) boundary conditions @xmath220 where @xmath221 the equations ( [ ch3collocationmethod])-([ch3collocationmethodcons ] ) define a dense system of linear equations to find the values of @xmath222 , @xmath223 .
however , the fact that , with relative few nodes for the spatial mesh we can achieve a very good precision , makes this method competitive with respect to a finite differences method , see section [ ch3numresults ] .
let us define @xmath224 which are explicit functions because @xmath225 is a known polynomial in @xmath126 . in ( [ ch3auxencheby ] ) ,
we compare ( see ( * ? ? ?
* subsection 3.5.3 ) ) whether it is better to not perform transactions or to buy the stock ( resp .
sell the stock ) .
if polynomial @xmath226 ( resp .
@xmath227 ) it is better to not perform transactions rather than buy ( resp .
sell ) the stock .
the numerical approximation to the buying and selling frontiers is given by : @xmath228 \right\ } , \\
\end{aligned}\ ] ] once we know the location of the frontiers and the function value in that points , we can compute the approximate function value through the following explicit formulas where we have used the notation @xmath229 , @xmath230 @xmath231^{\gamma},\ ] ] if @xmath232 , @xmath233 @xmath234^{\gamma},\ ] ] if @xmath235 .
then the complete algorithm reads as follows : * fix a number @xmath133 and a number @xmath202 big enough such that proposition [ ch3propoinclusnteninter ] holds .
+ compute @xmath236 and @xmath193 as in ( [ ch3timediscret ] ) . define @xmath175 . + set @xmath237 and compute @xmath238 with formula ( [ ch3forint1 ] ) + compute @xmath239 , @xmath240 , as the chebyshev interpolation polynomial in @xmath241 of function @xmath242 , given by ( [ ch3enunprobpolar2 ] ) , where @xmath241 denote the chebyshev nodes in @xmath238 .
* compute the polynomial @xmath215 solving the collocation equations ( [ ch3collocationmethod ] ) with final condition ( [ ch3valorcondvencheby ] ) and boundary conditions ( [ ch3collocationmethodcons ] ) . *
locate the buying and selling frontiers @xmath243 and @xmath244 using ( [ frontiers ] ) . *
compute the interval @xmath245 with ( [ ch3forint1 ] ) if @xmath246 or with ( [ ch3forint2 ] ) otherwise .
+ compute the numerical approximation @xmath247 at time @xmath248 with formulae ( [ vn1 ] ) , ( [ vn2 ] ) and ( [ vn3 ] ) . *
set @xmath249 and stop if @xmath250 or , otherwise , proceed to * step 1*.
we consider the parameter values as in the first experiment in @xcite . for
@xmath251 $ ] let : @xmath252 the following figure shows the numerical solution @xmath253 .
[ ch3v1valfun4anos ] for @xmath254}$ ] .
the colour code is blue if @xmath108 is in the buying , green in the no transactions and red in the selling region.,title="fig:",width=453,height=188 ] we have colored the function depending in whether @xmath108 is in the buying , selling or no transactions region . as in figure
[ ch3stationarycn2 ] , we can visually check the properties from proposition [ ch3tradpropertapolar ] .
first , we establish the criteria employed in the experiments to build the spatial mesh .
we have fixed the control parameter @xmath255 , so that , at least @xmath256% of the interval corresponds to the no transactions region .
the particular choice of @xmath166 does not affect the rate of convergence of the error . in order to compare the performance of the spectral method with other numerical methods ,
we have also implemented a central differences ( cd ) based method in order to solve the pde in step 2 ( see subsection [ ch3maccmcc ] ) .
the formal study of the error will be conducted for the cases where explicit formulas are available , comparing the results of the central differences and chebyshev methods .
the rest of the properties given in @xcite , although not included , were also checked .
we consider @xmath46 defined in ( [ checamborig ] ) . for @xmath258 , we can explicitly compute @xmath257 with ( * ? ? ?
* ( 3.9 ) ) . in figure [ ch3analiticalsolv0 t ]
we plot the value of @xmath257 for @xmath259 $ ] . analytical solution of @xmath260$].,width=245,height=188 ] the value @xmath258 corresponds in polar coordinates to @xmath261 .
a numerical solution @xmath262 can be computed explicitly using @xmath263 and formula ( [ ch3formularelacionpolarorig ] ) , which relates the function in polar coordinates and in the original variables . @xmath264
the following figure compares the difference between the analytical solution @xmath260 $ ] and the numerical solution obtained with the chebyshev method for @xmath265 ( left ) and @xmath266 ( right ) .
both pictures are in the same scale and we can observe that the error reduces for increasing value of @xmath202 .
value @xmath267 $ ] where @xmath268 was computed with the chebyshev method with @xmath265 ( left ) and @xmath266 ( right).,width=529,height=207 ] in both pictures of figure [ ch3valpimedcomcheb ] we can observe an error discontinuity at time @xmath269 . from ( * ? ? ?
* ( 3.9 ) ) , we know that the function @xmath257 is not derivable ( respect time ) at instant @xmath269 .
the same phenomena can be observed in the numerical experiments in @xcite .
we can also see that some oscillations appear at time @xmath270 where we change the kind of adaptive mesh @xmath271 ( see subsection [ ch3maccmam ] ) .
we proceed to check the rate of error convergence .
we define the root of the mean square error as @xmath272 figure [ ch3converrtesptempvalpimed ] shows the convergence of spatial error ( left ) for @xmath273 and different number of spatial nodes @xmath202 .
the right side shows the convergence of temporal error for @xmath202 fixed and different values of @xmath133 . spatial ( left ) and temporal ( right ) error convergence of @xmath268 in logarithmic scale of the central differences ( blue ) and chebyshev ( red ) methods.,width=472,height=200 ] in the left side of figure [ ch3converrtesptempvalpimed ]
we have plotted , in logarithmic scale , the number @xmath202 of spatial nodes versus the value @xmath274 .
the slope of the regression line of the cd method ( plotted in blue ) is @xmath275 and of the chebyshev method ( plotted in red ) is @xmath276 .
the spectral convergence that we could expect in the chebyshev method does not occur due to the regularity of the problem . in the right hand side of figure [ ch3converrtesptempvalpimed ] , we have plotted , in logarithmic scale , the number @xmath277 of time steps versus the value @xmath274 .
the slope of the regression line of the cd method ( solid - blue ) is @xmath278 as it could be expected from an order 2 method .
the slope of the chebyshev method ( solid - red ) is @xmath279 .
we note that for large values of @xmath133 we reach very soon the error limit marked by the size of @xmath202 .
we carry out a second experiment doubling the value of @xmath202 ( right - dashed - blue / red ) to check that the lowest value reached by the temporal error was given by the size of the spatial mesh . depending on the error tolerance
, we might need a big value for @xmath202 in the cd method but much smaller in the chebyshev method .
this makes that , depending on the required precision , chebyshev performs better in computational cost than cd .
this will be studied below . from proposition [ ch3tradpropertapolar ] , we know that in polar coordinates @xmath78 . given
a number of time steps @xmath133 , we look for @xmath280 which is nearest to @xmath269 and define the absolute error ( just for this experiment ) as : @xmath281 the next figure shows the convergence of spatial error ( left ) for @xmath273 and different number of spatial nodes @xmath202 . the right side shows the convergence of temporal error for @xmath266 ( chebyshev ) and @xmath282 ( cd ) and different values of @xmath133 .
spatial ( left ) and temporal ( right ) error ( semilogarithmic scale ) of instant when @xmath283 with the cd ( blue ) and chebyshev ( red ) methods.,width=472,height=200 ] the spatial error ( left ) reduces as we increase the value of @xmath202 . at equal number of nodes ,
the chebyshev method gives much smaller errors than the cd method . concerning the temporal error , the results are step shaped because of the definition of absolute error and the time partition when
@xmath284 is halved .
each time partition is included in the following one and @xmath285 sometimes changes and sometimes not .
the temporal error reduces as we increase the value of @xmath277 . as in the spatial error ,
the chebyshev method outperforms the cd method . from proposition [ ch3tradpropertapolar ] , we know that @xmath286 where @xmath270 is explicitly computable . given a number of time steps @xmath133 ,
we look for @xmath186 such that @xmath287 for the chebyshev method , the @xmath288 may be bigger than 0 a few time steps prior to @xmath289 .
we note that in the chebyshev method , the lower limit of @xmath290 $ ] is the buying frontier . in left picture of figure [ ch3v1buyingfrontiercheby ]
, we have plotted the numerical estimation of the buying frontier with the chebyshev method for , @xmath265 ( blue ) , @xmath291 ( red ) , @xmath292 ( green ) @xmath266 ( black ) . in the right picture
we zoom around @xmath270 .
numerically computed buying frontier with the chebyshev method for @xmath259 $ ] ( left ) and zoom around @xmath270 ( right).,width=491,height=207 ] let @xmath293 be the biggest value such that @xmath294 if @xmath295 , the location of the buying frontier oscillates around 0 for @xmath296 and for @xmath297 when it behaves as we could expect from proposition [ ch3tradpropertapolar ] .
numerical experiments show that it is better to let @xmath200 oscillate around 0 rather than imposing @xmath298 .
the oscillations observed in figure [ ch3v1buyingfrontiercheby ] are generated by the imposition of the neumann conditions .
the boundary error is controlled by @xmath133 and @xmath202 , but the spatial error is dominant in this experiment .
the instant when the numerical solution begins to oscillate is always very close to @xmath270 @xmath299 and the size of the oscillations reduces as @xmath202 increases .
these oscillations are the error that we are going to study .
they include all the negative values ( since the buying frontier must be always positive ) and any positive value for discrete times larger than @xmath270 .
thus , we define , for this method and experiment , the absolute error ( ae ) as @xmath300 we fix @xmath273 and compute the absolute error for several values for @xmath202 ) . in figure [ ch3v1convespfronnocerocheby ]
we plot , in logarithmic scale , the value of @xmath202 versus the absolute error . as we can see the error
is rapidly reduced by increasing @xmath301 .
spatial error convergence of the first instant when it is optimal to have a positive amount of stock ( chebyshev method).,width=226,height=188 ] @xmath302 and @xmath303 tend to a stationary state as @xmath304 that can also be computed explicitly ( see proposition [ ch3tradpropertapolar ] ) . computed with the same model parameters as before but for @xmath110 years ( see figure [ ch3stationarycn2 ] ) ,
frontiers have stabilized a few years before reaching @xmath115 at : @xmath305 computed with the chebyshev method ( @xmath306 , @xmath291 ) .
we define the absolute error ( for this experiment ) as @xmath307 we study the spatial ( @xmath308 and several values for @xmath202 ) and temporal ( @xmath309 for the cd , @xmath291 for the chebyshev method , and several values for @xmath133 ) error convergence .
in figure [ ch3v1estacconv ] we plot , in logarithmic scale , the value of @xmath301 ( left ) versus the absolute value of the error and the value of @xmath133 ( right ) versus the absolute value of the error for both methods .
spatial ( left ) and temporal ( right ) error convergence , in logarithmic scale , of the stationary state of the buying frontier for the cd(blue ) and chebyshev ( red ) methods.,width=510,height=226 ] in this experiment , temporal error is dominant compared with respect to the spatial error in the chebyshev method . in the case of the cd method , the error depends more in both the spatial and temporal discretizations . on the left side picture
, we can see that the chebyshev method reaches the error marked by the time discretization with the smallest number of nodes therefore , if a high precision is required , chebyshev will perform better than the central differences method .
the error behaviour of the selling frontier is similar to the one of the buying frontier . in this section
we compare the relative performance of the pseudospectral and finite difference methods .
first of all , we fix several time and spatial discretization parameters : a. @xmath310 $ ] b. @xmath311 $ ] ( chebyshev ) c. @xmath312 $ ] ( central differences ) and solve the problem with all the combinations of the different discretizations for both methods .
the lower and upper bounds of @xmath202 in the central differences method can be taken smaller or bigger .
the criteria that we have employed is such that the numerical error varies between @xmath313 and @xmath314 .
the same reads for the upper bound of @xmath202 in the chebyshev method .
we point that during the implementation of the method , we observed that if @xmath202 was not big enough , the location of the frontiers may oscillate ( due to the gibbs effect or to the fact that the polynomials are not accurate enough ) , complicating the location of @xmath315 and @xmath316 in ( [ frontiers ] ) .
the chebyshev spectral method is effective once enough resolution has been reached .
this behaviour is typical of high order methods , see @xcite .
the employment of the adaptive interval @xmath155 and an enough amount of interpolation nodes avoids the oscillations and allows to obtain just one numerical approximation of @xmath315 and @xmath316 in ( [ frontiers ] ) .
the oscillations may appear if the following ( empirical ) bounds are violated @xmath317 where @xmath318 and numerical experiments suggest that @xmath319 might be a growing function of @xmath202 .
the lowest value of @xmath202 in the chebyshev method was chosen so that no oscillations appear . if a smaller number of interpolation nodes is chosen , the solution oscillates and the error worsens .
we plot the value of @xmath320 ( [ ch3errrorcuadmed ] ) versus the computational time employed in computing @xmath268 for each different spatial and temporal meshes in logarithmic scale .
performance comparison of the error at @xmath321 . in logarithmic scale , we plot ( left ) , the value of rmse versus the total computational costs of cd ( blue ) and chebyshev ( red ) methods and their respective lower enveloping curves ( right).,width=510,height=226 ] the left - side picture of figure [ ch3v1performanceerror ] represents the cloud of results for the different discretizations of each method . the right - side , which is more visual , represents the lower convex enveloping curve . with the right - side picture
, we can obtain an approximate behaviour of the evolution of the error versus the required computational time to reach that precision .
we fix the error tolerance that we require for our problem and find which method and spatial and time discretization reaches it first . as we can see , the cd method ( blue in figure [ ch3v1performanceerror ] ) performs better if we do not require a high precision .
if a higher precision is required , chebyshev ( red in figure [ ch3v1performanceerror ] ) performs better than cd . a similar behaviour can be observed if we compare the errors of the rest of cases where we have explicit formulas .
the homothetic property of the potential utility function has been used to restate the investment problem in polar coordinates .
this has allowed us to give an equivalent formulation of the problem in a bounded spatial domain .
although some of the numerical difficulties that appear with the parabolic double obstacle problem are avoided , other problems may appear if we employ spectral methods .
the gibbs effect , which comes from the fact that the objective function is continuous but not differentiable at maturity , can complicate the location of the frontiers , but this issue can be circumvented by the employment of a time - adapted spatial mesh ( subsection [ ch3maccmam ] ) . further work may include the extension of the model including a consumption term or the design of spectral methods to optimal investment problems with other utility functions , like the exponential utility . furthermore , through the indifference pricing technique ( see @xcite and @xcite ) , these kind of models can be applied to option valuation .
chiarella , c. , el - hassan , n. and a. kucera , a _ evaluation of american option prices in a path integral framework using fourier - hermite series expansion _
, journal of economic dynamics and control 23 ( 1999 ) , 1387 - 1424 . | this paper concerns the numerical solution of the finite - horizon optimal investment problem with transaction costs under potential utility .
the problem is initially posed in terms of an evolutive hjb equation with gradient constraints . in @xcite , the problem is reformulated as a non - linear parabolic double obstacle problem posed in one spatial variable and defined in an unbounded domain where several explicit properties and formulas are obtained .
the restatement of the problem in polar coordinates allows to pose the problem in one spatial variable in a finite domain , avoiding some of the technical difficulties of the numerical solution of the previous statement of the problem .
if high precision is required , the spectral numerical method proposed becomes more efficient than simpler methods as finite differences for example .
* keywords : * optimal investment , potential utility , transaction costs , spectral method . |
one - dimensional _ analog signals _ are complex valued functions on the real line @xmath0 . in the same spirit ,
one - dimensional _ digital signals , _ also called _ sequences , _ might be considered as complex valued functions on the finite line @xmath1 , i.e. , the finite field with @xmath2 elements . in both situations the parameter of the line is denoted by @xmath3 and is referred to as _
_ time__@xmath4 _ _ _ _ in this work , we will consider digital signals only , which will be simply referred to as signals .
the space of signals @xmath5 is a hilbert space with the hermitian product given by @xmath6 a central problem is to construct interesting and useful systems of signals . given a system @xmath7 , there are various desired properties which appear in the engineering wish list . for example , in various situations 1 , 2 one requires that the signals will be weakly correlated , i.e. , that for every @xmath8 @xmath9this property is trivially satisfied if @xmath7 is an orthonormal basis .
such a system can not consist of more than @xmath10 signals , however , for certain applications , e.g. , cdma ( code division multiple access ) @xcite a larger number of signals is desired , in that case the orthogonality condition is relaxed . during the transmission process
, a signal @xmath11 might be distorted in various ways .
two basic types of distortions are _ time shift _ @xmath12 and _ phase shift _ @xmath13 , where @xmath14 .
the first type appears in asynchronous communication and the second type is a doppler effect due to relative velocity between the transmitting and receiving antennas . in conclusion ,
a general distortion is of the type @xmath15 suggesting that for every @xmath16 it is natural to require @xcite the following stronger condition @xmath17due to technical restrictions in the transmission process , signals are sometimes required to admit low peak - to - average power ratio @xcite , i.e. , that for every @xmath18 with @xmath19@xmath20finally , several schemes for digital communication require that the above properties will continue to hold also if we replace signals from @xmath21 by their fourier transform@xmath4 in this paper we construct a novel system of ( unit ) signals @xmath22 , consisting of @xmath23order of @xmath24 signals , where @xmath2 is an odd prime , called the _
oscillator system_. these signals constitute , in an appropriate formal sense , a finite analogue for the eigenfunctions of the harmonic oscillator in the real setting and , in accordance , they share many of the nice properties of the latter class . in particular , the system @xmath25 satisfies the following properties 1 .
_ autocorrelation ( ambiguity function ) .
_ for every @xmath26 we have @xmath27 2 .
_ crosscorrelation ( cross - ambiguity function ) .
_ for every @xmath28 we have @xmath29for every @xmath14 .
_ supremum .
_ for every signal @xmath30 we have@xmath31 4 . _
fourier invariance .
_ for every signal @xmath32 its fourier transform @xmath33 is ( up to multiplication by a unitary scalar ) also in @xmath34 explicit algorithm that generate the ( split ) oscillator system is given in the supporting text .
the oscillator system can be extended to a much larger system @xmath35 , consisting of order of @xmath36 signals if one is willing to compromise properties 1 and 2 for a weaker condition .
the extended system consists of all signals of the form @xmath37 for @xmath14 and @xmath30 .
it is not hard to show that @xmath38 @xmath39 . as a consequence of ( cross_eq ) and ( [ auto_eq ] ) for every @xmath40 we have @xmath41 the characterization and construction of the oscillator system is representation theoretic and we devote the rest of the paper to an intuitive explanation of the main underlying ideas . as a suggestive model example
we explain first the construction of the well known system of chirp ( heisenberg ) signals , deliberately taking a representation theoretic point of view ( see @xcite for a more comprehensive treatment ) .
let us denote by @xmath42 the character @xmath43 .
we consider the pair of orthonormal bases @xmath44 and @xmath45 , where @xmath46 and @xmath47 is the kronecker delta function , @xmath48 if @xmath49 and @xmath50 if @xmath51 let @xmath54 be the time shift operator @xmath55 .
this operator is unitary and it induces a homomorphism of groups @xmath56 given by @xmath57 for any @xmath58 .
elements of the basis @xmath53 are character vectors with respect to the action @xmath59 , i.e. , @xmath60 for any @xmath58 . in the same fashion ,
the basis @xmath52 consists of character vectors with respect to the homomorphism @xmath61 given by the phase shift operators @xmath62 .
the homomorphisms @xmath59 and @xmath63 can be combined into a single map @xmath64 which sends a pair @xmath65 to the unitary operator @xmath66 .
the plane @xmath67 is called the _ time - frequency plane _ and will be denoted by @xmath68 .
the map @xmath69 is not an homomorphism since , in general , the operators @xmath70 and @xmath71 do not commute . this deficiency can be corrected if we consider the group @xmath72 with multiplication given by @xmath73the map @xmath69 extends to a homomorphism @xmath74 given by @xmath75the group @xmath76 is called the _ heisenberg _ group and the homomorphism @xmath77 is called the _ _ heisenberg representation__@xmath4 the heisenberg group is no longer commutative , however , it contains various commutative subgroups which can be easily described . to every line
@xmath78 that pass through the origin , one can associate a maximal commutative subgroup @xmath79
. it will be convenient to identify the subgroup @xmath80 with the line @xmath81 . restricting the heisenberg representation @xmath82 to a subgroup @xmath81
yields a decomposition of the hilbert space @xmath83 into a direct sum of one - dimensional subspaces @xmath84 where @xmath85 runs in the set @xmath86 of ( complex valued ) characters of the group @xmath81 .
the subspace @xmath87 consists of vectors @xmath88 such that @xmath89 . in other words ,
the space @xmath87 consists of common eigenvectors with respect to the commutative system of unitary operators @xmath90 such that the operator @xmath91 has eigenvalue @xmath92 . choosing a unit vector @xmath93for every @xmath94 we obtain an orthonormal basis @xmath95 .
in particular , @xmath53 and @xmath52 are recovered as the bases associated with the lines @xmath96 and @xmath97 respectively . for a general @xmath98
the signals in @xmath99 are certain kind of chirps .
concluding , we associated with every line @xmath100 an orthonormal basis @xmath101 and overall we constructed a system of signals consisting of a union of orthonormal bases @xmath102for obvious reasons , the system @xmath103 will be called the _ heisenberg _ system__. _ _ it will be convenient to introduce the following general notion . given two signals @xmath104 , their matrix coefficient is the function @xmath105 given by @xmath106 . in coordinates , if we write @xmath107 then @xmath108 . when @xmath109 the function @xmath110 is called the _ ambiguity _ function of the vector @xmath11 and is denoted by @xmath111 .
the system @xmath103 consists of @xmath112 orthonormal bases is the number of lines in @xmath68 .
] , altogether @xmath113 signals and it satisfies the following properties @xcite 1 . _ _ autocorrelation__**. * * for every signal @xmath114 the function @xmath115 is the characteristic function of the line @xmath81 , i.e. , @xmath116 2 . _
crosscorrelation_. * * * * for every @xmath117 and @xmath118 where @xmath119 we have@xmath120for every @xmath121 . if @xmath122 then @xmath123 is the characteristic function of some translation of the line @xmath81 .
_ supremum__**. * * a signal @xmath124 is a unimodular function , i.e. , @xmath125 for every @xmath126 , in particular we have @xmath127 note the main differences between the heisenberg and the oscillator systems .
the oscillator system consists of order of @xmath24 signals , while the heisenberg system consists of order of @xmath128 signals .
signals in the oscillator system admits an ambiguity function concentrated at @xmath129 ( thumbtack pattern ) while signals in the heisenberg system admits ambiguity function concentrated on a line .
reflecting back on the heisenberg system we see that each vector @xmath130 is characterized in terms of action of the additive group @xmath131 .
roughly , in comparison , each vector in the oscillator system is characterized in terms of action of the multiplicative group @xmath132 .
our next goal is to explain the last assertion .
we begin by giving a model example . given a multiplicative character which satisfies @xmath133 for every @xmath134 @xmath135
, we define a vector @xmath136 by @xmath137we consider the system @xmath138 , where @xmath139 is the dual group of characters . for each element @xmath141
let @xmath142 be the unitary operator acting by scaling @xmath143 .
this collection of operators form a homomorphism @xmath144 .
elements of @xmath140 are character vectors with respect to @xmath145 , i.e. , the vector@xmath146 satisfies @xmath147 for every @xmath141 . in more conceptual terms , the action @xmath145 yields a decomposition of the hilbert space @xmath83 into character spaces @xmath148 , where @xmath85 runs in the group @xmath139 .
the system @xmath140 consists of a representative unit vector for each space @xmath87 , @xmath149 .
we would like to generalize the system @xmath140 in a similar fashion like we generalized the bases @xmath52 and @xmath53 in the heisenberg setting . in order to do this
we need to introduce several auxiliary operators .
let @xmath150 , @xmath151 be the operators acting by @xmath152 ( scaling ) , where @xmath153 is the unique quadratic character of @xmath154 , let @xmath155 be the operator acting by @xmath156 ( quadratic modulation ) , and finally let @xmath157 be the operator of fourier transform @xmath158where @xmath159 is a normalization constant @xcite .
the operators @xmath160 and @xmath161 are unitary .
let us consider the subgroup of unitary operators generated by @xmath162 and @xmath163 .
this group turns out to be isomorphic to the finite group @xmath164 , therefore we obtained a homomorphism @xmath165 . the representation @xmath145 is called the _
weil representation _ @xcite and it will play a prominent role in this paper .
the group @xmath166 consists of various types of commutative subgroups
. we will be interested in maximal _
diagonalizable _ commutative subgroups .
a subgroup of this type is called maximal _ split _ _ torus . _
the standard example is the subgroup consisting of all diagonal matrices @xmath167which is called the _
standard torus_. the restriction of the weil representation to a split torus @xmath168 yields a decomposition of the hilbert space @xmath83 into a direct sum of character spaces @xmath169 , where @xmath85 runs in the set of characters @xmath170 .
choosing a unit vector @xmath171 for every @xmath85 we obtain a collection of orthonormal vectors @xmath172 .
overall , we constructed a system @xmath173which will be referred to as the _ split oscillator system .
we note that _ our initial system @xmath140 is recovered as @xmath174 . from the point of view of this paper ,
the most interesting maximal commutative subgroups in @xmath166 are those which are diagonalizable over an extension field rather than over the base field @xmath1 .
a subgroup of this type is called maximal _ non - split torus .
_ it might be suggestive to first explain the analogue notion in the more familiar setting of the field @xmath0 . here
, the standard example of a maximal non - split torus is the circle group @xmath175 .
indeed , it is a maximal commutative subgroup which becomes diagonalizable when considered over the extension field @xmath176 of complex numbers .
the above analogy suggests a way to construct examples of maximal non - split tori in the finite field setting as well .
let us assume for simplicity that @xmath177 does not admit a square root in @xmath1 .
the group @xmath166 acts naturally on the plane @xmath178 .
consider the symmetric bilinear form @xmath179 on @xmath68 given by @xmath180 an example of maximal non - split torus is the subgroup @xmath181 consisting of all elements @xmath182 preserving the form @xmath179 , i.e. , @xmath183 if and only if @xmath184 for every @xmath185 . in the same fashion like in the split case , restricting the weil representation to a non - split torus @xmath186 yields a decomposition into character spaces @xmath187 .
choosing a unit vector @xmath188 for every @xmath189 we obtain an orthonormal basis @xmath190 .
overall , we constructed a system of signals @xmath191the system @xmath192 will be referred to as the _ non - split oscillator _
system__. _ _ the construction of the system @xmath193 @xmath194 together with the formulation of some of its properties are the main contribution of this paper .
the oscillator system is closed under the operation of fourier transform , i.e. , for every @xmath30 we have @xmath195 the fourier transform on the space @xmath196 appears as a specific operator @xmath197 in the weil representation , where @xmath198given a signal @xmath199 , its fourier transform @xmath200 is , up to a unitary scalar , a signal in @xmath201 where @xmath202 .
in fact , @xmath203 is closed under all the operators in the weil representation ! indeed , given an element @xmath182 and a signal @xmath204 we have , up to a unitary scalar , that @xmath205 @xmath206 , where @xmath207 .
in addition , the weyl element @xmath208 is an element in some maximal torus @xmath209 ( the split type of @xmath209 depends on the characteristic @xmath2 of the field ) and as a result signals @xmath210 are , in particular , eigenvectors of the fourier transform . as a consequences a signal @xmath211 and
its fourier transform @xmath33 differ by a unitary constant , therefore are practically the `` same '' for all essential matters .
these properties might be relevant for applications to ofdm ( orthogonal frequency division multiplexing ) @xcite where one requires good properties both from the signal and its fourier transform .
here we give the explanation why functions in the non - split oscillator system @xmath192 constitute a finite analogue of the eigenfunctions of the harmonic oscillator in the real setting .
the weil representation establishes the dictionary between these two , seemingly , unrelated objects .
the argument works as follows .
the one - dimensional harmonic oscillator is given by the differential operator @xmath212 .
the operator @xmath213 can be exponentiated to give a unitary representation of the circle group @xmath214 where @xmath215 .
eigenfunctions of @xmath213 are naturally identified with character vectors with respect to @xmath145 .
the crucial point is that @xmath145 is the restriction of the weil representation of @xmath216 to the maximal non - split torus @xmath217 . summarizing ,
the eigenfunctions of the harmonic oscillator and functions in @xmath218 are governed by the same mechanism , namely both are character vectors with respect to the restriction of the weil representation to a maximal non - split torus in @xmath219 .
the only difference appears to be the field of definition , which for the harmonic oscillator is the reals and for the oscillator functions is the finite field .
two applications of the oscillator system will be described .
the first application is to the theory of discrete radar . the second application is to cdma systems
. we will give a brief explanation of these problems , while emphasizing the relation to the heisenberg representation .
the theory of discrete radar is closely related @xcite to the finite heisenberg group @xmath220 a radar sends a signal @xmath221 and obtains an echo @xmath222 .
the goal @xcite is to reconstruct , in maximal accuracy , the target range and velocity . the signal @xmath221 and the echo @xmath222 are , principally , related by the transformation@xmath223where the time shift @xmath224 encodes the distance of the target from the radar and the phase shift encodes the velocity of the target .
equivalently saying , the transmitted signal @xmath11 and the received echo @xmath225 are related by an action of an element @xmath226 , i.e. , @xmath227 the problem of discrete radar can be described as follows . given a signal @xmath11 and an echo @xmath228 extract the value of @xmath229**. * * it is easy to show that @xmath230 and it obtains its maximum at @xmath231 .
this suggests that a desired signal @xmath11 for discrete radar should admit an ambiguity function @xmath232 which is highly concentrated around @xmath233 , which is a property satisfied by signals in the oscillator system ( property 2 ) .
* there exists a collection of users @xmath234 , each holding a _ bit _ of information @xmath235 ( usually @xmath236 is taken to be an @xmath237th root of unity ) . *
each user transmits his bit of information , say , to a central antenna . in order to do that
, he multiplies his bit @xmath236 by a private signal @xmath238 and forms a message @xmath239 . *
the transmission is carried through a single channel ( for example in the case of cellular communication the channel is the atmosphere ) , therefore the message received by the antenna is the sum@xmath240 the main problem @xcite is to extract the individual bits @xmath236 from the message @xmath241 .
the bit @xmath236 can be estimated by calculating the inner product@xmath242the last expression above should be considered as a sum of the information bit @xmath236 and an additional noise caused by the interference _ _ _ _ of the other messages .
this is the standard scenario also called the _ synchronous _ scenario . in practice ,
more complicated scenarios appear , e.g. , _ asynchronous scenario _ - in which _
_ _ _ each message @xmath243 is allowed to acquire an arbitrary time shift @xmath244 , _ phase shift scenario _ - in which each message @xmath243 is allowed to acquire an arbitrary phase shift @xmath245 and probably also a combination of the two where each message @xmath243 is allowed to acquire an arbitrary distortion of the form @xmath246
the previous discussion suggests that what we are seeking for is a large system @xmath7 of signals which will enable a reliable extraction of each bit @xmath236 for as many users transmitting through the channel simultaneously .
[ stability_def ] two unit signals @xmath247 @xmath11 are called * stably cross - correlated * if @xmath248 for every @xmath121 .
a unit signal @xmath249 is called * * stably autocorrelated * * if @xmath250 , for every @xmath251 . a system @xmath7 of signals
is called a * stable * system if every signal @xmath18 is stably autocorrelated and any two different signals @xmath252 are stably cross - correlated .
formally what we require for cdma is a stable system @xmath7 .
let us explain why this corresponds to a reasonable solution to our problem . at a certain time
@xmath3 the antenna receives a message @xmath253which is transmitted from a subset of users @xmath254 .
each message @xmath255 , @xmath256 @xmath23is of the form @xmath257 where @xmath258 .
in order to extract the bit @xmath236 we compute the matrix coefficient@xmath259 if the cardinality of the set @xmath262 is not too big then by evaluating @xmath263 at @xmath264 we can reconstruct the bit @xmath236 .
it follows from ( [ cross_eq ] ) and ( [ auto_eq ] ) that the oscillator system @xmath265can support order of @xmath24 users , enabling reliable reconstruction when order of @xmath266 users are transmitting simultaneously . * remark about field extensions . *
all the results in this paper were stated for the basic finite field @xmath1 for the reason of making the terminology more accessible . however , they are valid for any field extension of the form @xmath267 with @xmath268 complete proofs appear in @xcite . * acknowledgement . *
the authors would like to thank j. bernstein for his interest and guidance in the mathematical aspects of this work .
we are grateful to s. golomb and g. gong for their interest in this project .
we thank b. sturmfels for encouraging us to proceed in this line of research .
the authors would like to thank v. anantharam , a. grnbaum and a. sahai for interesting discussions . finally , the second author is indebted to b. porat for so many discussions where each tried to understand the cryptic terminology of the other .
howard s. d. , calderbank a. r. and moran w. , the finite heisenberg - weyl groups in radar and communications .
_ urasip journal on applied signal processing volume 2006 ( 2006 ) , article i d 85685 , 12 pages . _ | a novel system of functions ( signals ) on the finite line , called the oscillator system , is described and studied .
applications of this system for discrete radar and digital communication theory are explained . weil representation , commutative subgroups , eigenfunctions , random behavior , deterministic construction . |
investigating how active galactic nuclei ( agns ) are distributed in the universe provides a clue as to the physical conditions in which accretion onto supermassive black holes ( smbh ) takes place .
after the formation of galaxies in the universe , at any given time in the history of the universe , only a small fraction of galaxies show agn activity . it is known that almost all galaxies with a spheroidal component contain a smbh , and that the bh mass is closely related to the mass or the velocity dispersion of the spheroidal component ( e.g. , @xcite for review ) .
this means that most galaxies have likely had one or more brief agn periods , during which the central smbh grows in such a way that the growth of the smbh and the formation of the spheroidal component are tightly related .
the question of when and under what physical conditions the accretion takes place is important in understanding not only the origin and evolution of smbhs but also the origin and evolution of galaxies .
many models postulate that agn activity is merger - driven , especially those with high luminosities ( i.e. , qsos ) , ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . on the other hand , for lower - luminosity agns ,
processes internal to the galaxy such as galaxy disk instabilities , may be important @xcite .
different mechanisms may be responsible for triggering agn activity at different redshifts or luminosities .
large - scale clustering properties of agns provide important clues as to the physical process(es ) that are responsible for the smbh accretion .
the large - scale clustering amplitude reflects , through the bias parameter @xmath7 , the typical mass of the dark matter halos ( dmh ) in which agn reside @xcite .
thus exploring the clustering properties of agns at different redshifts , luminosities , and agn types to probe the masses of the dmhs that host them , provides us with important clues in understanding constraints on smbh growth in a cosmological context . clustering measurements of x - ray point sources
have been made whenever new large - scale surveys have been completed , through angular correlation functions ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
when redshifts for a complete sample become available , three - dimensional ( 3d ) correlation functions can then be computed ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
redshift information provides a huge leap in clustering measurements both in terms of the statistical accuracy as well as in removing systematic errors associated with model assumptions in making limber s de - projection @xcite .
however , even with redshift information , small number statistics has limited the accuracy of correlation function ( cf ) measurements , especially through the auto - correlation function ( acf ) of the agns themselves .
the situation can be improved by measuring the cross - correlation function ( ccf ) of agns with a galaxy sample that has a much higher space density . in order to facilitate a ccf analysis
, one needs an extensive galaxy redshift surveys with common sky and redshift coverage as the agn redshift surveys . with recent large - scale survey projects , measurements of agn clustering through ccf with
galaxies are now emerging ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
while the _ rosat _ all sky survey ( rass ) @xcite , most of which was conducted during the first half year of the _ rosat _ mission ( mostly in 1990 ) , produced a catalog of @xmath8 x - ray point sources , the availability of a comprehensive redshift survey of rass - selected agns was limited until a rass - sloan digital sky survey ( sdss ) matched agn catalog became available @xcite . in view of this
, we initiated a series of studies of the clustering properties of low redshift agns in the rass - sdss sample through the ccf analysis with sdss galaxies . in (
* hereafter paper i ) , we reported our results on our ccf analysis between broad - line agns in rass that have been identified with sdss @xcite and the sdss luminous red galaxies ( lrgs ) @xcite in the redshift range @xmath0 . through these efforts to measure the agn acf / ccf , the masses of the dmhs where agn activity occurs
are being gradually uncovered .
most results measuring the 3d correlation functions of x - ray selected agns indicate that the typical dmh mass in which these agn reside is in the range @xmath9\la 13.5 $ ] at both low ( @xmath10 ) and high ( @xmath11 ) redshifts ( e.g. , paper i ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
optically - selected qsos , typically representing a high redshift , high luminosity agn population , are associated with dmhs with a typical mass in the range @xmath9\la 13 $ ] ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
it appears that the luminosity dependence of agn clustering may be different at different redshifts . in paper
i , we show that among broad - line agn , high x - ray luminosity agns are more strongly clustered than lower x - ray luminosity agn at z@xmath12 . at @xmath11 , however , low luminosity x - ray selected agns are more strongly clustered than optical qsos @xcite .
this apparent difference may be caused by a non - monotonic luminosity dependence and/or agn type dependence of biasing rather than a dependence on redshift .
what is needed is to explore agn clustering across wider ranges in luminosity - redshift space to break this degeneracy . in interpreting the correlation function measurements ,
most previous studies use the large scale bias ( @xmath7 ) of agns to infer the associated typical dmh mass using linear growth and linear biasing schemes . strictly speaking ,
this is only valid for the correlation function measurements on sufficiently large scales ( @xmath13 mpc ) .
non - linear modeling through the halo occupation distribution ( hod ) framework ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) is imperative to accurately interpret and make full use of the correlation function measurements . in this framework ,
the mean number of the sample objects in the dmh ( @xmath14 ) is modeled as a function of the dmh mass ( @xmath1 ) .
then the two - point correlation function is modeled as the sum of the contributions of pairs from the same dmh ( 1-halo term ) and those from different dmhs ( 2-halo term ) .
this method has been used extensively to interpret galaxy correlation functions ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* hereafter z09 ) to constrain how various galaxy samples are distributed among dmhs as well as whether these galaxies occupy the centers of the dmhs or are satellite galaxies @xcite . partially due to the low number density of agns
, there have been few results in the literature interpreting agn correlation functions using hod modeling , where the small - scale clustering measurements are essential .
@xcite discussed qualitative hod constraints on their lrg - optical qso ccf , where they argued that the acceptable models include those in which @xmath15% of their qsos are satellites and those in which qsos are a random subsampling of a luminosity - threshold sample of galaxies .
@xcite also used the hod modeling approach to binary pairs of qsos at @xmath16 and conclude that they favor a model in which @xmath17% of the qsos are satellites .
in this paper , we apply the hod model to the ccfs of the rass broad - line agns ( hereafter , simply referred to as agns ) and sdss lrgs obtained in paper i. the hod modeling has allowed us to model the ccfs beyond simple power - law fits and to investigate constraints on how these agns are distributed among dmhs as a function of halo mass .
while almost all previous hod modeling studies apply the method to acfs , here we apply the hod modeling approach to the ccf . with this method ,
we use the previously estimated lrg hod to model the agn hod by fitting to the measured agn - lrg ccf . in our analysis , we take advantage of the lrg hod of z09 , which is based on the well - measured lrg acf from sdss @xcite .
this approach was taken by z09 in modeling lrg-@xmath18 galaxy ccfs , assuming that the lrgs are central galaxies of the dmhs and that the @xmath18 galaxies are satellite galaxies upon the calculation of the 1-halo term . in this paper , we present a comprehensive explanation of the hod analysis that can be applied to general cases , and then we apply the method to our agn - lrg ccf .
the scope of the paper is as follows . in sect .
[ sec : ccf_results ] , we summarize the samples used in paper i and the basic methods of ccf calculations . in sect .
[ sec : halomodel ] , we explain our basic modeling procedure of the lrg acf , which is used as a template to apply the hod modeling to the agn - lrg ccf . in sect .
[ sec : results ] , we show our results on the agn hods .
[ sec : disc ] discusses our results , including a comparison with the results from paper i and the astrophysical implications of our constraints on the agn hod obtained here .
finally , sect .
[ sec : concl ] summarizes important consequences from our analysis and concludes our discussion . throughout the paper , all distances
are measured in comoving coordinates and given in units of @xmath19mpc , where @xmath20kms@xmath21 .
we also use the symbol @xmath22kms@xmath21 for x - ray luminosities , while @xmath23 is used to express optical absolute magnitudes , for consistency with referenced articles .
we use a cosmology with @xmath24 , @xmath25 , @xmath26 , @xmath27 , @xmath28 and @xmath29 , which are consistent with the most updated wmap cosmology as of writing this paper @xcite .
the symbol @xmath30 signifies a base-10 logarithm , while a natural logarithm is expressed by an @xmath31 .
in paper i , we calculate the ccfs between agns and the lrg , as well as the lrg acf in the redshift range @xmath0 .
basic properties of the sample used in that paper are repeated here in table [ tab : samp ] .
the rationale and strategies of the sample selection are discussed in paper i in detail , including a redshift-@xmath32 diagram of the agn sample ( fig . 1 of paper i ) .
here we briefly summarize the sample selection and cf calculations below .
we use the agn sample from the rass - sdss matched agn catalog by @xcite , which is based on the sdss data release ( dr ) 5 . in order to make a well - defined , uniform agn sample , we only include broad lines agns . for the reference sample ,
we extract lrgs from the sdss catalog archive server jobs system using the flag `` galaxy_red '' , which is based on the selection criteria defined in @xcite .
we create a volume - limited spectroscopic lrg sample with @xmath33 and @xmath34 .
the agn sample is also limited to this redshift range . in order to obtain the largest common geometry between the agns and lrgs with public geometry and completeness files @xcite , we limit our sample to the sdss dr4 + geometry .
the selected geometry covers an area of 5468 deg@xmath35 .
the agn sample is further divided into high @xmath36 and low @xmath36 samples .
the numbers of lrgs and agns used in our ccf measurements are listed in table [ tab : samp ] .
ccccccc lrg sample & @xmath34 & @xmath33 & 45899 & @xmath37 & 0.28 & -21.71 + & & & & & & + & & @xmath38 range & & @xmath39 & & @xmath40 + & @xmath41-range & range [ @xmath42erg s@xmath21 ] & number & [ @xmath44 mpc@xmath45 & @xmath46 & [ @xmath42 erg s@xmath21 ] + + all rass - agn sample & @xmath34 & @xmath47 & 1552 & @xmath48 & 0.25 & 44.17 ( 44.16 ) +
high @xmath36 rass - agn sample & @xmath34 & @xmath49 & 562 & @xmath50 & 0.28 & 44.58 ( 44.53 ) + low @xmath36 rass - agn sample & @xmath34 & @xmath51 & 990 & @xmath52 & 0.24 & 43.95 ( 44.16 ) + in calculating the agn - lrg ccfs and the lrg acf , we use the classic @xcite estimator , in a two - dimensional ( 2d ) grid in the @xmath53 space , where @xmath54 is the projected distance and @xmath55 is the line - of - sight separation ( both in comoving coordinates ) : @xmath56 in which the subscripts 1 and 2 identify the sample , @xmath57 is the number of pairs between real samples , and @xmath58 is the number of pairs between real sample 1 and random sample 2 . for the agn - lrg ccf ,
the sample 1 is the agn sample and the sample 2 is the lrg sample , while for the lrg acf , both sample 1 and sample 2 are the same lrg sample .
we use eq .
[ eq : dp83 ] rather than that proposed by @xcite because eq .
[ eq : dp83 ] requires a random sample for only sample 2 ( in our case , lrgs ) .
this is important due of the difficulty in generating a random sample for a rass - based agn catalog .
_ rosat _ is sensitive to soft x - rays , which are subject to absorption by neutral hydrogen .
the variation in the column density across the sky due to neutral gas in our galaxy , combined with the diversity of the x - ray spectra of agn , makes it very difficult to model the spatial variation in the detection limit of the agns , which is essential in generating a random sample .
we then calculate the projected correlation function : @xmath59 where the upper bound of the integral ( @xmath60 ) is determined by the saturation of the integral .
we take @xmath61 and @xmath62 mpc for the lrg acf and agn - lrg ccf respectively .
the reasoning for these choices is explained in paper i. in paper i , we further calculate the _ inferred _ agn acf , @xmath63 , from the agn - lrg ccf @xmath64 and lrg acf @xmath65 , assuming a linear biasing scheme : @xmath66 ^ 2}{w_p(lrg|lrg)}. \label{eq : acf_infer}\ ] ] we then fit @xmath63 with a power - law form and derive the agn bias parameters and typical masses of the dmh in which the agns reside , based on the power - law fits . in this paper , instead of using the inferred agn acf as in paper i
, we directly model the ccf , @xmath64 , with an hod analysis , as @xmath64 has been directly derived from the observations and does not depend on a linear biasing approximation . the comparison of results from paper i and this paper are discussed below in sect .
[ sec : comp_paperi ] .
in performing the hod modeling , we consider that galaxies and agns are associated with dmhs , the mass function of which ( per comoving volume per @xmath67 ) is denoted by @xmath68 , where @xmath1 is the dark matter halo mass .
we use @xmath69 based on @xcite , which is in good agreement with later analyses by @xcite and @xcite . a dmh may contain one or more galaxies and/or agns that are included in our samples .
we then model the two - point projected correlation function as the sum of two terms , @xmath70 where the terms are : * the 1-halo term ( denoted by a subscript `` 1h '' ) , where both of the objects occupy the same dmh , and * the 2-halo term ( denoted by a subscript `` 2h '' ) , where each of the pair of objects occupies a different dmh .
recent articles prefer to define @xmath71 such that the total 3d correlation function is expressed by @xmath72 @xcite , instead of @xmath73 , as used in older articles .
this is because @xmath74 represents a quantity that is proportional to the number of pairs , and allows one to express each of the 1- and 2-halo terms consistently . in this case , the number of pairs is @xmath75+[1+\xi_{\rm 2h}]$ ] . in this new convention , our @xmath76 represents the projection of @xmath77 rather than @xmath71 , i.e. , @xmath78 d\pi.\ ] ] similarly , we express the power spectrum of the distribution of the objects in terms of the 1- and 2-halo term contributions : @xmath79 with @xmath80 where @xmath81 is the zeroth - order bessel function of the first kind . in the case of modeling the acf of a sample , the 2-halo term of the power spectrum can be approximated by @xcite @xmath82 where @xmath7 is the bias parameter of the sample , which we model as @xmath83 for the linear power spectrum , @xmath84 , we use the primordial power spectrum with @xmath29 and a transfer function calculated using the fitting formula of @xcite under our assumed cosmology ( sect . [
sec : intro ] ) . for the mass - dependent bias parameter of dmhs , @xmath85
, we use eq .
( 8) of @xcite with recalibrated parameter values by @xcite ( see their appendix a ) . in calculating the 1-halo term ,
the usual assumption is that the radial distribution of the involved objects that are not at the center of the halo objects that are satellites follows the mass profile of the dmh itself , i.e. , there is no internal biasing .
following @xcite , whose results will be used later in this work , we use the @xcite ( nfw ) profile for the dmh density distribution , which is still a popular choice for representing dmh profiles , although other parametrizations exist in the literature ( e.g. * ? ? ?
* ; * ? ? ?
we express the fourier transform of the nfw profile of the dmh with mass @xmath1 , normalized such that volume integral up to the virial radius is unity , by @xmath86 .
again , in the case of the acf , @xmath87dm_{\rm h } , \label{eq : p1h_acf}\end{aligned}\ ] ] where @xmath88 and @xmath89 are the numbers of the objects in the sample per dmh as a function of @xmath1 for those that are at the center of the halos ( central ) and those that are off center ( satellites ) respectively , while @xmath90 is their overall space density @xcite .
the hod model is calculated using a set of software we have developed , based partially on a code developed by j. peacock for the use in @xcite and @xcite .
our software includes a number of improvements over their code , most notably by including the application of the hod modeling to a ccf .
we note , however , that in our analysis we do not include some recent improvements in modeling the 2-halo term @xcite .
the neglected factors include the convolution of the dmh profile to the 2-halo term and the scale - dependent bias .
in addition , if a pair of objects are closer than the sum of the virial radii of their parent halos , the pair should not be counted in the 2-halo term ; this effect is neglected .
these improvements have enabled those authors to make accurate modelings of , e.g. , the acf of sdss galaxies , where the correlation functions can be measured to a few percent level at @xmath91 mpc . however , as agns have a much smaller number density , the errors on our ccf measurements are @xmath92% ( paper i ) at this @xmath54 , where these improvements are important .
therefore the effects of these recent improvements by other authors are estimated to be within the 1@xmath93 measurement errors of the ccf . in this subsection
, we explain the application of the hod modeling to the ccf between agns and lrgs ( or more generally , between any two different populations ) . using the hod of the lrgs as a template , which has been accurately constrained by z09 using the lrg acf measured by @xcite
, we constrain the hod of the agns by fitting to the agn - lrg ccf .
a similar approach has been made by z09 to model the ccf between lrgs and @xmath18 galaxies , where they assumed that the lrgs are at the halo centers and the @xmath18 galaxies are satellites in the 1-halo term calculation . since lrgs occupy the centers of almost all massive dmhs ( @xmath94 $ ] , z09 ) , this is a good approximation for their lrg-@xmath18 galaxy ccf as well as for our agn - lrg ccf .
however , for completeness , we develop a formulation of the hod modeling that includes more general cases . hereafter , the quantities for agns are represented by a subscript `` a '' , lrgs by `` g '' ( representing _ galaxies _ ) , and ccf between the two by `` ag '' .
we denote the lrg hods at the halo center and of satellites by @xmath95 and @xmath96 , respectively .
now @xmath97 .
likewise , the hods of the agns at the halo centers , of satellites , and the sum of the two are denoted by @xmath98 , @xmath99 and @xmath100 . in our approximate treatment
, the two halo term can be expressed as follows .
let the linear bias parameters of the lrgs and agns be @xmath101 and @xmath102 respectively : @xmath103 then the two halo term of the power spectrum corresponding to the ccf ( cross power spectrum ) can be expressed by @xmath104 the 1-halo term is composed of three terms : @xmath105\ , dm_{\rm h } , \label{eq : pag1h}\end{aligned}\ ] ] provided that no object is in common between the agn and lrg samples ( which is the case in our ccf ) . in this case , the relation @xmath106 holds , where the subscripts x and y represent any combination of the subscripts s and c , except for the case both are c s .
this relation is exact even if @xmath107 , where one has to use the exact poisson distribution for the probability distribution of having exactly @xmath108 objects in a halo ( sometimes called a sub - poissonian case ) . the relation @xmath109 also holds . even in cases where there are objects in common between the two samples ,
the above discussion does not change for the central - satellite pairs , because the two objects are surely different .
however , the existence of the common objects makes assumptions behind the third term in the integrand of eq .
[ eq : pag1h ] invalid .
a detailed discussion of this case is beyond the scope of the present paper , and will be made in a future paper .
the hod of the lrgs has been extensively studied by z09 based on the acf measurements of the lrgs by @xcite using sdss data release ( dr ) 3 .
their @xmath110 model has essentially five parameters , which are the values at five @xmath1 points , and they have spline - interpolation values between these points .
they take an elaborate parametrization of @xmath111 , which is an integration of a luminosity - dependent smoothed step - function , where the value increases from 0 ( at low @xmath1 ) to 1 ( at high @xmath1 ) , with a transition following a luminosity - dependent error function . in paper
i , we re - calculate the lrg acf for the @xmath0 , @xmath112 sample based on dr4 + , using @xmath11346,000 lrgs , instead of @xmath11330,000 lrgs from dr3 used by @xcite .
we obtain practically the identical acf to that calculated by @xcite .
since we want to use exactly the same model and the same lrg sample between the lrg acf and the lrg - agn ccf , we derive the best - fit lrg hod to our dr4 + sample using our software .
however , instead of newly exploring the full parameter space , we take advantage of the analysis in z09 to find the hods that give the best fit to our data by tweaking their lrg hods as follows .
1 ( b ) of z09 shows the @xmath114 upper and lower bounds of the hods separately for the central and satellite lrgs . for the satellites , we start with their @xmath114 upper and lower bounds , denoted by @xmath115 and @xmath116 respectively .
we then interpolate between these curves : @xmath117 where @xmath118 is a weight parameter in the linear interpolating procedure . for the central lrgs , we shift their central hod ( @xmath119 ) horizontally by @xmath120 in @xmath121 : @xmath122 we tweak @xmath123 horizontally because of the @xmath124 constraint . in calculating @xmath125 of the lrg acf based on the above hods
, we calculate the mean value using the exact poisson statistics in the small number case ( @xmath126 ) , while we use @xmath127 for larger values . as a whole ,
our model has two `` tweak '' parameters , @xmath118 and @xmath120 .
we calculate a series of model acfs at @xmath128 ( the average redshift of our sample , see table [ tab : samp ] ) in a parameter space grid exhaustively over a rectangular area in the @xmath129 space that is large enough to include all the region with @xmath130 .
the grid spacings are 0.03 and 0.003 for @xmath118 and @xmath120 respectively .
then we fit the model to the data by minimizing @xmath131 , taking the correlation of errors and the total number density of lrgs ( @xmath132 ) into account : @xmath133m^{-1}_{ij}\times\nonumber\\ & & [ w_{\rm
p}(r_{{\rm p},j})-w_{\rm p}^{\rm mdl}(r_{{\rm p},j})]\ } + \nonumber\\ & & ( n_{\rm lrg}-n_{\rm lrg}^{\rm mdl})^2/\sigma_{n_{\rm lrg}}^2 , \label{eq : acf_chi2}\end{aligned}\ ] ] where the quantities from the model are indicated by a superscript `` mdl '' , @xmath134 is the covariance matrix , and @xmath135 is the 1@xmath93 error of the lrg number density . the covariance matrix and @xmath135 are estimated using the jackknife resampling method ( see paper i ) . during the minimization process , the models @xmath136 are calculated by interpolating from the four nearest grid points in the parameter space computed above .
the minimization is performed using the minuit package distributed as a part of the cern program library . including the number density term in @xmath131 gives important parameter constraints , not only on the overall normalization , but also the shape
this is because @xmath123 , by definition , has the absolute maximum value of unity ( as there can not be more than one central galaxy in a halo ) , and for the lrgs , the value saturates at 1 for @xmath137 . with this constraint
, @xmath132 becomes sensitive to the combination of the relative normalizations between @xmath123 and @xmath138 and the `` cut - off '' mass of @xmath123 . for the lrg sample defined in paper i , @xmath139 mpc@xmath140 . in performing the fits ,
we use only the data points that are dominated by either the 1-halo or 2-halo terms , excluding the transition region ( @xmath141<2.8 $ ] ) . in this transition region
, our model acf becomes inaccurate due to our approximations , especially in neglecting the halo - exclusion effect in the 2-halo term , as described above in sect 3.1 .
we use the data points down to @xmath142 $ ] , which is slightly below the lower @xmath54 limit used in paper i. our measurements to this scale are consistent with those measure by @xcite .
our best fit model has ` tweak ' parameter values of @xmath143 and @xmath144 , where the 68% confidence ranges for two interesting parameters ( @xmath145 ) are given in parentheses .
our hods are slightly higher than those measured by z09 , probably because we include a data point at @xmath142 $ ] and we use a slightly different linear power spectrum . the best - fit model @xmath146 is compared with the observation as well as the z09 model in fig . [ fig : lrgacfhod](a ) .
the corresponding hods ( @xmath147 , @xmath148 and the total ) are shown in fig .
[ fig : lrgacfhod](b ) , and confidence contours in the ( @xmath120,@xmath118 ) space are shown in fig . [
fig : lrgacfhod](c ) .
-@xmath118 space ( see text for details ) .
the confidence levels are @xmath149 and 4.6 .
the cross shows the best - fit values . ] with the lrg hod in hand , we come to our main purpose of obtaining constraints on the hod of the agns in the rass - sdss survey . as a parametrized form of the agn hod
, we first try a simple truncated power - law form , assuming that all the agns are satellites ( model a ) :
@xmath150 where @xmath151 is the step function ( @xmath152 at @xmath153 ; @xmath154 at @xmath155 ) , @xmath156 is a critical dmh mass below which the hod is zero , and @xmath2 is a power - law slope of the hod above @xmath156 . while formally we are assuming that all agns are satellites , the hod constraints obtained from this assumption can be applied fairly to the sum of the central and satellite agns in cases where the agn activity in the central galaxy of high mass @xmath157\gtrsim 13.5 $ ] dmhs is suppressed .
for simplicity of explanation , we visualize the hod of the central lrg as a step function with a step at @xmath157=13.5 $ ] , and refer the dmh mass roughly above ( below ) this cut as higher ( lower ) mass .
there is no contribution to the 1-halo term at lower mass dmhs , since the 1-halo term requires both agns and lrgs to reside at the same dmh mass . since there is in practice no distinction between central and satellite agns in the 2-halo term ( eq . [ eq : p_ccf_2h ] ) , whether the agns in low mass dmhs are satellites or central makes no difference to the ccf , and indeed this model can fairly accommodate a case where a low - mass halo represents one galaxy ( halo ) at the center . at higher masses , among the three terms in eq .
[ eq : pag1h ] the satellite agn - central lrg pairs dominate the 1-halo term , because , as we show later , the present hod analysis is sensitive at @xmath157<14.0 $ ] , where there are in practice no satellite lrgs ( fig .
[ fig : lrgacfhod ] ) , and therefore the central agn - satellite lrg term as well as the satellite agn - satellite lrg term are negligible .
thus eq . [ eq : agnhod ] represents the sum of the central and satellite agn hod if there is no central agn at higher masses . the underlying assumption in model a is partially motivated by observations of agn `` downsizing '' @xcite , where the number density of high luminosity agns peaks earlier in the history of the universe and drops rapidly towards low redshift , while lower luminosity agn activity peaks later .
( we note , however , that a recent work by @xcite reports that this trend might be weaker than those reported previously . )
one possible implication of this trend is that black holes at the centers of massive galaxies occupying the centers of ( higher mass ) dmhs stopped accreting long before @xmath158 , while accretion occurs more frequently in lower mass galaxies in satellites or at the centers of lower mass halos in the redshift range of our sample .
thus model a can be a demonstrative case of this scenario .
additionally , central galaxies of massive halos are mostly early - type luminous galaxies , where agn activity is reported to be suppressed @xcite . in sect .
[ sec : modelb ] we consider in detail other models ( models b and c ) which uses well - studied galaxy hods ( central plus satellites ) as templates and treat the separate central and satellite hods explicitly . using a fixed set of @xmath95 and @xmath96 derived in the previous subsection and the formulations developed earlier in this section , we calculate the expected @xmath159 in the two parameter ( @xmath156,@xmath2 ) model of @xmath100 . due to much smaller errors of the lrg acf compared with those in the agn - lrg ccf
, we fixed the tweak parameter @xmath118 and @xmath120 at the best - fit values during the fits to the ccf . shifting the tweak parameters to any point on the @xmath160 contour in fig .
[ fig : lrgacfhod](c ) causes a shift of only @xmath161 to the same agn hod model .
this justifies the use of fixed lrg hods during the fit to the ccf .
we calculate a series of model ccfs in a parameter grid at @xmath128 with the spacings of 0.1 and 0.05 for @xmath162 and @xmath2 , respectively , over the rectangular region that we are interested in ( see confidence contours in the next section ) .
while the mean redshift of the agn samples range from @xmath163 to @xmath128 , the 1-halo and 2-halo terms vary by only @xmath164% and @xmath165% respectively between these redshifts , justifying our single redshift calculations .
we search for the best - fit model by minimizing the correlated @xmath131 : @xmath166m^{-1}_{ij}\times\nonumber\\ & & [ w_{\rm p , ag}(r_{{\rm p},j})-w_{\rm p , ag}^{\rm mdl}(r_{{\rm p},j})]\ } , \label{eq : chi2ccf}\end{aligned}\ ] ] where the errors and covariance matrix are estimated using the jackknife resampling method ( paper i ) . unlike in the case of lrgs , here we do not include the number density term in determining @xmath131 , because the ccf depends only on @xmath2 and @xmath156 in eq .
[ eq : agnhod ] but not the normalization .
thus the ccf constraints on @xmath2 and @xmath156 and the density constraints on the normalization can be separated .
after the constraints of these two parameters are made , the normalization can be determined by matching to the number density of the agns .
we use only data points with @xmath167mpc , as smaller scale bins contain only a small number of agn - lrg pairs ( @xmath168 pairs for the total and much less in the high and low @xmath36 rass - agn samples ) , which would not warrant the applicability of the @xmath131 statistic . unlike in the case of the lrg acf ,
we use all the data points in @xmath169\la 40 $ ] . due to the smaller mass for the dmh occupied by the agn compared to satellite lrgs
, the transition between 1-halo and 2-halo dominated regimes is very narrow for the agn - lrg ccfs , as will be seen in the next section . even within this small transition region ,
the effects of the major source of inaccuracy in our approximation , i.e. , halo - exclusion in the 2-halo term , is estimated be a few times smaller than our ccf measurement errors , based on the comparison of our lrg 2-halo term and that calculated by z09 .
using the methods described above , we obtain constraints on the agn hod parameters @xmath162 and @xmath2 in eq .
[ eq : agnhod ] . for each point in this two parameter space
, we can also calculate the mean dark matter halo mass occupied by the agn sample , @xmath170 as well as the effective bias parameter of agns @xmath102 ( eq . [ eq : bias_a ] ) .
figure [ fig : tot_ccf_con](a ) compares our agn - lrg ccf for the total rass - agn sample and our best - fit model .
figure [ fig : tot_ccf_con](b ) shows the confidence contours in the @xmath162
@xmath2 space , overlaid on underlying thin ( green ) contours showing the mean halo mass @xmath171 .
a line of constant bias , @xmath102 , roughly follows a constant @xmath171 contour , with small deviations caused by the non - linearity of the @xmath85 function .
the degree of the deviation is such that @xmath172 for a given @xmath7 decreases typically by @xmath173 when @xmath2 is decreased from @xmath174 to @xmath175 .
figures [ fig : hxlx_ccf_hod ] and [ fig : hxlx_con ] show our hod fits and confidence contours , respectively , for the high @xmath36 and low @xmath36 rass - agn samples , with an emphasis on the comparison between the two .
figures [ fig : tot_ccf_con](b ) and [ fig : hxlx_con ] show that the parameter pair @xmath176 is tightly constrained roughly along the constant @xmath171 line .
this primary constraint comes from the amplitude of the 2-halo term , which is proportional to the bias parameter of the agn sample ( for the fixed bias of the lrg sample ) . in the hod analysis
, however , the inclusion of the 1-halo term adds additional constraints in the two - parameter space . in any of the total ,
high @xmath36 and low @xmath36 rass - agn samples , models where the number occupation of agns is proportional to @xmath1 above @xmath156 ( @xmath177 ) are excluded , while a constant number of agns per halo above @xmath156 is preferred .
figs [ fig : tot_ccf_con](b ) and [ fig : hxlx_con ] put constraints on the values of @xmath156 for each sample . using the @xmath160 contour ( 68% confidence level for two parameters ) , the minimum mass is constrained to be @xmath178 \ga$ ] 11.9 , 11.9 , & 10.7 for the total , high @xmath36 and low @xmath36 rass - agn samples respectively . for the total rass - agn sample
, the contour is constrained to have @xmath179 , while for the high @xmath36 and low @xmath36 rass - agn samples , the contour allows the smallest @xmath2 that we have explored ( @xmath180 ) .
this essentially gives constraints on the _ width _ of the hod , and at least for the total rass - agn sample , a delta - function type hod could be marginally excluded .
and ( b ) low @xmath36 rass - agn samples .
the meanings of line styles and symbols are the same as those in fig .
[ fig : tot_ccf_con](a ) .
[ fig : hxlx_ccf_hod ] ] ( thick blue lines ) and the low @xmath36 ( thick red lines ) rass - agn samples .
the contours are at @xmath181 and @xmath182 levels .
the confidence contours are shaded for the ease of the visibility of the difference between the two samples .
the @xmath183 contours are shown in the same way as in fig .
[ fig : tot_ccf_con](b ) .
[ fig : hxlx_con ] ] ccccccc + total & 12.6 ( 11.9 ; 13.0 ) & -0.3 ( 0.4;-3.0 ) & 8@xmath184 & [email protected] & [email protected] & 6.9 + high @xmath36 & 12.9 ( 11.9 ; 13.2 ) & -0.5 ( 0.7;-3.25 ) & 3.3@xmath186&13.26@xmath187 & 1.44@xmath188 & 10.9 + low @xmath36 & 12.5 ( 10.7 ; 13.0 ) & -0.3 ( 0.7;-3.25 ) & 5.5@xmath184 & 12.97@xmath189 & [email protected] & 12.5 + the best fit hod parameters , the mean dmh mass occupied by the agns , and the linear bias from the fits are summarized in table [ tab : ccf_fit ] . for the fitting parameters @xmath162 and @xmath2
, we show the full range corresponding to 68% confidence range in the 2d parameter space ( @xmath160 ) . as shown in figs .
[ fig : tot_ccf_con ] and [ fig : hxlx_con ] , these two parameters are highly correlated and the confidence contours are skewed , thus one should be cautious in interpreting these ranges .
on the other hand , if we project the probability distribution on the variable @xmath190 or the variable @xmath7 ( bias ) , both of which are unique functions of our two fitting parameters , the projected probability distribution becomes roughly gaussian .
thus we take the @xmath191 range ( 68% error for one parameter ) to estimate the 1@xmath93 errors of these derived quantities , @xmath190 and @xmath102 , in table [ tab : ccf_fit ] .
we note that because of the exponential drop of the dmh mass function the hod at very high @xmath1 contributes little to the ccf . in order to check the maximum @xmath1 that our analysis can reasonably constrain , we recalculate the ccf with the truncated power - law model with an upper mass cutoff : @xmath192 we recalculate @xmath131 with varying @xmath193 while fixing ( @xmath156,@xmath2 ) at the best - fit value above .
the change of @xmath131 from the @xmath194 is larger than unity at @xmath195<14.0 $ ] .
thus our hod analysis is sensitive up to @xmath196 , i.e. , the mass of a poor cluster . in order to illustrate the range of acceptable hods , we plot a number of representative agn hod models accepted by our fits ( @xmath197 ) as function of dmh mass in fig .
[ fig : hods_dens](a ) . in fig .
[ fig : hods_dens](b ) , we also show the same sets of models in units of the spatial density per comoving volume per log of the dmh mass : @xmath198 these models have been normalized to the observed number densities ( see table [ tab : samp ] ) . in each panel , three curves are plotted , representing the best - fit case and two extreme cases .
these are ( 1 ) for the best - fit values ( solid lines ) , ( 2 ) the point on the @xmath199 contour that has the smallest @xmath156 value ( i.e. , the upper left tip of the two - parameter 68% confidence area in fig .
[ fig : tot_ccf_con](b ) or [ fig : hxlx_con ] ) ( dotted lines ) , and ( 3 ) the smallest @xmath2 point on the @xmath199 ( the bottom of the confidence region and narrowest possible hod distribution , dashed lines ) . the best - fit model with @xmath180 ( the smallest @xmath2 in our search grid )
is used for the high @xmath36 and low @xmath36 rass - agn samples , for which the @xmath199 contour continues below @xmath180 .
these three models have been plotted to illustrate the almost full range of possible hods statistically accepted by our analysis . in order to illustrate
the dhm mass range that can be occupied by both lrgs and agns , and therefore contributes to the 1-halo term of the ccf , we also show the hods of the lrgs for reference . of course , these plots are for our rather restrictive truncated power - law model and by no means are intended to illustrate an exhaustive set of possible hods .
however , this illustrates that these three models have roughly the same average @xmath1 , with varied widths of the hod .
an important constraint is that the agn hod can be no wider than the dotted line in [ fig : hods_dens](a)(b ) , which represents the point corresponding to the lowest extreme of @xmath156 and the roughly highest extreme of @xmath2 , among acceptable models with @xmath200 .
the meaning of this constraint is discussed in sect .
[ sec : disc_imp ] . ) of model
a for three representative points in the parameter space that are acceptable ( @xmath145 ) from our fits for each of the total ( top panel ) , high @xmath36 ( upper middle panel ) and low @xmath36 ( lower middle panel ) agn samples .
the solid , dotted , and dashed lines represent the ( 1 ) best - fit , ( 2 ) lowest @xmath156 extreme , and ( 3 ) lowest @xmath2 extreme respectively ( see text ) . for reference ,
the bottom panel shows the lrg hods for the central galaxies ( short - dashed line ) , satellites ( long - dashed line ) , and their sum ( solid line ) .
( b ) the same set of models are shown in terms of the spatial number density per @xmath201 .
the meaning of the line styles is the same as those in panel ( a ) . ] in this section we explore models in which central agns are explicitly included , using galaxy hods as a template .
galaxy hods have been extensively investigated using , e.g. , sdss data , with a high statistical accuracy .
a simple model for luminosity - threshold galaxy samples ( i.e. more luminous than a given limit ) , introduced by @xcite is a three parameter model including a step function for the central hod and a truncated power - law satellite hod .
we express the agn hods here with the same form , except that the global normalization is also a free parameter ( model b ) : @xmath202 where @xmath203 is the agn fraction ( duty cycle ) among central galaxies at @xmath204 .
we use the symbol @xmath205 to emphasize that it represents the hod slope for satellites .
@xmath206 is the dmh mass at which the number of central agns is equal to that of satellite agns . for luminosity - threshold galaxy samples , it is usually assumed that the central galaxy hod saturates at unity at the high @xmath1 end ( i.e. , @xmath207 in eq .
[ eq : step_pl ] ) , because central galaxies in the highest mass dmhs are expected to be luminous enough to be included in the sample ( e.g. sect .
[ sec : lrghod ] ) .
since only a small fraction of galaxies contain an agn , this has to be multiplied by a factor @xmath208 , the value of which does not affect the ccf as discussed above
. the value of @xmath203 can be determined by normalizing the hod to @xmath209 .
it is important to note the underlying physical assumption of eq .
[ eq : step_pl ] , which is that the agn duty cycle among central galaxies does not depend on the dmh mass .
we explore the consequences of this model as an illustrative contrast to model a , which assumes a suppression of the agn duty cycle in central galaxies of high mass dmhs .
we investigate the behavior of @xmath131 ( eq . [ eq : chi2ccf ] ) in the space of @xmath210 , @xmath206 , and @xmath205 . due to the low signal - to - noise ratio of the ccfs for our luminosity - divided subsamples , we limit our discussion to the total sample .
even with the total sample , our statistics do not allow us to constrain all three parameters ; therefore we investigate the constraints on two parameters by fixing the remaining parameter .
the hod analysis of luminosity - threshold samples with @xmath211 , studied by @xcite found @xmath212 and @xmath213 .
we consider the typical hod of sdss galaxy samples in this luminosity threshold range as a template because their @xmath210 range ( @xmath214\la 12.7 $ ] ) roughly coincides with the minimum dmh range for our agns , as we see from the range of @xmath156 in the confidence contour in fig .
[ fig : tot_ccf_con](b ) , which approximately coincide with the range of @xmath210 as we see below .
we therefore ( i ) fix @xmath215 and obtain constraints in the @xmath216-space , and ( ii ) fix @xmath217 and obtain constraints in the @xmath218-space . , @xmath205 ) space for @xmath215 .
( b ) the confidence contours of the same set of parameters for model c , for @xmath219=13.0 $ ] and @xmath215 .
the contours are at @xmath181 and @xmath182 levels .
[ fig : modelbccont ] ] cccccccc + b & total & 12.5 ( 12.2;13.0 ) & ... & 1.36 ( fixed ) & 0.55 ( 0.95;-1.50 ) & 2.9@xmath184 & 6.9 + b & total & 12.4 ( 11.9;12.6 ) & ... & 1.65 ( 1.51;1.83 ) & 1.20 ( fixed ) & 3.3@xmath184 & 8.4 + b & total & 12.3 ( 12.1;12.7 ) & ... & 1.55 ( 1.36;1.81 ) & 1.00 ( fixed ) & 2.6@xmath184 & 7.7 + c & total & 12.6 ( 12.2;13.0 ) & 13.0 ( fixed ) & 1.36 ( fixed ) & 0.0 ( 0.84;-1.50 ) & 3.7@xmath184 & 6.8 + the resulting confidence contours for ( i ) are shown in fig . [
fig : modelbccont](a ) .
the best - fit parameters and @xmath131 values are summarized in table [ tab : modelbc_fit ] for both ( i ) and ( ii ) , where we also show the @xmath220 case for ( ii ) , in which the number of satellites is proportional to the dmh mass . for the @xmath215 case
, the best - fit slope for the satellite is below unity ( @xmath221 ) and @xmath222 is identical to that of model a. the best - fit model with @xmath220 is marginally rejected ( @xmath223 ) and the model with @xmath224 is significantly rejected ( @xmath225 ) .
models that are consistent with @xmath224 can be found at @xmath226 , but they are not consistent with the hod of luminosity threshold galaxy samples in the @xmath211 range
. we also explore a model ( model c ) where only lower mass dmhs contain a central agn , where the physical picture is explained at the introduction of model a ( sect .
[ sec : agnhod_ex ] ) . in this model ,
the central hod in model b is replaced by @xmath227 while the form of @xmath228 stays the same .
as a representative case , we choose @xmath219=13.0 $ ] , below which the dmh center is more frequently occupied by blue galaxies than red galaxies in the hod analysis of color - separated samples by @xcite , while we keep @xmath229 ( but see also @xcite , which use a different parametrization on color - separated samples ) .
the confidence contours are shown in fig .
[ fig : modelbccont](b ) and the best - fit parameters and @xmath131 values are shown in table [ tab : modelbc_fit ] . in this scenario , the best - fit slope for the satellite is @xmath230 , while @xmath220 is rejected at the @xmath231 level .
cccccccc + total & [email protected] & 1.11@xmath232 & 1.52@xmath233 & [email protected] & [email protected] & 12.58@xmath234&12.70@xmath235 + high @xmath36&1.44@xmath188&1.44@xmath236 & 1.56 @xmath237 & 13.26@xmath187 & 13.19@xmath238 & 13.10@xmath239&13.27@xmath240 + low @xmath36 & [email protected] & 0.88@xmath241 & 1.6@xmath242 & 12.97@xmath189 & 12.87@xmath243 & 11.83@xmath244&12.03@xmath245 + table [ tab : papericomp ] compares our results with those from paper i. the meanings of the table columns are : ( 1 ) sample , ( 2 ) the bias parameter from the hod analysis ( eq . [ eq : bias_a ] ) , ( 3 ) the bias parameter from paper i , i.e. for the power - law fit in @xmath246<15 $ ] , ( 4 ) the bias parameter for the power - law fit in @xmath247<15 $ ] calculated in the same way as paper i ( see below ) , ( 5 ) the `` mean ''
dmh mass from eq .
[ eq : meanmass ] , ( 6 ) the `` typical '' dmh mass calculated from @xmath102 from the hod model , defined by @xmath248 calculated with the @xmath85 by @xcite ( see sect . [
sec : model_ing ] ) .
( 7 ) the `` typical '' dmh mass from paper i ( i.e. from power - law fits ) calculated with the @xmath85 by @xcite , and ( 8) the `` typical '' dmh mass from the paper i bias parameter re - calculated using the @xmath85 relation by @xcite .
while the results from the the hod analysis presented here and the simple analysis of paper i agree well for the high @xmath36 rass - agn sample , the hod analysis gives larger bias parameters than those derived in paper i for the total and low @xmath36 rass - agn samples at the level of 1.5 @xmath93 .
since these bias parameters have been derived from the same dataset with different methods , the discrepancies are systematic rather than statistical .
one of the major findings of paper i is the detection of the x - ray luminosity dependence of the correlation function .
the difference in correlation lengths ( @xmath249 ) , between the high @xmath36 and low @xmath36 rass - agn samples measured in paper i is @xmath250 , based on power - law fits with a fixed slope of @xmath251 .
the differences of the agn bias parameters @xmath102 and associated typical dmh mass @xmath252 derived from power - law fits are @xmath253 . in the hod analysis ,
the significance of the difference in @xmath102 decreases to @xmath254 , because it is constrained mainly by measurements at large scales where the 2-halo term dominates ( see below ) .
however , in terms of @xmath252 derived from the hod analysis , the difference is @xmath253 because of the additional constraints from the 1-halo term .
there are a number of differences in the processes used to derive the bias parameter between paper i and this work . in paper
i , the bias parameters are based on the power - law fits to eq .
[ eq : acf_infer ] over scales of @xmath255 < 15 $ ] , assuming a linear bias over the entire scale range .
then the power - law models are converted into the rms fluctuations over 8 @xmath19 mpc spheres ( @xmath256 ) , and the bias parameter is calculated using @xmath257 $ ] , where @xmath258 is the linear growth factor .
the main source of discrepancy is that , while paper i also uses data points down to a scale of @xmath259 mpc to obtain @xmath7 , in the hod analysis presented here the constraints on @xmath7 mainly come from the 2-halo term dominated region ( @xmath260 ) , while smaller scale data constrain other variables . fig .
[ fig : hxlx_ccf_hod ] shows that the difference between the high @xmath36 and low @xmath36 rass - agn samples is more prominent in the 1-halo term dominated regime , at @xmath261 , than in the 2-halo term dominated regime .
this means that , for the agn - lrg ccf , the 1-halo term is more sensitive to differences in the agn hod .
the reason for this is that while @xmath85 increases only weakly with increasing with @xmath1 , the 1-halo term is directly proportional to the number of agns that are in the same dmhs as those occupied by lrgs , i.e. , very massive ones ( @xmath262 $ ] ) , i.e. the 1-halo term is more sensitive to the existence of agns in high mass halos . in the literature ,
relative or absolute bias parameters are often calculated based on results from power - law fits ( e.g. , paper i ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the power - law models of @xmath263 are usually converted to the rms fluctuation over 8@xmath19 mpc spheres or are averaged up to the distance of 20@xmath19 mpc ( @xmath264 ) .
while some authors use only large scales ( @xmath265mpc ) to ensure that only the linear regime is used , others include smaller scales .
usually the rationale for including smaller scales is to have better statistics , combined with the empirical observation that the correlation functions are well fit with a power - law model over the scale of @xmath266mpc ( @xmath267 mpc in the case of paper i ) . as a check
, we have re - calculated the bias parameter for power - law fits to the @xmath63 used in paper i on scales of @xmath268 mpc , to limit ourselves to the linear regime .
these are shown in column ( 4 ) of table [ tab : papericomp ] .
the bias parameters obtained in this manner carry large statistical errors , but the values are consistent with those derived from the hod analysis within 1@xmath93 in all three samples , while they are more than @xmath269 above the bias parameters derived in paper i for the total and low @xmath36 rass - agn samples .
the agn bias parameters in our total and high @xmath32 samples from the hod analysis , as well as the power - law fit @xmath268 mpc , are consistent within combined 1@xmath93 errors with those measured by @xcite for their qso samples all0 and lstar0 , which covers a similar redshift range as ours ( @xmath270 ) .
however , their bias parameters for their higher redshift qso samples , all1 & lstar1 ( @xmath271 ) as well as all2 & lstar2 ( @xmath272 ) are lower . using the power - law fit results including the non - linear regime is useful to detect the different clustering properties between , e.g. , different source populations , luminosities , or types . in our case
, this gives the most statistically significant difference between the clustering properties of the high @xmath36 and low @xmath36 rass - agn samples .
however , one has to be careful in interpreting the results , e.g. , in terms of the dmh mass based on the bias parameter . in principle
, the hod analysis overcomes this limitation , because we are able to derive the pure linear bias parameter , while at the same time utilizing the data in the non - linear regime . with the hod analysis ,
we can also obtain constraints on at least one additional variable which is independent from the `` mean '' dmh mass , thanks to the constraints from the 1-halo term .
the importance of the 1-halo term constraints has also been noted by @xcite using a similar ccf analysis between photometric - redshift lrg samples and optically - selected qsos in sdss at @xmath273 .
they qualitatively concluded that at least some qsos must be satellites and also that models of the form of our model b ( eq . [ eq : step_pl ] ) with @xmath274 and @xmath275 are also acceptable , while lower @xmath2 values are also possible .
our analysis here gives qualitatively similar results .
the acceptable parameter space for model a is always in the @xmath276 regime for all of our three agn samples : the total , high @xmath36 , and low @xmath36 rass - agn samples .
this means that the average number of agns in a dmh does not increase proportionally with mass .
[ fig : tot_ccf_con](b ) and [ fig : hxlx_con ] show that models with @xmath277 are preferred .
it is interesting to compare this result with hod analyses of galaxies , that find @xmath278 for a wide range of absolute magnitudes and redshifts at least up to @xmath279 @xcite . in principle
, neither the results nor the assumptions of model a say anything about the central fraction among agns at the lower mass dmhs ( see sect .
[ sec : agnhod_ex ] ) , and thus they do not directly constrain the slope of the satellite agn hod @xmath205 . however , if we take the extreme case where all the agns are satellites , @xmath280 , this would imply that agn fraction among satellite galaxies strongly decrease with @xmath1 . in model b , where the central galaxy component is included explicitly under the assumption that the agn duty cycle in central galaxies is constant against varying @xmath1 above @xmath210 , the best - fit slope is still less than unity ( @xmath221 ) , and models with @xmath281
are only marginally rejected for @xmath215 ( a representative value of luminosity - threshold galaxy samples ) . in case of model c ,
where only @xmath282<13.0 $ ] dmhs can contain a central agn , the best - fit solution is @xmath230 , while @xmath281 is rejected at a @xmath283 level for @xmath215 . on the whole ,
our results tend to favor models in which the agn fraction among satellite galaxies _ decreases _ with increasing dmh mass ( or the richness of groups and clusters ) , in the mass range @xmath284\la 14.0 $ ] . since our agn sample is selected from x - ray point sources in rass , and since clusters of galaxies ( which represent the most massive dmhs ) are also x - ray sources , an observational bias that might affect our results is that agns in the clusters are _ selected against _ in the rass - agn sample due to confusion , especially considering that the mean point spread function ( psf ) has a fwhm@xmath164 arcminutes @xcite , which is not negligible when compared to the extent of the cluster x - ray emission .
we have estimated the significance of this effect as follows .
the detection limit of the cluster diffuse x - ray emission is higher than that of point sources .
a typical flux limit for clusters in rass is @xmath285 erg s@xmath21 @xmath286 , ( 0.1 - 2.4 kev)@xcite , corresponding to @xmath287 @xmath288 $ ] at the lower bound of our redshift range of @xmath289 .
this scales to a cluster virial mass of @xmath290 @xcite .
as we found at the end of section [ sec : results_fit ] , our ccf is only sensitive to the hod behavior at @xmath291 , and thus this selection effect should play negligible role in our results . additionally , a confusion due to two or more x - ray agn point sources being artifically blended into a single source by the rass point spread function ( psf ) is not important either , as @xmath292 is much less than unity over all masses ( see fig .
[ fig : hods_dens ] ) .
thus we can state with a good degree of confidence that our limits on @xmath2 ( model a ) or @xmath205 ( model b / c ) are not an artifact of x - ray source confusion . our results of @xmath293 , found in our extensive explorations in the parameter space of the models that we consider , may be interpreted in view of the long - suggested deficiency of emission line diagnostic - selected agns in rich clusters ( e.g. * ? ? ?
* ; * ? ? ?
* ) . while our ccf analysis is not sensitive to rich clusters ( sect .
[ sec : results_fit ] ) , a more recent work by @xcite verified that this trend extends to the group scale by finding an anti - correlation between agn fraction and the galaxy velocity dispersion in nearby ( @xmath294 ) groups / clusters of galaxies down to the velocity dispersion of @xmath295 $ ] .
@xcite also found that the x - ray selected agn ( @xmath296erg s@xmath21 , representing a much lower luminosity population than our agns ) fraction is larger in groups than in clusters by a factor of two at low redshifts ( @xmath297 ) .
various physical mechanisms have been suggested to explain the deficiency of star - formation galaxies in clusters of galaxies , and similar processes might be responsible for the lack of agns as well .
these include ram - pressure stripping @xcite or evaporation @xcite of a galaxy s interstellar medium , which is a required ingredient of the agn activity in galaxies , by the hot intracluster medium .
also , if agn activity is triggered by mergers between two gas - rich galaxies , another possible mechanism for this trend is the decreased cross - section for galaxy mergers in the galaxy - galaxy close encounters at high relative velocities @xcite . in high mass dmhs , representing richer groups / clusters ,
the velocity dispersion is higher and thus one may expect that there are fewer mergers than in groups of galaxies @xcite , although the density of galaxies increase at the center .
other mechanisms such as the accretion of small gas - rich dwarf galaxies ( minor mergers ) may play a role in fueling the agn host , but it is not clear how this would affect the slope @xmath205 . in order to fully understand the growth history of smbhs as well as the physical processes responsible for the agn activity in groups and clusters , we need to explore the statistics at different redshifts , luminosities , and agn types .
the results mentioned above that investigate agns in groups / clusters are concerned with agn at lower redshifts and with lower luminosities than the agns in our ccf analysis . in the aegis field , @xcite found that agns at @xmath11 are more frequently found in the group environment than the overall galaxy population .
however , deep field surveys such as aegis do not contain a large enough volume to explore the richness dependence of the agn fraction in groups / clusters .
@xcite investigate chandra images of a sample of 33 clusters at @xmath298 , finding that the luminous agn ( @xmath299erg s@xmath21 ) fraction increases with redshift .
they explore the redshift dependence of the agn fraction , but not the richness dependence . at @xmath300 , where the luminosity range of their agns is closer to ours
, they have only two agns in 17 clusters in the redshift range that is similar to our sample . a much larger sample of groups / clusters is needed to explore the richness dependence of the agn fraction in a sufficiently narrow redshift range to avoid any degeneracy with the redshift dependence . in general , directly investigating agn prevalence in individual groups and clusters is observationally and analytically intensive .
the hod analysis of galaxy - agn ccf provides a strong statistical tool with which to probe the agn population in groups and clusters , which can be made without identifying individual groups and clusters .
thus it is important to extend the hod analysis to agn clustering measurements at higher redshift , especially in view that the fraction of star - forming galaxies in clusters increase with redshift @xcite .
it is also important to have larger numbers of agns for the ccf measurements , which will enable measurements of small - scale clustering and produce stronger constraints on @xmath205 , as well as to break the model degeneracy . using non - lrg galaxies for the ccf to obtain the 1-halo term constraints in the lower @xmath1 regime would also improve the analysis .
linked to our constraints on @xmath2 in model a , we also have weak constraints on the range of @xmath156 .
the range of @xmath156 is similar to that of @xmath210 in model b. using the scaling relations and the luminosity range of our rass - agn samples , we can make an interesting comparison .
figure 1 of paper i shows that the minimum 0.1 - 2.4 kev luminosities of the total and high @xmath36 rass - agn samples are @xmath301 \approx 43.7 $ ] & 44.3 respectively , while the low @xmath36 sample has the same minimum luminosity as the total rass - agn sample . converting the 0.1 - 2.4 kev luminosity to a 0.5 - 2 kev luminosity , assuming a power - law spectrum with photon index @xmath302 and applying the luminosity - dependent bolometric correction from eq .
( 2 ) of @xcite , the corresponding bolometric luminosities are @xmath303 \approx$ ] 44.5 and 45.4 for the total and high @xmath36 rass - agn samples .
the mass of the smbh ( @xmath304 ) corresponding to the minimum luminosity is then @xmath305 and @xmath306 @xmath307 $ ] , where @xmath308 and @xmath309 is the eddington luminosity .
the relation between the black hole mass @xmath304 and the dmh of its host galaxy is explored by @xcite by comparing the rotation curves of spiral galaxies . while their @xmath304-dmh mass relation is subject to uncertainties due to different assumptions of the circular velocity and virial velocity by a factor of a few , using their eq .
( 6 ) with @xcite s recipe , the above black hole mass corresponds to a galaxy halo mass of @xmath310 ( @xmath311 ) @xmath312 $ ] ( for @xmath24 ) .
these masses are comparable or slightly lower than the lower bounds of @xmath156 derived above for those agns emitting at the eddington luminosity and are an order of magnitude lower than our best - fit values . in general
, the concept of the dmh of a galaxy explored by @xcite is different from the dmh derived in our ccf / hod analysis , since the large - scale correlation function reflects the dmh mass as the largest virialized structure the object belongs to .
if the largest virialized structure represents a group or cluster of galaxies , then there are multiple objects in a dmh , and the dmh that is related to @xcite s scaling relation represents a sub - halo in a larger mass halo . while our statistics do not allow for strong constraints , if @xmath156 is close to our nominal value , then the lowest luminosity agns among our samples should be accreting at a sub - eddington rate ( @xmath313 ) or reside in groups .
we have applied a halo occupation distribution ( hod ) model to the two - point cross - correlation function ( ccf ) between broad - line rass - agn and luminous red galaxies ( lrgs ) at @xmath0 measured in paper i. we have developed a method of applying the hod model directly to the ccf , where we have used the well - determined hod of the lrg as a reference to constrain the hod of the agns .
our findings are summarized as follows : 1 . from our hod analysis of the agn - lrg ccf
, we have found constraints on the hod of agns .
we have modeled the agn hod as a truncated power - law model ( model a ) with two parameters : the critical dark matter halo ( dmh ) mass ( @xmath156 ) below which the agn hod is zero ( i.e , there are no agn in halos below this mass threshold ) , and the slope @xmath2 , where the hod @xmath314 for @xmath315 .
we have obtained joint constraints in this two - parameter space , @xmath316 under the scenario that there is no central agns in high mass halos .
the constraints in this @xmath316 space lie mainly along lines of constant large - scale bias and constant mean dmh mass .
these provide a more accurate determination of the mean mass of the dmhs in which agns reside than the analysis based on power - law fits provided in paper i. the mean dmh mass we derive is @xmath317 $ ] @xmath318 , 13.26@xmath187 , and 12.97@xmath189 for the total , high @xmath36 and low @xmath36 rass - agn samples respectively .
the 2-halo term of the ccf primarily constrains the mean dmh mass .
3 . in all three of our rass - agn samples
, we have detected a significant contribution of the 1-halo term in the agn - lrg ccf , indicating that some agns are in the same dmhs occupied by lrgs .
this gives additional constraints on the agn hod beyond the mean dmh mass .
the range of acceptable hods of high @xmath36 and low @xmath36 rass - agn samples are significantly different in the @xmath316 space , confirming the paper i s results on the x - ray luminosity dependence of agn clustering properties at the @xmath319 level .
the most important constraint we have obtained using model a is an upper limit ( corresponding to @xmath160 ) , of @xmath320 for the total agn sample , rejecting @xmath321 . in the extreme case where all agns are satellites ,
the results show a sharp contrast with previously derived hods of galaxies , which show @xmath322 ( i.e. , the number of satellite galaxies approximately @xmath323 ) , across a range of galaxy luminosity and redshift .
taken at face value , this implies that _ the agn fraction among satellite galaxies decreases with increasing @xmath1 .
_ this is consistent with previous observations that the agn fraction is smaller in clusters than in groups in the nearby universe .
possible explanations include the ram - pressure stripping and/or evaporation of galactic gas by the hot intracluster medium and a decreased cross - section for galaxy merging in the high velocity dispersion environment of richer groups / clusters .
we also investigated a model ( model b ) which is composed of central and satellite agns , where the central hod is constant ( @xmath324 ) and the satellite hod has a power - law form @xmath325 , both at masses above @xmath210 .
this model is based on the results of hod analyses of galaxies . for @xmath215 , which is appropriate for galaxies ,
we obtain @xmath326 , finding a marginal preference for a picture in which the agn fraction among satellite galaxies decrease with dmh mass . with another model ( model c ) , which has the same form as model b except that only @xmath327<13.0 $ ]
dmhs are allowed to contain a central agn , we obtain slightly stronger constraints ( @xmath328 ) that give preference to this picture . 7 .
we have also obtained a lower limit ( corresponding to @xmath160 ) of @xmath329\la 11.9 $ ] for the total agn rass - sample .
if the lowest luminosity agns in our sample are emitting at the eddington luminosity , the black hole mass can be scaled to the dmh mass of the galaxy that is comparable to this lower limit of @xmath156 .
if the critical dmh mass is at the best - fit value ( @xmath330\approx 12.6 $ ] ) , then the lowest @xmath36 agns in our sample contain smbhs that accrete at low eddington ratios and/or reside in group environments .
we thank john peacock for providing us with his hod modeling code .
we also thank francesco shankar for stimulating discussions and the referee for extensive comments and suggestions , which have improved the paper very significantly .
this work has been supported by conacyt grant 83564 and dgapa - universidad nacional autnoma de mxico ( unam ) grants papiit in110209 and in109710 to ia - unam - ensenada as well as the nasa adp grant nnx07at02 g to ucsd .
this research made use of data from the _ rosat _ satellite , which was supported by the bundesministerium fr bildung und forschung ( bmbf / dlr ) and the max - planck - gesellschaft ( mpg ) , as well as the sloan digital sky survey ( sdss ) , the funding of which has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , the japanese monbukagakusho , and the max planck society .
the sdss web site is http://www.sdss.org/. this research also made use of computing facility available from departmento de supercmputo , dgsca , unam . | this is the second paper of a series that reports on our investigation of the clustering properties of agns in the _ rosat _ all - sky survey ( rass ) through cross - correlation functions ( ccfs ) with sloan digital sky survey ( sdss ) galaxies . in this paper , we apply the halo occupation distribution ( hod ) model to the ccfs between the rass broad - line agns with sdss luminous red galaxies ( lrgs ) in the redshift range @xmath0 that was calculated in paper i. in our hod modeling approach , we use the known hod of lrgs and constrain the hod of the agns by a model fit to the ccf . for the first time , we are able to go beyond quoting merely a ` typical ' agn host halo mass , @xmath1 , and model the full distribution function of agn host dark matter halos . in addition , we are able to determine the large - scale bias and the mean @xmath1 more accurately . we explore the behavior of three simple hod models .
our first model ( model a ) is a truncated power - law hod model in which all agns are satellites . with this model
, we find an upper limit to the slope ( @xmath2 ) of the agn hod that is far below unity .
the other two models have a central component , which has a step function form , where the hod is constant above a minimum mass , without ( model b ) or with ( model c ) an upper mass cutoff , in addition to the truncated power - law satellite component , similar to the hod that is found for galaxies . in these two models
we find that the upper limits on @xmath2 are still below unity , with @xmath3 and @xmath4 for model b and c respectively .
our analysis suggests that the satellite agn occupation increases slower than , or may even decrease with , @xmath1 in contrast to the satellite hods of luminosity - threshold samples of galaxies , which , in contrast , grow approximately as @xmath5 with @xmath6 .
these results are consistent with observations that the agn fraction in groups and clusters decreases with richness . |
from the very beginning of the investigation on the properties of the high - temperature superconductors @xcite it was obvious that strong local correlations together with long - ranged antiferromagnetic ( af ) exchange play an important role for the understanding of the anomalous normal state properties of these materials @xcite .
several models of strongly correlated electrons have been proposed since then to describe the essential physics of the cuprates @xcite .
the most common ones are the one- and three - band hubbard model @xcite and the @xmath0-@xmath1 model @xcite .
the latter is the most basic one taking into account explicitely both the strong local correlations and the magnetic interactions among itinerant electrons .
it can be viewed at as an effective model describing essentially the low energy physics of the more general hubbard models in the strong coupling limit @xcite . using standard notation
, the hamiltonian of the @xmath0-@xmath1 model reads @xmath4 where @xmath5 denotes the sum over nearest neighbours only and @xmath6 creates an electron at site @xmath7 with spin @xmath8 if and only if this site is unoccupied .
due to the constraint of no double occupancy imposed by the projected single - particle operators the hopping term is already nontrivial and contains strong correlation effects .
the second term describes the af exchange coupling of spins at nearest neighbour sites , denoted by the operators @xmath9 and @xmath10 respectively .
the @xmath11-term appears in the strong coupling expansion of the hubbard model .
we will drop it because it does not directly affect the magnetic properties we are interested in .
there is a large amount of knowledge about this model for the very special case of only one hole moving in an af background ( see e.g. @xcite and references therein . )
this limiting case proved to be of special interest since it could be solved exactly for @xmath12 @xcite .
the most important physical aspect is that the moving hole feels a binding potential growing linearly with the distance from its starting point due to the disturbance of the af background during its motion @xcite .
this linear potential leads to a sequence of discrete poles as spectrum for the one particle excitations with a distinct dependence on @xmath13 @xcite .
this rather simple picture appears to remain valid even for finite - dimensional systems @xcite , when transverse spin fluctuations are important . in order to compare with experiments
it is equally important to study the case of quasiparticle motion away from half filling at finite temperatures . to our knowledge
so far no results are available for this case in the thermodynamic limit . here
we want to study this situation in the framework of a dynamical mean field theory ( dmft ) , which is known to become exact in the limit of infinite spatial dimensions .
this approach provides an extension of the work of strack and vollhardt @xcite for one hole to the more general situation described above .
we find a multi - peak structure in the spectral functions that can be related straight forwardly to the discrete spectrum of the one hole case .
the paper is organized as follows .
we give a short review of our method and notation in the next section . in sec.3 the possible evaluation of a magnetic phase diagram is considered .
sec.4 shows resulting spectral functions and the first part of their interpretation . a brief review on the physics of one hole in the fully polarized af state together with an interpretation of the results from section 4 in connection with this picture
is then given in sec.5 , before the concluding remarks of sec.6 summarize the paper .
the ideas of the dmft for correlated electrons on a lattice are based on the substantial simplifications found in the limit of infinite dimensions @xcite . the most striking consequence of the @xmath14=@xmath15-limit concerns the irreducible one - particle selfenergy @xmath16 , which is due to the coulomb interaction in hubbard - like models .
this quantity becomes purely local , i.e. it looses all @xmath17 dependence in the limit @xmath14=@xmath15 @xcite .
nevertheless it keeps a highly nontrivial behaviour as a function of frequency . in order to perform the limit @xmath18 systematically the hopping matrix element @xmath0 has to be rescaled by @xmath19 with a constant @xmath20 @xcite .
we will chose @xmath21 as the unit of energy .
for a simple hypercubic lattice with nearest - neighbour hopping only the bare density of states is then found to be a gaussian @xcite : @xmath22 besides @xmath0 , also the next neighbour exchange interaction @xmath1 ( which is proportional to @xmath23 ) has to be rescaled according to @xmath24 it is well known that this type of interaction becomes trivial for @xmath18 in the sense that the hartree approximation becomes exact @xcite . in our case this means that the spin - interaction has to be treated at the mean field level : @xmath25 for a polarization in @xmath26-direction . here
@xmath27 @xcite .
obviously @xmath28 vanishes in the paramagnetic phase and in this case the @xmath0-@xmath1 model for @xmath14=@xmath15 is equivalent to the @xmath29=@xmath15 hubbard model .
here we are interested mainly in the one - particle properties of the @xmath0-@xmath1 model ( [ tjmodel ] ) _ in _ the antiferromagnetic state , i.e.when @xmath30 .
to allow for such a solution we must provide a small symmetry - breaking field and in addition formulate all quantities on an a - b lattice @xcite .
note that the use of two sublattices does not affect the locality of the self energies !
hence , the one - particle green s function may be represented as a matrix on the sublattices of the form @xmath31 for simplicity we introduced the abbreviation @xmath32 with the local selfenergies @xmath33 and the staggered magnetic field @xmath34 , which results from the exchange interaction ( [ mf ] ) on a hypercubic lattice . the local green s function is obtained by summing over @xmath17 with the result @xmath35 due to the additional symmetry @xcite @xmath36 it is sufficient to perform the calculations for the a - sublattice only and to use the spin index for the bookkeeping .
the actual calculation is a straight forward extension of the method used for the paramagnetic phase @xcite . due to the local nature of the selfenergies it is possible to reduce the solution of the lattice problem to that of a single site impurity problem with local energy @xmath37 in contact with an effective medium described by a green s function @xmath38 .
the latter contains all contributions from the local correlations except for those at the impurity site , i.e. @xmath39 this equation relates the green s function of the effective medium to the local green s function @xmath40 defined by ( [ gloc ] ) .
the remaining atomic problem is solved by using the resolvent method @xcite and the so - called non crossing approximation ( nca ) @xcite , from which we obtain the self energies of the effective impurity problem .
numerically exact methods like e.g. quantum monte carlo are not available for this particular problem due to the fact that we actually have to solve an impurity anderson model with @xmath41 @xcite .
thus , taking the impurity self energy as new approximation for the local self energies of the lattice problem ( [ gloc ] ) , we are able to determine @xmath42 selfconsistently by iteration . for details of this procedure
see e.g. @xcite .
in the following we are interested in solutions in the af state . therefore we choose slightly different starting conditions for up and down spins in our iteration procedure and look for solutions with a finite order parameter @xmath43 , which is nonvanishing only in the af phase .
there are of course other methods for obtaining the phase boundary , such as calculating the free energies for the para- and af - solutions or the susceptibility of the @xmath0-@xmath1 model .
a detailed discussion of the static magnetic properties following from the susceptibility will be presented in a forthcoming publication @xcite .
as it turns out , the task to calculate the complete magnetic phase diagram of the @xmath0-@xmath1 model as function of temperature , doping , and @xmath13 using the method outlined in the previous section can not be performed for the following reasons .
first , the self - consistency cycle for the order parameter @xmath43 converges extremely slow close to the phase boundary .
second , we could not set a reasonable boundary for @xmath43 below which one may safely assume that one is indeed in the paramagnetic phase and not in the af but with only a very small staggered magnetization .
the latter uncertainity for example leads to a comparatively large error in the exact position of the phase boundary .
we thus rest content with comparing a few points of the phase diagram obtained with the present calculations to the more accurate results obtained from the staggered susceptibility @xcite .
this is done in fig .
[ fig1 ] , where we show results for the critical @xmath1 as function of @xmath44 at fixed @xmath45 .
evidently both methods lead within the error bars to the same values of @xmath46 , as was to be expected .
( 12,8 ) ( 1,1 )
we now concentrate on the behaviour of the spectral function @xmath47 close to the transition from the paramagnetic to the af state .
[ fig2 ] shows the typical behaviour as the coupling constant @xmath48 is increased for fixed temperature @xmath49 and filling @xmath50 . for @xmath51
the system is in the paramagnetic state , in which case ( 12,8 ) ( 1,1 ) the model is equivalent to a @xmath29=@xmath15 hubbard model .
the structure near the chemical potential was earlier already related to some kind of single - band kondo effect @xcite . with increasing @xmath48 one eventually enters into the af regime , where the spectral functions of up and down spins ( full and dotted lines , respectively ) become inequivalent and show a pronounced multi - peak structure .
the latter will be discussed in the next section .
a similar series of pictures would be obtained , if the filling or the temperature were varied with the other parameters fixed . in order to show that the structures in both spin directions have the same origin
, we shift the energy scales about the staggered magnetic field @xmath52 that enters the local energy @xmath53 .
fig.3 shows that indeed the peak positions coincide .
note that in the following figures we will use energy scales for the different spectral functions , therefore , that are shifted in the same way .
the spectral function of the minority spins ( we will refer to it as @xmath54 in the following ) will sometimes also be rescaled in height by a suitable factor in order to be easier compared with @xmath55 .
note also , that the total spectral weight for the two spin directions may be different since we are looking only on the lower hubbard band for which no sumrule applies .
( 12,9 ) ( 1,1 ) an interesting problem concerns the question what happens to the kondo - like peak , observable in the paramagnetic state ( fig .
[ fig2 ] ) near the fermi energy ( 12,8.5 ) ( 1,1 ) in the case of a finite sublattice magnetization .
in fact , for low temperatures and small doping the spectral function shows an additional shoulder at the fermi energy if the spectra are shifted by the molecular field @xmath56 ( fig . [ fig3 ] ) .
this can be seen clearly in fig .
[ fig4 ] , where we compare the spectrum in the antiferromagnetic state whith the corresponding spectrum in the absence of the exchange interaction .
it is tempting to identify this feature with a kondo resonance in the antiferromagnetic state .
however , due to the restricted parameter and temperature range it is not possible to answer this question definitely and to exclude artefacts of the non - crossing approximation .
we now want to discuss in detail the multi - peak structure in the spectral function , which in our opinion can be related to bound states of a hole in a linear potential .
this becomes clear if we compare our results with the theory of strack and vollhardt @xcite .
there it is shown that the problem of one hole in the nel background can be solved exactly in the limit @xmath12 .
in addition , a particularly comprehensive deduction of the one - particle green s function is given .
before we start to present their results we should mention , that the definition of the green s function and the choice of the energy scale in @xcite are slightly different from ours : the zero of energy is the energy of the nel state with one hole @xmath57 and the hole s
green s function is defined by @xmath58 from this definition it can easily be seen that the sign of energy is changed and the energy scale is shifted about the chemical potential and the exchange field .
note also that a different way of scaling for the hopping energy in ( [ tjmodel ] ) was used , namely @xmath59 . according to this
we have to transform the results of @xcite in a proper way before comparing them to our thermodynamic green s functions . following the notation of @xcite the hole
s green s function reads @xmath60 it has poles at frequencies @xmath61^{\frac{2}{3}}\;\;,\ ] ] where @xmath62 are the zeros of the airy function @xmath63 .
@xmath64 has been calculated using the retraceable path approximation of brinkman and rice ( br ) @xcite , which becomes exact in the limit of high dimensions @xcite .
the poles correspond to bound states of a particle in a linear or string potential @xmath65 .
it is generated if the hole moves away from the initial site by @xmath66 steps , destroying pairs of antiferromagnetic bonds along its path , thus enhancing the energy roughly by @xmath67 ( cf .
fig.5 ) .
( 12,8 ) ( 0,-1 ) note , however , that even in @xmath14=@xmath15 the potential @xmath68 is only exact and well defined for the described situation of one hole moving in the af background .
therefore , the most important question that arises obviously is to what extent these bound states will survive and how the spectrum will be modified , as one turns to low but finite doping and to finite temperatures .
let us now return to our results for the one particle spectrum of the @xmath0-@xmath1 model in the af phase . in fig .
[ fig3 ] a set of distinct peaks can be seen , which we believe are directly related to the bound states of one hole .
of course , temperature and finite doping will lead to the observed broadening .
the more interesting question thus is to what extent the positions of the peaks in fig .
[ fig3 ] are related to those in equation ( [ gf_hole_poles ] ) . in fig .
[ fig6 ] we present a direct comparison for the following examples : a ) @xmath69 c ) @xmath70 b ) @xmath71 , @xmath72 , @xmath73 , @xmath74 d ) @xmath75 .
( 12,10 ) ( 0,1 ) the vertical lines represent the poles of the one - hole green s function ( [ gf_hole_rpa ] ) .
as mentioned before , the energy scale for the poles has to be inverted and shifted to allow a direct comparison with our data .
as we can not calculate accurately the chemical potential of the n - particle nel state , the shift is chosen to yield a reasonable fit to the peak positions of @xmath76 . in the four examples we find quite good agreement with the distance of the peak positions , particularly when the system is very close to half filling like in the cases b ) and c ) .
this also holds for other parameters we examined .
one can try to simulate the effect of finite doping in the theory of strack and vollhardt by assuming isolated holes , which move in a partially polarized background .
the analytic form of the green s function of this hole remains the same as in ( [ gf_hole_rpa ] ) , with the only difference that @xmath48 is replaced by an effective exchange interaction @xmath77 which is reduced by means of the polarization of the actual system .
this procedure leads to a better correspondence for the cases a ) and d ) but not for the other cases .
in this paper we presented results for the dynamics of the @xmath78 model in the antiferromagnetic state .
these were obtained in the framework of the dynamical mean field theory which becomes exact in the limit of infinite dimensions ( systems with many next neighbours ) .
the spectral function of single - particle excitations for finite doping and finite temperature has features ( a sequence of peaks ) which are similar to those found for a single hole in an antiferromagnetic groundstate and which can be understood by the motion of a hole in a string potential .
our results show that this special case is relevant also for finite hole concentrations and finite temperatures , at least for @xmath18 . in this limit
we were able to neglect spin - fluctuations in the antiferromagnetic state , which become more important in low dimensions .
then also the concept of a string potential breaks down . due to this effect
we expect a further broadening of the peak structure in the spectral function .
therefore it is not clear wether these structures will also be present in @xmath79 and @xmath80 for finite doping .
however , for one hole exact diagonalisation studies @xcite for one hole in @xmath79 show structures which can be related to the @xmath14=@xmath15 spectrum .
we wish to acknowledge helpfull discussions with d. vollhardt , w. metzner , f. gebhardt , p. van dongen , g. uhrig , k. becker , n. grewe , f. anders , r. eder , and many others .
one of us ( tp ) would like to thank the department of physics at the university of cincinnati , where part of this work was done , for its hospitality .
this work was supported by the deutsche forschungsgemeinschaft grant number .
j. g. bednorz , k. a. mller , z. phys .
b * 64 * , 189 ( 1986 ) p. w. anderson , science * 235 * , 1196 ( 1987 ) e. dagotto , rev .
phys . * 66 * , 763 ( 1994 ) j. hubbard , proc .
a * 276 * , 238 ( 1963 ) ; m. c. gutzwiller , phys .
rev . lett.*10 * , 159 ( 1963 ) ; j. kanamori , prog . theor .
phys . * 30 * , 257 ( 1963 ) v. j. emery , phys .
rev . lett.*58 * , 2794 ( 1987 ) f. c. zhang , t. m. rice , phys . rev .
b * 37 * , 3759 ( 1988 ) k. a. chao , j. spalek , a. m. oles , j. phys .
c * 58 * , l271 ( 1977 ) j. e. hirsch , phys .
rev . lett.*54 * , 1317 ( 1985 ) w. f. brinkman , t. m. rice , phys .
b * 2 * , 1324 ( 1970 ) c. l. kane , p. a. lee , n. read , phys .
b * 39 * , 6880 ( 1989 ) r. eder , k. w. becker , z. phys .
b * 78 * , 219 ( 1990 ) ; k. w. becker , r. eder , h. won , phys .
b * 45 * , 4864 ( 1992 ) j. igarashi , p. fulde , phys .
b bf 45 , 10419 ( 1992 ) ; phys .
b * 45 * , 12357 ( 1992 ) ; phys .
b * 48 * , 998 ( 1993 ) ; physica b * 186 - 188 * , 962 ( 1993 ) r. strack , d. vollhardt , phys . rev .
b * 46 * , 13852 ( 1992 ) for a survey on the subject see e.g. d. vollhardt : _ investigation of correlated electron systems using the limit of high dimensions _ in _ correlated electron systems _ , ed
. v. j. emery , world scientific singapore 1994 w. metzner , d. vollhardt , phys .
rev . lett.*62 * , 324 ( 1989 ) e. mller - hartmann , z.phys.b * 74 * , 507 ( 1989 ) c. itzykson , j. m. drouffe , _ statistical field theory _ ,
i & ii ( cambridge university press , cambridge 1989 ) note that on this level the neglected last term of the hamilonian ( [ tjmodel ] ) reduces to @xmath81 , i.e. it simply leads to a shift in the chemical potential unless we allow for a charge density wave .
u. brandt , c. mielsch , z. phys .
b * 79 * , 295 ( 1990 ) ; z. phys .
b * 75 * , 365 ( 1989 ) ; z. phys .
b * 82 * , 37 ( 1991 ) a. georges , g. kotliar , q. si , int . j. mod
* 6 * , 705 ( 1992 ) m. jarrell , t. pruschke z. phys .
b * 90 * , 187 ( 1993 ) ; t. pruschke , d. l. cox , m. jarrell phys . rev .
b * 47 * , 3553 ( 1993 ) g. hlsenbeck , f. stephan , z. phys .
b * 94 * , 281 ( 1994 ) h. keiter , j. c. kimball , phys .
rev . lett.*25 * , 672 ( 1970 ) n. e. bickers , d. l. cox , j. w. wilkins , phys .
b * 36 * , 2036 ( 1987 ) r. bulla , j. keller , t. pruschke z. phys .
b * 94 * , 195 ( 1994 ) q. qin , t. pruschke , t. obermeier , j. keller , m. jarrell _ to be published _
t.pruschke , m. jarrell , j. freericks , adv .
( in press ) w. metzner , p. schmit , d. vollhardt , phys . rev .
b * 45 * , 2237 ( 1992 ) r. eder , private communications | we study the one - particle properties of the @xmath0-@xmath1 model within the framework of vollhardt s dynamical mean field theory . by introducing an @xmath2 - sublattice structure
we explicitely allow for a broken symmetry for the spin degrees of freedom and are thus able to calculate the one - particle spectral function in the antiferromagnetic phase .
we observe surprisingly rich structures in the one - particle density of states for @xmath3 at finite doping up to 15% .
these structures can be related to the well known results for one single hole in the nel background .
we are thus able to establish the relevance of this at a first sight academic limit to physical properties of the @xmath0-@xmath1 model with a finite density of holes in the thermodynamical limit .
keywords : @xmath0-@xmath1 model , dynamical mean field theory , antiferromagnetism |
approximately 2% of main sequence a - type stars are similar to @xmath0 boo in having essentially solar carbon , nitrogen and oxygen abundances but subsolar abundances of heavier elements such as iron @xcite .
a plausible model to explain these distinctive abundances is that these stars have accreted gas but not dust @xcite . in this scenario ,
radiation pressure from the star prevents grains with their high opacity from accreting while gas is not so inhibited @xcite .
although the general picture that @xmath0 boo stars acquired their distinctive abundances by selective accretion is promising , there are objections which must be resolved @xcite . for example , most @xmath0 boo stars do not exhibit an infrared excess @xcite .
consequently , there is no evidence for either circumstellar dust or nearby interstellar dust . also ,
main - sequence a - type stars with dust do not always exhibit @xmath0 boo photospheric abundances @xcite . seemingly , unless there is undetected dust , some additional mechanism to explain selective elemental accretion should be considered .
the popular model for @xmath0 boo stars requires a substantial accretion rate . in particular , depending upon the amount of deep mixing and meridional circulation , required accretion rates range between 6 @xmath1 10@xmath2 g s@xmath3 and 6 @xmath1 10@xmath4 g s@xmath3 @xcite , comparable to the solar wind outflow of @xmath510@xmath6 g s@xmath3 . as described in detail below , these values are orders of magnitude greater than empirically - determined bounds on accretion rates by stars from the interstellar medium .
therefore , the thin outer envelope must be acquired from an orbiting source close to the star . a process that has been invoked to explain the anomalous circumstellar abundances of @xmath7 pic @xcite but not yet considered for @xmath0 boo stars is radiation pressure on individual ions .
this effect is highly element specific and therefore may be important in explaining the abundance pattern in @xmath0 boo stars .
if so , then dust need not be a constituent in the reservoir of the accreted material , allowing us to consider sources for accretion onto @xmath0 boo stars that were previously neglected .
in section 2 , we examine and discard the possibility of interstellar accretion . in section 3 ,
we assess the possibility that radiation pressure on atoms and ions can act to allow for selective accretion with the characteristic pattern seen in @xmath0 boo stars . in section 4 ,
we assess various possible sources of the matter accreted onto @xmath0 boo stars and suggest that one unappreciated possibility is the wind from an irradiated hot jupiter .
we discuss our results in section 5 and our conclusions in section 6 .
@xcite used bondi - hoyle theory to compute interstellar accretion onto @xmath0 boo stars . because the applicability of this approach is uncertain , @xcite , empirical measures of accretion rates should be considered . a sensitive upper bound on the time - averaged rates of interstellar accretion onto stars
can be derived from the amount of hydrogen found in the outer mixing zones of db white dwarfs , stars where the dominant element in the atmosphere is helium . because much or all of this hydrogen either had a circumstellar origin or was primordial @xcite , only upper bounds can be derived for the rate of interstellar accretion , @xmath8 . in a catalog of 57 dbs within 80 pc of the sun ,
the time - averaged value of @xmath8 typically is less than or equal to 10@xmath9 g s@xmath3 @xcite . for stars of mass , @xmath10 ,
the bondi - hoyle accretion rate varies as @xmath11 .
white dwarf stars typically have 1/3 the mass of a @xmath0 boo star @xcite , and therefore their accretion rates from the interstellar medium might be a factor of 10 lower .
however , even making this adjustment , the time - averaged accretion rate inferred for db white dwarfs is nearly a factor of 10@xmath12 lower than the minimum required rate . because @xmath550% of the 34 @xmath0 boo stars discussed by @xcite lie within 80 pc of the sun and therefore within the local interstellar bubble @xcite , this comparison with nearby white dwarfs is meaningful .
we also consider empirical measures of instantaneous " accretion .
we therefore use measures of accretion rates onto da white dwarfs ( stars where the atmosphere is dominated by hydrogen ) that are warmer than 13,000 k where the settling time of heavy elements below the photosphere is only days @xcite . in a sample of 87 such stars ,
the maximum carbon accretion rate is 2 @xmath1 10@xmath9 g s@xmath3 @xcite .
because solar abundances @xcite are assumed for the source material , @xmath0 boo stars are required to accrete carbon at a rate of 2 @xmath1 10@xmath13 g @xmath3 , a factor of 1000 greater than the maximum value measured for white dwarfs .
the true discrepancy is probably substantially larger because the most likely source of the white dwarf s photospheric carbon is circumstellar rather than interstellar @xcite .
we conclude that empirical bounds on interstellar accretion rates are much lower than what is required to explain @xmath0 boo stars .
before exploring the different possible sources of accretion onto @xmath0 boo stars , we consider an additional element - sorting mechanism for sources that are dust - free .
specifically , we extend models invoking radiation pressure on individual ions developed to explain the distinctive composition of @xmath7 pic s circumstellar gas to @xmath0 boo stars .
following notational convention , we define @xmath7 as the ratio of the outward force of radiation pressure compared to the inner force of gravity .
if the star has radius , @xmath14 , effective temperature , @xmath15 , mass , @xmath10 and emergent flux at the stellar surface , @xmath16 , then : @xmath17 here , @xmath18 denotes the charge of an electron of mass , @xmath19 , @xmath20 is the speed of light , @xmath21 is the gravitational acceleration at the photosphere and @xmath22 is the mass of the @xmath23th ion with resonance transitions of oscillator strength , @xmath24 .
we consider environments sufficiently far from the host star that ions mostly lie in their ground states . qualitatively , using the wien approximation to the planck curve and assuming that in the ultraviolet the opacity is so high that we observe only the upper layers of the atmosphere and that the photospheric lines are weak , we write that : @xmath25 where @xmath26 is the temperature minimum in the photosphere or @xmath5 0.75 @xmath15 @xcite .
a low value of @xmath7 requires a low value of @xmath16 which requires that the resonance lines lie at relatively short wavelengths .
important lines for c ii , n i and o i lie at 1335 , 1200 and 1302 , respectively , while , in contrast , important lines for mg , ca and fe lie at 2802 , 3933 and 2599 , respectively @xcite . therefore , c , n and o can be accreted while mg , ca and fe can not .
this qualitative expectation is borne out by detailed calculations @xcite for a model of @xmath7 pic with assumed values of @xmath15 = 8000 k and @xmath27 @xmath21 = 4.2 , stellar parameters appropriate for many @xmath0 boo stars @xcite .
we show in figure 1 a plot for @xmath0 boo of an element s abundance , compared to its solar abundance , @xmath28 , vs. the relative importance of outward radiation pressure , @xmath7 , where : @xmath29 there are two classes of elements : those with high values of @xmath28 and low values of @xmath7 in their likely dominant state of ionization and those with low values of @xmath28 and high values of @xmath7 .
this analysis extend to elements beyond those considered in figure 1 . both h and s
also are computed to have low values of @xmath7 and can be accreted while si and mn have high values of @xmath7 @xcite and would be inhibited from accreting , consistent with the general pattern of abundances among @xmath0 boo stars @xcite .
one complication to this picture is that there are some elements , such as na , where @xmath7 can be high in one state of ionization and low in another .
for such an element , the net balance among all the potential states must be determined in order to assess whether it is accreted .
although there are uncertainties , we argue that outward radiation pressure on individual elements might play a key role in explaining the abundance pattern of @xmath0 boo stars .
the alternative to interstellar accretion is accretion from some material gravitationally bound to the host star .
we now assess different possibilities .
a possible scenario to explain @xmath0 boo abundances is the accretion of the outflow from a stellar companion .
in fact , there are rare instances of spectroscopic binary systems where two stars orbit close to each other and one star exhibits a @xmath0 boo abundance pattern @xcite .
however , such situations probably do not explain most @xmath0 boo stars @xcite .
a binary configuration of two more distant stars which might account for some @xmath0 boo objects would be an a - type star accreting from the wind of a late type companion . the typical mass loss rate from a late - type main - sequence star is @xmath5 10@xmath6 g s@xmath3 @xcite , comparable to the mass accretion rate required to explain @xmath0 boo stars . because the characteristic wind speeds from late type stars likely
are near their escape speed of 600 km s@xmath3 , only a small fraction of the matter lost from the companion can be accreted by the @xmath0 boo star @xcite .
therefore , accretion of winds from late - type stellar companions seems unlikely to explain most @xmath0 boo stars .
the carbon to oxygen ratio measured in the atmospheres of @xmath0 boo stars typically is within a factor of three of the solar ratio @xcite .
in contrast , most matter from rocky minor planets accreted onto white dwarfs has a carbon to oxygen ratio that is typically a factor of 100 lower than solar @xcite .
this marked deficiency of carbon within extrasolar asteroids the building blocks of rocky planets argues against their being a major reservoir of matter accreted onto @xmath0 boo stars .
another source of accreted material might be analogs to kuiper belt objects or comets with carbon to oxygen abundance ratios that might be nearly solar @xcite although perhaps not @xcite .
@xcite have proposed that colliding comets explain the observed gas around 49 cet , a young main sequence a - type star .
the suggested co production rate is greater than @xmath1 10@xmath30 g s@xmath3 , large enough to induce @xmath0 boo abundances .
however , there are difficulties with this model .
it is not clear how matter initially at @xmath31100 au ultimately accretes onto the host star .
also , comets by consisting largely of water , may have very subsolar hydrogen to oxygen ratios .
accretion of such material might produce an oxygen overabundance which is not observed .
finally , 49 ceti has an infrared excess while most @xmath0 boo stars do not .
given these objections , alternative reservoirs for the accretion onto @xmath0 boo stars should be considered .
in the absence of accretion , mixing erases the abundance anomalies in @xmath510@xmath9 yr @xcite . while accretion from a primordial disk may lead to the @xmath0 boo abundance pattern measured in numerous pre - main - sequence stars @xcite , there are many @xmath0 boo stars that are sufficiently old that their photospheric abundances are not a relic from the era of their formation @xcite .
a recent source of accretion should be considered . because the source must be substantial and likely located relatively near the star to guarantee ultimate accretion ,
we propose that the source of matter accreted onto many @xmath0 boo stars is the wind from a close - in planet . as an example , consider wasp 33 ( or hd 15082 ) , an a5 main - sequence star orbited by a hot jupiter with a period of 1.22 d @xcite .
the mass loss rate from this planet may be 2 @xmath1 10@xmath30 g s@xmath3 @xcite , sufficiently high to explain @xmath0 boo stars . here
, we assume that all the mass lost from the planet ultimately is acquired by the star .
furthermore , we also assume that individual ions can separate from the general flow as supported by a toy model not presented here .
detailed calculations are required to assess these requirements .
for example , in the case of the outer circumstellar gas orbiting
@xmath7 pic , the effectiveness of the separation between different elements depends upon the viscosity in the gas @xcite . in models
where the upper atmosphere is heated by ultraviolet and x - ray photons , the wind rate , @xmath32 , from a planet at distance @xmath33 from the can star can be estimated as : @xmath34 where @xmath35 is the gravitational constant , @xmath36 is the total stellar euv luminosity in the spectral range @xmath0 @xmath37 912 , @xmath38 is the planet s mean density , and @xmath39 is an efficiency factor to describe the conversion of incident radiative energy to kinetic energy in the outflowing wind @xcite .
this formula may overestimate the planetary mass loss rate if the planet s magnetic field is sufficiently large @xcite . now consider
optimistic " values of the different parameters in equation ( 4 ) .
we take @xmath40 of and @xmath38 for wasp 17b of 1.3 @xmath1 10@xmath2 cm and 0.062 g @xmath41 , respectively @xcite , the most inflated hot jupiter currently known @xcite .
we assume @xmath33 = 6.0 @xmath1 10@xmath2 cm in order for the planet to be stable against tidal disruption @xcite .
we also take @xmath39 = 1 .
because of attenuation in the interstellar medium , the euv luminosity of most stars is unobservable and therefore highly uncertain .
indirect estimates of @xmath42 can be derived from observed subcoronal emission lines characteristic of gas with temperatures between 50,000 k and 300,000 k. such lines have been detected in stars with @xmath43 @xmath37 8200 k although stars as hot as 8800 k may possess coronae @xcite .
further , approximately 10 - 15% of main sequence a - type stars are detected x - ray sources ; the underlying physical explanation is not known @xcite . here , to be specific , we consider for comparison the well studied star , @xmath7 pic , with @xmath43 = 8200 k @xmath44 150 k @xcite , a representative temperature for many @xmath0 boo stars . for @xmath7 pic ,
the measured x - ray luminosity between 0.2 and 20 kev is 3 @xmath1 10@xmath45 erg s@xmath3 , and the extrapolated x - ray luminosity between 0.06 and 5 kev is 3 @xmath1 10@xmath46 erg s@xmath3 @xcite .
there must also be emitted photons with energies between 0.0136 and 0.06 kev . according to @xcite , the total chromospheric emission of @xmath7 pic is at least 9.9 @xmath1 10@xmath47 erg s@xmath3 .
while much of this energy is emitted with @xmath0 @xmath48 912 , an appreciable fraction is thought to be emitted at @xmath0 @xmath37 912 . here
, recognizing that there is great uncertainty , we adopt @xmath36 = 2 @xmath1 10@xmath47 erg s@xmath3 or @xmath49/l@xmath50 @xmath5 5 @xmath1 10@xmath51 .
if so , then @xmath32 @xmath52 10@xmath6 g s@xmath3 , sufficiently large to explain @xmath0 boo abundances .
one complication is that a hot jupiter may have a sufficiently short period that it can induce activity on its host star .
@xcite have reported evidence to support @xmath510@xmath46 erg s@xmath3 of such dissipation .
accordingly , through magnetically controlled interactions , the host planet might lose 10@xmath6 g s@xmath3 @xcite .
however , it is uncertain whether models developed for later type stars apply to a - type stars , but if they do , then the planetary outflow would be sufficiently great to explain many @xmath0 boo abundances .
this model might apply to to hr 8799 , a main - sequence a - type star orbited by planets in wide orbits @xcite which is also is a mild " @xmath0 boo star @xcite . because hr 8799 emits 1.3 @xmath1 10@xmath47 erg s@xmath3 in the 0.2 - 2 kev band @xcite , a wind from a hot jupiter an additional planet interior to those already known - might account for the photospheric abundances . in this analysis
, we assume that hot jupiters have an atmosphere with nearly a solar composition .
although this result is unestablished @xcite , at least na and k have been detected in the outer atmospheres of hot jupiters @xcite .
further , winds from hot jupiters display evidence for mg @xcite and possibly si @xcite .
current data are consistent with the hypothesis that the abundances in the atmospheres of hot jupiters are similar to the material accreted on @xmath0 boo stars .
winds from hot jupiters may be the source of accretion for many @xmath0 boo stars . for f , g and k stars ,
the frequency of hot jupiters is 1.2% @xmath44 0.38% @xcite .
although the frequency of hot jupiters orbiting main - sequence a - type stars is unknown , the fraction of stars with giant planets appears to peak for stellar masses near 1.9 m@xmath53 @xcite , and it is plausible that the frequency of hot jupiters around a - type stars is sufficiently high to explain many @xmath0 boo stars .
if the wind from a hot jupiter provides the material accreted onto a @xmath0 boo star , then it may be possible to detect transits or more subtle photometric variations for systems inclined less than 90@xmath54 .
@xcite has analyzed the kepler light curves for @xmath52000 a - type stars and found 166 candidate hot jupiters @xcite .
further study is required to determine how many of these systems truly harbor planets . in our proposal ,
some of the elements in a wind from hot a jupiter are expelled into the interstellar medium by the star s radiation pressure and therefore mimic a stellar wind .
this scenario might explain the observed outflow apparently from sirius @xcite which is not otherwise well understood .
usually the wind speed is comparable to the escape velocity , and the modest blueshift of between -20 km s@xmath3 to -80 km s@xmath3 in the outflow attributed to sirius may instead be from a planet .
one observational argument against our proposed scenario is that hd 15028 the host star of wasp 33 , an a - type star with a hot jupiter is classified as an am star @xcite and not a @xmath0 boo star .
however , if the wind from the planet is magnetically funneled onto the host star , then perhaps elemental separations would not occur .
it may be that only some winds from hot jupiters lead to the @xmath0 boo pattern of abundances .
another observational argument against our model is that two @xmath0 boo stars identified in the kepler survey @xcite are not obvious candidates for harboring a hot jupiter @xcite .
however , this sample is tiny and the result inconclusive . despite difficulties , the scenario of accretion from hot jupiters appears to be the least unlikely of all current proposals to explain @xmath0 boo stars .
future observational tests of @xmath0 boo models can be performed with the well defined set of stars provided by @xcite .
we argue that empirical bounds exclude interstellar matter as the source of accreted material onto @xmath0 boo stars .
we then propose two modifications to the popular model to explain the abundance pattern in @xmath0 boo stars .
first , we suggest that radiation pressure on individual gas - phase ions might be an important mechanism for selecting which elements can be accreted .
second , we suggest that the matter acquired by some @xmath0 boo stars may originate in the wind from a hot jupiter .
99 anderson , d. r. , smith , a. m. s. , lanotte , a. a. et al .
2011 , , 416 , 2108 ballester , g. e. & ben - jaffel , l. 2015 , 804 , 116 balona , l. a. 2013 , , 431 , 2240 balona , l. a. 2014 , , 441 , 3543 bento , j. wheatley , p. j. , copperwheat , c. m. et al .
2014 , , 437 , 1511 bergeron , p. , wesemael , f. , dufour , p. et al .
2011 , , 737 , 28 bertin , p. , lamers , h. j. g. l. m. , vidal - madjar , a. , ferlet , r. , & lallement , r. 1995 , , 302 , 899 bouret , j .- c . , deleuil , m. , lanz , t. et al .
2002 , , 390 , 1049 bourrier , v. , lecalelier des etangs , a. , & vidal - madjar , a. 2015 , , 573 , a11 burrows , a. 2014 , nature , 513 , 345 charbonneau , d. , brown , t. m. , noyes , r. w. , & gilliland , r. l. 2002 , , 568 , 377 collier - cameron , a. , guenther , e. , smalley , b. et al .
2010 , , 407 , 507 cowley , c. r. 2014 , in putting a stars into context : evolution , environment , and related stars , eds : g. mathys et al .
; moscow publishing house cowley , c. r.,hubrig , s. , & przybilla , n. 2014 , , 440 , 2457 debes , j. h. 2006 , , 652 , 636 farihi , j. , barstow , m. , a. , redfield , s. et al .
2010 , , 404 , 2123 fernandez , r. , brandeker , a. , & wu , y. 2006 , , 643 , 509 folsom , c.p . , bagnulo , s. , wade , g. a. et al .
2014 , , 422 , 2072 fossati , l. , haswell , c. a. , froning , c. et al .
2010 , , 714 , l222 gray , r. o. , & corbally , c. j. 1992 , , 124 , 989 gray , r. o. , & kaye , a. b. 1999 , , 118 , 2993 griffin , r. e. , gray , r. o. , & corbally , c.j .
2012 , , 547 , 8 gunther , h. m. , wolk , s. j. , drake , j. j. et al .
2012 , , 750 , 78 haswell , c. a. , fossatim , l. , ayres , t. et al . , 2012 , , 760 , 79 heiter , u. 2002 , , 381 , 959 heiter , u. , weiss , w. w. , & paunzen , e. 2002 , , 381 , 971 jura , m. , dufour , p. , xu , s. et al . 2015 , , 799 , 109 jura , m. & xu , s. 2012 , , 143 , 6 jura , m. & young , e. d. 2014 , ann .
earth planet .
, 42 , 45 kamp , i. , hempel , m. , holweger , h. 2002 , , 388 , 978 kamp , i. , & paunzen , e. 2002 , , 335 , l45 king , j. r. 1994 , , 269 , 209 kleinman , s. j. , kepler , s. o. , koester , d. et al .
2013 , , 204 , 5 koester , d. 1976 , , 52 , 415 koester , d. 2009 , , 498 , 517 koester , d. , gaensicke , b. , & farihi , j. 2014 , , 566 , 34 kurucz , r. l. 1979 , , 40 , 1 lallement , r. , welsh , b. y. , vergely , j. l. , crifo , f. , & sfeir , d. 2003 , , 411 , 447 lanz , t. , heap , s. r. , & hubeny , i. 1995 , , 447 , l41 lanza , a. f. 2013 , , 557 , a31 linsky , j. l. , yang , h. , france , k. et al .
2010 , , 717 , 1291 lodders , k. 2003 , , 591 , 1220 marois , c. , zuckerman , b. , konopacky , q. m. , macintosh , b. , & barman , t. 2010 , nature , 468 , 1080 martinez - galarza , j. r. , kamp , i. , su , k. y. l. , et al .
2009 , , 694 , 165 morton , d. c. 1991 , , 77 , 119 murphy , s. j. , corbally , c. j. , gray , r. o. et al .
2015 , , in press , astroph 1508.03633 murray - clay , r. , chiang , e. i. , & murray , n 2009 , , 693 , 23 narusawa , s. , ozaki , s. , kambe , e. , & sadakane , k. 2006 , , 58 , 617 niemczura , e. , murphy , s. j. , smalley , b. et al .
2015 , , 450 , 2764 owen , j. e. & adams , f. c. 2014 , , 444 , 3761 paczynski , b. 1971 , , 19 , 183 paunzen , e. 1991 , , 373 , 633 paunzen , e. , iliev , i. kh . , kamp , i. & barzova , i. s. 2002 , , 336 , 1030 paunzen , e. , kamp , i. , weiss , w. w. , & wiesemeyer , h. 2003 , , 404 , 579 reffert , s. , bergmann , c. , quirrenbach , a. , trifonov , t. , & kunstler , a. 2015 , , 574 , a116 robrade , j. & schmitt , h. h. m. m. 2010 , , 516 , a38 schroder , c. & schmitt , j. h. m. m. 2007 , , 475 , 677 shkolnik , e. , walker , g. a. h. , bohlender d. a. , bu , p .-
g . , & kurster , m. 2005 , , 622 , 1075 simon , t. , ayres , t. r. , redfield , s. , & linsky , j. 2002 , , 579 , 800 sing , d. k. , wakeford , h. r. , showman , a. p. et al .
2015 , , 446 , 2428 stutz , ch . , & paunzen , e. 2006 , , 458 , l17 turcotte , s. 2002 , , 573 , l129 turcotte , s. & charbonneau , p. 1993
, , 413 , 376 venn , k. a. , & lambert , d. l. 1990 , , 363 , 234 vidal - madjar , a. , huitson , c. m. , bourrier , v. et al .
2013 , , 560 , a54 xie , j .- w .
, brandeker , a. , & wu , y. 2013 , , 762 , 114 waters , l. b. f. m. , trams , n. r. , & waelkens , c. 1992 , , 262 , l37 wilson , p. a. , sing , d. k. , nikolov , n. et al .
2015 , , 450 , 192 wood , b. e. , muller , h .-
r . , zank , g. p , & linsky , j. l. 2002 , , 574 , 412 wright , j. t. , marcy , g. w. , howard , a. w. et al .
2012 , , 753 , 160 zuckerman , b. & song , i. 2012 , , 758 , 77 | the abundance anomalies in @xmath0 boo stars are popularly explained by element - specific mass inflows at rates that are much greater than empirically - inferred bounds for interstellar accretion .
therefore , a @xmath0 boo star s thin outer envelope must derive from a companion star , planet , analogs to kuiper belt objects or a circumstellar disk . because radiation pressure on gas - phase ions might selectively allow the accretion of carbon , nitrogen , and oxygen and inhibit the inflow of elements such as iron ,
the source of the acquired matter need not contain dust .
we propose that at least some @xmath0 boo stars accrete from the winds of hot jupiters . |
one of the earliest results from infrared ( ir ) astronomy was the discovery of ` excess ' ir radiation by heated circumstellar dust from a variety of stars having emission - line spectra .
amongst these were four wc9 type wolf rayet ( wr ) stars , ve245 (= wr104 ) , hd 313643 (= wr106 ) , ls 15 (= wr119 ) and as 320 (= wr121 ) , observed by allen , harvey & swings ( 1972 ) and , at longer wavelengths , by gehrz & hackwell ( 1974 ) .
cohen , barlow & kuhi ( 1975 ) measured optical
ir energy distributions of wr stars and showed that they could be matched by either free - free or graphite dust emission .
the enduring interest of dust formation by some wc type wr stars is the great difficulty of forming dust in such hostile environments : close to the stars , the stellar radiation fields heat any dust to above its evaporation temperature but , at greater distances where the radiation field is sufficiently diluted for the grains to survive in radiative equilibrium , the density in an isotropic wc stellar wind is too low to allow homonuclear dust condensation .
this was recognised by hackwell , gehrz & grasdalen ( 1979 ) in the case of the wc7+o5 type episodic dust - maker hd 193793 (= wr140 ) and discussed by williams , van der hucht & th ( 1987 ) for the wc89 stars making dust persistently .
high - density structures in the wc winds are required to allow dust formation .
two classes of high - density structure are suggested by the observed incidence and phenomenology of dust formation : large - scale structures generated by the compression of the wc wind where it collides with the wind of a luminous companion in a binary system a colliding wind binary ( cwb ) and isolated clumps in stellar winds probably related to those observed spectroscopically by moffat et al .
( 1988 ) , moffat & robert ( 1991 ) and lpine & moffat ( 1999 ) .
usov ( 1991 ) modelled cwbs and suggested that very high compression factors ( @xmath0 ) could be produced in wr140 if the heated and compressed wind was able to cool efficiently .
the ir light curves show that dust formation in wr140 occurs only very near periastron passage in its highly elliptical 8-y orbit ( williams et al .
1990 ) , presumably related to the increase by factor of @xmath1 40 of the pre - shock wind density at the wind - collision region ( wcr ) around this phase ( williams 1999 ) .
such phase - locking occurs in other episodic dust - making systems such as the @xmath2-yr wc7+o system hd 192641 (= wr137 ) ( williams et al .
2001 , lefvre et al .
2005 ) . the connection between dust formation by the wc89 stars and their movement in binary orbits comes from ir images of the dust which show rotating ` pinwheel ' structures around systems like wr104 , interpreted to be dust formed and ejected in a stream to one side of a rotating binary system observed at a low orbital inclination ( tuthill , monnier & danchi 1999 , tuthill et al .
further dust ` pinwheels ' have been observed around other wc89 stars , ( cf .
monnier et al .
2007 ) but binary orbits are required to confirm the picture .
dust formation in clumps in wr winds deduced from their eclipsing the light of the wr star was first suggested independently from spectroscopy of wr104 by crowther ( 1997 ) and from photometry of wr121 and wr103 by veen et al .
compared with earlier observations , the spectrum of wr104 observed in 1996 showed selective obscuration of the high - ionization emission lines , formed in the core of the wc9 wind , allowing crowther to set a stringent limit on the projected size ( @xmath3 ) of the obscuring dust clump using a model of the wr104 wind .
spectra observed by williams & van der hucht ( 2000 , hereafter wh00 ) showed that the strengths of the cii and ciii lines had partly recovered from crowther s 1996 eclipse by mid-1997 , and that they had been near ` normal ' in mid-1995 , a year before it .
tighter timing of a dust eclipse was provided by veen et al .
( 1998 ) , who observed wr121 during and emerging from an 0.8-mag eclipse in 1990 .
the flux took about 10 d. to recover to its normal level .
multi - colour observations showed that the eclipse was deeper at shorter wavelengths , suggesting that it was caused by extinction by dust particles .
a shallower ( 0.4-mag . ) eclipse of wr103 also showed colour changes indicative of varying dust extinction .
this star had previously shown isolated eclipses ( massey , lundstrm & stenholm 1984 ) , but they were not ascribed to dust formation .
another deep eclipse by wr121 in _ hipparcos _ @xmath4 data was observed by veen et al . and marchenko et al .
( 1998 ) , but its time - scale was not well defined because the observations of this star did not cover the whole event .
better defined is the deep ( @xmath5-mag . ) , apparently multiple , eclipse by wr106 observed by kato et al .
( 2002a ) in 2000 in a photometric study running from 1994 to mid-2002 .
the observations were in a single passband , like the @xmath4 data , so interpretation in terms of dust absorption is by analogue with the multi - wavelength observations of eclipses . on a smaller scale , variable
circumstellar extinction by dust was invoked by fahed , moffat & bonanos ( 2009 ) to explain the low amplitude ( 0.06-mag . ) variations of wr76 observed in their two - band ( @xmath6 and @xmath7 ) photometric survey of 20 wc89 stars .
kato et al . ( 2002b ) observed significant ( 2.7-mag . ) variations in the light from wr104 , with the star at its faintest near the time of crowther s spectroscopy , supporting the suggestion that the variations were caused by variable dust extinction .
they found a period of 241 d. in their data , close to that of the pinwheel orbit , and suggested that the dust eclipses were phase - locked to the orbit and caused by some of the dust plume creating the pinwheel entering our line of sight .
further optical photometry of wr104 in the all - sky automated survey ( asas-3 , pojmaski 2002 ) allowed reinvestigation of this suggestion .
a dataset combining the observations by kato et al . with the then available asas-3 observations , duly shifted to allow for the different passbands , showed ( williams 2008 ) no phasing of the optical photometry with the pinwheel period .
the asas-3 is now complete , prompting revisiting the variations of wr104 , and extension of the study to a search for eclipses by other dust - forming wc89 stars .
the prime goal is examination of the incidence of eclipses most of which had previously been discovered serendipitously as a step to investigating the significance of this mode of dust formation amongst wc89 stars as a possible alternative to formation in cwbs , with particular relevance to the wc89 stars without luminous companions .
the asas-3 database is well suited to this , covering the region of the galactic plane occupied by the dust - making wc89 stars and containing observations made over a long time span ( 20012009 ) which , allowing for the gaps between the stars observing seasons , provide the equivalent of @xmath1 5 yr continuous photometry of each . the limiting magnitude is almost 15 and stars as bright as @xmath8 can be observed without saturation ( pojmaski 2002 ) , allowing inclusion of the brighter wc89 stars . as already noted ,
the interpretation of abrupt falls in single - passband photometry as obscuration by dust clumps is by analogue with the multi - wavelength studies of similar objects ; but the present paper will proceed using this assumption , and follow veen et al . (
1998 ) in calling the phenomena eclipses .
such eclipses can tell us about varying dust absorption along a narrow column towards the stars ; in contrast , the ir emission gives the total amount of dust around the stars close enough to be heated by the stellar radiation and re - emit in the ir . in this paper , the designation of stars as dust - makers or dust - free is on the basis of their ir spectral energy distributions ( seds ) .
.stars observed , with number of a - quality observations , median @xmath6 magnitude , standard deviation and aperture ( pixels , 1 pixel = 15 arc sec ) used for the photometry .
[ cols=">,>,<,>,>,<,^",options="header " , ] wr numbers assigned by van der hucht et al .
( 1981 ) [ stars ] magnitudes ( @xmath9 ) as a function of median @xmath6 of the stars in table [ stars ] .
the line has the form of the equation for photometric error as a function of intensity derived by schnurr et al .
( 2009 ) , but fitted to the asas-3 data using the zero point , readout noise and flat - field noise as free parameters.,width=302 ] photometry for 16 dust - making wc89 stars was extracted from the asas-3 all star catalogue . to avoid bias , the accessible dust - free wc89 stars ( excluding wr88 , which is a transitional wn / wc9 object , williams , van der hucht & rauw 2005 ) were added , together with a number of wc67 stars to help characterise the dataset .
the asas-3 observations are flagged by quality ( ` a d ' ) , and only stars having at least 100 a - quality ( best ) data were used for this study .
the photometry is provided through a range of apertures ( 26 pixels , where a pixel = 15 arc sec ) , with a suggested ` default ' aperture chosen on the basis of the object s brightness , smaller for the fainter objects .
we began using the ` default ' apertures but in each case checked the scatter in the photomtery in other apertures ; in a very few cases the scatter was found to less in an aperture other than the default , and the data measured through that aperture were used .
the number of extracted observations , median @xmath6 magnitude , standard deviation and aperture size are given for each star in table [ stars ] .
the relation between the standard deviation and median @xmath6 magnitudes is shown in fig[fvmedsig ] , which is consistent with the corresponding figure , based on very many more stars , in pojmaski ( 2002 ) .
the photometric accuracy , 0.040.20 mag . , is lower than that of some of the studies cited above , but entirely adequate for searching for @xmath10 0.5-mag .
in addition to observational errors , including calibration , the dispersions include the effects of any dust eclipses and also of the low - amplitude intrinsic variation observed in many wc9 stars ( e.g. moffat , lamontagne & cerruti 1986 ; balona , egan & marang 1989 ; fahed et al .
2009 ; van genderen , veijen & van der hucht 2013 ) .
these low - level variations are generally below the threshold of the asas-3 dataset and are not part of the present study .
marks the median of the dataset and the dotted lines are spaced @xmath11 about it ( cf .
table [ stars ] ) .
the six points lying more than 3 @xmath12 below the median are outliers : see text.,width=321 ] for wr50 with appropriate median and errors .
the 2001 fading during @xmath1 30 d. is shown in detail in the inset ; the faint magnitudes in 2002 and 2003 are isolated and considered outliers.,width=321 ] with the exception of wr125 , these are stars which have never shown ir seds characteristic of dust formation .
the episodic dust - maker wr125 ( williams et al .
1994 ) was included in this group because ir photometry in the 2mass ( skrutskie et al .
2006 ) , _ spitzer _ glimpse ( churchwell et al . 2003 ) and _ wise _ ( wright et al . 2010 ) surveys , along with unpublished @xmath13 observations in 2006 and 2007 , showed no recurrence of the 199093 dust - formation episode during the time span of the present study .
it is possible that other stars in this group are unrecognised episodic dust makers observed during quiescence , since only a few wc89 stars have long ir photometric histories .
the light curve for each star was examined . to avoid clutter ,
the photometry was gridded to 1-d .
intervals . besides the dispersion ( @xmath12 ) , which increases with faintness ( cf .
fig.[fvmedsig ] ) , the light curves show a small number ( about one per cent of the observations ) of isolated points having @xmath14 mag fainter than the median @xmath6 .
they are more conspicuous in the light curves of the brighter stars having less scatter ; an example is shown in fig.[wr14all ] .
the light curves near each of the outliers were examined in detail to see if the outliers could represent real eclipses . where the low values were not supported by the preceeding or following observations , they were considered to be artefacts of the observations and unreliable apparently a characteristic of the dataset . with the single exception of wr50
, none of the dust - free stars in table [ stars ] showed evidence for eclipse - like variation .
the light curve of wr50 ( fig.[wr50comp ] ) shows a well defined fading of 0.5 mag . over @xmath1 30 d. in 2001
this was followed by a gap in the observing for @xmath1 40 d. , after which the flux had returned to its previous value .
short time scale , low amplitude photometric variation in 1988 was observed by van genderen , larsen & van der hucht ( 1990 ) and nightly fading in 1989 reported by van genderen et al .
( 1991 ) ; but reinvestigation of the latter by veen & wieringa ( 2000 ) found that most , if not all , of this variation was not real . from ir observations in the 2mass , glimpse and _ wise _ surveys , there is no evidence for dust emission in 200110 but this does not rule out the possibility that it is an episodic dust - maker observed during a long period of quiescence , like wr125 .
the spectrum of wr50 shows absorption lines ( van der hucht et al .
1981 ; smith , shara & moffat 1990 ) so wr50 is most probably a binary but apparently lacks an orbit . for wr106 ; in this case the @xmath12 for the dotted lines ( 0.06 mag . ) was that read from the dispersion curve ( fig.[fvmedsig ] ) for the brightness of wr106.,width=321 ] the asas-3 light curve ( fig.[wr106all ] ) of wr106 shows deep eclipses in 2003 and 2007 , the second eclipse in 2007 having a double minimum separated by @xmath1 46 d. with perhaps a third , lower amplitude , one between them ( fig.[ec106 ] ) .
this invites comparison with the complex eclipse ( in sparser data ) observed in 2000 by kato et al .
( 2002a ) , having at least two components , with minima separated by @xmath1 56 days , but a larger amplitude ( @xmath5 mag ) , than that of the 2007.7 minimum ( @xmath15 mag . ) .
kato et al .
( 2002a ) also reported the marginal detection of a shallower ( @xmath1 0.6-mag . )
eclipse in 2002 may ; unfortunately , the asas-3 observations in 2002 began in mid - june so this event can not be compared in the two datasets . a possible low - amplitude ( @xmath1 0.4-mag . )
eclipse in the asas-3 data could be the two fainter magnitudes in 2004.2 which were observed within 4 d. of each other and are therefore probably not observational outliers ; the flux was near normal 2 d. before the first observation and again 4 d. after the second , making it a very brief event .
the other low points in 2006 and 2008 in the light curve are isolated outliers .
the kato et al .
( 2002a ) and asas-3 datasets overlap in 20012 , so that the two taken together provide coverage from 1994.1 to 2009.8 , but with a number of gaps , mostly seasonal . to estimate the completeness of coverage with respect to discovery of deep eclipses like those observed
, we note that the 2003 eclipse and the first in 2007 each lasted about 60 d. , so that at least part of such an eclipse would still be observed if there was a gap of no more than 0.1 yr . in the observations . omitting the larger gaps in coverage
, we then have the equivalent of 9.9 yr of continous photometric monitoring during which six deep minima , four of them in close pairs , were observed .
if the six minima were independent events , their distribution is far from poissonian .
for example , if we divided the 9.9 period into 33 intervals of 0.3 yr , an interval long enough to include the double eclipses but not so long as to include both the single and double eclipses in 2007 , we have a rate of 6/33 = 0.182 eclipses per interval .
a poissonian distribution would have 27.5 intervals with no eclipse , 5.0 with one eclipse , 0.45 with two eclipse , etc .
far from what is observed .
the numbers are too small for a conclusion , but this suggests that the double eclipses are integral events rather than the close proximity of independent single eclipses .
this suggestion is supported by the apparent minor eclipse ( depending on two observations separated by 2 d. ) between the two components of the 2007 double eclipse , and would be strengthened if the apparent eclipses on the rising branches of the 2003.7 and 2007.2 eclipses were real ( fig.[ec106 ] ) .
the 2003 and the first 2007 eclipses show a slower recovery than fading , 36 and 44 d. , compared with fadings of 0.6 mag in 17 and 10 d. respectively .
the recovery from the double eclipse was much faster , 17 d. to rise by 0.9 mag .
if we interpret the rise times as the lifetimes of the dust clumps as they rise through the stellar wind in our direction , the simplest explanation of the shorter lifetimes is that the trajectories of the clumps make a small angle to our line of sight and they pass out of it sooner than if the trajectories were exactly aligned to our line of sight .
the particular interest of formation of dust in clumps by wr106 is that it is not obviously a cwb : wh00 searched for , but did not observe , balmer absorption lines attributable to an ob companion to the wc9 star in its spectrum .
if wr106 is not a cwb , its circumstellar dust must be formed in some other process .
is _ not _ the median , but an estimate of the uneclipsed flux from the light curve ; and the @xmath16 used for the dotted lines spaced @xmath11 about it is read from the dispersion curve for @xmath17 .
, width=321 ] , and the expected observational error read from the dispersion curve is marked in the corner.,width=302 ] , the faint lines through the points are drawn only to guide the eye.,width=302 ] the asas-3 light curve of wr104 ( fig.[wr104all ] ) is reminiscent of that observed by kato et al .
( 2002b ) during 19942001 in showing almost continuous variation , and it differs significantly from all the other stars observed in not having long runs of apparently constant flux .
evidently , wr104 suffers more eclipses than any other wc9 star observed in this or previous studies , apparently an almost continuous string of eclipses having a range of depths , and sometimes overlapping .
another possible explanation for the variations is variable extinction caused by some of the dust formed in the colliding - wind system entering our line of sight , as proposed by kato et al .
( 2002b ) .
they found a period equal to the ` pinwheel ' rotation period in their photometric data but , as noted above , this was not supported in a preliminary study using their photometry and some early asas-3 data ( williams 2008 ) .
it is now possible to repeat the period search using the whole asas-3 dataset for wr104 , 617 observations compared with 146 used by kato et al .
the datasets were not combined this time as the kato et al .
data are of lower accuracy ( 0.20.3 mag . ) than the asas-3 data , and are in a different photometric band ( unfiltered 400600 nm ) .
a search for periods in the range 200300 d. using phase dispersion minimization ( pdm , stellingwerf 1978 ) showed no evidence for a period near that ( 241.5 d. , tuthill et al .
2008 ) of the ` pinwheel ' rotation .
the data were then phase - folded with the 241.5-d . period ( fig.[ph104 ] ) , giving the pdm statistic @xmath18 , and do not show the large - scale phased variation of extinction found by kato et al .
the latter used 15 phase bins for their phased light curve in contrast to the 50 used here , which seems the minimum number needed to capture the variations of wr104 .
inspection of the light curve shows well - defined changes of 0.10.2 mag .
, compared with 0.1 mag . for the expected observational error , on the time scale of the p/50 phase bins ,
so wider phase bins would smother these variations .
phased light curves using narrower binning were also investigated , but none showed systematic variation .
the large dispersions seen in some of the phase bins is a consequence of stochastic occurence of deep eclipses independent of the orbital period .
there are so many eclipses that the median @xmath6 is not a good measure of the undisturbed flux level , which is estimated to lie near @xmath17 from inspection of the light curve . omitting gaps in the observing exceeding 0.1 yr
, we have the equivalent of 4.8 yr of continuous photometric coverage , during which wr104 was brighter than @xmath19 ( 1 @xmath12 below the estimated undisturbed level ) for only about 0.64 yr , implying that eclipses were occurring for seven - eighths of the time . from the light curves ,
14 eclipses were identified ; some completely observed , and others inferred from well - defined but isolated brightening or fading at the ends of observing runs , of which three had amplitudes in the range 1.52.0 mag .
, eight in 1.01.4 mag .
, and three ( probably an underestimate ) in 0.50.9 mag . , giving an average rate of 3.4 yr@xmath20 , also probably an underestimate .
the light curves show much wider minima than the eclipses shown by wr106 , but this might be caused by the partial overlapping of individual events .
for example ( fig.[ec104 ] ) , there was a reasonably well defined eclipse in 2005 , but its recovery appears to have been interrupted by another eclipse , so estimation of the recovery time and clump lifetime comes from measuring the brightening rates , giving @xmath1 40 d. in 2007 , there was an even broader minimum , but it is not obvious whether this was caused by a single clump or several the data in the second half of 2007 are very sparse , with a gap of 65 d. between the last two observations , and one of 110-d . between the last and the first observations in 2008 . during 2008 ,
the light curve looks like noise , but closer examination shows @xmath1 1-mag eclipses in 2008.4 and 2008.6 .
again it was not possible to measure recovery times or clump lifetimes as the recoveries were interupted .
gaps in the coverage make other events harder to characterise , but the steady rise of 0.9 mag . in 0.18
yr in 2002 , recovery from an earlier eclipse , suggests a clump lifetime of at least 66 d. for wr69 with appropriate median and errors .
the 2003.5 eclipse is shown in the inset ; the other low points are all isolated outliers.,width=321 ] the photometric dispersion in wr69 seems anomalously high for its brightness ( cf .
fig.[fvmedsig ] ) , and the light curve ( fig.[wr69comp ] ) shows rather more outliers than the others .
the dip in mid-2003 comes from seven observations ( some gridded ) and is probably a real , low - amplitude , eclipse .
inspection of all the other outliers showed that they were isolated observations .
low - amplitude ( @xmath21-mag . ) variability has been observed by balona et al .
( 1989 ) , and was found in _ hipparcos _
@xmath4 data by marchenko et al .
( 1998 ) and koen & eyer ( 2002 ) , with different periods , but phasing the asas-3 photometry to both of these periods did not reveal any systematic variation .
spectroscopy showed the wc9 star to have an early o - type companion ( wh00 ) , and a subsequent observation showed the lines in the two stars to have moved relative to each other , confirming wr69 as a spectroscopic binary ( williams et al .
2005 ) .
for wr117 with appropriate median and errors.,width=321 ] , the faint lines through the points are drawn only to guide the eye.,width=321 ] for wr95 with appropriate median and errors.,width=321 ] another star showing slightly high dispersion for its magnitude is wr117 , whose light curve ( fig.[wr117all ] ) shows a fair amount of scatter and a possible 0.5-mag .
eclipse defined by four points in 2005.4 .
this is shown at higher scale in fig.[ec11795 ] .
the light curve of wr95 ( fig.[wr95all ] ) shows a 0.6-mag .
eclipse in 2008.6 ( fig[ec11795 ] ) , with a slow recovery ( @xmath1 40 d. ) ; and a possible weak ( 0.3-mag . ) eclipse in 2002.3 .
no dust eclipses were identified in the light curves of wr103 or wr113 , which had shown eclipses in the past ( see above ) , based on the equivalents of 3.8 and 5.2 yr of continuous monitoring . phasing the photometry of wr113 ( cv ser ) , omitting the 7 out of 733 magnitudes identified as isolated outliers , to the elements of david - uraz et al .
( 2012 ) clearly showed a broad ( @xmath22 p ) 0.02-mag .
atmospheric eclipse at phase 0.5 , matching their observations and giving a measure of the quality of the asas-3 photometry of a bright star .
david - uraz et al .
suggested that scattering by dust in the wind of the wc8 component could be a significant contributor to the depth of the eclipse , but this idea suffers from the problem that any dust would also absorb stellar radiation which would heat it to evaporation , if too close to the star . from the @xmath23 measured by niemela et al .
( 1996 ) and @xmath24 given by david - uraz et al .
, the line of sight to the o star can be seen to pass within @xmath25 of the wc8 star .
this is very much closer to the star than the minimum distance at which carbon grains can survive the radiative heating , the inner edge of the wr113 dust cloud ( @xmath26 , where @xmath27 is the radius of a single star having the same luminosity as the wc8+o binary ) modelled by williams et al .
( 1987 ) , or the corresponding distances , 2030 au , observed and modelled for the wc9+b0 system wr104 by tuthill et al .
( 2008 ) and harries et al .
( 2004 ) , so it is most unlikely that dust can play a part in the atmospheric eclipses .
no well defined eclipse was shown by the third star to have shown eclipses in the past , wr121 , during the equivalent of 3.8 yr of continuous monitoring . two isolated points 0.2 mag .
below median observed on successive nights , supported by a b - quality observation four nights earlier , at the end of the 2008 observing season might point to an eclipse but the coverage was too sparse to be certain .
none of the other eight dust - makers showed evidence for an eclipse in their light curves or , indeed , any significant variation .
apart from wr104 and wr106 , in which at least 14 and 6 eclipses with amplitudes greater than 0.5 mag . were observed in the equivalents of 4.8 and 9.9 yr of continuous monitoring respectively , the other 14 dust - making wc89 stars accounted for about four eclipses of lower amplitude in the equivalent of @xmath28 yr of continuous monitoring .
this very great discrepancy in observed eclipse rates suggests a significant difference between the winds of wr104 and wr106 and those of the other wc89 stars observed : greater than the high end of a continuous distribution of properties .
an alternative to the very uneven distribution of clump formation in wc9 winds is that all the stars form clumps at rates comparable to wr104 and wr106 , but that the clumps are confined to narrowly collimated streams , and that only these two stars are favourably inclined to the observer . in this picture , the streams of clumps from the stars showing occasional eclipses have trajectories making slightly larger angles with our line of sight .
the problem with this explanation is that it runs contrary to the spectroscopic study of the radial velocities of clumps in wr winds , which show their trajectories to have a range of angles to the line of sight ( moffat & robert 1991 ; robert 1994 ; lpine & moffat 1999 ) , and that it is most unlikely that the two sorts of clump would be formed in such different ways .
the eclipse rates can be compared with the circumstellar dust clouds , as measured by the fractions of the stellar ultraviolet
visible flux which they absorb and re - radiate in the ir . for wr104 and wr106 , these are about 60 and 23 per cent respectively , compared with 110 per cent ( williams et al .
1987 ) for the other dust makers in this study . amongst the latter , wr80 and wr96 re - radiate approximately 9.5 per cent of their stellar flux in the ir but have not shown eclipses , whereas the stars showing a few eclipses , wr69 , wr95 , wr117 and perhaps wr121 have re - radiation rates between 1.7 and 5.8 per cent , and wr50 is not a known dust maker ; so the correlation is not strong .
unfortunately , the other stars shown by williams et al .
( 1987 ) to have high ir re - radiation rates , wr48a , wr76 and wr118 , are all too faint for the present study ( fahed et al .
( 2009 ) did not observe any eclipses by wr76 in their study , but the photometry covered only 18 d. ) fainter wc89 dust makers , including wr48a and wr76 , should be observable in the currently progressing surveys such as asas-4 and ogle - iv ( cf .
pietrucowicz et al .
the photometric coverage of at least some of the stars in the present study should be extended to improve our knowledge of the incidence of eclipses in stars like wr121 and to investigate the possible double or multiple eclipses apparently shown by wr106 . to study the connection between the dust formed in clumps and that making up the circumstellar dust reservoirs responsible for the ir emission , consider what happens to the dust .
the clump apparently starts close to the wc star ( crowther 1997 , veen et al .
1998 ) and moves out with the wind .
the dust condenses quickly , judging by the fading times of the eclipses .
as the dust absorbs stellar radiation , it heats up and re - radiates in the ir until either it gets so hot that the dust evaporates , or the clump has moved away from the star sufficiently quickly that the geometric dilution of the stellar radiation heating it is no longer enough to heat it above evaporation temperature . by the end of the eclipse , either the dust has been destroyed , or the dust has survived but the clump has dissipated sufficiently that it no longer provides enough visual absorption to produce a measurable eclipse but , of course , continues absorbing and re - radiating in the ir .
there are two time - scales : the kinematic time for the dust to reach a safe distance ( e.g. about 25 au in the case of wr104 from the modelling by harries et al .
2004 ) which would be @xmath29 d. at an average speed of 1200 km s@xmath20 , the terminal velocity of the wind .
this may be optimistic , the initial acceleration of clumps is rather slow ( e.g. lpine & moffat 1999 ) , but radiation pressure will accelerate the dust to move faster than the wind in which it condensed .
the other time - scale is that for the dust being heated in the stellar radiation field to reach its evaporation temperature .
this needs to be modelled , particularly the properties of the original clump and the shielding it must provide to the condensing and newly formed dust which is beyond the scope of the present study .
we can see , however , that the kinematic time - scale is comparable to the recovery times of eclipses , so it seems possible that dust formed in clumps is surviving long enough to reach the ` safe ' distance and contribute to the circumstellar dust reservoir .
simultaneous optical and ir observations during a dust eclipse would be invaluable in measuring the brightening in the ir and temperature of the dust clump as it moves out in the stellar wind .
one star , wr70 was observed frequently in the ir during the present study , varying with a range of @xmath30 ( williams et al .
2013 ) , but its asas-3 @xmath6 magnitudes showed no eclipses or other variation , so the variations in dust emission must be ascribed to changes in its wc9+b0i wind - collision system . the absence of variable extinction associated with the orbital motion in wr104 is unsurprising considering the geometry . we know from the spectroscopy in 199597 that dust was in front of the wc9 star and not the ob companion .
the winds collide between the stars , and the dust is formed further from the wc9 star , beyond the ob companion , in a well - collimated beam ( opening angle @xmath31 , cf .
harries et al .
2004 ) in the orbital plane , which lies almost in the plane of the sky given the low inclination , @xmath32 . only in an inclined system like wr140 ,
is some of the cwb dust modelled to pass in front of the star ( williams et al .
2009 ) and cause the minor eclipses observed shortly after periastron passage ( marchenko et al .
2003 ) , when the cwb dust formation was briefly active .
the high eccenticity of its orbit results in the projected position angle of the wind - collision system swinging through more than @xmath33 on the sky in only @xmath34 around periastron and the dust to be spread very unevenly around the stars .
the ir images of wr104 ( tuthill et al .
2008 ) can be used to infer the relative importance to the circumstellar dust reservoir of formation in clumps and formation in the wc9+ob wind - collision system responsible for the dust pinwheel .
if dust made in clumps responsible for the eclipses is significant , we would expect to see evidence in the form of extended , possibly clumpy , diffuse emission centred on the star in the ir images .
the very observation of pinwheel structures , however , demonstrates that most of the dust emission is in the pinwheel ; the composite 2-@xmath35 m image ( tuthill et al . 2008 , fig.8 ) shows no dust emission above 0.1 per cent of the peak between the arms .
this demonstrates that dust formation in clumps is not an important contributor to the total dust being formed by wr104 either because comparatively little dust is being made in this process , or because the dust in the clumps does not survive long enough to reach the ` safe ' distance . if instead of isotropic clump formation
, we consider that wr104 owes its high rate of observed eclipses to a collimated beam and our favoured viewing angle , we might expect the ir images to show a bright spot in the centre : emission by a column of surviving clump dust .
there is no evidence for this ; on the contrary , the images show a ` standoff distance ' of 13 mas ( tuthill et al .
2008 ) , the ` safe ' distance above , between the brightest dust pixel and the spiral centre , interpreted as the stars , where no emission is seen . in either case
, it is evident that the star having by far the highest rate of eclipses is not receiving a significant contribution to its circumstellar dust cloud by this process .
we need to know more about the wr104 eclipses from multi - colour photometry over a wide wavelength range to make sure that the extinction really is by dust and to measure its properties .
large amplitude ( @xmath36 ) variability was found in the first results from the bochum southern galactic disk survey ( haas et al .
2012 ) and the combination of such photometry with contemporaneous @xmath6 in the asas-4 survey over the next few years would be valuable . also , wr104 would be a good candidate for simultaneous ir and optical photometry to determine the behaviour of the dust on the recovery branches of the eclipses .
of the 16 dust - making wc89 stars surveyed , most of the eclipses were observed from just two stars , wr106 and wr104 , while the other 14 stars showed only 34 eclipses , none as deep as those in wr104 or wr106 .
it is evident from the ir images of wr104 that any dust formed in the eclipses is not a significant contributor to its circumstellar dust cloud .
the implication of this is that dust formed in occasional eclipses by the other wc89 dust makers is also an insignificant contributor to their dust clouds , so that the question of how wc9 stars which are not members of cwbs make their dust still remains open .
the two stars with the highest eclipse rates also have the most luminous dust clouds in the ir and , although dust formed in clumps may not contribute to the circumstellar dust directly , cwbs having higher rates of clumps in the wc9 winds might be more efficient dust makers .
this work is based on the all sky automated survey all star catalogue and it is a pleasure to thank grzegorz pojmaski for helpful correspondence . it has made use of the nasa / ipac infrared science archive , which is operated by the jet propulsion laboratory , california institute of technology , under contract with the nasa , and the simbad and vizier databases , operated at the cds , strasbourg .
i am grateful to the institute for astronomy and uk astronomy technology centre for continued hospitality and access to the facilities of the royal observatory edinburgh .
moffat a. f. j. , robert c. , 1991 , in in van der hucht k. a. , hidayat b. , eds , iau symp .
143 , wolf - rayet stars and interrelations with other massive stars in galaxies , kluwer academic publishers , dordrecht , p. 109
van genderen a. m. , et al .
, 1991 , in van der hucht k. a. , hidayat b. , eds , iau symp .
143 , wolf - rayet stars and interrelations with other massive stars in galaxies , kluwer academic publishers , dordrecht , p. 129
williams p. m. 1999 , in : van der hucht k. a. , koenigsberger g. , eenens p. r. j. , eds , proc .
193 , wolf - rayet phenomena in massive stars and starburst galaxies , astron .
pacific , san francisco , p. 267
williams p. m. , et al . , 2001 ,
mnras , 324 , 156 williams p. m. , van der hucht k. a. , rauw g. , 2005 , in : rauw g. , naz y. , blomme r. , gosset e. , eds , massive stars and high - energy emission in ob associations , a workshop of the jenam 2005 , ` distant worlds ' , lige 2005 , p. 65 | visual photometry of 16 wc89 dust - making wolf rayet ( wr ) stars during 20012009 was extracted from the all sky automated survey all star catalogue ( asas-3 ) to search for eclipses attributable to extinction by dust formed in clumps in our line of sight .
data for a comparable number of dust - free wc69 stars were also examined to help characterise the dataset .
frequent eclipses were observed from wr104 , and several from wr106 , extending the 19942001 studies by kato et al .
( 2002a , b ) , but not supporting their phasing the variations in wr104 with its ` pinwheel ' rotation period .
only four other stars showed eclipses , wr50 ( one of the dust - free stars ) , wr69 , wr95 and wr117 , and there may have been an eclipse by wr121 , which had shown two eclipses in the past .
no dust eclipses were shown by the ` historic ' eclipsers wr103 and wr113 . the atmospheric eclipses of the latter
were observed but the suggestion by david - uraz et al .
that dust may be partly responsible for these is not supported . despite its frequent
eclipses , there is no evidence in the infrared images of wr104 for dust made in its eclipses , demonstrating that any dust formed in this process is not a significant contributor to its circumstellar dust cloud and suggesting that the same applies to the other stars showing fewer eclipses .
[ firstpage ] stars : wolf rayet stars : individual : wr104 , wr106 . |
whether protons or electrons dominate particle acceleration and non - thermal energy transport in solar flares is still subject of recent debate [ e.g. see _ great debate _ articles by cargill ( 1996 ) , simnett ( 1996 ) , and emslie ( 1996 ) ] .
moreover , proponents of proton - dominated models argue that even the @xmath4 kev hxr bremsstrahlung is produced either by protons directly ( heristchi 1986 ; 1987 ) or by proton - energized electrons ( simnett & haines 1990 )
. the purpose of this _ letter _ is to complement these ongoing debates with powerful new arguments that are supported by observations of energy - dependent hard x - ray ( hxr ) timing delays in solar flares .
electrons have been favored for the interpretation of impulsive @xmath4 kev hxrs because the inversion of the hxr spectrum ( brown 1971 ) in the framework of energy loss by collisional bremsstrahlung ( in a relatively cold target ) could easily reproduce : ( 1 ) an electron spectrum that contains the necessary and sufficient energy content to explain the non - thermal hxr emission , ( 2 ) the gyrosynchrotron emission observed at microwave wavelengths , and ( 3 ) the soft x - ray emission resulting from the heated , upflowing chromospheric plasma .
more detailed arguments for the case of electron dominance in solar flare energy transport can be found in emslie ( 1996 ) , emslie et al .
( 1996 ) , and references therein . protons , on the other hand , have been invoked as primary energy carriers ( simnett 1986 ; 1995 ) mainly because it minimizes the number of energy - carrying particles ( thanks to their intrinsically higher rest mass , i.e. @xmath5 ) , which in turn also minimizes return currents ( at least for high - energy protons with @xmath6 kev , see , e.g. , simnett & strong 1984 , emslie & brown 1985 ) .
observational support for the proton hypothesis was mainly gathered by indirect effects , e.g. ( 1 ) acceleration mechanisms ( e.g. shocks ) can have greater efficiencies for protons because of their larger gyroradii , ( 2 ) the proton spectrum at energies @xmath7 mev could be considerably steeper than previously supposed ( ramaty et al .
1995 ) , based on recent measurements of share & murphy ( 1995 ) , ( 3 ) detection of interplanetary high - energy ( @xmath8 mev ) protons ( evenson et al .
1984 ) , etc .
a discussion of arguments in favor of protons can be found in a recent review by simnett ( 1995 ) , and further arguments of proton versus electron beam models are discussed in brown et al .
( 1990 ) .
a new aspect that we would like to contribute to the electron - versus - proton debate is based on new timing results from hxr observations in solar flares , analyzed in a series of recent studies ( aschwanden , schwartz & alt 1995 ; aschwanden & schwartz 1995 ; 1996 ; aschwanden 1996 ; aschwanden et al . 1996a ; 1996b ; 1996c ) .
the energy - dependent timing delays @xmath9 of fast ( @xmath0 s ) _ hxr pulses _ during the impulsive flare phase have been found to be consistent with electron time - of - flight differences , while the timing delays of the _ smooth hxr _ flux are comparable with electron trapping time scales . the interpretation of hard x - ray time structures in terms of these two components ( associated with directly - precipitating versus trapped particles ) is a basic assumption used in the following discussion , in contrast to the widely - held hypothesis that the lower envelope of the rapidly - varying hard x - ray flux consists of an unresolved pileup of elementary pulses with similar characteristics as the rapid fluctuations ( e.g. kiplinger et al .
1983 , 1984 ; machado et al . 1993 ) .
while protons have not been considered in our previous studies , we would like to explore in this _
letter _ whether proton theories could be reconciled with the observed hxr timing delays , and what the required proton energies would be . because the timing delays of hxrs are directly related to the particle velocities
, timing information provides a direct means of constraining the kinetic energy and rest mass of the involved particles . in section
2 we compare theoretical electron and proton time - of - flight differences with the observed delays of fast _ hxr pulses_. in section
3 we discuss theoretical trapping times in solar flare loops and compare them with the observed delays of the _ smooth hxr _ emission . section
4 contains conclusions on the question of which species of particles is more likely to reproduce the observed hxr timing delays .
both proponents of electron - dominated and proton - dominated flare models agree with the assumption that particles are accelerated in the coronal part of a flare loop and that the hxrs observed from the loop footpoints are produced by collisional bremsstrahlung in the chromosphere .
the main difference between the two models is the question of the particle species that carries the energy from the acceleration site to the energy loss site . in the case of proton - dominated models , either the protons produce hxr s directly by inverse bremsstrahlung ( emslie & brown 1985 ; heristchi 1986 ; 1987 ) , or a neutral beam is assumed to precipitate to the chromosphere , where an electric field arising from charge separation accelerates hxr - emitting electrons ( simnett & haines 1990 ) . in the latter model , electrons are thought to be stripped by coulomb scattering from a neutralized ion beam as it impacts the chromosphere .
this charge separation process may set up an electric field that is able to produce runaway acceleration of electrons responsible for the @xmath4 kev hxr bremsstrahlung .
this model has been severely criticized by emslie ( 1996 ) .
thus , the timing of the two models can be tested in first order by comparing the particle time - of - flight differences occurring during the propagation from the coronal acceleration site to the chromospheric energy loss site .
if the energy - dependent timing delays of the hxrs is dominated by such propagation effects , the hxr emission produced by the low - energy particles should always be delayed with respect to the high - energy particles , because of their lower velocity .
energy loss times and acceleration times seem not to dominate the timing delays of ( @xmath0 s ) hxr pulses in most of the flares , because they do not reproduce the correct sign in the energy - dependent timing delays ( aschwanden et al . 1996a ; aschwanden 1996 ) .
thus , the time - of - flight difference @xmath10 between two particles @xmath11 and @xmath12 arriving at the chromospheric energy loss - site is determined by their velocities @xmath13 ( or @xmath14 ) and the path length @xmath15 , @xmath16 the relativistic velocity @xmath17 is defined by the kinetic energy @xmath18 of the particles and the lorentz factor @xmath19 , i.e. @xmath20^{-2}}\ ] ] if we compare electrons with a kinetic energy @xmath21 and rest mass @xmath22 with protons of kinetic energy @xmath23 and rest mass @xmath24 , the velocity ratio is @xmath25^{-2 } } \over \sqrt { 1 - [ { e_e \over m_e c^2 } + 1]^{-2 } } } \approx \sqrt { { e_p \over e_e } \cdot { m_e \over m_p } } = { 1 \over 43 } \sqrt{e_p \over e_e } \ , \ ] ] where the right - hand approximation applies to the non - relativistic ( or mildly relativistic ) case .
for instance , to obtain identical velocities for electrons and protons , the kinetic energy of the protons has to be a factor of @xmath26 larger , e.g. 40 mev protons have the same velocity as 20 kev electrons . in the case of protons
, it was proposed that flare hxrs could be produced by an `` inverse bremsstrahlung process '' ( boldt and serlemitsos 1969 ; emslie & brown 1985 ; heristchi 1986 ; 1987 ) .
protons would produce hxr photons by collisional bremsstrahlung in collisional encounters with electrons .
the relative velocity of a proton of energy @xmath27 in an encounter with an ambient electron is the same as that of an electron of energy @xmath21 in an encounter with a stationary proton .
since @xmath8 mev protons have identical velocities as @xmath4 kev electrons , they would also show the same time - of - flight differences as observed .
however , the major problem with @xmath8 mev protons is that the required proton flux to produce the observed @xmath4 kev hxr flux would exceed the proton flux deduced from nuclear gamma - ray line observations ( at 2 - 10 mev ) by a factor of @xmath28 ( ramaty 1985 ) . to alleviate the problem of @xmath29 mev protons , simnett & haines ( 1990 ) proposed a neutral beam model , that requires only 0.1 - 1 mev protons to produce the observed @xmath4 kev hxr flux . after a neutral beam arrives at the chromosphere , hxr emission
is produced by secondary electrons accelerated in the electric field that arises from the charge separation of penetrating ions and thermalized primary electrons .
thus , the timing delays of hxr emission are mainly determined by the coronal propagation time of the 0.1 - 1 mev protons in the neutral beam .
however , these lower proton energies correspond to velocities that would produce far greater time - of - flight differences than for 40 mev protons or 20 kev electrons . according to eq.3 , the velocity of 1 mev protons would be a factor of @xmath30 smaller than for 20 kev electrons , and thus , the time - of - flight differences would be 10 times larger than for 20 kev electrons or 40 mev protons .
fig.1 shows typical hxr time delay measurements in the energy range of 40 - 300 kev for a major flare .
the _ pulsed hxr _ flux ( fig.1 bottom right ) exhibits time delays of @xmath31 kev)@xmath32 ms .
these delays show the correct sign as expected for particle time - of - flight differences , i.e. the lower energies lag the high energies .
also the functional form of the energy - dependent time delay is consistent with the expected model ( eq.1 , thick line on right - hand side bottom in fig.1 ) , within a high accuracy of @xmath33 ms . for the conversion of electron energies
@xmath34 into hxr photon energies @xmath35 , a conversion factor of @xmath36 was determined ( see second line on right - hand side from bottom in fig.1 , according to calculations in aschwanden & schwartz 1996 ) .
moreover , the time - of - flight distance , i.e. @xmath37 km , is found to be close to a loop half length ( @xmath38 km ) , if electron velocities are employed .
this ratio @xmath39 was established to be scale - invariant for a comprehensive set of loop sizes in different flares ( aschwanden et al .
1996b , 1996c ) .
if we employ protons to reproduce these observed hxr delays , we can either adjust the energy range or the time - of - flight distance .
if we adjust the energy range , protons with energies of @xmath8 mev would be required to produce matching velocities .
if we employ 1 mev protons , the velocities would be 10 times smaller , thus requiring a time - of - flight distance 10 times shorter to satisfy the same timing delays ( eq.1 ; also depicted in fig.1 middle right ) . based on the statistics of flare loop sizes ( typically with loop radii of @xmath40 km , aschwanden et al .
1996c ) , a time - of - flight distance 10 times smaller ( @xmath41 km ) would be comparable with the height of the chromosphere , requiring an acceleration site located near the transition region .
such a scenario would have difficulty to explain the observed simultaneity of hxr emission in conjugate loop footpoints , which were found to be coincident within @xmath42 s ( sakao 1994 ; kosugi 1996 ) , but are separated by considerably larger electron travel times of 0.5 - 1.0 s. therefore , it seems that all proton models face some fundamental difficulties to reconcile the observed timing delays of impulsive hxr emission .
the bulk of hxr emission ( @xmath43 , depending on the modulation depth of hxr pulses ) is contained in smooth , slowly - varying time structures , which moreover exhibit an opposite sign of the energy - dependent time delay , compared with the fast - fluctuating hxr pulses ( aschwanden et al . 1996b ; 1996c ) .
the timing delays of this _ smooth hxr _ component was therefore interpreted in terms of particle trapping or energy loss inside the flare loop .
this interpretation is mainly motivated by two reasons : ( 1 ) trapping is expected to smear out rapid fluctuations of the acceleration and injection process , and ( 2 ) the collisional deflection time or collisional energy loss time shows a functional dependence ( @xmath44 ) that reproduces the correct sign and fits the functional form of the observed delay @xmath45 ( see example in fig.1 , bottom left ) .
let us consider the collisional deflection time , which in any case should represent an upper limit of the particle trapping time in a flare loop .
the collisional deflection time is defined by ( e.g. benz 1993 ) , @xmath46\ ] ] with @xmath47 the kinetic energy , @xmath48 the rest mass , and @xmath49 the charge of the test particles , @xmath50 the electron density of the ambient plasma , and @xmath51 the coulomb logarithm .
if we employ @xmath52 kev electrons ( to produce @xmath53 kev hxr photons ) , the expected deflection time is about @xmath54 s in a flare loop with a typical electron density of @xmath55 @xmath56 . considering the fit of @xmath57 to the observed time delay for the flare shown in fig.1 ( bottom left )
, we find that the _ smooth hxr _
flux has timing delays that are consistent with a trapping model in terms of collisional deflection .
if we employ protons , the collisional deflection times would be a factor of @xmath58 longer than for electrons , which is far beyond any delay ever measured in hxrs or gamma rays . the collisional deflection time , though an upper limit , is therefore no reasonable estimator of trapping times for protons .
pitch - angle scattering by strong wave turbulence or cross - field diffusion times in twisted magnetic flux tubes are believed to restrict the trapping times of ions or protons in flare loops to time scales of @xmath59 s ( ramaty & mandzhavidze 1994 ) , which are a factor of @xmath60 longer than collisional deflection times of electrons ( @xmath61 s in typical flare loop densities ; aschwanden et al . 1996a ; 1996b ) . if we employ protons to generate the @xmath62 kev hxr emission ( e.g. simnett & haines 1990 ) , the precipitating protons are necessarily required to have an identical timing as the observed hxrs .
this requirement for the proton timing can not be reconciled with the observed timing delays of the smooth hxr flux for the following reasons : ( 1 ) if the smooth hxr flux would be produced by directly precipitating protons , the proton timing would have an opposite sign because of time - of - flight differences ; ( 2 ) if the smooth hxr flux would be produced by initially trapped and subsequently precipitating protons , the proton trapping times would be much longer ( a factor of @xmath60 ) than the observed hxr delays , which were found to fit electron collisional deflection times ( fig.1 , bottom left ) . in the latter case
, it would also be very difficult to understand that the trapping times of @xmath63 mev protons would have exactly the same energy - dependence as the collisional deflection time of @xmath4 kev electrons ( see fig.1 , bottom left ) .
thus , we see no suitable possibility to adjust the proton timing to the observed hxr delays , neither in terms of time - of - flight differences nor in terms of trapping time differences .
from previous work we found that the energy - dependent timing delays @xmath45 of solar flare hxr emission in the energy range of @xmath64 kev are consistent with two timing processes : ( 1 ) the timing delays of the _ pulsed _ hxr flux are consistent with electron time - of - flight differences over spatial scales that correspond roughly to half a loop length , and ( 2 ) the timing delays of
the _ smooth _ hxr flux are consistent with collisional deflection times for electrons in typical flare loop densities ( @xmath65 @xmath56 ) . in an attempt to explain the same timing results of _ hxr pulses _ in terms of protons
, we considered two energy ranges , which both face fundamental difficulties : ( 1 ) @xmath2 mev protons would match the same velocities and time - of - flight differences , but would require a proton flux far in excess of gamma - ray line observations , and ( 2 ) @xmath1 mev protons have a velocity @xmath66 times smaller than 20 kev electrons and can only satisfy the observed time - of - flight differences for flight distances @xmath66 times smaller , which are difficult to reconcile with the observed simultaneity of conjugate hxr footpoint emission . employing protons to explain the observed timing delays of the _ smooth _
hxr flux would require that the trapping times of protons have the same energy dependence as collisional deflection times for electrons .
this conflicts with estimates of proton trapping times , which are significantly longer than those of electrons , in the case of collisional deflection as well as for pitch - angle scattering by wave turbulence .
based on these timing arguments it appears to be very unlikely that protons are responsible for producing @xmath4 kev hxr bremsstrahlung , e.g. as proposed by simnett & haines ( 1990 ) .
however , these timing results do not exclude that an arbitrary amount of ions ( or protons ) are accelerated concomitantly with electrons and propagate to the chromosphere , as evidenced by the observed nuclear gamma - ray lines ( at 1 - 10 mev ) in large flares .
the main conclusion of this study is that * protons can not be responsible for @xmath4 kev hxr emission * , because their timing would be inconsistent with the observed hxr delays .
[ aschwanden , m.j .
1996 , in `` high energy solar physics '' , ( eds .
r.ramaty , n.mandzhavidze , and x .- m.hua ) , aip conf.proc . 374 , 300 ] [ aschwanden , m.j . ,
hudson , h . , kosugi , t . , and schwartz , r.a .
1996a , apj 464 , 985 ] [ aschwanden , m.j . ,
kosugi , t . ,
hudson , h.s . ,
wills , m.j . , and schwartz , r.a .
1996c , apj 470 , oct 20 issue , in press ] [ aschwanden , m.j . &
schwartz 1995 , apj 455 , 799 ] [ aschwanden , m.j . &
schwartz 1996 , apj 464 , 974 ] [ aschwanden , m.j . ,
wills , m.j . , hudson , h.s . ,
kosugi , t . , and schwartz , r.a .
1996b , apj 468 , sept 1 issue , in press ] [ aschwanden , m.j . , schwartz , r.a . , and alt , d.m .
1995 , apj 447 , 923 ] [ benz , a.o . 1993 , _ plasma astrophysics _
, dordrecht : kluwer ] [ boldt , e . , & serlemitsos , p .
1969 , apj 157 , 557 ] [ brown , j.c . 1971 , sp 18 , 489 ] [ brown , j.c . ,
karlicky , m . ,
mackinnon , a.l . , and van den oord g.h.j . 1990 , apjs 73 , 343 ] [ cargill , p.j .
1996 , eos , american geophys .
union , in press ] [ emslie , a.g .
1996 , eos , american geophys .
union , in press ] [ emslie , a.g . and
brown , j.c .
1985 , apj 295 , 648 ] [ emslie , a.g . ,
henoux , j .- c . ,
mariska , j.t . , and newton , e.k .
1996 , apj .
submitted ] [ evenson , p . , meyer , p . ,
yanagita , s . , and forrest , d.j . 1984 ,
apj 283 , 439 ] [ heristchi , d .
1986 , apj 311 , 474 ] [ heristchi , d .
1987 , apj 323 , 391 ] [ kiplinger , a.l . ,
dennis , b.r .
, emslie , a.g . , frost , k.j . , and orwig , l.e .
1983 , apj 265 , l99 ] [ kiplinger , a.l . , dennis , b.r . , frost , k.j . , and orwig.l.e .
1984 , apj 287 , l105 ] [ kosugi , t .
1996 , in `` high energy solar physics '' , ( eds .
r.ramaty , n.mandzhavidze , and x .- m.hua ) , aip conf.proc . 374 , 267 ] [ machado , m.e , ong , k.k . ,
emslie , a.g . , fishman , g.j .
, meegan , c . ,
wilson , r . , and paciesas , w.s .
1993 , adv.space res .
13(9 ) , 175 ] [ ramaty , r .
1985 , in physics of the sun ( eds .
p.a.sturrock et al . ) , chapter 14 , ( dordrecht : reidel ) ] [ ramaty , r .
& mandzhavidze , n .
1994 , in _ high - energy solar phenomena _ , eds . j.m.ryan and w.t.vestrand , aip conf.proc .
294 , p.26 ] [ ramaty , r .
, mandzhavidze , n . , kozlovsky , b . , and murphy , r.j .
1995 , apj 455 , l193 ] [ sakao , t .
1994 , phd thesis , univ .
tokyo ] [ share , g.h . and murphy , r.j . 1995 , apj 452 , 933 ]
[ simnett , g.m .
1986 , sp 106 , 165 ] [ simnett , g.m . 1995 , space sci.rev .
73 , no.3 - 4 , 837 ] [ simnett , g.m .
1996 , eos , american geophys .
union , in press ] [ simnett , g.m . and haines , m.g . 1990 ,
sp 130 , 253 ] [ simnett , g.m . and
strong , k.t .
1984 , apj 284 , 839 ] | both electrons and ions are accelerated in solar flares and carry nonthermal energy from the acceleration site to the chromospheric energy loss site , but the relative amount of energy carried by electrons versus ions is subject of debate . in this _
letter _ we test whether the observed energy - dependent timing delays of 20 - 200 kev hxr emission can be explained in terms of propagating electrons versus protons . for a typical flare ,
we show that the timing delays of fast ( @xmath0 s ) _ hxr pulses _ is consistent with time - of - flight differences of directly precipitating electrons , while the timing delays of the _ smooth hxr _ flux is consistent with collisional deflection times of trapped electrons .
we show that these hxr timing delays can not be explained either by @xmath1 mev protons ( as proposed in a model by simnett & haines 1990 ) , because of their longer propagation and trapping times , or by @xmath2 mev protons ( which have the same velocity as @xmath3 kev electrons ) , because of their longer trapping times and the excessive fluxes required to generate the hxrs .
thus , the hxr timing results clearly rule out protons as the primary generators of @xmath4 kev hxr emission .
# 1#2 * figure # 1 : * _ _ # 2 _ _ # 1 epsf _ the astrophysical journal letters _ |
cege@xmath9 appears to be one of the simple binary compounds that has a benefit of being of a interest for a number of years . over
50 years ago it was reported to be a ferromagnet with the curie temperature @xmath10 k that crystallizes in the orthorhombically distorted thsi@xmath9 structure .
@xcite the crystal structure of cege@xmath9 was studied further . @xcite it was identified either as a tetragonal @xmath11-thsi@xmath9 type @xcite , or as an orthorhombic @xmath11-gdsi@xmath9 type .
@xcite magnetization and heat capacity measurements @xcite suggested ferromagnetic order with @xmath12 k. similar @xmath12 k value and the kondo temperature @xmath13 k were reported in @xcite .
detailed magnetization study of cege@xmath9 @xcite proposed two magnetic transitions in this material , antiferromagnetic , at @xmath14 k and ferromagnetic , at lower temperature , @xmath15 k. the inconsistencies in determination of the crystal structure in early publications appeared to have a simple explanation . with further studies of the ce - ge binary phase diagram
it was realized that cege@xmath9 is not a stoichiometric , line , compound , but has a depleted structure with a width of formation , cege@xmath0 with @xmath16 , or larger ( see below ) @xcite a transition from high temperature , tetragonal , @xmath11-thsi@xmath9 type structure to low temperature , orthorhombic , @xmath11-gdsi@xmath9 type was reported at temperatures between 560@xmath17c ( @xmath18 ) and 490@xmath17c ( @xmath19 ) .
@xcite within the range of ge concentrations between 1.71 and 2.0 the orthorhombic crystal structure , and two magnetic transitions , with @xmath20 k and @xmath21 k , virtually independent on ge concentration , were reported .
@xcite for lower ge concentrations , between cege@xmath22 and cege@xmath23 , a tetragonal crystal structure and a single , antiferromagnetic , transition with @xmath20 k ( and no other magnetic transition down to 0.5 k ) were observed .
@xcite the latter magnetic results were challenged by conclusions in @xcite , where two magnetic transitions , one at 6.7 k and another , ferromagnetic , at 5.3 k were described for the tetragonal cege@xmath22 . given that majority of the studies , in particular for orthorhombic cege@xmath0 with higher ge concentration , were performed on polycrystalline samples ( as a result the anisotropies of physical properties were not accessed ) and that the literature data contains apparent inconsistencies , we report a detailed study of the thermodynamic and transport properties of cege@xmath0 ( @xmath1 ) single crystals .
single crystals of cege@xmath0 were grown using the high - temperature solution technique . @xcite the constituent elements , with an initial stoichiometry of ce@xmath24ge@xmath25 , were placed in an alumina crucible and sealed in a fused silica tube under a partial ar pressure .
the ampule was heated up to 1190@xmath17c and then was cooled down to 890@xmath17c over @xmath26 hours .
after that the flux was decanted , meaning that the resulting single crystals were effectively quenched to room temperature from to 890@xmath17c over several minutes .
the crystals grow as large plates , with the dimensions in excess of @xmath27 mm@xmath28 , with the @xmath7-axis perpendicular to the plate ( see below ) . from earlier work @xcite and the binary ce - ge phase diagram @xcite
the crystals are expected to have some ge - deficiency , cege@xmath0 .
elemental analysis was performed on these crystals using wavelength - dispersive x - ray spectroscopy ( wds ) in the electron probe microanalyzer of a jeol jxa-8200 electron microprobe .
ce@xmath9fe@xmath29 and elemental ge were used as the standards .
the stoichiometry of the samples was determined to be cege@xmath30 with the error bar in the value of @xmath31 of about or less than 0.01 . at first glance
, it appears that this result contradicts the established width of formation shown in the binary phase diagram.@xcite we argue that the spacing between the experimental points in the original papers @xcite and the error bars present in this work as well as in the original publications allow to remove this apparent contradiction .
given our growth method this result implies that the ge - rich side of the width of formation should be shifted from ce@xmath32ge@xmath33 to this composition ( ce@xmath34ge@xmath35 ) .
we will use the notation cege@xmath30 for the samples studied in this work .
room temperature powder x - ray diffraction measurements were taken using cu-@xmath36 radiation in a rigaku miniflex diffractometer .
the spectrum was refined using a rietica software .
@xcite the x - ray data ( fig .
[ f1 ] ) are consistent with the sample being orthorhombic with the lattice parameters @xmath37 , @xmath38 , and @xmath39 , the values similar to those reported in refs @xcite .
it has to be mentioned that a complex superstructure originating from a partial ordering of the ge vacancies was reported for a sample with a very close ge - concentration , cege@xmath40 @xcite .
the resolution of our x - ray diffraction data was not sufficient to detect such a superstructure , if present .
figure [ f2 ] shows the optical image of the surface of the crystal taken with optical , polarized light , microscope at room temperature .
the very clear domain structure is an evidence of a solid - solid phase transition above room temperature and is fully consistent with the reported tetragonal to orthorhombic structural phase transition near to 500@xmath17c .
@xcite the long orthorhombic domains with the very wide range of width , from @xmath41 m to @xmath42 m , are clearly seen .
the existence of such domains implies that without de - twinning of the crystals measured in - plane properties will represent some average between @xmath43 and @xmath44 directions .
additional scattering on the domain boundaries might worsen the in - plane resistivity values .
anisotropic , temperature dependent , and field dependent dc magnetization was measured using commercial quantum design mpms-5 and mpms-7 units . in - plane ac resistivity ( @xmath45 ma , @xmath46 hz )
was measured using an act option of a quantum design ppms-14 instrument .
the heat capacity data were taken using a hybrid adiabatic relaxation technique of the heat capacity option in a quantum design ppms-14 instrument .
thermoelectric power ( tep ) measurements were carried out by dc , alternating heating ( two - heaters - two - thermometers ) , technique @xcite over the temperature range from 2 to 300 k. thermal expansion and magnetostriction were measured using a capacitive dilatometer constructed of ofhc copper .
@xcite the dilatometer was mounted in a quantum design ppms-14 instrument and was operated over a temperature range of 1.8 - 300 k in magnetic field up to 140 koe . due to aforementioned presence of the orthorhombic domains , the @xmath47 dilation was measured along an arbitrary in - plane direction .
in - plane , temperature dependent resistivity of the cege@xmath30 crystal is shown in fig .
it features a very broad hump at about 50 k and a shallow minimum just above the sharp drop in resistivity below @xmath48 k. the residual resistivity ratio , @xmath49 is approximately 3.5 . such relatively high value of the @xmath50 ( compare e.g. with the @xmath50 values of ni - defficient _
r_nige@xmath51 , it r = rare earth , single crystals @xcite ) hints at possibility of the ordering of the ge vacansies suggested earlier , @xcite for a sample with a stoichiometry that is essentially cege@xmath40 .
this temperature behavior is similar to that observed in other heavy fermion or intermediate valence compounds with a low temperature magnetic order .
the temperature derivative of the resistivity data , @xmath52 , @xcite clearly shows two distinct features , at 7.1 k and 3.8 k , suggesting two magnetic transitions in this material .
large magnetic fields suppress the features , and low temperature magnetoresistance is negative .
the overall resistivity behavior is similar to what was reported for cege@xmath53 @xcite , tetragonal cege@xmath22 @xcite and cege@xmath9 @xcite , despite the differences is stoichometry that do affect the number and exact temperature of transitions detected in low temperature resistivity .
anisotropic , inverse , temperature - dependent dc susceptibility , @xmath54 , measured at 10 koe is shown in fig .
the polycrystalline average was calculated as @xmath55 .
the susceptibility is anisotropic with @xmath56 in the paramagnetic state .
the curie - weiss fits of the high temperature ( 150 k @xmath57 300 k ) susceptibility data , @xmath58 ( where @xmath59 and @xmath60 are curie constant and curie - weiss temperature , respectively ) yield the effective moment ( from @xmath61 ) @xmath62 , consistent with the theoretical value for ce@xmath63 , and the curie - weiss temperatures , @xmath64 k , @xmath65 k , and @xmath66 k. the signs of the curie - weiss temperatures suggest that antiferromagnetic interactions dominate in the @xmath47-plane whereas ferromagnetic ones are predominant along the @xmath7-axis . the low temperature , low field magnetization data are shown in the inset to fig . [ f4 ] .
these data further confirm magnetic transitions at @xmath48 k and @xmath67 k. in addition , the data for @xmath68 suggests one more transition , at @xmath69 k ( see fig . [ f10 ] and discussion below ) . the temperature dependent heat capacity ( fig .
[ f5 ] ) has a clear , distinct feature at @xmath70 k and somewhat less pronounces anomaly @xmath71 k ( seen as a knee when the data are plotted as @xmath72 vs @xmath73 , fig .
[ f5 ] , inset ) .
there appear to be other small deviations in @xmath74 behavior in the 3.7 k @xmath75 7 k temperature range .
they do not have corresponding anomalies in resistivity , however the one at @xmath69 k is consistent with the anomaly in the in - plane low field magnetization , mentioned above .
the sommerfeld coefficient , estimated from the linear fit of @xmath72 vs @xmath73 above the transitions , between the 150 k@xmath3 and 300 k@xmath3 values of @xmath73 ( 12 - 17 k in the units of @xmath76 ) is enhanced , @xmath77 mj / mol k@xmath3 .
this value of the sommerfeld coefficient is similar to that reported for cege@xmath9 and cege@xmath78 @xcite and is noticeably higher than the values reported for cege@xmath79 and cege@xmath22 .
@xcite a rough estimate of the magnetic entropy associated with the magnetic transitions ( integration up to @xmath80 k ) gives the value of @xmath81 .
anisotropic temperature - dependent dilation of cege@xmath30 is shown in fig .
the overall change of volume from the base temperature , 1.8 k to 300 k is close to 1% , that is similar to the change of volume in cu @xcite and is approximately factor of two larger than that of yni@xmath9b@xmath9c @xcite . in - plane
thermal dilation is positive over the whole temperature range and is close to linear in temperature above @xmath82 k with the thermal expansion coefficient @xmath83 k@xmath84 .
negative thermal expansion along the @xmath7-axis is observed in cege@xmath30 below @xmath85 k with the volume thermal expansion being positive in the whole temperature range .
the upper transition is clearly seen at @xmath86 k in both anisotropic dilation and thermal expansion coefficients ( fig .
[ f6 ] , inset ) .
the lower transition is seen as a small step - like feature at @xmath87 k in @xmath88 . in @xmath89 a well - defined minimum at temperature @xmath90 k is seen , in agreement with the positions of anomalies in the @xmath91 and the heat capacity .
since the magnetic transition at @xmath48 k appears to be the second order , we can use our heat capacity and thermal expansion results to evaluate the initial uniaxial and total pressure derivatives of this transition using the ehrenfest relation , @xcite : @xmath92 , where @xmath93 is the molar volume ( @xmath94 m@xmath28 using the lattice parameters from the x - ray diffraction ) , @xmath95 is the jump in the thermal expansion coefficient at the phase transition , and @xmath96 is the corresponding jump in the heat capacity . using the experimental data above , for cege@xmath30 we found out that in - plane uniaxial pressure derivative is positive , @xmath97 k / kbar , whereas the @xmath7-axis uniaxial pressure derivative is very similar in its absolute value but negative @xmath98 k / kbar .
the hydrostatic pressure derivative can be approximated as @xmath99 k / kbar .
so it appears that cege@xmath30 is equally sensitive to uniaxial pressure applied in the @xmath47 - plane and along the @xmath7-axis , however the upper magnetic transition temperature increases when the pressure is applied in the @xmath47 - plane and decreases when it is applied along the @xmath7-axis .
the absolute values of the pressure derivatives are fairly small .
the absolute value of the uniaxial pressure derivatives are close to those inferred for orthorhombic cevsb@xmath51 @xcite , however for cevsb@xmath51 all uniaxial pressure derivatives have the same , positive , sign .
cege@xmath30 , like cevsb@xmath51 , may be close to a local maximum in ordering temperature and pressures up to 7 gpa may well be able to suppress the magnetic transition to a quantum phase transition .
temperature - dependent , in - plane , tep data for cege@xmath30 is shown in fig .
[ f7 ] . on a gross level
these data are similar to those reported for cege@xmath9 . @xcite the data show two maxima , a broad one at @xmath100 k ( corresponding to the lowest crystal electric field splitting of the energy levels @xcite ) and a sharper one , at @xmath101 k and a minimum at @xmath102 k. since it is expected that @xmath103 , another minimum , below 2 k is anticipated . above @xmath104
k the tep is linear in temperature , indicating that in this temperature region the diffusion tep is dominant .
the tep changes sign several times , being negative above @xmath105 k , positive between @xmath106 k and @xmath105 k , then negative below @xmath106 k with a very short positive excursion between 3.0 k and 3.8 k. two low temperature transitions , at @xmath86 and @xmath107 k , are clearly seen in the @xmath108 ( fig .
[ f7 ] , inset ) , the upper one as a sharp minimum and the lower one as a change in slope .
there appears to be an additional , step - like feature in the @xmath108 at @xmath5 k , close to the temperature of small anomalies in the thermodynamic measurements discussed above .
low - field , five - quadrant , anisotropic magnetization loops for cege@xmath30 are shown in fig .
these data suggest that there is a ferromagnetic component in the magnetization along the @xmath7 - axis at 1.85 k , with a rather small , @xmath110 per cerium , saturated moment and very narrow hysteresis loop .
the width of the loop and the saturated moment along the @xmath7 - axis are smaller than those of stoichiometric and magnetically soft ceagsb@xmath9 @xcite .
this virtual absence of hysteresis again hints at possible order of vacancies ( perhaphs in a cege@xmath40 , or ce@xmath111ge@xmath112 , structure ) or at least lack of pinning , even on domain walls .
the apparent low field ferromagnetic loop measured for the magnetic field in the @xmath47 plane might be intrinsic , but very possibly is a result of a slight ( less than 0.5@xmath17 ) misallignment of the sample .
additional features , corresponding to metamagnetic transitions , are observed in the anisotropic @xmath113 data , taken up to higher , 10 koe , field at different temperatures ( fig .
magnetization data between 10 koe and 55 koe for both orientations ( not shown ) are monotonic , with no additional features .
the saturated magnetization in both directions is below @xmath8 .
the magnetization isotherms measured at different temperatures and low temperature , temperature - dependent magnetization measured at different fields ( fig . [ f10 ] ) allow for construction an anisotropic , @xmath109 , phase diagrams for cege@xmath30 ( fig .
[ f11 ] ) .
the points obtained from @xmath114 and @xmath113 data are consistent with each other . for @xmath115 the points obtained from resistivity , thermoelectric power , thermal expansion and heat capacity ( as described in the text above ) are added .
whereas the @xmath116 direction is unique , the phase diagram for @xmath68 is shown for some guidance only , since twinning and exact orientation might affect the number and the exact position of the phase lines .
phase i is the phase with a ferromagnetic component along the @xmath7 - axis , phase ii is a higher temperature antiferromagnetic phase .
phase iv , if indeed exists , is located @xmath117 k and @xmath118 k in zero field , and probably disappears by 0.5 - 1 koe .
phase v for @xmath68 is a metamagnetic phase.the phase marked as iii appears to be the phase corresponding to saturated paramagnetic state .
the magnetic field scales for cege@xmath30 are rather small , not exceding 10 koe , suggesting certain fragility of the magnetic order in this material .
we presented results of a comprehensive study of thermodynamic and transport measurements on cege@xmath30 single crystals .
although formally it is an extreme member of the cege@xmath0 family , the wds data are consistent with cege@xmath40 or ce@xmath111ge@xmath112 and both resistivity data and magnetization loop data are consistent with a well ordered compound .
this material has an enhanced electronic specific heat coefficient and two magnetic ordering transitions , at @xmath4 k and at @xmath6 k , and possibly a third one , at @xmath119 k that is observed as a small anomaly in a number of measurements .
the ground state characterized by a possibly compex magnetic structure with a ferromagnetic component along the @xmath7 - axis corresponding to a magnetic moment of @xmath120 per ce .
the response of the cege@xmath30 to the applied magnetic field is anisotropic , with metamagnetism ( spin reorientation ) observed in both directions .
this simple binary material is a good candidate for neutron scattering studies as well as for studies of the evolution of its magnetism under pressure . to some extent ,
cege@xmath30 has similarities to other ce - based , reduced moment ferromagnets , like ceagsb@xmath9 @xcite and ceru@xmath9ge@xmath9 @xcite , the studies of which yield interesting results related to quantum phase transitions and strongly correlated behavior .
help of w. e. straszheim in wds measurements on the sample is greatly appreciated .
s.l.b . thanks a. i. goldman for his abetment in finalyzing this manuscript .
work at the ames laboratory was supported by the us department of energy , basic energy sciences , division of materials sciences and engineering under contract no .
de - ac02 - 07ch11358 .
s.l.b . acknowledges partial support from the state of iowa through iowa state university .
99 hunter b , in : ( 1998 ) _ iucr cpd newsletter _ * 20 * , 21 .
( howard c j , hunter b a and swinkels d a j 1997 rietica iucr powder diffr .
22 21 ) ( howard c j and hunter b a 1998 a computer program for rietveld analysis of x - ray and neutron powder diffraction patterns ( nsw , australia : lucas heights research laboratories ) pp 127 ) schmiedeshoff g m , lounsbury a w , luna d j , tracy s j , schramm a j , tozer s w , correa v f , hannahs s t , murphy t p , palm e c , lacerda a h , budko s l , canfield p c , smith j l , lashley j c and cooley j c , 2006 _ rev .
instrum . _ * 77 * , 123907 . single crystal .
insets : low temperature resistivity in zero , 70 koe and 140 koe magnetic field applied along the @xmath7-axis ( upper right ) ; low temperature part of the derivative @xmath52 with arrows marking the transition temperatures ( lower left).,width=453 ] , for cege@xmath30 single crystal .
the polycrystalline average was calculated as @xmath55 .
inset : low temperature , low field anisotropic magnetization of cege@xmath30 single crystal plotted on a semi - log scale.,width=453 ] of cege@xmath30 single crystal .
arrows point to the features associated with the magnetic transitions .
inset : the same data plotted as @xmath72 vs @xmath73 . dashed line - linear extrapolation of @xmath72 vs. @xmath73 from above the magnetic transitions . , width=453 ] -plane and @xmath7-axis , dilation and the volume change of cege@xmath30 .
the data are normalized to the respective values at 1.8 k. inset : low - temperature @xmath47-plane and c - axis dilation and anisotropic thermal expansion coefficients .
, width=453 ] measured at several different temperatures for ( a ) @xmath123 , and ( b ) @xmath124.,title="fig:",width=377 ] measured at several different temperatures for ( a ) @xmath123 , and ( b ) @xmath124.,title="fig:",width=377 ] measured in several different applied fields for ( a ) @xmath123 , and ( b ) @xmath124 , shown in semi - log plots.,title="fig:",width=377 ] measured in several different applied fields for ( a ) @xmath123 , and ( b ) @xmath124 , shown in semi - log plots.,title="fig:",width=377 ] phase diagrams for ( a ) @xmath123 , and ( b ) @xmath124 .
symbols : squares - from @xmath113 data , stars - from @xmath114 data . for @xmath115 data from resistivity , thermoelectric power , thermal expansion and heat capacity are added .
roman numbers mark different magnetic phases , title="fig:",width=377 ] phase diagrams for ( a ) @xmath123 , and ( b ) @xmath124 .
symbols : squares - from @xmath113 data , stars - from @xmath114 data . for @xmath115 data from resistivity , thermoelectric power , thermal expansion and heat capacity are added .
roman numbers mark different magnetic phases , title="fig:",width=377 ] | we present data on the anisotropic magnetic properties , heat capacity and transport properties of cege@xmath0 ( @xmath1 ) single crystals .
the electronic coefficient of the heat capacity , @xmath2 mj / mol k@xmath3 , is enhanced ; three magnetic transitions , with critical temperatures of @xmath4 k , @xmath5 k , and @xmath6 k are observed in thermodynamic and transport measurements .
the ground state has a small ferromagnetic component along the @xmath7 - axis .
small applied field , below 10 koe , is enough to bring the material to an apparent saturated paramagnetic state ( with no further metamagnetic transitions up to 55 koe ) with a reduced , below @xmath8 , saturated moment . |
the neutrino fluxes emitted from a core collapse supernova ( sn ) contain information about the primary fluxes produced inside the star , the neutrino mixing pattern as well as the matter densities that the neutrinos have passed through .
the high statistics neutrino signal that one expects from a future galactic sn needs to be decoded in order to extract this information .
the neutrinos and antineutrinos produced inside the sn pass through the core , mantle and envelope of the progenitor star before escaping .
they encounter matter densities ranging from nuclear densities deep inside the core all the way to vanishingly small densities in the interstellar space .
neutrino masses and mixing angles , and hence the extent of their flavor conversions , depend crucially on the density of surrounding matter , hence it is important to study these matter effects in detail .
the matter effects @xcite give rise to msw resonances @xcite when the matter density corresponds to _
r = 2m_n / ( 2g_f y_e e ) .
[ msw ] here @xmath4 is the electron fraction , @xmath5 the average nucleon mass , and the plus and minus signs correspond to neutrinos and antineutrinos respectively . for neutrinos of energy @xmath6 , resonances are possible at two densities , the @xmath7 resonance corresponding to @xmath8 and the @xmath9 resonance corresponding to @xmath10 .
the energies of sn neutrinos are typically in the range @xmath11@xmath12 mev .
for these energies , the @xmath7 resonance takes place around @xmath13@xmath14 g / cc .
it occurs in neutrinos for normal hierarchy and in antineutrinos for inverted hierarchy .
the @xmath9 resonance that takes place around @xmath15@xmath16 g / cc always occurs in neutrinos .
however , since @xmath17 is large , significant flavor conversions of antineutrinos also take place at the @xmath9 resonance .
the adiabaticities at the @xmath7 and @xmath9 resonances determine the net neutrino conversion probabilities .
the @xmath9 resonance is always adiabatic , given the values of @xmath18 and the typical density profile of the progenitor star around @xmath19 .
the adiabaticity at the @xmath7 resonance is very sensitive to the value of @xmath3 and the density profile of the star in the resonance region .
indeed the neutrino conversion rates are crucially dependent on the value of @xmath3 , and whether the @xmath7 resonance is in neutrinos or antineutrinos
. the sn signal can therefore be an extremely sensitive probe of @xmath3 and whether the mass hierarchy is normal or inverted @xcite .
in addition to divulging the neutrino mixing scenario , sn neutrino fluxes can also allow us to have a peek at the propagation of the shock wave while it is still inside the mantle of the star .
the violent density fluctuations caused by the sn shock wave can change the adiabaticity at the @xmath7 resonance in a time dependent manner , thus leaving their imprint on the time dependent neutrino spectra @xcite . in particular , the observations of the time dependent neutrino spectra can confirm the presence of forward as well as reverse shock wave through the `` double dip '' feature @xcite , and in addition can track the positions of the shocks as they pass through the @xmath7 resonance region .
the feasibility of such a tracking at a water cherenkov detector has been explored in @xcite .
our understanding of the sn explosion mechanism is still unsatisfactory @xcite , which makes it very important to extract as much information about the shock wave as possible . in this paper , we demonstrate how the neutrinos that pass through the shock wave near the @xmath7 resonance carry information about the density profile of the shock wave . in addition to making the @xmath7 resonance
temporarily nonadiabatic , the shock wave also forces the neutrinos to encounter multiple @xmath7 resonances .
the relative phases that the neutrino mass eigenstates gain between two or more of such @xmath7 resonances manifest themselves as oscillations in the neutrino flavor conversion probabilities as a function of the neutrino energy .
these oscillations are related to the shock wave density profile , and in principle carry information about it .
we therefore study the relation between the oscillations and the shock wave profile in detail .
most of the analyses of sn neutrino conversions till now calculate the _ probabilities _ of conversions between neutrino mass eigenstates in matter in each resonance region , and combine the results at different resonances , assuming them to be independent .
this includes the implicit assumption that the information on the relative phase between the neutrino mass eigenstates is lost between successive resonances . for neutrinos that undergo a single @xmath7 and a single @xmath9 resonance ,
this is a valid assumption since the two resonance regions are well separated @xcite .
however , when the neutrinos encounter multiple @xmath7 resonances that are close to each other , coherence between the neutrino mass eigenstates is maintained , and one has to compute the _ amplitudes _ of neutrino flavor conversions at the resonances , keeping track of the relative phases .
this gives rise to the phase effects that we explore in this paper .
such effects were pointed out in the context of solar neutrinos in @xcite and also observed in the numerical studies of supernova neutrinos in @xcite .
we provide an analytical approximation to study these effects , relating them to the the density profile of the medium .
we also provide criteria to decide when decoherence or finite energy resolution of detectors renders these effects unimportant .
if the multiple resonances are semi - adiabatic , the phase effects may be strong and if the oscillations are indeed observed in the neutrino spectra , they can help us infer about the density profile of the shock wave , which in turn can provide us important clues about the shock wave propagation and the sn explosion .
the observability of these effects is dependent on factors like neutrino fluxes at the source , possible coherent development of neutrinos due to neutrino - neutrino interactions , the different density profiles encountered by neutrinos coming from different regions of the neutrinosphere , stochastic fluctuations of density , effective luminosity of the supernova , detector capabilities , et cetera .
we comment on different model predictions of source fluxes and point out the model independent features of the phase effect .
we investigate the number of events and detector capabilities required for the effect to be observable and find them to be demanding .
our emphasis in this paper is to first ascertain the origin of these phase effects , their dependence on various parameters in the problem , and to check whether they are important at least in a simplified analysis where effects like coherent development of neutrinos @xcite , anisotropy @xcite and turbulence @xcite of the medium are neglected .
we find that even in a simplified setting , the observation of these effects is challenging .
the paper is organised as follows . in sec .
[ analytic ] , we give an analytical approximation for calculating the neutrino conversion probability when multiple @xmath7 resonances are taken into account in a two - neutrino framework .
the results of this section are general and can be applied to any situation where multiple resonances are involved . in sec .
[ actual ] , we apply the results to the case of a sn shock wave , where the third neutrino and the @xmath9 resonance are included . we show the neutrino conversion probabilities obtained by using a realistic shock wave profile and study the feasibility of detecting the phase effects .
[ concl ] concludes .
in this section , we calculate the survival probability of @xmath20 when they pass through multiple resonances , keeping track of the relative phases between the mass eigenstates .
the calculations are performed in the @xmath21-@xmath22 framework .
the results are readily extended to the @xmath23-@xmath22 framework in the case of a sn shock wave , as will be shown in sec .
[ actual ] .
although all the arguments in this section are given for neutrinos , they are valid for antineutrinos just by changing the sign of the matter potential @xmath24
. however , the flavor conversion analysis at the @xmath7 resonance is applicable to neutrinos in the normal mass hierarchy and for antineutrinos in the inverted mass hierarchy .
therefore , while analyzing antineutrinos , the value of @xmath25 should be taken to be negative .
the relevant mixing angle at the resonance @xmath7 is @xmath27 , on which we currently have a strong bound : @xmath28 @xcite .
therefore , we try to solve the problem using a small angle approximation .
we follow the notation and framework outlined in @xcite and work in the flavor basis .
let @xmath29 be the relevant linear combination of @xmath30 and @xmath31 .
when neutrinos pass through matter , their propagation is described by the schrdinger equation i ( c _ e + _ + ) = h ( c _ e + _ + ) .
[ coupledeqns ] upto a matrix proportional to the unit matrix , the hamiltonian @xmath7 is given by , ( cc a(x)- 2 & 2 + 2 & -a(x)+ 2 + ) , [ hamiltonian ] where @xmath32 .
these two coupled first order equations give rise to the second order equation -_e - ( ^2 + i ) _ e = ^2 _ e [ nu_eeqn ] + [ order2 ] where ( x ) = ( a(x ) - 2 ) , = 2and prime ( @xmath33 ) denotes derivative with respect to @xmath34 . in order to find the survival probability of @xmath20 , we solve for the @xmath35 wavefunction with the initial conditions @xmath36 , @xmath37 .
these conditions are equivalent to _ e(0 ) = 1 , i.|_0 = ( 0 ) .
[ initial ] the `` logarithmic perturbation '' approximation @xcite solves the differential equation ( [ nu_eeqn ] ) for small mixing angles by choosing g 1 - 2[g - def ] as the small expansion parameter . denoting _
e = e^s(x ) , s(x ) = c_0(x ) + gc_1(x ) + o(g^2 ) [ sx - def ] so that _
e(x ) & = & ( _ 0^xdx_1 c_0(x_1 ) . + & & .+g_0^xdx_1 c_1(x_1)+ o(g^2 ) ) , the solution becomes _
e(x ) & = & + & & + o(g^2 ) .
[ og - soln ] here we have defined the `` accumulated phase '' q(x ) _ 0^x dx_1 ( a(x_1)- ) . [ q - def ] the survival probability @xmath38 at @xmath39 then becomes p_ee(x ) & = & + & & + o(g^2 ) .
[ pee ] the integral in the above expression can be evaluated using the stationary phase approximation .
the integral oscillates rapidly unless @xmath40 .
so the entire contribution to the integral can be taken to be from the saddle point @xmath41 , which is the point where @xmath42 , i.e. a(x_s ) = .
note that this is also the resonance point in the small angle limit . for a monotonic density profile ,
there is only one saddle point @xmath41 and the survival probability is p_ee , [ pee - mono ] which agrees with the landau - zener jump probability @xcite in the limit of small mixing angle , and hence small @xmath43 , even when @xmath44 . for a non - monotonic density profile
, neutrinos can experience more than one resonance at the same density but at different positions . in that case
@xmath45 at more than one point .
if the resonances are sufficiently far apart , the contributions from each of them may be added independently of each other .
their total contribution to the integral in ( [ pee ] ) is _ 0^x
dx e^iq(x ) _
ie^i_ie^iq(x_i ) ( ) ^ 1/2 where @xmath46 runs over all the saddle points . note that @xmath47 if @xmath48 and @xmath49 if @xmath50 .
the probability calculated using ( [ pee ] ) now has terms which depend on the differences between the integrated phases _ ij & & q(x_j)-q(x_i)+_j-_i + & = & _ x_i^x_j ( a(x)- ) dx + _ j-_i .
[ phi - ij - def ] in general , p_ee(x ) &
= & + & & + o(g^2 ) [ pee - general ] where a_i ( ) ^1/2 . [ a - def ] for example , when there are only two saddle points the survival probability is given by p_ee & = & ( -g a_1 ^ 2 ) ( -g a_2 ^ 2 ) ( -2 g a_1 a_2 _ 12 ) + & & + o(g^2 ) .
[ master - eqn ] the first two factors in ( [ master - eqn ] ) are the individual landau - zener jump probabilities for the two level crossings .
the last factor gives rise to oscillations in @xmath51 as a function of energy .
the oscillation pattern has its maxima at @xmath52 and minima at @xmath53 where @xmath54 is an integer .
this expression is valid as long as @xmath55 , so that the @xmath56 terms remain small , and none of the resonances overlap .
we illustrate the validity and limitations of the small angle approximation with a toy density profile ( x ) = \ { ll a + b_1 x^2 & ( x<0 ) + a + b_2 x^2 & ( x>0 ) + . ,
[ toy - ne ] as shown in figure [ densityplot ] .
we take @xmath57 and @xmath58 .
neutrinos are produced at @xmath59 and we calculate @xmath51 at @xmath60 .
we also show the positions of resonance densities for various energies , which are given by _
r[g / cc ] .
[ msw2 ] figure [ nonmonoplot ] shows the survival probability @xmath51 as a function of energy , both the exact numerical result and the result of our analytical approximation for small angles .
it can be seen that at such small angles ( @xmath61 rad @xmath62 ) , the approximation works extremely well .
note that the amplitude of the oscillations is comparable to the deviation of the average survival probability from unity .
that is , the oscillation effect is not a small effect .
indeed , the oscillation term is of the same order as the averaged effect , as can be seen from ( [ master - eqn ] ) .
figure [ nonmonoplot ] also shows the average value of @xmath51 that one would have obtained if one naively combined the jump probabilities at the two resonances .
our analysis gives additional oscillations in the survival probability as a function of neutrino energy about this average value .
this effect is what we call as the phase effect , and is clearly significant as can be seen from the figure
. an important feature of the oscillations is that the `` wavelengths , '' i.e. the distances between the consecutive maxima or minima , are larger at larger @xmath6 .
the resonances start overlapping at @xmath63 ( figure [ densityplot ] ) , which is where our approximation starts breaking down , as can be seen in figure [ nonmonoplot ] .
for @xmath64 , the neutrinos no longer encounter a strict resonance , and our approximation gives @xmath65 identically .
however , the resonances have finite widths which may affect the conversion probabilities of neutrinos with @xmath63 . the sharp jump observed in figure [ nonmonoplot ] at @xmath66 mev
is therefore not a real effect , but a limitation of our technique .
the small angle approximation starts failing for larger angles and lower energies , where @xmath67 .
figure [ pee - large ] shows that the amplitude at low energies is not calculated correctly for @xmath68 rad @xmath69 .
however , note that the positions of maxima and minima of @xmath51 are still predicted to a good accuracy .
we shall argue in the next subsection that these can be computed accurately for the whole allowed range of @xmath3 , given the non - monotonic density profile between the two resonances .
let us consider a density profile with a `` dip '' as in the toy model in the previous section .
a neutrino with energy @xmath6 encounters two resonances @xmath71 and @xmath72 at @xmath73 and @xmath74 respectively , so that _ r ( x_1 ) = ( x_2 ) .
we assume @xmath4 to be a constant throughout the region of interest .
we also assume that the propagation of neutrino mass eigenstates is adiabatic everywhere except in the resonance regions @xmath75 and @xmath76 around the resonance points @xmath77 and @xmath78 respectively . in the limit of a small mixing angle ,
the widths of the resonances are small : 2 .
therefore , @xmath79 and @xmath80 .
we shall work in this approximation , and shall use the notation @xmath81 only for the sake of clarity wherever needed . at @xmath82 , the density @xmath83 , so that the heavier mass eigenstate @xmath84 is approximately equal to the flavor eigenstate @xmath20
. let us start with @xmath20 as the initial state : _ e ( x x_1 ) _ h .
the mass eigenstate @xmath84 propagates adiabatically till it reaches the resonance region @xmath85 : _ e ( x_1- ) _ h . while passing through the resonance ,
unless the resonance is completely adiabatic , the state becomes a linear combination of @xmath84 and @xmath86 , the lighter mass eigenstate .
note that the phases of @xmath84 and @xmath86 can be defined to make their relative phase vanish at @xmath87 .
_ e(x_1 + ) = _ 1 _ h + _ 1 _ l , where @xmath88 is the `` jump probability '' at @xmath71 if it were an isolated resonance .
the two mass eigenstates @xmath84 and @xmath86 propagate to the other resonance @xmath72 , gaining a relative phase in the process ( the overall phase of the state is irrelevant ) : _
e(x_2- ) = _ 1 _ h + _ 1 ( i _ x_1^x_2 dx ) _
l , where @xmath89 is the mass squared difference between @xmath84 and @xmath86 in matter : m^2
= ( ( m^2 2- 2 e v(x))^2 + ( 2))^1/2 . [ dmtilde ] the effect of the resonance @xmath72 may be parametrised in general as ( c _ h(x_2 + ) + _ l(x_2 + ) + ) = ( cc _ 2 & _ 2 e^i + -_2 e^-i & _ 2 + ) ( c _
h(x_2- ) + _ l(x_2- ) + ) . where @xmath90 is the `` jump probability '' at @xmath72 if it were an isolated resonance . from ( [ og - soln ] ) , one can deduce that in the limit @xmath91 , we have @xmath92 .
the state @xmath93 can then be written as _ e(x_2 + ) & = & _ h + & & + _ l . for @xmath94 , the mass eigenstates travel independently and over sufficiently large distances , decohere from one another . at @xmath95 , since @xmath83 , the heavier mass eigenstate @xmath84 again coincides with @xmath20 and we get the @xmath20 survival probability as p_ee & = & ^2(_1-_2 ) + & & - 2_1 2_2 ^2 ( _ x_1^x_2 dx ) .
[ eq - freq ] if the phase information were lost , either due to decoherence or due to finite energy resolution of the detectors @xcite the survival probability would have been p_ee & = & p_1 p_2 + ( 1-p_1 ) ( 1-p_2 ) + & = & ^2 _ 1 ^2 _ 2 + ^2 _ 1 ^2 _ 2 , [ prob - combo ] which matches with ( [ eq - freq ] ) when the @xmath96 term is averaged out to @xmath97 .
the @xmath96 term in ( [ eq - freq ] ) gives rise to the oscillations in @xmath70 .
if two consecutive maxima of @xmath51 are at energies @xmath98 and @xmath99 such that @xmath100 , then the condition _
x_1(e_k+1)^x_2(e_k+1 ) dx & - & _ x_1(e_k)^x_2(e_k ) dx + & = & 2 [ integrals ] is satisfied .
the quantity @xmath101 is the `` wavelength '' of the oscillations .
note that @xmath102 is equal to @xmath103 in the small angle limit .
moreover , this quantity is rather insensitive to @xmath104 in the allowed range of @xmath3 .
therefore , it is not a surprise that the predictions of the positions of maxima and minima in the small angle approximation ( sec .
[ anal ] ) are accurate and robust in the whole range @xmath105 . since @xmath104 is small
, the left hand side of ( [ integrals ] ) is approximately equal to the area of the region in the density profile plot enclosed by the densities @xmath106 and @xmath107 : 2 .
the distance between the two resonances in the region @xmath108 is then r_k / ( _ e_k+1 - _ e_k ) .
[ res - dis ] this procedure may be repeated for various @xmath109 values to estimate the separation between the resonances at the corresponding densities , and hence to constrain the form of the density profile .
although this seems straightforward in our simplified analysis , the effect of density variations due to convection , turbulence and anisotropies would greatly complicate such a reconstruction in practice .
nonmonotonic density profiles are encountered by neutrinos escaping from a core collapse supernova during the shock wave propagation .
if the phase effects are observable at neutrino detectors , the above procedure may help us reconstruct the shock wave partially .
of course the procedure is somewhat crude , and the information obtained would only be on the density profile along the line of sight .
we have also assumed that neutrinos coming from different parts of the neutrinosphere encounter nearly the same density profiles .
however , since neutrinos are the only particles that can even in principle carry information from so deep inside the exploding star , it is important to check whether the detection of these phase effects is feasible even with these simplifying assumptions .
we shall do this in the next section .
in this section , we apply the results in the last section to the neutrinos travelling through a supernova shock wave . though we have to consider three - neutrino mixing in this case , the separation of @xmath7 and @xmath9 resonances @xcite means that we can calculate the transition probabilities at these resonances separately .
each of these resonances can then be treated as an effective two - neutrino level crossing .
the @xmath9 resonance that takes place in neutrinos is always adiabatic @xcite . if the net survival probability after passing through all the @xmath7 resonances is denoted by @xmath110 , the survival probability of @xmath26 after passing through all the @xmath7 and @xmath9 resonances is p = p_h ^2 _ p = ^2 _ [ p - def ] where nh and ih stand for normal and inverted mass hierarchy respectively . here
@xmath111 is the solar mixing angle .
similarly , the survival probability of @xmath2 after passing through all the @xmath7 and @xmath9 resonances is |p = p_h ^2 _ |p = ^2 _ .
[ pbar - def ] clearly , since the phase effects appear through @xmath110 , they will be visible only in @xmath20 for normal hierarchy and only in @xmath2 for inverted hierarchy . in present and planned water cherenkov @xcite and scintillation @xcite detectors ,
the main neutrino detection channel is the inverse beta decay reaction @xmath112 that allows the reconstruction of @xmath2 energies .
therefore we consider only the @xmath113 spectrum in our analysis .
however an analogous analysis can be easily performed in the neutrino channel for a detector able to measure the @xmath26 spectrum , for example using liquid argon @xcite . in order to illustrate the phase effects on @xmath114 or @xmath115
, we consider a typical snapshot of the density profile of a sn during a shock wave @xcite , as shown in figure [ snapshot ] . the forward shock f and the reverse shock r are sharp density discontinuities , the density change of a factor of two or more taking place over a distance of much less than a km .
the density variation in the `` contact discontinuity '' c , which is the transition region between the shock - accelerated and neutrino - heated sn ejecta , takes place more slowly , over a distance of more than @xmath116 km @xcite .
the mass accretion region a behind the forward shock wave , and the low density bubble b have gradually changing densities .
the region t is the tail of the shock wave .
the neutrinos , while passing through these regions , may undergo multiple level crossings .
the extent of flavor conversion in each region will depend on the value of @xmath3 and the steepness of the density profile in that region .
it is found that for @xmath117 rad or higher , the density variations in the mass accretion region a , the low density bubble b and the contact discontinuity c are too gradual for any non - adiabaticity .
we therefore concentrate on the forward shock f , the reverse shock r and the tail t. the coherence between mass eigenstates and the oscillations in the survival probability may be lost due to two major sources : the separation of mass eigenstate wavepackets and the finite energy resolution of the detector . the coherence length @xmath118 , defined as the distance over which the wavepackets separate , is given by @xcite l_coh ~4 e^2/ , [ coherence - crit ] where @xmath119 is the width of the wavepacket at source .
taking @xmath120 cm near the neutrinosphere @xcite in the relevant energy range of @xmath11@xmath121 mev , the coherence length for sn neutrinos is @xmath122@xmath123 km .
resonances separated by distances larger than @xmath118 may be taken to be incoherent .
since the distances involved are @xmath124 ( see figure [ snapshot ] ) , a definite conclusion about decoherence due to wavepacket separation can not be reached with this simple estimate . however , for observability the oscillations in survival probability must also occur over energy intervals much larger than @xmath125 , the uncertainity in energy measurement at the detector .
this turns out to be the dominant factor in smearing out the oscillations .
we can estimate from ( [ integrals ] ) the difference in energies at which successive maxima of the survival probability occur to be _
e_k e_k+1-e_k ~4 e_k^2/ r_k , [ ener - crit ] where @xmath126 is the distance between the two resonances encountered by a neutrino of energy @xmath127 . for the energy range of @xmath11@xmath121 mev , this gives @xmath128 of @xmath129@xmath130 mev for the t - r resonance pair with @xmath131@xmath132 km .
the oscillations are faster for the t - f and r - f pairs that occur about @xmath133@xmath134 km apart .
typically the energy resolution is @xmath129@xmath130 mev for a water cherenkov detector and @xmath135@xmath129 mev for a scintillation detector over this energy range .
moreover , the charged lepton energy is not of the same as the energy of the neutrino , which introduces additional smearing .
indeed as we shall see , the fast oscillations due to the t - f or r - f resonance pairs will be smeared and therefore clearly unobservable even at a scintillation detector .
note that the above arguments are only qualitative ; we have neglected the matter effects while estimating @xmath136 and @xmath137 .
figure [ pbarplot ] shows the value of @xmath115 as a function of energy for @xmath138 rad @xmath62 .
the rapid oscillations correspond to the relative phase @xmath139 that is accumulated by the mass eigenstates between resonance regions r and f ( t and f ) .
such high frequency oscillations are virtually impossible to observe , given the practical limits on the energy resolutions of neutrino detectors .
we smear the high frequency oscillations by taking a `` running average '' over the energy range corresponding to the typical energy resolution of a scintillation detector .
the low frequency oscillations that survive are found to correspond to the relative phase @xmath140 accumulated by the mass eigenstates between resonances in regions t and r. since these two resonances are closer compared to the resonance pairs r - f or t - f , the frequency of oscillations is smaller .
the same oscillation pattern is observed if the survival probability is computed by assuming that the resonance in region f is completely non - adiabatic , which confirms that the pattern is indeed due to the level crossings in regions t and r. note that @xmath115 goes to @xmath141 at its maximum where @xmath110 goes to unity , whereas at @xmath142 at high energies and low energies where @xmath110 goes to zero .
the oscillations in the low energy region ( @xmath143 mev ) are too rapid to be observable .
the fluctuations observed in the running average at low energies are not robust : they depend partly on the details of the density profile and are partly numerical artifacts . in figure [ analpbarplot ] , we plot the value of @xmath115 for the same parameter values , but calculated using ( [ pee - general ] ) at the t , r , f resonances .
the smeared probability curve is calculated by dropping the oscillatory terms due to the resonance at f and combining the survival probability at the t - r pair with that at f using ( [ prob - combo ] ) .
we see that the analytical expression agrees quite well with the numerical result of figure [ pbarplot ] .
in particular , the oscillation frequency matches quite well . moreover , the slope of @xmath24 at t is about @xmath135 times that at r , therefore ( [ a - def ] ) and ( [ master - eqn ] ) predict the amplitude of oscillations to be @xmath144 times the mean value of @xmath115 .
this estimate is also in good agreement with our numerical results in the figure .
the analytical approximation in ( [ pee - general ] ) breaks down when @xmath145 .
this happens below a certain value of energy , that is higher for larger mixing angles .
comparing figure [ pbarplot ] and figure [ analpbarplot ] we see that for @xmath146 rad the approximation is resonably accurate for neutrino energies above @xmath147 mev .
figure [ pbarlargetheta ] shows the smeared @xmath115 calculated numerically and analytically for @xmath148 rad .
it is clear that the approximation is valid above @xmath149 mev .
the approximation fails at low energies where @xmath115 calculated analytically does not tend to zero . for higher energies , the oscillations due to the phase effect are predicted quite accurately . in order to explore the observability of the phase effects , we use the parametrization for the primary fluxes given by @xcite : f^0__i= n()()^ , where @xmath150 . for illustration ,
we choose two models of neutrino fluxes , the garching model @xcite that uses the parameters _ = _ |_x =3 , e_0 ( ) = 15 , + e_0(|_x ) = 18 , _ 0()/_0(|_x ) = 0.8 , and the livermore model @xcite that uses _
|_x =3 , e_0 ( ) = 15 , + e_0(|_x ) = 24 , _
0()/_0(|_x ) = 1.6 .
we plot in figure [ fig - flux](a ) the quantity dn_/de _ & = & ( e _ ) f _ + & = & ( e _ ) [ f^0_|_x + |p ( f^0_-f^0_|_x ) ] [ sigma - e ] as a function of the neutrino energy @xmath151 , where we have normalized the spectrum such that the total number of events is @xmath133 .
here we use the differential cross section as computed in @xcite .
the phase effect is most prominent around @xmath152@xmath11 seconds postbounce when the tail t and reverse shock r pass through the @xmath7 resonance and the t - r resonance pair is about @xmath153 km apart .
we expect from ( [ ener - crit ] ) that the extrema in the survival probability will occur at energies separated by about @xmath129@xmath130 mev .
this is clearly visible in figure [ fig - flux](a ) .
moreover , the positions of the extrema are independent of the primary fluxes
. however , the spectrum shown in figure [ fig - flux](a ) is not directly observable : one can only observe the energy spectrum of positrons produced by the inverse beta reaction .
assuming quasielastic scattering , the positron energy is given by @xcite e_e= [ e - positron ] and p_e= , [ e - positron ] where @xmath154 is the angle of scattering , @xmath155 the proton mass , @xmath156 , and @xmath157 .
for @xmath158 mev , we have @xmath159 , so that the positron energy is spread over a range of @xmath160 mev depending on the scattering angle .
given that the successive maxima of the oscillation pattern are separated by only about @xmath21@xmath161 mev in this energy range , the oscillation pattern is significantly smeared out .
a further smearing of the oscillation pattern is caused by the finite energy resolution of the detector .
the energy resolution of a water cherenkov detector is typically @xmath162 and washes off the oscillations completely . for a scintillation detector ,
the resolution is much better , @xmath163 .
we show in figure [ fig - flux](b ) the spectrum of the observed positron energy @xmath164 after taking ( [ e - positron ] ) and ( [ e - positron ] ) into account and using the energy resolution of a scintillation detector .
it is observed that one or two extrema at high energies ( @xmath165 @xmath166@xmath167 mev ) may still survive for the livermore model where the spectrum extends to higher energies , but their clean identification would require @xmath168 events at a scintillation detector in a single time bin .
figure [ fig - flux](b ) also shows the positron energy spectrum for @xmath133 events , binned in @xmath169 mev energy intervals , which is approximately the energy resolution of a scintillation detector near @xmath170 mev .
the total number of events expected even at a future 50 kt scintillation detector is @xmath171 for a sn at a distance of @xmath130 kpc @xcite .
even with 1 sec time bins , the number of events in each bin will be @xmath172 .
this is a number too small for our purposes .
thus , the identification of the phase effects seems very unlikely , unless the sn is as close as a kpc .
the density profiles in the shock wave are quite uncertain , and one may expect that some possible profiles give rise to oscillations with larger amplitudes and wavelengths , which would be easier to observe .
the amplitude of oscillations is proportional to the ratio @xmath173 [ see ( [ a - def ] ) and ( [ master - eqn ] ) ] .
therefore , larger amplitudes need a larger ratio @xmath173 whereas larger wavelengths need a sharp tail , i.e. a larger @xmath174 . at the same time
, the adiabaticity parameter at the two resonances has to be @xmath175 for @xmath158 mev
. however , we find that even with such finely tuned density profiles , the improvement in observability is not significant .
when the neutrinos escaping from the core of a core collapse sn pass through the shock wave , they may encounter multiple `` @xmath7 '' resonances corresponding to @xmath176 and @xmath3 , when the shock wave is in the regions with densities around @xmath177@xmath14 g / cc .
we have shown that this necessarily gives rise to oscillations in the neutrino survival probabilities , which we have calculated as a function of energy .
we present an analytical approximation for small mixing angles and show that the oscillations are a significant effect : they can be of the same order as the non - oscillating terms .
the typical values of @xmath3 that gives rise to the oscillation features are in the `` transition region '' of the neutrino mixing parameter space @xcite that is significant in range ( @xmath1 ) but is usually neglected in the sn analysis for the sake of simplicity . the local maxima and minima in the survival probabilities of @xmath26 or @xmath2 are determined by the relative phase accumulated by the neutrino mass eigenstates between the multiple resonances .
the positions of these extrema depend on the density profile and are independent of the primary neutrino spectra .
if these extrema were identified they would reveal information on the propagation of the shock wave : its location as well as the density variation present around it . since neutrinos are the only particles that can carry information about the shock wave while it is still deep inside the exploding star , it is important to explore the observability of these phase effects .
moreover , the mere identification of these effects in the @xmath2 spectrum would establish the inverted hierarchy and nonzero @xmath3 , which are two of the most important quantities in neutrino phenomenology , and even faint chances of their determination should be explored .
it is interesting that for typical shock wave profiles , oscillations with `` wavelengths '' of @xmath21@xmath161 mev are indeed present in the neutrino survival probabilities in the energy range @xmath165 @xmath178@xmath167 mev .
these wavelengths are tantalizingly close to the resolving power of the neutrino detectors
. however , the inability of the detectors in reconstructing the energy of the incoming neutrino tends to wash out the oscillation pattern . for typical shock wave density profiles ,
the energy resolution of water cherenkov detectors is insufficient to detect the oscillations .
even a scintillation detector with a superior energy resolution will need @xmath168 events in the relevant time bin of @xmath152@xmath11 sec postbounce for identifying one or two extrema in the most optimistic scenario .
therefore , we expect that neglecting the phase effects is a valid practical approximation except under extreme cases
. note that it is still possible to observe other robust signatures of shock wave propagation like the dips in the time spectrum of number of events @xcite or the double dip feature in @xmath179 @xcite .
some of these features survive even in the presence of somewhat extreme stochastic density fluctuations , as discussed in @xcite .
the oscillatory effects might contribute additional scatter to these signals but will not spoil the signatures .
we must qualify the above conclusions by stating the effects ignored in our analysis to obtain the present results .
we have assumed a smooth spherically symmetric density profile and ignored anisotropies that are likely to be present @xcite .
this is a justified assumption only as long as the deviations from this assumed profile occur only on transverse distances larger than the size of the neutrinosphere so that the neutrinos in our line of sight do not experience the anisotropy .
we have not included the recently discussed collective effects of coherent flavor development @xcite , which may be important in the inverted hierarchy .
however , the extent and nature of its impact on flavor conversion has not been worked out in detail , which makes it difficult to be included in the present analysis .
similarly , the effects of a realistic spectrum of stochastic fluctuations in the medium density or turbulent convections behind the shock wave are yet to be calculated for sn neutrinos @xcite . we think that with a better understanding of the collective effects and density fluctuations one could include their effect on the realistic observability of the phase effects .
the phase effects pointed out here result from the interference between two or more msw resonances .
this phenomenon is not restricted to the sn alone , but may occur whenever neutrinos pass through nonmonotonic matter densities and the resonances are semiadiabatic .
the technique developed in this paper for treating coherent multiple resonances is applicable to such cases .
we would like to thank r. buras and l. scheck for help in understanding features of the shock wave .
ad would like to thank the max planck institute for physics for hospitality during the initial part of the work .
we would like to especially thank b. bhattacharya for useful discussions .
we are grateful to s. choubey , h. -th .
janka , e. lisi , g. g. raffelt and r. toms for comments on the manuscript .
this work was partly supported through the partner group program between the max planck institute for physics and tata institute of fundamental research .
l. wolfenstein , phys .
d * 17 * ( 1978 ) 2369 .
s. p. mikheev and a. yu . smirnov , sov .
j. nucl .
* 42 * ( 1985 ) 913 [ yad .
* 42 * ( 1985 ) 1441 ] .
a. s. dighe and a. yu . smirnov , phys .
d * 62 * ( 2000 ) 033007 [ hep - ph/9907423 ] .
l. oberauer , f. von feilitzsch and w. potzel , nucl .
suppl . * 138 * ( 2005 ) 108 .
a. bueno , i. gil - botella and a. rubbia , hep - ph/0307222 .
s. nussinov , phys . lett .
b * 63 * , ( 1976 ) 201 .
h. anada and h. nishimura , phys .
d * 41 * , ( 1990 ) 2379 .
m. t. keil , g. g. raffelt and h. t. janka , astrophys .
j. * 590 * ( 2003 ) 971 [ astro - ph/0208035 ] .
r. buras , h. -th .
janka , m. t. keil , g. g. raffelt and m. rampp , astrophys .
j. * 587 * ( 2003 ) 320 [ astro - ph/0205006 ] .
t. totani , k. sato , h. e. dalhed and j. r. wilson , astrophys .
j. * 496 * ( 1998 ) 216 [ astro - ph/9710203 ] .
a. strumia and f. vissani , phys .
b * 564 * ( 2003 ) 42 [ astro - ph/0302055 ] . | neutrinos escaping from a core collapse supernova a few seconds after bounce pass through the shock wave , where they may encounter one or more resonances corresponding to @xmath0 .
the neutrino mass eigenstates in matter may stay coherent between these multiple resonances , giving rise to oscillations in the survival probabilities of neutrino species .
we provide an analytical approximation to these inevitable phase effects , that relates the density profile of the shock wave to the oscillation pattern .
the phase effects are present only if the multiple resonances encountered by neutrinos are semi - adiabatic , which typically happens for @xmath1 .
the observability of these oscillations is severely limited by the inability of the detectors to reconstruct the neutrino energy faithfully . for typical shock wave profiles
, the detection of these phase effects seems rather unlikely .
however , if the effects are indeed identified in the @xmath2 spectra , they would establish inverted hierarchy and a nonzero value of @xmath3 . |
in @xcite c. geiss , b. leclerc , and j. schrer initiated the study of the ( generalized ) tilting theory for preprojective algebras of dynkin type @xmath4 vis vie the cluster algebra structure of the coordinate ring of the maximal unipotent group @xmath5 attached to @xmath4 under the cartan - killing classification ; see @xcite , @xcite , and @xcite regarding _
cluster algebras_. specifically they construct an explicit map @xmath0 from the module category of the preprojective algebra to the coordinate ring of the corresponding maximal unipotent group which transforms exceptional objects into cluster variables and maximal rigid modules into clusters .
the @xmath0-map can be interpreted ( and this is the view initially taken here ) as type of partition function which records the euler characteristics of _ generalized flag varieties _ attached to the module .
recently both @xcite and @xcite have independently made the new step of examining the gls @xmath0-map in the affine setting .
in particular @xcite studied examples of unipotent cells of the loop group @xmath3 and proved that their coordinate rings in accordance with the predictions made in @xcite are cluster algebras of geometric type .
this was accomplished in part by evaluating the type @xmath2 gls @xmath0-map for a fixed family of nilpotent finite dimensional modules over the preprojective algebra of type @xmath2 . for each unipotent cell in @xmath3
the authors of @xcite conjecture an explicit list of nilpotent @xmath1-modules whose images under the type @xmath2 gls @xmath0-map form the initial seed generating the cluster algebra structure of the coordinate ring of the unipotent cell . moreover
conjecture ( 4.3 ) of @xcite predicts determinantal expressions for these initial cluster variables .
this paper evaluates the type @xmath2 gls @xmath0-map over a class of nilpotent @xmath1-modules called _ shape modules _ indeed a class which properly contains those modules stipulated in conjectures ( 4.1)-(4.3 ) of @xcite and expresses the result determinantally in order to settle ( 4.3 ) of @xcite .
the proof entails computing the euler characteristic of any generalized flag variety attached to a shape module ; this is accomplished combinatorially by counting standard tableaux of requisite shape and parity .
we now go into more detail : recall that the preprojective algebra @xmath1 of type @xmath2 is defined as the quotient of the _ path algebra _ associated to the quiver @xmath6 @xmath7^{\alpha } \ar@/_2.5pc/[r]_{\ \beta^ * } & 1 \ar@/^1.1pc/[l]^{\beta } \ar@/_2.5pc/[l]_{\ \alpha^ * } } \ ] ] by the ideal @xmath8 generated by @xmath9 and @xmath10 .
let @xmath11 and @xmath12 denote the idempotents of @xmath1 . for a finite dimensional left @xmath1-module @xmath13 of dimension @xmath14 together with a choice of bit string @xmath15 in @xmath16 the _ generalized flag variety _
@xmath17 is the variety of all @xmath1-composition series @xmath18 with @xmath19 and @xmath20 such that @xmath21 whenever @xmath22 . here
@xmath23 and @xmath24 are the simple left @xmath1-modules associated to the vertices labeled @xmath25 and @xmath26 in the quiver @xmath6 .
let @xmath27 be an alternating bit string in @xmath28 and let @xmath29 denote for the moment formal variables . if @xmath13 is a finite dimensional left @xmath1-module then @xmath30 where @xmath31 is the bit string in @xmath16 given by @xmath32 with @xmath33 and where @xmath34 whenever @xmath22 .
the symbol @xmath35 denotes euler characteristic ( for cohomology with compact support ; see @xcite ) .
section ( 1 ) of this paper begins with a quick survey of partition and tableau combinatorics .
the notion of @xmath36-_parity _ of a standard tableau is defined together with the auxiliary notion of a _
chess tableau_. proposition ( 1 ) proves that the number of standard tableaux of shape @xmath37 and @xmath36-parity @xmath31 equals @xmath38 times the number of chess tableaux of shape @xmath37 and content @xmath39 where @xmath40 is an alternating bit string in @xmath28 and @xmath41 . in section
( 2 ) we give a construction which associates to a pair @xmath42 of ordered partitions and choice of parity @xmath43 a nilpotent left @xmath1-module called the _ skew - shape module_. the modules considered in chapter ( 4 ) of @xcite are particular examples .
_ shape modules _ are defined as skew - modules where the smaller partition @xmath44 is empty . following this section ( 3 )
details a proof of our first theorem : let @xmath13 be a shape module of shape @xmath37 , parity @xmath36 , and dimension @xmath45 . if @xmath46 is a bit string in @xmath16 then the euler characteristic @xmath47 equals the number of standard tableaux @xmath48 of shape @xmath37 whose @xmath36-parity equals @xmath49 .
conjecture ( 1 ) of section ( 3 ) refines theorem ( 1 ) and tallies the number of @xmath50-rational points of @xmath51 when @xmath1 is viewed as an algebra over a finite field @xmath50 with @xmath52 elements .
the _ algebraic loop group _ @xmath3 is the group consisting of all @xmath53-valued @xmath54 matrices @xmath55 with determinant @xmath26 where @xmath53 is the laurent polynomial ring @xmath56 $ ] .
an element @xmath57 is viewed as encoding a regular map @xmath58 given by @xmath59 ; a closed contour or `` loop '' in @xmath60 is obtained upon restricting the map to the circle group @xmath61 in @xmath62 and taking its image hence the name .
the _ maximal unipotent subgroup _
@xmath5 is the subgroup of @xmath3 containing all loops @xmath63 which extend to @xmath25 and for which @xmath64 is an upper triangular unipotent matrix . as @xmath53-valued matrices @xmath65 & \bbb{c}[t ] \\
\\ t\bbb{c}[t ] & 1 + t\bbb{c}[t ] \end{pmatrix } \ \bigg| \
\text{with $ \det(g ) = 1 $ } \
\bigg\}.\ ] ] for @xmath66 and @xmath67 let @xmath68 denote the 1-parameter subgroups defined by @xmath69 the set of elements @xmath70 which factorize as @xmath71 for some alternating bit string @xmath27 in @xmath28 , for some choice of parameters @xmath29 in @xmath62 , and for some @xmath72 is a zariski open subset within @xmath5 . consequently a regular function over @xmath5 is uniquely determined by its values over @xmath70 . in section ( 4 )
we define regular functions @xmath73 indexed by a choice of parity @xmath74 together with a pair of partitions @xmath44 and @xmath37 .
these functions are expressed as minors of the infinite block - toeplitz matrix @xmath75 associated to the argument @xmath57 .
in particular the minors consider in @xcite are of the form @xmath76 .
these minors are shown to satisfy an @xmath36-pieri rule reminiscent of the generalized pieri identities considered by @xcite and as such behave like schur polynomials which carry a parity .
this view is reinforced by remark ( 7 ) in section ( 5 ) which expresses @xmath77 as a generating function for chess tableaux of shape @xmath37 and parity @xmath36 .
section ( 4 ) ends with conjecture ( 2 ) which claims that the type @xmath2-_generalized minors _ of fomin - zelevinsky ( see @xcite ) are among these block - toeplitz matrix minors . in section ( 5 )
the restriction of the minors @xmath78 to the maximal unipotent subgroup @xmath5 of the loop group are studied combinatorially by means of pairwise non - crossing families of paths in a weighted planar network @xmath79 .
proposition ( 3 ) sets up a weight and content preserving bijection between families of non - crossing paths and chess tableaux which is then used to prove the main result of this paper : let @xmath13 be a shape module of shape @xmath37 , parity @xmath36 , and dimension @xmath45 .
let @xmath27 be an alternating bit string in @xmath28 and let @xmath29 be parameters in @xmath62 then @xmath80
by a partition @xmath37 we will mean a non - increasing infinite sequence of non - negative integers @xmath81 with is eventually zero ; as short hand we only write the non - zero terms .
the size @xmath82 of @xmath37 is defined as the sum @xmath83 and we say the @xmath37 is a partition of @xmath84 if @xmath85 . in the context of partitions the symbol @xmath86 will denote the _ empty partition _ defined by @xmath87 whenever @xmath88 . for
a partition @xmath37 let @xmath89 with the understanding that @xmath90 .
given a partition @xmath37 , a choice of parity @xmath74 , and a non - negative integer @xmath91 define the @xmath92 by @xmath93 in this paper a partition @xmath37 of @xmath45 will be synonymous with its young diagram or _ shape _ : this is a left - justified arrangement of @xmath45 boxes into rows whereby the top row contains @xmath94 boxes , the next @xmath95 , and so on .
for instance shape of the partition @xmath96 is @xmath97 each box in a shape will be coordinatized by its row and column positions @xmath98 measured from the top and left respectively .
we employ the convention that the top row and left - most column are counted as row and column zero .
let @xmath37 be a partition and let @xmath99 be a choice of parity .
the * @xmath36-parity * of a box with row and column coordinates @xmath98 in the the shape of @xmath37 is by definition the parity of @xmath100 . by definition two partitions @xmath101 and @xmath102
are ordered @xmath103 if @xmath104 for all integers @xmath105 .
the symbol @xmath106 is justified by the fact that @xmath103 if and only if the shape of @xmath44 is contained in the shape of @xmath37 . in this case let @xmath107 denote the _ skew - shape _ obtained by removing @xmath44 from @xmath37 as shapes .
for example : if @xmath108 and @xmath96 then @xmath44 , @xmath37 , and @xmath109 are depicted respectively by @xmath110 let @xmath37 be a partition and @xmath111 be a non - negative integer @xmath84-tuple in @xmath112 such that @xmath113 . for the purposes of this paper , a _
semi - standard tableau _
@xmath48 of shape @xmath37 and content @xmath114 is an assignment whereby each box in shape of @xmath37 is labeled by an integer @xmath115 in @xmath116 $ ] such that @xmath117 each index @xmath118 $ ] is used exactly @xmath119 times @xmath117 the indices @xmath115 strictly increase when read from left to right in each row @xmath117 the indices @xmath115 weakly increase when read from top to bottom in each column in the event @xmath120 and @xmath121 is a bit string in @xmath16 with @xmath122 we say that @xmath48 is a _
standard tableau of content @xmath114_. note that if in this case the indices @xmath115 of @xmath48 are read from top to bottom in each column they will necessarily strictly increase . by convention a standard tableau @xmath48 of shape @xmath37 without any reference to content
will be understood to be a standard tableau of shape @xmath37 and content @xmath123 where @xmath124 whenever @xmath125 .
below are examples of a standard and a semi - standard tableaux of shape @xmath96 with the later having content @xmath126 : @xmath127 for a partition @xmath37 and a bit string @xmath114 in @xmath16 with @xmath120 and @xmath122 the set of all standard tableaux of shape @xmath37 and content @xmath114 will be denoted @xmath128 ; the set of all standard tableaux of shape @xmath37 will be denoted @xmath129 given an arbitrary non - negative integer @xmath84-tuple @xmath130 in @xmath112 let @xmath131 $ ] denote the _ indicator set _
, i.e. @xmath132 } \ : = \ \big\ { t \ , \big| \ , q_t \ne 0 \big\}\ ] ] for @xmath133 let @xmath134 denote the set of bit strings @xmath114 such that @xmath135}$ ] has cardinality @xmath84
. we may express the indicator set of a bit string @xmath114 in @xmath134 as @xmath135 } = \big\ { r_1 < \cdots < r_m \big\}$ ] . for a partition @xmath37 of @xmath84 and a tableau @xmath48 in @xmath128 let @xmath136 be the tableau in @xmath129 obtained by replacing each @xmath137 in @xmath48 by @xmath115 whenever @xmath138 .
clearly the mapping @xmath139 defines a bijection between @xmath128 and @xmath129 .
assume @xmath37 is a partition of @xmath84 and fix a bit string @xmath140 and express it indicator set as @xmath135 } = \big\ { r_1 < \cdots < r_m \big\}$ ] . a _
@xmath114-step flag of partitions _ in @xmath37 is a weakly increasing sequence of partitions @xmath141 with @xmath142 and @xmath143 such that the shapes of @xmath144 and @xmath145 differ by exactly one corner box whenever @xmath124 and are equal whenever @xmath146 .
associate to a @xmath114-step flag of partitions @xmath147 a standard tableau @xmath48 with content @xmath114 obtained by labeling the box of @xmath37 which is deleted from @xmath145 in order to get @xmath144 by @xmath137 .
for example when @xmath148 and @xmath149 the @xmath114-step flag of partitions ( depicted here in terms of shapes ) @xmath150 corresponds to @xmath151 the correspondence between @xmath114-step flags of partitions in @xmath37 and standard tableaux of shape @xmath37 and content @xmath114 is clearly bijective .
let @xmath37 be a partition of @xmath45 and let let @xmath152 be a choice of parity .
the * @xmath36-parity * of a standard tableau @xmath48 of shape @xmath37 is by definition the bit string @xmath153 in @xmath16 where @xmath154 equals the @xmath36-parity of the box labeled by @xmath115 in @xmath48 .
let @xmath155 denote the set of all such tableaux .
let @xmath37 be a partition , let @xmath66 be a choice of parity , and let @xmath156 be a non - negative integer @xmath72-tuple in @xmath157 such that @xmath158 . a * chess tableau * @xmath48 of shape @xmath37 , parity @xmath36 , and content @xmath39 is a semi - standard tableau of shape @xmath37 and content @xmath39 with the added constraint that the @xmath36-parity of each box in @xmath48 equals the parity of the index @xmath115 labeling that box .
let @xmath159 denote the set of all chess tableaux of shape @xmath37 , parity @xmath36 , and content @xmath39 .
let @xmath160 denote the disjoint union @xmath161 the parity condition for chess tableaux forces the row and column entries to increase strictly .
chess tableaux seem to have been considered first by chow , erisson , and fan in @xcite . among other things ,
they independently interpret chess tableaux in terms of _ rat races _ which are essentially the non - crossing path configurations which we consider in section 5 .
let @xmath37 be a partition of @xmath45 , let @xmath162 be a choice of parity , let @xmath49 be a bit string in @xmath16 , let @xmath163 be an alternating bit string in @xmath16 , and let @xmath156 be a non - negative integer @xmath72-tuple in @xmath157 such that @xmath164 then @xmath165 where @xmath166 is the parity of @xmath167 .
recall the definition of @xmath168 given in equation ( 1 ) of the introduction .
note that @xmath169}$ ] whenever @xmath22 .
clearly @xmath170 whenever @xmath171 .
if @xmath48 be a tableau in @xmath172 then the filling @xmath173 of @xmath37 obtained by replacing each index @xmath115 in @xmath48 by @xmath168 is weakly increasing in both rows and columns .
the parity of @xmath168 equals the parity of @xmath174 whenever @xmath22 .
consequently the @xmath36-parity of the box labeled @xmath115 in @xmath48 equals @xmath154 if and only if the @xmath166-parity of the box labeled @xmath115 in @xmath48 equals the parity of @xmath168 . by assumption
the @xmath36-parity of the box in @xmath48 labeled @xmath115 equals @xmath154 whenever @xmath22 therefore the @xmath166-parity of the box labeled @xmath168 in @xmath173 equals the parity of @xmath168 ; hence @xmath175 .
moreover for each @xmath176 $ ] the cardinality of @xmath177 is exactly @xmath178 and so @xmath179 .
if @xmath180 is a chess tableau in @xmath181 let @xmath48 be a filling of @xmath37 obtained by replacing each label @xmath182 occurring in @xmath180 by some choice of element @xmath115 in @xmath177 with the rule that pre - images in @xmath177 are not reused once they are selected . as indicated in the remark above
, any chess tableau is in fact row and column strict and since @xmath170 whenever @xmath171 it follows that @xmath48 will be a standard tableau .
moreover the @xmath36-parity of the box labeled @xmath115 will be @xmath154 .
consequently the mapping @xmath183 is surjective .
since the construction described in the surjectivity argument prescribes that each index in @xmath184}$ ] is chosen once and since there are @xmath178-choices to be made it follows that are exactly @xmath185 possibilities for each @xmath182 .
the proposition now follows . for a standard tableau @xmath48 of shape @xmath37 and an index @xmath115 let @xmath186 and @xmath187 respectively denote the row and column positions of the box labeled by @xmath115 in @xmath48 let @xmath37 be a partition of @xmath45 and let @xmath15 be a bit string in @xmath16 .
a pair of indices @xmath188 is called a * @xmath49-transposition pair * for a standard tableau @xmath48 of shape @xmath37 if @xmath189 and the tableau @xmath190 obtained by exchanging the positions of @xmath182 and @xmath115 remains standard .
a @xmath49-transposition pair @xmath191 is said to be * grounded * if @xmath192 and @xmath193 and @xmath194 .
define the * ground state * @xmath195 of a standard tableau @xmath48 as the total number of grounded @xmath49-transposition pairs of @xmath48 .
we may always regard a finite dimensional left @xmath1-module @xmath13 as a module over the polynomial ring @xmath196 $ ] whereby @xmath197 . if in addition @xmath13 is nilpotent as a @xmath1-module then its isomorphism type when viewed as a @xmath196$]-module is determined by the partition @xmath37 of @xmath198 which encodes the jordan type of @xmath199 considered as a nilpotent endomorphism on @xmath13 ; we call @xmath37 the _ @xmath200$]-partition type _ of @xmath13 . in this case
we can visualize the @xmath196$]-module structure on @xmath13 using the shape of @xmath37 : each box in the shape corresponds to a basis vector in @xmath13 and each row corresponds to an indecomposable @xmath200$]-summand of @xmath13 with the convention that the action of @xmath201 is depicted as going from right to left in each row .
for example the shape and corresponding @xmath200$]-module associated to the partition @xmath202 are @xmath203_{{\scriptscriptstyle}\delta } v_{{\scriptscriptstyle}0,1 } & \ar[l]_{{\scriptscriptstyle}\delta } v_{{\scriptscriptstyle}0,2 } & \ar[l]_{{\scriptscriptstyle}\delta } v_{{\scriptscriptstyle}0,3 } \\
v_{{\scriptscriptstyle}1,0 } & \ar[l]_{{\scriptscriptstyle}\delta } v_{{\scriptscriptstyle}1,1 } & & } \ ] ] where @xmath204 is the basis vector corresponding to the box in located in row @xmath182 and column @xmath115 of the shape of @xmath37 with the proviso that the top row and left most column of @xmath37 both have coordinates equal to @xmath25 .
it is well known ( e.g. see @xcite chapter ii section 2 ) that if @xmath13 is a nilpotent @xmath196$]-module and @xmath91 is a proper @xmath200$]-submodule then @xmath42 where @xmath44 and @xmath37 are the @xmath200$]-partition types of @xmath91 and @xmath13 respectively . for a finite dimensional
left @xmath1-module @xmath13 of dimension @xmath14 let @xmath205 denote the variety of all @xmath196$]-composition series , i.e. complete flags @xmath18 with @xmath19 and @xmath20 such that @xmath206 is a maximal @xmath200$]-submodule of @xmath207 whenever @xmath22 .
if @xmath13 is a nilpotent @xmath200$]-module of dimension @xmath208 and @xmath200$]-partition type @xmath37 then any @xmath200$]-composition series @xmath209 in @xmath13 gives rise by the remarks made earlier to a flag of partitions @xmath210 in @xmath37 where @xmath145 is the @xmath200$]-partition type of @xmath207 .
the standard tableau @xmath48 of shape @xmath37 which records this flag of partitions will be called the @xmath200$]-_tableau type _ of the @xmath200$]-composition series .
let @xmath13 be a nilpotent left @xmath1-module of dimension @xmath14 and @xmath200$]-partition type @xmath37 . for a standard tableau @xmath48 of shape @xmath37 let @xmath211 denote the collection of all @xmath200$]-composition series of @xmath200$]-tableau type @xmath48 in @xmath205 .
set @xmath212 for @xmath213 .
macdonald relates in @xcite that spaltenstein proved in @xcite that @xmath214 is a smooth irreducible locally closed subvariety of @xmath215 ; moreover @xmath214 is a disjoint union of subvarieties each of which is isomorphic to an affine space .
@xmath216 the subset @xmath217 is a locally closed subvariety of @xmath218 owing to the fact that @xmath214 is locally closed in @xmath205 . the lemma must hold since euler characteristic @xmath35 ( for cohomology with compact support ; see @xcite ) is additive over disjoint unions of locally closed subvarieties and because @xmath219 let @xmath220 be a choice of parity and let @xmath221 and @xmath222 be two partitions such that @xmath42 .
the * skew - shape module * @xmath13 of * skew - shape * @xmath109 and * parity @xmath36 * is the nilpotent left @xmath1-module of dimension @xmath223 for which the @xmath224 , @xmath225 , @xmath226 , and @xmath227 actions are given by @xmath228 and zero otherwise .
if @xmath229 we call @xmath13 instead a * shape module * of shape @xmath37 and parity @xmath36 . in addition
the shape module associated to @xmath230 is by definition the zero module .
this module can be depicted by a grid of north and east pointing arrows , each representing the action of either @xmath224 , @xmath225 , @xmath226 , and @xmath227 , which is subordinate to the skew - shape @xmath109 .
for instance the case of @xmath231 , @xmath108 , and @xmath96 is shown here together with the corresponding skew - shape : @xmath232_{\beta } \\ & v_{{\scriptscriptstyle}1,1 } & v_{{\scriptscriptstyle}1,2 } \ar[l]_{\beta } \ar[u]_{\alpha^ * } & \\ v_{{\scriptscriptstyle}2,0 } & v_{{\scriptscriptstyle}2,1 } \ar[u]_{\alpha^ * } \ar[l]_{\beta } & & \\
v_{{\scriptscriptstyle}3,0 } \ar[u]^{\alpha^ * } & v_{{\scriptscriptstyle}3,1 } \ar[l]^{\alpha } \ar[u]_{\beta^ * } & & \\ v_{{\scriptscriptstyle}4,0 } \ar[u]^{\beta^ * } & & & } \ ] ] the preprojective relations @xmath233 and @xmath234 are clearly valid for any skew - shape module and any parity . clearly the @xmath200$]-partition type of the shape module @xmath13 of shape @xmath37 of any parity @xmath36 is @xmath37 .
the techniques used in this section are adopted from those described by geiss - leclerc - schrer in sections 2 and 3 of their publication @xcite .
in addition we have tried to reconcile our notation with theirs as much as possible .
let @xmath13 be a shape module of shape @xmath37 , parity @xmath36 , and dimension @xmath235 .
assume also that @xmath37 has @xmath72 parts .
let @xmath236 denote the @xmath1-submodule of @xmath13 spanned by the vectors @xmath237 associated to the first column of @xmath37 .
let @xmath6 denote the span of the remaining basis vectors @xmath238 .
evidently both @xmath236 and the quotient module @xmath239 are shape modules of parity @xmath36 and @xmath240 and shape @xmath241 and @xmath242 respectively where @xmath243 provided @xmath244 and zero otherwise .
note @xmath6 is a @xmath245-submodule of @xmath13 where @xmath245 is the @xmath246-subalgebra of @xmath1 generated by @xmath226 , @xmath227 , @xmath11 , and @xmath12 ; moreover @xmath6 and @xmath239 as isomorphic when regarded as @xmath245-modules .
let @xmath247 and @xmath248 denote the quotient and projection maps respectively .
let @xmath249 be a bit string in @xmath250 , let @xmath251 denote the bit string @xmath252 in @xmath253 , and let @xmath254 be an arbitrary bit string in @xmath16 .
let @xmath255 denote the subset of @xmath256 consisting of all @xmath1-composition series @xmath257 such that @xmath258 whenever @xmath22 .
clearly this condition is equivalent to @xmath259 whenever @xmath22 .
an easy application of the zassenhaus butterfly lemma for vector spaces shows that @xmath260 let @xmath13 be a finite dimensional nilpotent left @xmath1-module of @xmath200$]-partition type @xmath37 . for bit strings @xmath261 and @xmath213 @xmath262 where the union is taken over all standard tableaux @xmath48 of shape @xmath37 whose first column content is @xmath263}$ ] .
note first that if @xmath257 is a @xmath1-composition series in @xmath13 then @xmath264 equals the number of parts of the partition @xmath145 associated to @xmath207 and hence equal to the number of rows in the shape of @xmath145 .
recall that box is labeled @xmath115 in the standard tableau @xmath48 associated to @xmath257 if and only if the shape of @xmath145 contains the box while the shape of @xmath144 does not .
consequently a box in the first column of @xmath48 is labeled @xmath115 if and only if the number of rows in the shape of @xmath145 is one more than the number of rows in the shape of @xmath144 .
equivalently a box in the first column of @xmath48 is labeled @xmath115 if and only if @xmath265 .
therefor @xmath115 labels a box in the first column of @xmath48 if and only if @xmath266 .
it follows from lemma ( 2 ) that each @xmath267 is a locally closed subvariety of @xmath17 . for bit strings @xmath261 and @xmath213 let @xmath268 be the map given by @xmath269 where @xmath270 denotes the variety of all _ @xmath249-step @xmath1-composition series _ , i.e. @xmath271 in @xmath236 such that @xmath272 whenever @xmath22 .
bearing a slight abuse of terminology and notation @xmath273 denotes the variety of all @xmath251-step flags @xmath274 in @xmath6 such that @xmath275 is a @xmath1-submodule of @xmath239 and @xmath276 whenever @xmath22 .
express the indicator sets of @xmath249 and @xmath251 as @xmath263 } = \big\{r_1 < \dots < r_k \big\}$ ] and @xmath277 } = \big\ { s_1 < \dots < s_{n - k } \big\}$ ] respectively ; also set @xmath278 .
set @xmath279 with @xmath280 and @xmath281 with @xmath282 .
clearly @xmath283 and @xmath284 are isomorphic as varieties , and so are @xmath285 and @xmath286 . indeed the isomorphisms @xmath287 and @xmath288 in each case are given by @xmath289 any @xmath249-step composition series @xmath290 in @xmath291 determines a @xmath249-step flag of partitions @xmath292 where @xmath293 and @xmath294 .
this in turn corresponds to a standard tableau @xmath295 of shape @xmath241 of content @xmath249 .
let @xmath296 denote the set of all @xmath249-step composition series in @xmath291 associated to the standard tableau @xmath297 .
similarly any @xmath251-step composition series @xmath298 in @xmath299 determines a @xmath251-step flag of partitions @xmath300 in @xmath242 where @xmath301 is the @xmath200$]-partition type of the @xmath1-submodule @xmath275 of @xmath239 .
let @xmath180 be the standard tableau of shape @xmath242 and content @xmath251 corresponding to this @xmath251-step flag .
let @xmath302 denote the set of @xmath251-step composition series in @xmath303 associated to the standard tableau @xmath304 .
clearly @xmath305 and @xmath306 where @xmath307 and @xmath308 are standard tableaux obtained from @xmath295 and @xmath180 by replacing each index @xmath137 in @xmath295 by @xmath115 and each index @xmath309 in @xmath180 by @xmath115 respectively .
let @xmath13 be a shape module of shape @xmath37 and parity @xmath36 .
let @xmath48 be a standard tableau of shape @xmath37 whose left - most column is labeled by indices in @xmath263}$ ] .
let @xmath310 and @xmath311 be the pair of standard tableaux respectively obtained by taking the 1-st column and its complement in @xmath48 , then @xmath312 its is an immediate consequence of the definition of @xmath296 and @xmath302 that @xmath313 whenever @xmath295 and @xmath180 are respectively the left most column and its complement in a standard tableau @xmath48 whose left - most column is labeled by indices in @xmath263}$ ] .
conversely suppose @xmath314 where @xmath315 and @xmath316 .
let @xmath317 and @xmath318 denote the corresponding @xmath249-step and @xmath251-step flags of partitions where @xmath319 and @xmath294 .
note that @xmath295 and @xmath180 will be the left most column and complement in a standard tableau @xmath48 whose left - most column is labeled by indices in @xmath263}$ ] if and only if the number of parts of @xmath320 is less than or equal to the number of parts of @xmath321 whenever @xmath322 .
the latter condition holds @xmath323 @xmath324 whenever @xmath322 @xmath323 @xmath325 whenever @xmath322 @xmath323 @xmath326 whenever @xmath322 .
note that @xmath327 must be a @xmath245-submodule of @xmath13 since @xmath275 is @xmath1-submodule of @xmath239 and because @xmath6 itself is actually a @xmath245-submodule .
in addition @xmath328 contains both @xmath329 and @xmath330 whenever @xmath322 . to see this just note that we may express @xmath331 as @xmath332 and so @xmath333 and @xmath334 where @xmath335 are the idempotents of @xmath1 .
consequently the third implication listed above holds @xmath323 @xmath336 is a @xmath1-composition series in @xmath13 the proposition now follows from the observation @xmath337 let @xmath13 be a shape module of parity @xmath36 and shape @xmath37 and let @xmath48 be any standard tableau of shape @xmath37 . either @xmath338 is empty or else its euler characteristic @xmath339 equals one .
let @xmath14 and assume @xmath37 has @xmath340 parts .
the corollary is clearly valid when @xmath341 because the partition admits only one standard tableau and the corresponding shape module is uniserial in which case @xmath338 is either empty or else is a point and hence its euler characteristic equals one .
assume now that the number of columns of @xmath37 is @xmath342 and hypothesize inductively that the corollary holds for all partitions possessing strictly fewer than @xmath91 columns .
let @xmath48 be a standard tableau of shape @xmath37 and let @xmath249 be the unique bit string in @xmath250 corresponding to the content of the 1-st column of @xmath48 ; namely @xmath266 if and only if @xmath115 labels a box in the first column of @xmath48 . by proposition 1 above @xmath343 maps @xmath338 onto @xmath344 where @xmath315 and @xmath345 are the tableaux obtained by taking the 1-st column and its complement in @xmath48 . as mentioned before @xmath346 and @xmath347 where @xmath307 and @xmath308 are standard tableaux of shapes @xmath241 and @xmath242 respectively .
since both @xmath236 and @xmath239 are shape modules whose associated partitions have shapes with strictly less than @xmath91 columns then we may inductively conclude that both @xmath296 and @xmath302 are either empty or have euler characteristic one .
clearly @xmath338 is empty if and only if either @xmath296 is empty or @xmath302 is empty .
let us suppose that both @xmath296 and @xmath302 are non - empty and thus have euler characteristic equal to one . by adapting lemmas 3.1.1 and 3.2.2 of @xcite
it follows that a composition series @xmath257 inside @xmath338 will be a pre - image with respect to @xmath343 of a pair @xmath348 in @xmath344 if and only if there exists a linear map @xmath349 satisfying @xmath117 @xmath350 and @xmath351 = 0 $ ] for @xmath66 @xmath117 @xmath352\big(q_t\big ) \subseteq p_t$ ] and @xmath353\big(q_t \big ) \subseteq p_t$ ] whenever @xmath22 where @xmath354 such that @xmath355 whenever @xmath322 where @xmath356 note that the commutator identities are necessary and sufficient conditions to insure that @xmath357 is a @xmath1-composition series in @xmath13 .
there is a degree of redundancy in this description of the @xmath343-preimages since @xmath358 if and only if @xmath359 whenever @xmath322 . in this case
we declare @xmath360 and @xmath361 to be equivalent and write @xmath362 . as the authors of @xcite point out in lemma 3.2.2
the kernel and commutator constraints and the equivalence relation @xmath363 are all linear conditions and thus the fiber of preimages under @xmath343 must be affine . the map @xmath343 is known to be a morphism of varieties and thus it descends to a morphism between @xmath338 and @xmath344 .
moreover the euler characteristic of any of its fibers is equal to one owing to the fact that each fiber is isomorphic to an affine space . by proposition 7.4.1 in @xcite
we may conclude in this case that @xmath364 on the other hand euler characteristic is multiplicative therefore @xmath365 let @xmath13 be a shape module of parity @xmath36 , shape @xmath37 , and dimension @xmath45 .
let @xmath49 be a bit string in @xmath16 and let @xmath48 be a standard tableau of shape @xmath37 .
then @xmath366 will be non - empty if and only if the @xmath36-parity of @xmath48 equals @xmath49 .
assume @xmath37 has @xmath72 parts .
the corollary is clearly valid when @xmath341 because the partition admits only on standard tableau and the corresponding shape module is uniserial . in this case
@xmath366 will be non - empty if and only if @xmath154 equals the parity @xmath367 which is precisely the @xmath36-parity of the box in row @xmath115 .
assume now that the number of columns of @xmath37 is @xmath342 and hypothesize inductively that the corollary holds for partitions with strictly fewer than @xmath91 columns .
let @xmath48 be a standard tableau of shape @xmath37 and let @xmath249 be the unique bit string in @xmath250 corresponding to the content of the 1-st column of @xmath48 ; namely @xmath266 if and only if @xmath115 labels a box in the first column of @xmath48 .
write @xmath263 } = \big\ { r_1 , \dots , r_k \big\}$ ] and @xmath277 } = \big\ { s_1 , \dots , s_{n - k } \big\}$ ] . by proposition 1
we know @xmath338 is non - empty if and only if both @xmath296 and @xmath302 are non - empty where @xmath315 and @xmath368 are the tableaux obtained by taking the 1-st column and its complement in @xmath48 . on the other hand @xmath296 and @xmath302
are non - empty if and only if @xmath369 and @xmath370 are respectively non - empty ; recall that @xmath307 and @xmath308 are the standard tableaux obtained from @xmath295 and @xmath180 by replacing the entries @xmath137 and @xmath309 by @xmath115 respectively .
also @xmath280 whenever @xmath371 and @xmath282 whenever @xmath372 . since both @xmath236 and @xmath239 are shape modules of parities @xmath36 and @xmath240 and since their respective shapes @xmath241 and @xmath242 have fewer than @xmath91 columns we may apply the inductive assumption and conclude that @xmath373 equivalently @xmath374 bear in mind that the @xmath375-parity of a box in @xmath308 is equal to the @xmath36-parity of the same box in @xmath180 owing to the fact that first column in @xmath180 , which is situated to the immediate right of the first column in @xmath48 , is counted as column @xmath26 not @xmath25 .
the corollary follows given that @xmath376 = \big\{d_{r_t } \ , | \ , k \geq t \geq 1 \big\ } \sqcup \big\ { d_{s_t } \ , | \ , n - k \geq t \geq 1 \big\}$ ] .
* proof of theorem 1 : * by lemma ( 1 ) we know that @xmath216 by corollary ( 1 ) each euler characteristic @xmath377 contributes either @xmath26 or @xmath25 and corollary ( 2 ) stipulates that only standard tableau @xmath48 whose @xmath36-parity equals @xmath49 add a non - zero contribution . therefore the euler characteristic @xmath378 equals the number of standard tableaux of shape @xmath37 whose @xmath36-parity equals @xmath49 .
taking into account proposition ( 1 ) from the first section we can furthermore conclude : let @xmath13 be a shape module of shape @xmath37 , parity @xmath36 , and dimension @xmath45 .
let @xmath49 be a bit string @xmath16 which can be expressed as @xmath31 for some alternating bit string @xmath40 in @xmath28 and non - negative integer @xmath72-tuple @xmath39 in @xmath157 , then the euler characteristic @xmath379 equals @xmath380 where @xmath166 is the parity of @xmath167 .
let @xmath37 be a partition of @xmath45 , let @xmath74 a choice of parity , and let @xmath49 be a bit string in @xmath16 .
let @xmath13 be a shape module of shape @xmath37 and parity @xmath36 and let @xmath381 , then @xmath382 is an affine space of dimension @xmath383 . alternatively ,
if @xmath52 denotes a power of a prime and @xmath50 is a finite field with @xmath52 elements then the number of @xmath50-rational points of @xmath384 is @xmath385 where , accordinging to definition ( 5 ) of section ( 1 ) , @xmath383 is the number of grounded @xmath49-transposition pairs of @xmath48 .
in this section we recount a standard realization of the loop group in terms of infinite block - toeplitz matrices described in lecture 9 of @xcite .
consider the variable @xmath115 as a coordinate for the circle subgroup @xmath61 in @xmath62 and let @xmath386 be the hilbert space of all square integrable vector - valued functions @xmath387 expressed in the coordinate @xmath115 .
a loop @xmath55 in @xmath3 gives rise to a multiplication operator @xmath388 defined by @xmath389 note : the component functions of the result of this matrix multiplication are still square integrable since , over a compact domain such as @xmath61 , the product of a bounded integrable function ( in this case a polynomial function ) and a square integrable function remains square integrable .
the map @xmath75 is clearly linear and so we may write down the matrix representing @xmath75 with respect to the ( ordered ) fourier basis @xmath390 of @xmath391 where @xmath392 the matrix of @xmath75 will be the @xmath393 matrix whose @xmath394 entry is @xmath395 where @xmath396 and @xmath397 with @xmath398 and where @xmath399 means residue .
alternatively @xmath75 can be expressed in block form @xmath400 where @xmath401 for @xmath402 .
the @xmath54 matrix @xmath403 is precisely the coefficient matrix of @xmath404 in the fourier expansion @xmath405 .
such a @xmath393 matrix in block form with constant block diagonals will be called a _ block toeplitz matrix_. we will identify @xmath75 with its matrix and the map @xmath406 defines an injective homomorphism , denoted @xmath48 , from the loop group @xmath3 to the _ restricted general linear group _ @xmath407 ( see @xcite chapter 6 ) .
let @xmath103 be a pair of ordered partitions , let @xmath74 be a choice of parity , and set @xmath408 .
for an element @xmath409 in the loop group @xmath3 define @xmath410 to be the determinant of the @xmath411 submatrix of @xmath75 whose row and columns sets are @xmath412 and @xmath413 respectively . for each @xmath409 the minor @xmath78 is a polynomial in the matrix entries of @xmath75 and as such is a regular function over @xmath3 . for a non - negative integer @xmath45 let @xmath414 be a short hand notation for the minor @xmath415 where @xmath37 is the partition with @xmath416 and @xmath417 whenever @xmath418 .
evidently @xmath414 is the @xmath419-entry of @xmath75 . in view of the fact that the @xmath394 and @xmath420 entries of @xmath75 are equal
one observes that the following * @xmath36-pieri * rule must hold for @xmath421 : @xmath422 where @xmath423 and @xmath424 is the alternating bit string in @xmath425 starting with @xmath426 and @xmath427 whenever @xmath428 .
here we use the convention that @xmath429 whenever @xmath45 is negative .
for non - negative integers @xmath84 and @xmath45 let @xmath430 be short hand notation for @xmath78 where @xmath44 is the partition with @xmath431 whenever @xmath432 and @xmath433 whenever @xmath434 and where @xmath37 is the partition with @xmath435 whenever @xmath436 and @xmath417 whenever @xmath437 .
up to re - indexing and a _ twist _ as defined in @xcite the minors @xmath438 are the minors considered in conjecture 4.3 of @xcite .
the minors @xmath430 are precisely the * generalized minors * of fomin - zelevinsky defined in @xcite .
this section describes a locally finite but nevertheless infinite version of the fomin - zelevinsky graphical calculus which was originally developed in @xcite and @xcite as a means to parameterize the double bruhat cells of a simply connected simple algebraic group . for @xmath67 and a choice of parity @xmath99 the _ chip diagram _
@xmath439 is is a weighted directed planar graph whose vertices are either sources or sinks : the set of sources and the set of sinks are both indexed by @xmath440 obeying the rule that @xmath441 @xmath45 is connected to sink @xmath84 if and only if either @xmath442 or else @xmath443 and the parity of @xmath45 is @xmath36 .
finite portions of the chip diagrams @xmath444 and @xmath445 are depicted below : @xmath446^{\vdots } & \circ \ , 3 \\ 2 \ , \circ \ar[r ] \ar[ur]^{a } & \circ \ , 2 \\ 1 \ , \circ \ar[r ] & \circ \ , 1 \\ 0 \ , \circ \ar[r]_{\vdots } \ar[ur]^{a } & \circ \ , 0 } \qquad \qquad \qquad \xymatrix { 4 \ , \circ \ar[r]^{\vdots } & \circ \ , 4 \\ 3 \ , \circ \ar[r ] \ar[ur]^{a } & \circ \ , 3 \\ 2 \ , \circ \ar[r ] & \circ \ , 2 \\ 1 \ , \circ \ar[r]_{\vdots } \ar[ur]^{a } & \circ \ , 1 } \ ] ] the diagonal edges carry weight @xmath447 while all other edges are assumed to carry weight @xmath26 .
given an alternating bit string @xmath27 in @xmath28 and a @xmath72-tuple of parameters @xmath448 in @xmath449 let @xmath450 denote the graph obtained by _ concatenating _ the chip diagrams @xmath451 from left to right starting with @xmath452 .
to concatenate @xmath453 and @xmath454 simply graft each sink of @xmath453 with the source of @xmath454 having the same integer label .
the set of sources and sinks of the graph @xmath79 are both indexed by @xmath440 which will be respectively depicted on the left and right .
for example @xmath455 is shown here : @xmath456^{\vdots } & \circ \ar[r ] & \circ \ar[r ] & \circ \ar[r]^{\vdots } & \circ \ 3 \\ 2 \ \circ \ar[r ] & \circ \ar[r ] \ar[ur]^{a_2 } & \circ \ar[r ] & \circ \ar[r ] \ar[ur]^{a_4 } & \circ \ 2 \\ 1 \
\circ \ar[r ] \ar[ur]^{a_1 } & \circ \ar[r ] & \circ \ar[r ] \ar[ur]^{a_3 } & \circ \ar[r ] & \circ \ 1 \\ 0 \ \circ \ar[r]_{\vdots } & \circ \ar[r ] \ar[ur]^{a_2 } & \circ \ar[r ] & \circ \ar[r]_{\vdots } \ar[ur]^{a_4 } & \circ \ 0 } \ ] ] the _ weight matrix _
@xmath457 is the @xmath393 matrix whose @xmath458 entry is the sum of all weights @xmath459 of paths @xmath460 joining the @xmath36-th source to the @xmath461-th sink in @xmath79 . by definition
the weight @xmath459 of a path @xmath460 is the product of the weights of edges comprising the path @xmath460 .
note that the @xmath393 weight matrices @xmath462 are exactly the block toeplitz matrices @xmath463 associated with the element @xmath464 in the loop group @xmath465 for @xmath66 and @xmath466 .
weight matrices are multiplicative in the sense that given two finite subsets @xmath469 and @xmath470 of @xmath440 the famous _ lindstrm lemma _
( see @xcite lemma 1 ) asserts in the present context that the matrix minor @xmath471 of the weight matrix @xmath472 is given by the sum where the sum is taken over all families of paths @xmath474 in @xmath475 whose members are pairwise non - crossing ( i.e. sharing no vertices or edges ) and @xmath476 joins source @xmath477 to sink @xmath478 whenever @xmath479 .
@xmath484 & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] & \bullet \ 3 \\ 2 \ \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] \ar@{-->}[ur]^{a_2 } & \circ \ar@{-->}[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_4 } & \circ \ 2 \\ 1 \
\bullet \ar[r ] \ar@{-->}[ur]^{a_1 } & \bullet \ar[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_3 } & \circ \ar@{-->}[r ] & \bullet \ 1 \\ 0 \ \bullet \ar[r ] & \bullet \ar[r ] \ar@{-->}[ur]^{a_2 } & \bullet \ar[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_4 } & \circ \ 0 } \qquad \xymatrix { 3 \ \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] & \bullet \ 3 \\ 2 \ \circ \ar@{-->}[r ] & \bullet \ar[r ] \ar@{-->}[ur]^{a_2 } & \bullet \ar[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_4 } & \circ \ 2 \\ 1 \
\bullet \ar@{-->}[r ] \ar[ur]^{a_1 } & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] \ar@{-->}[ur]^{a_3 } & \circ \ar@{-->}[r ] & \bullet \ 1 \\ 0 \ \bullet \ar[r ] & \bullet \ar[r ] \ar@{-->}[ur]^{a_2 } & \bullet \ar[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_4 } & \circ \ 0 } \ ] ] @xmath485 & \circ \ar@{-->}[r ] & \bullet \ar[r ] & \bullet \ar[r ] & \bullet \ 3 \\ 2 \ \circ \ar@{-->}[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_2 } & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] \ar@{-->}[ur]^{a_4 } & \circ \ 2 \\ 1 \
\bullet \ar@{-->}[r ] \ar[ur]^{a_1 } & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] \ar@{-->}[ur]^{a_3 } & \circ \ar@{-->}[r ] & \bullet \ 1 \\ 0 \ \bullet \ar[r ] & \bullet \ar[r ] \ar@{-->}[ur]^{a_2 } & \bullet \ar[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_4 } & \circ \ 0 } \qquad \xymatrix { 3 \ \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] & \bullet \ar[r ] & \bullet \ar[r ] & \bullet \ 3 \\ 2 \ \circ \ar@{-->}[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_2 } & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] \ar@{-->}[ur]^{a_4 } & \circ \ 2 \\ 1 \
\bullet \ar@{-->}[r ] \ar[ur]^{a_1 } & \circ \ar@{-->}[r ] & \bullet \ar[r ] \ar@{-->}[ur]^{a_3 } & \bullet \ar[r ] & \bullet \ 1 \\ 0 \ \bullet \ar[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_2 } & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] \ar@{-->}[ur]^{a_4 } & \circ \ 0 } \ ] ] @xmath485 & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] & \bullet \ 3 \\ 2 \ \circ \ar@{-->}[r ] & \bullet \ar[r ] \ar@{-->}[ur]^{a_2 } & \bullet \ar[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_4 } & \circ \ 2 \\ 1 \
\bullet \ar@{-->}[r ] \ar[ur]^{a_1 } & \circ \ar@{-->}[r ] & \bullet \ar[r ] \ar@{-->}[ur]^{a_3 } & \bullet \ar[r ] & \bullet \ 1 \\ 0 \ \bullet \ar[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_2 } & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] \ar@{-->}[ur]^{a_4 } & \circ \ 0 } \ ] ] let @xmath103 be a pair of ordered partitions , set @xmath423 , and let @xmath74 be a choice of parity .
in addition let @xmath487 be an alternating bit string in @xmath28 and let @xmath488 .
define @xmath489 to be the collection of all families of non - crossing paths @xmath490 in @xmath79 which join sources in @xmath491 to sinks in @xmath492 . for @xmath493
let @xmath494 denote those families of non - crossing paths @xmath490 in @xmath495 with weight @xmath496 .
clearly consider now the case where @xmath229 .
as before set @xmath423 and let @xmath504 and @xmath505 whenever @xmath479
. choose @xmath506 in @xmath507 .
an index @xmath115 in @xmath508 $ ] will be called an _ ascent _ of a path @xmath460 in @xmath79 if @xmath460 ascends along a diagonal edge within the @xmath509 component . since @xmath510
it must be the case that the path @xmath476 makes exactly @xmath511 diagonal ascents as it travels from @xmath477 up to @xmath478 in @xmath79 .
let @xmath512 be the list of the ascents of @xmath476 .
since the paths are non - crossing it also follows that the position of the @xmath513-th ascent of @xmath514 must be directly under or to the right of the @xmath513-th ascent of @xmath515 ; equivalently @xmath516 whenever @xmath517 and whenever @xmath518 .
record ( in increasing order and from left to right ) the ascents of @xmath476 in the @xmath45-th row of @xmath37 and repeat this for each @xmath45 . in this way we obtain a semi - standard tableau @xmath519 of shape @xmath37 owing to the observations made above .
moreover @xmath496 if and only if the content of @xmath520 is @xmath39
. for example the following pair @xmath521 of ( highlighted ) non - crossing paths @xmath522 & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] & \bullet \ 4 \\ 3 \ \circ \ar@{-->}[r ] \ar@{-->}[ur ] & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] \ar@{-->}[ur ] & \circ \ar@{-->}[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_5 } & \circ \ 3 \\ 2 \ \circ \ar@{-->}[r ] & \bullet \ar[r ] \ar@{-->}[ur ] & \bullet \ar[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_4 } & \circ \ar@{-->}[r ] & \bullet \ 2 \\ 1 \
\bullet \ar@{-->}[r ] \ar[ur]^{a_1 } & \circ \ar@{-->}[r ] & \bullet \ar[r ] \ar@{-->}[ur ] & \bullet \ar[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_5 } & \circ \ 1 \\ 0 \ \bullet \ar[r ] & \bullet \ar@{-->}[r ] \ar[ur]^{a_2 } & \circ \ar@{-->}[r ] & \circ \ar@{-->}[r ] \ar@{-->}[ur ] & \circ \ar@{-->}[r ] & \circ \ 0 } \ ] ] since @xmath528 allows ascents only from vertices of parity @xmath529 and since @xmath79 is constructed as an alternating concatenation of such graphs it follows that ( 1 ) the parities of the ascents of any path must alternate when read from left to right and ( 2 ) the parity of the first ascent equals the parity of @xmath530 where @xmath513 is index of the source vertex .
the source vertices of @xmath531 are @xmath532 respectively .
consequently @xmath533 .
both injectivity and surjectivity follow from the fact that a path in @xmath79 is uniquely determined by it source , sink , and list of parity alternating ascents .
the refinement of these results to weight and content is obviously true .
is precisely the definition of the _ schur polynomial _
@xmath538 in the variables @xmath29 ; see @xcite .
this observation taken together with the remark made in section ( 4 ) about the @xmath36-pieri rule suggest that the block - toeplitz minors @xmath539 should be viewed as generalized schur polynomials . | we evaluate the geiss - leclerc - schrer @xmath0-map for _ shape modules _ over the preprojective algebra @xmath1 of type @xmath2 in terms of matrix minors arising from the block - toeplitz representation of the loop group @xmath3 .
conjecturally these minors are among the cluster variables for coordinate rings of unipotent cells within @xmath3 . in
so doing we compute the euler characteristic of any _ generalized flag variety _ attached to a shape module by counting standard tableaux of requisite shape and _ parity _ ; alternatively by counting _
chess tableaux _ of requisite shape and content . |
this paper introduces quanfruit v1.1 , a java application available@xcite for free .
( source code included in the distribution . ) recently , farhi - goldstone - gutmann ( fgg ) wrote a paper@xcite that proposes a quantum algorithm for evaluating nand formulas .
quanfruit outputs a quantum circuit for the ffg algorithm .
we say a unitary operator acting on a set of qubits has been compiled if it has been expressed as a seo ( sequence of elementary operations , like cnots and single - qubit operations ) .
seo s are often represented as quantum circuits .
there exist software ( quantum compilers ) like qubiter@xcite for compiling arbitrary unitary operators ( operators that have no a priori known structure ) .
quanfruit is a special purpose quantum compiler .
it is special purpose in the sense that it can only compile unitary operators that have a very definite , special structure .
the quanfruit application is part of a suite of java applications called quansuite .
quansuite applications are all based on a common class library called qwalk .
each quansuite application compiles a different kind of quantum evolution operator .
the applications output a quantum circuit that equals the input evolution operator .
we have introduced 6 other quansuite applications in 2 earlier papers . ref.@xcite
introduced quantree and quanlin .
ref.@xcite introduced quanfou , quanglue , quanoracle , and quanshi .
quanfruit calls methods from these 6 previous applications , so it may be viewed as a composite of them . before reading this paper ,
the reader should read refs.@xcite and @xcite .
many explanations in refs.@xcite and @xcite still apply to this paper .
rather than repeating such explanations in this paper , the reader will be frequently referred to refs.@xcite and @xcite .
the goal of all quansuite applications , including quanfruit , is to compile an input evolution operator @xmath0 .
@xmath0 can be specified either directly ( e.g. in quanfou , quanshi ) , or by giving a hamiltonian @xmath1 such that @xmath2 ( e.g. in quanglue and quanoracle ) .
the standard definition of the evolution operator in quantum mechanics is @xmath3 , where @xmath4 is time and @xmath1 is a hamiltonian . throughout this paper
, we will set @xmath5
so @xmath2 .
if @xmath1 is proportional to a coupling constant @xmath6 , reference to time can be restored easily by replacing the symbol @xmath6 by @xmath7 , and the symbol @xmath1 by @xmath8 .
the input evolution operator for quanfruit is @xmath9 , where h_fruit = .
@xmath10 where @xmath11 for some positive integer @xmath12 .
@xmath13 is proportional to the incidence matrix for a line graph , where the edges of the graph connect states that are consecutive in a gray order .
for example , for @xmath14 , the graph of fig.[fig - line - graph ] yields : h_line = g c|c|c|c|c|c|c|c|c| & & & & & & & & + & 0&1 & & & & & & + & 1&0 & & 1 & & & & + & & & 0&1 & & & 1 & + & & 1&1&0 & & & & + & & & & & 0&1 & & + & & & & & 1&0 & & 1 + & & & 1 & & & & 0&1 + & & & & & & 1&1&0 + , [ eq - h - line ] where @xmath6 is a real number that we will call the * coupling constant*. @xmath15 where @xmath16 for some positive integer @xmath17 .
@xmath18 is proportional to the incidence matrix for a balanced - binary tree graph .
for example , for @xmath19 , the graph of fig.[fig - tree - graph ] yields : h_tree = g c|c|c|c|c|c|c|c|c| & & & & & & & & + & & & & & & & & + & & & 1&1 & & & & + & & 1 & & & 1&1 & & + & & 1 & & & & & 1&1 + & & & 1 & & & & & + & & & 1 & & & & & + & & & & 1 & & & & + & & & & 1 & & & & + , [ eq - h - tr ] where @xmath6 is the same coupling constant as before . @xmath20 .
in fact , h_glue = g .
here @xmath21 labels the god state of the tree , the one with children but no parents .
( @xmath22 labels the dud node ) we will call @xmath23 the * line door*. if @xmath24 , then the tree is connected to a tail of states . for @xmath14 , from fig.[fig - line - graph ] , if @xmath25 , then the tree is connected to the midpoint of the line of states ( runway " ) .
the number of leaves in the tree is half the number of nodes in the tree : @xmath26 . also , @xmath27 for some positive integer @xmath28 .
@xmath29 . in fact , h_oracle = g , where @xmath30 are the inputs to the nand formula .
the dimension of the matrix @xmath31 is not generally a power of two . to represent it as a quantum circuit
, we need to extend it to @xmath32 , where = 2^ , = \{n : n_s , line + n_s , tree2^n } .
[ eq - nb ] define h_glue = h_glue + h_glue^ , h_ora = h_ora + h_ora^. [ eq - htoh ] ( this last equation is fine as an operator statement , but as a matrix statement , @xmath33 and @xmath34 must be padded " with zeros to make the equation true . by `` padding a matrix with zeros '' , we mean embedding it in a larger matrix , the new entries being zeros . )
one can split @xmath31 into two parts , which we call the * bulk hamiltonian * @xmath35 and the * boundary corrections hamiltonian * @xmath36 : h_fruit = h_bulk + h_corr , where h_bulk = h_line + h_tree , h_corr = h_glue + h_ora .
( again , this last equation requires zero padding if considered a matrix equation . ) note that @xmath37=0 $ ] and @xmath38=0 $ ] . for @xmath39 ,
if @xmath40 , we say @xmath41 * approximates ( or is an approximant ) of order * @xmath42 for @xmath0 . given an approximant @xmath43 of @xmath0 , and some @xmath44 , one can approximate @xmath0 by @xmath45 .
we will refer to this as * trotter s trick * , and to @xmath46 as the * number of trots*. for @xmath47 , quanfruit approximates @xmath48 with a suzuki approximant of order @xmath49 that is derived in ref.@xcite .
quanfruit also applies the trotter trick with @xmath50 trots to the @xmath47 approximant of @xmath48 . for @xmath51
, quanfruit always approximates @xmath52 with an approximant of order 3 , that is derived in ref.@xcite .
quanfruit also applies the trotter trick with @xmath53 trots to the @xmath51 approximant of @xmath52 .
ref.@xcite gives exact ( to numerical precision ) compilations of the glue and oracle parts of @xmath54 .
quanfruit uses these compilations , so the order of the suzuki ( or other ) approximant and the number of trots do not arise in quanfruit , for either the glue or the oracle . for @xmath55
, quanfruit also approximates @xmath56 with a suzuki approximant of order @xmath57 . recall that @xmath58 for @xmath59 is the second order suzuki approximant , and higher order ones are defined recursively from this one .
thus , all suzuki approximants are specified by giving two functions of @xmath4 , @xmath60 and @xmath61 . to get a meta " suzuki approximant , we set @xmath62 and @xmath63 .
quanfruit also applies the trotter trick with @xmath64 trots to the @xmath65 approximant of @xmath56 .
fig.[fig - qfruit - main ] shows the * control panel * for quanfruit . this is the main and only window of the application .
this window is open if and only if the application is running
. the control panel allows you to enter the following inputs : file prefix : : : prefix to the 3 output files that are written when you press the * write files * button .
for example , if you insert test in this text field , the following 3 files will be written : + * test_qfru_log.txt * test_qfru_eng.txt * test_qfru_pic.txt line : number of qubits : : : the parameter @xmath12 defined above .
tree : number of qubits : : : the parameter @xmath17 defined above .
coupling constant : : : the parameter @xmath66 defined above . line door : : : the parameter @xmath23 defined above . bands : : : you must enter here an even number of integers separated by any non - integer , non - white space symbols .
say you enter @xmath67 . if @xmath30 for @xmath68 are as defined above , then @xmath69 iff @xmath70 .
each set @xmath71 is a band " . if @xmath72 , the band has a single element .
quanfruit checks that @xmath73 , @xmath74 , and @xmath75 for all @xmath76 .
it also checks that @xmath77 .
( if @xmath78 , bands @xmath79 and @xmath76 can be merged . if @xmath80 , bands @xmath79 and @xmath76 overlap . )
line : number of trots : : : the parameter @xmath81 defined above . line : order of approximant : : : the parameter @xmath82 defined above . tree : number of trots : : : the parameter @xmath83 defined above . tree : order of approximant : : : this parameter is always 3 .
meta : number of trots : : : the parameter @xmath84 defined above .
meta : order of approximant : : : the parameter @xmath85 defined above . the control panel displays the following outputs : number of qubits : : : the parameter @xmath86 defined by eq.([eq - nb ] ) .
number of elementary operations : : : the number of elementary operations in the output quantum circuit .
if there are no loops , this is the number of lines in the english file , which equals the number of lines in the picture file .
when there are loops , the loop k reps:@xmath46 " and next k " lines are not counted , whereas the lines between loop k reps:@xmath46 " and next k " are counted @xmath46 times .
error : : : the distance in the frobenius norm between the input evolution operator and the output quantum circuit ( i.e. , the seo given in the english file ) .
for a nice review of matrix norms , see ref.@xcite . for any matrix @xmath87 ,
its frobenius norm is defined as @xmath88 .
another common matrix norm is the 2-norm .
the 2-norm @xmath89 of @xmath90 equals the largest singular value of @xmath90 .
the frobenius and 2-norm of @xmath90 are related by@xcite : @xmath91 .
message : : : a message appears in this text field if you press * write files * with a bad input . the message tries to explain the mistake in the input .
pressing the * write files * button of the control panel of quanfruit generates 3 files ( log , english , picture ) .
these files are analogous to their namesakes for quantree , quanlin and other quansuite applications .
ref.@xcite explains how to interpret them .
the quansuite applications , based on the qwalk class library , exhibit some code innovations that you will find very helpful . hopefully , these innovations will become commonplace in future quantum computer software . *
* qwalk class library does most of the work in all quansuite applications : * look in the source folder for any of the quansuite applications .
you ll find that it contains only 3 or 4 classes .
most of the classes are in the source folder for qwalk . that s because most of the work is done by the qwalk class library , which is independent of the quansuite application .
* * reusability of seo writers : * look at the class fruitseo_writer in the source folder for quanfruit .
you ll find that fruitseo_writer utilizes the methods glueseo_writer ( ) , oracleseo_writer ( ) , treeseo_writer ( ) , lineseo_writer ( ) , and shiftseo_writer ( ) .
thus , fruitseo_writer delegates its seo writing to methods from the quansuite applications : quanglue , quanoracle , quantree , quanlin and quanshi .
in fact , quanfruit can be viewed as a composite of these simpler quansuite applications .
this reusability of seo writers is made possible by the novel technique described in appendix [ app - pad ] . *
* nested loops : * the english and picture files of quansuite applications can have loops within loops .
this makes the english and picture files shorter , without loss of information .
however , if you want to multiply out all the operations in an english file ( this is what the class seo_reader in qwalk does ) , then having nested loops makes this task more difficult .
seo_reader of qwalk is sophisticated enough to understand nested loops . *
* painless object oriented implementation of suzuki approximants and trotter s trick : * higher order suzuki approximants can be implemented painlessly by using the classes : qwalk / src / suzfunctions and qwalk / src / suzwriter .
see the class quanlin / src / lineseo_writer for an example of how it s done .
essentially , all you have to do is to override the two abstract methods in qwalk / src / suzfunctions .
+ trotter s trick can also be easily implemented in a quansuite application , by using loop and next lines in the english file .
see the write ( ) method of quanlin / src / lineseo_writer for an example .
suppose we know how to compile @xmath93 .
is it possible to use this compilation to compile @xmath94 , where @xmath95 and @xmath96 are square matrices of zeros ?
the answer is yes , as we show next . _
_ s_s & = & + & = & ( _ b-1 ) ( _ b-2 ) ( + 1 ) ( ) h _ . as usual , @xmath102 .
we will say that @xmath1 has been padded with @xmath92 s to obtain @xmath101 .
now let @xmath103 be the unitary operation that shifts state @xmath104 to @xmath105 , with @xmath106 .
the application quanshi gives a compilation of @xmath103 .
using @xmath103 , one can define a matrix @xmath107 from @xmath101 as follows : | this paper introduces quanfruit v1.1 , a java application available for free .
( source code included in the distribution . ) recently , farhi - goldstone - gutmann ( fgg ) wrote a paper arxiv : quant - ph/0702144 that proposes a quantum algorithm for evaluating nand formulas .
quanfruit outputs a quantum circuit for the ffg algorithm . |
two - photon interference ( tpi ) and in particular the the hong - ou - mandel ( hom ) @xcite effect are relevant tools in photonic quantum technology .
the hom dip plays a central role in dense and superdense coding , since it allows to identify two of the four bell states @xcite .
here the crucial parameter is the visibility of the hom dip : the higher the visibility , the higher the mutual information @xcite .
tpi have found applications also in quantum metrology , where frequency entangled states of photons produced by downconversion ( pdc ) represent a resource for clock synchronization @xcite and optical coherence tomography @xcite .
however , the need to improve the resolution performances while using low - coherence interference leads to a halt : if one increases resolution widening the spectrum , the effect of dispersive media all along the path gets larger , thus preventing any further improvement @xcite .
nevertheless , since the width of the hom dip is not affected by dispersive media @xcite , tpi represents a promising candidate to face the resolution puzzle @xcite : the so - called quantum optical coherence tomography @xcite represents an interferometric technique for axial imaging offering several advantages over conventional methods .
this feature of the hom effect , i.e. the independence of the dip width on dispersion , however , holds only if the pump laser is nearly monochromatic .
when a broadband source is used to produce frequency entangled photons , the hom dip becomes sensitive to even order dispersions @xcite , and both width and visibility are degraded . in view of its relevance for quantum technology @xcite , theoretical @xcite and experimental @xcite studies
have been performed to characterize tpi and the hom dip . however , a full analysis of both the visibility and the width of the hom dip as a function of the pump spectrum and of the pdc spectrum is still missing .
the dispersive effects on the propagating photons can be described by a suitable phase function : this is the main topic of the present paper . in particular
, we develop an experimental apparatus to introduce and manipulate a custom phase function with a spatial light modulator , as well as to work with different pump and pdc spectrum . as a matter of fact
, we can simulate high - order dispersive effects of different transmission channels by exploiting the same setup .
indeed , we have performed a complete theoretical and experimental study of how the hom effect is affected by the presence of second and third order dispersions , together with a direct comparison between theoretical and experimental results .
the same setup can be used to compensate this kind of effect , paving the way for optimization in quantum technological applications involving pdc photons and optical fibers .
the paper is structured as follows . in section [
s : tpi ] we review the description of frequency entangled photons obtained by pdc and provide the basic principles of two - photon interference in dispersive media . in section [ s : expa ]
we describe in some details our experimental apparatus , whereas section [ s : expr ] is devoted to the experimental results .
we discuss our findings in section [ s : discuss ] and close the paper with some concluding remarks in section [ s : out ] .
the hom effect occurs in tpi when two frequency - entangled photons generated by pdc interfere at a beam splitter ( bs ) . in order to describe the tpi process
after the propagation in dispersive media , we need to introduce the two - photon state produced via pdc and traveling from the source to the bs . in the presence of a broadband source laser the quantum state of the two photons , i.e. the signal `` @xmath0 '' and the idler `` @xmath1 '' photons , which interfere at the bs ,
is given by ( see , e.g. , @xcite ) : @xmath2 where @xmath3 and @xmath4 are the frequency shifts from the central pump and the central pdc frequencies , respectively , whereas @xmath5 and @xmath6 are the pump and the pdc spectral amplitudes , respectively .
( [ si ] ) we introduced the two phase functions @xmath7 and @xmath8 denoting the phase functions induced by the propagation of the signal and the idler photon with frequencies @xmath9 and @xmath10 , respectively . since , in general , @xmath11 , the effects of the broadband pump on @xmath6 are negligible .
the phase functions are responsible for high - order effects on the propagation of state . in order to enlighten the different contributions , we expand @xmath12 up to the third order in @xmath13 , @xmath14 , namely : @xmath15 where @xmath16 . each coefficient @xmath17 of the expansion plays a well defined role in the photon propagation ( in our analysis we do not consider the constant phase @xmath18 since , as we will see
, it will not contribute to the quantities of interest considered throughout the rest of the paper ) .
the first order coefficient @xmath19 represents the `` time '' spent by the photons to reach the bs and is related to the group velocity .
the second order coefficient @xmath20 refers to the group velocity dispersion experienced by both photons and it is due to the presence of dispersive media . finally ,
the third order dispersion coefficient @xmath21 , though it has usually a small impact in common materials , may have a huge effect when using very long optical fibers . in each path ,
then , the two photons can experience a different dispersion , which may largely affect the tpi process and the occurrence of the hom effect .
more in details , if we introduce the `` delay time '' @xmath22 , the probability of having coincident counts after the bs as a function @xmath23 reads : @xmath24 } \nonumber\\ & -r f(-\omega ) \e^{\i [ \phi_s(\omegap,-\omega)+\phi_i(\omegap,\omega ) ] } \bigg|^2 , \label{eq : ptau } \end{aligned}\ ] ] where @xmath25 and @xmath26 are the transmission and reflection coefficients of the beam splitter . in the following
we will show how we can experimentally manipulate and control the different contributions described above and , in particular , we will investigate their effects on the hom dip and visibility .
in fig . [ f : appar ] we show a schematic diagram of the experimental apparatus . the two frequency - entangled photon state is generated with a pump laser @ @xmath27 nm by using a bbo crystal ( @xmath28 mm thick ) .
the signal and idler photons are then collected by two fiber couplers and sent into single spatial mode and polarization preserving fibers ( smf ) .
a @xmath28 mm thick crystal is enough to couple the maximum amounts of pdc radiation with our fiber couplers ( cs and ci ) , as the interaction length inside the crystal between the pump and the fiber mode is slightly lower than @xmath28 mm @xcite .
the residual of the pump reaches a homemade high - resolution spectrometer ( not shown in the figure ) .
the possibility to stabilize the laser to different temperatures allows to keep the same central wavelength while changing the current ( and , thus , the laser power ) and also to perform two - photon interference measurements with different pump laser widths ( 70 ma and 150 ma to have a narrower and broader pump respectively ) .
when the idler photon enters the coupler ci , it travels entirely through the fiber towards the bs .
conversely , the signal photon , after a short fiber , enters a 4f system @xcite , i.e. propagates in the air , through few optical devices ( the couplers c1 and c2 , the gratings g1 and g2 and lens l1 and l2 ) and a spatial light modulator ( slm ) . at the end of the 4f system the signal photon
is again coupled to a fiber and finally reaches the bs , where it interferes with the idler .
schematic diagram of our the two - photon interference apparatus . from left to right , the pump laser @ 405.5 nm ( blue arrow ) enters the bbo crystal and is downconverted .
pdc photons are collected with couplers ( cs , ci ) .
the idler photon travels until the bs entirely through the fiber ( smf ) .
the signal photon travels also through the 4f optical system before entering again the fiber and reaching the bs .
two single photon detectors , with a time - to - amplitude converter / single - channel analyzer ( tac / sca ) system , detect the coincident counts ( cc ) at the two bs outputs .
the shadowed region on the left , represents an ir laser beam that a lens ( l ) focalizes in the bbo position in order to perform an interference test .
the see the text for details . ]
the slm is a 1d liquid crystal mask ( 640 pixels , 100 @xmath29m / pixel ) and is placed on the fourier plane between the two lenses l1 and l2 of the 4f system ( see fig . [
f : appar ] ) .
the slm can be used to introduce and control a phase function @xmath30 on the signal photon .
the possibility to choose at will @xmath30 allows us to achieve several goals ( even at the same time ) : 1 ) to introduce from second to high - order dispersion effects on the hom dip ; 2 ) to perform a delay - time scan of the coincidence counts rate , avoiding any mechanical stress on the apparatus ( this method to perform a delay scan is crucial for the stability of the 4f coupling efficiency ) ; 3 ) to compensate any dispersion effects introduced by the fibers and the 4f system . as a matter of fact , also
the presence of the 4f optical system on the signal path introduces a phase function @xmath31 and , in particular , due to small shifts of the grating positions with respect to the optimal ones , it introduces second @xcite and third order dispersions on its own .
finally , the signal and idler optical fibers lead to the further phase functions @xmath32 and @xmath33 , respectively . taking into account all the mentioned contributions
, we have : @xmath34 \phi_{\text{i } } ( \omegap,\omega ) & = \phi_{\text{f , i } } ( \omegap,\omega),\end{aligned}\ ] ] and we can rewrite eq .
( [ eq : ptau ] ) as : @xmath35\big\},\\ \approx & \int\!\!\!\int\ ! d\omegap\ , d\omega\ , |a(\omegap)|^2 \big\{t^2 f(\omega)^2 + r^2 f(-\omega)^2 \nonumber\\ & -2\ , r\ , t\,\re\text{e}\big[f(\omega)\overline{f(-\omega)}\nonumber\\ & \hspace{1.5cm}\times\e^{\i 2\omega\tau+i\phi^{(2)}_{\text{tot}}(\omegap,\omega)+i\phi^{(3)}_{\text{tot}}(\omegap,\omega ) } \big]\big\ } , \label{eq : ptausvolta}\end{aligned}\ ] ] where @xmath36 , and the total second and third order dispersion terms are ( note that , as mentioned , the zero order terms cancel out ) : @xmath37 respectively , with : @xmath38 according to our formalism , the quantities appearing at the rhs of the last equation refer to the @xmath39-th order dispersion coefficients due to the fibers ( f ) , to the 4f system ( 4f ) and to the slm ( m ) .
it is worth noting that @xmath40 , where @xmath41 is the @xmath39-th order dispersion of the fiber and @xmath42 is the length difference between the signal and idler optical fiber .
it is straightforward to see that in the presence of an almost monochromatic pump , i.e. , @xmath43 , the only relevant dispersion effect comes from the second term of the rhs in eq .
( [ eq : phi3 ] ) , that is the second order dispersion does not affect the hom dip shape .
the presence of several optical components in the apparatus leads to an overall loss in purity of the polarization quantum state of the photons . in order to assess the purity
, we performed an interferometric measurement : after a strong filtering with neutral density filters , we focus a @xmath44 nm laser in the bbo position and couple the radiation of the resulting divergent beam with the fibers . then we obtain a mach - zehnder interferometer where the action of focusing plays the role of the first bs .
counts are then balanced by carefully positioning the focusing lens : this is crucial in order to have the same amount of direct counts on both paths , thus maximizing the interferogram visibility . if @xmath45 is the polarization mismatching between the two beams reaching the bs , the visibility of the interferogram is given by @xmath46 .
it is now straightforward to show that the visibility of the hom dip @xmath47 is directly linked to the interferogram visibility by @xmath48 in our experiment the measured visibility is @xmath49 , corresponding to a mismatching angle of @xmath50 and a purity factor @xmath51 .
since , given @xmath52 , the coincidence counts probability becomes : @xmath53 then the maximum dip visibility we can achieve with the measured @xmath54 , @xmath55 and a symmetric pdc spectrum , i.e. @xmath56 , turns out to be @xmath57 .
pump and pdc spectra .
panel ( a ) : measured pump spectrum shift about the central wavelength of @xmath27 nm , with 70 ma ( spectrum width @xmath58 nm ) and 150 ma ( spectrum width @xmath59 nm ) current . panel ( b ) : pdc spectrum data points and curve fit , with a slit on the fourier plane ( spectrum width @xmath60 nm ) and without it ( spectrum width @xmath61 nm ) . ] in fig .
[ f : spectra ] we show the pump and pdc spectra we used during our measurements . on the pump side , in fig .
[ f : spectra](a ) one can see the difference in the pump spectrum width between two different current settings : the narrower spectrum ( @xmath62 nm ) corresponds to a current of @xmath63 ma , while a current of @xmath64 ma leads to a broader spectrum ( @xmath65 nm ) .
we measured the pump spectrum with a homemade spectrometer as follows .
we collected the residual of the pump light with a fiber and delivered it towards a collimating lens .
then , a grating ( @xmath66 l / mm ) divides the collimated beam into its spectral components .
these are then focused with a @xmath67 mm lens onto the sensor of a cmos camera placed on the fourier plane .
figure [ f : spectra](b ) shows the pdc spectrum .
we can choose the nominal spectrum ( @xmath61 nm ) or the `` cut '' spectrum ( @xmath60 nm ) with a slit placed on the fourier plane of the 4f system .
we measured pdc spectra with a @xmath68 mm slit on the fourier plane of the 4f system .
we calibrated the slit using a graduated reference on the fourier plane : we found the @xmath44 nm component position for the slit , we measured a dispersion coefficient of @xmath69 nm for a @xmath68 mm slit displacement and finally , in order to measure the spectrum , for each slit position ( and therefore for each wavelength ) we recorded coincidence counts from the detectors .
as mentioned above , one usually desires to have symmetric pdc spectra .
however , fig .
[ f : spectra](b ) shows that we used a non symmetric signal photon spectrum .
it is worth noting that our apparatus allows us to center the pdc spectra on @xmath44 nm : this is possible by transversally moving a mirror that delivers the pump inside the bbo crystal . in this way
we make the pump axis ( that is the pdc cone axis ) be at the exactly same distance from both fiber couplers .
however , we found that the 4f system introduced an asymmetric cut on the pdc spectrum tails , that is translated in an antisymmetric cut on the idler photon spectrum .
this makes the two photons distinguishable and the visibility worsen .
the spectrum we report in fig .
[ f : spectra](b ) correspond to the spectrum giving the highest visibility possible in our configuration : the peaks are not superimposed at 811 nm but the tails are .
overall , this makes the two photons less distinguishable than the situation where the tails are not superimposed .
a slit placed under one of the fiber couplers is used to match signal and idler paths .
the slm delay - time scan method allows us to match the two paths with an error of less than @xmath28 @xmath29 m .
a range of several picoseconds can be covered with an slm delay - time scan .
therefore , before measuring , signal and idler paths were always perfectly matched . in fig .
[ f : prof ] , we report some measured dip profiles ( points ) together with the corresponding theoretical predictions ( lines ) . it is worth noting that the predictions are obtained from eq .
( [ eq : purity ] ) and from the experimental spectra of fig .
[ f : spectra ] without any fitting parameter .
the experimental data are normalized using the mean value at high delay time ( i.e. , outside the hom dip ) .
the time windows for the counts acquisition for each point are @xmath68 s and @xmath70 s , respectively , for the cases at @xmath64 ma and @xmath63 ma . the hom dip in the presence of second and third order dispersion . in panels ( a1 ) , ( a2 ) , and ( a3 ) we show the dip profiles for three different second order dispersion values , from the lower to the higher dip we set @xmath71 fs@xmath72 ( red dotted line and rhombus for theoretical prevision and data points , respectively ) , @xmath73 fs@xmath72 ( blue dashed line and empty circles ) , @xmath74 fs@xmath72 ( black solid line and filled circles ) . in panels ( b1 ) , ( b2 ) , and ( b3 ) we show the dip profiles for three different third order dispersion values , from the lower to the higher dip we set @xmath75 fs@xmath76 ( red dotted line and rhombus ) , @xmath77 fs@xmath76 ( blue dashed line and empty circles ) , @xmath78 fs@xmath76 ( black solid line and filled circles ) .
the corresponding values of @xmath79 and @xmath80 of the pump and pdc spectrum widths are reported in the panels .
all solid lines are theoretical predictions obtained from eq .
( [ eq : purity ] ) . ]
each line in fig .
[ f : prof ] corresponds to a different configuration . on the left column
, we show three different profiles related to three different values of the second order dispersion coefficient ; on the right column , we report the profiles for three different values of the third order dispersion coefficient . in the top panels , ( a1 ) and ( b1 ) , we have results for the broader pump spectrum ( @xmath28 nm ) and normal pdc spectrum ( @xmath81 nm ) ( from now on we refer to these values as to configuration s1 ) .
we found a huge dependence on second and third order dispersion : the dip visibility decreases and its width increases , with increasing dispersion . in the panels ( a2 ) and ( b2 ) , we show results for the case of the narrower pump spectrum ( @xmath82 nm ) and normal pdc spectrum ( @xmath61 nm , configuration s2 ) . in this case , the dependence of the dip shape on the second order dispersion is significantly weaker , while the dependency on the third order dispersion remains the same of the previous case .
this is in agreement with the fact that two - photon interference is affected by second order dispersion only if the pump is not monochromatic , while it is always affected by third order dispersion ( and all odd orders , including the first order ) . in the above cases ,
one may quantify the effects of the third order dispersion on the hom dip by looking at the visibility .
on the bottom line of fig .
[ f : prof ] , panels ( a3 ) and ( b3 ) , we have the case of the broader pump spectrum ( @xmath58 nm ) and the cut pdc spectrum ( @xmath60 nm , configuration s3 ) .
while we still have a visible dependence on the second order dispersion , the dependence on the third order dispersion disappears .
the width of the compensated dip , here , is much larger than in the cases above .
this comes directly from eq . .
a peculiar behaviour can be seen here : symmetric oscillation appear on the dip , caused by the clean cut of the pdc spectrum .
as we mentioned in the introduction , by a suitable choice of the phase function we can simulate the high order dispersion effects of a particular transmission channel . in particualr ,
the values of the second and third order contribution we used in our experiment are related to those obtained in current optical fibers .
for instance , a dispersion of @xmath83 fs@xmath72 is induced by a fused silica fiber @xmath84 mm long ( @xmath85 fs@xmath86 ) . for a third order dispersion of @xmath87 fs@xmath76 , on the other hand , @xmath88 m of fiber
are needed ( @xmath89 fs@xmath90 ) or , for example , @xmath91 cm of znse ( @xmath92 fs@xmath90 ) .
it is worth noting that these lengths are quite small compared to those to be used in telecommunications .
for this reason the full comprehension of these dispersion effects on the hom dip becomes a fundamental a prerequisite .
visibility and fwhm of the hom dip as a function of @xmath93 and @xmath94 for different pump and pdc spectra : pdc @xmath95 nm and @xmath96 nm ( solid lines ) ; dotted line : @xmath97 nm and @xmath96 nm ( dotted ) ; @xmath97 nm and @xmath98 nm ( dashed lines ) . ] in fig .
[ f : andamenti ] we summarize our main results : we can see the behaviour of the dip visibility and full width at half maximum ( fwhm ) as a function of second and third order dispersions .
the visibility and the fwhm are defined with respect to the normalized value and are obtained by using suitable fitting functions .
each plot reports the results for the three different experimental configurations introduce above : dotted red line for s1 , dashed blue line for s2 and solid black line for s3 . in panel ( a1 ) , on the upper left , we see how visibility decreases as the the second order dispersion increases . this worsening gets less sharp if we cut down either the pdc spectrum ( s3 ) or the pump spectrum ( s2 ) . in panel
( a2 ) , on the lower left , we see how the dip fwhm increases ( the dip gets broader ) when the second order dispersion coefficient increases . as already mentioned above , without dispersion the dip fwhm in s3
is larger than in s1 and s2 . in both s2 and s3 ,
the fwhm is much less sensitive to the second order dispersion compared to the case of broad pump and pdc spectra . in fig .
[ f : andamenti](b1 ) and ( b2 ) , on the right , we studied again the changes in the dip visibility and fwhm as a function of the third order dispersion .
we can see how the cut of the pdc spectrum keeps frozen the dip parameters , making them insensitive to any third order dispersion coefficient .
however , we can also clearly see that the visibility and the fwhm do not depend on the pump spectrum .
in this paper we have designed and developed an experimental setup to simulate second and third order dispersion effects on the propagation of pdc photons . by using our innovative apparatus , which is able to introduce and manipulate a custom phase function with a slm , we have analyzed high - order dispersion effects in two - photon interference .
in particular , we have studied theoretically and experimentally , the dependence of both the visibility and the width of the hom dip on both the pump spectrum and on the downconverted photon spectrum .
though we have mostly used our system to introduce high - order dispersion effects in the propagation of the states , the same setup can be used to _ compensate _ such effect .
therefore , on the one hand , our results clarify the role of the different dispersion coefficients and , on the other hand , pave the way for optimization procedures in communications protocols and other applications in quantum technology involving pdc photons and optical fibers .
this work has been supported by unimi through the unimi14 grant 15 - 6 - 3008000 - 609 and the h2020 transition grant 15 - 6 - 3008000 - 625 , and by eu through the collaborative project quprocs ( grant agreement 641277 ) . 20 c. k. hong , z. y. ou , and l. mandel , phys .
lett . * 59 * , 2044 ( 1987 ) .
h. weinfurter , europhys . lett . * 25 * , 559 ( 1994 ) s. l. braunstein , and a. mann , phys .
a * 51 * , r1727 ( 1995 ) .
k. mattle , h. weinfurter , p. g. kwiat , and a. zeilinger , phys .
lett . * 76 * , 4656 ( 1996 ) .
j. t. barreiro , t .- c .
wei , p. g. kwiat , nature * 4 * , 282 ( 2008 ) .
v. giovannetti , s. lloyd , l. maccone , and f. n. c. wong phys .
* 87 * , 117902 ( 2001 ) .
d. huang , e. a. swanson , c. p. lin , j. s. schuman , w. g. stinson , w. chang , m. r. hee , t. flotte , k. gregory , c. a. puliafito , and j. g. fujimoto , science * 254 * , 1178 ( 1991 ) .
a. g. van engen , s. a. diddams , and t. s. clement , appl . opt .
* 37 * , 5679 ( 1998 ) .
a. m. steinberg , p. g. kwiat , r. y. chiao , phys .
lett . * 68 * , 2421 ( 1992 ) .
a. f. abouraddy , m. b. nasr , b. e. a. saleh , a. v. sergienko , and m. c. teich , phys . rev . a * 65 * , 053817 ( 2002 ) . m. b. nasr , b. e. a. saleh , a. v. sergienko , and m. c. teich , phys . rev
. lett . * 91 * , 083601 ( 2003 ) .
k. j. resch , r. kaltenbaek , r. lavoie , and d. n. biggerstaff , proc .
spie * 7465 * , 74650n ( 2009 ) .
k. mattle , h. weinfurter , p. g. kwiat , and a. zeilinger phys .
lett . * 76 * , 4656 ( 1996 ) .
b. dayan , phys .
a. * 76 * , 043813 ( 2007 ) .
j. peina , jr .
, a. v. sergienko , b. m. jost , b. e. a. saleh , and m. c. teich , phys . rev . a * 59
* , 2359 ( 1999 ) .
r. erdmann , d. branning , w. grice , and i. a. walmsley , phys .
a * 62 * , 053810 ( 2000 ) .
m. okano , r. okamoto , a. tanaka , s. ishida , n. nishizawa , and s. takeuchi , phys .
a * 88 * , 043845 ( 2013 ) .
m. okano , h. h. lim , r. okamoto , n. nishizawa , s. kurimura , and s. takeuchi , sc . rep . * 5 * , 18042 ( 2015 ) .
m. atatre , a. v. sergienko , b. m. jost , b. e. a. saleh , and m. c. teich , phys .
* 83 * , 1323 ( 1999 ) .
r. okamoto , s. takeuchi , and k. sasaki , phys .
a. * 74 * , 011801(r ) ( 2006 ) .
l. e. vicent , a. b. uren , r. rangarajan , c. i osorio , j. p torres , l. zhang , and i. a walmsley , new j. phys .
* 12 * , 093027 ( 2010 ) .
s. cialdi , c. vicario , m. petrarca , and p. musumeci , appl . opt . * 46 * , 4959 ( 2007 ) .
a. m. weiner , rev .
. instrum . * 71 * , 1929 ( 2000 ) .
a. smirne , s. cialdi , g. anelli , m. g. a. paris , and b. vacchini phys .
a * 88 * , 012108 ( 2013 ) . | two - photon interference and hong - ou - mandel ( hom ) effect are relevant tools for quantum metrology and quantum information processing . in optical coherence tomography ,
hom effect is exploited to achieve high - resolution measurements with the width of the hom dip being the main parameter . on the other hand ,
applications like dense coding require high - visibility performances . here
we address high - order dispersion effects in two - photon interference and study , theoretically and experimentally , the dependence of the visibility and the width of the hom dip on _ both _ the pump spectrum and the downconverted photon spectrum . in particular , a spatial light modulator is exploited to experimentally introduce and manipulate a custom phase function to simulate the high - order dispersion effects . |
the determination of the equation of state of neutron stars has been an important goal of nuclear physics for more than two decades .
progress toward this goal can be made by establishing astrophysical constraints as well as by improving our understanding of nuclear forces .
many ways of deriving constraints from astrophysics have been explored .
one of the best known is pulse timing of pulsars in binary systems . although binary pulsar timing has made possible stringent tests of general relativity ( see , e.g. , taylor 1992 ) , the @xmath3 masses derived from timing ( see thorsett et al .
1993 ) are allowed by all equations of state based on realistic nuclear physics , and hence these observations have not eliminated any of the equations of state currently being considered .
the highest known neutron star spin frequency , the 643 hz frequency of , is also allowed by all equations of state currently under consideration .
radius estimates based on the energy spectra of type 1 x - ray bursts and on observations of thermal emission from the surfaces of neutron stars are more restrictive in principle but currently have large systematic uncertainties ( see lewin , van paradijs , & taam 1993 ; gelman 1995 ) .
the discovery of high - frequency brightness oscillations from sixteen neutron stars in low - mass x - ray binaries with the _ rossi _ x - ray timing explorer ( rxte ) holds great promise for providing important new constraints .
oscillations are observed both in the persistent x - ray emission and during type 1 x - ray bursts .
the kilohertz quasi - periodic oscillations ( qpos ) observed in the persistent emission have high amplitudes and relatively high coherences ( see van der klis 1997 ) .
a pair of kilohertz qpos is commonly observed from a given source .
although the frequencies of these qpos vary by up to a factor of @xmath42 , the frequency separation @xmath5 between a pair of kilohertz qpos appears to be constant in almost all cases . in both the sonic - point ( miller , lamb , & psaltis 1996 ) and magnetospheric ( strohmayer et al .
1996 ) beat - frequency interpretations , the higher frequency in a pair is the orbital frequency at the inner edge of the keplerian flow , whereas the lower frequency is the beat of the stellar spin frequency with this frequency .
such high orbital frequencies yield interesting bounds on the masses and radii of these neutron stars and interesting constraints on the equation of state of neutron star matter ( miller et al .
1996 ) .
the brightness oscillations observed during type 1 x - ray bursts are different in character from the qpos observed in the persistent emission . only a single oscillation has been observed from each source during a type 1 x - ray burst , and the oscillations in the tails of bursts appear to be highly coherent ( see , e.g. , smith , morgan , & bradt 1997 ) , with frequencies that are always the same for a given source ( comparison of burst oscillations from over about a year shows that the timescale for any variation in the oscillation frequency is @xmath6 yr ; strohmayer 1997 ) . the burst oscillations in ( strohmayer et al .
1996 ; strohmayer 1997 ) and ( swank 1997 ) have frequencies that are consistent with the separation frequencies of their kilohertz qpo pairs . the burst oscillations in ( zhang et al .
1996 ) and ( smith et al . 1997 ; wijnands & van der klis 1997 ) , have frequencies that are consistent with twice the separation frequencies of their kilohertz qpo pairs .
the burst oscillations are thought to be caused by emission from one or two nearly identical emitting poles , producing oscillations at the stellar spin frequency or its first overtone , respectively .
if so , the anisotropy observed at infinity of the radiation emitted from the surface , and hence the amplitude of the burst oscillations , typically decreases with increasing gravitational light deflection .
the burst oscillations can therefore be used to constrain the compactness of neutron stars and the equation of state of neutron star matter ( strohmayer 1997 ) . in this letter we present the results of general relativistic calculations to determine the maximum relative amplitudes of burst oscillations , as a function of stellar compactness , following the procedure described by pechenick , ftaclas , & cohen ( 1983 ; see also chen & shaham 1989 ) .
we go beyond these previous treatments by using a more realistic angular intensity distribution at the stellar surface , by including the angle dependence of the energy spectrum , by considering the effects of x - ray color oscillations , and by computing the amplitudes at overtones .
we show that gravitational light deflection typically reduces the relative amplitude of the oscillations produced by two poles compared to one pole and that effects other than doppler shifts and aberration are second - order in the rotation rate .
we describe our assumptions and method in 2 . in
3 we present our results and discuss the observational signatures that allow the derivation of robust limits on the compactness of the burst sources .
the main purpose of our calculations is to derive constraints on the compactness of the neutron stars that display oscillations during type 1 x - ray bursts . to do this
, we must determine the largest possible amplitude of an oscillation for a given compactness .
we therefore make the following assumptions , which maximize the oscillation amplitude ( these are discussed further in 3 ) : ( 1 ) the radiation that we see comes directly from the stellar surface .
( 2 ) the star radiates from one or two emitting poles ; the rest of the surface is dark .
( 3 ) the pole or poles are pointlike and located in the rotational equator .
( 4 ) if there are two poles , they are identical and antipodal .
( 5 ) the observer s line of sight is in the rotational equator .
the effects on the spacetime of the star s rotation change the oscillation amplitude only to second order ( see lamb & miller 1995 and miller & lamb 1996 for details ) , and are much smaller than first- and second - order doppler effects for the relatively slow rotation rates of interest .
we therefore neglect them , performing all calculations in the schwarzschild spacetime .
we compute the intensity seen by an observer when an emitting pole is at a given rotational phase using the procedure discussed by pechenick et al .
suppose that the emitting pole is at azimuthal angle @xmath7 and that the azimuthal angle of the photon when it reaches infinity is @xmath8 .
the angle @xmath8 is a maximum when the photon is emitted tangentially to the stellar surface , and is then ( pechenick et al .
1983 , eq .
[ 2.13 ] ) @xmath9^{-1/2}\,du\ ; .\ ] ] here and below we use geometrized units in which @xmath10 .
suppose that the observer is at infinity at azimuthal angle @xmath11 .
if @xmath12 the observer can not see the pole . if @xmath13 we solve for @xmath14 , the reciprocal of the impact parameter to the pole , using ( pechenick et al .
1983 , eq .
[ 2.12 ] ) : @xmath15^{-1/2}\,du\ ; .\ ] ] in the schwarzschild spacetime , the angle @xmath16 to the normal at the stellar surface of the light ray that has impact parameter @xmath17 at infinity is given implicitly by @xmath18 , where the maximum impact parameter is @xmath19 ( see , e.g. , abramowicz , ellis , & lanza 1990 ; miller & lamb 1996 ) .
the flux measured by the observer is proportional to @xmath20\cos\psi ( \phi_{\rm obs})$ ] , where @xmath21 is the specific intensity at photon frequency @xmath22 and angle @xmath16 from the normal to the pole . as usual , @xmath23 and @xmath24 are constant , where @xmath25 is the frequency - integrated specific intensity .
hence , changes in @xmath26 and @xmath25 caused by doppler shifts and redshifts are easily computed by calculating the change in the frequency of the photons .
if the star were not rotating and the emission from the pole were isotropic , then @xmath27 would be constant .
however , because the star is rotating , the intensity distribution is aberrated and the photon frequency is doppler - shifted by the factor @xmath28 $ ] , where @xmath29 is the velocity at the stellar equator and @xmath30 . also , the specific intensity is not expected to be isotropic even in the rest frame of the stellar surface , because radiation that diffuses outward through an optically thick scattering atmosphere emerges with a specific intensity distribution @xmath31 that is nearly three times as bright along the normal as in directions tangential to the surface ( see chandrasekhar 1960 , chapter 3 ) .
moreover , the energy spectrum will also , in general , depend on the angle from the normal .
the angle dependence of the actual spectrum is likely to be complicated ( see , e.g. , lewin , van paradijs , & taam 1993 ) .
however , the qualitative effects of angle dependence are illustrated by a simplified emission model in which the spectrum emerging at angle @xmath16 is a blackbody spectrum with a temperature @xmath32 given by @xmath33 , where @xmath34 is the effective temperature ; this approximates the angle dependence of the energy spectrum from a grey atmosphere . using this approach
, we compute the brightness seen by an observer at infinity as a function of rotational phase , for a given specific intensity distribution and rotation rate .
intensities can be added linearly , so addition of a second pole is straightforward .
we fourier analyze the resulting light curve to determine the oscillation amplitudes at different harmonics of the spin frequency .
the results of our calculations are summarized in figure 1 , which shows the fractional rms amplitudes of countrate oscillations in various situations ( the amplitudes of oscillation in the photon number and total energy fluxes are equal for nonrotating stars but are generally unequal for rotating stars ) .
figure 1a compares the oscillation amplitude for isotropic emission and for the peaked emission @xmath31 from a single pole .
the rms amplitude of a non - sinusoidal waveform can be greater than unity , as shown .
figure 1b shows the fractional rms amplitude at the first overtone under the same assumptions , for two identical , antipodal emitting poles .
figure 1c shows the amplitude as a function of the angle between the observer s line of sight and the rotation equator for a single emitting pole .
figure 1d illustrates the generation of harmonics by doppler shifts and aberration .
figure 1e shows the amplitude as a function of photon energy for different assumptions about the surface velocity and the rest - frame angle - dependence of the energy spectrum .
figure 1f plots the amplitude from two emitting poles as detected by instruments with different bandpasses , as a function of the effective temperature of the surface .
these results illustrate several points .
( 1 ) the maximum relative amplitude is significantly greater when one pole is emitting than when two poles are emitting .
( 2 ) the amplitude tends to be higher when the neutron star is less compact , i.e. , when @xmath35 is larger .
the exception occurs for isotropic emission and very compact neutron stars , @xmath36 , in which case gravitational focusing of emission from the pole on the opposite side of the star from the observer can increase the amplitude ( this effect was discussed by pechenick et al .
[ 1983 ] ) .
( 3 ) the angular distribution of the specific intensity makes a large difference .
for example , for @xmath37 , @xmath38 , two emitting poles , and @xmath39 , the maximum amplitude is nearly three times greater than the maximum amplitude for isotropic emission .
( 4 ) stellar rotation affects the _ bolometric _ amplitude only to second order .
however , rotation shifts the _ energy spectrum _ by an amount that is first order in the rotational velocity . as a result
, the measured relative countrate amplitude depends strongly on photon energy if the energy spectrum within the bandpass of the instrument is steep , and may be large even for two emitting poles .
as figure 1f shows , the measured rms amplitude can exceed @xmath470% for two emitting poles .
the amplitude will also depend strongly on the photon energy if , as expected , the rest - frame energy spectrum depends on the angle from the normal , and hence the amplitude is likely to rise steeply with increasing photon energy even if doppler shifts and aberration are unimportant .
( 5 ) rotation generates significant power at overtones of the spin frequency .
( 6 ) the oscillation amplitude is proportional to the cosine of the angle @xmath40 between the line of sight of the observer and the rotation equator and hence the observed amplitude is near its maximum value over a large solid angle : for example , the observed amplitude for @xmath40 between @xmath41 and @xmath42 , which occurs 50% of the time for randomly distributed observers , is within 5% of the maximum .
these results have been computed assuming that the only emission from the surface comes from pointlike poles , that if there are two poles then they are identical and antipodal , and that the emitting pole or poles are in the spin equator .
these assumptions produce the largest possible amplitude .
small deviations from them affect the amplitude of the oscillation only to second order in the deviation , as is also evident from the results of pechenick et al .
for example , if emission at the pole comes from a cap of angular radius @xmath43 instead of from a point , then the amplitude is decreased by an amount @xmath44 .
if there are two poles and they are offset from antipodal by a small angle @xmath45 , then the amplitude at twice the spin frequency drops by @xmath46 . if the axis of the pole or poles is offset from the spin equator by a small angle @xmath47 , the amplitude is diminished by @xmath48 .
if the burst is near its onset or its end , gas will be accreting onto the neutron star and is likely to comptonize the radiation coming from the stellar surface ( see psaltis , lamb , & miller 1995 ) .
this introduces two competing effects .
first , scattering tends to isotropize the radiation , diminishing the observed oscillation amplitude ( see , e.g. , brainerd & lamb 1987 ; kylafis & phinney 1989 ; miller 1997 ) .
second , relatively small changes in the optical depth or temperature of the accreting gas with azimuth can change the shape of the spectrum , producing large changes in the flux at high energies with azimuth , thereby increasing the observed amplitude ( see miller , lamb , & psaltis 1997 ; miller & lamb 1992 ) . in principle , either effect could dominate .
the bounds on stellar compactness that we have computed apply only when we are seeing emission coming directly from the stellar surface .
strong evidence that we _ are _ observing such emission is an energy spectrum that is close to a modified blackbody spectrum , because if the radiation from the surface undergoes significant comptonization by the accreting gas , the spectrum generally has an extended high - energy tail ( see , e.g. , psaltis , lamb , & miller 1995 ) .
thus , if a modified blackbody spectrum is observed , the amplitudes of oscillations during x - ray bursts may be used to derive bounds on the compactness of the star .
we are grateful to dimitrios psaltis for many helpful discussions of the amplitudes of burst oscillations , especially about the role of x - ray color oscillations .
we also thank tod strohmayer and jean swank for discussing their burst observations in advance of publication .
this work was supported in part by nsf grants ast 93 - 15133 and ast 96 - 18524 and nasa grant nag 5 - 2925 at the university of illinois , and nasa grant nag 5 - 2868 at the university of chicago . | the discovery of high - amplitude brightness oscillations at the spin frequency or its first overtone in six neutron stars in low - mass x - ray binaries during type 1 x - ray bursts provides a powerful new way to constrain the compactness of these stars , and hence to constrain the equation of state of the dense matter in all neutron stars . here
we present the results of general relativistic calculations of the maximum fractional rms amplitudes that can be observed during bursts . in particular , we determine the dependence of the amplitude on the compactness of the star , the angular dependence of the emission from the surface , the rotational velocity at the stellar surface , and whether there are one or two emitting poles . we show that if two poles are emitting , as is strongly indicated by independent evidence in and , the resulting limits on the compactness of the star can be extremely restrictive .
we also discuss the expected amplitudes of x - ray color oscillations and the observational signatures necessary to derive convincing constraints on neutron star compactness from the amplitudes of burst oscillations .
# 1 10^#1 @xmath0 # 1_#1 @xmath1 # 1^#1 @xmath2 * = " 002a # 1#1 ps . # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 |
since the successful realization of bose - einstein condensation in a bosonic @xmath1 system @xcite , ultracold atomic systems have attracted considerable interest .
@xcite one of the most active topics in this field is an optical lattice system , @xcite which is formed by loading the ultracold atoms in a periodic potential .
this provides a clean system with quantum parameters which can be tuned in a controlled fashion from weak to strong coupling limits .
in fact , remarkable phenomena have been observed such as the phase transition between a mott insulator and a superfluid in bosonic systems @xcite .
in addition , the superfluid state @xcite and the mott insulating state @xcite have been observed in the fermionic optical lattices , which stimulates theoretical investigations on the quantum states in the optical lattice systems . among them , the possibilty of the supersolid state has been discussed as one of the interesting problems in optical lattice systems .
the existence of the supersolid state was experimentally suggested in a bosonic @xmath2he system , @xcite and was theoretically discussed in the strongly correlated systems such as bosonic systems @xcite and bose - fermi mixtures @xcite . as for fermionic systems
, it is known that a density wave ( dw ) state and an @xmath0-wave superfluid ( ssf ) state are degenerate in the half - filled attractive hubbard model on the bipartite lattice except for one dimension , @xcite which means that the supersolid state might be realizable in principle . however , the degenerate ground states are unstable against perturbations .
in fact , the hole doping immediately drives the system to a genuine ssf state .
therefore , it is difficult to realize the supersolid state in the homogeneous bulk system .
by contrast , in the optical lattice , an additional confining potential makes the situation different . @xcite in our previous paper,@xcite we studied the attractive hubbard model on square lattice with harmonic potential to clarify that the supersolid state is indeed realized at low temperatures .
however , we were not able to systematically deal with large clusters to discuss how the supersolid state depends on the particle density , the system size , etc
. this might be important for experimental observations of the supersolid state in the optical lattice . in this paper
, we address this problem by combining the real - space dynamical mean - field theory ( r - dmft ) with a two - site impurity solver .
we then discuss how stable the dw , ssf and supersolid states are in the optical lattice system .
we also clarify the role of the confining potential in stabilizing the supersolid state .
the paper is organized as follows . in sec .
[ 2 ] , we introduce the model hamiltonian and explain the detail of r - dmft and its impurity solver .
we demonstrate that the supersolid state is indeed realized in a fermionic optical lattice with attractive interactions in sec . [ 3 ] . in sec .
[ 4 ] , we discuss the stability of the supersolid state in large clusters .
we also examine how the phase diagram depends on the particle number .
a brief summary is given in sec .
let us consider ultracold fermionic atoms in the optical lattice with confinement , which may be described by the following attractive hubbard model , @xcite @xmath3 where @xmath4 annihilates ( creates ) a fermion at the @xmath5th site with spin @xmath6 and @xmath7 .
@xmath8 is a nearest neighbor hopping , @xmath9 an attractive interaction , @xmath10 $ ] a harmonic potential and the term @xmath11 indicates that the sum is restricted to nearest neighbors .
@xmath12 is a distance measured from the center of the system and @xmath13 is lattice spacing . here , we define the characteristic length of the harmonic potential as @xmath14 , which satisfies the condition @xmath15 .
the ground - state properties of the hubbard model on inhomogeneous lattices have theoretically been studied by various methods such as the bogoljubov - de gennes equations @xcite , the gutzwiller approximation @xcite , the slave - boson mean - field approach @xcite , variational monte carlo simulations @xcite , local density approximation.@xcite although magnetically ordered and superfluid states are described properly in these approaches , it may be difficult to describe the coexisting phase like a supersolid state in the inhomogeneous system .
the density matrix renormalization group method @xcite and the quantum monte carlo method @xcite are efficient for one - dimensional systems , but it may be difficult to apply them to higher dimensional systems with large clusters .
we here use r - dmft @xcite , where local particle correlations are taken into account precisely . this treatment is formally exact for the homogeneous lattice model in infinite dimensions@xcite and the method has successfully been applied to some inhomogeneous correlated systems such as the surface @xcite or the interface of the mott insulators @xcite , the repulsive fermionic atoms @xcite .
furthermore , it has an advantage in treating the ssf state and the dw state on an equal footing in the strong coupling regime , which allows us to discuss the supersolid state in the optical lattice . in r - dmft , the lattice model is mapped to an effective impurity model , where local electron correlations are taken into account precisely .
the lattice green function is then obtained via self - consistent conditions imposed on the impurity problem .
when one describes the superfluid state in the framework of r - dmft,@xcite the lattice green s function for the system size @xmath16 should be represented in the nambu - gorkov formalism .
it is explicitly given by the @xmath17 matrix , @xmath18_{ij}= -t\delta_{\langle ij \rangle}\hat{\sigma}_z + \delta_{ij}\left [ i\omega_n \hat{\sigma}_0 + \left\{\mu - v\left ( r_i\right ) \right\}\hat{\sigma}_z -\hat{\sigma}_i\left(i\omega_n\right)\right ] , \label{eq : lat}\end{aligned}\ ] ] where @xmath19 is the @xmath20th component of the @xmath21 pauli matrix , @xmath22 the identity matrix , @xmath23 the chemical potential , @xmath24 the matsubara frequency , and @xmath25 the temperature .
the site - diagonal self - energy at @xmath5th site is given by the following @xmath21 matrix , @xmath26 where @xmath27 $ ] is the normal ( anomalous ) part of the self - energy .
in r - dmft , the self - energy at the @xmath5th site is obtained by solving the effective impurity model , which is explicitly given by the following anderson hamiltonian , @xcite @xmath28 where @xmath29 annihilates ( creates ) a fermion with spin @xmath6 in the effective bath and @xmath30 is the impurity level .
we have here introduced the effective parameters in the impurity model such as the spectrum of host particles @xmath31 , the pair potential @xmath32 and the hybridization @xmath33 . by solving the effective impurity model eq .
( [ eq : imp ] ) for each site , we obtain the site - diagonal self - energy and the local green s function .
the r - dmft self - consistent loop of calculations is iterated under the condition that the site - diagonal component of the lattice green s function is equal to the local green s function obtained from the effective impurity model as @xmath34_{ii } = \hat{g}_{imp , i}\left(i\omega_n\right)$ ] . when r - dmft is applied to our inhomogeneous system
, it is necessary to solve the effective impurity models @xmath16 times by iteration .
therefore , numerically powerful methods such as quantum monte carlo simulations , the exact diagonalization method , and the numerical renormalization group method may not be efficient since they require long time to perform r - dmft calculations . in this paper
, we use a two - site approximation @xcite , where the effective bath is replaced by only one site . in spite of this simplicity
, it has an advantage in taking into account both low- and high - energy properties reasonably well within restricted numerical resources @xcite . in the two - site approximation , a non - interacting green s function for the impurity model at the @xmath5th site
is simplified as , @xmath35^{-1}&= & i\omega_n \hat{\sigma}_0-\epsilon_i\hat{\sigma}_z\nonumber\\ & -&v_i \hat{\sigma}_z \frac{1 } { i\omega_n \hat{\sigma}_0-e_i \hat{\sigma}_z - d_i \hat{\sigma}_x } v_i\hat{\sigma}_z,\end{aligned}\ ] ] where the index @xmath36 was omitted .
the effective parameters @xmath37 should be determined self - consistently so that the obtained results properly reproduce the original lattice problem . here
, we use the following equations , @xmath38^{-1 } _ { 11}\right|_{n\rightarrow \infty}\\ v_i&=&\sqrt{\left(a-1\right)\left(\pi^2 t^2+e_i^2+d_i^2\right)}\\ d_i&=&\frac{b}{1-a},\end{aligned}\ ] ] where @xmath39^{-1}_{11}/\pi t$ ] and @xmath40^{-1}_{12}$ ] .
furthermore , the number of particles is fixed in the non - interacting green s function , as @xmath41_{11 } + \frac{1}{2}.\end{aligned}\ ] ] we can determine the effective parameters @xmath37 in terms of these equations . here , the effective particle density is defined as @xmath42 , where @xmath43 is the total number of particles .
this density relates the systems with different sites , number of particles , and curvatures of the confining potentials in the same way as the particle density does for periodic systems with different sites .
we set @xmath44 as a unit of energy and calculate the density profile @xmath45+\frac{1}{2}$ ] and the distribution of the pair potential @xmath46 $ ] , where @xmath47 $ ] is the normal ( anomalous ) green s function for the @xmath5th site . note
that @xmath48 represents the order parameter for the ssf state for the @xmath5th site . in the following , we consider the attractive hubbard model on square lattice with harmonic confinement as a simple model for the supersolid . in this case , it is known that the symmetry of the square lattice is not broken in the dw , ssf , and supersolid states @xcite . therefore , the point group @xmath49 is useful to deal with the system on the inhomogeneous lattice .
for example , when the system with 5513 sites @xmath50 is treated , one can deal with only 725 inequivalent sites .
this allows us to discuss the low temperature properties in larger clusters , in comparison with those with @xmath51 treated in our previous paper .
-2 mm in the optical lattice system with @xmath52 at @xmath53 when @xmath54 and @xmath55 ( from the top to the bottom ) .
, width=264 ] -6 mm -2 mm in the optical lattice system with @xmath52 at @xmath53 when @xmath54 and @xmath55 ( from the top to the bottom ) . ,
width=264 ] -6 mm
by means of r - dmft with the two - site impurity solver , we obtain the results for the system with @xmath52 and @xmath56
. figures [ fig : fig1 ] and [ fig : fig2 ] show the profiles of the local density and the pair potential at @xmath53 . in the non - interacting case @xmath57 ,
fermionic atoms are smoothly distributed up to @xmath58 , as shown in fig .
[ fig : fig1 ] ( a ) . increasing the attractive interaction @xmath59 ,
fermions tend to gather around the bottom of the harmonic potential , as seen in fig .
[ fig : fig1 ] ( b ) . in these cases ,
the pair potential is not yet developed , as shown in figs .
[ fig : fig2 ] ( a ) and ( b ) , and thereby the normal metallic state with short - range pair correlations emerges in the region ( @xmath60 ) .
further increase in the interaction @xmath59 leads to different behavior , where the pair potential @xmath48 is induced in the region with @xmath61 .
thus , the ssf state is induced by the attractive interaction , which is consistent with the results obtained from the bogoljubov - de gennes equation @xcite . in the case with @xmath62 , another remarkable feature is found around the center of the harmonic potential @xmath63 , where a checkerboard structure appears in the density profile @xmath64 , as shown in fig . [
fig : fig1 ] ( c ) .
this implies that the dw state is realized in the region . on the other hand ,
the pair potential @xmath48 is not suppressed completely even in the dw region , as shown in fig .
[ fig : fig2 ] ( c ) .
this suggests that the dw state coexists with the ssf state , _
i.e. _ _ a supersolid state _ appears in our optical lattice system .
the profile characteristic of the supersolid state is clearly seen in the case of @xmath65 .
figures [ fig : fig1](d ) and [ fig : fig2](d ) show that the dw state of checkerboard structure coexists with the ssf state in the doughnut - like region @xmath66 .
by contrast , the genuine ssf state appears inside and outside of the region @xmath67 .
further increase in the interaction excludes the dw state out of the center since fermionic atoms are concentrated around the bottom of the potential for large @xmath59 . in the region ,
two particles with opposite spins are strongly coupled by the attractive interaction to form a hard - core boson , giving rise to the band insulator with @xmath68 , instead of the ssf state .
therefore , the ssf state survives only in the narrow circular region surrounded among the empty and fully occupied states .
we see such behavior more clearly in figs .
[ fig : fig1 ] ( f ) and [ fig : fig2 ] ( f ) .
note that in the strong coupling limit @xmath69 , all particles are condensed in the region @xmath70 . in this section ,
we have studied the attractive hubbard model with the harmonic potential to clarify that the supersolid state is realized in a certain parameter region .
however , it is not clear how the supersolid state depends on the system size and the number of particles . to make this point clear
, we deal with large clusters to clarify that the supersolid state is indeed realized in the following .
in this section , we discuss the stability of the supersolid state in fermionic optical lattice systems , which may be important for experimental observations .
first , we clarify how low - temperature properties depend on the system size , by performing r - dmft for several clusters with different @xmath71 .
we here fix @xmath65 and @xmath72 to obtain the profiles of the local particle density and the pair potential , which are shown in fig .
[ fig : thermodynamic ] . and pair potential @xmath48 as a function of @xmath73 with fixed @xmath74 and @xmath75 , when @xmath65 . , width=264 ]
we note that the distance @xmath76 is normalized by @xmath71 in the figure .
it is found that @xmath77 and @xmath48 describe smooth curves for @xmath78 and @xmath79 , where the genuine ssf state is realized . on the other hand , for @xmath80 , two distinct magnitudes appear in @xmath77 , reflecting the fact that the dw state with two sublattices is realized . since the pair potential is also finite in the region , the supersolid state is realized . in this case
, we deal with finite systems , and thereby all data are discrete in @xmath76 .
nevertheless , it is found that the obtained results are well scaled by @xmath71 although some fluctuations appear due to finite - size effects in the small @xmath71 case .
the effective particle density @xmath81 is almost constant in the above cases .
therefore , we conclude that when @xmath82 , the supersolid state discussed here is stable in the limit with @xmath83 . this result does not imply that the supersolid state is realized in the homogeneous system with arbitrary fillings .
in fact , the supersolid state might be realizable only at half filling @xcite .
therefore , we can say that a confining potential is essential to stabilize the supersolid state in the optical lattice system .
next , we focus on the system with @xmath65 and @xmath52 to discuss in detail how the supersolid state depends on the effective particle density @xmath81 .
the dw state is characterized by the checkerboard structure in the density profile @xmath64 , so that the fourier transform @xmath84 at @xmath85 is appropriate to characterize the existence of the dw state . on the other hand ,
the fourier transform @xmath86 at @xmath87 may represent the rigidity of the ssf state in the system . and @xmath88 as a function of the effective particle density @xmath89 when @xmath90 and @xmath91 .
a broken line represents the local particle density at the center of the lattice in the noninteracting case .
, width=264 ] in fig .
[ fig : rho ] , we show the semilog plots of the parameters normalized by @xmath92 .
it is found that the normalized parameter @xmath93 is always finite although the increase in the attractive interaction monotonically decreases it .
this implies that the ssf state appears in the system with the arbitrary particle number .
in contrast to this ssf state , the dw state is sensitive to the effective particle density as shown in fig .
[ fig : rho ] . these may be explained by the fact that in the system without a harmonic confinement @xmath94 , the dw state is realized only at half filling @xmath95 , while the ssf state is always realized . to clarify this
, we also show the local particle density in the noninteracting case at the center of the system as the broken line in fig .
[ fig : rho ] .
it is found that when the quantity approaches half - filling @xmath96 , @xmath97 takes its maximum value , where the ssf state coexists with the dw state .
therefore , we can say that the supersolid state is stable around this condition . increasing the effective particle density , the band insulating states become spread around the center , while the dw and ssf states should be realized in a certain circular region surrounded among the empty and fully - occupied regions .
therefore , the normalized parameters @xmath98 and @xmath93 are decreased with increase in @xmath81 . on the other hand , in the case with low density @xmath99 ,
the local particle density @xmath100 at each site is far from half filling even when @xmath65 .
therefore , the dw state does not appear in the system , but the genuine ssf state is realized .
these facts imply that the condition @xmath101 is still important to stabilize the supersolid state even in fermionic systems confined by a harmonic potential.@xcite by performing similar calculations for the systems with low , intermediate , and high particle densities ( @xmath102 and @xmath103 ) , we end up with the phase diagrams , as shown in fig .
[ fig : phases ] . and
the density plot represents the profiles of the @xmath0-wave pair potential as a function of the attractive interaction @xmath104 .
the dw state is realized in the shaded area .
the broken lines give a guide to eyes which distinguishes the region with a fractional particle density from the empty and fully - occupied regions .
, width=264 ] we find that increasing the attractive interaction , fermionic particles gradually gather around the center of the system , where the empty state is stabilized away from the center and the band insulating state with fully occupied sites is stabilized .
it is found that the region surrounded among these states strongly depend on the effective particle density @xmath81 .
the increase in the effective particle density shrinks the region , which affects the stability of the ssf , dw and their coexisting states .
in particular , the dw region , which is shown as the shaded area in fig . [
fig : phases ] , is sensitive to the effective particle density , as discussed above .
namely , the local pair potential @xmath48 takes its maximum value around @xmath105 , which may give a rough guide for the crossover region between the bcs - type and the bec - type states .
we note that the dw state appears only in the bcs region @xmath106 .
this implies that the condition @xmath107 is not sufficient , but necessary to stabilize the supersolid state in the attractive hubbard model with an inhomogeneous potential .
we wish to comment on the conditions to observe the supersolid state in the fermionic optical lattice system .
needless to say , one of the most important conditions is the low temperature.@xcite second is the tuning of the effective particle density @xmath108 , which depends on the curvature of the harmonic potential as well as the total number of particles .
this implies that a confined potential play a crucial role in stabilizing the supersolid state in fermionic optical lattice systems .
in addition to this , an appropriate attractive interaction is necessary to stabilize the dw state in the bcs - type ssf state .
when these conditions are satisfied , the supersolid state is expected to be realized at low temperatures .
we have investigated the fermionic attractive hubbard model in the optical lattice with harmonic confinement . by combining r - dmft with a two - site impurity solver
, we have obtained the rich phase diagram on the square lattice , which has a remarkable domain structure including the ssf state in the wide parameter region . by performing systematic calculations ,
we have then confirmed that the supersolid state , where the ssf state coexists with the dw state , is stabilized even in the limit with @xmath109 and @xmath110 const .
we have also elucidated that a confining potential plays a key role in stabilizing the supersolid state .
this work was supported by the grant - in - aid for the global coe program `` the next generation of physics , spun from universality and emergence '' and scientific research [ 20740194 ( a.k . ) , 20540390 ( s.s . ) , 19014013 and 20029013 ( n.k . ) ] from the ministry of education , culture , sports , science and technology ( mext ) of japan .
some of the computations were performed at the supercomputer center at the institute for solid state physics , university of tokyo .
m. greiner , o. mandel , t. esslinger , t. w. hnsch and i. bloch : nature * 415 * ( 2003 ) 39 .
j. k. chin , d. e. miller , y. liu , c. stan , w. setiawan , c. sanner , k. xu , and w. ketterle : nature * 443 * ( 2006 ) 961 .
p. sengupta , l. p. pryadko , f. alet , m. troyer , and g. schmid : phys .
* 94 * ( 2005 ) 207202 ; d. l. kovrizhin , g. v. pai , and s. sinha : europhysics letters * 72 * ( 2005 ) 162 ; v. w. scarola and s. das sarma : phys . rev . lett .
* 95 * ( 2005 ) 033003 ; s. wessel and m. troyer : phys .
* 95 * ( 2005 ) 127205 ; d. heidarian and k. damle : phys
lett . * 95 * ( 2005 ) 127206 ; r. g. melko , a. paramekanti , a. a. burkov , a. vishwanath , d. n. sheng , and l. balents : phys . rev . lett .
* 95 * ( 2005 ) 127207 ; g. g. batrouni , f. hbert , and r. t. scalettar : phys .
* 97 * ( 2006 ) 087209 .
w. metzner and d. vollhardt : phys .
* 62 * ( 1989 ) 324 ; e. mller - hartmann : z. phys .
b * 74 * ( 1989 ) 507 ; t. pruschke , m. jarrell , and j. k. freericks : adv .
* 42 * ( 1995 ) 187 ; a. georges , g. kotliar , w. krauth and m. j. rozenberg : rev .
* 68 * ( 1996 ) 13 . | we study ultracold fermionic atoms trapped in an optical lattice with harmonic confinement by combining the real - space dynamical mean - field theory with a two - site impurity solver . by calculating the local particle density and the pair potential in the systems with different clusters ,
we discuss the stability of a supersolid state , where an @xmath0-wave superfluid coexists with a density - wave state of checkerboard pattern .
it is clarified that a confining potential plays an essential role in stabilizing the supersolid state .
the phase diagrams are obtained for several effective particle densities . |
even if nonlinear distortions are often present , many systems can be approximated by a linear model . when the nonlinear distortions are too large , a nonlinear model is required .
one option is to use a block - oriented model @xcite , which consists of interconnections of linear dynamic and nonlinear static systems .
one of the simplest block - oriented models is the wiener model ( see fig .
[ fig : wiener model ] ) .
this is the cascade of a linear dynamic and a nonlinear static system .
wiener models have been used before to model e.g. biological systems @xcite , a ph process @xcite , and a distillation column @xcite .
some methods have been proposed to identify wiener models , see e.g. @xcite for a nonparametric approach where the nonlinearity is assumed to be invertible , @xcite for the maximum likelihood estimator ( mle ) , @xcite for an approach built on the concept of model complexity control , and @xcite for a frequency domain identification method where memory nonlinearities are considered .
more complex parallel wiener systems are identified in @xcite using a subspace based method , and in @xcite using a parametric approach that needs experiments at several input excitation levels .
some more wiener identification methods can be found in the book edited by giri and bai @xcite .
this paper considers a wiener - schetzen model ( a type of parallel wiener model , see fig .
[ fig : wiener - schetzen model ] ) , but , without loss of generality , the focus is on the approximation of a single - branch wiener system , to keep the notation simple .
the recent book @xcite also provides some industrial relevant examples of wiener , hammerstein , and wiener - hammerstein models , that can be handled by the wiener - schetzen model structure due to its parallel nature .
in fact , a wiener - schetzen model can describe a large class of nonlinear systems arbitrarily well in mean - square sense ( see @xcite for the theory and other practical examples ) . if the system has fading memory , then for bounded slew - limited inputs , a uniform convergence is obtained @xcite . in a wiener - schetzen model ,
the dynamics are described in terms of orthonormal basis functions ( obfs ) .
the nonlinearity is described by a multivariate polynomial .
though any complete set of obfs can be chosen , the choice is important for the convergence rate .
if the pole locations of the obfs match the poles of the underlying linear dynamic system closely , this linear dynamic system can be described accurately with only a limited number of obfs @xcite . in the original ideas of wiener
@xcite , laguerre obfs were used .
laguerre obfs are characterized by a real - valued pole , making them suitable for describing well - damped systems with dominant first - order dynamics . for moderately damped systems with dominant second - order dynamics ,
using kautz obfs is more appropriate @xcite .
we choose generalized obfs ( gobfs ) , since they can deal with multiple real and complex valued poles @xcite .
this paper considers the approximation of a wiener system with finite - order iir ( infinite impulse response ) dynamics and a polynomial nonlinearity by a wiener - schetzen model that contains a limited number of gobfs .
the system poles are first estimated using the best linear approximation ( bla ) @xcite of the system .
next , the gobfs are constructed using these pole estimates .
the coefficients of the multivariate polynomial are determined with a linear regression .
the approach can be applied to parallel wiener systems as well .
as the estimation is linear - in - the - parameters , the wiener - schetzen model is well suited to provide an initial guess for nonlinear optimization algorithms , and for modeling time - varying and parameter - varying systems .
the analysis in this paper is a starting point to tackle these problems .
the contributions of this paper are : * the proposal of an identification method for wiener systems with finite - order iir dynamics ( the initial ideas were presented in @xcite ) , * a convergence analysis for the proposed method .
the paper is organized as follows .
the basic setup is described in section [ sec : setup ] .
section [ sec : identification procedure ] presents the identification procedure and the convergence analysis .
section [ sec : discussion ] discusses the sensitivity to output noise .
the identification procedure is illustrated on two simulation examples in section [ sec : illustration ] .
finally , the conclusions are drawn in section [ sec : conclusion ] .
this section first introduces some notation , and next defines the considered system class and the class of excitation signals .
afterwards , the wiener - schetzen model is discussed in more detail .
a brief discussion of the bla concludes this section .
this section defines the notations @xmath3 , @xmath4 , and @xmath5 .
the sequence @xmath6 , converges to @xmath7 in probability if , for every there exists an @xmath8 such that for every .
we write @xmath9 the notation @xmath10 is an @xmath11 indicates that for @xmath12 big enough , , where @xmath13 is a strictly positive real number .
the notation @xmath14 is an @xmath15 indicates that the sequence @xmath16 is bounded in probability at the rate @xmath17 .
more precisely , , where @xmath18 is a sequence that is bounded in probability . the data - generating system is assumed to be a single input single output ( siso ) discrete - time wiener system ( see fig .
[ fig : wiener model ] ) .
this is the cascade of a linear time - invariant ( lti ) system @xmath19 and a static nonlinear system @xmath20 . in this paper
, @xmath19 is restricted to be a stable , rational transfer function , and @xmath20 to be a polynomial .
the class @xmath21 is the set of proper , finite - dimensional , rational transfer functions , that are analytic in and squared integrable on the unit circle .
[ assumption : g is a rational tf ] the lti system . [ assumption : known order of g ] the order of is known .
the poles of @xmath19 are denoted by @xmath22 ( ) .
[ assumption : f non - even ] the function @xmath20 is non - even around the operating point . [
assumption : f polynomial of known degree ] the function @xmath20 is a polynomial of known degree @xmath23 : @xmath24 more general functions can be approximated arbitrarily well in mean - square sense by over any finite interval .
the input @xmath25 and the output are measured at time instants ( ) . in this paper , random - phase multisine
@xcite excitations are considered .
[ def : random - phase multisine ] a signal @xmath25 is a random - phase multisine if @xcite @xmath26 with , the maximum frequency of the excitation signal , the number of frequency components , and the phases @xmath27 uniformly distributed in the interval @xmath28
. the amplitudes @xmath29 can be chosen by the user , and are normalized such that @xmath25 has finite power as @xcite .
[ def : excitation signal class ] the class of excitation signals @xmath30 is the set of random - phase multisines @xmath25 , having normalized amplitudes , where is a uniformly bounded function with a countable number of discontinuities , and if or .
[ assumption : excitation signal ] the excitation signal . for simplicity ,
the excitations in this paper are restricted to random - phase multisines . however , as shown in @xcite , the theory applies for a much wider class of riemann - equivalent signals . in this case , these are the extended gaussian signals , which among others include gaussian noise . the system is modeled with a wiener - schetzen model ( fig . [
fig : wiener - schetzen model ] ) , where we choose to be gobfs . in @xcite
, it is shown how a set of poles gives rise to a set of obfs [ eq : gobfs ] @xmath31 \quad .\ ] ] if the poles @xmath32 result from a periodic repetition of a finite set of poles , the gobfs are obtained , with poles @xmath33 the gobfs form an orthonormal basis for the set of ( strictly proper ) rational transfer functions in @xmath21 @xcite .
one extra basis function is introduced , namely , to enable the estimation of a feed - through term and as such also to enable the estimation of static systems @xcite .
this extra basis function is still orthogonal with respect to the other basis functions ( see appendix [ app : orthonormal basis ] ) .
the lti system @xmath19 can thus be represented exactly as a series expansion in terms of the basis functions @xmath34 : @xmath35 let be a truncated series expansion , with , and @xmath2 the number of repetitions of the finite set of poles . recall that @xmath22 ( ) are the true poles of @xmath19 , and let @xmath36 then there exists a finite , such that for any , @xcite @xmath37 which shows that @xmath19 can be well approximated with a small number of gobfs if the poles @xmath38 are close to the true poles @xmath22 .
the pole locations @xmath22 will be estimated by means of the bla of the system .
the bla of a system is defined as the linear system that approximates the system s output best in mean - square sense @xcite .
the bla of the considered wiener system is equal to @xcite @xmath39 where the constant @xmath40 depends upon the odd nonlinearities in @xmath20 and the power spectrum of the input signal .
a similar result for gaussian noise excitations results from bussgang s theorem @xcite . under assumption [ assumption : f non -
even ] , @xmath40 is non - zero .
this section formulates the identification procedure and provides a convergence analysis in the noise - free case .
the influence of output noise is analyzed in section [ sec : discussion ] .
the basic idea is that the asymptotic bla ( ) has the same poles as @xmath19 ( see ) .
the poles calculated from the estimated bla are thus excellent candidates to be used in constructing the gobfs .
since the bla will be estimated from a finite data set ( @xmath1 finite ) , the poles calculated from the estimated bla will differ from the true poles .
extensions of the basis functions ( ) will be used to compensate for these errors ( see ) .
the identification procedure can be summarized as : 1 .
estimate the bla and calculate its poles .
2 . use these pole estimates to construct the gobfs .
3 . estimate the multivariate polynomial coefficients .
these steps are now formalized and the asymptotic behavior ( ) of the estimator is analyzed .
first , the situation without disturbing noise is considered .
the influence of disturbing noise is discussed in section [ sec : discussion ] .
first , a nonparametric estimate of the bla is calculated .
since the input is periodic , the bla is estimated as , in which @xmath41 and @xmath42 are the discrete fourier transforms ( dfts ) of the output and the input . for random excitations , the classical frequency response estimates ( division of cross - power and auto - power spectra ) can be used @xcite , or more advanced frf measurement techniques can be used , like the local polynomial method @xcite .
next , a parametric model is identified using a weighted least - squares estimator @xcite [ eq : theta_hat ] @xmath43 where the cost function @xmath44 is equal to @xmath45 here , is a deterministic , @xmath46-independent weighting sequence , and @xmath47 is a parametric transfer function model @xmath48 with the constraint to obtain a unique parameterization . under assumption [ assumption : known order of g ] , we put .
eventually , the poles @xmath49 ( ) of the parametric model @xmath50 are calculated .
before we derive a bound on in lemma [ lemma : bound on delta_p ] , a regularity condition on the parameter set @xmath46 is needed .
[ assumption : regularity condition theta ] the parameter set @xmath46 is identifiable if the system is excited by . the existence of a uniformly bounded convergent volterra series @xcite is needed as well ( see appendix [ app : volterra series ] for more details ) .
[ assumption : convergent volterra series ] there exists a uniformly bounded volterra series whose output converges in least - squares sense to the true system output for .
[ lemma : bound on delta_p ] consider a discrete - time wiener system , with an lti system @xmath19 and a static nonlinear system @xmath20 .
let @xmath22 ( ) be the poles of @xmath19 and @xmath49 be the pole estimates , obtained using the weighted least - squares estimator . then under assumptions , and , is an .
let @xmath51 be the `` true '' model parameters , such that @xmath52 with @xmath53 and @xmath54 polynomials of degree @xmath55 .
the roots of @xmath56 are equal to the true poles @xmath22 ( ) .
if these poles are all distinct , the first - order taylor expansion of @xmath56 results in @xcite [ eq : sensitivity poles ] @xmath57 where @xmath58 , and where @xmath59 follows from @xmath60 under assumptions , it is shown in @xcite that . for the considered output disturbances ( no noise in this section , filtered white noise in section [ sec : discussion ] ) , the least - squares estimator is a mle @xcite . from the properties of the mle , it follows that @xcite @xmath61 then from and , it follows that @xmath62 this concludes the proof of lemma [ lemma : bound on delta_p ] .
lemma [ lemma : bound on delta_p ] shows that good pole estimates are obtained .
though no external noise is considered in this section , the probability limits in this paper are w.r.t .
the random phase realizations of the excitation signal .
next , the gobfs are constructed with these pole estimates ( see , with ) , and the intermediate signals ( ) ( see fig . [
fig : wiener - schetzen model ] ) are calculated .
the following lemma shows that the true intermediate signal @xmath63 can be approximated arbitrarily well by a linear combination of the calculated signals @xmath64 .
[ lemma : bound on delta_x ] consider the situation of lemma [ lemma : bound on delta_p ] .
let , and let @xmath34 ( ) be gobfs , constructed from the finite set of poles .
let , and denote .
then under assumptions , and , is an . from and
, it follows that @xmath65 is an @xmath66 .
it then follows from that , and thus is an .
finally , the coefficients of the multivariate polynomial are estimated .
let [ eq : y_beta ] @xmath67 be the output of , where the coefficients of the polynomial are chosen to be @xmath68 they are estimated using linear least - squares regression : [ eq : beta_hat ] @xmath69 where the regression matrix @xmath70 is equal to @xmath71 we will now show that the estimated output converges in probability to @xmath72 as . [
theorem : convergence y_hat ] consider the situation of lemma [ lemma : bound on delta_x ] .
let and , where the coefficients @xmath73 are obtained from the least - squares regression . then under assumptions , is an .
the exact output @xmath74 + \delta y\\ & = \begin{aligned}[t ] & \tilde{\beta}_{dc}\\ & + \sum_{p = 1}^{q } \left ( \sum_{i_1 = 0}^{n } \sum_{i_2 = i_1}^{n } \cdots \sum_{i_p = i_{p - 1}}^{n } \tilde{\beta}_{i_1 , \ldots , i_p } x_{i_1 } \cdots x_{i_p } \right)\\ & + \delta y \end{aligned}\\ & = y_{\tilde{\beta } } + \delta y \end{aligned}\ ] ] where @xmath75 is the output of a multivariate polynomial , in which the coefficients @xmath76 follow from the true coefficients @xmath77 of @xmath20 and the true coefficients @xmath78 of the series expansion of @xmath19 .
the truncation of this series expansion is taken into account by the term , which just as is an @xmath0 .
note that the coefficients @xmath76 can be obtained as the minimizers of the artificial least - squares problem @xmath79 the estimated coefficients are equal to @xmath80 we now show that is an .
each element in @xmath70 is an @xmath81 , due to the normalization of the excitation signal ( see assumption [ assumption : excitation signal ] ) .
consequently , each element in the matrix is the sum of @xmath12 terms that are an @xmath81 , so is an @xmath82 .
the elements in the matrix are thus an @xmath83 .
each element in the vector is an @xmath0 .
consequently , each element in the vector is the sum of @xmath12 terms that are the product of an @xmath81 and an @xmath0 .
the elements in the vector are thus an . as a consequence , . and
thus @xmath84 this concludes the proof of theorem [ theorem : convergence y_hat ] .
theorem [ theorem : convergence y_hat ] shows that the estimated output converges in probability to the exact output with only a finite number of basis functions .
the convergence rate increases if @xmath2 is increased .
the multivariate polynomial @xmath85 is implemented in terms of hermite polynomials @xcite to improve the numerical conditioning of the least - squares estimation in @xcite . as this has no consequences for the result in theorem [ theorem : convergence y_hat ] ,
ordinary polynomials are used throughout the paper to keep the notation simple .
in this section , the sensitivity of the identification procedure to output noise is discussed .
noise on the intermediate signal @xmath63 is not considered here . in general
, this would result in biased estimates of the nonlinearity .
more involved estimators , e.g. the mle , are needed to obtain an unbiased estimate @xcite . in the case of filtered white output noise , with @xmath86 a sequence of independent random variables , independent of @xmath25 , with
zero mean and variance @xmath87 , and with @xmath88 a stable monic filter , the exact output .
the estimated coefficients are then equal to ( cfr . ) @xmath89 the columns of @xmath70 are filtered versions of the known input signal @xmath25 , which was assumed independent of @xmath90 .
it is thus clear that the noise @xmath90 is uncorrelated with the columns of @xmath70 .
consequently , each element in the vector is the sum of @xmath12 uncorrelated terms that are the product of an @xmath81 and an @xmath91 .
the elements in the vector are thus an @xmath92 , as a consequence , . the error on the estimated output due to the noise is thus independent of the number of repetitions @xmath2 .
increasing @xmath2 allows to tune the model error such that it disappears in the noise floor .
in this section , the approach is illustrated on two simulation examples .
the first one considers the noise - free case , and illustrates the convergence rate predicted by theorem [ theorem : convergence y_hat ] .
the second one compares the proposed method to the so - called approximative prediction error method ( pem ) @xcite , as implemented in the matlab system identification toolbox @xcite .
consider a siso discrete - time wiener system with @xmath93 and @xmath94 the system is excited with a random - phase multisine ( see ) with and @xmath95 the sampling frequency ; ; and the amplitudes chosen equal to each other and such that the rms value of @xmath25 is equal to @xmath96 .
the system is identified using the identification procedure described in section [ sec : identification procedure ] .
no weighting is used to obtain a parametric estimate of the bla , i.e. . a random - phase multisine with is used for the validation .
fifty monte carlo simulations are performed , with each time a different realization of the random phases of the excitation signals . the results in fig .
[ fig : monte carlo ] show that the convergence rate of agrees with what is predicted by theorem [ theorem : convergence y_hat ] . the convergence rate increases with an increasing number of repetitions of the basis functions .
these results generalize to parallel wiener systems as well . along the @xmath97 monte carlo simulations ( full line ) and its one standard deviation confidence interval ( filled zone ) .
the predicted convergence rate is indicated by the dashed lines , which are an @xmath98 .
[ fig : monte carlo],scaledwidth=40.0% ] the second example is inspired by the second example in @xcite .
it is a discrete - time siso wiener system with a saturation nonlinearity .
the system is given by @xmath99 where the input @xmath100 and the output noise @xmath101 are gaussian , with zero mean , and with variances and respectively .
the coefficients @xmath102 and @xmath103 are equal to @xmath104 and @xmath105 , respectively . compared to the example in @xcite
, no process noise was added to @xmath63 since the gobf approach can not deal with process noise .
moreover , the output noise variance was lowered from to as the large output noise would otherwise dominate so much that no sensible conclusions could be made .
one thousand monte carlo simulations are performed , with each time an estimation and a validation data set of data points each . in case of the approximative pem method @xcite
, the true model structure is assumed to be known , and the true system belongs to the considered model set . in case of the proposed gobf approach , the order of the linear dynamics is assumed to be known . a model with and one with
is estimated .
the local polynomial method @xcite is used to estimate the bla .
the nonlinearity is approximated via a multivariate polynomial of degree @xmath106 , in order to capture both even and odd nonlinearities .
note that in this case , the true system is not in the model set . the results in fig .
[ fig : histogram nrmse output noise ] show that the approximative pem method performs significantly better than the gobf approach .
note that the approximative pem method used full prior knowledge of the model structure , while no prior knowledge on the nonlinearity was used in the gobf approach .
still , it is able to find a decent approximation . a better approximation can be obtained by representing the nonlinearity with another basis function expansion that is more appropriate to the nonlinearity at hand .
this , however , is out of the scope of this paper .
finally , for a single - branch wiener system , the shape of the output nonlinearity can be determined as follows .
motivated by lemma [ lemma : bound on delta_x ] and , an estimate of the coefficients @xmath107 can be obtained as @xmath108 the signal is then , up to an unknown scale factor @xmath40 , approximately equal to @xmath63 .
the shape of the nonlinear function @xmath109 can then be determined from a scatter plot of @xmath110 and @xmath72 ( see fig . [
fig : scatter plot ] ) . on the validation data sets for the @xmath111 monte carlo simulations .
pem indicates the results for the approximative prediction error method , while gobf ( 0 ) and gobf ( 1 ) indicate the results for the proposed gobf approach with and .
[ fig : histogram nrmse output noise],scaledwidth=40.0% ] and @xmath72 reveals the shape of the saturation nonlinearity ( gobf approach with , last monte carlo simulation ) .
[ fig : scatter plot],scaledwidth=40.0% ] to make a fair comparison , the example is modified such that the system is in the model class for both of the considered approaches .
the nonlinearity in is changed to a third - degree polynomial that best approximates the saturation nonlinearity on all the estimation data sets .
the approximative pem method now estimates a third - degree polynomial nonlinearity . in this case , the gobf approaches have a similar performance as the approximative pem approach ( see fig .
[ fig : histogram nrmse polynomial ] ) .
finally , in order to determine the number of repetitions @xmath2 , one can easily estimate several models for an increasing @xmath2 , and compare the simulation errors on a validation data set .
once the simulation error increases , the variance error outweighs the model error , and one should select less repetitions . here , the normalized rms error for is lower than the normalized rms error for in @xmath112 out of the @xmath111 cases . in the remaining @xmath113 cases , one would select a model with . on the validation data sets for the @xmath111 monte carlo simulations ( polynomial nonlinearity ) .
[ fig : histogram nrmse polynomial],scaledwidth=40.0% ]
an identification procedure for siso wiener systems with finite - order iir dynamics and a polynomial nonlinearity was formulated and its asymptotic behavior was analyzed in an output - error framework .
it is shown that the estimated output converges in probability to the true system output .
fast convergence rates can be obtained .
the identification procedure is mainly linear in the parameters .
the proposed identification procedure is thus well suited to provide an initial guess for nonlinear optimization algorithms .
the approach can be applied to parallel wiener systems as well .
this work was supported by the erc advanced grant snlsid , under contract 320378 .
the orthogonality of @xmath114 with respect to the set @xmath115 ( ) can be shown either by working out the inner products @xmath116 ( see below ) or by choosing a pole structure in the shifted basis functions .
consider the basis functions @xmath117 , given by for and .
we now prove that they are orthogonal to , by showing that the inner product @xmath118 is equal to zero .
@xmath119 denotes the unit circle .
1 . @xmath120 + @xmath121 2 .
@xmath122 + @xmath123 the obfs @xmath34 are linear combinations of the basis functions @xmath117
@xcite and are thus orthogonal to . since the norm of @xmath124 is equal to one , the set @xmath115 ( ) is a set of obfs .
a volterra series generalizes the impulse response of an lti system to a nonlinear time - invariant system via multidimensional impulse responses .
the input - output relation of a volterra series is split in different contributions of increasing degree of nonlinearity @xcite for periodic excitations , the output fourier coefficient @xmath129 at frequency @xmath130 is @xcite @xmath131 with @xmath132 . here , @xmath133 is the symmetrized frequency domain representation of the volterra kernel of degree @xmath128 .
the volterra series is uniformly bounded if @xcite @xmath134 with , and @xmath135 as in definition [ def : excitation signal class ] . | many nonlinear systems can be described by a wiener - schetzen model . in this model ,
the linear dynamics are formulated in terms of orthonormal basis functions ( obfs ) .
the nonlinearity is modeled by a multivariate polynomial . in general , an infinite number of obfs is needed for an exact representation of the system .
this paper considers the approximation of a wiener system with finite - order infinite impulse response dynamics and a polynomial nonlinearity .
we propose to use a limited number of generalized obfs ( gobfs ) .
the pole locations , needed to construct the gobfs , are estimated via the best linear approximation of the system .
the coefficients of the multivariate polynomial are determined with a linear regression .
this paper provides a convergence analysis for the proposed identification scheme .
it is shown that the estimated output converges in probability to the exact output .
fast convergence rates , in the order @xmath0 , can be achieved , with @xmath1 the number of excited frequencies and @xmath2 the number of repetitions of the gobfs . , dynamic systems ; nonlinear systems ; orthonormal basis functions ; system identification ; wiener systems . |
sequential estimation continues observations until the observed sample satisfies a certain prescribed criterion .
its properties have been shown to be superior on the average to those of nonsequential estimation in which the number of observations is fixed a priori .
specifically the developments of higher - order asymptotic theory have suggested that the information loss due to the exponential curvature of the statistical model might be recovered by a sequential estimation procedure which makes use of the ancillary statistic .
such an estimator is expected to have a uniformly better characteristic on the average ( see e.g. srensen ( 1986 ) ) .
takeuchi and akahira ( 1988 ) formulated this scheme rigorously and analyzed the higher - order efficiency of sequential estimation procedures in the scalar parameter case ( see also akahira and takeuchi ( 1989 ) ) .
they showed that in the sequential case the exponential curvature term in the second - order variance can be eliminated by a second - order efficient estimator , and the maximum likelihood estimator with an appropriate stopping rule gives such a sequential estimator .
this also implies that appropriately designed sequential estimators are superior to nonsequential estimators in the asymptotic sense .
following the work of takeuchi and akahira ( 1988 , 1989 ) , okamoto , amari and takeuchi ( 1991 ) generalized the results to the multiparameter case by using the geometrical method , and studied characteristics of more general sequential estimation procedures . in the nonsequential case ,
a statistical manifold is uniformly enlarged by @xmath0 times when we use @xmath0 observations , keeping the intrinsic features of the manifold unchanged . in a sequential estimation procedure with a certain stopping rule
, the observed sample size @xmath0 is a random variable depending on the position of a statistical manifold .
this causes a nonuniform expansion of the statistical manifold .
such an expansion is called the conformal transformation in geometry , since it changes the scale locally and isotropically but it does not change the shape of a figure ( it does not change the orthogonality ) .
the result of takeuchi and akahira can be interpreted such that it is possible to reduce the exponential curvature of a statistical manifold to zero by a suitable conformal transformation .
the conformal geometry thus is an adequate framework for the analysis of the sequential inferential procedures if we extend the concept of the conformal transformation to the statistical manifold which is the riemannian manifold with a dual couple of affine connections . as a sequel to okamoto , amari and takeuchi ( 1991 )
, this paper investigates the sequential estimating procedures from the information geometrical viewpoint .
the novelty of this paper is the introduction of the dual conformal weyl - schouten curvature of a statistical manifold , and this quantity will be proved to play a central role when considering the problem of covariance minimization under the sequential estimating procedures .
information geometry was originated with the work of amari ( 1985 ) , and it has been establishing a solid status as a mathematical methodology for a variety of statistical sciences ( see e.g. amari et al ( 1987 ) , amari and nagaoka ( 2000 ) , kumon ( 2009 , 2010 ) ) . in line with these developments , the present paper also intends to provide a starting point for studying the conformal geometry of a statistical manifold itself succeeding to the work of lauritzen ( 1987 ) .
the organization of the paper is as follows .
the established known results are cited as propositions , and the results obtained in this paper are stated as theorems . in the next section , we prepare some statistical notations and preliminary results which will be relevant in this paper . in section 3 , we formulate a conformal transformation of a statistical manifold , where a set of dual weyl - schouten curvature tensors is introduced .
then we elucidate their implications in the structures of statistical manifolds . in this connection ,
the meaning of conjugate symmetry is also explained , which is the notion first introduced by lauritzen ( 1987 ) . in section 4
, the general result in the previous section is used to write out the structure of multidimentional exponential family . in section 5 ,
the general result is used to delineate the structure of multidimensional curved exponential family , where the dual euler - schouten curvatures are introduced .
then related notions such as totally exponential umbilic and dual quadric hypersurface are shown to involve key elements in studying statistical submanifolds .
in section 6 , the geometrical results obtained in the previous sections are applied to the sequential estimation in a multidimensional curved exponential family , where we give a concrete procedure for the covariance minimization . in section 7 , the results in section 6 are numerically examined by two typical curved exponential families called the von mises - fisher model and the hyperboloid model .
section 8 is devoted to some additional discussions and a perspective of future work .
let us denote by @xmath1 a @xmath2-dimentional random process defined on the probability space @xmath3 $ ] with values in @xmath4 $ ] , where @xmath5 and @xmath6 is the @xmath7-field of all borel sets in @xmath8 .
the time parameter @xmath9 runs over all non - negative integers @xmath10 or over all non - negative real numbers @xmath11
. moreover , let @xmath12 denote the @xmath7-field generated by the random vectors @xmath13 .
we assume that the probability measure @xmath14 depends on an unknown parameter @xmath15 , @xmath16 , where @xmath17 is homeomorphic to @xmath18 , and we shall consider the case where the following conditions are fullfiled : ( i ) @xmath19 is continuous in probability , has stationary independent increments and @xmath20 .
( ii ) the probability distributions at any time @xmath21 are dominated by a @xmath7-finite measure @xmath22 and have the densities @xmath23 with respect to @xmath22 @xmath24 we say that @xmath25 is an @xmath26-dimensional full regular minimally represented exponential family ( f.r.m . exponential family ) when the densities ( [ eq:1 ] ) can be written as @xmath27 where @xmath28 , @xmath29 is the natural parameter , and @xmath30 is the expectation parameter with @xmath31 a smooth ( infinitely differentiable ) convex function of @xmath29 . in the right - hand side of ( [ eq:2 ] ) and hereafter the einstein summation convention will be assumed , so that summation will be automatically taken over indices repeated twice in the sense e.g. @xmath32 .
sequential statistical procedures are characterized by a random sample size , where stopping times are used to stop the observations of the process .
we denote by @xmath33 an arbitrary stopping time , i.e. , a random variable @xmath33 defined on @xmath34 with values in @xmath35 and possessing the property @xmath36 .
we consider the case to which the sudakov lemma applies , where the stopped process @xmath37 has the densities @xmath38 , and these can be regarded as the same functions in ( [ eq:1 ] ) ( cf . e.g. magiera ( 1974 ) ) .
a typical statistical problem is the unbiased estimation of a given @xmath26-dimensional real vector valued function @xmath39 of the parameter @xmath29 by using observations on the @xmath40 .
an estimating procedure for @xmath41 is defined by a pair @xmath42 , where @xmath33 is a stopping variable and @xmath43 is a @xmath18-valued function defined on @xmath44 , which is an unbiased estimator of @xmath41 , i.e. , @xmath45 = h(\theta)$ ] .
let us look at all estimation procedures satisfying the following regularity conditions : ( i ) the @xmath41 gives a smooth one - to - one transformation from @xmath17 to @xmath46 in the sense @xmath47 ( ii ) @xmath48 < \infty,\ and the relation @xmath45 = h(\theta)$ ] can be differentiated with respect to @xmath29 under the expectation sign .
then we have the so - called cramr - rao inequality for the covariance matrix of the unbiased estimators .
if @xmath49 is an unbiased estimator of @xmath41 , then the the covariance matrix of @xmath49 is bounded below as @xmath50 \ge { \bar g}(h(\theta))^{-1},\end{aligned}\ ] ] where @xmath51,\ \ { \bar g}^{\alpha \beta}(h(\theta ) ) = e_\theta[\partial^\alpha { \bar l}_\tau \partial^\beta { \bar l}_\tau ] = c^{\alpha i}c^{\beta j}{\bar g}_{ij}(\theta),\\ & c^{\alpha i}(h ) = \frac{\partial \theta^i}{\partial h_\alpha},\ \ { \bar g}_{ij}(\theta ) = e_\theta[\partial_i { \bar l}_\tau \partial_j { \bar l}_\tau],\ { \bar l}_\tau = \log { \bar f}(x(\tau ) , \tau , h(\theta)),\quad \partial^\alpha = \frac{\partial}{\partial h_\alpha},\ \partial_i = \frac{\partial}{\partial \theta^i } , \end{aligned}\ ] ] and for symmetric matrices @xmath52 and @xmath53 , the inequality @xmath54 implies that @xmath55 is positive semi - definite . the equality in ( [ eq:3 ] ) holds if and only if @xmath49 can be represented almost everywhere as @xmath56 the condition ( [ eq:4 ] ) is also written as @xmath57 or in component form @xmath58 the above is a partial differential equation for @xmath59 , of which integrability condition @xmath60 is @xmath61 ( z_\gamma(\tau , x(\tau ) ) - h_\gamma(\theta ) ) = 0,\end{aligned}\ ] ] and hence the requirement for integrability is @xmath62 in this case , the log likelihood function @xmath63 is expressed as @xmath64 which implies that @xmath65 must be a f.r.m . exponential family with @xmath66 and @xmath67 the natual and the expectation parameters , respectively .
suppose that the original @xmath68 is not a f.r.m .
exponential family , then clearly @xmath65 is not a f.r.m .
exponential one , either .
hence we can restrict attention to the case ( [ eq:2 ] ) , when considering the attainment of the lower bound given by proposition 1.1 .
however we should remark that it is only a necessary condition for the attainment of the lower bound .
in fact even in the f.r.m . exponential family , only some restricted cases can exactly attain the lower bound due to the problem of the `` overshooting '' at the efficient stopping times ( see ghosh ( 1987 ) ) .
let @xmath69 be an original family of probability densities of unit time , where @xmath17 is homeomorphic to @xmath18 .
the family @xmath70 can be regarded as a statistical manifold , where the @xmath26-dimensiomal vector parameter @xmath29 serves as a coordinate system to specify a point , that is , a density @xmath71 .
the geometry of @xmath70 is determined by the following two tensor quantities ( cf .
amari ( 1985 ) , amari and nagaoka ( 2000 ) ) @xmath72,\quad t_{ijk}(\theta ) = e_\theta[\partial_il_1\partial_jl_1\partial_kl_1],\ \
l_1 = \log f(x , 1 , \theta),\ \
\partial_i = \frac{\partial}{\partial \theta^i}.\end{aligned}\ ] ] the first is the fisher information metric and the second is called the skewness tensor .
one parameter family of affine connections named the @xmath73-connection is defined by @xmath74 + \frac{1-\alpha}{2}t_{ijk}(\theta),\end{aligned}\ ] ] and then the @xmath73-riemann - christoffel curvature tensor is given by @xmath75 the @xmath73- and the @xmath76-connections are mutually dual @xmath77 and the @xmath78-rc curvature tensors are in the dual relation @xmath79 let @xmath80 be an @xmath26-dimensional extended statistical manifold under a sequential statistical procedure . from the wald identity the metric and the skewness tensors of @xmath81 are given by ( see akahira and takeuchi ( 1989 ) ) @xmath82,\\ & \nu(\theta ) = e_\theta[\tau],\quad s_k(\theta ) = \partial_k \log \nu(\theta),\quad 3g_{(ij}s_{k ) } = g_{ij}s_{k } + g_{jk}s_{i } + g_{ki}s_{j}.\end{aligned}\ ] ] these relations show that a sequential statistical procedure induces a conformal transformation @xmath83 by the gauge function @xmath84 .
the conformal transformation of a riemannian manifold implies that the manifold is expanded or contracted isotropically but that an expansion rate depends on each point .
our transformation is a statistical counterpart of this one .
a conformal transformation changes the @xmath73-connection into @xmath85,\\ & { \bar \gamma}_{ij}^{(\alpha)k } = { \bar \gamma}_{ijl}^{(\alpha)}{\bar g}^{lk } = \gamma_{ij}^{(\alpha)k } + \frac{1-\alpha}{2}(\delta_i^ks_j + \delta_j^ks_i ) - \frac{1+\alpha}{2}g_{ij}s_lg^{lk},\end{aligned}\ ] ] this is obtained by substituting ( [ eq:6 ] ) into @xmath86 and by noting that @xmath87 is the conformal change of the riemannian connection @xmath88 .
then a conformal transformation changes the @xmath73-rc curvature tensor into @xmath89,\\ & { \bar r}_{ijk}^{(\alpha)l } = { \bar r}_{ijkr}^{(\alpha)}{\bar g}^{rl } = r_{ijk}^{(\alpha)l } - \delta_i^ls_{jk}^{(\alpha ) } + \delta_j^ls_{ik}^{(\alpha ) } - g_{jk}s_{ir}^{(-\alpha)}g^{rl } + g_{ik}s_{jr}^{(-\alpha)}g^{rl},\\ & s_{ij}^{(\alpha ) } = \frac{1-\alpha}{2 } [ \nabla_i^{(\alpha)}s_j - \frac{1-\alpha}{2}s_is_j + \frac{1+\alpha}{4}g_{ij}s_ks_lg^{kl}],\quad \nabla_i^{(\alpha)}s_j = \partial_is_j - \gamma_{ij}^{(\alpha)k}s_k.\end{aligned}\ ] ] this is obtained by substituting ( [ eq:8 ] ) into @xmath90 we note that under a conformal transformation the mutual duality of @xmath91-connections is preserved @xmath92 and also the dual relation of the @xmath78-rc curvature tensors is preserved @xmath93 these are confirmed by the direct calculations with ( [ eq:6 ] ) , ( [ eq:8 ] ) and ( [ eq:10 ] ) .
one of the concerns about the conformal transformation is whether a given manifold can be transformed into a desirable space in some sense .
the main objective from the geometrical viewpoint is the flatness or the straightness , and it has been investigated usually in terms of the riemannian connection . from the statistical viewpoint ,
the main objective is the flatness or the straightness in terms of the mutually dual @xmath94-connections .
thus we say that a statistical manifold @xmath70 is _ conformally mixture ( exponential ) flat _ when there exists a gauge function @xmath84 such that @xmath95 holds .
note that by ( [ eq:14 ] ) @xmath70 is conformally mixture flat if and only if @xmath70 is conformally exponential flat . in view of these observations and also
the work of okamoto ( 1988 ) , we introduce the set of @xmath96-weyl - schouten curvature tensors as follows .
@xmath97 from ( 11 ) and ( 12 ) we have @xmath98 for the case @xmath99 we can also directly check @xmath100 , and hence @xmath101 . then we obtain the following result as to the conditions for the conformal mixture ( exponential ) flatness . a statistical manifold @xmath70 is conformally mixture flat ( or equivalently exponential flat ) if and only if _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ( i ) @xmath102 when @xmath103 .
+ ( ii ) @xmath104 and @xmath105 when @xmath106 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for the sake of simplicity
we hereafter express the notion such as conformally mixture ( or equivalently exponential ) flat as conformally @xmath107-flat .
consider the relation @xmath108 when @xmath109 we note by the integrability condition that @xmath110 we first prove the necessity .
suppose that @xmath70 is conformally @xmath26(@xmath111)-flat . then from @xmath112 , when @xmath113 we have @xmath114 .
when @xmath99 since @xmath100 we have @xmath115 , and since there exists a log gauge function @xmath116 we have @xmath117 .
we next prove the sufficiency when @xmath113 . from the bianchi s second identity ( cf .
schouten ( 1954 ) , p.147 ) @xmath118 and from @xmath102 we have @xmath119 so that @xmath104 .
then as noted before , there exists a covariant vector field @xmath120 such that @xmath121 . from the bianchi
s first identity ( cf .
schouten ( 1954 ) , p.144 ) @xmath122 and from the duality of @xmath123 @xmath124 we have @xmath125 on the other hand , from @xmath102 we have @xmath126 by combining these two relations we obtain @xmath127 and hence there exists a log gauge function @xmath116 such that @xmath128 . finally we prove the sufficiency when @xmath99 . from @xmath129 ,
as noted above , there exists a covariant vector field @xmath120 such that @xmath121 .
then from @xmath130 there exists a log gauge function @xmath116 such that @xmath128 .
this completes the proof of the theorem .
we further investigate the implications of theorem 3.1 .
suppose that the @xmath96-rc curvature tensor of a statistical manifold @xmath70 is expressed as @xmath131 where @xmath132 is constant on @xmath70 .
in this case we have @xmath133 a statistical manifold @xmath70 satisfying ( [ eq:19 ] ) is said to be _ a space of constant mixture ( exponential ) curvature _ , and @xmath70 satisfying ( [ eq:20 ] ) is said to be _
conjugate mixture ( exponential ) symmetric_. the conjugate @xmath26(@xmath111)-symmetry is the special notion of the conjugate @xmath91-symmetry introduced by lauritzen ( 1987 ) . by definition we know @xmath134 the connections among these notions are summarized in the following theorem . for the conformal @xmath26(@xmath111)-flatness of a statistical manifold @xmath70
, the following relations hold . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ( i ) @xmath70 is conjugate @xmath26(@xmath111)-symmetric and is conformally @xmath26(@xmath111)-flat if and only if @xmath70 is a space of constant @xmath26(@xmath111)-curvature . + ( ii ) a f.r.m .
exponential family @xmath135 is always conformally @xmath26(@xmath111)-flat .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ we first prove the sufficiency of ( i ) .
suppose that @xmath70 is a space of constant @xmath26(@xmath111)-curvature .
from @xmath136 we have @xmath137 then we obtain @xmath138 we next prove the necessity of ( i ) .
suppose that @xmath70 is conjugate @xmath26(@xmath111)-symmetric and is conformally @xmath26(@xmath111)-flat .
when @xmath113 , we have @xmath139 by substituting this expression into the bianchi s second identity , we have @xmath140 so that @xmath141 is constant on @xmath70 . when @xmath99 we have @xmath142 and again @xmath143 is constant on @xmath70 .
this completes the proof of ( i ) .
since @xmath135 is a space of zero @xmath26(@xmath111)-curvature @xmath144 , ( ii ) is obtained from ( i ) .
figure 1 illustrates the relations among several notions in theorem 3.2 . ,
based on theorem 3.2 ( ii ) , we seek a concrete conformal transformation @xmath145 such that @xmath146 is @xmath94-flat . when @xmath135 is a f.r.m
. exponential family , it is @xmath94-flat , i.e. , @xmath144 , in which the natural parameter @xmath29 and the expectation parameter @xmath147 provide the @xmath94-affine coordinate systems of @xmath135 in the sense ( cf .
amari ( 1985 ) , amari and nagaoka ( 2000 ) ) @xmath148 = 0,\quad \partial_i = \frac{\partial}{\partial \theta^i},\\ & \gamma^{(-1)ijk}(\eta ) = e_\eta[(\partial^i\partial^jl_1 + \partial^il_1\partial^jl_1)\partial^kl_1 ] = 0,\quad \partial^i = \frac{\partial}{\partial \eta_i},\end{aligned}\ ] ] and there exist two potential functions @xmath31 and @xmath149 such that @xmath150 by the formula ( [ eq:8 ] ) a conformal transformation @xmath145 with gauge function @xmath151 changes @xmath152 into @xmath153,\quad g^{ki } = e_\eta[\partial^kl_1\partial^il_1],\ \ s^j = \partial^j\log \nu.\end{aligned}\ ] ] we consider a coordinate transformation from @xmath147 to @xmath67 which will provide a @xmath96-affine coordinate system of @xmath146 .
the @xmath96-connection @xmath154 transforms to @xmath155,\end{aligned}\ ] ] and hence @xmath156 where @xmath157 is the integrability condition of the first equation on the right - hand side .
we can solve the above two partial differential equations for @xmath158 and @xmath159 as shown in the following theorem .
when a statistical manifold @xmath135 is an @xmath26-dimensional f.r.m . exponential family with @xmath96-affine coordinate system @xmath147 , it is conformally @xmath26(@xmath111)-flat by the gauge function @xmath151 and the new @xmath96-affine coordinate system @xmath67 given by @xmath160 where @xmath161 are constants , and @xmath162 . the new @xmath163-affine coordinate system @xmath164 , two potential functions @xmath165 and @xmath166 are respectively given as @xmath167 we first prove ( [ eq:21 ] ) . by putting @xmath168
we have @xmath169 and by putting @xmath170 we have @xmath171 we next prove ( 22 ) , ( 23 ) . by the direct calculation we can confirm @xmath172 and
then the others are immediately obtained .
this completes the proof of the theorem .
we remark that @xmath173 and @xmath174 in ( [ eq:21 ] ) cover the general solution and these are the same as those given in winkler and franz ( 1979 ) , which were derived from the statistical considerations of the efficient sequential estimators attaining the cramr - rao bound .
we first introduce a curved exponential family .
a family of probability densities @xmath175 parameterized by an @xmath26-dimensional vector parameter @xmath176 is said to be an @xmath177-curved exponential family when it is smoothly imbedded in an @xmath178-dimensional f.r.m .
exponential family @xmath179 in the sense @xmath180 where @xmath181 is homeomorphic to @xmath182 and @xmath183 is a smooth function of @xmath184 having a full rank jacobian matrix .
we use indices @xmath185 and so on to denote quantities in terms of the coordinate system @xmath29 or @xmath147 of @xmath135 , and indices @xmath186 and so on to denote quantities in terms of the coordinate system @xmath184 of @xmath187 . for analyzing the geometrical properties of @xmath187 imbedded in @xmath135 , it is convenient to introduce a new coordinate system @xmath188 of @xmath135 in the following manner .
we attach to each point @xmath189 an @xmath190-dimensional smooth submanifold @xmath191 of @xmath135 which transverses @xmath187 at @xmath192 or equivalently at @xmath193 .
we assume that the family @xmath194 fills up at least a neighborhood of @xmath187 in @xmath135 , that is , @xmath52 is a foliation of the tubular neighborhood of @xmath187 in @xmath135 .
such an @xmath191 is called an ancillary submanifold rigging @xmath184 , and @xmath52 is called an ancillary family rigging @xmath187 .
we introduce a coordinate system @xmath195 to each @xmath191 such that the origin @xmath196 is at the intersection of @xmath191 and @xmath187 . then the combined system @xmath197 gives a new local coordinate system of @xmath135 .
we use indices @xmath198 and so on for quantities related to the coordinate system @xmath199 , and indices @xmath200 and so on for quantities related to the coordinate system @xmath201 .
the basic tensors of @xmath135 are written as @xmath202 and the @xmath73-connection is given by @xmath203 in the @xmath199-coordinate system .
when we evaluate a quantity @xmath204 on @xmath187 , i.e. , at @xmath196 , we often denote it by @xmath205 instead of by @xmath206 for brevity s sake .
the metric tensors of @xmath187 and @xmath191 are given by @xmath207 and then indices can be lowered or uppered by using these metric tensors or their inverses @xmath208 .
the @xmath209-connections of @xmath187 are given by @xmath210 we call @xmath211 an orthogonal ancillary family when @xmath212 , and we assume this property in the following .
the mixed parts @xmath213 play central roles in the evaluation of statistical inferences , which are defined as follows .
@xmath214 and we call @xmath215 the @xmath209-euler - schouten curvature tensors of @xmath187 .
the @xmath209-rc curvature tensors and the @xmath216-es curvature tensors of @xmath187 are connected by the equations of gauss ( cf .
schouten ( 1954 ) , p.266 ) @xmath217 suppose that the @xmath209-es curvature tensors of @xmath187 are related as @xmath218 where @xmath219 is a constant .
in this case we have @xmath220 so that @xmath187 is conjugate @xmath26(@xmath111)-symmetric .
thus we say that @xmath187 satisfying ( [ eq:26 ] ) is _ es conjugate @xmath26(@xmath111)-symmetric_. suppose further that the @xmath163-es curvature tensor of @xmath187 is written as @xmath221 where @xmath222 is called the mean @xmath163-es curvature of @xmath187 , and @xmath187 satisfying ( [ eq:27 ] ) is said to be _ totally exponential umbilic ( e - umbilic)_. the implications of these notions are summarized in the following manner . for an @xmath177-curved exponential family @xmath187 , the following relation holds .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ let @xmath113 or @xmath223 , and suppose that @xmath187 is es conjugate @xmath26(@xmath111)-symmetric and totally @xmath111-umbilic .
then @xmath187 is a space of constant @xmath26(@xmath111)-curvature . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ suppose that @xmath187 is es conjugate @xmath26(@xmath111)-symmetric and totally @xmath111-umbilic .
then from ( [ eq:26 ] ) and ( [ eq:27 ] ) we have @xmath224 when @xmath113 , as noted in the proof of theorem 3.2 , @xmath225 is constant on @xmath187 . when @xmath226 , from the equation of codazzi ( cf .
schouten ( 1954 ) , p.266 ) @xmath227 and without loss of generality we can set @xmath228 , so that @xmath229 is again constant on @xmath187 .
we further deal with the case of @xmath226 .
suppose that @xmath187 satisfies the equations @xmath230 where @xmath231 are non - zero constants and @xmath232 are constant vectors . in this case
@xmath187 is expressed as @xmath233 and we call @xmath187 satisfying ( [ eq:28 ] ) _ a dual quadric hypersurface_. in section 7 it will be shown that the von mises - fisher model and the hyperboloid model are the examples of the dual quadric hypersurface .
the meaning of this hypersurface is described in the following theorem . for an @xmath234-curved exponential family @xmath187 , the following two conditions are equivalent .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ( i ) @xmath187 is a dual quadric hypersurface .
+ ( ii ) @xmath187 is es congugate @xmath26(@xmath111)-symmetric and totally @xmath111-umbilic with constant + @xmath26(@xmath111)-curvature @xmath235 , and @xmath236 on @xmath187 . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ we first prove @xmath237 . by the definition ( [ eq:28 ] ) we have @xmath238 on the other hand @xmath239 and hence @xmath240 we next prove @xmath241 . by the definitions ( [ eq:26 ] ) and ( [ eq:27 ] ) we have @xmath242 on the ther hand @xmath243 and hence @xmath244 this completes the proof of the theorem .
figure 2 illustrates the relations among several notions in theorems 5.1 and 5.2 .
, width=377,height=264 ] from theorems 3.2 and 5.2 , the dual quadric hypersurface is conformally @xmath107-flat , and we obtain the following result as to its dual structure .
when @xmath187 is an @xmath26-dimensional dual quadric hypersurface , it is conformally @xmath107-flat by the log gauge function @xmath245 and the @xmath96-affine coordinate system @xmath246 , satisfying @xmath247 where @xmath248 are constants and @xmath249 . the @xmath163-affine coordinate system @xmath250 , two potential functions @xmath251 and @xmath252 of @xmath253 are respectively given as @xmath254 we first prove ( [ eq:29 ] ) . as noted in the proof of theorem 3.1 , the partial differential equation for @xmath245 is @xmath255 when @xmath187 is an @xmath26-dimensional dual quadric hypersurface , from @xmath256 we have the first relation of ( [ eq:29 ] ) .
the partial differential equation for @xmath257 is given as @xmath258 = 0,\quad { \bar \gamma}_{abc}^{(-1 ) } = \nu[\gamma_{abc}^{(-1 ) } + g_{ac}s_b + g_{bc}s_a]\\ & \leftrightarrow\ \partial_bc_a^{\bar a } - { \bar \gamma}_{ba}^{(-1)c}c_c^{\bar a } = 0,\quad c_a^{\bar a } = \frac{\partial { \bar u}^{\bar a}}{\partial u^a},\quad { \bar \gamma}_{ba}^{(-1)c } = \gamma_{ba}^{(-1)c } + \delta_a^cs_b + \delta_b^cs_a.\end{aligned}\ ] ] when @xmath187 is a dual quadric hypersurface , we can directly show that the above is satisfied by the second relation of ( [ eq:29 ] ) .
we next prove ( [ eq:30 ] ) , ( 31 ) . by the direct calculation
we can confirm @xmath259 and then the others are immediately obtained .
this completes the proof of the theorem .
we next consider a conformal transformation @xmath145 by the gauge function @xmath260 .
as shown by ( [ eq:8 ] ) , the @xmath73-connection in terms of the @xmath199-coordinate system is changed into @xmath261.\end{aligned}\ ] ] then we can express the change of quantities related to @xmath253 , that is , the @xmath96-connection of @xmath187 , the @xmath163-es curvature of @xmath187 and the @xmath96-es curvature of @xmath191 are respectively changed into @xmath262,\quad { \bar h}_{ab\kappa}^{(1 ) } = \nu[h_{ab\kappa}^{(1 ) } - g_{ab}s_\kappa],\quad { \bar h}_{\kappa \lambda a}^{(-1 ) } = \nu
h_{\kappa \lambda a}^{(-1)}.\end{aligned}\ ] ] it is also seen that @xmath263 and we call @xmath264 _ the conformal @xmath163-es curvature tensor_. note that the change of the @xmath96-connection @xmath265 induces the projective transformation at the same time , which implies that the mixture geodesic is preserved under the transformation ( cf .
schouten ( 1954 ) , p.287 ) .
the effect of constant @xmath26(@xmath111)-curvature is given in the following theorem .
suppose that a curved exponential family @xmath187 is a space of constant @xmath26(@xmath111)-curvature .
then there exists a conformal transformation @xmath145 and a coordinate system @xmath266 , of @xmath253 such that the followings hold . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ( i ) @xmath267 + ( ii ) if @xmath187 is totally @xmath111-umbilic , then @xmath268 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ we first prove ( i ) . when @xmath187 is a space of constant @xmath26(@xmath111)-curvature , from theorem 3.2 , @xmath187 is conformally @xmath26(@xmath111)-flat , so that we have @xmath269 we next prove ( ii ) .
let us take @xmath270 ( see okamoto , amari and takeuchi ( 1991 ) ) .
then for the totally @xmath111-umbilic @xmath187 , from ( [ eq:27 ] ) , ( [ eq:32 ] ) and ( [ eq:33 ] ) we have @xmath271
we consider sequential estimations in an @xmath177-curved exponential family @xmath187 .
let @xmath272 be a parameter cotrolling the average sample size , and let @xmath273 be a smooth gauge function defined on @xmath135 in the @xmath147-(@xmath199-)coordinate system .
we denote by @xmath274 the sample mean up to time @xmath21 .
it has the same value in the @xmath147-coordinate system , and its value in the @xmath199-coordinate system is denoted by @xmath275 .
the random stopping time @xmath33 is assumed to satisfy ( see okamoto , amari and takeuchi ( 1991 ) ) @xmath276 = o(1),\quad e_u[\tau ] = k\nu(u),\ v_u[\tau ] = o(k).\end{aligned}\ ] ] the term @xmath277 is due to the bias of @xmath278 from the true @xmath279 , which is obtained by the requirement @xmath280 = k\nu(u)$ ] .
the term @xmath281 includes a rounding error and the `` overshooting '' at the stopping time @xmath33 .
we cite the established results concerning the asymptotics of sequential estimators of @xmath184 from okamoto , amari and takeuchi ( 1991 ) . for a consistent sequential estimator @xmath282 of @xmath184 ,
the following relations hold .
( i ) the estimator @xmath282 is first - order efficient , that is , @xmath283 as @xmath284 , if and only if @xmath285 is an orthogonal ancillary family .
+ ( ii ) the bias - corrected estimator @xmath286 of @xmath282 is given by @xmath287 ( iii ) the asymptotic covariance of @xmath286 is given by @xmath288 = g^{ab } + \frac{1}{k\nu}\bigg\{\frac{1}{2 } ( ^{'}\!\gamma_{m_c}^{(-1)})^{2ab } + ( ^{'}\!h_{m_c}^{(1)})^{2ab } + \frac{1}{2}(h_a^{(-1)})^{2ab } \bigg\ } + o(k^{-2}),\end{aligned}\ ] ] where @xmath289 based on theorem 5.4 we obtain the following result for the possibility of covariance minimization .
suppose that a curved exponential family @xmath187 is a space of constant @xmath26(@xmath111)-curvature and is totally @xmath111-umblic .
then there exists a conformal transformation @xmath145 and a coordinate system @xmath266 , of @xmath253 such that the following holds for the maximum likelihood estimator @xmath290 of @xmath291 without bias - correction : @xmath292 = g^{{\bar a}{\bar b } } + o(k^{-2}).\end{aligned}\ ] ] when @xmath187 itself is a f.r.m .
exponential family , ( [ eq:36 ] ) holds by ( [ eq:21 ] ) given in theorem 4.1 .
when @xmath187 is a dual quadric hypersurface , ( [ eq:36 ] ) holds by ( [ eq:29 ] ) given in theorem 5.3 . since @xmath293 holds for the maximum likelihood estimator ( m.l.e . ) , from ( [ eq:34 ] ) and theorem 5.4 , we have for the bias of the m.l.e.@xmath294 and the expression ( [ eq:36 ] ) is derived . when @xmath187 itself is a f.r.m .
exponential family with expectation parameter @xmath184 , the partial differential equations for @xmath245 and for @xmath257 are @xmath295 as noted in theorem 4.1 .
when @xmath187 is a dual quadric hypersurface , the partial differential equations for @xmath245 and for @xmath257 are @xmath296 as noted in theorem 5.3 .
this completes the proof of the theorem .
this is an @xmath234-curved exponential family , of which density functions with respect to the invariant measure on the @xmath26-dimensional unit sphere under rotational transformations are given by ( cf .
barndorff - nielsen et al ( 1989 ) , p.76 ) @xmath297 where @xmath298 is the modified bessel function of the first kind and of order @xmath299 .
we assume that the concentration parameter @xmath300 is assumed to be a given positive constant .
the parametric representations @xmath301 and @xmath302 are given by + [ cols= " < , < " , ] the above expressions show that @xmath303 and again this model is a dual quadric hypersurface . from theorem 5.2
we also see that this model is es conjugate @xmath26(@xmath111)-symmetric and totally @xmath111-umblic with constant @xmath26(@xmath111)-curvature @xmath304 .
the related geometrical quantities are given below .
@xmath305 from theorem 3.2 ( i ) this @xmath187 is conformally @xmath26(@xmath111)-flat , so that there exist a gauge function @xmath306 and a @xmath96-affine coordinate system @xmath307 such that @xmath308 . as given by ( [ eq:29 ] ) ,
the partial differential equation for @xmath245 is @xmath309 of which one solution is @xmath310 and then @xmath311 .
we examine our theoretical results numerically by using the von mises - fisher and the hyperboloid models .
we take 10 kinds of number @xmath0 ( nonsequential case ) and @xmath312 ( sequential case ) of observations , and for each @xmath0 or @xmath312 , we generate @xmath313 random simulated data .
then the empirical means of covariances @xmath314 $ ] ( nonsequential case ) and @xmath315 $ ] ( sequential case ) of the m.l.e . over this 500 sample size are used for evaluation , where @xmath316 and @xmath317 denote the true values of @xmath318 and @xmath319 .
the stopping times @xmath33 for the sequential estimations are determined by ( see okamoto , amari and takeuchi ( 1991 ) ) @xmath320 as for the von mises - fisher model , numerical results are based on the following set of values @xmath321 and for the hyperboloid model , numerical results are based on the following set of values @xmath322 figures 3 - 8 show the von mises - fisher model , and figures 9 - 14 show the hyperboloid model .
the notations in the figures indicate the following quantities .
@xmath323,\quad a , b = 1 , 2\\ & \quad ocrbab = g^{ab}(u_0),\\ & \quad oalbab = g^{ab}(u_0 ) + \frac{1}{n}\bigg\{\frac{1}{2 } ( \gamma_{m_c}^{(-1)}(u_0))^{2ab } + ( h_{m_c}^{(1)}(u_0))^{2ab } \bigg\}\\ & \textrm{$\cdot$ sequential case}\\ & \quad ccov{\bar a}{\bar b } = e_{emp}(\tau)e_{emp}[({\hat { \bar u}}^{\bar a}_{mle } - { \bar u}_0^{\bar a } ) ( { \hat { \bar u}}^{\bar b}_{mle } - { \bar u}_0^{\bar b})],\quad { \bar a } , { \bar b } = 1 , 2\\ & \quad ccrb{\bar a}{\bar b } = g^{{\bar a}{\bar b}}({\bar u}_0),\\ & \quad mst = e_{emp}(\tau)\ : \ \textrm{empirical mean of $ \tau$}\\ & \quad sdst = \sqrt{v_{emp}(\tau)}\ : \ \textrm{empirical standard deviation of $ \tau$}.\end{aligned}\ ] ] we see that in the nonsequential case @xmath324 approach to the asymptotic lower bound @xmath325 exhibiting the differential geometrical loss @xmath326 , and in the sequential case @xmath327 nearly attain the cramr - rao lower bound @xmath328 as if the model were a f.r.m . exponential family .
figures 8 , 14 confirm that the assumptions @xmath329 are satisfied in each model .
von mises - fisher , width=302,height=264 ] von mises - fisher , width=302,height=264 ] von mises - fisher , width=302,height=264 ] von mises - fisher , width=302,height=264 ] von mises - fisher , width=302,height=264 ] von mises - fisher , width=302,height=264 ] hyperboloid , width=302,height=264 ] hyperboloid , width=302,height=264 ] hyperboloid , width=302,height=264 ] hyperboloid , width=302,height=264 ] hyperboloid , width=302,height=264 ] hyperboloid , width=302,height=264 ]
we have analyzed sequential estimation procedures in terms of the conformal geometry of statistical manifolds .
we have also constructed a concrete procedure for the covariance mininization in a multidimensional curved exponential family @xmath187 .
the method is divided into two separate stages : one is to choose a stopping rule which is effective for reducing the @xmath163-es curvature @xmath330 and the other is to choose a gauge function @xmath331 on @xmath187 effective for reducing the @xmath96-connection @xmath332 .
another typical choice of @xmath331 is the one effective for the covariance stabilization , as suggested in okamoto , amari and takeuchi ( 1991 ) .
these choices contradict each other in general multidimensional cases , and this fact reflects the difference between the ordinary riemannian geometry and the mutually dual geometry as exhibited in several geometrical notions introduced in this paper .
the present method is also applicable to investigating sequential testing procedures .
the geometrical theory of higher - order asymptotics of testing hypothesis in nonsequential case was developed by kumon and amari ( 1983 ) and amari ( 1985 ) .
the main results are summarized as follows .
the power function @xmath333 of a test @xmath334 is expanded as @xmath335 where @xmath0 denotes the number of observations , and @xmath336 indicates the geodesic distance between the null hypothesis and the point in the alternative hypothesis . +
( i ) the first - order power function @xmath337 and the second - order power function @xmath338 are maximized uniformly in @xmath21 if and only if the ancillary family ( boundaries of the critical region ) associated with a test @xmath334 is asymptotically an orthogonal family .
+ ( ii ) the third - order power loss function @xmath339 is expressed as the weighted sum of two kinds of the square of the @xmath163-es curvatures @xmath330 , the square of the @xmath96-es mixture curvature @xmath340 of the associated ancillary family , and also the square of the @xmath96-mixture connection @xmath332 ( when there are unknown nuisance parameters ) .
based on these nonsequential results , we can utilize the conformal geometry to the analysis and the construction of most powerful sequential tests . specifically when a statistical manifold is a f.r.m . exponential family or a dual quadric hypersurface
, it is expected that one can design sequential tests without any power loss .
this is a subject which will be treated in a future work .
barndorff - nielsen , o. e. , blsild , p. and eriksen , p. s. ( 1989 ) .
_ decomposition and invariance of measures , and statistical transformation models .
lecture notes in statist . _ * 58 * , springer - verlag , new york .
lauritzen , s. l. ( 1987 ) .
statistical manifolds . in _
differential geometry in statistical inference .
_ ( s. amari , o. e. barndorff - nielsen , r. e. kass , s. l. lauritzen and c. r. rao eds . )
163 - 216 , ims , hayward , calif .
takeuchi , k. and akahira , m.(1988 ) .
second order asymptotic efficiency in terms of asymptotic variances of the sequential maximum likelihood estimation procedures . in _
statistical theory and data analysis ii : proceedings of the second pacific area statistical conference . _ ( k. matusita eds . )
191 - 196 .
north - holland , amsterdam . | we present a geometrical method for analyzing sequential estimating procedures .
it is based on the design principle of the second - order efficient sequential estimation provided in okamoto , amari and takeuchi ( 1991 ) . by introducing a dual conformal curvature quantity , we clarify the conditions for the covariance minimization of sequential estimators .
these conditions are further elabolated for the multidimensional curved exponential family .
the theoretical results are then numerically examined by using typical statistical models , von mises - fisher and hyperboloid models .
_ keywords and phrases : _
affine connections , curved exponential family , hyperboloid distribution , information geometry , projective transformation , riemannian metric , space of constant curvature , totally umbilic , von mises - fisher distribution . |
it often happened in high energy physics that previously discovered particles later became the `` standard candles '' , and led to the discoveries of new particles .
for example , the @xmath0 and @xmath2 gauge bosons were discovered by identifying the leptons they decay to ; they ( especially the @xmath2 boson ) in turn played important roles in the discovery of the higgs boson at the large hadron collider ( lhc ) @xcite .
it is our hope that at the lhc and future colliders , we will be able to find new physics beyond the standard model ( sm ) , again by first identifying particles that we are already familiar with .
therefore , it is crucial that we are able to efficiently identify all known particles , or , equivalently , their masses and quantum numbers .
this is a difficult task when the particle is a qcd parton , namely a quark or a gluon , or when it decays to qcd partons . in these cases ,
what we see in a collider detector is one or more energetic jets that contain multiple hadrons , and their identities are concealed .
there are several handles one can utilize to distinguish jets from different origins : the jet mass or multi - jet invariant mass is determined by the original parton s mass ( for convenience , in this article we use the term `` parton '' to refer to either a qcd parton or a massive particle that decays to qcd partons ) ; a massive hadronically decaying particle , such as a @xmath0 or @xmath2 boson , results in more than one jets with similar momenta , which is different from a usual qcd jet originated from a single quark or gluon ; finally , the qcd quantum number determines the amount of jet radiation and how it is distributed .
the first two handles have been used routinely in high energy experiments , while the last , though receiving significant amount of studies recently , has not been used as commonly .
part of the reason is , compared with the hard kinematics , the radiation information of a jet is often associated with soft particles , making it more vulnerable to various contaminations , including the underlying event , pileup , as well as radiation from nearby jets .
we will focus on tackling this difficulty in this article .
variables sensitive to the jet radiation are calculated from jet constituents . on the one hand
, we would like to choose a larger jet radius such that more radiation from the initial parton is included to give us more color information ; on the other hand , we would like to avoid introducing excessive contamination , and a smaller radius is preferred .
therefore , a careful choice of the jet radius is important .
similar considerations have been made in refs .
@xcite . in refs .
@xcite , it is shown that the optimum jet radius is around @xmath3 for jet @xmath4 at a hadron collider , where the effects of the underlying event and hadronization are minimized , and it yields the optimum @xmath5 mass measurement for the ua2 experiment .
the study has been extended to the lhc in refs .
@xcite where the criterion for optimizing the jet radius is to obtain the best mass resolution for new massive hadronically decaying resonances .
the purpose of those studies is to properly reconstruct the kinematics , _
i.e. _ , the momenta of the partons initiating the observed jets .
our goal is different : we want to choose the jet radius such that the discrimination power , including that from the jet radiation patterns , is maximized . for this purpose , it is important to study the distribution of the `` intrinsic '' radiation , _ i.e. _ , the radiation originated from the initial parton instead of the contaminations .
in particular , it is helpful to know what is the amount of the intrinsic radiation contained in a cone of a particular size around the jet axis .
nonetheless , there is not a unique definition of the `` amount '' of radiation , and the radiation distribution depends on various factors such as the color configuration , the boost to the system , _
etc._. in this article , we give a rigorous definition of the jet _ radiation radius _ , taking the simplest case as the start point : a dijet color singlet system .
roughly speaking , the jet radiation radius , @xmath6 , is the jet radius one needs to include on average a fraction of @xmath7 of the total radiation in the jet .
a variable that quantifies the amount of radiation is needed in this definition , such as thrust , girth @xcite , or @xmath8-subjettiness @xcite .
we study the dependence of @xmath6 on the jet qcd quantum number and the jet momentum .
we examine two applications , quark - gluon discrimination and @xmath0 jet tagging , and show that knowledge of the jet radiation radius can help us achieve the optimum discrimination power .
interestingly , in both cases , the jet radiation radius , and thus the optimum ( sub)jet radius decreases with increasing jet momentum , albeit for different reasons . for a quark or gluon jet produced from a color singlet system in its ( nearly ) rest frame ,
the jet momentum is roughly proportional to the center of mass energy , which determines the amount of radiation and how it is distributed .
it is well known that the qcd coupling constant decreases when the energy scale increases , resulting in smaller radiation radius .
we will see that the radiation radius approximately follows a power law dependence on the jet momentum , with the power being around @xmath9 .
therefore , to reduce the contaminations , we are able to use a smaller cone size to evaluate jet radiation variables for higher jet momentum . for the case of a boosted @xmath0 boson ( or other massive color singlet particles ) , the radiation scale is fixed to the @xmath0 mass , while the boost makes the radiated particles collimated .
this results in a radiation radius inversely proportional to the jet momentum , or subjet momentum if the @xmath0 is identified as a single jet . in light of that
, we propose a new @xmath0 tagging algorithm to deal with high pileup . in this algorithm , we start as usual with a fat jet that contains a @xmath0 boson or its qcd counterpart , and use a jet grooming method @xcite to reconstruct the kinematics of the @xmath0 decay .
then , when calculating a jet radiation variable , we use two much smaller cones around the two leading subjets axes , with the cone sizes inversely proportional to the subjets @xmath1 s . with this `` shrinking cone '' algorithm
, we find a 30% ( 60% ) improvement in the statistical significance for jet @xmath10 ( @xmath11 ) , when the average number of pileup events is 60 , compared with methods not using the shrinking cones .
the rest of the article is organized as follows . in section [ sec : definition ] , we give the definition of the jet radiation radius . in section [ sec : quark - gluon ] , we discuss the difference in the radiation radius between a quark jet and a gluon jet in a dijet color singlet system , and demonstrate that the optimum jet radius for quark - gluon discrimination decreases for increasing jet @xmath1 . section [ sec : wtag ] is devoted to the discussion of the radiation radius in a boosted system , and in particular its role in @xmath0 jet tagging .
section [ sec : discussions ] contains some discussions .
a jet has a finite radius precisely because of the qcd radiation and hadronization .
therefore , it seems redundant to have a definition of a _
radiation radius_. however , the usual purpose of using a finite radius for jet clustering is to ensure most of the jet radiation is included in the jet such that one can infer the initial partons momenta correctly . for this purpose ,
the radiation gives us difficulties rather than provides us useful information .
moreover , due to collinear singularity , the momentum of a jet is concentrated in the core , that is , the center of the jet .
therefore , it is not essential to control precisely the jet radius , and multiple choices of jet radius coexist in high energy experiments . for the purpose of determining the quantum number of a jet ,
the information from the qcd radiation is essential and one should be more careful about choosing the optimum jet radius . as mentioned in the introduction , this is particularly important when the jet is contaminated by other hadronic activities in the event , including the underlying event , pileup and radiation from nearby jets .
therefore , it is useful to define a jet radiation radius , which is a measure of how the radiation is distributed . knowing the `` intrinsic '' , _
i.e. , _ uncontaminated , jet radiation radii corresponding to partons with different quantum numbers will allow us to have a generic understanding of how to choose a jet radius to better deal with the contaminations .
as we will show later , it also allows us to engineer new algorithms that provide better discrimination powers .
the first jet definition was given by sterman and weinberg @xcite . in ref .
@xcite , a hadronic event from an @xmath12 collision is classified as a two jet event if at least a fraction of @xmath13 of the event s total energy is contained in two cones of opening half - angle @xmath14 , where @xmath15 is a small fraction .
obviously , for fixed @xmath15 , when we increase @xmath14 , the fraction of hadronic events that are classified as two jets , @xmath16 , will also increase . at the next leading order
, @xmath16 is given by @xcite @xmath17+\frac{\pi^2}{12}-\frac{7}{16}-\epsilon+\frac{3}{2}\epsilon^2+o(\delta^2\ln\epsilon)\right\}. \label{eq : f2}\ ] ] when @xmath15 is small , we keep only the leading term and find @xmath18 .
\label{eq : deltaq}\ ] ] in eq .
( [ eq : deltaq ] ) , we have added a subscript `` @xmath19 '' because hadronic events in an @xmath12 machine are dominated by quark - antiquark pairs . in such a machine ,
the event is not contaminated by the underlying event and initial state qcd radiation , and two - jet events are dominantly initiated from a pair of back - to - back high energy quarks . therefore , @xmath20 can be viewed as the `` intrinsic '' size of a quark jet . from eq .
( [ eq : deltaq ] ) , we see the definition of the size @xmath20 depends on two parameters , @xmath15 and @xmath16 .
similarly , if the system under study is initiated from two hard gluons ( in the color singlet configuration ) , we obtain the angular size , @xmath21 , of a gluon jet . at the leading order , @xmath22 @xcite .
therefore , @xmath20 or @xmath21 is a measure of how the radiation is distributed , which is sensitive to the partons color structure .
this motivates us to give a general definition of a jet _ radiation _ radius , by replacing the energy fraction @xmath13 with a measure of the amount of radiation .
the definition of the amount of radiation is not unique , and suitable choices include thrust ( denoted @xmath23 ) , charged particle multiplicity ( denoted @xmath24 ) , @xmath8-subjettiness , _
etc_. while a possible definition of the jet radiation radius could be given following the definition of the variable @xmath14 , in practice it is perhaps more convenient to use the following definition .
_ in a 2(or @xmath8)-jet color singlet system in its center of mass frame , a jet radiation radius is defined as the jet radius that the average amount of radiation within the 2 ( or n ) jets is a fraction of @xmath7 of the average total radiation . _ in practice , the jet radiation radius can be calculated as follows . for a given ensemble of events , to get the fraction @xmath7
, we first calculate the average `` total '' radiation : for an @xmath12 machine , it may include all particles in the event ; while for a hadron collider , we may use a set of jets clustered with a large radius as the start point .
then we use a ( smaller ) radius @xmath25 to recluster the jets and find the amount of radiation within the leading 2 ( or n ) jets , also averaged over the ensemble .
the fraction @xmath7 is the radio of the two averages , and the corresponding @xmath25 is our jet radiation radius .
we denote the radiation radius by @xmath26 , where @xmath27 , is a measure of the amount of radiation .
a few comments of the above definition is in place .
first , there are multiple choices of the variables that can quantify the amount of radiation . to be a good candidate
, the variable should satisfy : @xmath28 it is ( near ) zero when there is no radiation .
@xmath29 it increases monotonously when the jet radius is increased , and approaches a finite value when the jet radius is taken to be infinity , _
i.e. _ , when all particles in the event ( or the part of the event being studied ) are included .
we easily see that among the most commonly used variables , jet mass is a good radiation variable , while jet momentum is not .
other radiation variables include @xmath30 where @xmath23 is the thrust , charged particle multiplicity , @xmath8-subjettiness , _
etc_. second , the requirement of a color singlet system is to guarantee the jet radius is not affected by particles outside the system , otherwise the definition is ambiguous .
moreover , we have also required the system to be in the center of mass frame such that it is reasonable to use a common jet radius for all jets in the system .
even with this requirement , when the system contains more than 2 jets with different momenta , we still need a prescription on how to choose the jet radii this is not a concern in this article because we will only discuss dijet systems .
as we will see , when a dijet system is boosted , the two jets will have different momenta and it is desirable to use different radii for them . in this article , we will rely on pythia 8 @xcite simulations to calculate the jet radiation radius , leaving analytical calculations to future studies .
we start from the simplest case : dijet color - singlet systems .
there are two possibilities for the initial partons , @xmath31 or @xmath32 . given the large number of variables that have been proposed in the literature
, we will only consider a few of them that are particular useful for our later discussions : charged particle multiplicity , girth and @xmath8-subjettiness . charged particle multiplicity is simply the number of charged particles passing a certain @xmath1 threshold , or the number of tracks in the jet .
girth is defined as @xcite @xmath33 where @xmath34 is the distance in the @xmath35 plane between particle ( or jet constituent ) @xmath36 and the jet axis . here
, @xmath37 is the rapidity .
note the variable contains a normalization factor , @xmath38 , which is the jet @xmath1 . as described in the previous section
, we will need to calculate the `` total radiation '' starting from a fat jet ( or , for the @xmath12 case , starting from all particles in a hemisphere ) , as well as the radiation for a variety of smaller jet radius .
then it is a question whether we should use the @xmath1 of the fat jet ( or half of the center of mass energy for an @xmath12 machine ) as a common normalization factor for all jet radii , or use the jet @xmath1 of the corresponding radius .
both choices are valid , although there are subtle differences between them .
we find the latter is more convenient and more commonly used , which we will stick to in the following discussions .
@xmath8-subjettiness is derived from @xmath8-jettiness @xcite , and defined as follows @xcite .
we define a distance measure between a particle @xmath39 and an axis @xmath40 as @xmath41 , where @xmath42 , and @xmath43 is a constant . then for @xmath8-axes
we define @xmath44 namely , for each particle in the jet , we find the nearest axis and calculate the @xmath1 weighted distance .
then @xmath45 is a sum of the distances over all particles in the jet , normalized by a factor @xmath46 where @xmath47 is the jet radius . finally , we find the @xmath8-axes such that @xmath45 is minimized , and the minimum value is called @xmath8-subjettiness and denoted @xmath48 . when @xmath49 and @xmath50 , we see that 1-subjettiness , @xmath51 , is defined very similar to girth , with slight differences in the distance measure ( girth uses the rapidity @xmath37 while @xmath51 uses the pseudorapidity @xmath52 ) and the normalization factor .
therefore , the discussion on girth applies to @xmath51 as well . in general , these variables are examples of a set of variables called _ radial moments _ @xcite that are all valid radiation variables : @xmath53 where @xmath54 is a function of @xmath55 , such as @xmath55 , @xmath56 , @xmath57 .
although it will be infrared / collinear unsafe , we may also generalize the definition to allow a non - linear dependence on @xmath58 , for example , by making the sum over @xmath59 .
then the particle multiplicity can be viewed as a special case where @xmath60 .
as mentioned in the previous section , it is convenient to study events from @xmath12 collisions , where we only have final state qcd radiation emitted from the partons being studied .
therefore , we study the processes @xmath61 and @xmath62 .
we fix the pseudorapidity of the two outgoing partons to @xmath63 , _
i.e. _ , perpendicular to the beam , and use pythia 8 for showering and hadronization . for an @xmath12 machine ,
it is better to use the spherical coordinates .
nevertheless , we eventually would like to extend our results to a hadron collider , so we will use the @xmath64 coordinates for convenience .
we then apply the anti-@xmath65 algorithm with various @xmath25 s for jet clustering , using fastjet @xcite .
it is known that the anti-@xmath65 algorithm gives us circular jets , which mimic cones of size @xmath25 around the jet axes . since all contaminations are proportional to the area of the jet , we draw the radiation variables as a function of @xmath66 in fig . [
fig : vars_ee ] for better illustration .
it is clear from fig .
[ fig : vars_ee ] that the radiation variables grow much faster at smaller @xmath66 s than larger @xmath66 s , and eventually saturate at very large @xmath66 s .
this indicates that increasing @xmath25 beyond a certain value will only introduce more contaminations rather than provide more information , therefore , an optimum radius should exist . [
cols="^,^ " , ] .[fig : roc_150],scaledwidth=70.0% ]
in this article , we have given a definition of the jet radiation radius , which quantifies the size of a jet due to its qcd radiation .
this definition is closely related to the jet shape ( also known as jet profile ) variable which measures , on average , the fraction of momentum that is included in a cone of size @xmath25 around the jet axis . for the purpose of studying the jet radiation distribution ,
momentum is not a good measure because a large ( small ) momentum does not correspond to large ( small ) amount of radiation .
therefore , in our definition , we have replaced it with variables that directly measure the amount of radiation .
moreover , we have emphasized the nature of the radiation radius being _ intrinsic _ , _ i.e. _ , it is a characteristic of a parton with a definite qcd quantum , and should not change across different production processes and experimental setups . in particular , it is defined before various contaminations are included . it is our hope that by `` factorizing '' the contributions to jet shape variables into intrinsic and environmental ones , we can simplify jet substructure studies and use jet radiation patterns more efficiently to distinguish jets with different quantum numbers .
a key observation in this article is : in order to efficiently use jet radiation variables , a smaller than usual jet radius is often preferred .
this is particularly true for the @xmath0 jet tagging method we proposed , where shrinking cones are used when calculating jet radiation variables . in a high energy experiment , either a simpler or a more complicated method may be adopted . on the one hand , due to the limitation in granularities , especially those of the hadronic calorimeter , one may not be able to use a jet radius smaller than @xmath67 , or / and may not be able to use continuous jet radii in the calculation . in that case , we may choose to simplify the method by choosing a few typical , but small , cone sizes .
it is shown in fig .
[ fig : roc_300 ] and fig .
[ fig : roc_150 ] that increasing the jet radius within a sizable region will not significantly hurt the discrimination power . on the other hand , one does see the advantages of using small cones , for example , @xmath68 is preferred for subjet @xmath69 when the average number of pileup events is 60 . an even smaller radius might be preferred if the pileup level is higher . therefore , ideally we would want to use the finest granularity for jet constituents , including the information from the electromagnetic calorimeter and the tracking system as in the particle flow approach @xcite .
moreover , we have sticked to a single choice of ( sub)jet radius for each ( sub)jet in this article .
similar to ref .
@xcite , one may benefit from using two or more radii for each ( sub)jet , which not only gives us the information of how much the radiation is , but also captures how it grows with increasing @xmath25 .
the shrinking cone algorithm we proposed for @xmath0 tagging is parallel and complementary to other pileup reduction methods and may be combined to obtain the optimum results .
we have already used it with the pileup subtraction method proposed in ref .
@xcite , where we see the subtraction method is convenient to extract the kinematic information while the shrinking cone method is more useful to obtain the radiation information .
another set of useful techniques utilize the fact that a charged particle from a pileup event leaves a track not originated from the primary vertex , thus it can be subtracted from the jet .
these methods include charged hadron subtraction @xcite , using a jet vertex fraction @xcite cut and jet cleansing @xcite .
one may even use the charged particles from the primary vertex alone when calculation radiation variables , which still provides us a lot of information for the color structure @xcite . to improve over these method , one may simply apply them for cones with sizes determined by the ( sub)jet @xmath1 s , as discussed in this article . here , we emphasize that even if we only use tracks from the primary vertex to avoid most of the contamination from pileup events , it is still useful to optimize the jet cone sizes . we have seen that it is the case for quark - gluon discrimination when pileup is turned off .
we expect this consideration to be more important when dealing with events with many hard partons , where jets can be easily contaminated by nearby radiation and a large jet radius should be avoided .
this happens in , for example , susy cascades with long decay chains .
the author thanks dave soper for numerous useful discussions and comments on the manuscript .
the author is in part supported by us department of energy under grant numbers de - fg02 - 96er40969 and de - fg02 - 13er41986 .
99 s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 716 * , 30 ( 2012 ) [ arxiv:1207.7235 [ hep - ex ] ] .
g. aad _ et al .
_ [ atlas collaboration ] , phys .
b * 716 * , 1 ( 2012 ) [ arxiv:1207.7214 [ hep - ex ] ] .
j. alitti _ et al . _
[ ua2 collaboration ] , z. phys .
c * 49 * , 17 ( 1991 ) .
j. huth _ et al .
_ in e. l. berger , `` research directions for the decade .
proceedings , 1990 summer study on high - energy physics , snowmass , usa , june 25 - july 13 , 1990 , '' singapore , singapore : world scientific ( 1992 ) .
m. cacciari , j. rojo , g. p. salam and g. soyez , jhep * 0812 * , 032 ( 2008 ) [ arxiv:0810.1304 [ hep - ph ] ] . g. soyez , jhep * 1007 * , 075 ( 2010 ) [ arxiv:1006.3634 [ hep - ph ] ] .
j. gallicchio and m. d. schwartz , jhep * 1110 * , 103 ( 2011 ) [ arxiv:1104.1175 [ hep - ph ] ] ; j. gallicchio and m. d. schwartz , phys .
* 107 * , 172001 ( 2011 ) [ arxiv:1106.3076 [ hep - ph ] ] ; j. gallicchio and m. d. schwartz , jhep * 1304 * , 090 ( 2013 ) [ arxiv:1211.7038 [ hep - ph ] ] .
j. thaler and k. van tilburg , arxiv:1011.2268 [ hep - ph ] ; j. thaler and k. van tilburg , arxiv:1108.2701 [ hep - ph ] .
j. m. butterworth , a. r. davison , m. rubin and g. p. salam , phys .
lett . * 100 * , 242001 ( 2008 ) [ arxiv:0802.2470 [ hep - ph ] ] .
s. d. ellis , c. k. vermilion and j. r. walsh , phys .
d * 80 * , 051501 ( 2009 ) [ arxiv:0903.5081 [ hep - ph ] ] ; s. d. ellis , c. k. vermilion and j. r. walsh , phys .
d * 81 * , 094023 ( 2010 ) [ arxiv:0912.0033 [ hep - ph ] ] .
d. krohn , j. thaler and l. t. wang , jhep * 1002 * , 084 ( 2010 ) [ arxiv:0912.1342 [ hep - ph ] ] .
g. f. sterman and s. weinberg , phys .
lett . * 39 * , 1436 ( 1977 ) .
r. k. ellis , w. j. stirling and b. r. webber , camb .
monogr . part .
cosmol .
* 8 * , 1 ( 1996 ) .
t. sjostrand , s. mrenna and p. z. skands , comput .
commun . * 178 * , 852 ( 2008 ) [ arxiv:0710.3820 [ hep - ph ] ] .
y. cui , z. han and m. d. schwartz , phys .
d * 83 * , 074023 ( 2011 ) [ arxiv:1012.2077 [ hep - ph ] ] .
z. han , phys .
d * 86 * , 014026 ( 2012 ) [ arxiv:1112.3378 [ hep - ph ] ] .
g. soyez , g. p. salam , j. kim , s. dutta and m. cacciari , phys . rev
. lett . *
110 * , 162001 ( 2013 ) [ arxiv:1211.2811 [ hep - ph ] ] .
[ cms collaboration ] , cms - pas - pft-09 - 001 .
for example , a. perloff [ cms collaboration ] , j. phys .
404 * , 012045 ( 2012 ) .
for example , the atlas collaboration , atlas - conf-2013 - 083 .
d. krohn , m. low , m. d. schwartz and l. -t .
wang , arxiv:1309.4777 [ hep - ph ] . | jet radiation patterns are indispensable for the purpose of discriminating partons with different quantum numbers
. however , they are also vulnerable to various contaminations from the underlying event , pileup , and radiation of adjacent jets . in order to maximize the discrimination power , it is essential to optimize the jet radius used when analyzing the radiation patterns .
we introduce the concept of jet radiation radius which quantifies how the jet radiation is distributed around the jet axes .
we study the color and momentum dependence of the jet radiation radius , and discuss two applications : quark - gluon discrimination and @xmath0 jet tagging . in both cases ,
smaller ( sub)jet radii are preferred for jets with higher @xmath1 s , albeit due to different mechanisms : the running of the qcd coupling constant and the boost to a color singlet system .
a shrinking cone @xmath0 jet tagging algorithm is proposed to achieve better discrimination than previous methods . |
an asteroid family is an group of asteroidal objects in the proper orbital element space ( _ a _ , _ e _ , and _ i _ ) , considered to have been produced by a disruption of a large parent body through a catastrophic collision ( first identified by hirayama 1918 ; see cellino et al .
2009 and references therein ) .
family members have usually similar surface properties such as spectral taxonomy types @xcite , sdss colors @xcite , and visible geometric albedo @xcite .
therefore an asteroid family can be seen as a natural solar system experiment and is regarded as a powerful tool to investigate space weathering @xcite and non - gravitational phenomena such as the yarkovsky and yorp ( yarkovsky - okeefe - radzievskii - paddack ) effects .
the maria asteroid family has long been known as one of the hirayama families @xcite , and is a typical old population ( @xmath03 @xmath1 1 gyr ) @xcite that is expected to have experienced significant collisional and dynamical evolution in the history of the inner solar system .
the maria family is located close to the outer border of the 3:1 mean - motion resonance ( mmr ) ( @xmath02.5 au ) with jupiter , thus it might be regarded as a promising source region candidate for a couple of giant s - type near - earth asteroids , 433 eros and 1036 ganymed @xcite .
our knowledge about the properties of the maria family , however , are still limited . to date ,
rotational periods of the family members among 3,230 cataloged objects @xcite have been known only for 58 of the relatively large asteroids ( lcdb , march 2013 ) , accounting for less than 2 percent of the family .
the study of the rotational properties of an asteroid family , i.e. rotational period , pole orientation , and overall shape of lightcurve , can offer a unique opportunity to obtain insight both on the collisional breakup process and on the dynamical evolution of asteroids .
@xcite proposed that the statistical properties of the spin period and pole orientation contain information on the collisional transfer of angular momentum of small bodies .
laboratory experiments ( * ? ? ?
* and reference therein ) from catastrophic disruption events show that the distribution of asteroidal spin periods resembles a maxwellian , which implies that members have reached an approximate equilibrium status after a chain of isotropic collisions .
there must be differences between the behavior of laboratory experiments and the catastrophic collisions of asteroids in their natural environment due to various approaches on scaling law ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
@xcite have detected that , on average , the smaller asteroids rotate faster from the catastrophic disruption experiments , and it has already been predicted by the semi - empirical model of catastrophic impact processes by @xcite .
in addition , this tendency is also found in spin rate - size distribution from the asteroid lightcurve database ( lcdb ) @xcite as well as on a statistical analysis of c and s - type mbas @xcite .
a typical asteroid family shows a v - shape distribution in the proper semi - major axis vs. absolute magnitude plane @xcite .
the dynamical evolution of the family members due to the yarkovsky effect tends to spread the members out in semi - major axis , over the evolution time . for this reason
, the dispersions of the semi - major axes can be used as a clue to estimate the age of an asteroid family @xcite .
the lightcurve inversion method @xcite is a powerful tool to obtain rotational properties , especially pole orientation , from time - series photometric data .
@xcite derived 80 new asteroid models based on the combination of classical dense - in - time ( hereafter dense ) and sparse - in - time ( hereafter sparse ) photometric lightcurve data .
they found the distribution of pole latitudes for small asteroids ( d @xmath3 km ) in the main - belt to be depleted near the ecliptic plane , an effect attributed to the yorp thermal effect .
there are several observational studies on the rotational properties of asteroid families . @xcite and @xcite discovered an alignment of the spin vectors in the koronis family , which is commonly referred to as the slivan effect .
@xcite showed that the distribution of rotational periods in the flora asteroid family is non - maxwellian .
taking advantage of the analysis of lightcurves obtained from flora family members , @xcite recently found the slivan effect in the flora family .
@xcite tried to find correlations among rotation periods , lightcurve amplitudes , and sizes of objects for the themis , eos , and maria families , but no specific relationship was found .
@xcite focused on the observation of young asteroid families , which are expected to have experienced little collisional and orbital evolution . in general , we have some difficulty in obtaining high s / n lightcurves , especially for smallest and hence , faintest family members . the purpose of this paper is to investigate the rotational properties of the maria family ; rotation periods , orientations of the spin axis , and shapes , mainly based on our observations but also with data available in the literature . in sect .
2 we describe the photometric observations of the maria family asteroids with an introduction on the several telescopes we used . in sect .
3 we explain the method of data reduction and acquisition from the available dense and sparse datasets from the astdys ( asteroids dynamic site ) database .
lightcurve plots for each asteroid are given in appendix figs .
a. 1 @xmath4 61 . using the lightcurve inversion method
, we derive the orientations of spin axis for 13 maria asteroids .
the results of this analysis and related discussions about the rotational properties of the maria family are given in sect .
4 . finally , in sect . 5
, we summarize our conclusions .
observations of the maria family asteroids were conducted during 134 nights from 2008 july to 2013 may , using 0.5 m- to 2 m- class telescopes at 7 observatories in the northern hemisphere .
we used ccd cameras on the wise observatory ( wo ) 0.46 m telescope in negev desert , israel , the tubitak ulusal gozlemevi ( tug ) 1.0 m telescope in bakirlitepe , turkey , the bohyunsan optical astronomy observatory ( boao ) 1.8 m telescope on bohyunsan , korea , the sobaeksan optical astronomy observatory ( soao ) 0.6 m telescope on sobaeksan , korea , the korea astronomy and space science institute ( kasi ) 0.6 m telescope in jincheon , korea , the lemmonsan optical astronomy observatory ( loao ) 1.0 m telescope on mt .
lemmon , usa , and the calar alto ( ca ) astronomical observatory 1.2 m telescope in almeria , spain .
the details of the observatories and instruments are shown in table 1 .
all telescopes except for the boao 1.8 m and the ca 1.2 m telescopes were guided at sidereal tracking rate .
for sidereal tracking of asteroids , we considered two factors : ( 1 ) the apparent motion of the asteroid should be less than the fwhm of the stellar profiles on each observatory ; ( 2 ) the signal - to - noise ratio of the asteroid measurement to be at least 50 .
accordingly , the maximum exposure time did not exceed 300 seconds while tracking at sidereal rate during the observations .
because several asteroids were relatively faint , some images taken with the boao 1.8 m telescope were acquired through a non - sidereal tracking mode that corresponded to the predicted motion of the objects . on the other hand ,
the ca 1.2 m telescope was tracked with a tracking vector halfway between sidereal rate and that of asteroids . in this way
both asteroids and the background stars were trailed by the small amount .
we made time - series observations mainly with johnson r filter in order to characterize the rotational states of the asteroid , since the r - band filter with an optical imager is the most sensitive to small bodies in the solar system .
the maria asteroid family consists of 3,230 known members with identification based on hierarchical clustering method ( hcm ; * ? ? ?
the latest family classification is available at the astdys web siteastdys2/propsynth / numb.members ] . as of november 2013
, the database lists 2,085 maria members based on synthetic proper orbital elements of numbered asteroids .
we checked the membership as listed in the astdys , and found that most of our target asteroids were confirmed as maria members except for asteroids 652 , 4122 , 5977 , 13679 , 19184 , 19333 , 32116 , 43174 , 50511 , 109792 , and 114819 .
however , memberships of an asteroid family can be subject to some unavoidable uncertainties , since they depend on the datasets of proper orbital elements available at the epoch of family search , on the adopted identification technique and membership assignment criteria .
we investigated the cumulative distribution n ( @xmath5 ) of absolute magnitudes h for all the members of the maria family .
we used a power - law approximation of n ( @xmath5 ) @xmath6 10@xmath7 in the magnitude range between 12 and 14.5 , and obtained @xmath8 0.54 ( see fig .
the value of @xmath9 obtained is close to the steady - state mass distribution of collisional fragmentation for which @xcite derived @xmath9 = 0.5 , but discordant with a considerably steeper distribution constructed by pareto power laws ( e.g. * ? ? ?
this power law index implies that maria family has undergone a significant collisional and dynamical evolution and has currently reached an equilibrium state ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the slope depends highly on the number of objects brighter than 12th absolute magnitude , while the slope is thought to have been contaminated by observational bias for objects fainter than h = 14.5 magnitude .
we chose our targets in the h range between 12 and 14.5 ; the total number of targets in this magnitude range is 981 .
the catalogued absolute magnitudes in the standard ( h , g ) system have large uncertainties due to the transformation from an observed magnitude system to the johnson v system , and to the phase correction ( see * ? ? ? * for more details ) .
it might either overestimate or underestimate the size distribution of a family .
we generated ephemeris for each object from the 981 target asteroids using the jpl horizons service , then produced target lists of observable asteroids during any given night .
because the rotational periods of most targets are unknown , we decided to observe one or two target asteroids per night at least , assuming typical rotational periods between 2 and 24 hours .
since a preliminary rotational lightcurve of an asteroid is determined only after a prompt data reduction , it is better to allocate follow - up observation so as to cover its full - phased lightcurve . for the sake of increasing the observational efficiency , we developed an observation scheduler to carry out asteroid follow - up observation in a timely manner .
when we input observational parameters such as initial rotational period , observed time , and observatory code , the scheduling program suggests to the user the next proper observing time to cover gaps in the lightcurve .
our observations were focused on asteroids in the maria family that lack a known rotational period . to increase the number of observable asteroids as much as possible , we observed two objects alternately on a single night with the boao 1.8 m , tug 1.0 m , and ca 1.2 m telescopes .
moreover , further observations during the years in 2012 - 2013 were also performed for 21 targets at the wise telescope in 2008 , and spin axis information was obtained from published lightcurves ( see sect . 3.2 . for more details ) .
it is difficult to cover entire rotational phase of asteroids with rotational periods longer than 8 hours during a single night .
in addition , if the rotational period of an asteroid aliases with 24 hours @xmath4 the rotational period of the earth , it is essentially impossible to have a fully covered lightcurve from a single observatory .
a network of follow - up telescopes can be used to solve these problems .
in order to maximize phase coverage of an asteroid , taking advantage of a network observation at different time zones , we organized observation campaigns with three observatories in korea ( ut + 9 h ) , turkey ( ut + 2 h ) , and usa ( ut @xmath4 8 h ) .
these network observations were carried out in 2012 with tug and loao on june 28 , 30 ; boao and tug on oct .
12 , 14 , 19 , and soao and tug on oct .
12 , 14 , 17 . in order to calibrate all the data gathered from various observatories ,
the same ccd fields obtained during the previous observations were taken in the next runs .
the detailed observational circumstances of each asteroid are listed in table 2 : ut date corresponding to the mid time of the observation , the topocentric equatorial coordinates ( ra and dec , j2000 ) , the heliocentric ( r ) and the topocentric distances ( @xmath10 ) , the solar phase ( sun - asteroid - observer s ) angles ( @xmath11 ) , the apparent predicted magnitude ( v ) , and the telescopes used .
all observational data reduction procedures were performed using the image reduction and analysis facility ( iraf ) software package .
individual images were calibrated using the standard processing routine of the iraf task _
noao.imred.ccdred.ccdproc_. bias and dark frames with relatively high standard deviations were not used for analysis .
twilight sky flats were acquired before sunrise and after sunset , and combined to produce a master flat image for each night .
instrumental magnitudes of asteroids were obtained using the iraf _ apphot _ package ; aperture radii were set to be equal to the fwhm of the typical stellar profile on each frame in order to maximize the s / n ratio @xcite .
the lightcurves of most asteroids were constructed based on relative magnitudes , that is the difference between the instrumental magnitude of an asteroid and the average magnitude of each comparison star . in order to choose a set of comparison stars
, we inspected whether there exists any short - term variability in brightness of a star , according to the following method .
we overlapped each frame that was taken during the same night in order to calculate differences in magnitude of each star and the median value of an ensemble of the remaining stars .
in such a way , we were able to determine standard deviations of each star in the same frame , and hence checked if they showed variability during the course of a single night s observation .
likewise , we repeatedly applied the same procedure to the other ccd fields , and successfully selected stars with minimum standard deviation that show the least amount of light variation in each field .
finally we selected 3 to 5 comparison stars with a typical scatter of 0.01 to 0.02 mag .
the observation time ( ut ) was corrected for light travel time .
out of 134 nights observations in total , we obtained 218 individual lightcurves for 74 maria family members and derived synodic rotational periods for 51 objects including obtained periods for 34 asteroids for the first time . in order to find the periodicity of the lightcurve , the fast chi - squared ( f@xmath12 )
technique @xcite was employed .
in addition , we checked the result with the discrete fourier transform algorithm @xcite . in most cases for rotational periods ,
these two different techniques present similar results within the statistical errors .
the f@xmath12 technique presented here represents the observed magnitudes as a fourier time series truncated at the harmonic h : @xmath13 in practice , we fit the fourth - order fourier function and also obtained the highest spectral power with the discrete fourier transform algorithm . as a result ,
the final rotational period was determined assuming a double - peaked lightcurve .
we present the resultant composite lightcurves of observed asteroids in appendix figs .
a. 1 @xmath4 61 , folded with their synodic periods . in the approximation of a triaxial body with uniform albedo ,
the peak - to - peak variations in magnitude are caused by the change in apparent cross - section of the rotating body , with semi - axes a , b , and c , where @xmath14 ( the body rotates along the c axis ) . according to @xcite ,
the lightcurve amplitude varies as a function of the polar aspect viewing angle @xmath15 ( the angle between the rotation axis and the line of sight ) : @xmath16 + the lower limit of axis ratio a / b can be expressed as @xmath17 , which corresponds to an equatorial view ( @xmath18 ) .
the peak - to - peak variation in lightcurve becomes larger when we increase the solar phase angle .
the following empirical relationship between those two parameters was found by @xcite .
@xmath19 + where @xmath20 is the lghtcurve amplitude at zero phase angle and @xmath21 is the amplitude measured at solar phase angle @xmath22 .
these authors also found that average _ m_-values are 0.030 , 0.015 , and 0.013 degree@xmath23 for s , c , and m - type asteroids , respectively .
we adopted the constant _ m _ of 0.03 for our analysis , as maria family is known as a typical s - type asteroid family . in order to derive the orientation of the spin axis using disk - integrated photometric data , dense lightcurves obtained over three or four
apparitions are essential @xcite .
however , due to limited amount of dense datasets , the lightcurve inversion method using only sparse data @xcite and a combination of sparse and dense data @xcite has been improved during the past decades .
we obtained the available dense dataset by matching objects with the maria family members from the following three data sources . in the mpc ( minor planet center ) light curve database , we downloaded lightcurves for the 17 maria asteroids . from the online website of courbes de rotation dastrodes
et de comtes ( cdrbehrend / page_cou.html ] , operated by r. behrend ) , a total of 140 individual lightcurves for 13 object were acquired .
another 3 lightcurves were found in the asteroid photometric catalog ( apc , * ? ? ?
* ) from the nasa s planetary data system ( pds ) .
most sparse photometric datasets are a by - product of world - wide astrometric surveys , such as the catalina sky survey and the siding spring survey .
detailed instructions for obtaining the sparse data from the astdys site are in @xcite ; we follow their methodology . according to very recent analysis by the same authors ( priv .
communication ) , only sparse datasets from the usno in flagstaff ( mpc code 689 ) , the catalina sky survey ( 703 ) , and the la palma ( 950 ) were useful for determining the shape modeling of asteroids
. combining the dense and the sparse lightcurve datasets from the astdys , we computed the pole orientation for 16 objects , and determined 13 unique solutions .
our observations from 7 observatories yielded 218 individual lightcurves for 74 maria asteroids . among them
we derived synodic rotational periods for the 51 members with a reliability code @xmath24 2 .
for the 23 objects with a reliability code of 1 , we set a lower bound to the periods except for 879 ricarda for which the rotation period is known .
the reliability parameters follow the definition by @xcite : + 1 .
very tentative result , may be completely wrong . + 2 . reasonably secure result , based on over half coverage of the lightcurve .
secure result , full lightcurve coverage , no ambiguity of period .
multiple apparition coverage , pole position reported . the information on rotational characteristics from our lightcurve datasets with other physical properties is summarized in table 3 ; diameter and albedo information were mainly acquired from the wise ir data @xcite with the exception of those marked with a ( acua , asteroid catalog using akari ; * ? ? ?
* ) and m ( mean albedo value of 0.254 for the maria family asteroid ) , while the taxonomic information is compiled from @xcite and @xcite .
we used the datasets only with a reliability code @xmath24 2 for our analysis .
the lightcurves for 23 objects denoted with q notes of 1 did not cover even half - phased periods due to following three reasons : ( 1 ) very low amplitude of the lightcurve ( smaller than 3-sigma ) ; ( 2 ) variation of brightness much longer than observing time ; ( 3 ) data interruption due to bad weather or instrument conditions .
we examined a lower bound to the periods for those 23 objects to consider the influence of selection bias , i.e. , favoring observations of faster rotators .
although all 23 objects rotate slower than 5 hours , they do not affect the overall tendency of our results , and besides , the peak - to - peak variation derived from completely wrong lightcurves could negatively affect the statistics .
this is the reason why we only select periods only with a reliability code @xmath24 2 for our analysis . in order to improve the significance of the statistical analysis of the rotational properties in the maria family
, we searched the published lightcurve data from the asteroid lightcurve database ( lcdb ) released in march 2013 and matched 58 objects with the maria family asteroids . out of 58 lcdb data sets , there are 17 objects overlapping with our observations .
therefore , we checked those original lightcurves from literatures , and then adopted lightcurves with minimum period gaps between the lcdb and our observations .
figure 2 shows the distribution of rotational rates for 92 maria family members used in this study and the corresponding best fit maxwellian curve .
the 51 objects obtained from our observations are marked with shaded bars .
it can be obviously seen that asteroids rotating faster and slower considerably exceed the fitted distribution . from the kolmogorov - smirnov test ,
the compatibility between the observed distribution and the fitted maxwellian is completely inconsistent at a 92% confidence level .
this non - maxwellian distribution is also found in other old - type asteroid families , such as the koronis family @xcite and the flora family @xcite , with ages of 2.5 @xmath1 1.0 gyr and 1.0 @xmath1 0.5 gyr @xcite , respectively .
laboratory experiments of catastrophic collisions ( * ? ? ?
* and references therein ) and numerical simulations @xcite showed that the rotational frequency of fragments approximate the maxwellian distribution . accordingly , this inconsistency with our measurements suggests that the members belong to an old family and have had their rotational properties modified by non - gravitational forces operating over a long period of time ; the spin rate change , in particular , can be attributed to the yorp effect .
@xcite found that the spin states of pseudo - gaspra and pseudo - eros can significantly evolve on a 100 myr timescale , which is far shorter when compared to their break - up and collisional processes .
@xcite found that the spin vectors of the main - belt asteroids with d @xmath3 km were significantly affected by the yorp thermal effect .
the size range of most of the 92 maria members used in this study is distributed between 1.5 km and 30 km ( see fig .
we examine the correlation between the rotational frequency ( cycle / day ) and diameter ( km ) in fig .
4 . among the 92 maria family asteroids used for this study
, four objects might be regarded as interlopers . in terms of the visible geometric albedo @xmath25 , two asteroids 3094 chukokkala and 4860 gubbio possess very low albedos ; 0.068 and 0.037
, respectively , provided that the mean value of albedo for the maria family asteroids is 0.254 from the matched 1,152 wise ir data .
the other two members have a different spectral type of xe for 1098 hakone @xcite , c - type for 71145 ( 1999
xa183 ) @xcite .
those possible interlopers are marked with filled boxes , however , do not affect the overall tendency in fig .
4 . in the results obtained from our observations
, there is an apparent systematic trend toward larger dispersion of rotational frequency with decreasing size .
the rotation of small asteroids can be easily accelerated or decelerated due to the yorp effect .
moreover , all asteroids larger than 22 km rotate more slowly than 4.8 hr ( 5 cycles per day ) , quite similar to what was found for the flora family @xcite .
figure 5 shows the correlation between the rotational frequencies ( cycle / day ) and the amplitudes of lightcurve ( peak - to - peak variation magnitude ) that represents the overall shapes of asteroids .
we derived amplitudes of lightcurves using the equation ( 3 ) for our observations .
for lightcurve data from the lcdb , we adopted the maximum amplitudes of lightcurves .
the various sizes of circles indicate the diameter of each asteroid relatively ; asteroids larger than a diameter of 15 km are marked with blue circles .
we can see two features in fig .
5 : there is no object with both fast rotation ( faster than 6 cycles per day ) and large lightcurve amplitudes ( @xmath26 mag ) . despite insufficient number for large peak - to - peak variation , few highly elongated objects tend to rotate slowly .
this could be explained by the break - up limit of elongated rubble pile ; it makes sense that the more elongated an object is , the easier it can be shattered during its spin - up process .
the colored curves in fig . 5 approximate the critical rotational period ( @xmath27 ) for bulk densities of 1.5 , 2.0 , and 2.5 g/@xmath28 , respectively , adopted from @xcite : @xmath29 + where @xmath27 is the period in hours
, @xmath30 is the bulk density in g/@xmath28 , @xmath10_m _ is the amplitude of the lightcurve variation in magnitude .
it is apparent from fig .
5 that there is no object against `` spin barrier '' @xcite for bulk density of 2.5 g/@xmath28 .
in addition , elongated objects are located far from the spin rate limit . the other feature seen in fig .
5 is that no objects larger than 15 km have amplitude of lightcurve larger than 0.5 magnitude except for 4860 gubbio , regarded as an interloper .
small objects ( @xmath31 km ) with various shapes are spread out in this plot .
if an elongated rubble pile asteroid is disrupted by a non - gravitational force such as the yorp effect , it might have less elongated , namely more spherical shape as a result .
figure 5 separates into two plots ( fig .
a. 62 and fig .
a. 63 ) in appendix a. the yarkovsky effect plays a significant role in the dynamical evolution of asteroid orbits .
the study of this non - gravitational thermal force on the asteroid family has been improved dramatically during the past decades in several families , such as koronis @xcite , flora @xcite , and eos @xcite .
semi - major axis drift , in accordance with either the sense of rotation or various sizes of asteroids , by the yarkovsky effect results in a v - shape in the proper semi - major axis and absolute magnitude plane @xcite .
the analysis for drift in semi - major axis allows us to estimate the age of an asteroid family @xcite . to find the yarkovsky footprints in the maria asteroid family , we investigate the pole orientation of the members .
we examined sense of rotation for the 16 members which were observed during at least three apparitions , and successfully determined the pole orientations for the 13 objects by the combination with sparse data from the astdys site .
the summary of pole solutions for the maria family asteroids is listed in table 4 . due to an intrinsic symmetry of the problem @xcite , ground - based observations of objects with a small orbital inclination
are affected by an ambiguity in the determination of the ecliptic longitude of the pole axis direction , resulting in two solutions that are placed 180 degrees apart and are statistically indistinguishable .
we adopted the lower chi - squared solution marked with boldface for this study and estimated the uncertainty of the pole orientation to be 5 @xmath4 20 degrees , depending on the number of dense and sparse datasets .
out of 13 asteroids , the information on spin axis of 4 objects was also found in damit ( database of asteroid models from inversion techniques ; * ? ? ?
the result of our analysis for those 4 objects is quite consistent with that of damit .
concerning the pole orientation , we concentrate on the determination of the sense of rotation . as a result , we found six prograde rotators ( 652 jubilatrix , 787 moskva , 875 nymphe , 1158 luda , 1996 adams , and 3786 yamada ) ; while two objects ( 695 bella and 897 lysistrata ) rotate in the opposite sense .
we show the position of prograde and retrograde asteroids on a plane of proper semi - major axis ( au ) versus absolute magnitude ( h ) in fig .
in addition , we found five rotators with the pole along the ecliptic ; 170 maria , 575 renate , 660 crescentia , 727 nipponia , and 6458 nouda marked with open circles in fig .
the spin axes of those five objects are very close to the ecliptic plane within their statistical errors , between 5 and 20 degrees .
although an observational bias can not be excluded , the excess of prograde rotators with respect to retrograde rotators could be explained with a long - term evolution by the yarkovsky effect . in case of retrograde rotators ,
the yarkovsky drift causes the semi - major axis of family members to decrease ; consequently , they could have been ejected by the 3:1 mmr to the inner solar system .
@xcite found a similar result for the flora family , which is located near the outer border of the @xmath32 resonance area .
recent statistical studies by @xcite indicate that there is an excess of prograde versus retrograde rotators in the main - belt for asteroids smaller than 100 km .
however , these authors did not find any convincing explanations for this excess and hence this remains an open problem . on the contrary , a strong excess for retrograde rotation
was found in neas that is completely consistent with a theoretical ratio of 2 @xmath1 0.2 with respect to prograde rotators @xcite . regarding this prograde - retrograde asymmetry in mbas and neas
, they inferred that there is a connection between the excess of retrograde neas and the deficiency of retrograde mba due to the yarkovsky effect .
however , it might be more complicated to generalize the trend to mbas as they are too big to have been affected by the yarkovsky drift .
another supporting example we can find is the largest nea , 1036 ganymed that is regarded to be originated from the maria family .
@xcite carried out an extensive analysis of the maria family , and suggested that about 10 objects with the size between 15 and 30 km have probably been injected into the 3:1 mean - motion resonance with jupiter .
they proposed that a couple of giant s - type neas , 433 eros and 1036 ganymed have probably been originated from the maria family .
the spectral reflectance of this asteroid is quite similar to the average spectrum of the maria family members @xcite ; interestingly its spin orientation is retrograde @xcite , moreover , it is on a high - inclination orbit of 26.7 degrees .
we can easily distinguish three interlopers ( that is , two lower albedo objects of 3094 chukokkala , 4860 gubbio and xe - type asteroid of 1098 hakone ) from fig .
6 , while 71145 ( 1999 xa183 ) is not close to the border of the family . one distinct property of the maria family compared to other densely populated asteroid families such as flora , themis , and eunomia , is that there is no prominent large body among the family members . the largest body in this family is not 170 maria but 472 roma with a diameter 50.3 km as calculated from the wise ir data .
in the recent paper by @xcite , the roma family is substituted for the name of the maria family .
figure 7 shows the spin vector obliquity of the maria family members with respect to the rotational frequency ( cycle / day ) , compared with the koronis @xcite and flora families @xcite . in those two families , there is a conspicuous group of prograde objects .
furthermore , retrograde objects share almost the same obliquities , which is referred to as the slivan effect @xcite . in the maria family ,
on the other hand , no prominent prograde groups being in the slivan state were found , while the data for retrograde rotation is not sufficient for a conclusion .
@xcite investigated several test prograde rotators chosen with various ranges of either semi - major axis or inclination in the main - belt .
their preliminary results indicate that asteroids with low - inclination in the outer main - belt , such as 24 themis , might be trapped efficiently in the slivan state .
this is the reason why they suggest observations of either high - inclination or inner main - belt asteroid and numerical simulations for their dynamical evolution .
@xcite found that orbital inclination of main - belt asteroids is the most important factor that affects the magnitude of the spin vector variation .
based on their dynamical spin evolution , if the orbital inclination is increased , also the maximum obliquity variation increases linearly .
the other feature seen in fig . 7
is that there are several objects with spin vector obliquities around 90 degrees , that is , spin axes approximately parallel to the orbital plane .
the change of semi - major axis by the diurnal component of the yarkovsky effect is proportional to the cosine of the obliquity of an asteroid , so it will vanish when the obliquity approaches 90 degrees .
we can predict that the semi - major axis drift of these five objects plotted as open circles in fig .
6 and fig .
7 will not occur .
however , we can expect the spin vector can be modified by the yorp effect .
the 3:1 mean - motion resonance ( @xmath0 2.5 au ) is the most prominent jovian resonance region in the main - belt , and , as such , it is an important source region of neas and meteors .
the maria family is located quite close to the outer border of the 3:1 mmr with jupiter .
when members of the maria family enter this unstable region due to the yarkovsky drift , they can escape from its membership and become neas .
for this reason , the distribution of proper semi - major axis vs. absolute magnitude is asymmetric as shown in fig .
the @xcite model suggests that with 20 % probability , the population of all near - earth objects ( with absolute magnitude h @xmath33 ) originated from the 3:1 mmr region .
they estimated that the number of kilometer - sized neos ( h @xmath33 ) escaped from the 3:1 mmr every million year is 100 @xmath1 50 bodies .
similar results independently derived by @xcite show that 100 to 160 objects larger than 1 km enter the 3:1 mmr per million years . in order to make an order - of - magnitude estimation of the flux of the maria members into the 3:1 mmr region ,
we try to determine the number of missing maria family members .
we do this by taking the difference of the original members ( assuming the distribution was originally entirely symmetric ) and the number of currently known objects ( the absolute magnitude of all members is brighter than 18th mag ) . to obtain the number of original members , we assume that the center of the maria family is at about 2.562 au ( denoted with the red dashed line ) since the number density with respect to the semi - major axis is at a minimum
. then we double the number of family member , with proper semi - major axis @xmath34 au as in fig .
this results in a total of 5,368 original members . with this procedure
, our preliminary result implies that 2,138 maria asteroids had been injected into the 3:1 mmr region through its dynamical evolution age of 3 @xmath1 1 gyr .
it is well known that the dynamical lifetime of objects placed in orbital resonances is only a few million years @xcite .
their numerical simulations for 3:1 mmr injected particles from the maria family show that the fraction of particles experiencing the end - states of 1 ) being injected into the inner solar system and eventually colliding with the sun or 2 ) being injected into a jupiter - crossing orbit and being eventually removed from the solar system are 70 % and 29 % , respectively .
therefore , we may conclude that roughly 1,500 objects from the maria family had been injected into the inner solar system during 3 @xmath1 1 gyr , i.e. , 37 to 75 maria asteroids larger than 1 km every 100 myr .
we also look into the 3:1 resonance neighborhood region to find possible candidates among maria members residing in this unstable area . according to @xcite
, the boundary of the 3:1 mmr region can be defined approximately by the following formulas ( only for the right side of the 3:1 resonance ) : @xmath35 they found that the number density of asteroids sharply increases up to @xmath00.015 au from the resonance boundary , yet it is not changed considerably over the next 0.025 au .
@xcite also pointed out that there is a chaotic diffusion region in the range of 0.01 @xmath4 0.02 au in semi - major axis from the 3:1 resonance border .
asteroids residing in this region could enter the resonance on a time scale of 100 myr .
the oscillation of the border of the resonance could also have taken place @xcite .
figure 8 represents all 3,230 maria asteroids projected onto a plane of proper semi - major axis and proper eccentricity .
the 3:1 mmr boundary denoted with black dashed line is defined by eq .
the chaotic diffusion region that has been shifted to 0.015 au from the resonance border is represented by the red dashed line .
the number of objects residing in this region is 34 ; they could fall into the 3:1 resonance in the coming 100 myr , especially 114123 ( 2002 vx49 ) and 137063 ( 1998 wk1 ) marked with green open circles are the most promising candidates for becoming new neas .
the number of 34 is surprisingly consistent with the range 37 @xmath4 75 we obtained before , within an error range of age estimation .
large filled blue and red circles stand for prograde and retrograde rotators , respectively .
prograde objects tend to position outward the semi - major axis compared with retrograde , due to the yarkovsky effect .
this can be also seen in fig .
we performed observations of the maria family asteroids from 2008 july to 2013 may . from our observations for 134 nights , synodic rotational periods for 51 objects
were acquired including newly derived periods of 34 asteroids .
the rotational rates distribution of the maria family members obtained from our datasets in addition to published lightcurve data from the lcdb ( asteroid lightcurve database ) indicates that there is a considerable excess of fast and slow rotators .
the two - sample kolmogorov - smirnov test rejects the hypothesis that the spin distribution matches that of a maxwellian at a confidence level of 92% . from correlations among spin rates , sizes , and overall shapes of asteroids ,
we conclude that rotational properties of the maria family have been altered significantly by the yorp effect .
such a substantial change of the rotational characteristic is also observed in old asteroid families such as koronis and flora ( e.g. * ? ? ?
* ; * ? ? ?
* ) . on the other hand , however , we could not detect the slivan state in the maria family members that have relatively higher inclination orbits between 12 @xmath36 and 17 @xmath36 than those of koronis and flora families .
the yorp effect is most effective in the spin - up or spin - down of small asteroids .
consequently , there is an apparent trend toward larger dispersion of rotational frequency with decreasing size in fig .
4 . regarding the correlation between rotational frequencies and the lightcurve amplitude ( see fig .
5 ) , we confirm that there is no object being both fast rotating ( faster than 6 cycles per day ) and having a large lightcurve amplitude ( more than 0.6 mag ) thus verifying that no evidence of objects having high tensile strengths . based on the lightcurve inversion method , we obtained pole orientations for 13 members to trace non - gravitational forces such as the yarkovsky effect .
the excess of prograde rotators with respect to retrograde is found with a ratio ( @xmath2 ) of 3 .
the retrograde objects might have been injected into the inner solar system by entering the 3:1 mean - motion resonance with jupiter .
therefore , the overall trend of proper semi - major axis and absolute magnitude plot is non - symmetric . from our simplest arithmetic of the flux for the 3:1 mmr resonant objects from the maria family
, we estimate that approximately 37 to 75 maria asteroids larger than 1 km have entered the inner solar system every 100 myr . for understanding these non - gravitational forces that are affecting the maria asteroid family in detail , we need more samples in order to improve our statistical result .
this is the reason why we are planning to extend our observational campaign with a network of telescopes in both hemispheres .
the yarkovsky / yorp model based on high accurate parameters such as spin status , size , albedo , density , and thermal conductivity is expected to shed new light on the dynamical evolution in the history of the inner solar system .
we thank the refree of this paper , alberto cellino , for his valuable comments and suggestions that significantly improved the quality of this work .
m.j.k . would like to thank josef hanus for his valuable comments in the asteroid shape modeling using sparse datasets .
m.j.k . would like to thank fumihiko usui for matching family members with the wise data .
was partly supported by korea research council of fundamental science & technology .
work at yonsei was supported by the nrf grant 2011 - 0030875 .
was partly supported by the collaborative research project between nrf - jsps ( 2012010270 ) .
gratefully acknowledges financial support from the korea research council for fundamental science and technology ( krcf ) through the young research scientist fellowship program , and also the support of the korea astronomy and space science institute ( kasi ) grant 2013 - 9 - 400 - 00 .
we also thank the following ( mostly amateur ) astronomers for their observational material collected in the cdr - database and used in this paper : p. antonini , m. audejean , l. bernasconi , j .-
bosch , c. cavadore , m. conjat , m. diebold , j .-
godard , f. hormuth , m. husarik , l. kurtze , a. leroy , j. linder , f. mallman , f. manzini , a. martin , e. morelle , r. montaigut , r. naves , r. poncy , c. rinner , g. rousseau , r. roy , o. thizy , m. tlouzeau , and b. warner .
we present 61 individual lightcurves ( with a reliability code @xmath24 2 ) for a total of 51 maria family members ( fig .
a.1 @xmath4 61 ) .
all figures were drawn as composite lightcurve folded with their synodic rotational period .
horizontal axes of each lightcurve represent rotational phase and vertical axes are differential magnitude between instrumental magnitude of the asteroid and an average magnitude of comparison stars .
alvarez - candal , a. , duffard , r. , angeli , c. a. , et al .
2004 , , 172 , 388 asphaug , e. & scheeres , d. j. 1999 , , 139 , 383 binzel , r. p. , farinella , p. , zappal , v. , & cellino , a. 1989 , in asteroid ii , ed .
r. p. binzel , t. gehrels , & m. s. matthews ( tucson , az : univ .
arizona press ) , 416 bottke , w. f. , vokrouhlick , d. , broz , m. , et al .
2001 , science , 294 , 1693 bottke , w. f. , morbidelli , a. , jedicke , r. , et al .
2002 , , 156 , 399 bottke , w. f. , durda , d. , nesvorn , d. , et al .
2005 , the origin and evolution of stony meteorites . in : dynamics of populations of planetary systems ( z. knezevic & a. milani , eds . ) , pp .
357 - 374 .
cambridge university press , cambridge carbognani , a. 2010 , , 205 , 497 cellino , a. , bus , s. j. , doressoundiram , a. , & lazzaro , d. 2002 , in asteroid iii , ed .
w. bottke , a. cellino , p. paolicchi , & r. binzel ( tucson , az : univ .
arizona press ) , 633 cellino , a. , delloro , a. , & tedesco , e. f. 2009 , , 57 , 173 davis , d. r. , ryan , e. v. , & farinella , p , 1994 , , 42 , 599 durech , j. , kaasalainen , m. , warner , b. d. , et al .
2009 , , 493 , 291 durech , j. , sidorin , v. , & kaasalainen , m. 2010 , , 513 , 46 dohnanyi , j. s. 1969 , , 74 , 2531 fieber - beyer , s. k. , gaffey , m. j. , kelley , m. s. , et al .
2011 , , 213 , 524 giblin , i. , martelli , g. , farinella , p. , et al .
1998 , , 134 , 77 gladman , b. j. , migliorini , f. , morbidelli , a. , et al .
1997 , science , 277 , 197 guillens , s. a. , vieira martins , r. , & gomes , r. s. 2002 , , 124 , 2322 hanus , j. , durech , j. , broz , m. , et al .
2011 , , 530 , 134 hasselmann , p. h. , carvano , j. m. , & lazzaro , d. 2012 , sdss - based asteroid taxonomy v1.1 , ear - a - i0035 - 5-sdsstax - v1.1 , nasa planetary data system hirayama , k. 1918 , , 31 , 185 hirayama , k. 1922 , jajag , 1 , 55 holsapple , k. , giblin , i. , housen , k. , et al .
2002 , in asteroid iii , ed .
w. bottke , a. cellino , p. paolicchi , & r. binzel ( tucson , az : univ .
arizona press ) , 443 housen , k. r. & holsapple , k. a. 1990 , , 84 , 226 housen , k. r. , schmidt , r. m. , & holsapple , k. a. 1991 , , 94 , 180 howell , s. b. 1989 , , 101 , 616 ito , t. & yoshida , f. 2010 , advgeo , 25 , 161 ivezi , z. , lupton , r. h. , juri , m. , et al .
2002 , , 124 , 2943 kaasalainen , m. & torppa , j. 2001 , , 153 , 24 kaasalainen , m. , torppa , j. , & muinonen , k. 2001 , , 153 , 37 kaasalainen , m. , torppa , j. , & piironen , j. 2002 , , 159 , 369 kaasalainen , m. 2004 , , 422 , l39 kryszczyska , a. , colas , f. , poli , m. , et al .
2012 , , 546 , 72 kryszczyska , a. 2013 , , 551 , 102 la spina , a. , paolicchi , p. , kryszczyska , a. , & pravec , p. 2004
, , 428 , 400 lagerkvist , c .- i . , harris , a. w. , & zappal , v. 1989 , in asteroid ii , ed .
r. p. binzel , t. gehrels , & m. s. matthews ( tucson , az : univ .
arizona press ) , 1162 lagerkvist , c .- i . & magnusson , p. 2011
, asteroid photometric catalog v1.1 , ear - a-3-ddr - apc - lightcurve - v1.1 , nasa planetary data system lenz , p. & breger , m. 2005 , coast , 146 , 53 masiero , j. r. , mainzer , a. k. , grav , t. , et al .
2011 , , 741 , 68 masiero , j. r. , mainzer , a. k. , bauer , j. m. , et al .
2013 , , 770 , 7 michel , p. , benz , w. , tanga , p. , et al .
2001 , science , 294 , 1696 milani , a. , knezevic , z. , novakovic , b. , & cellino , a. 2010 , 207 , 769 mizutani , h. , takagi , y. , & kawakami , s. 1990 , , 87 , 307 morbidelli , a. & moons , m. 1995 , , 115 , 60 morbidelli , a. , zappal , v. , moons , m. , et al .
1995 , , 118 , 132 morbidelli , a. & vokrouhlick , d. 2003 , , 163 , 120 neese , c. 2010 , asteroid taxonomy v6.0 , ear - a-5-ddr - taxonomy - v6.0 , nasa planetary data system nesvorn , d. , morbidelli , a. , vokrouhlick , d. , et al .
2002 , , 157 , 155 nesvorn , d. & bottke , w. f. 2004 , , 170 , 324 nesvorn , d. , jedicke , r. , whiteley , r. , & ivezi , z. 2005 , , 173 , 132 nesvorn , d. 2010 , nesvorn hcm asteroid families v1.0 , ear - a - vargbdet-5-nesvornyfam - v1.0 , nasa planetary data system obrien , d. p. & greenberg , r. 2003 , , 164 , 334 palmer , d. m. 2009 , , 695 , 496 paolicchi , p. , verlicchi , a. , & cellino , a. , , 121 , 126 paolicchi , p. 2005
sait suppl . ,
6 , 110 paolicchi , p. & kryszczyska , a. 2012 , , 73 , 70 parker , a. , ivezi , z. , juri , m. , lupton , r. h. , et al .
2008 , , 198 , 138 pravec , p. & harris , a. w. 2000 , , 148 , 12 pravec , p. , harris , a. w. , & michalowski , t. 2002 , in asteroid iii , ed .
w. bottke , a. cellino , p. paolicchi , & r. binzel ( tucson , az : univ .
arizona press ) , 113 robutel , p. & laskar , j. 2001 , , 152 , 4 rubincam , d. p. 2000 , , 148 , 2 skoglv , e. & erikson , a. 2002 , , 160 , 24 slivan , s. m. 2002 , , 419 , 49 slivan , s. m. , binzel , r. p. , crespo da silva , l. d. , et al .
2003 , , 162 , 285 slivan , s. m. , binzel , r. p. , boroumand , s. c. , et al .
2008 , , 195 , 226 slivan , s. m. , binzel , r. p. , kaasalainen , m. , et al .
2009 , , 200 , 514 tanga , p. , cellino , a. , michel , p. , et al .
1999 , , 141 , 65 usui , f. , kuroda , d. , mueller , t. g. , et al .
2011 , , 63 , 1117 vokrouhlick , d. , nesvorn , d. , & bottke , w. f. 2003 , , 425 , 147 vokrouhlick , d. , broz , m. , bottke , w. f. , et al .
2006 , , 182 , 118 warner , b. d. , harris , a. w. , & pravec , p. 2009
, , 202 , 134 zappal , v. , cellino , a. , barucci , a. m. , et al . 1990 , , 231 , 548 zappal , v. , cellino , a. , di martino , m. , & migliorini , f. 1997 , , 129 , 1 ccccccc wo 0.46 m & 2008 jul . 6 @xmath4 dec .
4 & 34:45:48 & + 30:35:45 & 875 & sbig st-10 & 1.10 + tug 1.0 m & 2012 jun .
20 @xmath4 2013 mar .
6 & 30:19:59 & + 36:49:31 & 2538.6 & si 4k ccd & 0.62 + boao 1.8 m & 2012 apr .
15 @xmath4 2013 mar .
10 & 128:58:36 & + 36:09:53 & 1143 & e2v 4k ccd & 0.43 + soao 0.6 m & 2012 feb .
4 @xmath4 2013 may 8 & 128:27:27 & + 36:56:04 & 1354 & e2v 2k ccd & 1.02 + kasi 0.6 m & 2012 jan .
10 @xmath4 oct .
12 & 127:28:31 & + 36:46:53 & 87 & sbig st-8 & 1.06 + loao 1.0 m & 2012 jun .
28 @xmath4 2013 apr . 3 & 249:12:41 & + 32:26:32 & 2776 & e2v 4k ccd & 0.80 + ca 1.2 m & 2012 jul
. 12 @xmath4 19 & 2:32:45 & + 37:13:25 & 2173.1 & e2v 4k ccd & 0.63 + ccccccccc 575 renate & 2012/12/26.7 & 6.99 & + 44.61 & 2.760 & 1.826 & 7.86 & 14.97 & soao + 652 jubilatrix & 2012/05/20.7 & 16.53 & -12.46 & 2.675 & 1.678 & 4.68 & 15.07 & soao + 660 crescentia & 2012/01/10.4 & 2.07 & -4.80 & 2.750 & 2.444 & 20.76 & 14.30 & kasi + & 2012/01/11.4 & 2.08 & -4.66 & 2.751 & 2.458 & 20.79 & 14.32 & kasi + 695 bella & 2012/01/28.7 & 10.17 & -5.59 & 2.885 & 2.004 & 10.51 & 13.78 & kasi + & 2012/01/29.7 & 10.16 & -5.61 & 2.886 & 1.999 & 10.22 & 13.76 & kasi + & 2012/01/30.7 & 10.14 & -5.63 & 2.887 & 1.994 & 9.93 & 13.75 & kasi + & 2012/02/01.7 & 10.11 & -5.64 & 2.889 & 1.984 & 9.35 & 13.71 & kasi + 727 nipponia & 2008/07/06.8 & 17.07 & -6.07 & 2.770 & 1.864 & 11.67 & 13.90 & wo + & 2008/07/07.8 & 17.06 & -6.17 & 2.769 & 1.870 & 11.96 & 13.92 & wo + & 2008/07/29.7 & 16.91 & -8.59 & 2.753 & 2.040 & 17.53 & 14.29 & wo + & 2008/08/12.7 & 16.91 & -10.32 & 2.742 & 2.185 & 19.90 & 14.51 & wo + & 2012/05/18.7 & 16.97 & -1.59 & 2.817 & 1.871 & 8.98 & 13.84 & soao + 787 moskva & 2008/07/07.0 & 0.10 & + 9.41 & 2.240 & 1.829 & 26.55 & 14.16 & wo + & 2008/08/05.0 & 0.38 & + 9.56 & 2.263 & 1.538 & 21.92 & 13.67 & wo + & 2008/08/11.0 & 0.40 & + 9.17 & 2.268 & 1.486 & 20.23 & 13.55 & wo + & 2012/10/10.5 & 23.34 & -2.97 & 2.313 & 1.382 & 11.59 & 13.13 & kasi + & 2012/12/25.4 & 24.00 & -5.31 & 2.407 & 2.301 & 23.97 & 14.74 & soao + & 2012/12/26.4 & 0.02 & -5.22 & 2.408 & 2.316 & 23.92 & 14.75 & soao + 875 nymphe & 2008/08/10.0 & 3.51 & + 15.37 & 2.414 & 2.308 & 24.65 & 16.37 & wo + & 2008/09/25.0 & 4.07 & + 12.12 & 2.489 & 1.826 & 20.34 & 15.80 & wo + & 2008/11/10.9 & 3.71 & + 5.25 & 2.565 & 1.596 & 5.81 & 15.02 & wo + & 2008/12/04.8 & 3.37 & + 2.81 & 2.603 & 1.692 & 10.31 & 15.38 & wo + & 2012/10/12.7 & 3.32 & + 9.27 & 2.467 & 1.560 & 12.15 & 15.16 & soao + & 2012/10/13.7 & 3.31 & + 9.10 & 2.469 & 1.556 & 11.75 & 15.14 & soao + 879 ricarda & 2012/02/22.3 & 9.48 & -3.95 & 2.818 & 1.865 & 6.59 & 16.00 & loao + & 2012/02/25.6 & 9.42 & -3.74 & 2.822 & 1.875 & 6.07 & 16.05 & kasi + 897 lysistrata & 2008/08/05.0 & 3.32 & + 29.02 & 2.604 & 2.618 & 22.39 & 15.61 & wo + & 2008/08/11.0 & 3.44 & + 29.49 & 2.610 & 2.548 & 22.62 & 15.57 & wo + & 2008/11/10.9 & 3.49 & + 28.71 & 2.688 & 1.715 & 4.85 & 14.10 & wo + & 2008/12/04.8 & 3.10 & + 25.08 & 2.706 & 1.772 & 8.16 & 14.35 & wo + & 2012/10/14.7 & 3.39 & + 31.17 & 2.650 & 1.782 & 12.97 & 14.50 & soao + & 2012/10/15.7 & 3.37 & + 31.10 & 2.651 & 1.775 & 12.63 & 14.48 & soao + 1158 luda & 2008/09/23.0 & 7.90 & + 34.29 & 2.354 & 2.556 & 23.11 & 15.79 & wo + & 2008/09/25.0 & 7.96 & + 34.22 & 2.356 & 2.536 & 23.28 & 15.78 & wo + & 2008/11/10.1 & 9.04 & + 33.02 & 2.402 & 2.036 & 24.01 & 15.37 & wo + & 2008/11/24.1 & 9.21 & + 33.10 & 2.417 & 1.885 & 22.46 & 15.17 & wo + & 2008/12/04.1 & 9.27 & + 33.37 & 2.428 & 1.785 & 20.66 & 15.01 & wo + & 2013/04/01.1 & 7.34 & + 31.85 & 2.523 & 2.218 & 23.19 & 15.64 & loao + & 2013/04/02.1 & 7.36 & + 31.69 & 2.524 & 2.232 & 23.22 & 15.66 & loao + 1160
illyria & 2012/02/14.4 & 1.71 & + 22.27 & 2.336 & 2.532 & 22.95 & 16.05 & kasi + & 2012/02/16.4 & 1.76 & + 22.56 & 2.338 & 2.556 & 22.72 & 16.07 & kasi + 1215 boyer & 2008/08/05.7 & 16.88 & -11.37 & 2.701 & 2.058 & 19.15 & 15.84 & wo + & 2008/08/12.7 & 16.89 & -12.27 & 2.692 & 2.131 & 20.28 & 15.94 & wo + & 2008/09/02.7 & 17.05 & -14.93 & 2.664 & 2.369 & 22.13 & 16.21 & wo + & 2012/05/19.6 & 16.77 & -1.47 & 2.838 & 1.881 & 8.16 & 15.35 & soao + 1996 adams & 2012/02/26.6 & 9.38 & + 30.63 & 2.832 & 1.918 & 9.32 & 16.39 & soao + 2151 hadwiger & 2012/01/28.7 & 12.79 & + 10.30 & 2.466 & 1.818 & 20.17 & 15.36 & kasi + & 2012/01/30.7 & 12.79 & + 10.31 & 2.467 & 1.797 & 19.75 & 15.33 & kasi + & 2012/01/31.7 & 12.80 & + 10.32 & 2.467 & 1.787 & 19.53 & 15.31 & kasi + & 2012/02/01.7 & 12.80 & + 10.33 & 2.468 & 1.776 & 19.31 & 15.29 & kasi + 2429 schurer & 2012/02/20.7 & 11.41 & + 20.27 & 2.641 & 1.688 & 7.17 & 15.97 & soao + 3055 annapavlova & 2013/03/20.6 & 12.05 & + 1.11 & 2.647 & 1.651 & 0.53 & 15.81 & soao + 3158 anga & 2012/06/30.9 & 19.38 & + 0.68 & 2.698 & 1.747 & 9.50 & 16.49 & tug + & 2012/07/01.9 & 19.36 & + 0.69 & 2.697 & 1.743 & 9.31 & 16.48 & tug + 3786 yamada & 2012/01/30.4 & 3.56 & + 20.93 & 2.754 & 2.310 & 20.06 & 16.22 & kasi + & 2012/02/07.4 & 3.63 & + 20.64 & 2.755 & 2.418 & 20.69 & 16.34 & kasi + & 2013/01/10.8 & 10.87 & -12.00 & 2.599 & 1.968 & 19.19 & 15.72 & kasi + & 2013/01/13.8 & 10.87 & -12.41 & 2.597 & 1.933 & 18.66 & 15.66 & soao + & 2013/01/14.8 & 10.86 & -12.53 & 2.596 & 1.922 & 18.46 & 15.64 & soao + 3970 herran & 2008/08/26.0 & 3.09 & + 18.29 & 2.351 & 1.892 & 24.65 & 16.78 & wo + & 2008/12/04.8 & 2.04 & + 24.51 & 2.488 & 1.632 & 16.07 & 16.24 & wo + & 2012/10/11.6 & 2.11 & + 15.88 & 2.367 & 1.393 & 6.93 & 15.51 & kasi + 4104 alu & 2013/01/22.9 & 9.70 & + 38.46 & 2.800 & 1.882 & 8.84 & 16.71 & tug + & 2013/02/12.9 & 9.32 & + 40.38 & 2.793 & 1.880 & 9.40 & 16.72 & tug + 4122 ferrari & 2012/12/25.6 & 7.24 & + 14.73 & 2.597 & 1.637 & 5.96 & 15.91 & soao + 4673 bortle & 2012/02/07.7 & 10.72 & + 27.46 & 2.470 & 1.529 & 8.71 & 15.28 & kasi + & 2012/02/10.7 & 10.68 & + 28.03 & 2.472 & 1.524 & 8.08 & 15.25 & kasi + 4851 vodopyanova & 2012/04/27.6 & 13.51 & -10.76 & 2.593 & 1.602 & 4.89 & 16.31 & soao + 5326 ( 1988 rt6 ) & 2012/06/24.9 & 19.77 & + 2.51 & 2.229 & 1.312 & 14.79 & 16.06 & tug + & 2012/06/28.9 & 19.73 & + 2.46 & 2.228 & 1.294 & 13.61 & 15.99 & tug + 5977 ( 1992 th1 ) & 2012/07/01.1 & 22.30 & -6.23 & 2.373 & 1.636 & 20.47 & 16.26 & tug + & 2012/07/04.0 & 22.31 & -6.43 & 2.368 & 1.602 & 19.76 & 16.19 & tug + 6458 nouda & 2008/08/06.9 & 22.27 & + 10.86 & 2.193 & 1.265 & 14.17 & 16.12 & wo + & 2008/08/10.9 & 22.22 & + 10.53 & 2.190 & 1.246 & 12.79 & 16.04 & wo + & 2008/09/23.8 & 21.81 & + 2.32 & 2.176 & 1.269 & 14.86 & 16.14 & wo + & 2008/09/25.8 & 21.80 & + 1.90 & 2.175 & 1.280 & 15.57 & 16.18 & wo + & 2008/10/02.7 & 21.80 & + 0.46 & 2.175 & 1.326 & 17.96 & 16.34 & wo + & 2012/08/08.9 & 21.13 & + 7.75 & 2.214 & 1.247 & 10.60 & 15.98 & tug + & 2012/10/15.7 & 21.01 & -4.02 & 2.178 & 1.584 & 24.86 & 16.94 & tug + 7601 ( 1994 us1 ) & 2013/03/08.6 & 11.11 & + 27.54 & 2.925 & 1.985 & 7.59 & 16.97 & boao + 7644 cslewis & 2012/06/23.9 & 17.55 & -6.51 & 2.548 & 1.566 & 7.558 & 17.25 & tug + & 2012/06/29.8 & 17.46 & -6.20 & 2.539 & 1.575 & 9.243 & 17.33 & tug + 8653 ( 1990 ke ) & 2012/08/10.9 & 23.00 & + 8.34 & 2.371 & 1.459 & 13.69 & 16.48 & tug + & 2012/08/11.9 & 22.99 & + 8.25 & 2.370 & 1.453 & 13.32 & 16.46 & tug + 9175 graun & 2012/12/18.6 & 6.52 & + 42.69 & 2.792 & 1.853 & 7.38 & 16.51 & soao + & 2013/01/10.6 & 6.04 & + 41.46 & 2.819 & 1.904 & 8.96 & 16.65 & soao + & 2013/01/13.5 & 5.99 & + 41.17 & 2.822 & 1.922 & 9.75 & 16.71 & soao + & 2013/03/02.4 & 5.81 & + 35.59 & 2.870 & 2.442 & 19.43 & 17.61 & boao + & 2013/03/03.4 & 5.82 & + 35.47 & 2.871 & 2.456 & 19.51 & 17.63 & boao + & 2013/03/04.4 & 5.83 & + 35.37 & 2.872 & 2.470 & 19.59 & 17.64 & boao + 11129 hayachine & 2012/10/14.5 & 1.92 & + 22.20 & 2.687 &
1.715 & 5.95 & 16.79 & boao + & 2012/10/14.9 & 1.91 & + 22.17 & 2.688 & 1.715 & 5.85 & 16.78 & tug + 11931 ( 1993 dd2 ) & 2012/11/27.6 & 5.34 & + 44.68 & 2.406 & 1.485 & 10.74 & 16.24 & soao + & 2012/12/19.6 & 4.86 & + 45.39 & 2.407 & 1.470 & 9.96 & 16.20 & soao + & 2013/03/31.4 & 5.66 & + 37.81 & 2.429 & 2.476 & 23.47 & 17.80 & boao + 12740 ( 1992 ex8 ) & 2013/03/16.7 & 14.41 & + 6.09 & 2.579 & 1.736 & 14.33 & 16.57 & soao + 13679 shinanogawa & 2013/05/31.9 & 16.58 & + 3.81 & 2.350 & 1.396 & 10.82 & 16.36 & tug + 15288 ( 1991 rn27 ) & 2012/10/16.9 & 3.42 & + 39.12 & 2.606 & 1.759 & 14.14 & 16.81 & tug + & 2012/10/17.9 & 3.41 & + 39.19 & 2.605 & 1.752 & 13.86 & 16.79 & tug + 17157 ( 1999 kp6 ) & 2012/06/28.9 & 19.41 & -1.00 & 2.282 & 1.323 & 11.09 & 15.99 & tug + & 2012/06/30.3 & 19.39 & -1.08 & 2.282 & 1.318 & 10.71 & 15.97 & loao + & 2012/06/30.9 & 19.39 & -1.12 & 2.282 & 1.317 & 10.55 & 15.96 & tug + 18144 ( 2000 oo48 ) & 2008/08/07.0 & 23.44 & -6.17 & 2.185 & 1.277 & 15.48 & 16.98 & wo + & 2008/08/26.9 & 23.32 & -10.68 & 2.178 & 1.183 & 6.31 & 16.44 & wo + & 2012/10/13.5 & 22.29 & -20.03 & 2.175 & 1.413 & 21.17 & 17.38 & boao + & 2012/10/14.5 & 22.29 & -20.08 & 2.175 & 1.423 & 21.45 & 17.40 & boao + 18841 hruska & 2013/03/09.6 & 11.14 & + 28.02 & 2.463 & 1.521 & 9.36 & 16.49 & boao + & 2013/03/10.6 & 11.13 & + 28.00 & 2.464 & 1.524 & 9.57 & 16.50 & boao + 18881 ( 1999 xl195 ) & 2013/03/16.6 & 10.37 & -4.14 & 2.485 & 1.533 & 8.44 & 16.09 & soao + 19184 ( 1991 tb6 ) & 2012/10/18.9 & 3.52 & + 21.92 & 2.357 & 1.441 & 12.22 & 17.09 & tug + 19333 ( 1996 yt1 ) & 2012/08/10.9 & 21.58 & -2.86 & 2.416 & 1.417 & 5.29 & 16.41 & tug + & 2012/08/11.9 & 21.56 & -2.83 & 2.415 & 1.415 & 5.11 & 16.40 & tug + 19495 ( 1998 kz8 ) & 2012/12/24.6 & 6.11 & + 21.39 & 2.396 & 1.413 & 1.04 & 15.11 & soao + & 2013/01/18.5 & 5.71 & + 24.42 & 2.407 & 1.518 & 12.66 & 15.87 & soao + & 2013/03/10.4 & 5.84 & + 28.49 & 2.436 & 2.090 & 23.80 & 16.95 & boao + 19557 ( 1999 jc79 ) & 2012/06/23.9 & 19.82 & -7.05 & 2.297 & 1.352 & 12.22 & 16.50 & tug + & 2012/06/29.9 & 19.76 & -7.68 & 2.289 & 1.316 & 9.84 & 16.34 & tug + 20378 ( 1998 kz46 ) & 2012/11/26.6 & 4.70 & + 26.29 & 2.337 & 1.357 & 3.73 & 15.76 & soao + 21816 ( 1999 te31 ) & 2012/10/19.9 & 1.37 & + 16.35 & 2.356 & 1.365 & 3.07 & 16.85 & tug + 24004 ( 1999 rq57 ) & 2012/10/14.8 & 1.03 & + 22.64 & 2.552 & 1.575 & 5.74 & 16.78 & tug + & 2012/10/16.8 & 1.00 & + 22.59 & 2.554 & 1.578 & 5.78 & 16.79 & tug + 29393 ( 1996 na3 ) & 2012/04/28.6 & 15.24 & + 7.03 & 2.535 & 1.585 & 9.56 & 16.65 & soao + 31554 ( 1999 ej2 ) & 2012/06/24.9 & 18.13 & + 1.58 & 2.393 & 1.433 & 10.35 & 16.34 & tug + & 2012/06/30.9 & 18.05 & + 1.55 & 2.396 & 1.442 & 10.83 & 16.37 & tug + 32116 ( 2000 ld4 ) & 2012/06/21.9 & 17.05 & -4.46 & 2.709 & 1.746 & 8.53 & 17.16 & tug + & 2012/06/29.8 & 16.91 & -5.28 & 2.710 & 1.788 & 11.16 & 17.32 & tug + 33229 ( 1998 fc124 ) & 2012/10/16.8 & 23.85 & -6.34 & 2.280 & 1.355 & 12.08 & 16.18 & tug + 33489 ( 1999 gf9 ) & 2012/08/09.9 & 21.98 & + 10.12 & 2.200 & 1.250 & 12.24 & 16.13 & tug + & 2012/08/11.9 & 21.96 & + 9.93 & 2.200 & 1.244 & 11.63 & 16.10 & tug + 33548 ( 1999 jc13 ) & 2012/08/09.9 & 21.25 & + 4.05 & 2.337 & 1.357 & 8.39 & 16.99 & tug + 33646 ( 1999 jx82 ) & 2012/07/12.9 & 16.92 & -5.27 & 2.394 & 1.530 & 16.11 & 17.29 & ca + & 2012/07/13.9 & 16.92 & -5.41 & 2.393 & 1.537 & 16.40 & 17.31 & ca + & 2012/07/14.9 & 16.91 & -5.56 & 2.393 & 1.544 & 16.69 & 17.33 & ca + & 2012/07/15.9 & 16.90 & -5.69 & 2.393 & 1.550 & 16.98 & 17.35 & ca + & 2012/07/16.9 & 16.90 & -5.85 & 2.392 & 1.557 & 17.26 & 17.37 & ca + & 2012/07/17.9 & 16.89 & -6.00 & 2.392 & 1.565 & 17.54 & 17.39 & ca + & 2012/07/18.9 & 16.89 & -6.14 & 2.392 & 1.572 & 17.82 & 17.41 & ca + & 2012/07/19.9 & 16.89 & -6.29 & 2.391 & 1.580 & 18.09 & 17.43 & ca + 34035 ( 2000 ov27 ) & 2012/07/01.0 & 21.92 & -3.71 & 2.221 & 1.434 & 20.76 & 16.74 & tug + & 2012/07/04.0 & 21.93 & -3.92 & 2.219 & 1.406 & 19.89 & 16.67 & tug + 34502 ( 2000 se157 ) & 2012/10/17.8 & 23.15 & + 17.93 & 2.640 & 1.765 & 12.65 & 17.09 & tug + 34529 ( 2000 sd212 ) & 2008/08/25.0 & 1.27 & + 0.90 & 2.647 & 1.844 & 15.99 & 17.91 & wo + & 2008/09/25.9 & 1.02 & -4.42 & 2.667 & 1.683 & 5.09 & 17.29 & wo + & 2008/10/02.9 & 0.93 & -5.64 & 2.672 & 1.682 & 3.99 & 17.23 & wo + & 2012/10/12.6 & 0.30 & -8.39 & 2.664 & 1.708 & 7.66 & 17.44 & boao + & 2012/10/19.8 & 0.23 & -9.29 & 2.669 & 1.753 & 10.38 & 17.61 & tug + 34572 ( 2000 sy310 ) & 2012/10/12.5 & 23.42 & + 21.14 & 2.403 & 1.481 & 11.61 & 16.17 & soao + & 2012/10/14.5 & 23.40 & + 20.92 & 2.401 & 1.488 & 12.15 & 16.20 & soao + & 2012/10/17.8 & 23.37 & + 20.53 & 2.399 & 1.501 & 13.07 & 16.25 & tug + 39148 ( 2000 wm93 ) & 2012/08/09.9 & 21.13 & + 4.03 & 2.492 & 1.515 & 7.93 & 16.65 & tug + 40664 ( 1999 rf196 ) & 2012/10/18.9 & 2.74 & + 7.24 & 2.291 & 1.320 & 7.36 & 17.14 & tug + 41510 ( 2000 qu171 ) & 2012/08/05.9 & 20.55 & + 4.59 & 2.400 & 1.433 & 9.37 & 17.25 & tug + & 2012/08/08.9 & 20.51 & + 4.47 & 2.397 & 1.432 & 9.60 & 17.25 & tug + 42704 ( 1998 mb32 ) & 2013/01/14.2 & 7.45 & + 29.02 & 2.532 & 1.556 & 3.39 & 16.49 & soao + 42835 ( 1999 ns56 ) & 2012/10/14.9 & 1.91 & + 25.92 & 2.301 & 1.336 & 8.20 & 16.71 & tug + & 2012/10/15.9 & 1.90 & + 25.76 & 2.302 & 1.334 & 7.81 & 16.69 & tug + 43174 ( 1999 xf180 ) & 2013/01/22.9 & 9.82 & + 35.25 & 2.513 & 1.586 & 9.54 & 16.81 & tug + & 2013/02/12.9 & 9.48 & + 38.28 & 2.506 & 1.576 & 9.61 & 16.79 & tug + 49653 ( 1999 jo85 ) & 2012/06/24.8 & 17.34 & -12.18 & 2.439 & 1.449 & 7.07 & 16.76 & tug + 49923 ( 1999 xq174 ) & 2013/03/08.6 & 11.44 & + 27.97 & 2.700 & 1.760 & 8.37 & 17.09 & boao + 50511 ( 2000 dz101 ) & 2013/05/31.9 & 16.46 & + 4.41 & 2.398 & 1.447 & 10.86 & 17.51 & tug + 59748 ( 1999 le14 ) & 2012/07/12.9 & 19.06 & -3.24 & 2.374 & 1.392 & 8.28 & 16.67 & ca + & 2012/07/13.9 & 19.04 & -3.20 & 2.373 & 1.393 & 8.41 & 16.68 & ca + & 2012/07/14.9 & 19.03 & -3.17 & 2.373 & 1.394 & 8.56 & 16.69 & ca + & 2012/07/15.9 & 19.02 & -3.15 & 2.373 & 1.395 & 8.74 & 16.70 & ca + & 2012/07/16.9 & 19.00 & -3.12 & 2.372 & 1.396 & 8.93 & 16.71 & ca + & 2012/07/17.9 & 18.99 & -3.10 & 2.372 & 1.398 & 9.15 & 16.72 & ca + & 2012/07/18.9 & 18.97 & -3.08 & 2.372 & 1.400 & 9.38 & 16.73 & ca + & 2012/07/19.9 & 18.96 & -3.07 & 2.371 & 1.402 & 9.628 & 16.74 & ca + 66584 ( 1999 rm161 ) & 2012/10/11.6 & 1.95 & + 34.61 & 2.298 & 1.377 & 12.35 & 17.64 & boao + 68045 ( 2000 yn44 ) & 2012/10/11.6 & 1.51 & + 30.88 & 2.732 & 1.790 & 8.60 & 18.54 & boao + 71001 natspasoc & 2013/02/04.2 & 6.97 & + 18.32 & 2.460 & 1.567 & 12.19 & 17.26 & loao + & 2013/02/05.2 & 6.96 & + 18.47 & 2.461 & 1.575 & 12.58 & 17.29 & loao + & 2013/03/09.5 & 6.93 & + 22.12 & 2.486 & 1.911 & 21.39 & 18.03 & boao + 71145 ( 1999 xa183 ) & 2013/03/10.6 & 11.21 & + 27.61 & 2.861 & 1.923 & 8.04 & 17.53 & boao + 91533 ( 1999 ry199 ) & 2012/10/12.6 & 1.10 & + 30.27 & 2.821 & 1.873 & 7.83 & 17.87 & boao + & 2013/10/13.6 & 1.08 & + 30.17 & 2.820 & 1.872 & 7.71 & 17.86 & boao + 109792 ( 2001 rn91 ) & 2012/02/20.6 & 9.84 & + 29.42 & 2.865 & 1.912 & 6.48 & 19.11 & boao + 114819 ( 2003 ou11 ) & 2012/10/14.6 & 1.29 & + 27.89 & 2.725 & 1.764 & 7.02 & 18.64 & boao ccccccc 575 renate & 3.67 & 0.14 & 3 & 21.2 & 0.175 & ls@xmath37 + 652 jubilatrix & 2.66 & 0.21 & 3 & 17.1 & 0.168 & - + 660 crescentia & 8.09 & 0.10 & 3 & 43.6 & 0.205 & s@xmath37 + 695 bella & 14.20 & 0.39 & 4 & 41.2 & 0.245 & s@xmath37 + 727 nipponia & 5.07 & 0.12 & 3 & 34.6@xmath38 & 0.212@xmath38 & dt@xmath39 + 787 moskva & 6.06 & 0.39 & 4 & 40.3 & 0.120 & s@xmath39 + 875 nymphe & 12.44 & 0.11 & 4 & 15.2 & 0.192 & - + 879 ricarda & 82.9 & 0.37 & 1 & 17.9 & 0.243 & s@xmath39 + 897 lysistrata & 11.26 & 0.09 & 2 & 29.4 & 0.146 & s@xmath39 + 1158 luda & 6.86 & 0.08 & 3 & 21.1 & 0.192 & s@xmath37 + 1160 illyria & 4.29 & 0.29 & 3 & 14.8 & 0.224 & - + * 1215 boyer * & * 10.36 * & 0.39 & 2 & 14.7 & 0.301 & s@xmath39 + 1996 adams & 3.27 & 0.23 & 4 & 13.5 & 0.141 & - + 2151 hadwiger & 5.87 & 0.08 & 3 & 17.5@xmath38 & 0.209@xmath38 & s@xmath37 + 2429 schurer & 6.66 & 0.57 & 3 & 12.5 & 0.198 & - + 3055 annapavlova & @xmath40 7 & - & 1 & 9.4 & 0.199 & - + 3158 anga & @xmath40 8 & - & 1 & 8.1 & 0.269 & s@xmath39 + 3786 yamada & 4.03 & 0.35 & 3 & 16.7 & 0.234 & s@xmath39 + * 3970 herran * & * 8.09 * & 0.26 & 3 & 8.6 & 0.264 & - + 4104 alu & @xmath40 6 & - & 1 & 8.8 & 0.230 & s@xmath39 + * 4122 ferrari * & * 8.45 * & 0.18 & 3 & 11.8 & 0.221 & - + 4673 bortle & 2.64 & 0.13 & 3 & 12.5 & 0.290 & s@xmath39 + * 4851 vodopyanova * & * 4.90 * & 0.17 & 3 & 7.2@xmath41 & 0.254@xmath41 & - + 5326 ( 1988 rt6 ) & @xmath40 6 & - & 1 & 7.4 & 0.296 & ls@xmath37 + 5977 ( 1992 th1 ) & @xmath40 12 & - & 1 & 11.8 & 0.154 & - + 6458 nouda & 4.203 & 0.56 & 3 & 7.5 & 0.181 & - + * 7601 ( 1994 us1 ) * & * 3.74 * & 0.31 & 3 & 7.3 & 0.190 & s@xmath37 + * 7644 cslewis * & * 2.31 * & 0.22 & 3 & 5.8 & 0.174 & - + 8653 ( 1990 ke ) & @xmath40 14 & - & 1 & 6.6@xmath41 & 0.254@xmath41 & s@xmath37 + 9175 graun & 25.8 & 0.12 & 3 & 7.9 & 0.308 & - + * 11129 hayachine * & * 17.57 * & 0.17 & 2 & 6.2 & 0.350 & - + * 11931 ( 1993 dd2 ) * & * 19.87 * & 0.09 & 3 & 6.7 & 0.363 & - + 12740 ( 1992 ex8 ) & @xmath40 10 & - & 1 & 9.1 & 0.285 & - + 13679 shinanogawa & @xmath40 8 & - & 1 & 5.7 & 0.308 & s@xmath37 + * 15288 ( 1991 rn27 ) * & * 7.14 * & 0.13 & 3 & 9.3 & 0.188 & l@xmath37 + * 17157 ( 1999 kp6 ) * & * 3.72 * & 0.14 & 3 & 6.4 & 0.296 & l@xmath37 + * 18144 ( 2000 oo48 ) * & * 3.10 * & 0.24 & 3 & 3.5 & 0.437 & sq@xmath37 + * 18841 hruska * & * 6 * & 0.11 & 2 & 7.3 & 0.212 & - + * 18881 ( 1999 xl195 ) * & * 2.98 * & 0.18 & 3 & 9.7 & 0.173 & - + * 19184 ( 1991 tb6 ) * & * 4.99 * & 0.36 & 3 & 6.7 & 0.129 & - + * 19333 ( 1996 yt1 ) * & * 2.83 * & 0.16 & 3 & 5.6 & 0.268 & - + * 19495 ( 1998 kz8 ) * & * 2.33 * & 0.12 & 3 & 6.3 & 0.342 & - + 19557 ( 1999 jc79 ) & @xmath40 5 & - & 1 & 8.4 & 0.230 & s@xmath37 + * 20378 ( 1998 kz46 ) * & * 5.14 * & 0.31 & 3 & 7.9 & 0.236 & ls@xmath37 + 21816 ( 1999 te31 ) & @xmath40 20 & - & 1 & 4.6 & 0.206 & - + * 24004 ( 1999 rq57 ) * & * 5.68 * & 0.43 & 3 & 9.1@xmath41 & 0.254@xmath41 & - + 29393 ( 1996 na3 ) & @xmath40 20 & - & 1 & 6.6@xmath41 & 0.254@xmath41 & ls@xmath37 + 31554 ( 1999 ej2 ) & @xmath40 12 & - & 1 & 7.5 & 0.240 & ls@xmath37 + 32116 ( 2000 ld4 ) & @xmath40 8 & - & 1 & 6.0@xmath41 & 0.254@xmath41 & - + 33229 ( 1998 fc124 ) & @xmath40 5 & - & 1 & 5.4 & 0.377 & - + * 33489 ( 1999 gf9 ) * & * 6.74 * & 0.55 & 3 & 6.0 & 0.198 & l@xmath37 + * 33548 ( 1999 jc13 ) * & * 3.97 * & 0.27 & 3 & 4.8 & 0.227 & l@xmath37 + * 33646 ( 1999 jx82 ) * & * 3.33 * & 0.09 & 3 & 5.2 & 0.218 & l@xmath37 + 34035 ( 2000 ov27 ) & @xmath40 10 & - & 1 & 5.8 & 0.276 & - + * 34502 ( 2000 se157 ) * & * 4.71 * & 0.05 & 3 & 5.7 & 0.344 & q@xmath37 + * 34529 ( 2000 sd212 ) * & * 3.71 * & 0.50 & 3 & 5.2 & 0.261 & - + * 34572 ( 2000 sy310 ) * & * 6.64 * & 0.40 & 3 & 7.3 & 0.252 & - + * 39148 ( 2000 wm93 ) * & * 7.55 * & 0.13 & 2 & 4.1 & 0.315 & - + * 40664 ( 1999 rf196 ) * & * 9.61 * & 0.17 & 2 & 3.5 & 0.368 & - + 41510 ( 2000 qu171 ) & @xmath40 24 & - & 1 & 4.8 & 0.207 & - + 42704 ( 1998 mb32 ) & @xmath40 12 & - & 1 & 5.3 & 0.333 & - + * 42835 ( 1999 ns56 ) * & * 2.49 * & 0.14 & 3 & 5.4 & 0.222 & - + 43174 ( 1999 xf180 ) & @xmath40 16 & - & 1 & 5.5 & 0.214 & - + 49653 ( 1999 jo85 ) & @xmath40 8 & - & 1 & 5.5 & 0.209 & - + * 49923 ( 1999 xq174 ) * & * 5.21 * & 0.15 & 3 & 5.3 & 0.329 & l@xmath37 + 50511 ( 2000 dz101 ) & @xmath40 8 & - & 1 & 3.7 & 0.205 & - + * 59748 ( 1999 le14 ) * & * 6.65 * & 0.05 & 3 & 5.6 & 0.223 & - + * 66584 ( 1999 rm161 ) * & * 9.06 * & 0.09 & 3 & 3.6 & 0.137 & - + * 68045 ( 2000 yn44 ) * & * 4.35 * & 0.12 & 3 & 3.3@xmath41 & 0.254@xmath41 & - + * 71001 natspasoc * & * 3.15 * & 0.14 & 3 & 5.3 & 0.271 & ls@xmath37 + * 71145 ( 1999 xa183 ) * & * 12 * & 0.09 & 2 & 6.1 & 0.246 & c@xmath37 + * 91533 ( 1999 ry199 ) * & * 3.85 * & 0.53 & 3 & 5.0 & 0.283 & - + 109792 ( 2001 rn91 ) & @xmath40 16 & - & 1 & 2.5 & 0.101 & s@xmath37 + 114819 ( 2003 ou11 ) & @xmath40 10 & - & 1 & 3.3 & 0.211 & - ccccccccccc 170 maria & * 103 * & * -18 * & 289 & -5 & * 112 * & 13.1313212 & 10 & 183 & 112 & - + 575 renate & 6 & + 3 & * 183 * & * -19 * & * 105 * & 3.6603283 & 5 & 108 & 191 & - + 652 jubilatrix & * 143 * & * + 34 * & 321 & + 27 & * 69 * & 2.6626693 & 5 & 84 & 163 & - + 660 crescentia & 58 & + 1 & * 241 * & * + 15 * & * 90 * & 7.9140789 & 10 & 220 & 158 & - + 695 bella & * 81 * & * -58 * & 308 & -49 & * 149 * & 14.2190116 & 15 & 203 & 144 & 87 , -55 + & & & & & & & & & & 314 , -56 + 727 nipponia & * 172 * & * + 11 * & 333 & + 5 & * 89 * & 5.0687459 & 10 & 239 & 185 & - + 787 moskva & 132 & + 11 & * 334 * & * + 44 * & * 55 * & 6.0558066 & 30 & 164 & 132 & 126 , + 27 + & & & & & & & & & & 331 , + 59 + 875 nymphe & * 47 * & * + 32 * & 200 & + 46 & * 52 * & 12.6212926 & 18 & 95 & 176 & 42 , + 29 + & & & & & & & & & & 196 , + 41 + 897 lysistrata & * 109 * & * -55 * & 299 & -29 & * 136 * & 11.2742816 & 7 & 186 & 134 & - + 1158 luda & 88 & + 67 & * 267 * & * + 21 * & * 54 * & 6.8750690 & 12 & 106 & 177 & - + 1996 adams & * 106 * & * + 56 * & 281 & + 17 & * 48 * & 3.3111390 & 3 & 82 & 172 & 107 , + 55 + 3786 yamada & * 77 * & * + 61 * & 236 & + 63 & * 33 * & 4.0329422 & 9 & 18 & 131 & - + 6458 nouda & * 63 * & * -2 * & 240 & + 21 & * 81 * & 4.2030968 & 9 & - & 163 & - | maria family is regarded as an old - type ( @xmath03 @xmath1 1 gyr ) asteroid family which has experienced substantial collisional and dynamical evolution in the main - belt .
it is located nearby the 3:1 jupter mean motion resonance area that supplies near - earth asteroids ( neas ) to the inner solar system .
we carried out observations of maria family asteroids during 134 nights from 2008 july to 2013 may , and derived synodic rotational periods for 51 objects , including newly obtained periods of 34 asteroids .
we found that there is a significant excess of fast and slow rotators in observed rotation rate distribution .
the two - sample kolmogorov - smirnov test confirms that the spin rate distribution is not consistent with a maxwellian at a 92% confidence level . from correlations among rotational periods , amplitudes of lightcurves , and sizes ,
we conclude that the rotational properties of maria family asteroids have been changed considerably by non - gravitational forces such as the yorp effect . using a lightcurve inversion method @xcite , we successfully determined the pole orientations for 13 maria members , and found an excess of prograde versus retrograde spins with a ratio ( @xmath2 ) of 3 .
this implies that the retrograde rotators could have been ejected by the 3:1 resonance into the inner solar system since the formation of maria family .
we estimate that approximately 37 to 75 maria family asteroids larger than 1 km have entered the near - earth space every 100 myr . |
the pibeta experiment@xcite at the paul scherrer institute ( psi ) is a comprehensive set of precision measurements of the rare decays of the pion and the muon .
the goals of the experiment s first phase are : to improve the experimental precision of the pion beta decay branching ratio , @xmath4 ( also referred to as @xmath5 , or @xmath6 ) , from the present @xmath7 to @xmath8 . to measure the branching ratio of the radiative pion decay @xmath9 ( @xmath1 , or rpd ) , enabling a precise evaluation of the pion axial - vector form factor @xmath10 , and limits on the tensor form factor @xmath11 , predicted to vanish in the standard model ( sm ) .
an extensive measurement of the radiative muon decay rate , @xmath12 , with broad phase space coverage , enabling a search for non-(v@xmath13a ) admixtures in the weak lagrangian .
the experiment s second phase calls for a precise measurement of the @xmath14 ( known as @xmath15 ) decay rate , used for normalization in the first phase .
the current 0.33% accuracy would be improved to under 0.2% , in order to provide a precise test of lepton universality , and , hence , of certain extensions to the standard model . in this report
we focus mainly on parts ( i ) and ( ii ) above , and briefly discuss the first muon radiative results , ( iii ) . the rare pion beta decay , @xmath16 ( branching ratio @xmath17 ) , is one of the most basic semileptonic electroweak processes .
it is a pure vector transition between two spin - zero members of an isospin triplet , and is therefore analogous to superallowed fermi ( sf ) transitions in nuclear beta decay .
the conserved vector current ( cvc ) hypothesis@xcite and quark - lepton universality relate the rate of the pure vector pion beta decay to that of muon decay via the cabibbo - kobayashi - maskawa ( ckm ) quark mixing matrix element @xmath18@xcite in a theoretically exceptionally clean way.@xcite hence , pion beta decay presents an excellent means for a precise experimental determination of the ckm matrix element @xmath18,@xcite hindered only by the low branching ratio of the decay .
the ckm quark mixing matrix has a special significance in modern physics as a cornerstone of a unified description of the weak interactions of mesons , baryons and nuclei . in a universe with three quark generations
the @xmath19 ckm matrix must be unitary , barring certain classes of hitherto undiscovered processes not contained in the standard model .
thus , an accurate experimental evaluation of the ckm matrix unitarity provides an independent check of possible deviations from the sm . as the best studied element of the ckm matrix
, @xmath18 plays an important role in all tests of its unitarity .
however , evaluations of @xmath18 from neutron decay have , for the most part , not been consistent with results from nuclear sf decays.@xcite clearly , a precise evaluation of @xmath18 from pion beta decay , the theoretically cleanest choice , is of interest .
radiative pion decay offers unparalleled access to information on the pion s structure .
given the unique role of the pion as the quasi - goldstone boson of the strong interaction , the implications are far reaching . in the standard model description@xcite of the @xmath2 decay , where @xmath20 is a real or virtual photon ( @xmath21 pair ) , the decay amplitude @xmath22 depends on the vector v and axial vector a weak hadronic currents .
both currents give rise to structure - dependent terms sd@xmath23 and
sd@xmath24 associated with virtual hadronic states , while the axial - vector current alone causes the inner bremsstrahlung process ib from the pion and positron . the ib contribution to the decay probability can be calculated in a straightforward manner using qed methods . the structure - dependent amplitude is parameterized by the vector form factor @xmath25 [ constrained by cvc to 0.0259(5 ) ] and the axial vector form factor @xmath10 that have to be extracted from experiments .
the ratio @xmath26 in @xmath9 decay directly determines the chiral perturbation theory parameter sum @xmath27 , or , equivalently , @xmath28 , the pion polarizability.@xcite these quantities are of longstanding interest since they are among the few unambiguous predictions of chiral symmetry and qcd at low energies .
the current status of these measurements is not satisfactory , as there is considerable scatter among the various experimental determinations of @xmath29 , and the accepted particle data group ( pdg ) average value has a 14% uncertainty.@xcite moreover , the statistical accuracy of the present experimental data on the radiative pion decay@xcite can not rule out contributions from other allowed terms in the interaction lagrangian , namely the scalar s , pseudoscalar p , and tensor t admixtures.@xcite nonzero values of any of these amplitudes would imply new physics outside the standard model . in particular
, reports from the istra collaboration@xcite have indicated a nonzero tensor term , @xmath30 .
a careful analysis by herczeg of the existing beta decay data set could not rule out such a value of @xmath11 , which he presumed could be due to leptoquarks@xcite .
in contrast , chizhov proposed a new intermediate chiral boson with an anomalous interaction with matter , in order to account for the apparent non-(v@xmath13a ) behavior in rpd.@xcite finally , radiative muon decay ( rmd ) provides an even better testing ground for non-(v@xmath13a ) interaction terms than does rpd , thanks to the absence of internal structure in the muon .
measuring photon positron energy distributions in rmd allows an evaluation of the @xmath31 decay parameter @xmath32 that is predicted to be zero in the v@xmath13a standard model .
the current limit on @xmath32 is loose , @xmath33 .
it is of considerable interest to bring the uncertainties down to a level competitive with the constraints from standard muon decay . to
the pibeta apparatus is a large - acceptance non - magnetic detector optimized for detection of photons and electrons in the energy range of 5150@xmath34mev with high efficiency , energy resolution and solid angle .
the main sensitive components of the apparatus , shown and labeled in fig .
[ fig : det : xsect ] , are : beam defining plastic scintillator detectors : bc , a thin forward beam counter , ac@xmath35 and ac@xmath36 , cylindrical active collimators , ad , an active degrader , at , a 9-element segmented active target that stops the @xmath37 beam ; charged particle tracking and i d : mwpc@xmath35 and mwpc@xmath36 , cylindrical chambers , and pv , a 20-bar segmented thin plastic scintillator hodoscope ; a 240-element segmented spherical pure - csi shower calorimeter , subtending a solid angle of @xmath38% of @xmath39 .
the entire system is enclosed in a temperature - controlled pb house lined with cosmic muon veto detectors . to collect the available @xmath40 decay events we recorded all non - prompt large - energy ( above the @xmath41 endpoint ) electromagnetic shower pairs occurring in opposite detector hemispheres ( non - prompt two - arm events ) .
in addition , we recorded a large prescaled sample of non - prompt single shower ( one - arm ) events . using these minimum - bias sets , we extract @xmath6 and @xmath15 events , using the latter for normalization . in a stopped pion experiment
these two channels have nearly the same detector acceptance , and have much of the systematics in common .
these two event classes are part a full complement of twelve fast analog triggers comprising all relevant logic combinations of one- or two - arm , low- or high calorimeter threshold , prompt and delayed ( with respect to @xmath37 stop time ) , as well as a random and a three - arm trigger , all of which were implemented in order to obtain maximally comprehensive and unbiased data samples .
details of the method are explained in detail in ref . .
the list of decays measured and/or used for normalization in our work is given in table [ tab : decays ] .
the building and testing of the detector components were completed in 1998 , followed by the assembly and commissioning of the full detector apparatus .
data acquisition with the pibeta detector started in the second half of 1999 , initially at a reduced pion stopping rate , as planned .
the experiment ran subsequently during 2000 and 2001 at @xmath42mhz @xmath37 stopping rate .
we review below the current status of the analysis of the data acquired during this running period . in spite of its low branching ratio ,
the pion beta decay signal , marked by a nearly back - to - back energetic pair of neutral showers from the @xmath43 decay , is strong and without appreciable background .
this is demonstrated in fig .
[ fig : pb : sn : tim ] which shows virtually background - free histograms of @xmath20-@xmath20 relative timing , and of the event time distribution following the pion stop time , @xmath44(deg ) .
the corresponding csi calorimeter energy distributions for the @xmath6 events and for the normalizing @xmath15 events are also given in the same figure we note the excellent agreement between the measured distributions and those simulated by the monte carlo program geant3 .
to to details of the analysis method of the @xmath6 decay channel can be found in refs . and .
applying our method and using the pdg 2004 recommended value of @xmath45 , @xcite we extract the pion beta decay branching ratio : @xmath46 \times 10^{-8}\ , , \label{eq : pb_br}\ ] ] or , in terms of the decay rate , @xmath47\,s^{-1}\,.\ ] ] in both expressions the first uncertainty is statistical , the second systematic , and the third is the @xmath15 branching ratio uncertainty .
our result represents a six - fold improvement in accuracy over the most precise previous measurement@xcite .
furthermore , our result is in excellent agreement with predictions of the sm and cvc given the pdg recommended value range for @xmath18:@xcite @xmath48 and represents the most accurate test of cvc and cabibbo universality in a meson to date .
our result confirms the validity of the radiative corrections for the process at the level of @xmath49 , since , excluding loop corrections , the sm would predict @xmath50 at 90% c.l . using our measured branching ratio @xmath51
, we can calculate a new value of @xmath18 from pion beta decay , @xmath52 , which is in excellent agreement with the pdg 2004 average , @xmath53 .
we will continue to improve the overall accuracy of the @xmath54 decay branching ratio to @xmath550.5% by further refining the experiment simulation and analysis , and by adding new data . during our 1999 - 2001 run we recorded over 40,000 @xmath1 radiative
pion decays ( rpd ) events . besides its intrinsic interest
, the @xmath1 process is an important physics background to other decays under study , particularly the @xmath6 decay .
the pion radiative decay analysis has given us the most surprising result to date , and has commanded significant effort on our part to resolve the issue .
the different event triggers used in our experiment are sensitive to three distinct regions in the @xmath1 phase space : region a with @xmath57 and @xmath20 emitted into opposite hemispheres , each with energy exceeding that of the michel edge ( @xmath58mev ) , recorded in the main two - arm trigger , region b with an energetic photon ( @xmath59 ) , and @xmath60mev , recorded in the one - arm trigger , and region c with an energetic positron ( @xmath61 ) , and @xmath62mev , also recorded in the one - arm trigger .
together , the three regions overconstrain the standard model parameters describing the decay , and thus allow us to examine possible new information about the pion s hadronic structure , or non-(v@xmath13a ) interactions .
the rpd data are of a similar quality to our @xmath6 event set , particularly in region a , as is readily verified in fig .
[ fig : pe2g : sn : tim ] , which shows histograms of @xmath63-@xmath20 time differences and event timing with respect to the pion stop gate time , @xmath64 .
the analysis of these data is involved ; more details can be found in refs . and .
to the dependence of the region-@xmath65 experimental and theoretical branching ratio on the value of @xmath20 is shown in fig . [
fig : rp_br ] ( left ) , indicating two solutions .
the positive @xmath20 solution is preferred by a @xmath66 ratio of @xmath67:1 once data from regions @xmath68 and @xmath69 are included in the analysis ( right panel ) .
we compare the experimental and theoretical branching ratios for the three phase space regions in table [ tab : p2eg : br ] .
we note that due to the large statistical and systematic uncertainties present in all older experiments , our values are consistent with previously published measurements .
the best cvc fit to our data yields @xmath70 this result represents a four - fold improvement in precision over the previous world average @xmath71 @xcite .
it is consistent with chiral lagrangian calculations @xcite , and will lead to a correspondingly improved precision in the order @xmath72 chiral constant @xmath73 @xcite . to thus , our experimental @xmath2 branching ratios and energy distributions in kinematic regions @xmath65 and @xmath69 are compatible with the ( @xmath74 ) interaction .
the sizable 19% shortfall in the measured branching ratio in region @xmath68 dominates the high value of @xmath75 ( table [ tab : p2eg : br ] ) , and is disconcerting . in a fit restricted to region @xmath65 data only ,
we obtain @xmath76 ; this result remains unchanged if region @xmath69 data are added to the fit .
significantly , all previous studies except one@xcite ( which , too , found an anomaly ) , have analyzed only data with kinematics compatible with our region @xmath65 .
an illustration of the nature of the observed discrepancy is given in the region-@xmath68 plots of a conveniently defined kinematic variable @xmath77 in fig .
[ fig : rp2a_l12prop ] .
addition of a small negative tensor term , @xmath78 , improves agreement with the data .
[ b ] to as of this writing the pibeta collaboration is pursuing a dedicated run to determine the extent and nature of the observed discrepancy more precisely .
the 1999 - 2001 run produced a set of some 300,000 radiative muon decay events , increasing the world set by two orders of magnitude .
due to an emphasis in the early analysis on the pion rare decay channels , the rmd analysis has not yet reached the same sub-1% level of precision .
the analysis is currently at the @xmath80% level and both integrated and differential branching ratios are in good agreement with the v@xmath13a sm predictions at this level.@xcite of particular interest are the shapes of differential distributions of @xmath81 , a suitably defined kinematic variable ( analogous to @xmath77 in rpd ) , shown in fig .
[ fig : rmd : delta ] .
clearly , nothing like the @xmath82% discrepancy of region @xmath68 in rpd is observed here . to
this work is in progress , and the 2004 dedicated rpd run is expected to more than double the rmd event set , adding high - quality new data .
interpretation of the pibeta rmd data below the 1% accuracy level will not be possible without new reliable radiative corrections for this process .
over a decade after the last precision measurements of the rare pion and muon decays were completed , the pibeta experiment at psi is revisiting the field , and producing new results . unlike any experiment exploring the field before it , the pibeta project measures simultaneously practically all of the rare @xmath83 and @xmath84 decays .
this gives it an unusually powerful set of built - in consistency checks .
the pion beta decay precision has been improved six - fold , resulting in a first definitive test of the cvc and radiative corrections in a meson .
work on this decay channel is continuing , and will reach @xmath85% accuracy in this first phase of the project .
the current result , with a combined systematic and statistical uncertainty of 0.64% is in excellent agreement with the sm predictions .
the pibeta radiative pion decay measurements cover a broader region of phase space than previous experiments , with more than an order of magnitude higher statistics , and have brought about a four - fold improvement in the precision of the pion axial form facto .
however , the unexpected and pronounced deviation from the v@xmath13a description of the process is limiting this precision , and raising important questions .
this matter is being addressed in a current dedicated run .
finally , in the following , second phase of the project , the pibeta collaboration will turn its attention to the @xmath86 process which provides the best test of lepton universality and a selective check on possible new physics . | after a decade of no measurements of pion and muon rare decays , pibeta , a new experimental program is producing its first results .
we report on a new experimental study of the pion beta decay , @xmath0 , the @xmath1 radiative decay , @xmath2 , and muon radiative decay , @xmath3 .
the new results represent four- to six - fold improvements in precision over the previous measurements . excellent agreement with standard model predictions
is observed in all channels except for one kinematic region of the @xmath1 radiative decay involving energetic photons and lower - energy positrons . |
experimental evidence for oscillations among the three neutrino generations has been recently reported @xcite . since quasielastic ( qe ) scattering forms an important component of neutrino scattering at low energies
, we have undertaken to investigate qe neutrino scattering using the latest information on nucleon form factors .
recent experiments at slac and jefferson lab ( jlab ) have given precise measurements of the vector electromagnetic form factors for the proton and neutron .
these form factors can be related to the form factors for qe neutrino scattering by conserved vector current hypothesis , cvc .
these more recent form factors can be used to give better predictions for qe neutrino scattering and better determination of the axial form factor , @xmath4 .
the hadronic current for qe neutrino scattering is given by @xcite @xmath6u(p_1),\end{aligned}\ ] ] where @xmath7 , @xmath8 , and @xmath9 .
here , @xmath10 and @xmath11 are the proton and neutron magnetic moments .
we assume that there are no second class currents , so the scalar form factor @xmath12 and the tensor form factor @xmath13 need not be included .
= 6.51 in the form factors @xmath14 and @xmath15 are given by : @xmath16 @xmath17 we use the cvc to determine @xmath18 and @xmath19 from the electron scattering form factors @xmath20 , @xmath21 , @xmath22 , and @xmath23 : @xmath24 @xmath25 previously , many neutrino experiment have assumed that the vector form factors are described by the dipole approximation .
@xmath26 @xmath27 @xmath28 we refer to the above combination of form factors as ` dipole form factors ' .
it is an approximation that has been improved by us in a previous publication @xcite .
we use our updated form factors which we refer as ` bba-2003 form factors ' @xcite @xcite ( budd , bodek , arrington ) .
the axial form factor is given by @xmath29 we have used our updated value of @xmath30=1.00 @xmath31 0.020 gev @xcite which is in good agreement with the theoretically corrected value from pion electroproduction of 1.014 @xmath31 0.016 gev @xcite . for extraction of @xmath4
we use the value of @xmath30 = 1.014 , since it is independent of qe scattering measurements .
figures [ elas_jhakjhaj_nu ] shows the qe cross section for @xmath5 using bba-2003 form factors and @xmath30=1.00 gev the normalization uncertainty in the data is approximately 10% .
the solid curve uses no nuclear correction , while the dotted curve @xcite uses a nuance @xcite calculation of a smith and moniz @xcite based fermi gas model for carbon .
this nuclear model includes pauli blocking and fermi motion , but not final state interactions .
the fermi gas model was run with a 25 mev binding energy and 220 mev fermi momentum .
the dotted curve is the prediction for carbon including both fermi gas pauli blocking and the effect of nuclear binding on the nucleon form factors as modeled by tsushima _
et al . _
the ratio of bounded form factors to free form factors is set to 1 for @xmath32 .
the updated form factors improve the agreement with neutrino qe cross section data and give a reasonable description of the cross sections from deuterium .
we plan to study the nuclear corrections , adopting models which have been used in precision electron scattering measurements from nuclei at slac and jlab .
a substantial fraction of the cross section comes from the form factor @xmath4 .
therefore , we can extract @xmath4 from the differential cross section .
figure [ ddsigma_dq_dff_nu ] and [ ff_contribution_nu ] show the contribution of @xmath4 to @xmath33 .
figure [ ddsigma_dq_dff_nu ] shows the percent change in the neutrino cross section for a 1% change in the form factors .
figure [ ff_contribution_nu ] shows the fractional contribution of the form factor determined by setting the form factor to zero and by determining the fractional decrease in the differential cross section . since some terms are products of different form factors , the sum of the curves do not have be 1 .
to extract @xmath34 , we write the equation for @xmath35 in terms of a quadratic function of @xmath4 .
@xmath36 for each @xmath37 bin , we integrate the above equation over the @xmath37 bin and the neutrino flux .
@xmath38 the above equation can be written as a quadratic equation in @xmath34 at the bin value @xmath39 .
@xmath40 the terms of this equation are given below : @xmath41 @xmath42 @xmath43 to find @xmath39 , we assume a nominal @xmath4 , written @xmath44 .
we determine @xmath39 from @xmath45 @xmath46 is a bin center correction term which also uses @xmath44 .
@xmath46 is determined by @xmath47 the number of events in the bin is given by @xmath48 .
the number of events in the bin from theory is @xmath49 the errors in the points are given by @xmath50 figure [ f_a_lin_mva ] and [ f_a_log_mva ] show our extracted values of @xmath4 for the three deuterium bubble chamber experiments . for these plots
the curve shown in the figures is a dipole with @xmath51=1.014 , the value extracted from pion - electro production .
the data and fluxes given in their papers are used in the extraction of @xmath4 .
these plots show the previous data is not sufficient to determine the form for @xmath4 .
in addition , we have shown the expected values for miner@xmath5aand its errors .
we have plotted miner@xmath5a assuming it is a dipole .
we have assumed a 4 year run with 3 tons of fiducial volume and included the effects of inefficiencies and backgrounds .
resolution smearing and systematic errors are not included .
figure [ f_a_mva ] plots @xmath4/dipole to show how well miner@xmath5a can measure @xmath4 .
@xmath20 from electron scattering experiments depends upon the measuring technique @xcite .
for miner@xmath5a we show @xmath34 under the assumption @xmath4@xmath52dipole = @xmath20@xmath52dipole from the cross section technique ( rosenbuth separation ) and @xmath4@xmath52dipole = @xmath20@xmath52dipole from polarization transfer technique .
the miner@xmath5a errors are plotted assuming the plotted @xmath4 is the nominal @xmath4 .
we see that the measurement of @xmath4 from miner@xmath5acan distinguish between these to the two possible forms .
in addition , miner@xmath5a can determine whether @xmath4 is a dipole or not .
the determination of @xmath4 will have systematic errors from the flux , nuclear effects , qe identifications , background determination , etc .
anti - neutrino data can provide a check on @xmath4 .
figure [ ddsigma_dq_dff_nub ] and [ ff_contribution_nub ] show the contribution of @xmath4to the cross section vs @xmath53 for anti - neutrinos .
figure [ ddsigma_dq_dff_nub ] shows the percent change in the anti - neutrino cross section for a 1% change in the form factors .
the plot shows that @xmath4 has a different contribution to the cross section for anti - neutrinos than neutrinos . at @xmath54
, @xmath34 is not contributing to the cross section , and the cross section becomes independent of @xmath4 .
hence , at higher @xmath53 the cross section can be predicted and compared to the data to determine errors to the neutrino extraction .
figure [ ff_contribution_nub ] shows the fractional contribution of the form factor determined by setting the form factor to zero and by determining the fractional decrease in the differential cross section .
note , since some terms are products of different form factors the sum of the curves do not have to sum to 1 .
figure [ f_a_dipole_nub ] shows the errors on @xmath34/dipole for anti - neutrinos .
the overall errors scale is arbitrary . as we expect
, the errors on @xmath4 become large at @xmath53 around 3 @xmath55 when the derivative of the cross section with respect to @xmath4 goes to 0 .
we have used new form factors to show the cross sections for qe neutrino scattering .
the cross sections give a reasonable description of the deuterium data , but the nuclear data is low .
we have shown how to extract @xmath34 and have shown show well miner@xmath5acan measure @xmath4 . for anti - neutrino data at high @xmath53
, @xmath34 has a different contribution , so anti - neutrinos provides a check for the extraction of @xmath34 from neutrinos .
this work is supported in part by the u. s. department of energy , nuclear physics division , under contract w-31 - 109-eng-38 ( argonne ) and high energy physics division under grant de - fg02 - 91er40685 ( rochester ) . | we calculate the quasielastic cross sections for neutrino scattering on nucleons using up to date fits to the nucleon elastic electromagnetic form factors @xmath0 , @xmath1 , @xmath2 , @xmath3 , and weak form factors .
we show the extraction of @xmath4 for neutrino experiments .
we show how well miner@xmath5a , a new approved experiment at fnal , can measure @xmath4 .
we show the that @xmath4 has a different contribution to the anti - neutrino cross section , and how the anti - neutrino data can be used to check @xmath4 extracted from neutrino scattering .
( presented by howard budd at nuint04 , mar .
2004 , laboratori nazionali del gran sasso - infn - assergi , italy @xcite ) |
word sense disambiguation is the process of identifying the correct sense of words in particular contexts .
the solving of wsd seems to be ai complete ( that means its solution requires a solution to all the general ai problems of representing and reasoning about arbitrary ) and it is one of the most important open problems in nlp @xcite,@xcite,@xcite , @xcite,@xcite,@xcite . in the electronical on - line dictionary wordnet ,
the most well - developed and widely used lexical database for english , the polysemy of different category of words is presented in order as : the highest for verbs , then for nouns , and the lowest for adjectives and adverbs .
usually , the process of disambiguation is realized for a single , target word .
one would expect the words closest to the target word to be of greater semantical importance for it than the other words in the text .
the context is hence a source of information to identify the meaning of the polysemous words .
the contexts may be used in two ways : a ) as _ bag of words _ , without consideration of relationships with the target word in terms of distance , grammatical relations , etc . ;
b ) with relational information .
the _ bag of words _
approach works better for nouns than verbs but is less effective than methods that take other relations in consideration .
studies about syntactic relations determined some interesting conclusions : verbs derive more disambiguation information from their objects than from their subjects , adjectives derive almost all disambiguation information from the nouns they modify , and nouns are best disambiguated by directly adjacent adjectives or nouns @xcite .
all these advocate that a global approach ( disambiguation of all words ) helps to disambiguate each pos .
in this paper we propose a global disambiguation algorithm called * chain algorithm * for disambiguation , chad , which presents elements of both points of view about a context : because this algorithm is @xmath0 it belongs to the class of algorithms which depend of relational information ; in the same time it does nt require syntactic analysis and syntactic parsing .
in section 2 of this paper we review lesk s algorithm for wsd . in section 3
we present `` triplet '' algorithm for three words and chad algorithm . in section 4
we describe some experiments and evaluations with chad .
section 5 introduces some conclusions of using the chad for translation ( here from romanian language to english ) and for text entailment verification .
section 6 draws some conclusions and further work .
work in wsd reached a turning point in the 1980s when large - scale lexical resources , such as machine readable dictionaries , became widely available .
one of the best known dictionary - based method is that of lesk ( 1986 ) .
it starts from the idea that a word s dictionary definition is a good indicator for the senses of this word and uses the definition in the dictionary directly . let us remember basic algorithm of lesk @xcite : suppose that for a polysemic target word @xmath1 there are in a dictionary @xmath2 senses @xmath3 given in an equal number of definitions @xmath4 .
here we mean by @xmath5 the set of words contained in the @xmath6-th definition .
consider that the new context to be disambiguated is @xmath7 .
the * reduced form * of lesk s algorithm is : @xmath8 the score of a sense is the number of words that are shared by the different sense definitions ( glosses ) and the context .
a target word is assigned that sense whose gloss shares the largest number of words .
the algorithm of lesk was successfully developed in @xcite by using wordnet dictionary for english .
it was created by hand in 1990s and includes definitions ( glosses ) for individual senses of words , as in a dictionary .
additionally it defines groups of synonymous words representing the same lexical concept ( synset ) and organizes them into a conceptual hierarchy .
the paper @xcite uses this conceptual hierarchy for improving the original lesk s method by augmenting the definitions with non - gloss information : synonyms , examples and glosses of related words ( hypernyms , hyponyms ) . also , the authors introduced a novel overlap measure between glosses which favorites multi - word matching .
first of all we present an algorithm for disambiguation of a triplet . in a sense , our triplet algorithm is similar with global disambiguation algorithm for a window of two words around a target word given @xcite .
instead , our chad realizes disambiguation of all - words in a text with any length , ignoring the notion of `` window '' and `` target word '' and target word in similar studies , all that without increasing the computational complexity .
the algorithm for disambiguation of a triplet of words @xmath9 for dice measure is the following : begin for each sense @xmath10 do for each sense @xmath11 do for each sense @xmath12 do @xmath13 endfor endfor endfor @xmath14 / * sense of @xmath15 is @xmath16 , sense of @xmath17 is @xmath18 , sense of @xmath19 is @xmath20 * / end for the overlap measure the score is calculated as : @xmath21 for the jaccard measure the score is calculates as : @xmath22 shortly , chad begins with the disambiguation of a triplet @xmath23 and then adds to the right the following word to be disambiguated .
hence it disambiguates at a time a new triplet , where first two words are already associated with the best senses and the disambiguation of the third word depends on these first two words .
chad algorithm for disambiguation of the sentence @xmath24 is : begin disambiguate triplet @xmath23 @xmath25 while @xmath26 do calculate @xmath27 calculate @xmath28 @xmath29 endwhile end due to the brevity of definitions in wn many values of @xmath30 are 0 .
we attributed the first sense in wn for @xmath31 in this cases .
in this section we shortly describe some experiments that we have made in order to validate the proposed chain algorithm * chad*. we have developed an application that implements * chad * and can be used to : * disambiguate words ( [ res ] ) ; * translate words into romanian language ( [ app2 ] ) ; * text entailment verification ( 5.2 ) .
the application is written in jdk 1.5.0 . and uses _ httpunit _ 1.6.2 api @xcite . written in java
, httpunit is a free software that emulates the relevant portions of browser behavior , including form submission , javascript , basic http authentication , cookies and automatic page redirection , and allows java test code to examine returned pages either as text , an xml dom , or containers of forms , tables , and links @xcite .
we have used _ httpunit _ in order to search wordnet through the dictionary from @xcite .
more specifically , the following java classes from @xcite are used : * _ webconversation_. it represents the context for a series of http requests .
this class manages cookies used to maintain session context , computes relative urls , and generally emulates the browser behavior needed to build an automated test of a web site . * _
webresponse_. this class represents a response to a web request from a web server . * _
webform_. this class represents a form in an html page . using this class
we can examine the parameters defined for the form , the structure of the form ( as a dom ) , and the text of the form .
we have used _ webform _ class in order to simulate the submission of the form with corresponding parameters .
we tested our chad on 10 files of brown corpus , which are pos tagged .
recall that wn stores only stems of words .
so , we first preprocessed the glosses and the input files , replacing inflected words with their stems .
the reason for choosing brown corpus was the possibility offered by semcor corpus ( the best known publicly available corpus hand tagged with wn senses ) to evaluate the results .
the correct disambiguated words means the disambiguated words as in semcor .
we ran separately chad for : 1 . nouns , 2 .
verbs , and 3 .
nouns , verbs , adjectives and adverbs . in the case of chad
addressed to nouns , the output is the sequence of nouns tagged with senses .
the tag @xmath32 means that for noun @xmath33 the wn sense @xmath6 was found .
analogously for the case of disambiguation on verbs and of all pos .
the results are presented in tables 1 and 2 .
as our chad algorithm is dependent on the length of glosses , and as nouns have the longest glosses , the highest precision is obtained for nouns . in figure 3 , the precision progress can be traced . by dropping and rising
, the precision finally stabilizes to value 0.767 ( for the file br - a01 ) .
the most interesting part of this graph is that he shows how this chain algorithm works and how the correct or incorrect disambiguation of first two words from the first triplet influences the disambiguation of the next words .
it is known that , at senseval 2 contest , only 2 out of the 7 teams ( with the unsupervised methods ) achieved higher precision than the wordnet @xmath34 sense baseline .
we compared in figures 1 , 2 and 3 the precision of chad for 10 files in brown corpus , for dice , overlap and jaccard measures with wordnet @xmath34 sense . comparing the precision obtained with the overlap measure and the precision given by the wordnet @xmath34 sense for 10 files of brown corpus ( br - a01 , br - a02 , br-11 , br-12 , br-13 , br-14 , br - a15 , br - b13 , br - b20 and br - c01 ) , we obtained the following results : * for nouns , the minimum difference was 0.0077 , the maximum difference was 0.0706 , the average difference was 0.0338 ; * as a whole , for 4 files difference was greater or equal to 0.04 , and for 6 files was lower ; * in case of all parts of speech , the minimum difference was 0.0313 , the maximum difference was 0.0681 , the average difference was 0.0491 ; * as a whole , for 7 files difference was greater or equal to 0.04 , and for 3 files was lower ; * relatively to verbs , the minimum difference was 0.0078 , the maximum difference was 0.0591 , the average difference was 0.0340 ; * as a whole , for 4 files difference was greater or equal to 0.04 , and for 6 files was lower .
let us remark that in our chad the standard concept of windows better size parameter @xcite is not working : simply , a window is the variable space between the previous and the following word in respect to the current word .
file & words & dice & jaccard & overlap & wn1 + bra01 & 486&0.758&0.758&0.767&0.800 + bra02 & 479&0.735&0.731&0.758&0.808 + bra14 & 401&0.736&0.736&0.754&0.769 + bra11 & 413&0.724&0.726&0.746&0.773 + brb20 & 394&0.740&0.740&0.743&0.751 + bra13 & 399&0.734&0.734&0.739&0.746 + brb13 & 467&0.708&0.708&0.717&0.732 + bra12 & 433&0.696&0.696&0.710&0.781 + bra15 & 354&0.677&0.674&0.682&0.725 + brc01 & 434&0.653&0.653&0.661&0.728 + file & words & dice & jaccard & overlap & wn1 + bra14 & 931&0.699&0.701&0.711&0.742 + bra02 & 959&0.637&0.685&0.697&0.753 + brb20 & 930&0.672&0.674&0.693&0.731 + bra15 & 1071&0.653&0.651&0.684&0.732 + bra13 & 924&0.667&0.673&0.682&0.735 + bra01 & 1033&0.650&0.648&0.674&0.714 + brb13 & 947&0.649&0.650&0.674&0.722 + bra12 & 1163&0.626&0.622&0.649&0.717 + bra11 & 1043&0.634&0.639&0.648&0.708 + brc01 & 1100&0.625&0.627&0.638&0.688 +
wsd is only an intermediate task in nlp . in machine translation wsd
is required for lexical choise for words that have different translation for different senses and that are potentially ambiguous within a given document .
however , most machine translation models do not use explicit wsd @xcite ( in introduction ) . the algorithm implemented by us
consists in the translation word by word of a romanian text ( using dictionary at http://lit.csci.unt.edu/ rada / downloads / ronlp / r.e .
tralexand ) , then the application of chain algorithm to the english text .
as the translation of a romanian word in english is multiple , the disambiguation of a triplet is modified as following .
let be the word @xmath15 with @xmath35 translations @xmath36 , the word @xmath17 with @xmath37 translations @xmath38 and the word @xmath19 with @xmath39 translations @xmath40 .
each triplet @xmath41 is disambiguated with the triplet disambiguation algorithm and then the triplet with the maxim score is selected : begin for @xmath42 do for @xmath43 do for @xmath44 do disambiguate triplet @xmath45 in @xmath46 calculate @xmath47 endfor endfor endfor calculate @xmath48 optimal translation of triplet is @xmath49 end let us remark that @xmath50 , for example , is a synset which corresponds to the best translation for @xmath15 produced by chad algorithm .
however , since in romanian are used many words linked by different spelling signs , these composed words are not found in the romanian - english dictionary .
accordingly , not each romanian word produces an english correspondent as output of the above algorithm .
however , many translations are still correct .
for example , the translation of expression _ vreme trece _ ( in the poem `` glossa '' of our national poet mihai eminescu ) , is _ word : ( rom)vreme ( eng)age@xmath51 , word : ( rom)trece ( eng)@xmath52 _ . as another example from the same poem , where the synset of a word occurs ( as an output of our application ) , _ ine toate minte _ ,
is translated in _ word : ( rom ) tine ( eng ) @xmath53 : \{keep , maintain } , word : ( rom ) toate ( eng ) @xmath54 : \{wholly , entirely , completely , totally , all , altogether , whole } , word : ( rom ) minte ( eng ) @xmath55 : \{judgment , judgement , assessment}_. the recognition of text entailment is one of the most complex task in natural language understanding @xcite .
thus , a very important problem in some computational linguistic applications ( as question answering , summarization , segmentation of discourse , and others ) is to establish if a text _ follows _ from another text .
for example , a qa system has to identify texts that entail the expected answer .
similarly , in ir the concept denoted by a query expression should be entailed from relevant retrieved documents . in summarization
, a redundant sentence should be entailed from other sentences in the summary .
the application of wsd to text entailment verification is treated by authors in the paper `` text entailment verification with text similarity '' in this volume .
in this paper we presented a new algorithm of word sense disambiguation .
the algorithm is parametrized for : 1 .
all words ( that means nouns , verbs , adjectives , adverbs ) ; 2 .
all nouns ; 3 .
all verbs .
some experiments with this algorithm for ten files of brown corpus are presented in section 4.2 .
the stemming was realized using the list from http://snowball.tartarus.org/algorithms/porter/diffs.txt .
the precision is calculated relative to the corresponding annotated files in semcor corpus .
some details of implementation are given in 4.1 .
we showed in section 5 how the disambiguation of a text helps in automated translation of a text from a language into another language : each word in the first text is translated into the most appropriated word in the second text .
this appropriateness is considered from two points of view : 1 .
the point of view of possible translation and 2 .
the point of view of the real sense ( disambiguated sense ) of the second text .
some experiments with romanian - english translations and text entailment verification are given ( section 5 ) .
another problem which we intend to address in the further work is that of optimization of a query in information retrieval . finding whether a particular sense is connected with an instance of a word is likely the ir task of finding whether a document is relevant to a query .
it is established that a good wsd program can improve performance of retrieval .
as ir is used by millions of users , an average of some percentages of improvement could be seen as very significant . | a large class of unsupervised algorithms for word sense disambiguation ( wsd ) is that of dictionary - based methods .
various algorithms have as the root lesk s algorithm , which exploits the sense definitions in the dictionary directly .
our approach uses the lexical base wordnet for a new algorithm originated in lesk s , namely _ chain algorithm for disambiguation _ of all words ( chad ) .
we show how translation from a language into another one and also text entailment verification could be accomplished by this disambiguation . |
broad emission lines ( bels ) are the distinguishing feature of agn spectra ; and the variability , absorption , profile shape and strength of these lines are our most powerful probes of the inner structure of active galactic nuclei . the basic photoionization model and the evidence for variability in the both the lines and continuum fluxes of agns led blandford & mckee ( 1982 ) to propose `` reverberation mapping '' as a means of constraining the geometry and kinematics of the broad line region ( blr ) .
the principle underlying reverberation mapping is that the light curve of a given line , @xmath2 , can be related to the continuum light curve , @xmath3 , by a causal , linear transformation .
the frequency - dependent version of this transformation is @xmath4 where the transfer function @xmath5 is the emission line response to a @xmath6-function outburst of the ionizing continuum . in principle
, the convolution theorem and fourier inversion can be used to determine @xmath5 from data . the actual geometry and velocity structure of the blr
can then be inferred from some model for the kinematics of the blr and the fact that the surfaces of constant delay are paraboloids aligned along the observer line - of - sight with the central continuum source at the foci .
the difficulties associated with actually performing such an inversion have motivated long - term observations of agn in order to obtain high quality spectra and light curves .
the first of these extensive monitoring programs was the combined iue ( international ultraviolet explorer)/ground - based observations of ngc 5548 in 1988 - 89 ( clavel et al .
1991 ) . several major insights into the blr resulted from the analysis of these ngc 5548 observations .
first of all , the size of the blr was found to be substantially smaller than the size indicated by photoionization calculations of cloud models of the blr ( krolik et al 1991 ) .
furthermore , cross - correlation analysis of light curves of various lines with the continuum has shown that the blr is stratified : longer lag times were found for lower ionization lines such as h@xmath7 and @xmath8 $ ] than for higher ionization lines such as nv , civ and heii ( krolik et al . 1991 ) . also , because of the higher quality spectral data , the response functions of separate components of a single line could be studied .
crenshaw & blackwell ( 1990 ) found evidence for a faster response from the red - shifted half of the civ line versus the blue - shifted half .
the complexity of the blr which these studies revealed called for even better data . the recent hst / iue / ground - based effort to monitor ngc 5548 ( korista et al .
1995 ; hereafter k95 ) has produced the highest quality data yet obtained for probing the blr via reverberation mapping techniques .
k95 have confirmed the difference between the red - shifted and blue - shifted responses of the civ line .
they have also provided better constraints on the ionization stratification for some of the higher ionization state lines .
the most direct evidence of these effects has been obtained by examining the cross - correlation of the emission line light curves with the continuum .
recently , efforts have been made towards deconvolutions of the actual line response functions using various techniques ( horne 1994 ; pijpers 1994 ; krolik & done 1995 ) . underscoring the importance of modeling , `` echo images '' in the @xmath9-@xmath10 plane for various models of the broad line region ( including keplerian disks and spherically symmetric models with radial infall or out - flow )
have been produced , anticipating frequency - resolved response function deconvolutions ( welsh & horne 1991 ; perez , robinson , & de la fuente 1992 ) .
done & krolik ( 1995 ) and wanders et al .
( 1995 ) have made the first attempts at such a deconvolution for the civ line of ngc 5548 .
following k95 , they divided the line into four frequency bands two `` core '' components and two `` wing '' components ; they derived separate response functions for each , confirming the results of the cross - correlation results of k95 , and were able to study specific blr models in greater detail .
current thinking about the blr has been dominated for the past decade by the cloud picture ( see peterson 1994 for a review ) . in these models ,
dense clouds appropriately distributed in velocity are photoionized by the central continuum and emit the line photons which form the broad lines .
since there is very little agreement about how such clouds are formed , how they are confined , and how they attain their velocity distribution , cloud models have the advantage of considerable flexibility in describing the observations .
reverberation mapping and photoionization models have placed some constraints on the nature of the bel cloud region : the characteristic radius of the blr is likely @xmath11-@xmath12 lt - days ; the clouds are almost certainly optically thick , though the absence of lyman edges suggests that individual clouds are significantly smaller than the size of the central continuum emitting region ; the smoothness of the line profiles implies that there are of order @xmath13 clouds contributing to any given line profile .
furthermore , if the clouds are distributed spherically or nearly so , then the tendency for non - zero lags in the line response implies that the cloud emission is anisotropic , with the emission directed primarily towards the central source .
this is consistent with the clouds being optically thick , and ferland et al .
( 1992 ) have performed photoionization calculations to determine the expected emission profile for such clouds . finally
, the faster response times for the red wing of the civ line of ngc 5548 versus the blue wing imply that any radial motion which the clouds possess must be predominantly inward rather than away from the central source .
it is this last feature which has led some to dismiss the disk model of the blr .
because the velocity field in a disk , which is usually taken to be keplerian , is symmetric with respect to red and blue shifted emission , it is expected that the timing response for both sides of the line should also symmetric ( done & krolik 1995 ) . however , this is only necessarily the case if the line emission from the surface of the disk is _ isotropic _ as is generally assumed ( welsh & horne 1991 ; dumont et al .
1990 ; blandford & mckee 1982 ) . in this work ,
we show that in the disk wind model of murray et al .
( 1995 ; hereafter m95 ) the combination of the shears in a keplerian disk and a radiatively driven disk wind produces an anisotropic opacity for the line emission .
this emission anisotropy has several important consequences : it solves the double - peak problem of disk line profiles ( mathews 1982 ) without requiring unrealistically large blr radii ; it provides a natural explanation for the faster response of the red - shifted line emission despite the symmetry in the disk velocity structure ; and it establishes a connection with the observed blueshifted absorption in the civ line of ngc 5548 and the disk wind models of broad absorption line qsos .
furthermore , this model for the broad line region allows the mass of the central object to be constrained , and thus provides additional evidence for the massive black hole hypothesis for the centers of active galactic nuclei .
the plan for the remainder of this paper is as follows : in 2 , we review the disk wind model of m95 , discuss its observational connection with warm absorbers and seyfert galaxies , derive the expressions for computing time - averaged line profiles , and compare a model line profile with data from ngc 5548 . in 3 , we derive the frequency - resolved response functions , discuss how the red / blue timing asymmetry arises , apply our calculations to the ngc 5548 data and compare our results with the deconvolved response functions of done & krolik ( 1995 ) and wanders et al .
( 1995 ) . in 4 ,
we discuss some observational consequences of our model , address possible objections and discuss prospects for more detailed work in the future .
finally , we summarize our conclusions in 5 .
our model for the bel region of agns is based upon the broad absorption line ( bal ) qso model of m95 . in balqsos ,
the broad line spectra are similar to that of ordinary qsos except for the deep , broad absorption troughs blueward of line center . based on this similarity , weymann et al .
( 1991 ) have concluded that ordinary qsos and balqsos belong to the same class of object only the observer s line - of - sight to the central source distinguishes them , passing through absorbing material for the latter objects . because some of the line emission is also absorbed
, the cloud picture requires that two distinct sets of clouds exist in the broad line region .
the broad emission line clouds exist at the inner radii , closer to the central continuum , while the broad absorption line clouds surround them , with appropriate covering factor to allow the absorption line emission .
the complexities and difficulties associated with cloud models of balqsos led m95 to consider a different approach . noting the similarity in the bal profiles to the p - cygni profiles found in o star spectra , they considered the possibility that a radiatively driven wind provides the observed absorption . in o stars
, it has been well established that such winds are responsible for the shape of the resonance lines ( castor , abbott & klein 1975 ; pauldrach et al .
1994 ) . following the formalism of castor et al .
( 1975 ) for radiatively driven winds , m95 calculated the dynamics of a wind which is driven from the surface of an accretion disk .
they found that the velocity and ionization structure of these winds can account for the width and shape of the absorption troughs as well as providing a natural explanation of the inferred covering factor ( @xmath14% ) of the bal region .
the relevance of this model for lower luminosity agn became clearer when a connection was established between the discovery in x - rays of photoionized bulk outflows , the so - called `` warm absorbers '' ( turner et al .
1993 ; mathur et al .
1994 ) , in seyfert galaxies , and the presence of blueward absorption features in resonance line spectra in these same objects ( murray & chiang 1995a ) .
these results suggest that radiatively driven winds may be a ubiquitous feature of all agn .
since broad emission lines are also universal , it is natural that a connection exists between them as well . in order to illustrate the applicability of the disk wind model to the broad emission line region , we consider a simplified description of the velocity structure of the wind .
the wind consists of nearly radial streamlines which have footpoints at various radii on the surface of the disk . near the footpoint radius , @xmath15 , for a given streamline
, the gas is initially driven upwards by the surface radiation of the disk . at a vertical height which is approximately the scale height of the disk ,
the radiation pressure from the central continuum source becomes strong enough to push the gas predominantly radially outwards .
m95 have integrated the equations of motion using the radiation force law determined from calculations using the photoionization code cloudy ( ferland 1993 ) and the line list of verner , barthel , & tytler ( 1994 ) .
they found that the radial component of the gas velocity along a streamline can be described approximately by @xmath16 where the terminal velocity of the streamline , @xmath17 , is within factors of order unity of the local escape velocity , @xmath18 , and the exponent @xmath19-@xmath20 .
it should be noted that this exponent differs from the result for o stars ( @xmath21 , castor et al .
1975 ) because of the differing force multiplier law , as derived from the photoionization calculations , and the importance of the centrifugal support of the rotating keplerian disk . in the following discussion
, we will use @xmath22 .
m95 have performed dynamical and photoionization calculations which show that the inner edge of the wind should be located at @xmath23 cm for an accretion disk around a @xmath24 black hole .
it is important to note that at these radii , the shear in the disk itself is sufficient to make the sobolev approximation an appropriate description of the resonance line transfer in the disk ( cf castor 1970 ) .
murray & chiang ( 1995b ) discuss the validity of this approximation in greater detail . for our purposes , it is sufficient to note that the sobolev length scale associated with the shear due to the differential rotation of the disk is @xmath25 where @xmath26 cm s@xmath27 is the thermal velocity of the disk gas , and @xmath28 is the azimuthal disk velocity .
this length is much smaller than the length scale of the shakura - sunyaev ( 1973 ) thin - disk solution of @xmath29 cm for a @xmath24 black hole ; this implies that , as within the wind itself , the sobolev approximation also applies to the disk .
we expect that the largest contribution to the line emission will be from regions of high density .
therefore , we will assume for now that the line emission originates primarily near the wind footpoints at the surface of the accretion disk where the densities are high ( @xmath30@xmath31 ) , but the continuum optical depths , which are largely due to electron scattering , are still no more than order unity . at these locations ,
the velocity of the gas is composed of a radial component due to the wind and an azimuthal component due to the disk motion : @xmath32 for an observer located in the @xmath33-@xmath34 plane with a line - of - sight angle @xmath35 with respect to the disk symmetry axis ( see figure [ diskschematic ] ) , the projected velocity shift at a location @xmath36 on the disk is @xmath37 note that we have neglected the finite thickness of the disk ( @xmath38 cm ) which is much smaller than the inferred radius of the blr ( @xmath1 cm ) .
because of the azimuthal symmetry of the problem , we model the line source function , @xmath39 , to be a function only of the disk radius .
the contribution to the line profile at a given frequency is @xmath40 where @xmath41 and @xmath42 are the radial boundaries of the blr , @xmath43 is the doppler - shifted frequency of the line at the disk element which has velocity @xmath44 : @xmath45 and @xmath46 is the frequency of the line .
the optical depth @xmath9 is angle dependent and is _ locally _ determined in the sobolev approximation , and @xmath47 is the integrated line opacity ( rybicki & hummer 1983 ) . in the notation of hamann , korista & morris ( 1993 ) the quantity @xmath48
is known as the _ directional _ escape probability . transforming the @xmath6-function and evaluating the @xmath49-integral , we obtain for the integrand of equation [ lnuintegral ] @xmath50 where @xmath49 satisfies @xmath51 for a given radius @xmath52 . the line optical depth in the sobolev approximation is @xmath53 where @xmath54 is the line absorption coefficient , @xmath55 is the density , @xmath56 is the line - of - sight vector to the observer and @xmath57 is the symmetric strain tensor ( castor 1970 ; rybicki & hummer 1978 ) . for the civ line , the line absorption coefficient is @xmath58 assuming a cosmic abundance of carbon and where @xmath59 is the fractional abundance of the c@xmath60 ion .
we expect @xmath59 to be of order unity throughout the civ belr . for the velocity structure given by equations [ radial_vlaw ] & [ disk_velocity ]
, we have @xmath61 \\ & = & \sin^2i\left[v_\infty\frac{r_f}{r^2}\sin^2\phi - \frac{3}{2}\frac{v_\phi}{r}\sin\phi\cos\phi + \frac{v_r}{r}\cos^2\phi\right ] \label { shear}.\end{aligned}\ ] ] evaluated at the high density regions near the footpoints of the wind streamlines , this factor is of order @xmath62 . combining equations [ opticaldepth ] , [ kappa ] , and [ shear ] , the optical depth in the line is of order @xmath63 so we can use @xmath64 the final expression for the luminosity per unit radius is @xmath65 where the final form is obtained by evaluating equation [ dlnudr ] near the wind streamline footpoints , @xmath66 , and neglecting terms which are proportional to @xmath67 .
we note that the remaining @xmath52-dependence is contained in the source function @xmath39 given that @xmath17 and @xmath68 both scale as @xmath69 .
it is not entirely clear how the central continuum source illuminates the outer regions of the disk where the broad emission lines are formed .
the standard picture of agns is that the uv and x - ray continuum emission is emitted from the hot inner regions of the accretion disk .
however , the shakura - sunyaev thin disk solution , which is thought to apply to these inner regions , gives a disk height which is essentially constant from the hot inner regions out to the disk blr .
therefore , in the standard thin - disk solution , there would not be any lines - of - sight from the continuum source to the blr .
however , in disk models of the blr , it is conventional to assume that either that a continuum source exists a moderate distance above the inner regions of the disk ( matt , fabian & ross 1993 ) or that there is a hot , spherical region at the center of the disk ( rokaki & magnan 1994 ) .
murray & chiang ( 1995b ) propose a physical explanation for the latter case wherein resonance line pressure makes an additional contribution to the opacity at the inner regions of the disk causing it to `` puff up . '' in any case , for the purposes of the reverberation mapping study , we will assume that there is a central source of ionizing continuum which is situated not far above the disk plane ( @xmath70 cm ) , and which can illuminate the disk blr .
in order to reproduce a reasonable line profile shape ( cf figure [ lineprofiles ] ) , we have found that the radial dependence in equation [ simdlnudr ] can be adequately modeled by @xmath71 with a value of @xmath72 .
the inner radial boundary of the blr , @xmath41 , is constrained by the reverberation mapping results and photoionization calculations .
the outer radial boundary , @xmath42 , is determined by the location of the strmgren radius for the ion in question .
preliminary photoionization results give @xmath73 for the civ line in ngc 5548 . using this radial dependence and these boundaries ,
we have integrated equation [ lnuintegral ] and find the line profile ( solid line ) for civ shown in figure [ lineprofiles ] . for comparison ,
the dashed line is the mean gex - extracted civ line profile of ngc 5548 from the recent hst observations ( k95 ) and the dot - dashed line results from the same model calculation except that isotropic emission is assumed .
the effect of the anisotropic optical depth is clear : the disk wind shears result in single - peaked profiles , and the shape of the line , except for the blue - ward absorption and a possible contribution from a narrow - line component , are well - matched .
we have used a disk inclination angle @xmath74 for this calculation .
this choice is motivated by the presence of the warm absorber and the expectation that it is due to a highly ionized component of the disk wind which lies in the observer line - of - sight ( murray & chiang 1995a ) .
we note that the value of @xmath35 and the index of the flux radial dependence given in equation [ flux_scaling ] are correlated : the civ line will be better fit by values of @xmath75 for smaller values of @xmath35 .
it should also be noted that these results only apply to the _ shape _ of the line from our model .
accordingly , the model lines shown in figure [ lineprofiles ] are normalized so that their equivalent widths match that of the observed line .
more complete photoionization calculations would have to be performed using the detailed disk wind velocity and ionization structure and the actual ionizing continuum of ngc 5548 for a full comparison with the data to be made .
the preceding arguments are presented to show that the scalings we have inferred are reasonable and that there is physical motivation for the disk wind model .
so far , the line profiles we have presented are for a steady continuum source .
we now derive the transfer function in order to consider variability .
we have @xmath76 for each point on the disk , @xmath36 , there is associated a time - delay and frequency , @xmath77 . transforming the integral over the disk area to one over @xmath9 and @xmath10
, we obtain @xmath78 the denominator is the jacobian of the transformation , and the entire expression is evaluated at the corresponding locations @xmath79 on the disk which yield the desired time - delay and frequency .
figure [ diskmap ] is an image depicting the value of the transfer function as a function of location on the disk ( cf figure [ diskschematic ] ) .
also plotted are the lines of constant time delay ( dashed lines ) and the lines of constant frequency ( solid lines ) to illustrate the mapping from @xmath36 to @xmath77 .
the lines of constant time delay are given by the intersection of the constant time delay paraboloids and the disk surface .
near the disk surface , the radial component of the velocity , @xmath80 cm s@xmath27 , is small compared with the azimuthal velocity @xmath81-@xmath82 cm s@xmath27 throughout the blr .
therefore the projected velocity ( cf equation [ vproj ] ) is approximately @xmath83 so that for a given frequency shift @xmath84 , we have @xmath85 from this expression , we see that the loci of constant absolute shift @xmath86 look like lobes which are nearly red - blue symmetric .
the red - blue asymmetry of the disk response function can be attributed to the radiative transfer effects due to the radial and azimuthal shear of the disk wind . from equation [ simdlnudr ] , we see that the @xmath49-dependence of the disk contribution to the line luminosity is @xmath87 because @xmath88 and both are intrinsically positive quantities , the transfer function attains small values near the median angles ( @xmath89 ) in the upper right and lower left quadrants of the disk as shown in figure [ diskmap ] .
these smaller values of the emission seen by the observer are depicted by the lighter pixels in the image .
figure [ echoimage ] is an echo - image of the response function in @xmath77-space ; note the similarities and differences with respect to the echo - image maps for disk blr models with isotropic emission given in welsh & horne ( 1991 ) and perez et al .
( 1992 ) .
following k95 , we consider the timing response for the blue / red wing and core components of the civ line .
the core components are defined to extend from doppler shifts of 0 to @xmath90 km / s and the wing components from @xmath90 to @xmath91 km / s .
the small arrows in figure [ lineprofiles ] indicate these boundaries at the corresponding wavelengths . we have integrated equation [ response_func ] over frequency for these four bands and found the response functions shown in figure [ responsefunctions ] . as with the response functions found by the deconvolutions ( done & krolik 1995 ; wanders et al .
1995 ) , the core components are nearly identical and are very similar to the blue wing component except at very short time scales , @xmath92 day . most significantly , the red wing has significantly stronger response at shorter time scales ( @xmath93 - 4 days ) than the other three components .
in addition , the secondary peaks found by both wanders et al .
( 1995 ) and done & krolik ( 1995 ) at delays of 8 - 15 days are also a natural consequence of the disk model
. the faster response of the red wing can be understood by considering figure [ diskmap ] .
the blue and red wing frequency ranges are bounded by the solid lines , and the 0.5 , 5 and 10 day time delay lines are indicated by the dashed lines . within the earlier region , between 0.5 and 5 days , the red wing response is stronger than the response of the blue wing .
this occurs because the locations of weaker emission due to the near cancellation of two @xmath49-dependent terms in equation [ phi_dependence ] pass through the blue wing region , but not through the corresponding red wing region . at later times , from 5 to 10 days , the situation is reversed the line of near cancellation of the @xmath49-dependent terms now passes predominantly through the red wing region and does not suppress as much emission in the corresponding blue region . in effect
, the response in the blue wing `` catches - up '' to and surpasses that of the red wing this is evident in the second peak of the blue wing response function at @xmath94 days . in figure
[ echoimage ] , the difference in the red and blue response functions can also be clearly seen .
in order to underscore the applicability of this model to actual data , we have used the combined iue / hst spectra of k95 to create light curves for the four line components and cross - correlation functions of these model light curves with the continuum .
figure [ continuumlightcurve ] shows the 1350 continuum ( mean subtracted ) as measured by iue ( swp ) and hst over an @xmath95 day period .
the solid line is the interpolated continuum we have used for the convolutions , and figure [ wingcorelightcurves ] shows the resulting light curves for each of the four line components .
the earlier response of the red wing relative to the blue wing is clearly evident .
figure [ ccfs ] shows the cross - correlation functions for these light curves with the interpolated 1350 continuum ( cf figure 15 in k95 ) .
table 1 shows lists the peak and centroid values of the inferred lags for each component .
the detailed structure of the frequency - resolved response functions may also be useful in explaining asymmetries in the broad line profiles as well as long term variability in the line structure .
the larger second peak in the blue - wing response at @xmath96 days may either `` alias '' or `` anti - alias '' structure present in the continuum light curve into the mean line profile for a typical observation .
the result would be line profiles which would skewed toward the either the blue or the red depending on the phase and strength of the continuum light curve variations relative to the time window(s ) of the observation .
on the other hand , perry , van groningen , & wanders ( 1994 ) have shown that the variations seen in the shape of the civ line of ngc 4151 are preceded by similar continuum light curves .
they argue that the changes in the line shape may instead be due to structural changes in the blr , an extended or anisotropic continuum source , or a significantly larger blr than that deduced from cross - correlation analysis . in any case , the structure of the response functions will still be reflected in the shape of the emission line in the way we have described .
if these other effects are not significant , then examining the line profile structure and the preceding continuum variations may be an additional means of discriminating between different kinematic models of the blr .
the radial extent of the blr is determined by the reverberation mapping results , and the velocity field is determined by the width of the resonance lines and the photoionization results for the ionization structure of the disk wind .
thus , the observations essentially fix the quantities @xmath52 and @xmath68 .
the assumption of keplerian motion relates these two quantities by @xmath97 so that the mass of the central object is actually constrained by these observations .
now , since @xmath98 for a fixed line width ( i.e. @xmath99 constant ) and since the lags are related to radii on the disk linearly , @xmath100 , we have @xmath101 the lags which we infer from the cross - correlation functions ( figure [ ccfs ] & table 1 ) agree with the values given by k95 to within 20% , and we have determined the mass of the central object of ngc 5548 to be @xmath102 . furthermore , for objects with line widths similar to ngc 5548 , the observed lags for those lines give a direct measurement of the mass of the central object if this model for the blr is correct .
one of the difficulties of this model of the broad line region is that observations of broad , redshifted iron k@xmath0 lines in seyfert 1s seem to indicate that the disks are being viewed predominantly face - on with inclinations of @xmath103 - 30@xmath104 . in these objects ,
the iron emission is believed to originate near the inner edge of the accretion disk at radii @xmath105 , where @xmath106 is the schwarzschild radius of the central black hole ( matt et al .
mushotzky et al .
( 1995 ) report that _ ginga _ observations of ngc 5548 have found that the equivalent width of the fe k@xmath0 line is @xmath107 ev and that this is best fit by disk inclinations of @xmath108-@xmath109 . on the other hand , as done & krolik ( 1995 ) and wanders et al .
( 1995 ) have pointed out , in order for the response functions of a disk with inner radius @xmath1 cm to peak near zero delay , the disk must be highly inclined .
we do not believe that these observations of the iron k@xmath0 lines and our disk - model for the uv blr are irreconcilable .
first of all , the timing resolution of the ngc 5548 observations is no better than 1 day , and we can still produce response functions with the peak response at time delays @xmath110 day for disk inclinations as small as @xmath111 . secondly , mushotzky et al .
( 1995 ) state that only three of the five parameters inner and outer radii of the emitting region , initial line energy , disk inclination , and line flux which are necessary to describe the iron line are actually constrained by the ngc 5548 data .
thirdly , tanaka et al . ( 1995 ) and matt et al . (
1993 ) have indicated that the emitting material in the seyfert 1s may be in a higher ionization state than they have considered and that this can increase the relative yield of fe k@xmath0 photons , thus relaxing the constraint on larger inclination angles . finally , it should be noted that the constraints which are also based on the _ shape _ of the fe k@xmath0 arrived at by tanaka et al .
( 1995 ) for the seyfert galaxy mcg-6 - 30 - 15 are based upon radiative transfer calculations which , to our knowledge , ignore the effect of shears in the disk .
we have seen that these effects can have a significant impact on the shape of the emission line ( cf figure [ lineprofiles ] ) and may also affect the amount of line flux which escapes .
we have not presented here a detailed model of the line emission from ngc 5548 .
the photoionization calculations and our simplified model of the disk wind structure are too crude to be expected to match the observations in detail .
hence , we have not attempted to fit the ngc 5548 light curves of the civ components . in order for that to be accomplished
, photoionization calculations using a more detailed physical model of the disk wind structure must be performed .
this will enable us to determine better the radial dependence for the emission from each ion , and also probe for non - linear response of the line emission to the continuum .
we would also like to use more realistic velocities and shears in equation [ dlnudr ] , and improve the radiative transfer to take into account possible non - local effects from multiple resonant sobolev surfaces ( rybicki & hummer 1978 ) .
despite the incompleteness of this model , we have offered a plausible qualitative explanation of the timing response seen in ngc 5548 using a physically motivated model for the blr . in particular
, the response functions we obtain possess all the features found in the response functions obtained by the deconvolution procedures of wanders et al .
( 1995 ) and done & krolik ( 1995 ) : they peak near zero time delay ; the red wing response is stronger than the response of the other components at time delays of @xmath112 - 4 days ; and the secondary peak at 8 - 14 days is also reproduced .
furthermore , the line profile which we obtain is single - peaked in agreement with the profiles seen for most broad lines in agn spectra , and we have shown that by using the width of the line profile and the timing response of the red and blue components that we can estimate the mass of the central black hole for ngc 5548 we find the mass to be @xmath113 .
the most significant finding of this work is that the radiative transfer effects which reproduce the observed line profile and timing response _ require _ that a wind with large radial shears be present .
therefore , if it is accepted that the broad lines are produced from the surface of a disk , then something which acts like a radiatively driven wind must exist .
we would like to thank kirk korista for generously supplying us with hst spectra of ngc 5548 , and julian krolik for providing us with a manuscript of his work in advance of publication .
this work was supported by nserc of canada and by the connaught fund of the university of toronto .
arav , n. , li , z. , & begelman , m.c .
1994 , apj , 432 , 62 blandford , r. d. , & mckee , c. f. 1982 , apj , 255 , 419 castor , j. i. 1970 , mnras , 149 , 111 castor , j. i. , abbott , d. c. , & klein , r. i. 1975 , apj , 195 , 157 clavel , j. , et al .
1991 , apj , 366 , 64 crenshaw , d. m. , & blackwell , j. h. , jr . 1990 , apj , 358 , l37 dumont , a. m. , & collin - souffrin , s. 1990 , a&a , 229 , 313 done , c. , & krolik , j. h. 1995 apj , submitted emmering , r.t . , blandford , r.d . , & schlossman , i. 1992 , apj , 385 , 460 ferland , g. j. 1993 , university of kentucky department of physics and astronomy internal report ferland , g. j. , peterson , b. m. , horne , k. , welsh , w. f. , & nahar , s. n. 1992 , apj , 387 , 95 hamann , f. , korista , k. t. & morris , s. l. 1993 , apj , 415 , 541 horne , k. 1994 , in reverberation mapping of the broad - line region in active galactic nuclei , eds .
m. gondhalekar , k. horne , and b. m. peterson , ( asp : san francisco ) p. 23 korista , k. , et al .
1995 , apjs , 97 , 285 ( k95 ) krolik , j. h. , horne , k. , kallman , t. r. , malkan , m. a. , edelson , r. a. , & kriss , g. a. 1991 , apj , 371 , 541 krolik , j. h. , & done , c. 1995 , apj , in press mathews , w. g. 1982 , apj , 258 , 425 mathur , s. , wilkes , b. , elvis , m. , & fiore , f. 1994 , apj , 434 , 493 matt , g. , fabian , a. c. , & ross , r. r. 1993 , mnras , 262 , 179 murray , n. , & chiang , j. 1995a , apj , in press murray , n. , & chiang , j. 1995b , in preparation murray , n. , chiang , j. , grossman , s. , & voit , g. m. 1995 apj , 451 , 498 ( m95 ) mushotzky , r. , fabian , a. c. , iwasawa , k. , kunieda , h. , matsuoka , m. , nandra , k. & tanaka , y. 1995 , mnras , 272 , l9 nandra , k. , & pounds , k. a. 1994 , mnras , 268 , 405 pauldrach , a.w.a . , kudritzki , r.p . , puls , j. , butler , k. , & hunsinger , j. 1994 , a&a , 283 , 525 perez , e. , robinson , a. , & de la fuente , l. 1992 , mnras , 256 , 103 perry , j. j. , van groningen , e. , & wanders , i. 1994 , mnras , 271 , 561 peterson , b. m. 1994 , in reverberation mapping of the broad - line region in active galactic nuclei , eds .
p. m. gondhalekar , k. horne , and b. m. peterson , ( asp : san francisco ) p. 1 pijpers , f. p. 1994 , in reverberation mapping of the broad - line region in active galactic nuclei , eds .
p. m. gondhalekar , k. horne , and b. m. peterson , ( asp : san francisco ) p. 69 rokaki , e. & magnan , c. 1992 , a&a , 261,41 rybicki , g. b. , & hummer , d. g. 1978 , apj , 219 , 654 shakura , n. i. , & sunyaev , r. a. 1973 , a&a , 24 , 337 tanaka , y. , et al .
1995 , nature , 375 , 659 turner , t. j. , nandra , k. , george , i. m. , fabian , a. c. , & pounds , k. a. 1993 , apj , 419 , 127 verner , d. a. , barthel , p. d. , & tytler , d. 1994 , a&ass , 108 , 287 wanders , i. et al .
1995 , apj , in press welsh , w. f. , & horne , k. 1991 , apj , 379 , 586 weymann , r. j. , morris , s. l. , foltz , c. b. , & hewett , p. c. 1991 , apj , 373 , 23 | using the disk wind model of murray et al .
( 1995 ) , we calculate line profiles and frequency - resolved response functions for broad line emission from the surface of an accretion disk in an agn in the presence of a radiatively driven wind . we find that the combined effects of the shears in the wind and in the disk itself produce anisotropic line emission which solves several well - known problems connected with disk models of the broad line region . in particular , the broadening of resonance lines such as civ , ly@xmath0 , and nv can be attributed to orbital motion of the disk gas at radii as close as @xmath1 cm in seyferts without requiring unrealistically large emission regions in order to produce single - peaked profiles .
furthermore , the anisotropy of the line emission results in frequency - dependent response functions which are no longer red - blue symmetric so that the time delays inferred for the various red and blue components of the line agree qualitatively with recent reverberation mapping observations of ngc 5548 . |
the investigation of structure and dynamics of networks has been a powerful strategy to analyze interacting many - body problems present in many different areas : biological , ecological , economical and social systems , to name some of them .
the map of these systems into graphs is a fruitful old idea , and the knowledge of the interconnection between its vertices is a necessary condition that allows us to examine a myriad of pratical problems @xcite .
nowadays , there are several research interests involving complex networks .
we can say , for instance , that there is an effort to obtain a better understanding of networks from some of its internal structures like the formation of communities @xcite , or a more complex interconnection of graphs like the multilayer networks @xcite .
the complexity of the internal structure reflects on the entropy of the network @xcite , which shows the possibility of classifying several internal structures . at the same time
, we still have progress on important questions that use complex networks as a framework to define other problems on it ; for instance , we can cite the active area of epidemiological models @xcite , or statistical models on complex networks to analyze critical phenomena @xcite . at this point
, it is worth mentioning that despite the progress in several directions , it is natural that analytical results are less frequent than numerical ones , which is understandable due to the technical complexities presented by many relevant questions .
furthermore , many existing analytical results come from stationary regime . in this scenario ,
we propose a scheme that estimates the time - dependent degree distribution . in order to illustrate our idea
, we revisited the watts - strogatz model @xcite .
although not being a complex networkin the sense that it does not display a heterogeneous degree distribution @xcite , it has small - world property and has high clustering @xcite , two properties shared with many real networks .
the model was originally defined as an intermediate configuration between a regular lattice and a graph where all their nodes are randomly linked , and we will present a slightly modified version from the original one in order to capture its dynamical evolution analytically . the layout of this work is as follows . in purpose of illustrating the main idea of the work , we start with a dynamical version of the erds - rnyi model @xcite in section 2 and we introduce the main model , the time - dependent watts - strogatz graph , in section 3 . then , we present the main idea that allow one to achieve an analytical form for the dynamic degree distribution in section 4 . some final comments are presented in section 5 .
the initial condition of the model consists of @xmath0 vertices and no edges at time @xmath1 . at each time step ,
two vertices are randomly chosen and linked ; this includes the possibility of having a loop ( an edge that connects a vertex to itself ) .
it is clear that each end of an edge links to a vertex with probability @xmath2 .
therefore , defining @xmath3 as the probability that a vertex @xmath4 has degree @xmath5 at time @xmath6 , one can represent the dynamics as @xmath7 with @xmath8 as the initial condition , where @xmath9 is the kronecker symbol ( @xmath10 when @xmath11 , and @xmath12 otherwise ) .
furthermore , @xmath13 is the time - independent conditional probability of changing the degree of a vertex from @xmath14 to @xmath5 ; in the present case , @xmath15 by introducing the time - dependent degree distribution , @xmath16 the time evolution equation ( [ er_me ] ) can be written as @xmath17 if now one introduces the z - transform @xmath18 it is straightforward that @xmath19 where the initial condition is @xmath20 and @xmath21 from ( [ l_er ] ) , one sees that the right - hand side of ( [ phiphi0 ] ) is a polynomial in @xmath22 , and the time - dependent degree distribution @xmath23 is the coefficient of the term of order @xmath24 ( which we will refer as @xmath24-term ) in the right hand side of ( [ phiphi0 ] ) .
hence , since @xmath25 , one can show by a direct calculation that @xmath26 this result will be revisited in section [ rw ] , where we will treat the problem of finding the time - dependent degree distribution as a random walk in degree space . when @xmath27 , which is the time equivalent to the number of possible distinct edges , one recovers the poisson distribution from the usual erds - rnyi model for @xmath28 , @xmath29 the exact form of the time - dependent degree distribution can be used to investigate the shannon entropy , @xmath30\ , .
\label{shannon}\end{aligned}\ ] ] the profile of the entropy can be investigated numerically and is presented in figure 1 .
it starts from a low value and achieves the maximum for @xmath31 , which is when the original ( static ) erds - rnyi model realizes , and the inclusion of more connections decreases the entropy , as one can see from ( [ poisson ] ) .
this phenomenon can be heuristically understood by realizing that the inclusion of edges randomly ( with uniform probability to each possible pair of nodes ) leads the distribution to converge to a kronecker delta , _
i.e. _ , the vertices tend to have all the same degree ( that increases with time ) from the statistical standpoint .
figure 1 : entropy of time - dependent erds - rnyi model ( @xmath32 and @xmath33 ) .
the watts - strogatz model @xcite is a small - world network that , unlike the erds - rnyi graph , keeps high clustering .
the analytical approach treats it as a static model , despite the fact that it is obtained as an intermediate configuration in rewiring process between a regular lattice and a random graph .
we will define a dynamical model that generates a small - world network similar to the one introduced by watts and strogatz . although being slightly different from the original watts - strogatz model , it is statistically equivalent and suitable for analytical investigations .
the initial condition of our model consists of a ring with @xmath0 vertices , and each vertex has degree @xmath34 by having a single link to its @xmath35 next - neighbors as in watts - strogatz model .
the model has , therefore , @xmath36 edges with total degree @xmath37 .
the dynamics obeys the following scheme : \(i ) an edge end is chosen with uniform probability @xmath38 .
\(ii ) this extremity is reconnected with probability @xmath39 ( and kept without reconnection with probability @xmath40 ) .
\(iii ) back to ( i ) ( repetition for a fixed number of iterations ) .
therefore , the probability @xmath3 of a vertex @xmath4 having degree @xmath5 at time @xmath6 obeys the discrete time recurrent equation @xmath41 where @xmath42 stands for the discrete - time transition rate ( conditional probability ) from the state of degree @xmath14 to degree @xmath5 , as in the previous section .
furthermore , the initial condition is @xmath43 consider now a vertex @xmath4 at time @xmath6 ; it can have degree @xmath5 at time @xmath44 in the following scenarios : \i ) the vertex @xmath4 has degree @xmath45 at time @xmath6 and degree @xmath5 at time @xmath44 : an edge - end , which is not connected to @xmath4 , is chosen with probability @xmath46 .
then , it rewires with probability @xmath39 , and links to @xmath4 with probability @xmath47 ; therefore , one has @xmath48
\ii ) the vertex @xmath4 has degree @xmath49 at time @xmath6 and degree @xmath5 at time @xmath44 : an edge - end connected to @xmath4 is chosen with probability @xmath50 .
then , it rewires with probability @xmath39 , and links to another vertex , say @xmath51 , with probability @xmath52 ; therefore , one has @xmath53
\iii ) the vertex @xmath4 has degree @xmath5 at time @xmath6 and remains with degree @xmath5 at time @xmath44 : this scenario is divided in four cases , as follows .
iiia ) an edge - end connected to @xmath4 is chosen with probability @xmath54 , rewires with probability @xmath39 , and links again to @xmath4 with probability @xmath47 ; iiib ) an edge - end connected to @xmath4 is chosen with probability @xmath54 , but does not rewire ( this happens with probability @xmath40 ) ; iiic ) an edge - end not connected to @xmath4 is chosen with probability @xmath55 , rewires with probability @xmath39 , and links to a vertex that is not @xmath4 with probability @xmath52 ; iiid ) an edge - end not connected to @xmath4 is chosen with probability @xmath55 , but does not rewire ( this happens with probability @xmath40 ) ; the conditional probability associated to the union of disjoint events iiia to iiid is @xmath56 the dynamics defined above can generate a graph similar to the watts - strogatz model . for @xmath57 ,
one has a interval of @xmath39 where the system displays high clustering and low mean shortest path length , as shown in figure 2 .
figure 2 : clustering @xmath58 and shortest path length @xmath59 ( normalized by @xmath60 and @xmath61 , respectively ) of the graph generated by the dynamics of section [ tdws ] .
the parameters are @xmath62 , @xmath63 and @xmath64 with @xmath65 realizations of the simulations ; the error bars are smaller than the size of the points
. the time - dependent degree distribution can be evaluated iteratively from the recurrent equation ( [ me ] ) and ( [ pkt ] ) , and this allows one to compute the entropy @xmath66 of the model , which is shown in figure 3 .
figure 3 : entropy of the time - dependent watts - strogatz model for @xmath32 , @xmath67 and @xmath33 .
the entropy starts from a low value , as expected since the initial condition of watts - strogatz model is a regular lattice with @xmath68 .
the entropy , then , grows with time , but reaches a constant value : differently from the erds - rnyi model , the watts - strogatz graph has no new connection being added , and the system converges to a stationary degree distribution different from a kronecker - delta - like as in the erds - rnyi case .
introducing , again , the z - transform ( [ z ] ) to the recurrent equation of the time - dependent degree distribution obtained by combining ( [ me ] ) and ( [ pkt ] ) , one has @xmath69 where the initial condition @xmath70 stands for each vertex having exactly @xmath34 connections .
the explicit form of the operator @xmath71 , which acts on this polynomial , will be presented in the next section . for
now , it is sufficient to state that the analytical form of the time - dependent degree distribution is not well explored in the literature .
this section is devoted to develop the arguments that will establish analytic results concerning the time - dependent degree distribution of the two models above .
the erds - rnyi case will support and illustrate our arguments , since its a simpler laboratory and the exact form ( [ pkt_er ] ) is already known . as seen in section [ tder ] ,
the time - dependent degree distribution @xmath23 is the coefficient of the @xmath24-term in @xmath72 , as one can see from ( [ z ] ) .
moreover , from ( [ phiphi0 ] ) and @xmath73 , we have @xmath74 .
this means that one should search for the @xmath24-term of a polynomial resulted from the application of @xmath75 for @xmath6 times on @xmath76 .
the operator @xmath75 , however , can be divided into a sum of three operators , @xmath77 , @xmath78 and @xmath79 .
this separation is convenient , since when these operators are applied on a monomial @xmath80 ( @xmath81 ) , one has the following behavior : @xmath82 hence , starting from degree @xmath83 , one can see the procedure of applying @xmath6 times the operator @xmath84 as follows . since @xmath85 the @xmath24-term is a sum of many terms , each of them a product of @xmath77 , @xmath78 and @xmath79 .
let us consider @xmath86 as an example ; in this case , the @xmath87-term of @xmath88 is @xmath89 and this is @xmath90 . in the first term ,
the system remains with degree zero at time @xmath91 and increases two unities at @xmath92 ; similar interpretation can be made for the second and third terms .
the time - dependent degree distribution is , therefore , a sum of all trajectories , which are random walks in degree space ( see figure 4 ) , that leads @xmath93 at @xmath1 to degree @xmath5 at time @xmath6 . at each time step
, the degree can increase one unity , or two unities , or stay constant with probabilities @xmath94 , @xmath95 and @xmath96 , respectively ( note that @xmath97 and @xmath98 ) .
hence , denoting by @xmath99 the degree at time @xmath14 , it is straightforward that @xmath100 where @xmath101 for @xmath102 and the first two kronecker deltas refer to the initial and final conditions ; each term inside the parenthesis indicates if the degree at time @xmath14 remains constant or increases ( with one or two unities ) when compared to the degree at the previous instant , @xmath103 .
figure 4 : three examples of possible evolution of the degree ( these examples do not apply for the erds - rnyi model , where the degree never decreses ) . the initial and final degrees should be @xmath34 and @xmath5 , respectively . the continuous version of ( [ er_rw ] ) is a path - integral formulation of the problem .
nevertheless , it does not lead to an expression that can be trivially tackled by the usual methods .
the time - dependent degree distribution can be evaluated explicitely by exploring the property that @xmath77 , @xmath78 and @xmath79 are @xmath104-numbers . during the time interval @xmath6
, there should be @xmath105 , @xmath106 and @xmath107 terms of @xmath96 , @xmath94 and @xmath95 , respectively , such that @xmath108 and @xmath109 . therefore ,
@xmath110 which yields the same result of ( [ pkt_er ] ) , as expected . in ( [ er_step ] ) , @xmath111 is the largest integer equal or less than @xmath112 , and the last equality can be shown after a lengthy induction argument . finally , one can also restate the recurrent equation @xmath113 as @xmath114 , where now we have two types of operators , @xmath115 that act for an interval of time equal to @xmath116 on the initial condition .
similarly as in the previous case , the time - dependent watts - strogatz degree distribution is the @xmath24-term of @xmath117 , where now the initial condition is @xmath70 and @xmath118 with @xmath119 the form of these operators , which are not @xmath104-numbers anymore , can be deduced by ( [ w(k|k-1 ) ] ) , ( [ w(k|k+1 ) ] ) , ( [ w(k|k ) ] ) and the z - transform of ( [ me ] ) .
when these operators are applied on a polynomial of degree @xmath14 , one has @xmath120 with @xmath121 note that now the coefficients @xmath122 , @xmath123 and @xmath124 are not constants and the operators @xmath125 , @xmath126 and @xmath127 do not commute as in erds - rnyi case .
following the same argument that has led to ( [ er_rw ] ) , we have @xmath128 for the watts - strogatz model .
the degree starts with @xmath129 at time @xmath1 and ends with @xmath130 at time @xmath6 . between these boundaries ,
the variable @xmath99 performs a random walk .
this expression is not analytically treatable , and we will invoke some simplifications , which consist of choosing the dominant contributions ( paths ) to @xmath23 . in this section , we will concentrate on the dominant contributions to the degree distribution @xmath23 .
this follows by choosing a class of paths that starts at @xmath129 and ends at @xmath130 . by noticing that @xmath131 , the dominant contributions come from terms that maximize the number of @xmath122-factors
this implies minimizing the number of @xmath123-factors or @xmath124-factors such that they should appear only to change the degree from @xmath34 to @xmath5 . in other terms
, we have @xmath132 terms of @xmath123 ( @xmath124 ) type if @xmath133 ( @xmath134 ) , and the remaining @xmath135 terms are of @xmath122 type . note that these are monotonic paths in the sense that the degree only increases ( if @xmath133 ) or decreases ( if @xmath134 ) .
let us consider initially , the case @xmath136 .
writing the sum of all monotonic paths as being equal to the time - dependent degree distribution leads to @xmath137 the @xmath123 terms are functions of the degree @xmath99 ( see equation ( [ bad ] ) ) , and not on the instant they appear . in the monotonic crescent path , therefore , each term , @xmath138 should appear one and only one time in this order .
the remaining @xmath139 segments of the path are filled by @xmath122-terms , and there should be @xmath140 of them that are @xmath141 , @xmath105 of them that are @xmath142 , and so on ( see figure 5 ) .
firstly it is immediate from ( [ bad ] ) that @xmath143 figure 5 : increasing monotonic paths .
all the monotonic paths , when @xmath133 , are located inside the envelope defined by the dashed lines .
the upper dashed line corresponds to the path @xmath144 , and the lower dashed line is the monotonic path @xmath145 . on the other hand , by using @xmath146 one has @xmath147 where we have performed the change of variables @xmath148 , @xmath149 , @xmath150 up to @xmath151 in the last passage .
therefore , one has @xmath152 and by ( [ bbb > ] ) and ( [ aaa>2 ] ) one finally finds @xmath153^{\delta } \\ & & ( k\geq k_{0})\ , .
\label{pkt>}\end{aligned}\ ] ] the monotonic paths when @xmath134 is such that @xmath154 since now the @xmath124-terms are needed to decrease the degree .
since @xmath155 by following a similar procedure as before , one has @xmath156^{\left|\delta\right| } \\ & & ( k < k_{0})\ , .
\label{pkt<}\end{aligned}\ ] ] for @xmath157 . the comparison between the ( exact ) numerical time - dependent degree distribution obtained from the recurrent equation and the estimations ( [ pkt > ] ) and ( [ pkt < ] ) are shown in figure 6 . the formulas ( [ pkt > ] ) and ( [ pkt < ] ) should be asymptotically exact for @xmath158 and @xmath159 .
the reason for this statement comes from a simple analysis of the order of magnitude of the paths . remembering that @xmath160 , @xmath161 and @xmath162 , a monotonic path is @xmath163 , while there are @xmath164 of them .
the first correction is due terms that have a @xmath123 and @xmath124 terms more than the monotonic paths terms ( and two @xmath122 terms less ) .
each one of its first correction terms are @xmath165 , and there are @xmath166 of them .
the contribution of the first correction is roughly @xmath167 times the contribution of the monotonic paths .
this argument can be extended to corrections of all orders .
therefore , for @xmath168 , one expects that the formulas from the monotonic paths only are asymptotically exact for @xmath158 and @xmath159 .
naturally , the same argument concludes that our estiamtions fail in the case @xmath169 .
the numerical solution in figure 6 shows that our estimations apply in the case @xmath170 , while the same comment can not be made for @xmath169 , as expected .
figure 6 : time - dependent degree distribution .
the points are associated to numerically exact results , and were obtained from the recurrent equation ( [ me ] ) .
the points generated from equations ( [ pkt > ] ) and ( [ pkt < ] ) were interpolated with lines for better visualization .
inset : a detailed visualization of the time - dependent degree distribution ( logarithmic scale for the vertical axis ) for @xmath171 and @xmath172 .
in this work , we have formulated the erds - rnyi and watts - strogatz graphs as a dynamic model and characterized their behavior from the standpoint of their entropies .
we have also examined their time - dependent degree distribution analytically .
the erds - rnyi model is analytically accessible , while the same does not extend to the watts - strogatz model .
we have , nevertheless , obtained a formula that is asymptotically exact for @xmath173 and confirmed this validity numerically .
the main ideia to achieve this result was to consider the evolution of the degree distribution as a random walk in degree space and select the paths that have dominant contribution .
we have also presented the argument that support the range of validity of our formula , which is based on the estimation of the order of magnitude of contribution of relevant terms .
hlcg thanks pnpd / capes ( ed .
82/2014 ) for financial support .
mc acknowledges the oea scholarship program and capes for financial support . | in this work , we propose a scheme that provides an analytical estimate for the time - dependent degree distribution of some networks .
this scheme maps the problem into a random walk in degree space , and then we choose the paths that are responsible for the dominant contributions .
the method is illustrated on the dynamical versions of the erds - rnyi and watts - strogatz graphs , which were introduced as static models in the original formulation .
we have succeeded in obtaining an analytical form for the dynamics watts - strogatz model , which is asymptotically exact for some regimes . |
* dag realization problem : * given is a finite sequence @xmath0 with @xmath7 does there exist an acyclic digraph ( without parallel arcs ) @xmath8 with the labeled vertex set @xmath2 such that we have indegree @xmath9 and outdegree @xmath10 for all @xmath3 ? + if the answer is `` yes '' , we call sequence @xmath11 _ dag sequence _ and the acyclic digraph @xmath12 ( a so - called `` dag '' ) a _ dag realization_. a relaxation of this problem not demanding the acyclicity of digraph @xmath12 is called _ digraph realization problem_. in this case , we call @xmath12 _ digraph realization _ and @xmath11 _ digraph sequence_. the digraph realization problem can be solved in linear - time using an algorithm by wang and kleitman @xcite . unless explicitly stated , we assume that a sequence does not contain any _ zero tuples _ @xmath13 .
moreover , we will tacitly assume that @xmath14 , as this is obviously a necessary condition for any realization to exist , since the number of ingoing arcs must equal the number of outgoing arcs .
furthermore , we denote tuples @xmath15 with @xmath16 and @xmath17 as _ sink tuples _ , those with @xmath18 and @xmath19 as _ source tuples _ , and the remaining ones with @xmath16 and @xmath19 as _ stream tuples_. we call a sequence only consisting of source and sink tuples , _ source - sink - sequence_. a sequence @xmath20 with @xmath21 source tuples and @xmath22 sink tuples is denoted as _ canonically sorted _ , if and only if the first @xmath21 tuples in this labeling are decreasingly sorted source tuples ( with respect to the @xmath6 ) and the last @xmath22 tuples are increasingly sorted sink tuples ( with respect to the @xmath5 ) . + * hardness and efficiently solvable special cases .
* nichterlein very recently showed that the dag realization problem is np - complete @xcite . on the other hand ,
there are several classes of sequences for which the problem is not hard .
one of these sequences are source - sink - sequences , for which one only has to find a digraph realization .
the latter is already a dag realization , since no vertex has incoming as well as outgoing arcs .
furthermore , sparse sequences with @xmath23 are polynomial - time solvable as we will show below .
we denote such sequences by _ forest sequences_. the main difficulty for the dag realization problem is to find out a `` topological ordering of the sequence '' . in the case where we have one , our problem is nothing else but a directed @xmath24-factor problem on a complete dag .
the labeled vertices of this complete dag are ordered in the given topological order .
this problem can be reduced to a bipartite undirected @xmath24-factor problem which can be solved in polynomial time via a further famous reduction by tutte @xcite to a bipartite perfect matching problem . in a previous paper @xcite
, we proved that a certain ordering of a special class of sequences _opposed sequences_ always leads to a topological ordering of the tuples for at least one dag realization of a given dag sequence . on the other hand , it is not necessary to apply the reduction via tutte if we possess one possible topological ordering of a dag sequence .
the solution is much easier .
next , we describe our approach . + * realization with a prescribed topological order .
* we denote a dag sequence @xmath0 which possesses a dag realization with a topological numbering corresponding to the increasing numbering of its tuples by _ dag sequence for a given topological order _ and analogously the digraph @xmath8 by _ dag realization for a given topological order_. without loss of generality , we may assume that the source tuples come first in the prescribed numbering and are ordered decreasingly with respect to their @xmath6 values .
a realization algorithm works as follows .
consider the first tuple @xmath25 from the prescribed topological order which is not a source tuple .
then there must exist @xmath26 source tuples with a smaller number in the given dag sequence .
reduce the @xmath26 first ( i.e. with largest @xmath6 ) source tuples by one and set the indegree of tuple @xmath25 to @xmath27 that means , we reduce sequence @xmath28 to sequence @xmath29 if we get zero tuples in @xmath30 then we delete them and denote the new sequence for simplicity also by @xmath31 furthermore , we label this sequence with a new numbering starting from one to its length and consider this sorting as the given topological ordering for @xmath32 .
we repeat this process until we get an empty sequence ( corresponding to the realizability of @xmath11 ) or get stuck ( corresponding to the non - realizability of @xmath11 ) .
the correctness of our algorithm is proven in lemma [ th : topologicalrealization ] .
[ th : topologicalrealization ] @xmath11 is a dag sequence for a given topological order @xmath33 @xmath32 is a dag sequence for its corresponding topological order .
* discussion of our main theorem and its corresponding algorithm .
* we do not know how to determine a feasible topological ordering ( i.e. , one corresponding to a realization ) for an arbitrary dag sequence .
however , we are able to restrict the types of possible permutations of the tuples . for that ,
we need the following order relation @xmath34 , introduced in @xcite . given are @xmath35 and @xmath36 we define : @xmath37 note , that a pair @xmath38 equals @xmath39 with respect to the opposed relation if and only if @xmath40 and @xmath41 the opposed relation is reflexive , transitive and antisymmetric and therefore a partial , but not a total order .
our following theorem leads to a recursive algorithm with exponential running time and results in corollary [ korollartopologischerealisation ] which proves the existence of a special type of possible topological sortings provided that sequence @xmath11 is a dag sequence .
[ realisationdagsequenzen ] let @xmath11 be a canonically sorted sequence containing @xmath42 source tuples .
furthermore , we assume that @xmath11 is not a source - sink - sequence .
we define the set @xmath43 @xmath11 is a dag sequence if and only if @xmath44 and there exists an element @xmath45 such that @xmath46 @xmath47 is a dag sequence .
( ) return true sequence @xmath32 may contain zero tuples . if this is the case , we delete them and call the new sequence for simplicity also @xmath32 .
theorem [ realisationdagsequenzen ] ensures the possibility for reducing a dag sequence into a source - sink - sequence .
the latter can be realized by using the algorithm for realizing digraph sequences @xcite .
the whole algorithm is summarized in algorithm [ alg : dag realization ] , where we consider the maximum subset @xmath48 of @xmath49 only containing pairwise disjoint stream tuples .
the bottleneck of this approach is the size of set @xmath50 we give an example for the execution of algorithm [ alg : dag realization ] in the appendix .
our pseudocode does not specify the order in which we process the elements of @xmath48 in line 3 .
several strategies are possible which have a significant influence on the overall performance . the most promising deterministic strategy
( as we will learn in the next sections ) is to use the lexicographic order , starting with the lexicographic maximum element within @xmath48 . in @xcite
we introduced a special class of dag sequences _ opposed sequences _ where we have @xmath51 if sequence @xmath11 is not a source - sink - sequence .
we call a sequence @xmath11 _ opposed sequence _ , if it is possible to sort its stream tuples in such a way , that @xmath52 and @xmath53 is valid for stream tuples with indices @xmath54 and @xmath55 in this case , we have the property @xmath56 for all stream tuples .
at the beginning of the sequence we insert all source tuples such that the @xmath6 build a decreasing sequence and at the end of sequence @xmath11 we put all sink tuples in increasing ordering with respect to the corresponding @xmath57 the notion _ opposed sequence _ describes a sequence , where it is possible to compare all stream tuples among each other and to put them in a `` chain '' .
indeed , this is not always possible because the opposed order is not a total order .
however , for opposed sequences line ( 3 ) to line ( 9 ) in algorithm [ alg : dag realization ] are executed at most once in each recursive call , because we have always @xmath58 overall , we obtain a linear - time algorithm for opposed sequences . however , there are many sequences which are not opposed , but theorem [ realisationdagsequenzen ] still yields a polynomial decision time .
consider for example dag sequence @xmath59 which is not an opposed sequence , because stream tuples @xmath60 and @xmath61 are not comparable with respect to the opposed ordering .
however , we have @xmath62 and so we reduce @xmath11 to @xmath63 , leading to the realizable source - sink - sequence @xmath64 theorem [ realisationdagsequenzen ] leads to further interesting insights
. we can prove the existence of special topological sortings .
[ korollartopologischerealisation ] for every dag sequence @xmath11 , there exists a dag realization @xmath8 with a topological ordering @xmath65 of all @xmath66 vertices corresponding to stream tuples , such that we can not find @xmath67 for @xmath68 we call a topological ordering of a dag sequence obeying the conditions in corollary [ korollartopologischerealisation ] an _ opposed topological sorting_. at the beginning of our work ( when the complexity of the dag realization problem was still open ) , we conjectured that the choice of the lexicographical largest tuple from @xmath48 in line ( 3 ) would solve our problem in polynomial time .
we call this approach _ lexmax strategy _ and a dag sequence which is realizable with this strategy _ lexmax sequence _ , otherwise we call it _
non - lexmax sequence_. hence , we conjectured the following .
[ co : lexmaxconjecture ] each dag sequence is a lexmax sequence .
we soon disproved our own conjecture by a counter - example ( example [ example ] , described in the following section and in appendix [ app : examples ] ) . in systematic experiments we found out that a large fraction of sequences can be solved by this strategy in polynomial time .
we tell this story in the next section [ storyexperiments ] .
moreover , we use the structural insights from our main theorem to develop a randomized algorithm which performs well in practice ( section [ randomisiertealgorithmen ] ) .
proofs and further supporting material can be found in the appendix and in @xcite .
* why we became curious .
* to see whether our lexmax conjecture [ co : lexmaxconjecture ] might be true , we generated a set of dag sequences , called _ randomly generated sequences _ in the sequel , by the following principle : starting with a complete acyclic digraph , delete @xmath69 of its arcs uniformly at random .
we take the degree sequence from the resulting graph .
note that we only sample uniformly with respect to random dags but not uniformly degree sequences since degree sequences have different numbers of corresponding dag realizations . in a first experiment we created with the described process one million dag sequences with 20 tuples each , and @xmath70 .
likewise , we built up another million dag sequences with 25 tuples and @xmath71 .
the fact that the lexmax strategy realized all these test instances without a single failure was quite encouraging .
the lexmax conjecture [ co : lexmaxconjecture ] seemed to be true , only a correctness proof was missing .
but quite soon , in an attempt to prove the conjecture , we artificially constructed a first counter - example , a dag sequence which is definitely no lexmax sequence , as can easily be verified : even worse : we also found an example ( example [ ex : strategiesichtweise ] ) showing that no fixed strategy which chooses an element from @xmath48 in algorithm [ alg : dag realization ] and does not consider the corresponding set of sinks , will fail in general .
[ ex : strategiesichtweise ] we consider the two sequences @xmath73 and @xmath74 only differing in their sink tuples .
sequence @xmath75 can only be realized by the lexmax strategy , while several strategies but not the lexmax strategy work for @xmath76 .
thus , there is no strategy which can be applied in both cases .
these observations give rise to several immediate questions : why did we construct by our sampling method ( for @xmath77 and @xmath78 ) only dag sequences which are lexmax sequences ?
how many dag sequences are not lexmax sequences ?
therefore , we started with systematic experiments . for small instances with @xmath79 tuples we generated systematically the set of all dag sequences with all possible
@xmath80 , see for an example the case @xmath81 in figure [ fig : percentagelexmaxsequences ] and appendix [ app : material ] .
more precisely , we considered only _ non - trivial sequences _ , i.e. we eliminated all source - sink sequences and all sequences with only one stream tuple .
we denote this set by _ systematically generated sequences_. note that the number of sequences grows so fast in @xmath82 that a systematic construction of all sequences with a larger size is impossible .
we observed the following : 1 .
the fraction of lexmax sequences among the systematically generated sequences is quite high .
for all @xmath83 it is above @xmath84 , see figure [ fig : percentagelexmaxsequences ] ( blue squares ) .
the fraction of lexmax sequences strongly depends on @xmath83 .
it is largest for sparse and dense dags .
lexmax sequences are overrepresented among one million randomly generated sequences ( for each @xmath83 ) , we observe more than @xmath85 for all densities of dags , see figure [ fig : percentagelexmaxsequences ] ( red triangles ) .
this leads to the following questions : given a sequence for which we seek a dag realization .
how should we proceed in practice ? as we have seen , the huge majority of dag sequences are lexmax sequences .
is it possible to find characteristic properties for lexmax sequences or non - lexmax sequences , respectively ?
+ * distance to opposed sequences .
* let us exploit our characterization that opposed sequences are efficiently solvable .
we propose the _ distance to opposed _
@xmath88 for each dag sequence @xmath89 consider for that the topological order of a dag realization @xmath12 given by algorithm [ alg : dag realization ] , if in line ( 3 ) elements are chosen in decreasing lexicographical order .
this ordering corresponds to exactly one path of the recursion tree .
thus , we obtain one unique dag realization @xmath12 for @xmath11 , if existing .
now , we renumber dag sequence @xmath11 such that it follows the topological order induced by the execution by this algorithm , i.e. by the sequence of choices of elements from @xmath50 then the distance to opposed is defined as the number of pairwise incomparable stream tuples with respect to this order , more precisely , @xmath90 [ [
question-1-do - randomly - generated - sequences - possess - a - preference - to - a - small - distance - to - opposed - in - comparison - with - systematically - generated - sequences ] ] question 1 : do randomly generated sequences possess a preference to a `` small '' distance to opposed in comparison with systematically generated sequences ? + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in figure [ fig : systematicdifferencetoopposed ] ( left ) , we show the distribution of systematically generated sequences ( in % ) with their distance to opposed , depending on @xmath91 we compare this scenario with the same setting for randomly generated sequences , shown in figure [ fig : systematicdifferencetoopposed ] ( right ) . percentage of systematically generated sequences @xmath11 ( left ) and randomized generated sequences ( right ) with their difference @xmath88 to opposed for @xmath81 tuples and @xmath87.,scaledwidth=115.0% ] percentage of systematically generated sequences @xmath11 ( left ) and randomized generated sequences ( right ) with their difference @xmath88 to opposed for @xmath81 tuples and @xmath87.,scaledwidth=115.0% ] _ observations : _ systematically generated sequences have a slightly larger range of the `` distance to opposed '' than randomly generated sequences .
moreover , when we generate dag sequences systematically , we obtain a significantly larger fraction of instances with a larger distance to opposed than for randomly generated sequences , and this phenomenon can be observed for all @xmath83 .
[ [ question-2-do - non - lexmax - sequences - possess - a - preference - for - large - opposed - distances ] ] question 2 : do non - lexmax sequences possess a preference for large opposed distances ? + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + since opposed sequences are easily solvable @xcite , we conjecture that sequences with a small distance to opposed might be easier solvable by the lexmax strategy than those with a large distance to opposed .
if this conjecture were true , it would give us together with our findings from question 1 one possible explanation for the observation that the randomly generated sequences have a larger fraction of efficiently solvable sequences by the lexmax strategy .
_ observations : _ a separate analysis of non - lexmax sequences ( that is , the subset of unsolved instances by the lexmax strategy ) , displayed in figure [ fig : opposeddistancenonlexmax ] , gives a clear picture : yes ! for systematically generated sequences with @xmath81 , we observe in particular for instances with a middle density that the fraction of non - lexmax sequences becomes maximal for a relatively large distance to opposed .
[ [ question-3-can - we - solve - real - world - instances - by - the - lexmax - strategy ] ] question 3 : can we solve real - world instances by the lexmax strategy ? + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + 1 .
ordered binary decision diagrams ( obdds ) : in such networks the outdegree is two , that is constant .
this immediately implies that the corresponding sequences are opposed sequences , and hence can provably be solved by the lexmax strategy .
food webs : such networks are almost hierarchical and therefore have a strong tendency to be acyclic ( `` larger animals eat smaller animals '' ) . in our experiments we analyzed food webs from the pajek network library @xcite .
train timetable network : we use timetable data of german railways from 2011 and form a time - expanded network .
its vertices correspond to departure and arrival events of trains , a departure vertex is connected by an arc with the arrival event corresponding to the very next train stop .
moreover , arrival and departure events at the same station are connected whenever a transfer between trains is possible or if the two events correspond to the very same train .
flight timetable network : we use the european flight schedule of 2010 and form a time - expanded network as in c ) . the characteristics of our real - world networks b ) - d ) are summarized in table [ tab : real - world ] . the _ dag density _
@xmath92 of a network is defined as @xmath93 . to compare the distance to opposed for instances of different sizes , we normalize this value by the theoretical maximum @xmath94 , where @xmath95 denotes the number of stream tuples , and
so obtain a _
normalized distance to opposed_. without any exception , all real - world instances have been realized by the lexmax strategy . | we study the following fundamental realization problem of directed acyclic graphs ( dags ) . given a sequence @xmath0 with @xmath1 , does there exist a dag ( no parallel arcs allowed ) with labeled vertex set @xmath2 such that for all @xmath3 indegree and outdegree of @xmath4 match exactly the given numbers @xmath5 and @xmath6 , respectively ? recently this decision problem has been shown to be np - complete by nichterlein @xcite .
however , we can show that several important classes of sequences are efficiently solvable . in previous work @xcite
, we have proved that yes - instances always have a special kind of topological order which allows us to reduce the number of possible topological orderings in most cases drastically .
this leads to an exact exponential - time algorithm which significantly improves upon a straightforward approach .
moreover , a combination of this exponential - time algorithm with a special strategy gives a linear - time algorithm .
interestingly , in systematic experiments we observed that we could solve a huge majority of all instances by the linear - time heuristic .
this motivates us to develop characteristics like dag density and `` distance to provably easy sequences '' which can give us an indicator how easy or difficult a given sequence can be realized .
furthermore , we propose a randomized algorithm which exploits our structural insight on topological sortings and uses a number of reduction rules .
we compare this algorithm with other straightforward randomized algorithms in extensive experiments .
we observe that it clearly outperforms all other variants and behaves surprisingly well for almost all instances .
another striking observation is that our simple linear - time algorithm solves a set of real - world instances from different domains , namely ordered binary decision diagrams ( obdds ) , train and flight schedules , as well as instances derived from food - web networks without any exception . |
s0 galaxies are enigmatic objects whose formation and evolution is still not well understood .
originally , these galaxies were postulated to exist as a transition class between the elliptical and spiral sequence @xcite .
both elliptical and s0 galaxies lack spiral arms or major dust features .
however , early work on the ellipticity distributions of galaxies showed that s0 galaxies were disk - dominated systems with ellipticity distributions that differed from ellipticals , being more similar to spiral galaxies in their intrinsic shapes ( _ e.g. _ , * ? ? ? * ; * ? ? ? * ) . the early work on the morphology - density relation by @xcite emphasized that s0s and ellipticals both occur with higher frequency in higher density environments , while the pioneering studies of @xcite and @xcite began to provide hints about how the early - type population of elliptical and s0 galaxies might evolve out to redshifts around 0.5 and earlier .
however , it was not until _ hst _ allowed comprehensive high - resolution imaging of distant clusters of galaxies that @xcite , for example , and others began to show directly that the s0 fraction was changing at high redshift .
this early work , along with more recent studies ( e.g. , * ? ? ? * ; * ? ? ? * ) found smaller s0 galaxy fractions in clusters of galaxies at higher redshifts , @xmath13 .
the implication is that the s0 galaxy population forms with different time - scales and later than the elliptical population .
@xcite , @xcite , @xcite , @xcite , @xcite , and , most recently , @xcite all find that the majority of evolution occurs since @xmath14 , i.e. , in the past 4 gyrs , and above those redshifts there is little or no evolution in the early - type galaxy fraction in clusters of galaxies out to @xmath9 ( see * ? ? ?
* for a different point of view ) .
since s0 and elliptical galaxies have different ellipticity distributions and bulge - to - disk distributions , two simple ( and related ) predictions that can be drawn from the observed changes in s0 fraction with redshift is that both the mean bulge - to - disk ratio of early - type cluster galaxies and the ellipticity distribution should have changed over relatively recent epochs ( since @xmath15 ) .
the evidence for evolution in the s0 population of clusters rests primarily on morphological classifications of galaxies .
the separation of the early - type galaxy population into s0 and elliptical galaxies has long been recognized as being a challenging task ( see , for example , @xcite for a discussion of the systematic errors in classification and how misclassification mimics evolution . )
s0 galaxies are defined as multicomponent disk galaxies , while ellipticals are defined as single component systems . however
, a number of quantitative studies found that a substantial fraction of ellipticals have `` disky '' isophotes ( see * ? ? ? * for a summary ) in contrast with this definition .
the analysis of @xcite exemplifies the challenge of establishing the relative contributions of elliptical and s0 galaxies .
they showed , for a @xmath10 galaxy , that detecting a disk component in a spheroidal galaxy is increasingly difficult as the disk becomes more face - on in projection .
@xcite found that a disk containing 20% of the total light of a galaxy is impossible to detect over half of the range of @xmath16 where @xmath17 is the inclination angle .
specifically , they note `` since the cos @xmath17 axis can be interpreted as a probability axis , this implies that 50% of all disks with @xmath18 can not be detected by photometric means . ''
these , however , results for a single galaxy . for an ensemble of galaxies , the bulge - to - disk ratio distribution can be constrained by the intrinsic ellipticity distribution , with the average observed ellipticity being directly related to the intrinsic ellipticity @xcite . therefore , examining the distribution of galaxy ellipticities provides a direct measure of the evolution in the distribution of the bulge - to - disk ratios of that galaxy population .
@xcite investigated the ellipticity distribution of elliptical and s0 galaxies in the coma cluster , and suggested that elliptical and s0 galaxies were not distinct classes but were part of a continuum of objects of varying bulge - to - disk ratios .
this result was given additional support by a recent study by @xcite who found that 69% of elliptical galaxies have multiple kinematic components , generally disk - like components , while 92% of s0 galaxies have disk - like components .
these results give emphasis to the view that elliptical and s0 galaxies form a continuous distribution of disk fraction as opposed to two distinct classes .
in particular , @xcite constructed a model of the ellipticity distribution of elliptical and s0 galaxies using a continuum of bulge and disk components viewed from a variety of angles .
@xcite found a deficit of round s0 galaxies , in contrast with what was expected from their model , suggesting that some face - on s0 galaxies had been classified as ellipticals . @xcite and
@xcite also found a lack of round s0 galaxies in three @xmath19 clusters of galaxies when compared to what was expected for a disk population viewed at a variety of angles .
these three studies illustrate a potential bias in the visual classification of galaxies , namely that nearly face - on s0 galaxies galaxies may be incorrectly classified as ellipticals .
a number of studies suggest that quantitative measurements do not suffer the same orientation bias as visual classifications @xcite , but robust , bias - free galaxy classification , either visual or quantitative , still remains an elusive goal .
the importance of the work showing an apparent evolution in the s0 fraction in clusters and the knowledge of potential classification biases led @xcite and @xcite to investigate the ellipticity distributions of the s0 and elliptical galaxies in their samples . in general
, the authors found that the ellipticity distributions of s0 and elliptical galaxies show no evolution over the broad redshift ranges in their samples ( * ? ? ?
* compared clusters from @xmath20 to @xmath21 ) . also , the ellipticity distributions of elliptical and s0 galaxies differ from each other , providing evidence for the existence of two distinct classes of galaxies . however , these previous studies do not use a consistent measure of the ellipticity as compared with the @xmath10 efforts such as @xcite or @xcite , making a comparison between these @xmath22 samples observed with _ hst _ and @xmath5 samples observed from the ground difficult .
traditionally , ellipticities for galaxies at @xmath10 were measured by visual estimates @xcite or by fitting models to the elliptical isophotes @xcite . at higher redshifts , @xmath22 , the ellipticities in @xcite or @xcite are determined by the second - order flux - weighted moments of a detection isophote .
both the data and the techniques have matured so that we can now evaluate the ellipticity distribution , as a function of redshift , quantitatively , and even more importantly , in a consistent way with minimal systematic error .
we show that some approaches used previously that could not correct for the point - spread function ( psf ) are probably subject to significant systematic error .
essentially , the `` blurring '' effect of the psf will lead to galaxies being measured as rounder than they actually are .
we will use a single consistent and robust approach for measuring the ellipticities at all redshifts .
the ellipticities will be measured by fitting models convolved with the psf to galaxy surface brightness profiles .
this will eliminate some of the previously - reported uncertainties found when comparing ground - based imaging data taken under different seeing conditions ( see * ? ? ? * for some discussion ) .
in addition , this approach essentially eliminates the systematic error associated with the psf `` blurring '' .
our ellipticity measurements will provide an assessment of the evolution in the distribution of the overall bulge - to - disk ratio of early - type cluster galaxies from the present day to redshifts @xmath19 .
we will discuss how we compiled our samples of early - type cluster galaxies with morphological classifications in [ sample ] , and then how we measured their total magnitudes , colors , and ellipticities in [ data ] .
one of the advantages of our approach of using an automated measurement technique is that we can simulate the measurement process .
we discuss this in [ ellmeas ] and in appendix a1 . from our measurements ,
we find no evolution in the distribution of ellipticities of cluster early - type galaxies , which we show in [ elldist ] . in [ disc ]
, we discuss the implications of this result and contrast with previous measurements of the evolution in the overall distribution of the bulge - to - disk ratio of cluster early - type galaxy population .
we follow this with a summary and some discussion of the broader implications in [ conclusion ] . throughout this paper , we assume @xmath23 , @xmath24 and @xmath25 .
all @xmath26 magnitudes we list use the vega zeropoint ( @xmath27 . ) other magnitudes where given are ab mags .
to carry out our study of the ellipticities of early - type galaxies from @xmath9 to the current epoch , we have assembled a sample of morphologically - selected , cluster early - type galaxies ranging in redshifts from @xmath28 to @xmath29 .
the clusters used are tabulated below , as is the source of the morphologies for the galaxies samples ( [ morph ] ) .
the early - type galaxies are chosen to lie within the projected @xmath3 , which we use to define cluster membership .
the measurement for each galaxy of its color , magnitude , and ellipticity , is described in [ data ] .
lrrrr aco 119 & 0.0440 & 744 & 1.13 & 41 + aco 168 & 0.0452 & 524 & 0.80 & 54 + aco 194 & 0.0178 & 435 & 0.67 & 9 + aco 957 & 0.0440 & 691 & 1.05 & 32 + aco 1139 & 0.0383 & 436 & 0.67 & 23 + aco 1142 & 0.0353 & 417 & 0.64 & 17 + aco 1656 & 0.023 & 1008 & 1.55 & 203 + aco 1983 & 0.0441 & 433 & 0.66 & 18 + aco 2040 & 0.0456 & 602 & 0.92 & 50 + aco 2063 & 0.0337 & 521 & 0.80 & 50 + aco 2151 & 0.0371 & 786 & 1.20 & 58 + [ lowz_num ] our @xmath2 sample is selected from abell clusters in the sloan digital sky survey fifth data release ( * ? ? ? * sdss dr5 ) .
we summarize this sample in table [ lowz_num ] , where we list the abell number @xcite and the redshift of the cluster from @xcite .
the last column lists the number of redshift - selected members with early - type classification , which lie within the projected @xmath3 and are on the @xmath30 red sequence in the sdss dr5 imaging . to determine @xmath3
, we will use the formula give in @xcite , or @xmath31 where @xmath32 is the one - dimensional velocity dispersion and @xmath33 is the hubble constant at the redshift of observation .
@xcite uses the relation from @xcite , which is functionally the same , so effectively our selection radii are similar to @xcite .
we selected @xmath34 , instead of @xmath3 , as a galaxy at @xmath3 from the cluster center will be , on average , projected to appear at the distance @xmath35 ( see * ? ? ?
* for example ) .
for the high redshift clusters , we required _ hubble space telescope _
( _ hst _ ) imaging with the advanced camera for surveys ( acs ) or the wide field planetary camera 2 ( wfpc2 ) in multiple filters .
we also required that the morphological classifications were done in a way consistent with the original @xcite and @xcite work .
for the three clusters where @xmath36 was larger than the field of view over which we had imaging data ( generally the clusters in the sample of * ? ? ?
* ) , we simply used all available galaxies .
we mark those clusters in table [ summary ] with the superscript b. below we will detail the sources for the morphological classifications and how we measured the total magnitudes , colors , and ellipticities .
we tabulate the @xmath37 clusters for which we have early - type galaxy samples in table [ summary ] . in that table
, we list the dispersions , inferred radii , and the final sample sizes . for cl
j1226 + 33 we have compiled a new catalog of members which we will discuss in appendix [ ctw ] . for rx j0849 + 4452 , we compute a new dispersion .
we use both redshifts from previous work @xcite and unpublished ones we have recently collected ( which we plan to publish in a future paper . ) there are a total of 18 galaxies in rx j0849 + 4452 with redshifts within @xmath3 .
the biweight center of the distribution is @xmath38 and the dispersion is 798 @xmath8 208 km s@xmath39 .
the errors for both the redshift and the dispersion are estimated by bootstrapping the redshift distribution .
for the clusters rx j0152 - 13 , ms 1054 - 03 , ms 2053 - 04 , and cl 1358 + 62 , we used the redshift catalogs from @xcite .
the faint magnitude limit we adopted corresponds to the completeness limits for our high redshift samples @xcite . for the remaining clusters
, we included all galaxies that met our magnitude , color , morphology and radial requirements , rejecting only those galaxies with redshifts outside of the clusters .
this will mean that some cluster samples will be contaminated , i.e. , some red - sequence early - type galaxies will not be members .
the level of contamination is discussed below .
our @xmath1 sample consists of 487 early - type galaxies while our @xmath2 sample has 210 galaxies .
many of the cluster red - sequence selections we use are tabulated in other studies .
for those clusters studied by the acs instrument definition team , these red - sequence relations are summarized in @xcite . for the remaining clusters ,
we derive the early - type red - sequence from the data . whenever possible , we use existing redshift catalogs to determine the red - sequence of spectroscopic determined members .
we then accept all galaxies that lie within the 2@xmath40 of that sequence , rejecting those galaxies that are known not to be members .
we know that there will be some contamination by field galaxies in our catalogs .
to measure the level of contamination , we used our redshift catalogs for ms 1054 - 03 and rx j0152 - 13 along with the catalogs of ediscs from @xcite .
rx j0152 - 13 has a known group in the foreground with colors very similar to that of the cluster ( see * ? ? ?
* for a discussion ) . nonetheless , when we consider all early - type galaxies that lie on the red - sequence for ms 1054 - 03 and rx j0152 - 13 , regardless of redshift , we find a 3 @xmath8 1% contamination rate of nonmembers
when we examine the less extremely rich sample of the ediscs , we find a contamination fraction of 10 @xmath8 3% , which is likely more representative of the typical clusters in our sample .
the small size of this contamination means we need not compute a statistical background correction for the cluster red - sequence .
lrrrrrrr cl 1358 + 62 & 0.328 & 13 59 50.6 & + 62 59 05 & @xmath41 & 1027@xmath42 & 1.4 & 67 + cl 0024 + 16 & 0.395 & 00 26 35.7 & + 17 09 46 &
@xmath43 & 650@xmath44 & 0.83 & 26 + cl 0016 + 16 & 0.541 & 00 18 33.5 & + 16 26 14 & @xmath45 & 1234@xmath46 & 1.46 & 42 + ecl j1232 - 12 & 0.541 & 12 32 30.3 & -12 50 36 & @xmath47 @xmath48 , @xmath49 & @xmath50 & 1.28 & 21 + cl 0054 - 27 & 0.56 & 00 56 56.9 & -27 40 30 & @xmath41 & 1180 & 1.38 & 19 + ms 2053 - 04 & 0.587 & 20 56 21.3 & -04 37 51 & @xmath41 & 865@xmath51 & 1.00 & 31 + ecl j1054 - 11 & 0.697 & 10 54 24.4 & -11 46 19 & @xmath47 @xmath48 , @xmath49 & @xmath52 & 0.64 & 5 + ecl j1040 - 11 & 0.704 & 10 40 40.3 & -11 56 04 & @xmath47 @xmath48 , @xmath49 & @xmath53 & 0.45 & 2 + ecl j1054 - 12 & 0.750 & 10 54 43.5 & -12 45 52 & @xmath47 @xmath48 , @xmath49 & @xmath54 & 0.53 & 13 + ecl j1216 - 12 & 0.794 & 12 16 45.3 & -12 01 18 & @xmath47 @xmath48 , @xmath49 & @xmath55 & 1.04 & 31 + ms 1054 - 03 & 0.831 & 10 57 00.0 & -03 37 36 & @xmath56 & 1156@xmath57 & 1.16 & 59 + rx j0152 - 13 & 0.834 & 01 52 43.8 & -13 57 19 & @xmath58 & 919@xmath59 & 0.92 & 36 + cl j1226 + 33 & 0.890 & 12 26 58.2 & + 33 32 49 & @xmath41 & 1143@xmath60 & 1.11 & 46 + cl 1604 + 4304 & 0.897 & 16 04 24.0 & + 43 04 38 & @xmath41 & 962@xmath61 & 0.93 & 26 + cl 1604 + 4321 & 0.924 & 16 04 33.6 & + 43 21 04 & @xmath41 & 640@xmath51 & 0.61 & 15 + rx j0910 + 5422 & 1.106 & 09 10 44.9 & 54 22 08.9 & @xmath45 & 675@xmath62 & 0.58 & 12 + rx j1252 - 2927 & 1.237 & 12 52 48 & -29 27 00 & @xmath45 & 747@xmath63 & 0.60 & 16 + rx j0849 + 4452 & 1.260 & 08 48 58.66 & 44 51 57.0 & @xmath45 & 798@xmath64 & 0.63 & 8 + [ summary ] galaxy morphologies were obtained from the literature for all the galaxies in our sample .
we use the morphologies from @xcite for the @xmath2 cluster members .
we removed all late - type galaxies from that low redshift sample , i.e. , all those classified as spiral , irregular or unknown classifications .
we considered galaxies classified as s0/a as s0 , and galaxies classified as sa/0 as spirals .
we list in table [ lowz_num ] the number of early - type galaxies e through s0/a in the last column .
the numbers are smaller than the number available in @xcite because not all of the galaxies in each cluster are imaged in the sdss dr5 . for our higher redshift clusters
, we used the classifications of @xcite as tabulated in @xcite , @xcite , and @xcite .
each cluster was imaged with either acs or wfpc2 with morphological classifications done in a manner consistent with the previous work of @xcite .
the clusters ms 1054 - 03 , rx j0152 - 13 , cl j1226 + 33 , cl 1604 + 4304 , cl 1604 + 4321 , rx j0910 + 5422 , rx j1252 - 2927 , and rx j0849 + 4452 all come from @xcite .
the morphs survey clusters cl 0016 + 16 , cl 0054 - 27 , cl 0024 + 16 are part of @xcite sample .
five clusters from the ediscs survey
ecl j1040 - 11 , ecl j1054 - 11 ,
ecl j1054 - 12 , ecl j1216 - 12 , and ecl j1232 - 12 have morphologies from @xcite .
two additional , cl 1358 + 62 and ms 2053 - 04 , have morphologies in @xcite and @xcite , respectively .
our sample of early - type galaxies is based on a range in luminosity , corrected for the observed passive evolution to z @xmath65 1 .
the derivation of the total magnitudes and colors used for our sample selection is discussed below . taking that luminosity - selected sample
, we then discuss the derivation of their ellipticities .
these constitute the key observable for this work , and the implications of those ellipticity measurements are discussed in the remainder of the paper .
our final selection will consist of morphologically - selected early - type galaxies on the red - sequence within a well - defined magnitude range ( @xmath4 at @xmath66 ) lying within the cluster core that is defined by @xmath3 .
these magnitude limits correspond to @xmath67 and to galaxies with stellar masses roughly between @xmath68 , assuming a `` diet '' salpeter initial mass function ( imf ) ( see * ? ? ?
* for a discussion of the imf and the procedure we use to estimate the stellar masses of galaxies ) .
our samples were selected based on the rest - frame @xmath26 magnitude . to estimate these magnitudes
, we used the total magnitude in the passband closest to the rest - frame @xmath26 .
we also needed a color to correct the apparent magnitude in the observed passband to a rest - frame @xmath26 magnitude ; how this was done is outlined below .
the srsic model fits that we used to measure the ellipticities also were used to determine the total magnitudes and the color apertures .
the total magnitude is the normalization of the srsic model fit . for the color aperture
, we used the circularized half - light radius , @xmath69 , where @xmath70 is the ratio of the minor to major axis , or @xmath71 , and @xmath72 is the half - light radius along the major axis of the best - fitting elliptical model as determined by galfit .
these are the same apertures used in @xcite ( see that paper for more detail ) .
we adjust the magnitudes of the galaxies by @xmath73 as measured for early - type galaxies using the fundamental plane @xcite .
this compensates for the mean passive evolution of the old stellar population .
our magnitude range covers @xmath74 to @xmath75 using the @xmath76 from @xcite after converting the @xmath77 used by @xcite to the @xmath26 of the johnson - morgan system @xcite that we use in this paper .
we trim our sample at @xmath78 , as our high redshift samples become incomplete fainter than that magnitude .
the brighter magnitude limit is set to be @xmath79 to exclude the most luminous galaxies ( whose formation and evolution may differ , as is explained later in [ ellm ] ) .
thus our adopted magnitude range is @xmath80 . using our simulations of real galaxies that we discuss in more detail in appendix [ simreal ]
, we estimate the typical error and offset for these total magnitude measurements . we find that total magnitudes as measured with acs have an error of @xmath81 mag while magnitudes measured with wfpc2 have @xmath82 mag .
there is a small offset at most redshifts , such that we measure a magnitude brighter than the actual magnitude of the simulated galaxy .
for the clusters at @xmath83 , this is only 0.02 mag , while it increases to 0.06 mag at @xmath84 and 0.10 mag at @xmath85 .
this offset is the same for galaxies regardless of size , morphological type or the value of @xmath86 from the srsic model fit .
we do not apply this offset to the measured magnitudes . to ensure a reliable measure of the color
, we applied the clean algorithm @xcite to the original images in all passbands . using clean mitigates
the effects of the different - sized psf in different passbands , see @xcite for examples involving acs .
the final `` cleaned ''
images were used for measuring the galaxy colors within the @xmath87 given above .
we transformed the observed magnitudes into redshifted magnitudes using the same process as @xcite , @xcite , @xcite and @xcite .
we calculated the magnitudes of templates in the rest - frame filters .
we then redshifted the templates , and computed the magnitudes in the observed filters .
for the templates , we used exponentially decaying star - formation rate models from ( * ? ? ? * bc03 ) ; the same models were used in @xcite .
these models had exponential time - scales of 0.1- 5 gyr , covering a range of ages from 0.5 gyr to 12 gyr and three metal abundances , 2.5 solar , 1.0 solar and 0.4 solar . for the rest - frame filters
, we used the @xmath26 and @xmath47 curves from @xcite , specifically the b3 curve for the @xmath88 as tabulated by bc03 .
we use the same templates and procedure for all of the clusters in our sample .
we measured the ellipticities for our galaxies using the results from galfit @xcite .
galfit fits an elliptical model to the surface brightness profile .
effectively , the ellipticity measurements we use are ellipticities at the half - light radius .
the model is convolved with a psf before it is compared with the data .
the advantages of the approach we have used is that the galfit fit procedure minimizes the effect of the smoothing from the psf .
psf `` blurring '' will make galaxies appear rounder than they actually are , unless the galaxies being fitted have sizes much greater than the psf . in appendix [ ellrob ] , we discuss in more detail the robustness of the ellipticity measurements .
this procedure is performed for galaxies at all redshifts , but requires an estimate of the psf . for the clusters of galaxies at @xmath2 , our imaging data came sdss
we use the software tools provided by the sdss to extract the psf appropriate for each galaxy . for the higher redshift galaxies , the pipeline processing system we use for acs ( apsis acs pipeline science investigation software ) provided us with suitable psf @xcite .
we used empirical psf models constructed from multiple acs observations of 47 tuc . for the wfpc2 data ,
the psf still has a strong positional dependence and is significantly under sampled .
for these data , we estimated the psf using the tinytim software package @xcite for each galaxy we fit .
we also used the option in galfit to convolve the model of the galaxy with an over - sampled psf . before comparing the psf convolved model with the data , galfit rebinned the model to the wfpc pixel scale and smoothed the model with the charge diffusion kernel .
we constructed a set of simulations using real galaxy images and placed those images in the data of the high redshift clusters in our sample .
this is discussed in detail in appendix [ simreal ] .
to summarize , we made multiple measurements of each simulated individual galaxy to assess both the impact of noise and any systematic effects , and each galaxy was simulated over a range in magnitudes .
we found that the scatter in the ellipticity measurements of the images of real galaxies was @xmath89 at magnitudes typical of those in our samples , and only increased ( to @xmath90 ) at or below the magnitude limit of our samples .
these estimates of the uncertainty are in good agreement with the errors estimated by the model fitting process used to measure the ellipticity .
we found no large systematic trends in the ellipticity measurements .
this can be see in figure [ cl_e ] where we show the recovered ellipticity as a function of the input ellipticity for three redshift bins .
there was only a very small overall shift of @xmath91 , or 3% for the typical galaxy , such that the typical @xmath1 galaxy is measured to be slightly rounder than it would appear in the @xmath2 sample .
we found the shift was the same , regardless of the redshift of the galaxy .
no systematic changes were seen as a function of ellipticity .
for the rest of the paper , we ignore the small systematic shift of @xmath92 ( noted above and in figure 1 ) , and simply quote the observed values .
we also tested the robustness of the ellipticity measurements by using incorrect psfs .
we found that this error could result in a systematic offset of @xmath93 to @xmath94 depending on how bad the mismatch between the psfs were .
the small size of our statistical errors means that we are sensitive to systematic errors of this order , and we will discuss the implications of this in later sections .
magnitude or stellar mass for the @xmath2 clusters , as described in the text in [ data ] .
the dashed lines show magnitude limits we will use for this paper , @xmath4 . below @xmath95 our higher redshift
cluster samples become incomplete , while above brighter than @xmath96 , the ellipticity distribution of the early - type galaxy population changes , becoming rounder , possibly the result of the somewhat different evolutionary history of the most massive galaxies .
the galaxies within our selection limits are filled circles , while all of the remaining data are shown as open circles .
the median ellipticity of the galaxies in our sample is shown by a solid line . the mass - selected sample and
the luminosity - selected samples yield a very similar selection with similar median ellipticity , suggesting that mass - dependent effects are likely to be small . for the rest of the paper
, we will use the @xmath26 magnitude limits in the left panel to select galaxies at higher redshifts.,width=326 ] in this section , we define our selection of @xmath2 galaxies , and discuss how we will characterize the ellipticity distributions in order to compare them with the @xmath1 cluster galaxy sample . in figure [ mass_e ]
, we plot the ellipticity as a function of @xmath97 , and also by stellar mass , for all galaxies in the @xmath2 sample that are within @xmath98 , and classified as an early - type galaxy .
we show the median ellipticity of the whole population with a solid line .
the stellar masses are derived using the prescription from @xcite , which uses the rest - frame colors to estimate the mass - to - light ratio of the stellar population , assuming a `` diet '' salpeter imf .
the `` diet '' salpeter imf is a salpeter imf with truncation at very low masses , resulting in a mass - to - light ratio of the stellar population 0.15 dex smaller than a salpeter imf .
these stellar mass estimates agree well with the mass estimates from the fundamental plane ( see * ? ? ?
* ) . at the highest masses or brightest magnitudes ,
the population becomes rounder , a result seen in other work ( see * ? ? ?
* ; * ? ? ?
* for example ) , likely a result of a different morphological mix among the most luminous galaxies which may have a somewhat different formation and evolution history .
we will exclude these most luminous galaxies , @xmath79 , from our sample .
we illustrate our sample range @xmath4 with dashed lines in figure [ mass_e ] . at low redshifts , we know that elliptical and s0 galaxies have different ellipticity distributions . in figure [ type_e ] , we show that we reproduce those different distributions with our @xmath2 cluster galaxy sample .
the elliptical population , shown in red in figure [ type_e ] , is much rounder than the s0 population , shown in blue .
the whole of the population is shown in green .
because s0 galaxies dominate the population , the green line appears closer to the s0 distribution than the elliptical distribution .
the median ellipticities are different , with the median ellipticity of s0 galaxies @xmath99 while the median ellipticity of elliptical galaxies is @xmath100 . in figure [ type_e ] ,
there is a hint of a deficit of round s0 galaxies , a result found by @xcite .
we fit to the distribution of s0 ellipticities with the disk galaxy model used in @xcite .
this disk galaxy distribution is that of an oblate spheroid with the minor to major axis ratios drawn from a gaussian distribution with mean @xmath101 and a dispersion of @xmath102 .
we find that the values for the gaussian that best describes the data are similar to that of @xcite .
our values of @xmath103 and @xmath104 are consistent with @xmath105 and @xmath106 from @xcite a difference @xmath107 . at small ellipticity values ,
the ellipticity distribution of visually classified s0 galaxies fall consistently below the expectation of the model , which indicates a possible lack of round s0 galaxies .
we show the elliptical and s0 population separately ( red - ellipticals , blue - s0 galaxies , green - all early - type galaxies ) .
the two sets of galaxies show different ellipticity distributions in our data , as expected from the results of previous work . the elliptical galaxy population shows a peaked distribution , with a median ellipticity of @xmath100 .
in contrast , the s0 galaxies have a broader ellipticity distribution , with @xmath108 , consistent with a more disk dominated population .
we show , with a black line , the best - fitting disk population drawn from a gaussian distribution which has a mean thickness of @xmath109 giving a mean ellipticity of @xmath110 , with a standard deviation of @xmath111 .
if the s0 fraction decreases with redshift , the overall ellipticity distribution of the cluster population should also evolve as there will be fewer galaxies with large ellipticities ..
,width=326 ] coma ( a1656 ) is a dominant contributor to the early - type sample at @xmath2 , with about 37% of the total .
we investigated the impact of the removal of coma from the sample .
as can be seen in table [ coma ] doing so makes surprisingly little difference to the ellipticity measures of the early - types or of the e and s0s separately .
the ellipticity is consistent within the uncertainties .
the s0 fraction of the @xmath2 sample is @xmath112% for the whole sample and @xmath113% without coma .
lrrr all @xmath2 clusters & 0.29 @xmath8 0.02 & 0.38 @xmath8 0.02 & 0.18 @xmath8 0.01 + without coma & 0.29 @xmath8 0.02 & 0.34 @xmath8 0.03 & 0.21 @xmath8 0.02 + [ coma ] both @xcite and @xcite tabulate their ellipticities .
the median ellipticities in the coma sample of @xcite are @xmath114 for the s0 galaxies and @xmath115 for ellipticals .
both medians are slightly rounder than our measured low redshift values , though the differences are small .
the lower values are possibly due to the use of the psf in our analysis .
@xcite did not remove the smoothing of the psf , and so will be offset to slightly rounder values as a result .
@xcite derive ellipticies for the s0 galaxies and ellipticals .
these values have been used as a low - redshift comparison set by other higher redshift studies @xcite , it is useful to understand the difference between these measurements and ours .
these ellipticities are measured at a fixed @xmath116 isophote .
thus measurement of the ellipticity from @xcite is not directly comparable to ours and is not the optimal approach for our study .
we plot in figure [ hiz_mass_e ] the distribution of ellipticities as a function of absolute @xmath26 magnitude for four redshift bins . as before
, we select only those galaxies in a fixed magnitude range .
we adjust the magnitudes , however , by @xmath73 mag to reflect the amount of passive galaxy evolution that is measured from the fundamental plane @xcite .
we combine each cluster s sample into three redshift bins , @xmath83 , @xmath117 and @xmath85 .
we show , with a solid line , the median ellipticity for the selected sample in each bin . ,
for all of the galaxies in @xmath2 cluster sample is shown by the blue square and line .
the other squares are the medians for all of the cluster galaxies in the redshift ranges @xmath83 ( green ) , @xmath118 ( orange ) , and @xmath85 ( red ) samples , respectively .
the median ellipticity for the whole of the @xmath1 sample is @xmath119 .
the median ellipticity for cluster galaxies in the range of @xmath120 is @xmath121 , while the median ellipticity for cluster galaxies in the range of @xmath118 is @xmath122 . at @xmath85 ,
the median ellipticity is @xmath123 .
all values are all in excellent agreement with the @xmath124 we find for the @xmath2 sample .
we find no individual clusters that have drastically different ellipticity distributions .
the lack of any trend in the ellipticity , and minimal cluster - to - cluster variance , is a striking result .
the median ellipticity at @xmath1 is statistically identical to that at @xmath2 , being higher by only 0.01 @xmath8 0.02.,width=336 ] we find no evolution in the median ellipticity with redshift . in figure
[ avgell ] , we plot the median ellipticity for each cluster . in this plot
we show the median ellipticity for each cluster as open circles . for our @xmath2 sample ,
we show just the median value ( the blue square ) , @xmath125 .
the median values for the individual @xmath1 clusters show larger scatter , but in general are quite consistent with the low redshift value , with the whole @xmath1 sample having @xmath119 .
the green , orange , and red points are the median ellipticities ( shown also in figure [ hiz_mass_e ] as the horizontal lines ) at @xmath83 and at @xmath118 and @xmath85 , @xmath126 , @xmath127 , @xmath123 , respectively .
the errors we quote , here and later in the paper , are the errors on the median , or @xmath128 the error on the mean .
we confirmed these error estimates with bootstrap resampling .
we compared the scatter in the high redshift sample by computing the @xmath129 around the low redshift median value .
we find a @xmath130 for @xmath131 degrees of freedom , confirming both the good visual agreement between the high redshift data and that the scatter is not higher than expected from random errors . as we discuss in appendix [ simreal ]
, we expect the largest systematic errors on the median ellipticity values to be @xmath132 , which are not large enough to shift the high redshift data to significantly smaller ellipticity values .
the lack of any change in ellipticity with redshift is striking .
formally , _ the median ellipticity of our sample of 487 early - type galaxies at @xmath133 is statistically identical to that of the 210 early - type galaxies at @xmath134 , being higher by only 0.01 @xmath8 0.02 or 3 @xmath8 6% .
_ to add a more detailed assessment of the changes and to provide a more quantitative basis for the results seen above , we plot the ellipticity distributions of the samples in figure [ sumz_edist ] , both differentially and in cumulative form . in each figure
, the blue line shows the @xmath2 comparison sample , with the normalization rescaled to match the higher redshift samples . at no redshift
do we find a statistically significant change in the ellipticity distribution .
we use a number of tests to quantify this , including a kolmogorov - smirnov , wilcoxon - mann - whitney rank sum test and a kuiper test .
taking the results from these tests we find that _ the distribution of ellipticities at @xmath1 agrees with the shape of the @xmath2 distribution at the 1 - 2% level ( i.e. , the probability that they are drawn from the same distribution is 98 - 99% ) .
_ this is a remarkable demonstration of the consistency of the ellipticity distributions over a time span of more than half the age of the universe .
there is a hint , in our @xmath85 sample , of a lack of high ellipticity galaxies , though it is not significant in any of our tests .
this is , in large part , because of the small sample size .
there are only 48 galaxies in the @xmath85 cluster sample . in
[ ellm ] , we showed that the median ellipticity for the @xmath2 sample was essentially unchanged when coma ( a1656 ) was removed ( see table [ coma ] ) .
as expected from that test , there is essentially no change in these results on the ellipticity trends whether coma is included or removed from the @xmath2 sample .
we find that neither the median ellipticity nor the shape of the high redshift ellipticity distribution of early - type cluster galaxies evolves with redshift , implying no change in the overall distribution of the bulge - to - disk ratio of early - type galaxies with redshift . as we show in figure [ type_e ]
, the s0 population at low redshift ( @xmath2 ) has a different ellipticity distribution than the elliptical population , with s0 galaxies having a higher median ellipticity .
if the s0 fraction of the early - type galaxy population decreases , then the median ellipticity of the early - type population should decrease .
the lack of evolution we observe in the median ellipticity and in the shape of the ellipticity distribution implies little or no evolution in the s0 fraction .
this differs from the expectation from previous work , such as that of @xcite , which finds a decrease in the s0 fraction with redshift .
one possible reason why we may not find any evolution in the ellipticity distribution of the cluster early - type population could be because of the nature of our sample .
the fraction of galaxies morphologically selected as s0 galaxies may not change in our sample as has been found by other authors . however , when we look at the s0 fraction of our sample with redshift , using the visual classifications for our sample from the literature , we find a similar trend to what has been reported in other papers , i.e. , a lower fraction of s0 galaxies at redshifts @xmath11 . in figure
[ s0frac ] , we plot the fraction of morphologically identified s0 galaxies with redshift ( from the studies discussed in [ morph ] ) .
our sample has properties consistent with previous work @xcite as would be expected since our sample largely overlaps with previous studies and uses visual classifications from those studies .
we find 42 @xmath8 2% of the early - type galaxy population are s0 galaxies at @xmath11 . for comparison ,
the fraction of s0 galaxies within @xmath35 for the @xcite `` @xmath9 composite '' sample is 35 @xmath8 3% .
this shows that our red - sequence selection is consistent , being less than 2@xmath40 different ( the postman _
et al . _
2005 value is derived by summing the individual listings in that paper s table 4 under the item labeled `` @xmath9 composite '' ) .
our @xmath2 sample of early - type cluster galaxies , selected in the same manner as our @xmath1 sample , has a s0 fraction of @xmath135% .
this fraction is consistent with other analyses . examining the whole sample of early - type galaxies within @xmath35 , we find a s0 fraction of @xmath136% for all of the early - type galaxies within for the 10 clusters we use from @xcite , in good agreement with that found by @xcite . therefore
, our red - sequence selection produces a comparable sample of early - type cluster galaxies to those from previous efforts .
we find no statistically significant evidence for evolution within the @xmath11 cluster sample , which is also consistent with most of the previous work ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . using the visual classifications
, our sample also shows that evolution in the s0 fraction occurs between @xmath5 and @xmath15 as the authors above have noted . note that the fractions we are discussing here are the fraction of s0 galaxies in the early - type galaxy population ( see *
* for a discussion of the size of the systematic error in separating the elliptical and s0 populations . )
the result from our measurements of the ellipticity distributions in our clusters is that there is essentially no evolution in the ellipticity distributions from @xmath9 to @xmath5 , and thus that there is no change in the overall distribution of the bulge - to - disk ratio of early - type galaxies .
if we assume no evolution in the bulge - to - disk ratio distributions of the elliptical and s0 population separately as others have done , we can conclude that the e / s0 ratio does not evolve over this interval .
this contrasts with the clear and significant evolution in the s0 fraction seen from many studies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) based on visual classifications . while one explanation is that there is a problem with the visual classifications , this raises an interesting issue . in @xcite and @xcite , what evolves at @xmath11 is the fraction of s0 galaxies and the fraction of spiral and irregular galaxies .
in contrast , the fraction of elliptical galaxies does not evolve .
if the misclassification of round s0 galaxies as ellipticals alone causes the fraction of s0 galaxies to decrease , it is puzzling that there is no corresponding increase in the fraction of elliptical galaxies in @xcite and @xcite .
rather , in previous work , the spiral fraction increases , and there is little expectation that face - on s0 galaxies would be misclassified as spirals .
the source of this difference remains to be resolved .
our study does not provide an opportunity for a resolution of this issue , but we wanted to highlight it for further work by others . to the total number of elliptical and s0 galaxies .
the fraction of s0 cluster galaxies at @xmath11 is lower than that seen at @xmath2 , as seen in previous work .
we compute the average fraction of s0 galaxies in the early - type population for all of the @xmath1 clusters and plot that value , @xmath137% as a solid black line .
for contrast , we show the @xmath2 value , @xmath135% , as a solid blue line .
we plot , as squares , the average values of the s0 fraction of early - type galaxies in three redshift bins , @xmath83 ( green ) , @xmath118 ( orange ) and @xmath138 ( red ) .
our sample shows the same trend in s0 fraction found by other work .
however , the very different fraction of galaxies classified as s0 s at @xmath11 is in seeming contradiction with the lack of evolution in the ellipticity distribution.,width=326 ] our simulations show that the estimates we use for ellipticities are robust , and do not show a large systematic change with redshift ( see figure [ cl_e ] . ) some of the observed change in the fraction of s0 galaxies in figure [ s0frac ] may come about from a misclassification of round galaxies as ellipticals , as others @xcite have suggested and as we have discussed earlier
. it would be useful to examine a sample of s0s for which the elliptical contamination was , likely , very small .
we examined the fraction of galaxies with ellipticities above the median of the @xmath2 s0 population , @xmath139 .
measuring the fraction of galaxies with @xmath140 should give us an estimate of the s0 population that we expect to be modestly contaminated by elliptical galaxies .
in our @xmath2 sample , we find that @xmath141% of galaxies with @xmath142 are ellipticals .
we plot the fraction of galaxies with @xmath140 in figure [ sharp ] as a function of redshift .
this once again , shows no change with redshift , implying a s0 fraction that does not evolve under the assumption that the ellipticity distribution of ellipticals and s0 galaxies do not separately evolve .
s0 galaxies .
the open circles are the fractions of galaxies with @xmath140 , in each cluster .
the squares are the fractions for the @xmath143 ( blue ) , @xmath144 ( green ) , @xmath145 ( orange ) , and @xmath85 ( red ) samples respectively . at low redshift
, this fraction has only a small contamination from elliptical galaxies , such that the fraction of galaxies with @xmath140 is the half fraction of s0 galaxies with a 5% contamination from elliptical galaxies .
if the observed lack of evolution in the median ellipticity of the @xmath1 population is a result of evolution in the ellipticities of elliptical galaxies masking the decline in the s0 population , we should still see a change in the population of galaxies with @xmath140 .
galaxies with @xmath140 represent the most inclined of the disk - dominated early - type galaxies at low redshift .
the lack of evolution in this fraction suggests little evolution in the fraction of disk - dominated early - type galaxies as a whole.,width=336 ] if the s0 fraction decreases with redshift while the median ellipticity of the early - type population stays the same , we expect that the ellipticity distribution for the s0 galaxies should increase . in figure [ avge_type ] ,
we plot the median ellipticity of galaxies that are classified visually as s0 galaxies .
we show that the median ellipticity of the s0 population is higher in the higher redshift clusters . in the @xmath133 sample , we find that the median ellipticity of the s0 galaxy population is @xmath146 as compared with @xmath99 in our @xmath147 low redshift sample .
the error we quote is the error on the median , or @xmath128 the error on the mean .
we confirm these errors with bootstrapping which yields good agreement . ] .
the difference between the @xmath133 and the @xmath148 s0 samples medians is significant at the @xmath149 when using a student s t - test . restricting the redshift range to @xmath11 , the redshift where other authors have reported that the s0 fraction significant evolves , we find @xmath150 , which is also a @xmath149 difference .
the significance remains @xmath149 regardless of whether coma is included in the @xmath2 sample or not .
cluster sample .
the other squares are the medians for the @xmath144 ( green ) , @xmath118 ( orange ) , and @xmath85 ( red ) samples , respectively .
we also show the median value for the @xmath2 s0 sample as a blue line .
we show , with a black line , the median ellipticities of all the s0 galaxies in our @xmath1 sample .
the median ellipticity of the s0 population increases at @xmath1 , with @xmath151 .
the median ellipticity of the @xmath2 s0 galaxy is @xmath152 .
such a change in the ellipticity of galaxies classified as s0s can explain how the fraction of galaxies classified as s0 s decreases while , at the same time , the median ellipticity of the cluster galaxy population as a whole stays the same.,width=326 ] in figure [ hiz_mass_etype ] ( similar to figure [ hiz_mass_e ] , but with just the complete sample with e and s0 identified ) , we plot the ellipticities as a function of absolute magnitude in four redshift bins .
it appears from this figure that there is a deficit of rounder s0 galaxies ( blue points ) at higher redshifts ( excluding the @xmath85 bin which is less complete and has poor statistics ) , implying that the s0 galaxies at @xmath133 are drawn from a different distribution than those at @xmath148 .
the progression of higher ellipticities for the s0 population thus appears to come about from a change in the shape of the ellipticity distribution of s0 galaxies at higher redshift as compared with those at lower redshifts .
we discuss this in
[ evols0 ] . in figure
[ hizell_type_dist ] , we quantify what we see in figure [ hiz_mass_etype ] .
we plot the distributions of the ellipticities for the @xmath133 elliptical and s0 galaxies separately .
we find that the shape of the @xmath1 s0 distribution differs from the @xmath2 distribution , in contrast with the elliptical distributions which show little change between the two samples .
for example , if we examine the number of s0 galaxies with @xmath153 , we find that there are 56 in the @xmath1 sample , whereas we expect 95 from the scaled low redshift sample of s0 galaxies .
we compared the shapes of the two s0 ellipticity distributions using a number of statistical tests .
the probability of such a change in the ellipticity distribution of s0 galaxies is @xmath149 when using a kolmogorov - smirnov , a wilcoxon - mann - whitney rank sum test , and a kuiper test .
these are the same tests that showed no evolution in the ellipticity distribution in [ hizelldist ] .
these tests reinforce the result , from the analysis of the trends in figure [ avge_type ] , that the cluster s0 galaxies have a different ellipticity distribution at @xmath133 than at @xmath148 .
these trends , both in the median and in the shape of the distribution , are present regardless of how we select the higher redshift sample , whether @xmath1 , @xmath11 or @xmath154 . in all cases ,
the distributions differ from the @xmath2 sample at @xmath149 .
our work is not the only one that finds a different median ellipticity for s0 galaxies at higher redshift in cluster samples .
@xcite give the morphological classifications and ellipticities of a sample of s0 galaxies at @xmath155 .
the authors measured the ellipticities of the cluster galaxies using the same approach and software that we use ( and thus their ellipticities are psf - corrected . ) for their cluster s0 galaxies , @xcite find @xmath156 , in excellent agreement with our value of @xmath157 .
previous work found little or no evolution in the ellipticities of s0 and elliptical galaxies .
however , most of the previous work relied on measuring the ellipticities of galaxies using flux - weighted moments of galaxies , without taking into account the `` blurring '' effects of the psf or the lower resolution of observations at higher redshifts .
we show , in appendix [ ellrob ] , that the ellipticity measurements of galaxies based on simple flux - weighted moments are biased quite significantly towards rounder values for the acs and wfpc2 data that has been used without correcting for the effects of smoothing by the psf . the observed lack of any evolution in the ellipticity distribution for early - type galaxies has interesting implications for the distribution of the bulge - to - disk ratios of the early - type population from @xmath158 to @xmath159 .
the average observed ellipticity is directly related to the average intrinsic ellipticity @xcite . if the average bulge - to - disk of a population of galaxies changes , the average intrinsic ellipticity must change too , and vice versa .
our measurement of an unchanging ellipticity distribution implies that the average bulge - to - disk remains constant .
we performed simple simulations to derive the average ellipticities using models of galaxies with a variety of bulge - to - disk ratios projected over all possible viewing angles with a fixed scale length for the bulges and disk .
we find a roughly linear relation between the average bulge - to - total ratio and the average ellipticity of the population .
for example , for these simple models , a change in the average bulge - to - total ratio from 0.1 to 0.3 , or from 0.3 to 0.6 , would be observed as a shift of @xmath160 in the average ellipticity .
such a shift would be a @xmath161 change for our samples ( see figure [ avgell ] . )
while one could potentially develop a model whereby the underlying scale lengths relatively evolve with redshift , this seems rather contrived .
`` occam s razor '' would lead one to prefer a model where a constant ellipticity distribution implies a constant distribution of bulge - to - disk ratios over the redshift range from @xmath9 to @xmath10 .
we performed simulations to see if the observed s0 fraction of 42% , for the sample of clusters with @xmath11 , is consistent with the lack of evolution we see in the ellipticity distribution .
we constructed 10,000 mock catalogs for the clusters in the @xmath11 sample . for each cluster
, we assumed that the fraction of s0 galaxies was 42% .
we randomly drew the ellipticities for the s0s from the observed distribution of s0s at @xmath2 , and drew the ellipticities for the es from the observed low redshift distribution for es ( see figure [ type_e ] . )
we added the uncertainties and systematics applicable at high redshift as detailed in
[ measure ] and appendix [ simreal ] the resulting ellipticity distributions had median ellipticities lower than the observed value ( 0.29 ) for all 10,000 of the mock catalogs , with a typical value of @xmath162 .
this rules out at @xmath163 the assumption that the underlying ellipticities and s0 fraction are both constant with redshift . our others tests which we use to compare the @xmath2 sample and the @xmath1 sample give similar results ( the kolmogorov - smirnov statistic , a wilcoxon - mann - whitney rank sum statistic and a kuiper statistic ) .
none of the mock catalogs show the same good agreement between the @xmath1 sample and @xmath2 sample in the ellipticity distribution as we observe .
we repeated these simulations for a variety of s0 fractions .
we find that the s0 fraction value that best matches the @xmath11 ellipticity distribution is 66% with a @xmath164 , or 68% confidence limits , of @xmath86.5% .
another way to consider this is a @xmath165 change in the median ellipticity at @xmath1 , a change of @xmath166 , would imply a change in the s0 fraction of 6.5% .
we can rule out , at the 95% confidence limit , s0 fractions below 53% , and we can rule out the observed fraction of visually classified s0 galaxies of 42 @xmath8 2% at the @xmath167 confidence limit or at the 4@xmath40 level . to reconcile
the evolution seen in the s0 fraction in the visually - classified samples with the lack of evolution in the ellipticity distribution requires that either the s0 and elliptical populations evolve in ellipticity in such a way that the combined samples shows no evolution , or that some fraction of the s0 population has been misclassified as other morphological types ( cf .
* ; * ? ? ?
one possibility is that the higher ellipticity , or more edge - on , s0 population is contaminated with misclassified red spirals .
this would cause the s0 fraction to be overestimated , while , at the same time change the shape of the ellipticity distribution .
the spiral fraction among galaxies in the mass range of the cluster sample in our work is @xmath168% @xcite while the fraction of dusty , red objects is @xmath169% @xcite .
since the misclassification would be predominately for edge - on systems , these fractions are likely upper limits for the fraction of spirals misclassified as s0 galaxies .
is it possible that such a small fraction of `` spiral '' galaxies could contaminate the population visually classified as s0 galaxies and change the ellipticity distribution at higher redshift ?
we examined the ellipticities of the galaxies classified as spirals in both the @xmath2 and @xmath1 sample to see if there is a deficit of edge - on spirals .
we found the opposite , that there is a shift in the high redshift spiral population to larger ellipticities , @xmath170 at @xmath2 ( as expected , see * ? ? ?
* ) to @xmath171 at @xmath1 , but it is not statistically significant . from this
we conclude that a red spiral population is not a major source of contamination , especially at large ellipticities .
given the quantitative nature of the current study we feel that it provides a key datum for consideration of the evolutionary history of early - type galaxies over the last 8 gyr .
the result that the ellipticity distribution is essentially unchanged from @xmath5 to @xmath19 , combined with the result that the fraction of highly elongated disk - dominated systems does not evolve ( see figure [ sharp ] ) , provides strong evidence that the overall bulge - to - disk ratio distribution of the population does not evolve .
we have compiled a sample of 10 @xmath172 clusters of galaxies and a comparable sample of 17 @xmath1 clusters of galaxies with _ hst _ imaging . for each cluster , we selected a subsample of galaxies that lie on the red sequence and
have been classified as an early - type galaxy ( elliptical or s0 galaxy ) . to derive a robust sample at all redshifts , we selected galaxies to fall within a magnitude range of @xmath173 and required the galaxies to lie within @xmath36 ( to ensure that the de - projected sample lies within @xmath3 ) of the cluster center .
we change the rest - frame @xmath26 magnitudes by @xmath73 to match the mass - to - light evolution of cluster early - type galaxies as measured by the fundamental plane @xcite .
the magnitude limits correspond to @xmath67 or galaxies with stellar masses roughly between @xmath68 , assuming a `` diet '' salpeter imf .
we performed extensive simulations of the ellipticity measurements to test their robustness and to minimize systematics with redshift .
we found that it is crucial to use model fits ( e.g. , galfit ) where the effect of the psf can be modeled and removed . from our extensive simulations we find that we can robustly measure the ellipticity of cluster early - type galaxies ( by using psf - corrected model fits ) out to @xmath174 at magnitudes that allow the galaxies to also be morphologically classified .
the systematic error we find in our simulations is only @xmath175 , or a 3% change for the median galaxy .
we do not apply this small correction to our data .
we find that not including the effect of the psf can cause systematic shifts of @xmath176 , or a 30% change for the median galaxy .
using our two samples of cluster early - type galaxies , we measure the evolution in the ellipticity of @xmath1 cluster early - type galaxies compared with a substantial sample at @xmath2 . 1 .
we find no evolution in the median ellipticity of @xmath1 cluster early - type ( elliptical and s0 ) galaxies relative to the low redshift sample .
the median ellipticity at @xmath2 is @xmath125 , and is @xmath177 at @xmath1 . _
the median ellipticity of our sample of 487 early - type galaxies at @xmath133 is statistically the same as that of the 210 early - type galaxies at @xmath148 , being higher by only 0.01 @xmath8 0.02 or 3 @xmath8 6% .
the shape of the ellipticity distribution of @xmath1 galaxies also does not evolve .
_ the distribution of ellipticities at @xmath1 agrees with the shape of the @xmath2 distribution at the 1 - 2% level ( i.e. , the probability that they are drawn from the same distribution is 98 - 99% ) . _
3 . using visual classifications from previous reference studies
, we find that our sample shows a similar decrease in the fraction of early - type galaxies classified as disk - dominated systems ( s0s ) at @xmath11 .
as other studies have found , t he s0 fraction is 68@xmath83% at @xmath2 and decreases to 42@xmath82% at @xmath11 .
4 . for the s0
fraction to decrease with increasing redshift , while the median ellipticity of the overall cluster early - type population stays the same , the median s0 ellipticity needs to be larger at higher redshift .
there is a trend in our data for @xmath1 in this sense .
the median s0 galaxy at @xmath2 is @xmath178 while at @xmath133 , we find a @xmath179 different @xmath180 for galaxies classified as s0s . whether this change is large enough to account for the change seen in the morphological sample
remains to be determined .
we explored if the change in the ellipticity distribution of s0 galaxies could come from a deficit of round s0 galaxies or from the addition of misclassified spiral galaxies .
however , we found that neither of these simple scenarios can explain our results .
our results on the unchanging ellipticity distributions lead us to conclude that there has been little or no evolution in the overall distribution of bulge - to - disk ratio of early - type galaxies over the redshift range @xmath181 for morphologically - selected samples of early - type red - sequence galaxies with @xmath182 in the dense cores of clusters inside @xmath3 .
in particular , our results allow us to rule out a s0 fraction of @xmath18347 - 51% at @xmath1 at the 3@xmath40 level assuming no evolution in the ellipticity distributions of elliptical and s0 galaxies . _
if we assume , as in all previous studies , that the intrinsic ellipticity distribution of both elliptical and s0 galaxies remains constant , then we conclude from the lack of evolution in the observed early - type ellipticity distribution that the relative fractions of ellipticals and s0s do not evolve over the last @xmath7gyr , or from @xmath9 to @xmath12 , for a red - sequence selected sample of early - type galaxies in the cores of clusters of galaxies .
_ these results do highlight an inconsistency within the wide range of studies that have occurred on the evolution of early - type galaxies over the last decade . over this redshift range , and particularly since @xmath14 , the fraction of morphologically - identified s0 galaxies from visual classifications has been found to be significantly lower than the @xmath10 value ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
reconciling the trends seen in the elliptical , s0 and spiral fractions , as discussed in [ morphevol ] remains to be understood .
the early - type galaxy population , both in the field and in clusters , does not evolve purely passively .
the volume - averaged number density of red galaxies , which are mostly early - type galaxies @xcite , grows by a factor @xmath184 between @xmath19 and the present @xcite . for the cluster population
such evolution is harder to quantify , but mergers @xcite and filaments around massive clusters at @xmath185 ( e.g. , * ? ? ?
* ; * ? ? ?
* ) suggest that interactions and infalling galaxies enhance and modify over time the galaxy population in the cluster core .
besides evolution in the population of early - type galaxies , individual early - type galaxies also change over time .
as pointed out by , e.g. , @xcite , the evolution of the line strengths is not compatible with purely passive evolution .
moreover , early - type galaxies at @xmath9 were recently demonstrated to be significantly smaller than today by a factor of two @xcite .
hence , the resulting picture is complicated : early - type galaxies , both as individual objects and as a population , undergo substantial changes between @xmath19 and the present .
on the other hand , many basic properties of the early - type population have remained unchanged over the past @xmath7gyrs , providing useful means for characterizing and quantifying the observed evolution . at any redshift
@xmath9 early - type galaxies occupy a tight color - magnitude relation ( e.g. , * ? ? ?
* ) and fundamental plane ( e.g. , * ? ? ?
* ) , which suggests smooth and regular evolution .
moreover , neither in the field nor in clusters has the early - type _ fraction _ changed significantly in mass - selected samples @xcite . together with these characteristics , the results presented in this paper fit into a picture in which the field and cluster early - type galaxy populations grow and change , but only while leaving many basic characteristics the same . in particular , we have demonstrated from the constancy of the ellipticities of early - type galaxies over the last @xmath7gyr that the bulge - to - disk ratio distribution of the cluster population remains constant .
this suggests that processes that change the bulge - to - disk ratio of individual early - type galaxies and the bulge - to - disk ratios of newly formed or accreted early - type galaxies are balanced such that the overall bulge - to - disk ratio distribution remains the same .
it remains an open question why this is the case , but it is clear that much can be learned about the formation process of early - type galaxies by studying their properties in even more detail and extending their observation to higher redshifts . the authors would like to thank s. adam stanford for useful comments .
we would like to thank the anonymous referee for suggestions that improved this paper . finally , we would like to thank stefano andreon for pointing out previously published related papers , and dave wilman for showing us his work ahead of publication .
acs was developed under nasa contract nas5 - 32865 , this research was supported by nasa grant nag5 - 7697 .
we are grateful to k. anderson , j. mccann , s. busching , a. framarini , s. barkhouser , and t. allen for their invaluable contributions to the acs project at jhu .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
the analysis pipeline used to reduce the deimos data was developed at uc berkeley with support from nsf grant ast-0071048 .
we made a number of additional tests of the robustness of the ellipticity measurements with the goal of estimating the magnitude of any systematic errors .
these include adding fake galaxies into our imaging data , comparing the ellipticity measurements made using sextractor to those using galfit , and using the a range of ( incorrect ) psfs to estimate the possible sensitivity to systematic errors .
we performed a number of simulations using fake galaxies to test the reliability of the ellipticity measurements .
these simulations were done in order to establish the best approach to use and to assess how well we can measure the ellipticity , given our data .
the artificial galaxies were made by simulating two srsic models with the same centroid , one with @xmath186 and one with @xmath187 .
the two components had the same luminosity , but different effective radii .
the first effective radius was chosen to be three times the resolution of the acs camera , while the second was nine times the acs resolution .
each component was given the same ellipticity .
these represent well - resolved lenticular galaxies .
these fake images were then simulated at magnitudes typical of the early - type galaxies in the @xmath188 clusters in our acs imaging , and then added to a real acs image .
reassuringly , the simulations recovered the input ellipticities to a high accuracy , with a scatter of just @xmath189 and an offset of @xmath190 .
the scatter and the offset was highest for the most round objects . for objects with an input ellipticity of 0 , the average recovered
ellipticity was 0.03 with a scatter of @xmath191 .
as expected , objects that would be seen as perfectly round given an arbitrarily high signal - to - noise , are not measured to be round at typical signal - to - noise levels , but the offset , as noted above , was quite small .
these simulations were made for a variety of bulge - to - disk ratios , by varying the relative luminosities of the two components .
in addition , we assumed that the intrinsic ellipticity of the @xmath186 bulge was @xmath192 , or @xmath193 when averaged over all projections or over all @xmath16 .
we compiled a number of images of low redshift elliptical and s0 galaxies .
we began with the catalog of @xcite .
that catalog contains 24 elliptical and s0 galaxies , including two classified as s0/a .
we expanded upon this by including 49 additional early - type galaxies in the virgo cluster .
each was selected to be in the virgo cluster sample discussed in @xcite .
most of the galaxies in that sample have accurate distances from surface brightness fluctuations .
those that do not have distances were assumed to be at the center of the main virgo concentration .
once we selected the list of early - type galaxies , we extracted images of those not in @xcite from the sdss dr5 in the @xmath194 filter .
for those images from the sdss , we removed nearby bright stars and other galaxies , replacing the pixels with those randomly selected from nearby blank regions .
this produced a final collection of 73 images of nearby galaxies imaged in either @xmath194 or @xmath26 .
we added these galaxies , appropriately scaled in size and magnitude , to the imaging data used for each of the clusters , and replicated the detection and measuring process .
the galaxies were resized as appropriate for the redshift of the cluster , convolved with a representative psf , and then noise was added to each .
each galaxy was simulated at a variety of input magnitudes , and multiple realizations were done at each magnitude to evaluate the effect of noise and binning .
this is the same process used in @xcite .
the magnitudes were selected to span the range of observed galaxy magnitudes , to 0.4 mag below the limit used for classifying galaxies in each sample .
this is important to ensure that the selection at the limit is not influenced by the magnitude cutoff .
we compared the offset between the ellipticity measurement from the original image and that of the simulated high redshift image .
in general , the offset between the low redshift original image and high redshift scaled image ellipticities are consistent . for galaxies within our magnitude limits
, we found that the typical systematic offset is from @xmath195 to @xmath196 , where the simulated galaxies are measured to be less elliptical at high redshift than in their original , low redshift images . for galaxies at or below our selection limits @xmath197 to @xmath198 .
the median systematic offset was @xmath199 for our whole sample of simulated galaxy images .
we show the ellipticities measured in the output of the simulations as a function of the input ellipticities in figure [ cl_e ] .
the scatter scales with the input signal - to - noise as expected .
one exception is that , in general , the measurements made with wfpc2 are worse than those made with acs at the same redshift .
this is expected , because wfpc2 is under - sampled and the images are typically at a lower s / n .
we used the multiple realizations of the galaxy simulations to examine the random errors .
the scatter in the measurements for a given input galaxy in a given cluster observation is , on the whole , small .
the acs imaging has measurements with a median scatter of @xmath200 in ellipticity , while the wfpc2 imaging shows larger median scatter of @xmath201 in ellipticity .
these numbers are close to the typical errors reported by galfit .
galfit s reported errors scale with the magnitude of the galaxy , which we also find in our simulations .
at the faintest magnitudes , at or below the magnitude limits for classification , we find a scatter on the ellipticities as high as @xmath202 , once again matching the scatter in the ellipticity as estimated by galfit . because the psf so important , we performed simulations where galfit was run using a different psf than the psf that was used to convolve with the data .
we reran the simulations for ms 1054 - 03 using the correct @xmath203 psf when simulating the @xmath204 galaxy , but we ran galfit using the @xmath49 psf .
the differences between all three psfs in terms of encircled energy are small @xcite .
however , there is an asymmetry in the psf of the @xmath49 filter @xcite that causes a noticeable difference in the resulting psfs .
we show the results of using the @xmath49 psf to fit in figure [ badpsf ] .
we find an overall offset of @xmath205 between the average ellipticities , which is of order the same size as the bins we use in the top panel .
examining the lower panel , we can see that the offset is not constant with ellipticity , but also depends on the input galaxy used for the simulation .
for example , along the x - axis at @xmath206 , there are a number of points where the incorrect psf yields @xmath207 while other galaxies have similar ellipticities when using the correct or incorrect psf . from this
we can conclude that typical systematic errors on the ellipticity measurements are on the order of @xmath208 , but this depends on the shape of the galaxy .
the psf is more asymmetric in the redder passbands , so the systematic error is likely to be largest for the highest redshift clusters .
the statistic we use the most is the median ellipticity . using these simulations , we find that the offset in the median ellipticity is @xmath209 for the @xmath49data .
this is a quite a small effect , given the noticeable changes that we have made to the psf .
which were convolved with the @xmath203 psf , but the modeling of the galaxy luminosity profiles was done using the @xmath49 psf has some asymmetry .
the use of the incorrect psf in this case has only a modest impact , biasing the ellipticities towards higher values by @xmath210 at the highest ellipticities and becomes smaller , @xmath209 or a @xmath656% change , at the median input ellipticity . , width=326 ] previous work , such as @xcite and @xcite , use sextractor s flux - weighted moments to estimate the ellipticities of galaxies
this approach , however , does not remove the effect of the psf .
below we will compare the ellipticity measurements we use with those used by previous work in the literature .
we will also use our simulations to evaluate how well these techniques reproduce the underlying ellipticity distribution . in figure [ e_comp ] , we compare our ellipticity measurements for cl 0016 + 16 cl 0024 + 16 and cl 0054 - 27 with those from @xcite . our measurements , as we stated above , are based on fitting elliptical models of srsic profiles to the data , including the effects of the psf .
the results from @xcite are based on using sextractor @xcite .
sextractor estimates the ellipse by computing the flux - weighted second - order moments of the galaxy .
sextractor includes those pixels above the threshold defined by an isophote as measured in a smoothed image .
@xcite used a detection isophote of @xmath211 and a 03 diameter , or 3 pixels with wfpc2 , top - hat filter for the smoothing .
we find that our ellipticities are less round than those from @xcite .
we hypothesize that this comes about for two reasons .
first , our approach removes the smoothing from the psf which tends to circularize the ellipticity measurements .
second , much of our data was observed with acs , which has a more compact psf than wfpc2 .
the median offset in the ellipticities for s0 galaxies is @xmath212 .
this is a very large change , much larger than any of the systematic error we found in our simulations of our ellipticity measurements .
this highlights the importance of including the psf in the measuring process , and does indicate the challenge of using wfpc2 data for measurements at higher redshift . , and uses the same symbols , but compares ellipticities from measurements of the second - order flux - weighted moments made with acs imaging , as opposed to wfpc2 as in fig .
[ e_comp ] , to those from fitting models to the galaxies . as we see in fig.[e_comp ] , the luminosity - weighted moment estimates are generally rounder than the ones measured by galfit model fitting , despite the higher resolution and better sampling of acs as compared with wfpc2 .
this figure shows that modeling the psf when measuring the ellipticity is particularly important for deriving a robust estimate of the underlying value , and that the resolution of the instrument is a lesser factor ( at least in this case between wfpc2 and acs ) . as in fig .
[ e_comp ] , the galaxies classified as s0s are more likely to have small ellipticities when they are larger in size .
the median offset for s0 galaxies is @xmath213 , smaller than we found in fig .
[ e_comp ] , but still much larger than the statistical and systematic errors we find for our psf - corrected galfit model approach . , width=432 ] in figure [ mye_comp ] , we plot the ellipticities we measure using sextractor on the acs images of ms 1054 - 03 and cl 0024 + 16 in comparison to the ellipticities we find by fitting models using galfit .
from these plots , we conclude that the more compact psf for acs is not the dominate reason for the different ellipticities between our work and the work of previous authors .
removing the effect of the psf is the most important component .
finally , we plot the comparison between the ellipticities as derived by galfit and those from sextractor in figure [ compare_e ] for our simulations of high - redshift galaxies .
these are images of real galaxies that have been redshifted and to which we have added noise . as with the previous plots showing trends in the data , the simulated galaxies are measured to be rounder with sextractor than by fitting models using galfit .
the average offset is @xmath214 between the sextractor ellipticity and the galfit ellipticity , in good agreement with the @xmath215 we measure in figure [ mye_comp ] . as we show in figure [ cl_e ] , our galfit ellipticities reproduces the underlying ellipticity distribution . because our galfit method reproduces the input ellipticities so well , we conclude that the sextractor method of measuring ellipticities underestimates the ellipticities of galaxies at higher redshifts .
we would advise against the use of sextractor ellipticities for quantitative studies of high redshift galaxies using acs and wfpc2 where the psf is a modest , but significant fraction of the size of the object . and [ mye_comp ] , namely that the ellipticities as estimated by sextractor are rounder than those from galfit .
the galfit measurements of the simulated @xmath1 galaxies recover the actual ellipticities with an average offset of @xmath216 ( see fig . [ cl_e ] ) , much smaller than the offset of @xmath217 we find between the galfit and sextractor ellipticities .
the higher density of points in the @xmath218 redshift bin simply reflects the larger number of clusters included in that bin.,width=576 ]
of the clusters discussed here , only cl j1226 + 33 has no previously published redshift catalog beyond the six galaxies listed in the discovery paper @xcite .
we have collected redshifts for 21 new members in the central region of the cluster , along with 10 more galaxies outside of the region covered by the acs imaging .
30 of these 31 galaxies were selected to have acs @xmath49 magnitudes @xmath219 mag ab and colors in the range @xmath220 , the remaining one was a `` filler '' object with a bluer color .
the data were taken using the deimos spectrograph on keck ii , in a 1 hr exposure with the 600 line mm@xmath39 grating in variable conditions .
we list the catalog of redshifts in table [ cl1226cat ] .
we measured redshifts by centroiding emission features and cross - correlating with a variety of templates .
we computed the biweight center and scale of all galaxies with redshifts within @xmath221 .
we found @xmath222 with a dispersion of 1143 @xmath8 162 km s@xmath39 , in good agreement with the 1270 km s@xmath39 quoted in @xcite .
we used a biweight to estimate both the redshift and the velocity dispersion of the cluster with errors from the jackknife of the galaxy redshifts .
the redshift range we used for membership excludes two galaxies at @xmath223 . including those two objects
raises the dispersion to 1322 @xmath8 221 km s@xmath39 .
it is often the case that , with small numbers of galaxies that the dispersion is overestimated see for example the dispersions quoted for cl 1604 + 4304 and cl 1604 + 4321 in @xcite versus those from the much larger sample in @xcite .
this arises because nearby large scale structure or groups of galaxies can project into line of sight of the cluster .
we opted to remove the two galaxies at @xmath223 to provide conservative estimate of the dispersion , which we also list in table [ summary ] .
gal12263974 + 3329315 & 12:26:39.74 & 33:29:31.51 & 21.82 @xmath8 0.02 & 1.75 @xmath8 0.03 & 1.035 + gal12264240 + 3330094 & 12:26:42.40 & 33:30:09.40 & 22.11 @xmath8 0.01 & 1.67 @xmath8 0.01 & 0.929 + gal12263338 + 3330277 & 12:26:33.38 & 33:30:27.77 & 21.68 @xmath8 0.01 & 1.82 @xmath8 0.02 & 0.911 + gal12264935 + 3330312 & 12:26:49.35 & 33:30:31.25 & 21.61 @xmath8 0.01 & 1.82 @xmath8 0.01 & 0.891 + gal12263934 + 3330353 & 12:26:39.34 & 33:30:35.35 & 22.40 @xmath8 0.01 & 1.96 @xmath8 0.03 & 0.943 + gal12264600 + 3330347 & 12:26:46.00 & 33:30:34.74 & 21.43 @xmath8 0.01 & 1.84 @xmath8 0.01 & 0.892 + gal12263157 + 3330473 & 12:26:31.57 & 33:30:47.38 & 22.99 @xmath8 0.02 & 0.72 @xmath8 0.02 & 0.963 + gal12262950 + 3330580 &
12:26:29.50 & 33:30:58.01 & 21.59 @xmath8 0.01 & 1.81 @xmath8 0.02 & 1.211 + gal12264172 + 3331041 & 12:26:41.72 & 33:31:04.15 & 22.26 @xmath8 0.01 & 0.59 @xmath8 0.01 & 0.690 + gal12264671 + 3331037 & 12:26:46.71 & 33:31:03.73 & 22.07 @xmath8 0.01 & 1.60 @xmath8 0.01 & 0.767 + gal12263998 + 3331047 & 12:26:39.98 & 33:31:04.72 & 21.70 @xmath8 0.01 & 1.85 @xmath8 0.02 & 0.929 + gal12263648 + 3331124 & 12:26:36.48 & 33:31:12.47 & 22.47 @xmath8 0.02 & 0.99 @xmath8 0.02 & 0.846 + gal12263107 + 3331349 & 12:26:31.07 & 33:31:34.97 & 22.73 @xmath8 0.01 & 0.60 @xmath8 0.01 & 0.965 + gal12263402 + 3331356 & 12:26:34.02 & 33:31:35.62 & 21.84 @xmath8 0.01 & 1.84 @xmath8 0.02 & 0.966 + gal12265120 + 3331385 & 12:26:51.20 & 33:31:38.54 & 22.68 @xmath8 0.01 & 1.90 @xmath8 0.02 & 0.881 + gal12265293 + 3331465 & 12:26:52.93 & 33:31:46.54 & 22.78 @xmath8 0.01 & 1.76 @xmath8 0.02 & 0.964 + gal12263832 + 3331482 & 12:26:38.32 & 33:31:48.21 & 23.27 @xmath8 0.06 & 0.63 @xmath8 0.09 & 0.542 + gal12262701 + 3331580 & 12:26:27.01 & 33:31:58.09 & 22.27 @xmath8 0.08 & 1.82 @xmath8 0.08 & 0.645 + gal12272134 + 3332083 & 12:27:21.34 & 33:32:08.32 & 21.86 @xmath8 0.01 & 1.81 @xmath8 0.01 & 0.893 + gal12262817 + 3332232 & 12:26:28.17 & 33:32:23.23 & 22.93 @xmath8 0.02 & 1.13 @xmath8 0.03 & 0.673 + gal12271941 + 3332269 & 12:27:19.41 & 33:32:26.91 & 23.08 @xmath8 0.02 & 0.70 @xmath8 0.02 & 0.885 + gal12265525 + 3332324 & 12:26:55.25 & 33:32:32.43 & 22.28 @xmath8 0.01 & 1.90 @xmath8 0.01 & 0.895 + gal12271460 + 3332377 & 12:27:14.60 & 33:32:37.70 & 22.10 @xmath8 0.01 & 1.89 @xmath8 0.01 & 0.881 + gal12265629 + 3332414 & 12:26:56.29 & 33:32:41.48 & 22.53 @xmath8 0.01 & 1.78 @xmath8 0.01 & 0.893 + gal12265995 + 3332405 & 12:26:59.95 & 33:32:40.54 & 22.48 @xmath8 0.01 & 1.76 @xmath8 0.02 & 0.892 + gal12265923 + 3332405 & 12:26:59.23 & 33:32:40.59 & 22.21 @xmath8 0.01 & 1.82 @xmath8 0.01 & 0.897 + gal12265689 + 3332437 & 12:26:56.89 & 33:32:43.75 & 22.86 @xmath8 0.01 & 1.84 @xmath8 0.02 & 0.896 + gal12270510 + 3332475 & 12:27:05.10 & 33:32:47.53 & 23.01 @xmath8 0.01 & 0.77 @xmath8 0.01 & 0.798 + gal12265060 + 3332461 & 12:26:50.60 & 33:32:46.18 & 21.50 @xmath8 0.01 & 1.83 @xmath8 0.01 & 0.875 + gal12265825 + 3332485 & 12:26:58.25 & 33:32:48.57 & 19.13 @xmath8 0.00 & 1.96 @xmath8 0.00 & 0.891 + gal12271547 + 3332539 & 12:27:15.47 & 33:32:53.91 & 22.87 @xmath8 0.02 & 0.69 @xmath8 0.02 & 1.034 + gal12270083 + 3333019 & 12:27:00.83 & 33:33:01.97 & 22.50 @xmath8 0.01 & 1.35 @xmath8 0.02 & 0.914 + gal12265214 + 3333071 & 12:26:52.14 & 33:33:07.12 & 23.02 @xmath8 0.02 & 1.76 @xmath8 0.02 & 0.881 + gal12265714 + 3333046 & 12:26:57.14 & 33:33:04.61 & 21.65 @xmath8 0.01 & 1.95 @xmath8 0.02 & 0.930 + gal12270134 + 3333044 & 12:27:01.34 & 33:33:04.42 & 21.27 @xmath8 0.00 & 1.89 @xmath8 0.01 & 0.897 + gal12264489 + 3333094 & 12:26:44.89 & 33:33:09.46 & 22.95 @xmath8 0.01 & 1.83 @xmath8 0.02 & 1.196 + gal12271887 + 3333137 & 12:27:18.87 & 33:33:13.72 & 21.73 @xmath8 0.01 & 1.85 @xmath8 0.01 & 0.897 + gal12271572 + 3333111 & 12:27:15.72 & 33:33:11.11 & 21.60 @xmath8 0.01 & 1.73 @xmath8 0.01 & 0.892 + gal12270766 + 3333138 & 12:27:07.66 & 33:33:13.82 & 22.61 @xmath8 0.01 & 1.65 @xmath8 0.02 & 0.902 + gal12272080 + 3333163 & 12:27:20.80 & 33:33:16.37 & 21.60 @xmath8 0.01 & 1.83 @xmath8 0.01 & 0.819 + gal12271607 + 3333192 & 12:27:16.07 & 33:33:19.25 & 22.47 @xmath8 0.02 & 1.85 @xmath8 0.02 & 0.901 + gal12264296 + 3333195 & 12:26:42.96 & 33:33:19.59 & 22.77 @xmath8 0.01 & 1.00 @xmath8 0.02 & 0.849 + gal12265244 + 3333239 & 12:26:52.44 & 33:33:23.90 & 22.22 @xmath8 0.02 & 1.86 @xmath8 0.02 & 0.893 + gal12270671 + 3333266 & 12:27:06.71 & 33:33:26.65 & 21.87 @xmath8 0.01 & 1.72 @xmath8 0.01 & 0.883 + gal12271695 + 3333261 & 12:27:16.95 & 33:33:26.19 & 21.45 @xmath8 0.01 & 1.87 @xmath8 0.02 & 0.891 + gal12265312 + 3333310 & 12:26:53.12 & 33:33:31.08 & 21.90 @xmath8 0.01 & 1.87 @xmath8 0.02 & 0.896 + gal12264799 + 3333348 & 12:26:47.99 & 33:33:34.80 & 22.29 @xmath8 0.01 & 0.78 @xmath8 0.01 & 0.330 + gal12270750 + 3333439 & 12:27:07.50 & 33:33:43.96 & 22.62 @xmath8 0.01 & 0.90 @xmath8 0.01 & 0.850 + gal12271102 + 3333471 & 12:27:11.02 & 33:33:47.10 & 21.71 @xmath8 0.01 & 1.80 @xmath8 0.01 & 0.755 + gal12270440 + 3334063 & 12:27:04.40 & 33:34:06.35 & 22.74 @xmath8 0.01 & 1.82 @xmath8 0.02 & 0.887 + gal12270214 + 3334060 & 12:27:02.14 & 33:34:06.06 & 21.48 @xmath8 0.01 & 1.61 @xmath8 0.02 & 0.896 + gal12271313 + 3334130 & 12:27:13.13 & 33:34:13.02 & 21.63 @xmath8 0.01 & 1.78 @xmath8 0.01 & 0.763 + gal12271082 + 3334195 & 12:27:10.82 & 33:34:19.58 & 22.40 @xmath8 0.01 & 1.06 @xmath8 0.02 & 0.767 + gal12270730 + 3334237 & 12:27:07.30 & 33:34:23.78 & 22.07 @xmath8 0.01 & 1.82 @xmath8 0.02 & 0.892 + gal12271975 + 3334272 & 12:27:19.75 & 33:34:27.26 & 21.30 @xmath8 0.01 & 0.80 @xmath8 0.01 & 0.530 + gal12272074 + 3334437 & 12:27:20.74 & 33:34:43.71 & 21.30 @xmath8 0.01 & 1.79 @xmath8 0.01 & 0.762 + gal12271545 + 3334558 & 12:27:15.45 & 33:34:55.82 & 21.68 @xmath8 0.01 & 1.61 @xmath8 0.01 & 0.768 + gal12265239 + 3335006 & 12:26:52.39 & 33:35:00.60 & 22.75 @xmath8 0.01 & 1.67 @xmath8 0.02 & 0.764 + gal12270219 + 3335182 & 12:27:02.19 & 33:35:18.22 & 22.64 @xmath8 0.02 & 1.55 @xmath8 0.02 & 0.891 + gal12270676 + 3335289 & 12:27:06.76 & 33:35:28.92 & 21.58 @xmath8 0.01 & 1.77 @xmath8 0.01 & 0.755 + [ cl1226cat ] | we have compiled a sample of early - type cluster galaxies from @xmath0 and measured the evolution of their ellipticity distributions .
our sample contains 487 galaxies in 17 @xmath1 clusters with high - quality space - based imaging and a comparable sample of 210 galaxies in 10 clusters at @xmath2 .
we select early - type galaxies ( elliptical and s0 galaxies ) that fall within the cluster @xmath3 , and which lie on the red - sequence in the magnitude range @xmath4 , after correcting for luminosity evolution as measured by the fundamental plane .
our ellipticity measurements are made in a consistent manner over our whole sample .
we perform extensive simulations to quantify the systematic and statistical errors , and find that it is crucial to use point - spread function ( psf)-corrected model fits ; determinations of the ellipticity from _ hubble space telescope _ ( hst ) image data that do not account for the psf `` blurring '' are systematically and significantly biased to rounder ellipticities at redshifts @xmath1 . we find that neither the median ellipticity , nor the shape of the ellipticity distribution of cluster early - type galaxies evolves with redshift from @xmath5 to @xmath6 ( i.e. , over the last @xmath7gyr ) .
the median ellipticity at @xmath1 is statistically identical with that at @xmath2 , being higher by only 0.01 @xmath8 0.02 or 3 @xmath8 6% , while the distribution of ellipticities at @xmath1 agrees with the shape of the @xmath2 distribution at the 1 - 2% level ( i.e. , the probability that they are drawn from the same distribution is 98 - 99% ) .
these results are strongly suggestive of an unchanging overall bulge - to - disk ratio distribution for cluster early - type galaxies over the last @xmath7gyr from @xmath9 to @xmath10 .
this result contrasts with that from visual classifications which show that the fraction of morphologically - selected disk - dominated early - type galaxies , or s0s , is significantly lower at @xmath11 than at @xmath10 .
we find that the median disk - dominated early - type , or s0 , galaxy has a somewhat higher ellipticity at @xmath1 , suggesting that rounder s0s are being assigned as ellipticals .
taking the ellipticity measurements and assuming , as in all previous studies , that the intrinsic ellipticity distribution of both elliptical and s0 galaxies remains constant , then we conclude from the lack of evolution in the observed early - type ellipticity distribution that the relative fractions of ellipticals and s0s do not evolve from @xmath9 to @xmath12 for a red - sequence selected samples of galaxies in the cores of clusters of galaxies . |
quantum chromodynamics ( qcd ) is generally accepted as the theory of strong interactions between quarks and gluons .
if the quark masses are fixed , there is only one free parameter in the theory , the strong coupling constant .
thus measurements of this parameter are of paramount importance .
such measurements have been performed in a large variety of experiments , for various initial states , processes and energy scales . during recent years good
consistency has been obtained between the different determinations and the theoretical prediction for the running ( energy dependence ) of the coupling @xcite . here a summary of the measurements performed at is given , which contribute in an important way to the achieved precision on the world average value of @xmath0 @xcite .
at the strong coupling constant is measured in @xmath1 annihilations into quarks which subsequently fragment into hadrons , or in @xmath2 , where one or both tau leptons decay hadronically . for inclusive observables such as the total cross section
the hadronic final state is not analyzed w.r.t .
its particular properties such as particle content or topology . in practice
a ratio of cross sections or partial decay widths is measured , in order to take advantage of the cancellation of common systematic uncertainties , @xmath3 @xmath4 .
the qcd correction is known up to next - to - next - to - leading order ( nnlo ) , _ ie .
_ , @xmath5 , with known coefficients @xmath6 .
in addition there are corrections for quark mass and non - perturbative effects . in the case of @xmath7
the combination turns out to be @xmath8 @xcite , which translates into @xmath9 @xcite .
the error is dominated by experimental uncertainties , mainly the statistical error in the muon cross section .
mass corrections and non - perturbative effects are very small , since they are suppressed as @xmath10 and @xmath11 .
theoretical uncertainties arise from the missing knowledge of the higgs mass and the uncertainty on the top mass , as well as from renormalization scale and scheme variations
. for tau decays and have measured the hadronic mass spectrum and determined its moments .
for example , @xmath12 corresponds to the zeroth moment . by measuring several moments , not only can be fitted , but at the same time also some non - perturbative parameters . a combination of the results from the two experiments yields @xmath13 @xcite , which represents one of the most precise measurements .
in contrast to @xmath7 , here the experimental error is negligible , whereas the theoretical uncertainty dominates .
this uncertainty stems from different estimates of higher order corrections .
event shape distributions are observables which are sensitive to the topology of an hadronic event , and thus give direct sensitivity to since gluon radiation influences the topology . for a particular set of infrared and collinear safe variables
the cross section is known at nlo , and large logarithms have been resummed to all orders in ( see @xcite for a review on this topic ) .
this set of observables comprises thrust , heavy jet mass , the differential two - jet rate computed in the durham scheme , the c parameter , as well as the total and wide jet broadening .
these distributions have been measured by all collaborations and used for determinations , for a large set of centre - of - mass energies at i and ii ( for a review see _ eg . _
recently , also the four - jet rate ( durham scheme ) has been employed for measurements @xcite , since a nlo calculation as well as the resummation of large logarithms is at hand . for this kind of observables the dominant uncertainties on are of theoretical origin .
firstly the purely perturbative predictions have to be corrected for hadronization effects , which are larger than for inclusive observables .
these corrections are taken from monte carlo programs based on various phenomenological models , leading therefore to ambiguities .
secondly , there are uncertainties stemming from the estimates of missing higher order terms . as an example , in a recent preliminary analysis by @xcite of the data at 206 gev centre - of - mass energy the following uncertainties on are quoted when combining the results from the six event shapes mentioned above : @xmath14 , @xmath15 , @xmath16 , @xmath17 , @xmath18 .
`` match '' stands for an ambiguity in the combination of fixed order and resummation calculations , and `` scale '' for variations of the renormalization scale .
the statistical error can be substantially reduced by combining with the results from the z resonance and from other experiments .
therefore these findings clearly indicate that a considerable improvement can only be obtained if the theoretical uncertainties are reduced
. this might be possible with the advent of nnlo calculations for three - jet observables .
a somewhat different approach to the problem of missing higher orders has been attempted by @xcite .
using only the nlo prediction for a large set of event shape variables , and setting the renormalization scale equal to the centre - of - mass energy , a large scatter in the values from different observables is found . however , this scatter is considerably reduced if both and the renormalization scale are fitted at the same time .
the price to pay is a large scatter in renormalization scales , from @xmath19 to @xmath20 gev .
these extreme values might not be natural from a theoretical point of view . in any case
, a clear indication for large missing higher orders is obtained , which depend strongly on the observable .
during recent years there has been a lot of activity on the phenomenology of power law corrections to event shape variables . for an overview the interested reader
is referred to @xcite . instead of obtaining the non - perturbative corrections from monte carlo models
it turns out that additive terms in the case of moments or shifts in the case of event shape distributions of the type @xmath21 are capable of giving a good description of the data . using this power law behaviour , can be determined together with a non - perturbative parameter @xmath22 which is expected to be universal .
a recent analysis of a large set of data from a wide range of energies @xcite indicates that indeed universality is obtained at the 20% level for the mean values of event shapes .
however , in the case of distributions still some larger fluctuations are observed .
several effects might be at the origin of this .
it might be that hemisphere variables such as heavy jet mass and jet broadenings need a more careful analysis than event variables such as thrust .
an approach in this direction could be the introduction of a non - perturbative shape function ( see @xcite for a recent review ) .
in addition , further understanding is needed for the effects of hadron masses and resonance decays @xcite . in any case
, a high precision measurement of based on power law corrections will need further theoretical as well as experimental investigations . [ cols="^,^ " , ]
a large set of measurements of is available from .
these determinations contribute in an important way to the world average value and are nicely consistent with the predicted running of the strong coupling , as shown in fig .
[ summary ] .
the results from event shape variables have shown that further theoretical work is needed in order to improve the precision , such as the completion of nnlo calculations for three - jet quantities and a better understanding of power law corrections .
since these developments might not be finished within a few years from now , it is of great importance to preserve the data well , in order to facilitate a possible re - analysis in the future .
i would like to thank s. kluth and s. tafat for providing me with useful information on power law corrections . | a summary is given of the current status of measurements of the strong coupling constant performed at .
these include measurements from inclusive observables as well as from event shape variables .
recent results based on power law corrections are discussed . |
featuring simultaneous tracking and targeting of tumours , image - guided radiation therapy ( igrt ) paves a promising path towards a safer , more efficient and more accurate treatment . in autumn 2004 , varian medical systems , inc .
( vms ) , palo alto , ca , installed the first clinically - applicable solution for igrt : the vms clinac accelerator equipped with imaging functionality . unless a number of calibrations have been performed on such a complex system , the device is not operable .
the aim of the present paper is to describe one such important process , the geometric calibration .
the method outlined here applies equally to imaging and delivery units where the source and a flat - panel detector ( mounted in such a way as to face one another ) move in circular orbits around the irradiated object .
the rotation plane of the source is assumed to contain the geometric centre of the detector ; modifications are needed in case of tilted geometry , e.g. , in a c - arm configuration .
the aim of the geometric calibration is threefold .
( i ) to yield the values of the important parameters defining the geometry of the system .
if an imaging unit is calibrated , these values are subsequently used in the processing of the scan data ( reconstruction phase ) .
if a delivery unit is calibrated , the output may lead to the suspension of the operation of the system in case , for instance , that the deviations of the parameter values from the nominal ones are beyond the tolerance limits .
( ii ) to provide an assessment of the deviation from the ideal world , where the machine components move smoothly and rigidly in exact circular orbits ; these effects have appeared in the literature as ` mechanical flex ' , ` nonidealities ' , ` nonrigid motion of the system components ' , etc .
( iii ) to enable the investigation of the long - term mechanical stability of the system ; this examination involves the creation of a database containing the output of each calibration . on top of the arguments which were just put forth concerning the necessity of the geometric calibration separately for imaging and delivery units ,
there is one additional remark which specifically applies to igrt systems : the geometric calibration links together the two units comprising the machine . with the relationship between the two components of the system having been set ,
the updated anatomic information ( obtained from the imaging unit ) may be quickly processed and the dose distribution re - evaluated in the area which is subsequently subjected to radiation ( delivery unit ) ; thus , modifications in the treatment plan are enabled on the daily basis , reflecting the most up - to - date information in the region of interest . concerning earlier methods pertaining to the calibration of the geometry in computed tomography , an overview may be obtained from the paper of noo ( 2000 ) ; that work set forth an approach to extract the parameters of the geometry from the elliptical projections of fixed points , and inspired additional research in the field .
fahrig and holdsworth ( 2000 ) employed a small steel ball bearing ( bb ) placed at the isocentre , and traced its projection across a series of images ; corrections to the scan data were subsequently calculated ( as a function of the gantry angle ) from the movement of the image centroid on the detector .
a similar approach was followed by jaffray ( 2002 ) , whereas mitschke and navab ( 2003 ) developed a method featuring a ccd camera attached to the head of the x - ray source .
siewerdsen ( 2005 ) use a phantom consisting of a helical pattern of bbs ; the data analysis results in the creation of their ` flex maps ' , which then lead to the assessment of the mechanical nonidealities .
finally , the interesting paper of cho ( 2005 ) introduces a phantom consisting of an arrangement of @xmath0 steel bbs in two circular patterns and achieves the description of the geometry of the system via a set of spatial lengths and distortion angles . in this paper , we present one of the methods which are currently in use in the geometric calibration of the vms devices ; a shorter description of this image - based calibration technique has appeared in matsinos ( 2005 ) .
contrary to the bulky cylinders which are generally needed in other schemes in order to extract the parameters of the geometry , we use an easy - to - handle ( i.e. , to mount and dismount ) and light needle phantom which , additionally , can easily be manufactured .
the extraction of the values of the model parameters is achieved via a method which has been developed with attention to the robustness of the final outcome .
the needle phantom comprises five cylindrical @xmath1-mm - long metallic needles ( @xmath2 @xmath3 mm ) embedded in a urethane compound ( obomodulan 500 , obo - werke gmbh & co. kg , stadthagen , germany ) of cylindrical form .
the dimensions of the urethane housing are : @xmath4 mm ( diameter ) and @xmath5 mm ( height ) .
the weight of the needle phantom is about @xmath6 gr .
the needles are made of a chromium - nickel alloy ( 1.4305 , edelstahlwerke sdwestfalen gmbh , siegen , germany ) and are coplanar , parallel and equidistant ( @xmath7-mm separation ) ; the axis of the central needle coincides with the symmetry axis of the needle phantom .
marks have been incised into its surface to enable the proper alignment with a laser system .
the needle phantom may easily be mounted directly onto the vms couch ( exact couch ) ; it is placed in such a way that the needles are parallel to the rotation axis .
the vms paxscan 4030cb amorphous - silicon flat - panel detector , currently used in the data acquisition , is a real - time digital x - ray imaging device comprising @xmath8 square elements ( pixels ) ; the detector spans an approximate area of @xmath9 @xmath10 . in order to expedite the data transfer and processing , the so - called half - resolution ( @xmath11-binning ) mode is used in almost all applications
; thus , the detector is assumed to consist of @xmath12 ( logical ) pixels ( pitch : @xmath13 @xmath14 m ) .
the detector is connected to the body of the system via a set of robotic arms enabling three - dimensional ( 3d ) movement .
the experimental data which are analysed in the present paper were obtained at the vms laboratory in baden , switzerland , and involved two vms devices : the acuity simulator ( as ) , a machine dedicated to imaging ( figure [ fig : placementatisocentre ] ) , and the ` on - board imager ' ( obi ) system , the imaging unit of the vms igrt devices . the description of these two units may be obtained directly from the website of the manufacturer ( ` www.varian.com ' ) . in both devices ,
the x - ray pulses are produced by an x - ray tube , the vms model g242 .
the position of the x - ray source is fixed in the as device , but adjustable ( 2d movement ) in the obi unit . to correct for the ( large ) anisotropy in the radiation field ,
wedge filters have been mounted at the head of the gantry in the as device ; there are no such filters in the obi unit .
the ideal geometry and placement of the needle phantom are shown in figure [ fig : placementatisocentre ] , where @xmath15 and @xmath16 denote the source - isocentre ( often referred to as sad ) and isocentre - detector distances ; @xmath17 is the gantry angle , identifying the position of the source . in this ideal world , the isocentre , defined as the intersection of the rotation axis and the plane on which the x - ray - source locus s lies during a scan , is a fixed point in space ; a point source is assumed ( cone - beam geometry ) , circumscribing an exact circle .
the projection of the isocentre onto the detector ( point m ) coincides with the geometric centre of the detector ; ideally , the line sm is perpendicular to the detector surface . concerning the perfect orientation of the detector , two of its sides are paraller to the rotation plane ( and two perpendicular to it ) .
in the isocentre reference frame , a point will be represented as ( @xmath18,@xmath19,@xmath20 ) ; facing the gantry , the @xmath18 axis points to the right , the @xmath19 axis towards the gantry and the @xmath20 axis upwards .
the second coordinate system pertains to the machine .
the coordinates of points in this reference frame will carry the subscript @xmath21 , thus , a point will be represented as ( @xmath22,@xmath23,@xmath24 ) .
the origin of the machine reference frame is the isocentre ; additionally , @xmath25 .
the isocentre and machine reference frames are related via a simple rotation around the @xmath19 axis , the rotation angle being denoted as @xmath26 ( figure [ fig : coordinatesystems]a ) ; in the ideal case ( @xmath27 ) , the machine and isocentre reference frames are identical .
the third coordinate system is attached to the needle phantom .
the coordinates of points in this reference frame will carry the subscript @xmath28 , thus , a point will be represented as ( @xmath29,@xmath30,@xmath31 ) .
the couch angle @xmath32 will relate this reference frame and an auxiliary one ( figure [ fig : coordinatesystems]b ) , denoted as ( @xmath33,@xmath34,@xmath35 ) , having the same origin ( the geometric centre of the needle phantom ) , but being parallel to the isocentre reference frame ( @xmath18,@xmath19,@xmath20 ) . from figure
[ fig : coordinatesystems]b , it is evident that @xmath36 the auxiliary and isocentre reference frames are parallel to one another . denoting the isocentre coordinates in the auxiliary reference frame as @xmath37 , @xmath38 and @xmath39 , one obtains @xmath40 from figure [ fig : coordinatesystems]a , it is deduced that @xmath41 these three transformations relate the sets of coordinates of a point in the needle - phantom and machine reference frames .
two parameters ( @xmath32 and @xmath26 ) have been introduced to account for the miscalibration of the couch and gantry angles .
additionally , the parameter @xmath26 takes account of the movement of the source between the time instant corresponding to the retrieval of the gantry - angle value from the system and an average time associated with the actual exposure . including also the quantities @xmath37 , @xmath38 and @xmath39 , the transformation of the coordinates ( from the needle - phantom to the machine reference frame ) involves five parameters in total .
the ideal geometry has already been shown in figure [ fig : placementatisocentre ] ; we will now introduce two deviations from this hypothetical situation _ on the rotation plane_. first , the straight line drawn from the isocentre perpendicular to the detector might not contain the source ; in figure [ fig : deviationsonplane ] , this imperfection is represented by the angle @xmath42 ( positive clockwise ) .
second , the projection m@xmath43 of the isocentre onto the detector might not coincide with the middle of the segment defined by the intersection of the detector and the rotation plane ; the lateral ( tangential to the circular orbit ) displacement m@xmath43 m of the detector will be denoted as @xmath44 ( positive to the right as one faces the gantry at @xmath45 ) .
starting from a point p(@xmath22,@xmath24 ) on the rotation plane and introducing its polar coordinates as @xmath46 and @xmath47 , the projected length @xmath48 onto the detector is given by the formula @xmath49 as the geometric centre of the detector is the origin of the detector reference frame , the trace q corresponds to a lateral reading equivalent to @xmath50 .
the projected length @xmath51 onto the detector in the longitudinal direction ( parallel to the rotation axis ) for an abritrary point ( @xmath22,@xmath23,@xmath24 ) is given by the formula @xmath52 in analogy to the lateral direction , one must introduce a parameter representing the longitudinal displacement of the detector ; this parameter will be denoted as @xmath53 .
thus , the trace p@xmath43 corresponds to a longitudinal reading equivalent to the length @xmath54 .
the deviation in the orientation of the detector from the ideal geometry is a source of systematic effects .
this misorientation may easily be described in terms of three rotations around the principal axes of the detector , corresponding to the lateral and longitudinal directions , as well as to the one which is perpendicular to its surface . due to the fact that some of our parameters show sensitivity to the last rotation ( i.e. , around the detector normal )
, we will introduce one parameter ( @xmath55 ) to account for this degree of freedom ; concerning the two former rotations ( i.e. , around the lateral and longitudinal directions ) , our results show no sensitivity , at least up to the level at which these distortions are present in the vms devices which were calibrated .
on the contrary , the inclusion of the parameter @xmath55 leads to an improved description of the longitudinal residuals ; we will address this point later on . denoting the coordinates of a trace on the detector as @xmath56 and @xmath57 in the nonrotated reference frame ,
the coordinates @xmath58 and @xmath59 in the rotated reference frame are obtained via the tranformation @xmath60 equations ( [ eq : mprimeq])-([eq : detectorrotation ] ) establish a relationship between the coordinates @xmath22 , @xmath23 and @xmath24 of an arbitrary point and its corresponding trace on the detector ; the trace is obtained from ( [ eq : mprimeq ] ) and ( [ eq : qpprime ] ) after involving the detector displacement ( that is , the parameters @xmath44 and @xmath53 ) and the transformation ( [ eq : detectorrotation ] ) on the detector plane . by using also the transformations ( [ eq : relationbetweenfandt])-([eq : relationbetweenmandi ] ) , one may determine the projection of an arbitrary point onto the detector plane starting from its coordinates @xmath29 , @xmath30 and @xmath31 in the needle - phantom reference frame .
a number of parameters have been introduced in three steps : five parameters ( @xmath32 , @xmath26 , @xmath37 , @xmath38 and @xmath39 ) are associated with the transformation of the coordinates from the needle - phantom to the machine reference frame ; another five parameters ( @xmath15 , @xmath16 , @xmath44 , @xmath53 and @xmath42 ) pertain to the geometry in the machine reference frame ; finally , the parameter @xmath55 describes the orientation of the detector .
all these parameters , save for @xmath15 and @xmath16 , vanish in an ideal world , devoid of mechanical imperfection and inaccurate calibrations .
the input data to the geometric calibration comprise one complete scan of the needle phantom .
the traces corresponding to the end - points of the needles are identified in the acquired images ; these two coordinates ( for each needle , in each input image ) represent the ` experimental ' values . for a given set of parameter values , the projections of the needle end - points onto the detector plane
are calculated using the chain of equations ( [ eq : relationbetweenfandt])-([eq : detectorrotation ] ) ; these calculated values comprise the ` predictions ' .
an optimisation scheme is set , by varying the model parameters and seeking the best description of the input data ; a standard @xmath61 function is minimised .
we will now touch upon four issues which we consider important in our work .
an image of the needle phantom is shown in figure [ fig : needlephantomprojection ] .
the detection of the traces of the needle end - points is done in two steps . to suppress noise , the average signal along the detector midline ( along the @xmath62 axis ) is created by averaging the contents of @xmath63 pixels in the @xmath64 direction , @xmath65 on either side of the midline . to remove the urethane background reliably , a fixed - width running window
is applied to this average - signal data , creating the difference of the integrated signal to the ( linear ) background ( defined by the pixel values at the window limits ) .
when the window covers places where only the urethane housing is projected , the values of the transformed signal are nearly @xmath66 .
as the window approaches the projected axis of a needle , the transformed signal first attains a positive ( local ) maximum , then a negative ( local ) minimum ( at the position corresponding to maximum attenuation ) ; this signature is easily identified via simple software cuts .
the algorithm is very efficient in detecting the signals corresponding to the needles , save for gantry - angle values around @xmath67 where the projected signals ( of the needles ) overlap .
after the signal modes ( the peaks in the transformed - signal spectrum ) are assigned to the needles , the projection of each needle is followed towards the couch ( along the negative @xmath64 direction in figure [ fig : needlephantomprojection ] ) .
two signal levels are identified : one corresponding to the projection of the needle , one to the urethane background . the projection of the needle end - point is assumed to correspond to the position where the signal is equal to the geometric mean of these two values .
two coordinates ( @xmath58 and @xmath59 ) are thus determined for each needle in each input image . if more or fewer than five signals are detected in the lateral direction , the entire image is rejected . if a needle axis was found , but the identification of the end - point failed , the information relating to the particular needle is removed from the contribution of the current image to the database . to decrease the correlations among the model parameters in the fits ,
it was decided to fix the distance @xmath15 to the nominal value corresponding to the unit being calibrated .
for one thing , the description of the data with variable @xmath15 hardly improves ; for another , if @xmath15 and @xmath16 are both treated as free parameters , the largeness of their correlation might result in cases where the fit ` drifts ' .
due to the different sensitivity of the two directions ( @xmath62 and @xmath64 in figure [ fig : needlephantomprojection ] ) to the model parameters , the fit in the lateral direction ( @xmath62 coordinates ) is performed first ; this fit achieves the extraction of the values of the parameters @xmath26 , @xmath37 and @xmath39 ( associated with the relationship among the various coordinate systems ) , and @xmath16 , @xmath44 and @xmath42 ( associated with the geometry in the machine reference frame ) .
the remaining parameters are determined from the fit to the @xmath64 coordinates , which is performed with the values of the aforementioned six parameters fixed from the @xmath62-direction fit . for small @xmath55 ,
the movement of a trace in the @xmath64 direction is always small , whereas in the @xmath62 direction it may be small or large , depending on the position of the point being projected ; this is one of the reasons for the larger uncertainties in the determination of @xmath38 and @xmath53 ( compared to @xmath37 and @xmath44 ) , the other one being their strong correlation .
the minuit package of the cern library , see james ( 1994 ) , has been invariably used in the present paper .
all output uncertainties contain the birge factor @xmath68 , which adjusts the output uncertainties for the goodness of each fit ; ndf denotes the number of degrees of freedom .
the output shows no sensitivity to the values of the model parameters which are used in the first iteration of the optimisation scheme .
the dependence of the parameters of our model on the gantry angle can only indirectly be assessed ; in our scheme , their values are assumed constant within one scan .
if present at a significant level , a departure from constancy will manifest itself in the creation of large residual effects ; the point will be discussed later on . experience has shown that the presence of noise in the input data might lead to the extraction of erroneous signal modes in the transformed - signal spectrum ; this failure rate never exceeded the @xmath69 level .
however , to safeguard against such cases , it was decided to precede the main optimisation by one performed on the data of each needle _ separately _ ; this approach is more efficient in detecting and excluding the outliers .
since the results obtained with logarithmic forms of the objective function are considerably less sensitive to the presence of outliers ( compared with those of the standard @xmath61 optimisation ) , the minimisation of a simple logarithmic form has been implemented in this part of the analysis .
the outliers are identified via a software cut corresponding to a @xmath70 effect for the normal distribution ; other values ( e.g. , @xmath3 and @xmath71 cuts ) have also been used , resulting in tiny differences in the numbers quoted here .
the residuals are defined as the differences between the experimental values and the corresponding predictions obtained when using the optimal values of the model parameters .
there are three reasons why the residuals are not identically zero : ( i ) the statistical fluctuation ( random noise ) which is always present in measurements , ( ii ) the inclusion of erroneous data in the input database and ( iii ) the use of an insufficient ( incomplete ) model in the description of the measurements .
concerning potential sources of systematic effects in the present work , one may recall some assumptions on which our model is based ; for instance , it may be that ( within one scan ) the rotation axis is not constant , the movement of the x - ray source is irregular , the connecting arms are distorted under the weight of the gantry and/or of the detector , etc .
additional effects may have been introduced by approximations assumed in the geometry of the systems ; for example , it could be that additional degrees of freedom are needed in describing the orientation of the detector . in any case ,
given the smallness of these effects ( which will presently be shown ) , the question one has to answer is whether the complete description of the observations in terms of a model is called for , or it is adequate to make use of a simple model , grasp the main features of the geometry and attempt the empirical description of the residual effects ; a number of reasons compel us to adopt the latter strategy .
examples of the residuals in the lateral and longitudinal directions are shown in figures [ fig : residuals ] . in both directions , when plotted versus the gantry angle separately for each needle , the residual distributions overlap ; therefore , two average numbers ( i.e. , one per direction ) may be used in each image . to demonstrate the reproducibility of the effect ,
the results of five scans taken in identical conditions are shown .
two conclusions can be drawn from figures [ fig : residuals ] : ( i ) the residual effects are small , the variation being of the order of one pixel size of the detector in the lateral and two in the longitudinal direction and ( ii ) the residual effects are systematic , hence they can be modelled
. the description of the lateral ( @xmath72 ) and longitudinal ( @xmath73 ) residuals will be attempted by using the empirical formulae : @xmath74 and @xmath75
the experimental data were acquired on january 31 , 2006 .
the validity of the conclusions drawn from those data was confirmed by an analysis of additional measurements obtained on april 26 , 2006 ; the results of the analysis of the april data will not be given here .
the ( adjustable ) distance @xmath16 was set to @xmath76 mm both in the as device and the obi unit . the nominal @xmath15 value in the
as is @xmath77 mm ; @xmath15 was set to @xmath77 mm in the obi . on each machine , the following steps were taken .
the voltage of the x - ray tube was set ( @xmath78 kv in the as , @xmath5 kv in the obi ) .
the dark- and flood - field calibrations were performed , so that flat images may be obtained in open - field geometry .
the needle phantom was placed close to the ideal geometry of figure [ fig : placementatisocentre ] .
the acquisition settings ( x - ray - tube current and pulse width ) were chosen in such a way as to yield a good - quality signal on the detector .
the frame rate was set to @xmath65 images per second on both machines , thus resulting in about @xmath79 images on the as device and @xmath80 on the obi unit ( the rotation of the clinac is slower ) . to investigate the short - term reproducibility of the results of the geometric calibration
, five counterclockwise ( ccw ) and five clockwise ( cw ) scans were acquired in identical geometry ( except for the rotation of the gantry , no other movement of the system components was allowed ) on each of the two machines .
both series started with a ccw scan ; successive scans were taken in opposite directions .
it is noteworthy that , save for a trivial difference in the values of parameter @xmath26 , no other effect ( between ccw and cw scans ) is expected in systems behaving identically in the two rotation modes .
we have already mentioned that @xmath26 takes account of two effects : ( i ) the miscalibration of the gantry angle and ( ii ) the movement of the source during the time interval from the retrieval of the gantry - angle value ( from the system ) to the instant associated with the ` average ' of the actual exposure ; as the gantry - angle value is obtained a few micro - seconds after the ` beam - on ' condition ( i.e. , at the beginning of the emission of the radiation pulse ) , the time delay of case ( ii ) above may be thought of as being equal to half the acquisition setting for the pulse width .
since @xmath81-msec pulses were used , the difference in the @xmath26 values between ccw and cw scans is expected to be about @xmath82 in the as device and @xmath83 in the obi unit .
the average of the two @xmath26 values between ccw and cw scans provides an estimate of the miscalibration of the gantry angle .
our results for the optimal values of the model parameters are shown in table [ tab:1-table ] , separately for ccw and cw scans .
inspection of this table leads to the following conclusions . *
the values of the model parameters come out reasonable .
the values of the isocentre - detector distance @xmath16 are almost identical in ccw and cw scans and close to the expectation value .
the values of the distortion angles are small . *
the variation in the values of the model parameters within each rotation mode is , in most cases , smaller than the corresponding statistical uncertainty . *
the @xmath26 and @xmath42 values come out different in the two rotation modes in the as device .
we will prove , however , that these differences are the product of the strong correlation between these two parameters .
concerning the obi data , the difference in the @xmath26 values may entirely be attributed to the delay time . * the largest difference between the two rotation modes in the as device relates to the description of the residuals in the longitudinal direction .
the two amplitudes @xmath84 of equation ( [ eq : residualsy ] ) come out reasonably close , but the phase shifts @xmath85 do not .
this discrepancy has been systematic for a long period of time and has been verified using other techniques ; it is _ not _ a result of strong correlations among the model parameters . due to the lack of obvious structure in the distribution of the residuals in the lateral direction in the as device
( some structure was observed in part only of each scan ) , it was decided not to attempt a fit using equation ( [ eq : residualsx ] ) there .
the difference in the @xmath85 values for the obi unit is statistically significant , but not so pronounced as in the case of the as device . *
the remaining differences are small in absolute value ; for instance , the effect in @xmath37 , the largest one observed , corresponds to one - ninth of the pixel size of the detector in the as device and between one - sixth and one - seventh in the obi unit .
notwithstanding the smallness of the mismatch , it is evident that the systems do not behave identically in the two rotation modes , suggesting the introduction of two sets of corrections to the scan data .
the issue of the correlation among the model parameters has to be properly addressed .
the parameters @xmath26 and @xmath42 are strongly correlated in the fit to the @xmath62 coordinates ( of the traces of the needle end - points ) , whereas @xmath38 and @xmath53 are correlated in the @xmath64-direction fits ; the remaining elements of the hessian matrix are ( in absolute value ) smaller than @xmath86 , indicating insignificant correlations . due to the fact that the differences in the values of the parameters @xmath38 and @xmath53 in the two rotation modes are not statistically significant and , additionally ,
these two parameters are not correlated with any other , we will investigate the correlation only between the parameters @xmath26 and @xmath42 . from the theoretical point of view
, these two parameters are independent ; the distortion angle @xmath42 defines the offset of the source with respect to the detector in the machine reference frame , whereas @xmath26 , pertaining to the relationship between the isocentre and machine reference frames , applies equally to all machine components . in reality however , possibly coupled with the smallness of these distortion angles , a strong correlation between @xmath26 and @xmath42 has been observed . to investigate whether the differences between the two rotation modes may be attributed to the correlation between these two parameters , we analysed the measurements acquired in the as device further , after fixing @xmath42 at @xmath87 ; the results for the model parameters are given in table [ tab:2-table ] .
we found that the description of the data with fixed @xmath42 is as good as when it is allowed to vary freely .
the changes induced by fixing @xmath42 are entirely absorbed by parameter @xmath26 , the values of which turn out now to be in good agreement with the expectation , based on the movement of the source within the delay time ( ` beam - on ' condition to average exposure ) .
the values of all other parameters are almost intact .
our conclusion is that the correlation between the model parameters @xmath26 and @xmath42 affects only the values of these two parameters .
a good measure of the mechanical stability of a system may be obtained from the unexplained variation in the data , that is , from whichever fluctuation survives after all model contributions have been deducted .
as previously mentioned , in the case of the as runs , the unexplained variation in the lateral direction is represented by the entire fluctuation contained in these residuals , whereas in the longitudinal direction the empirical formula ( [ eq : residualsy ] ) is assumed to be part of the model .
for the ccw scans , the rms of the unexplained variation in the lateral direction is equal to @xmath88 @xmath14 m ; for the cw scans , it is @xmath5 @xmath14 m
. therefore , the ccw scans seem to be somewhat smoother in the as device .
the two values of the rms of the unexplained variation in the longitudinal direction come out almost identical : @xmath89 @xmath14 m . as the residuals in both directions
are structured in the case of the obi unit , they have been fitted to via equations ( [ eq : residualsx ] ) and ( [ eq : residualsy ] ) .
figures [ fig : residuals ] correspond to the five ccw scans in the obi unit . in case of the ccw scans ,
the rms of the unexplained variation in the lateral direction is equal to @xmath90 @xmath14 m ; in the cw scans , it is @xmath91 @xmath14 m .
the corresponding numbers , assuming no modelling of the lateral residuals , are @xmath92 and @xmath93 @xmath14 m ; therefore , the unexplained variation in the lateral direction drops significantly when involving the empirical modelling of these residuals . in both rotation modes ,
the rms values of the unexplained variation in the longitudinal direction come out identical : @xmath94 @xmath14 m .
comparing the results obtained in the as device with those extracted from the obi unit , one notices that the as scans are less noisy in the longitudinal direction .
the values of the unexplained variation in the data in the longitudinal direction come out the same for ccw and cw scans , smaller than one - fifth of the pixel size of the detector in the as device , one - third in the obi unit . in the lateral direction ,
the unexplained variation is smaller than one - sixth of the pixel size of the detector in both systems .
it is interesting to note that , after equation ( [ eq : residualsx ] ) has been invoked in the description of the residuals in the lateral direction , the ccw scans on obi correspond to a significantly smaller value of the unexplained variation ; a similar effect is observed in the as device , where equation ( [ eq : residualsx ] ) was not used . in any case , the description of the data has been achieved at the sub - pixel level in both systems which were calibrated .
the aim of the geometric calibration of cone - beam imaging and delivery systems is threefold : to yield the values of important parameters in relation to the geometry of the system , to provide an assessment of the deviation from the ideal world ( where the machine components move smoothly and rigidly in exact circular orbits as the gantry rotates ) and to enable the investigation of the long - term mechanical stability of the system .
the method described here applies to devices where an x - ray source and a flat - panel detector ( facing each other ) move in circular orbits around the irradiated object .
contrary to the bulky cylinders , which are generally needed in other works in order to extract the parameters of the geometry , we introduce a light needle phantom which is easy to manufacture .
a model has been set up to describe the geometry and the mechanical imperfections of the system being calibrated .
the model contains five parameters associated with the transformation of the coordinates from the needle - phantom to the machine reference frame ; another five parameters account for the geometry in the machine reference frame ; finally , one parameter is introduced to account for the deviation in the orientation of the detector from the ideal geometry . to avoid strong correlations among the important parameters of the model , the source - isocentre distance
is set to the nominal value of the device being calibrated .
the input data comprise one complete scan of the needle phantom .
the end - points of the needles are identified in the acquired images and comprise the ` experimental ' values
. a robust optimisation scheme has been put forth to enable the extraction of the model parameters from the entirety of the input data . the application of the approach to two sets of five counterclockwise ( ccw ) and five clockwise ( cw ) scans , acquired in two imaging devices manufactured by varian medical systems , inc .
, yielded consistent and reproducible results .
the values of the model parameters come out reasonable .
the description of the data has been achieved at the sub - pixel level .
a number of differences have been seen between ccw and cw scans , suggesting that the devices do not behave identically in the two rotation modes ; we are not aware of other papers which have investigated and reported this effect . as a indispensable part of our project , we have introduced and implemented a calibration scheme in which different parameter sets apply to the scan data depending on the rotation mode used .
we would like to draw attention to this effect , since the differences , albeit small , are systematic .
mitschke m and navab n 2003 recovering the x - ray projection geometry for three - dimensional tomographic reconstruction with additional sensors : attached camera versus external navigation system _ med .
* 7 * 65 - 78 noo f ,
clackdoyle r , mennessier c , white t a and roney t j 2000 analytic method based on identification of ellipse parameters for scanner calibration in cone - beam tomography _ phys .
biol . _ * 45 * 3489 - 508 siewerdsen j h , moseley d j , burch s , bisland s k , bogaards a , wilson b c and jaffray d a 2005 volume ct with a flat - panel detector on a mobile , isocentric c - arm : pre - clinical investigation in guidance of minimally invasive surgery _ med .
phys . _ * 32 * 241 - 54 the optimal values of the model parameters in the vms acuity - simulator ( as ) and ` on - board imager ' ( obi ) units .
the values quoted represent averages of five counterclockwise ( ccw ) and five clockwise ( cw ) scans .
all lengths are in mm , all angles in degrees .
the first uncertainties are systematic , the second statistical .
left : the vms acuity - simulator device ( vms laboratory , baden , switzerland ) .
the x - ray source is at a position corresponding to about @xmath95 .
right : a schematic view of the various elements of a system on the rotation plane ( facing the gantry , as in the figure on the left ) .
shown is the placement of the needle phantom in the ideal geometry ; the axis of the central needle coincides with the rotation axis . @xmath15 and
@xmath16 denote the source - isocentre and isocentre - detector distances.,title="fig:",width=283 ] left : the vms acuity - simulator device ( vms laboratory , baden , switzerland ) .
the x - ray source is at a position corresponding to about @xmath95 .
right : a schematic view of the various elements of a system on the rotation plane ( facing the gantry , as in the figure on the left ) .
shown is the placement of the needle phantom in the ideal geometry ; the axis of the central needle coincides with the rotation axis . @xmath15 and @xmath16
denote the source - isocentre and isocentre - detector distances.,title="fig:",width=283 ] the coordinate systems used in the present paper .
a ) the isocentre ( subscript @xmath96 ) and the machine ( subscript @xmath21 ) reference frames ; they are related via a simple rotation ( angle @xmath26 ) .
b ) the needle - phantom ( subscript @xmath28 ) and the auxiliary ( subscript @xmath97 ) reference frames ; they are related via a simple rotation ( angle @xmath32 ) .
the auxiliary and isocentre reference frames are related via a simple translation involving the vector ( @xmath37,@xmath38,@xmath39).,width=302 ] machine reference frame : derivation of the projection of a point p ( on the rotation plane ) onto the detector ; the quantity @xmath46 represents the length ip .
two deviations from the ideal geometry ( on the rotation plane ) have been introduced : the angle @xmath42 and the lateral displacement of the detector @xmath44 .
the isocentre position is denoted by i.,width=283 ] an example of an image of the needle phantom .
the origin of the coordinate system shown coincides with the geometric centre of the detector ; ideally , @xmath64 is parallel to @xmath23 and @xmath62 to @xmath22 at @xmath45 .
the shift of the axis of the central needle to the right is due to the nonzero values of the parameters @xmath37 and @xmath44 .
a small tilt , hardly visible , is due to the nonzero values of the couch angle @xmath32 and of the detector orientation angle @xmath55.,height=377 ] the residuals in the lateral direction ( left ) and those in the longitudinal direction ( right ) versus the gantry angle @xmath17 for the vms ` on - board imager ' ( obi ) unit ; each point represents the average value over the needles whose end - points were successfully identified in the corresponding image . shown
are the results of five counterclockwise scans taken in identical conditions ; these scans are represented by different symbols.,title="fig:",height=283 ] the residuals in the lateral direction ( left ) and those in the longitudinal direction ( right ) versus the gantry angle @xmath17 for the vms ` on - board imager ' ( obi ) unit ; each point represents the average value over the needles whose end - points were successfully identified in the corresponding image . shown
are the results of five counterclockwise scans taken in identical conditions ; these scans are represented by different symbols.,title="fig:",height=283 ] | we propose a method to achieve the geometric calibration of cone - beam imaging and delivery systems in radiation therapy ; our approach applies to devices where an x - ray source and a flat - panel detector , facing each other , move in circular orbits around the irradiated object . in order to extract the parameters of the geometry from the data
, we use a light needle phantom which is easy to manufacture .
a model with ten free parameters ( spatial lengths and distortion angles ) has been put forth to describe the geometry and the mechanical imperfections of the units being calibrated ; a few additional parameters are introduced to account for residual effects ( small effects which lie beyond our model ) .
the values of the model parameters are determined from one complete scan of the needle phantom via a robust optimisation scheme .
the application of this method to two sets of five counterclockwise ( ccw ) and five clockwise ( cw ) scans yielded consistent and reproducible results .
a number of differences have been observed between the ccw and cw scans , suggesting a dissimilar behaviour of the devices calibrated in the two rotation modes .
the description of the geometry of the devices was achieved at the sub - pixel level .
+ _ keywords _ : geometric calibration + |
recently there is great interest in the study of the ultrahigh energy cosmic ray ( uhecr ) and the tev - photon paradoxes .
the first paradox is that some experiments @xcite-@xcite observed many hundreds of events with energies above @xmath1ev and about 20 events above @xmath2ev which are above the greisen - zatsepin - kuzmin ( gzk ) threshold . in principle , photopion production with the microwave background radiation photons should decrease the energies of these protons to the level below the corresponding threshold .
the second paradox is the fact that experiments detected 20tev photons from mrk 501 ( a bl lac object at a distance of 150mpc ) .
similarly to the first case , due to the interaction with the ir background photons , the 20tev photons should have disappeared in the ground - based detections .
these two puzzles have a common feature that both of them can be seen as some threshold anomalies : energy of an expected threshold is reached but the threshold is not observed .
there are numerous solutions such as @xcite proposed for the uhecr and the tev-@xmath3 paradoxes . in particular
, most authors @xcite-@xcite have suggested that lorentz - invariance violation can be the origin of these anomalies and have obtained many developments . in this paper
, we try to deal with these paradoxes by a modified dispersion relation which is obtained from @xmath0-deformed noncommutative theory .
it is well - known that noncommutative geometry @xcite plays an important role in the trans - planck physics , therefore , for phenomena related to the ultrahigh energy physics such as these two paradoxes , it is more reasonable to consider the lorentz - invariance violation is from noncommutative geometry and make some explanations for them in the background of noncommutative theory . the possibility has been mentioned that noncommutative geometry may be responsible for the ultrahigh energy cosmic ray and tev - photon paradoxes in some papers@xcite . here through a complete process based on a brand - new method which is from @xmath0-deformed noncommutative theory
, we will connect threshold anomalies with noncommutative theory to not only unravel these threshold anomalies but also present a strong support for noncommutative theory in turn .
our paper is organized as follows : we derive a general modified dispersion relation from @xmath0-deformed noncommutative theory in section 2 . in the following section we will apply this dispersion relation to the ultrahigh energy cosmic ray and the tev - photon paradoxes , and with the observed energies we get a common region of the ( @xmath4 , @xmath5 ) parameter space to resolve these two anomalies . in section 4
, we will give some implications and remarks for our work .
therefore , from @xcite , we can directly get : @xmath15_q-[\widehat{1-n}]_q\right).\ ] ] applying ( [ fer - n - p ] ) to ( [ fermi - h - n ] ) , we can obtain @xmath16_q-\left[\frac{1}{2}-\frac{\sqrt{m^2+p^2}}{\omega}\right]_q\right).\ ] ] expand the right hand side of ( [ q - fermi - dispersion ] ) and ( [ q - bose - dispersion ] ) in the neighborhood of @xmath23 to the second order , we derive the common approximate deformed dispersion relation for a fermion and a boson : @xmath24
now we will compute the @xmath0-deformed threshold for the uhecrs and the tev - photons .
we consider the head - on collision between a soft photon of energy @xmath25 , momentum @xmath26 and a high energy particle of energy @xmath27 , momentum @xmath28 , which leads to the production of two particles with energies @xmath29 , @xmath30 and momenta @xmath31 , @xmath32 respectively @xcite . from the energy conservation and momentum conservation
, we have : @xmath33 @xmath34 in the c. o. m. frame , @xmath35 and @xmath36 are rest at threshold , so they have the same velocity in the lab frame .
it s easy to give the following equation : @xmath37 applying our resulting dispersion relation ( [ q - dispersion ] ) and ( [ p3p2 ] ) to ( [ threshold1 ] ) and ( [ threshold2 ] ) , we can obtain the expression for @xmath38 in this case : @xmath39 in the course of deriving ( [ q - threshold ] ) , we have used the simplified form of ( [ q - dispersion ] ) : @xmath40 @xmath41 while in the non - deformed case , the threshold has the form as below : @xmath42 hereafter , we will denote the @xmath43 in the non - deformed case and in the @xmath0-deformed case by @xmath44 and @xmath38 separately . in the uhecr paradox @xmath45 , we set @xmath46 , in the other one @xmath47 , we have @xmath48 and similarly set @xmath49 .
inputting the values of @xmath43 in @xcite @xmath50 and @xmath51 to ( [ q - threshold ] ) and ( [ nq - threshold ] ) respectively , we can get two curves for the relation between @xmath4 and @xmath5 as plotted in fig .
1 ( the unit of @xmath4 is mev ) .
figure 1 : * the region of parameter space * the region of the @xmath52 parameter space provides a solution to both the uhecr and the tev-@xmath3 threshold anomalies .
while applying the @xmath0-deformed dispersion relation to the present observed anomalies , the relation between @xmath4 and @xmath5 can be described by the two curves in fig .
1 which give the lower bound on the @xmath0-deformation . in the region below both of the curves , the uhecr and the tev - photon paradoxes are resolved together .
the top left corner is excluded , which is easily understood for at a fixed @xmath4 the conventional dispersion relation can be reproduced if we take @xmath53 . under the consideration of @xmath54 mev , the parameters must be within the shaded area .
@xmath54 mev is set according to the smallness of @xmath5 , which is the requirement of the @xmath0-deformed noncommutative theory .
the relation between the two parameters @xmath4 and @xmath0 depends strongly on the experimental input . in this paper
we just made use of the highest energies of the ultrahigh energy cosmic rays and the tev - photons arriving on earth observed so far .
because both threshold anomalies of the uhecrs and the mrk 501 tev - photons are still under the situation that the data should be proceeded very cautiously , our resulting relation of the parameters is to be tested . if in the future better data prove the assumption of these threshold anomalies is incorrect , the explanation for them in terms of the @xmath0-deformed noncommutative theory has to be excluded , the same case as other interpretations such as the quantum - gravity - motivated models @xcite will encounter . if , however , the threshold anomalies are confirmed , the enhanced constraints on the available region of the parameter space can result from future cosmic - ray observations and experiments performed at very high engines . on the other hand , it has long been recognized that cosmology is a natural laboratory for testing the possible theories of extremely high energy physics such as string theory and noncommutative theory , and many authors @xcite have tried to search for some supports for these theories from the cosmological observations and experiments directly or indirectly .
one purpose of this paper is also following this train of thought , and if it can be verified that the permitted region of the parameter space to resolve the threshold anomalies is coincident with the parameter range from the noncommutative theory , it will be more confident for us to consider the noncommutative theory as the suitable theory for the trans - planck physics . 999 m. takeda _
et al_. , phys . rev
. lett . * 81 * , 1163 ( 1998 ) .
m. takeda _
et al_. , arxiv : astro - ph/9902239 . n. hayashida _
et al_. , phys .
lett . * 73 * , 3491 ( 1994 ) .
d. j. bird _
et al_. , astrophys .
j. * 441 * , 144 ( 1995 ) ; d. j. bird _ et al_. , phys .
lett . * 71 * , 3401 ( 1993 ) ; d. j. bird _ et al_. , astrophys .
j. * 424 * , 491 ( 1994 ) .
m. a. lawrence , r. j. o. reid and a. a. watson , j. phys .
g * 17 * , 773 ( 1991 ) .
n. n. efimov _
et al_. , ref .
proc . international symposium on _ astrophysical aspects of the most energetic cosmic rays _ , world scientific , singapore , pp .
20 , ( 1991 ) .
et al_. , hires collaboration , proceeds .
26th icrc ( salt lake 1999 ) .
g. amelino - camelia and t. piran , phys .
d * 64 * , 036005 ( 2001 ) .
j. m. carmona , j. l. cort@xmath55s , j. gamboa and f. m@xmath55ndez , arxiv : hep - th/0207158 . l. g. mestres , arxiv : physics/9704017 .
s. coleman and s. l. glashow , phys .
d * 59 * , 116008 ( 1999 ) .
r. aloisio , p. blasi , p. l. ghia , and a. f. grillo , phys .
d * 62 * , 053010 ( 2000 ) .
o. bertolami and c. s. carvalho , phys .
d * 61 * , 103002 ( 2000 ) .
h. sato , arxiv : astro - ph/0005218 .
t. kifune , astrophys .
j. * 518 * , l21 ( 1999 ) .
w. kluzniak , arxiv : astro - ph/9905308 .
r. j. protheroe and h. meyer , phys . lett .
b * 493 * , 1 ( 2000 ) .
a. connes , _ noncommutative geometry _
, academic press , london ( 1994 ) .
j. wess , int . j. mod
a * 12 * , 4997 ( 1997 ) .
g. amelino - camelia , int .
d * 11 * , 35 ( 2002 ) .
y. j. ng , j. phys .
a * 23 * , 1023 ( 1990 ) .
m. l. ge and g. su , j. phys . a * 24 * , l721 ( 1991 )
. m. a. m. delgado , j. phys .
a * 24 * , l1285 ( 1991 ) .
s. vokos and c. zachos , mod .
lett . a * 9 * , 1 ( 1994 ) . j. a. tuszy@xmath56ski , j. l. rubin , j. meyer , and m. kibler , phys . lett .
a * 175 * , 173 ( 1993 ) .
z. chang and s. x. chen , arxiv : cond - mat/0205208 .
y. j. ng , d. s. lee , m. c. oh , and h. van dam , phys . lett .
b * 507 * , 236 ( 2001 ) .
j. a. gu , p. m. ho and s. ramgoolam , arxiv : hep - th/0101058 . s. f. hassan and m. s. sloth , arxiv : hep - th/0204110 . c. s. chu , b. r. greene and g. shiu , mod .
a * 16 * , 2231 ( 2001 ) .
f. lizzi , g. mangano , g. miele , and m. peloso , jhep * 0206 * , 049 ( 2002 ) .
r. h. brandenberger and p. m. ho , phys .
d * 66 * , 023517 ( 2002 ) . | in this paper , we present a general modified dispersion relation derived from @xmath0-deformed noncommutative theory and apply it to the ultrahigh energy cosmic ray and the tev - photon paradoxes threshold anomalies .
our purpose is not only trying to solve these puzzles by noncommutative theory but also to support noncommutative theory through the coincidence of the region in the parameter space for resolving the threshold anomalies with the one from the @xmath0-deformed noncommutative theory .
22.5cm-0.2 in 15.5 cm bihep - th-2002 - 53 * noncommutativity as a possible origin of the ultrahigh energy cosmic ray and the tev - photon paradoxes * shao - xia chen and zhao - yu yang + _ institute of high energy physics , chinese academy of sciences _ + _
p.o.box 918(4 ) , 100039 beijing , china _
pacs numbers : 02.40.gh , 02.20.uw , 98.70.sa . |
the phenomenon of induced fermion number arises due to the interaction of fermions with nontrivial topological backgrounds ( e.g. , solitons , vortices , monopoles , skyrmions ) , and has many applications ranging from polymer physics to particle physics @xcite . at zero temperature ,
the induced fermion number is a topological quantity , and is related to the spectral asymmetry of the relevant dirac operator , which counts the difference between the number of positive and negative energy states in the fermion spectrum .
mathematical results , such as index theorems and levinson s theorem , imply that the zero temperature induced fermion number is determined by the asymptotic topological properties of the background fields @xcite .
this topological character of the induced fermion number is a key feature of its application in certain model field theories . at finite temperature ,
the situation is very different @xcite . in this talk
i explain why at finite t the induced fermion number is generically nontopological , and is not a sharp observable .
this work has also been partly motivated by recent results that certain anomalies in @xmath0 field theory become much more subtle at nonzero @xmath1 .
for example @xcite , in even dim .
spacetime , finite t anomalous @xmath2 decay amplitudes are t dependent even though the chiral anomaly ( a topological object that is related to the anomalous @xmath2 decay amplitude at @xmath0 ) is independent of @xmath1 . and , in odd dim .
spacetime , the topological chern - simons term is the only parity - violating term induced in the @xmath0 effective action , but at finite @xmath1 there are infinitely many parity - violating terms @xcite .
the induced fermion number is an expectation value of the fermion number operator , @xmath3 $ ] . for a given classical static background field configuration
, the fermion field operator @xmath4 can be expanded in a complete set of eigenstates of the dirac hamiltonian @xmath5 .
the @xmath0 fermion number , @xmath6 , is related to the spectral asymmetry of the dirac hamiltonian : @xmath7 here @xmath8 is the spectral function of @xmath5 : @xmath9 so , @xmath10 essentially counts the number of positive energy states minus the number of negative energy states . at nonzero temperature @xmath1 ,
the induced fermion number is a _ thermal _ expectation value @xmath11 where @xmath12 .
this finite t expression ( [ nt ] ) reduces smoothly to the zero t expression ( [ asymmetry ] ) as @xmath13 ; in fact , finite temperature provides a physically meaningful and mathematically elegant regularization of the spectral asymmetry .
note that the spectrum of the fermions in the static background has nothing to do with the temperature .
all information about the fermion spectrum is encoded in the spectral function .
thus , if one knows @xmath8 , one immediately has integral representations for both @xmath14 and @xmath15 .
in fact , one only needs to know the odd part of @xmath8 , as is clear from ( [ asymmetry ] ) and ( [ nt ] ) .
it is convenient to use the explicit form ( [ spectral ] ) of the spectral function to write ( [ nt ] ) as a contour integral involving the resolvent of the dirac hamiltonian @xmath16 where @xmath17 is the contour @xmath18 and @xmath19 in the complex energy plane .
thus the resolvent is the key to the computation .
before going to the general case , it is instructive to consider two important cases in 1 + 1 dim field theory .
these two examples are very similar at t=0 , but turn out to be very different for @xmath20 . consider an abelian model of fermions interacting via scalar and pseudoscalar couplings to two ( classical , static ) bosonic fields @xmath21 and @xmath22 : @xmath23 we distinguish between two different cases : \(a ) kink case @xcite : @xmath24 is constant , but @xmath25 has a kink - like shape ( for example @xmath26 ) , with @xmath27 .
\(b ) sigma - model case @xcite : @xmath28 and @xmath25 are constrained to the `` chiral circle '' : @xmath29 the kink case ( a ) is relevant to the polymer physics applications , while the sigma - model case ( b ) is a low - dimensional analogue of sigma models used in 3 + 1 dim .
particle physics models @xcite . in each case , we define the angular field @xmath30 \label{theta}\end{aligned}\ ] ] in the sigma model case ( b ) , the angular field @xmath31 has the interpretation of a local chiral angle , and is chosen to have a kink - like shape , with asymptotic values @xmath32 .
note that in the kink case ( a ) , the angular field @xmath33 $ ] also has a kink - like shape . for _ both _ the kink case and the sigma - model case , at zero t
, the induced fermion number is @xmath34 which is _ topological _ in the sense that it only depends on the asymptotic values of @xmath31 , not on its detailed shape . in the kink case ,
this calculation is easy because the even part of the corresponding dirac resolvent is known exactly @xcite @xmath35_{\rm e } = { - m \hat{\phi}_2 \over ( m^2-z^2)\,\sqrt{m^2+\hat{\phi}_2 ^ 2-z^2 } } \label{callias}\end{aligned}\ ] ] but in the sigma - model case the resolvent is not known exactly , and one must resort to an approximation such as the derivative expansion . nevertheless , both cases lead to the same result ( [ nt1 ] ) . at finite t ,
the kink case is also easy because we know the relevant part of the resolvent exactly @xcite . inserting ( [ callias ] ) into the contour integral expression ( [ contour ] ) one finds @xmath36 note that @xmath37 is temperature dependent , but it is still topological in the sense that it only depends on the background field through its asymptotic value @xmath38 $ ] .
also , even though the expression ( [ kt ] ) looks complicated , it smoothly reduces to ( [ nt1 ] ) as @xmath39 .
it is also interesting that the angular nature of @xmath40 is more obvious at finite t than at zero t. at finite t , the sigma - model case is much more complicated , because there is no simple closed formula for the relevant part of the resolvent . in @xcite
the derivative expansion was used , and it was shown that for @xmath41 the derivative expansion can be resummed to all orders , leading to the following simple expression @xmath42 the correction term is temperature dependent and nontopological , as it is sensitive not just to the asymptotic values of @xmath31 , but also to the details of the shape of @xmath31 .
this nontopological term has a simple physical interpretation as follows @xcite . for the sigma model case we can write the interaction term in ( [ lag ] ) as @xmath43 .
then make the local chiral rotation @xmath44 to obtain @xmath45 which describes a fermion of mass @xmath46 in an inhomogeneous electric field @xmath47 . in this language
the zero t induced fermion number ( [ nt1 ] ) can be understood as the vacuum polarization , due to the inhomogeneous electric field background , of virtual vacuum @xmath48 dipoles .
this effect is independent of the temperature as the virtual vacuum dipoles do not live long enough to thermalize .
however , at finite t there is another effect of the background electric field .
the real particles and antiparticles respond to this electric field in a way that can be described by linear response theory . in the derivative expansion limit , where @xmath49 , the leading effect of the electric field background can be represented as a local chemical potential @xmath50 . with local fermi distributions @xmath51 the linear response charge density
is @xmath52 , where @xmath53 . for @xmath41 , @xmath54 in perfect agreement with the non - topological term in ( [ st ] ) .
so , while at zero t , the kink and sigma - model cases each have induced fermion number given by the same expression ( [ nt1 ] ) , at nonzero t they have very different induced fermion number , given by ( [ kt ] ) and ( [ st ] ) respectively . in each case
@xmath55 is temperature dependent , but in the kink case it is topological , while in the sigma - model case it is nontopological . the physical reason for this difference is explained in the next section .
the straightforward mathematical reason why @xmath55 is topological for the kink case is that the odd part of the spectral function is itself topological , as shown by the callias result ( [ callias ] ) . but a much more physical understanding can be obtained by considering the symmetries of the fermionic spectrum @xcite .
and this perspective generalizes immediately to higher dimensional models .
begin with the simple trigonometric identity , @xmath56 , where @xmath57 is the fermi - dirac distribution function
. then the finite t fermion number ( [ nt ] ) separates as @xmath58 the first term is just the zero temperature contribution , which can be expressed in terms of the spectral asymmetry , and is known to be topological .
but the second term is highly sensitive to the details of the spectrum , because of the presence of the fermi weighting factor @xmath59 .
thus , the only way this term can be topological is if the odd part of @xmath8 is itself topological . in terms of the spectrum
, this can only happen if the spectrum is symmetric under @xmath60 .
remarkably , this does sometimes happen , but only for extremely special backgrounds .
one example of such a special background is the one - dim kink background discussed in the previous section . indeed , for any kink background ( with @xmath40 positive )
, the spectrum of the dirac hamiltonian always has the form shown in fig . 1 .
there are continuum states for @xmath61 .
in addition , there is always a bound state at @xmath62 .
there may or may not be additional bound states with @xmath63 .
but if these additional bound states are present , they necessarily occur in @xmath64 pairs , because of the quantum mechanical susy of the dirac hamiltonian in the kink case @xcite .
thus , the contributions of these paired bound states to the integral in ( [ separation ] ) cancel in pairs , while the unpaired bound state at @xmath62 leads to a contribution @xmath65 . the quantum mechanical susy of the dirac hamiltonian is likewise the key to the exact callias index theorem result @xcite for the odd part of the spectral function [ which is equivalent to knowing the even part of the resolvent in ( [ callias ] ) ] .
this means that the remaining integral over the continuum states can be expressed as an integral of the odd part of the spectral function , weighted by the fermi factor @xmath66 , over the positive continuum beginning at the threshold energy @xmath67 . putting all this together
, we obtain @xmath68 one can check that this is , in fact , the sommerfeld - watson transform of the summation expression found before in ( [ kt ] ) . on the other hand , for the sigma - model case
the fermion spectrum has no obvious symmetry . in this case , the dirac spectrum has the form shown in fig .
the only general thing we can say is that the continuum thresholds are at @xmath69 .
there may or may not be one or more bound states for @xmath70 .
the existence of ( and the precise location of ) bound states is highly sensitive to the actual _ shape _ of the @xmath31 . as there is no special symmetry in the spectrum ,
there is nothing general that can be said about the odd part of the spectral function .
nevertheless , the first term in ( [ separation ] ) , which is @xmath10 , yields a topological result since the integral can be written , using levinson s theorem , in terms of the phase shift at infinity , which is topological @xcite .
this application of levinson s theorem does nt work for the t dependent second term in ( [ separation ] ) because of the fermi factor .
this is another way to understand why the finite temperature induced fermion number is generically nontopological .
these arguments based on the symmetry of the fermion spectrum ( in the given background ) clearly generalize to any dimension of spacetime .
so , we conclude that the generic situation is that the finite t induced fermion number is non - topological ; but , if the background is such that the fermion spectrum is highly symmetric , then the induced fermion number may be topological even though it is temperature dependent .
fluctuations in the induced fermion number can be analyzed in an analogous manner @xcite : @xmath71 the trigonometric identity , @xmath72 , re - expresses this as @xmath73 which is simply a sum over the fermion spectrum of @xmath74 , as it should be for the thermal occupation of independent single - particle states @xcite . from ( [ fluc ] )
it is clear that the fluctuations are always non - topological , as they are sensitive to the details of the fermion spectrum .
another way to see this is to note that the expressions ( [ fluc1],[fluc ] ) require knowledge of the even part of the spectral function , which is not accessible to an index theorem .
so the fluctuations must typically be calculated approximately ; however , for a special class of kink backgrounds , @xmath75 , where @xmath76 is an integer , it is possible to compute @xmath77 exactly , and one sees the explicit ( nontopological ) dependence on the scale factor @xmath76 @xcite .
these expressions also make it clear why the fluctuations vanish at zero t , because the @xmath78 factors vanish exponentially fast as @xmath39 .
thus , the induced fermion number is a sharp observable at @xmath0 , in agreement with previous results @xcite , but at nonzero t it is not a sharp observable , simply because the thermal expectation value mixes states other just the ground state .
another example of non - sharp fractional fermion number ( but not in the context of temperature dependence ) has been discussed recently in the context of liquid helium bubbles @xcite .
to conclude , the finite temperature induced fermion number @xmath15 is generically nontopological , and is not a sharp observable .
the induced fermion number naturally splits into a topological temperature - independent piece that represents the effect of vacuum polarization on the dirac sea , and a nontopological temperature - dependent piece that represents the thermal population of the available states in the fermion spectrum , weighted with the appropriate fermi - dirac factors . as the temperature approaches zero ,
the nontopological terms vanish exponentially fast , and we regain smoothly the familiar topological results at zero temperature . but at nonzero temperature , the nontopological contribution is more sensitive to the details of the spectrum , and so is generically nontopological .
this argument applies in any dimension .
consider , for example , for a chiral @xmath79 sigma model with a skyrme background in 3 + 1 dim .
, @xmath80 where @xmath81 and @xmath82 , which is the analogue of the 1 + 1 dim .
sigma - model case analyzed in this paper . from numerical work
@xcite , the dirac energy spectrum of the fermions is not symmetric , and the energy of a possible bound state ( and indeed the number of bound states ) is highly sensitive to the details of the shape of the radial hedgehog field .
this shows that in this case the finite temperature induced fermion number is nontopological .
thus , for a ( large ) skyrme background in 3 + 1 dim . , the finite temperature fermion number is not simply the ( topological ) winding number of the skyrme field , but a much more complicated nontopological object .
g. dunne , k. lee and c. lu , phys .
* 78 * ( 1997 ) 3434 ; g. dunne , `` aspects of chern - simons theories '' , les houches 1998 , _ topological aspects of low dimensional systems _ , a. comtet et al ( eds ) , ( springer , 1999 ) . | the induced fractional fermion number at zero temperature is topological ( in the sense that it is only sensitive to global asymptotic properties of the background field ) , and is a sharp observable ( in the sense that it has vanishing rms fluctuations ) . in contrast , at finite temperature , it is shown to be generically nontopological , and not a sharp observable . |
the geometric brownian motion is the stochastic process described by the differential equation @xmath0 where @xmath1 is a wiener process and @xmath2 are constants describing the drift and the variance of the noise , respectively .
the solution can be written as @xmath3 geometric brownian motion is used for modelling many phenomena in a variety of contexts @xcite .
a prominent role is played in financial applications , where the distribution of returns can be approximated by a log - normal distribution @xcite , at least in specific regimes . for the computation of certain properties ,
it is necessary to compute the integral of @xmath4 over a time interval @xmath5= \int_0^t f(w_s , s ) ds.\ ] ] the evaluation of this functional is also involved in the solution of the geometric brownian motion with logistic corrections . in general , averages of the form @xmath6 ) \rangle = \sum_{k=0}^\infty a_k \langle f[w , t]^k\rangle \equiv \sum_{k=0}^\infty a_k r_k.\ ] ] are quite common .
the evaluation of averages of powers of the integrated exponential brownian motion , then , is instrumental for the computation of these observables .
detailed studies of this functional and of its powers @xcite are already available in the literature . in this paper , we will derive exact formulas for the evaluation of these integrals , under the assumption of the ito formulation for the wiener process .
similar results have been given in @xcite .
motivated by obtaining exact formulas for asian options , in @xcite yor obtained an exact formula in terms of polynomials for the following moments : @xmath7 using girsanov s theorem @xcite , one can derive a series of identities , in which the last is bougerol s formula @xmath8 where @xmath9 and @xmath10 in this work , we take a different route with the use of combinatorics .
we prove a recurrence relation for the integrals involved at the @xmath11-th order in terms of integrals at the @xmath12-th order , and after resummation , we get an identity in terms of a determinant .
the central quantity of interest in the present paper is given by the average over the wiener process @xmath13 : @xmath14 ^k \rangle .
\label{eq : cumulant}\ ] ] if we expand eq .
( [ eq : cumulant ] ) , we obtain @xmath15 ^k \rangle=\int_0^t d { \tilde t } _ k \cdots \int_0^t d { \tilde t } _ 1 \left\langle e^{\sum_{i=1}^k [ ( \mu-\frac{\sigma^2}{2}){\tilde t}_i+\sigma w_{{\tilde t}_i } ] } \right\rangle.\ ] ] we will use the following formula due to the properties of integrals with gaussian measure @xcite , and in which we assume that @xmath16 is of the ito type .
this implies @xmath17 by using this property , we can now prove the following fact : [ lem1 ] for the average over the wiener process @xmath13 of ito type , the following formula holds true : @xmath18 by a direct application of eq .
( [ eq : ito ] ) @xmath19 } d{\tilde t}_1 \cdots
d{\tilde t}_k \right\rangle \nonumber \\ & = & \int_0^t \cdots \int_0^t e^{\sum_{i=1}^k[(\mu-\frac{\sigma^2}{2 } ) { \tilde t}_i+ \frac{\sigma^2}{2 } \sum_j \text{min}({\tilde t}_i,{\tilde t}_j ) } dt_1 \cdots dt_k.\end{aligned}\ ] ] due to symmetry of integrand , we can order the integration variables as @xmath20 , obtaining @xmath21 } d{\tilde t}_1 \cdots d{\tilde t}_k , \end{aligned}\ ] ] whence : @xmath22 } d{\tilde t}_1 \cdots
d{\tilde t}_k \nonumber \\ & = & \gamma(k ) \int_0^t \int_0^{{\tilde t}_{k-1 } } \cdots \int_0^{{\tilde t}_{2 } } e^{\sum_{i=1}^k \mu { \tilde
t}_i+ \sigma^2 \sum_{i=1}^k ( k - i){\tilde t}_i ] } d{\tilde t}_1 \cdots d{\tilde t}_k.\end{aligned}\ ] ] after rearranging carefully the terms , we arrive at the final result : @xmath23 let us now expand further on eq .
( [ eq : formula ] ) .
it is convenient to first perform the rescaling @xmath24 .
then , by defining @xmath25 , we obtain @xmath26 therefore , the computation of reduces to the computation of @xmath27 a very similar formula had been obtained in @xcite ; the main difference between the treatment made there and our relies on what follows .
an important observation will enable us to evaluate these integrals exactly by means of combinatorics : it is possible in fact to prove the following result , which establishes a recursion relation among the @xmath28 . for the quantity @xmath29 , we have @xmath30 with @xmath31 .
let us write @xmath32 with @xmath33 .
integrating by parts , @xmath34 where @xmath35 finally , the identity @xmath36 gives the desired result .
( [ eq : recursion ] ) suggests the evaluation of the averages by means of combinatorial considerations .
indeed , the evaluation of the integral can proceed graphically , for any fixed order of the moment @xmath11 , as in fig .
[ fig : rec ] .
starting from the top and using the properties of the recurrence relation , at each order @xmath11 , we organize the recurrence on a binary tree . in fig .
[ fig : rec ] , each left branch will pull out a factor @xmath37 , where @xmath38 is the @xmath38 obtained from the previous order .
one has to consider the fact that however , in each right branch one pulls out a factor @xmath39 , and sums the two factors of @xmath38 s . to obtain the final formula , once the empty set has been reached , one multiplies the final term by all the factors in the branch .
-th moments of the exponential integrated gaussian process to the case @xmath40 . ]
we now give exact formulas for all the terms obtained from the recurrence . to lighten the notation
, we define @xmath41 and @xmath42 . by iterating the second term of the recurrence ,
we obtain @xmath43 therefore , @xmath44 [ lem2 ] let @xmath45 with @xmath46 .
then @xmath47 we proceed the proof by induction on @xmath11 .
we have @xmath48 , which holds for @xmath49 .
we assume that the claim holds for @xmath11 and let us prove it for @xmath50 . by the recurrence relation , we have @xmath51 by induction hypothesis , we have @xmath52 which completes the induction step .
if @xmath53 and @xmath54 , then eq . can be written as @xmath55 applying lemma [ lem2 ] , we obtain @xmath56 which leads to the following result . for all @xmath57 , @xmath58 is given by @xmath59 the above theorem gives @xmath60 hence , the coefficient of @xmath61 in @xmath62 is given by @xmath63 a cleaner formula can be obtained in terms of determinants . if we define @xmath42 for all @xmath57 , where we define @xmath64 .
if @xmath53 and @xmath54 , then eq . can be written as @xmath65 using theorem 4.20 in @xcite ( see also @xcite ) , we obtain the following result : for all @xmath57 , @xmath42 is given by @xmath66 where @xmath53 and @xmath54 .
this expression can be elucidated by means of an example .
consider the case @xmath67 and @xmath68 . in this case
, we have an exact formula for @xmath69 : @xmath70 \right ) = 1/6\,{\frac { \left ( { e^{\mu\,t}}-1 \right ) ^{3}}{{\mu}^{3 } { t}^{3}}},\ ] ] which provides the exact value obtained from the deterministic logistic equation .
the previous results are general enough to hold also for averages of the form @xmath71^n\rangle = e^{(\mu-\frac{\sigma^2}{2})t } \langle e^{\sigma w_t}f[w , t]^n\rangle.\ ] ] it is easy to see that eq .
( [ eq : det ] ) applies .
we just substitute @xmath72 and multiply by a factor : @xmath73 the above expression will be used in the following section as an application to averages in the logistic stochastic differential equation .
appendix [ sec : appendix ] contains the analytical values of the functions @xmath74 and @xmath75 for the cases @xmath76 .
it is immediate that , when @xmath77 , we have @xmath78 this facts implies that eq .
( [ eq : detgen ] ) is analytic in @xmath79 .
[ fig : plotsn ] is a plot of @xmath80 for @xmath81 , @xmath82 . and
@xmath83 provided in appendix [ sec : appendix ] as a function of @xmath84 for @xmath85 , @xmath86 for @xmath87.,title="fig : " ] and @xmath83 provided in appendix [ sec : appendix ] as a function of @xmath84 for @xmath85 , @xmath86 for @xmath87.,title="fig : " ]
as an application of the formula ( [ eq : detgen ] ) , we focus on the solution of the logistic stochastic differential equation motion , @xmath88 , \label{eq : diffeq}\ ] ] given by ( we follow @xcite ) @xmath89 we evaluate the average of the solution @xmath90 , @xmath91 and observe that it involves the moments of ( [ eq : detgen ] ) . in the limit @xmath92 , eq .
( [ eq : pertth ] ) reduces to the average of the geometric brownian motion .
we consider now truncations of the mean of @xmath90 at @xmath11-th order , @xmath93 , and compare the truncated solution to the one obtained numerically . in fig .
[ fig : simulations ] , we plot @xmath94 for @xmath95 obtained for @xmath81 , @xmath82 , @xmath96 by means of a stochastic euler method with @xmath97 , and the averages then obtained using monte carlo over @xmath98 samples .
we observe that the higher the order of the approximation , the closer we are to the solution obtained by monte carlo . , @xmath82 , @xmath96 , and with @xmath97 , solved numerically using a stochastic euler method and averaged over 1000 simulations ( solid red ) , versus the analytical solution obtained at @xmath11-th order , considering @xmath99 the perturbative parameter .
we can observe that at higher order we obtain a solution closer to the one simulated . ]
in this paper , we have presented new exact formulas for the moments of the integrated exponential brownian motion in terms of sums and determinants , and based on recent results obtained in @xcite .
we described a simple graphical method to evaluate them , based on a recurrence relation .
exact formulas were proved in @xcite in terms of polynomials . in this paper however , we have taken an alternative route based on combinatorics . after realizing that the mean can be evaluated exactly using the properties of gaussian integrals , and after observing that these moments feature a recurrence relation , we have shown that exact expressions can be obtained via a combinatorial argument .
these exact expressions were then observe to be equivalent to evaluating the determinant of a specific linear operator which depends on the order of the moment . to complete the presentation ,
we have applied the formulas to the exact solution of the logistic stochastic differential equation . there
, the evaluation of the ensemble expectation values of certain observables can be carried out with our method , via taylor expansion . in particular
, the comparison of the mean solution obtained by means of monte carlo simulations with our method shows that our formula permits to approximate to a higher precision some ensemble averages of properties of the solution of the stochastic differential equation .
f. c. would like to thank the london institute for mathematical sciences for hospitality .
s. s. is supported by the royal society and epsrc .
99 c. gardiner , `` stochastic methods '' , springer - verlag , berlin ( 2009 ) m. yor , `` exponential functionals of brownian motion and related processes '' , springer - verlag , berlin ( 2001 ) h. matsumoto , m. yor , probability surveys , vol . 2 ( 2005 ) , 312 - 347 h. matsumoto , m. yor , probability surveys , vol . 2 ( 2005 ) , 348 - 384 m. yor , adv .
( 1992 ) , 509 - 531 i. v. girsanov , theory of probability and its applications 5 ( 3 ) : 285 - 301 ( 1960 ) m. salmhofer , `` renormalization - an introduction '' , springer - verlag , berlin ( 1999 ) r. vein , p. dale , `` determinants and their applications in mathematical physics '' , springer - verlag , berlin ( 1999 ) m. janji , determinants and recurrence sequences , _ j. int
. seq . _ * 15 * ( 2012 ) p. e. kloeden , e. platen , numerical solution of stochastic differential equations springer science & business media ( 2013 ) f. caravelli , l. sindoni , f. caccioli , c. ududec , _ upcoming _ | we present new exact expressions for a class of moments for the geometric brownian motion , in terms of determinants , obtained using a recurrence relation and combinatorial arguments for the case of a ito s wiener process .
we then apply the obtained exact formulas to computing averages of the solution of the logistic stochastic differential equation via a series expansion , and compare the results to the solution obtained via monte carlo . |
in practical implementations of the kohn - sham density functional theory @xcite the exchange - correlation ( xc ) energy is usually described by a suitable explicit functional of the electron density and the parameters characterizing the density inhomogeneity . a very promising step to improve the accuracy of density functional calculations relies on the use of orbital - dependent density functionals , in which the xc energy is expressed as an explicit functional of kohn - sham ( ks )
orbitals @xcite . a rigorous approach to implement the orbital - dependent functionals within
the ks formalism is the optimized effective potential ( oep ) method @xcite , where the xc potential is described by a local multiplicative term and the total energy functional is orbital - dependent . by virtue of the hohenberg - kohn theorems @xcite ,
the oep solution is equivalent to the minimization of the total electronic energy with respect to the density @xcite . in the case of the exchange - only
i. e. , the hartree - fock ( hf ) energy functional , the corresponding oep method ( xoep ) was first formulated by sharp and horton @xcite and numerically solved in real space for atoms by talman and shadwick @xcite .
applications of real space xoep formalism to atoms , molecules and solids , thanks to the use of pseudopotentials , have been reported in the literature . a number of approximations to the xoep method , such as the method of krieger - li - iafrate @xcite , the local hf @xcite and the effective local potential @xcite methods have been developed .
the real space resolution of the xoep method has only been efficiently applied to highly - symmetric systems , such as spherically symmetric atoms and diatomic molecules @xcite .
application of the xoep formalism to polyatomic molecules requires its formulation in terms of basis sets suitable for molecular calculations .
currently , there exist several formulations of the xoep method in terms of basis sets of local gaussian - type ( gto ) functions @xcite
. the most popular implementation of the xoep formalism employs two different basis sets , one for the expansion of the ks orbitals and another one for representing the local multiplicative potential @xcite . within this approach
, a special care must be taken when selecting the auxiliary basis set for the potential , thus leading to the concept of a balanced basis set firmly connected to the orbital basis set @xcite .
alternatively , a set of the products of the occupied and virtual ks orbitals can be employed for the solution of the xoep equations for the local potential @xcite .
the computational complexity of the xoep method in a basis set representation depends critically on the size of the orbital expansion basis set . for obtaining faithful solutions of the xoep equations , the orbital basis set should support the linear dependence in the space of the occupied - virtual orbital products @xcite . with the use of gto basis functions
, this requirement leads to very large orbital basis sets with hundreds of basis functions even for small molecules .
on the other hand , the use of slater - type ( sto ) basis functions give considerably more compact orbital basis sets which can be beneficial for the application of the xoep formalism , and a very efficient implementation of quantum chemical formalisms with slater - type basis functions has been achieved in the smiles suite of programs @xcite .
it is the primary purpose of the present work to implement the xoep method within the smiles package and to analyze the advantages which can be obtained from the use of the sto basis sets , testing it for a small number of atoms and diatomic molecules , for which both numerical solutions @xcite and stos are available @xcite . in this work
we employ the xoep algorithm outlined in refs . [ ] and [ ] and the xoep equations , formulated in terms of the stos .
the solution of the xoep equation will be carried out through the truncated singular value decomposition ( tsvd ) technique .
it will be demonstrated that the use of the slater - type basis sets leads to considerable computational savings in every step of the selfconsistent procedure , without deteriorating the accuracy of the calculated xoep total and orbital energies .
in this section the main features of the oep method will be outlined . in the oep method one
seeks for a local multiplicative potential @xmath0 such that its eigenfunctions ( atomic units will be used in this paper ) @xmath1 minimize the total energy functional given by @xmath2 & = & \sum_{\sigma } \sum_{i } \int \phi_{i\sigma}^ { * } ( { \mathbf r } ) \left ( -\frac{1}{2 } \nabla^{2 } \right ) \phi_{i\sigma } ( { \mathbf r } ) d{\mathbf r } \\ & + & \int \rho ( { \mathbf r } ) v_{ext } ( { \mathbf r } ) d{\mathbf r } + \frac{1}{2 } \int \rho ( { \mathbf r } ) \int \rho ( { \mathbf r } ' ) \frac{1}{|{\mathbf r}-{\mathbf r}'| } d{\mathbf r } d{\mathbf r } ' + e_{xc}\left [ \ { \phi_{i\sigma } \ } \right ] , \end{aligned}\ ] ] where @xmath3 runs over occupied orbitals and @xmath4 over the spin , being @xmath5 the electron density and @xmath6 $ ] the orbital - dependent exchange - correlation energy . here and below we use indices @xmath7 for occupied orbitals , @xmath8 for unoccupied orbitals and @xmath9 for general ( i.e. occupied or unoccupied ) orbitals . in the case of the exchange - only formalism , xoep , the xc energy @xmath6 $ ]
is replaced with the hf exchange energy , @xmath10 = - \frac{1}{2 } \sum_{i , j } \sum_{\sigma } \int \frac { \phi_{i\sigma}^ { * } ( { \mathbf r } ) \phi_{j\sigma } ( { \mathbf r } ) \phi_{j\sigma}^ { * } ( { \mathbf r } ' ) \phi_{i\sigma } ( { \mathbf r}')}{|{\mathbf r}-{\mathbf r}|'}.\ ] ] the local multiplicative potential @xmath11 is splitted into the external potential @xmath12 ( i. e. the potential due to the nuclei ) , the coulomb potential of the electron cloud and the local exchange potential @xmath13 .
the xoep equations in a basis set representation are obtained from the minimization of the total energy presented in eq .
( [ etot ] ) with respect to the local potential @xmath14 @xcite , being this minimization equivalent to the minimization respect to the density , by virtue of the sham - schlter condition @xcite and the hohenberg - kohn theorems @xcite .
an expansion of the exchange part of the local potential is assumed in terms of an appropiate set of functions @xcite , @xmath15 where @xmath16 labels the spin , @xmath17 labels all needed values of the basis index and @xmath18 are the expansion coefficients of the exchange potential in this basis . following the literature @xcite the expansion functions
are conveniently defined as @xmath19 where @xmath20 are square integrable functions .
note that this definition implies that the expansion functions @xmath21 are not necessarily square integrable ; this is not a problem as the local potential does not satisfy this condition .
the requirement that the total energy to be stationary under the variations of the local potential , i. e. @xmath22 , is then equivalent to finding a minimum of the total xoep energy with respect to the set of the expansion coefficients of the the local potential , @xmath23 .
if we work with a real orbital basis and if we introduce a scalar product of two functions @xmath24 and @xmath25 of our expansion space as @xmath26 the minimization of the xoep energy i. e. eq .
( [ etot ] ) with @xmath27 defined as the fock exchange energy given in eq .
( [ xenergy ] ) leads to the equation @xcite @xmath28 \phi_{a\sigma } ( { \mathbf r } ' ) = 0 .\ ] ] here the non - local potential , @xmath29 , is defined as @xmath30 } % { \delta \phi_{j \sigma}(\mathbf{r } ' ) } , \ ] ] being @xmath31 the solutions of eq .
( [ eigen ] ) .
using the matrix @xmath32 in eq .
( [ eqpot ] ) we get a matrix equation equivalent to the minimization of the xoep energy , @xmath33 where @xmath34 is the vector of the expansion coefficients for the local exchange potential and @xmath35 is the projection in the chosen basis set of the non - local hf potential , @xmath36 . in this work
our expansion basis set will be a scaled form of the occupied - virtual products , specifically @xcite @xmath37 thus , the elements of the vector @xmath35 are @xmath38 and the matrix elements @xmath39 reduce to @xmath40 note that with the basis set of occupied - virtual products it is not possible to get any term having a @xmath41 asymptotic decay .
this is corrected by the addition to our basis set @xmath42 of the fermi - amaldi function @xmath43 , where @xmath44 is the number of electrons of the system .
that procedure reproduces the fermi - amaldi potential for long distances and will make the xoep homo eigenenergies to be very close , but not equal , to those found using the hf method .
this prescription @xcite is very different to the one adopted when the exchange potential is expanded in an auxiliary basis set @xcite .
as the products of the occupied and virtual states are linearly dependent , the matrix @xmath45 appearing in eq .
( [ nablaoep ] ) is singular and the equation can not be solved by inversion @xcite . following the argument given in refs .
@xcite and @xcite , only the linearly independent functions @xmath21 ( or @xmath20 ) can be used in the expansion of the xoep , i. e. for a faithful solution of the xoep equation , eq .
( [ nablaoep ] ) , linearly - independent orbital products must be employed .
we will then apply the truncated singular value decomposition ( tsvd ) technique to separate the linear dependent occupied - virtual products and the independent ones , seeking for the linear independent set of products by diagonalization of the @xmath45 matrix . in order to fulfill this condition ,
a threshold is chosen to discriminate the elements of the @xmath45 matrix that correspond to the linearly independent functions .
perfectly invertible ) , then eq . ( [ nablaoep ] ) would have had a unique solution that would correspond to the lowest variational energy obtainable with the functional given in eq .
( [ xenergy ] ) , that is the hf energy @xmath46 . ] in this way , the mapping between the density and density matrix becomes non - unique and a solution with the energy @xmath47 is obtained @xcite . in our case ,
the tsvd method requires the diagonalization of the matrix @xmath45 , in general a very large one .
but as the stos represent the unoccupied orbitals in a much more efficient way than the gtos , when slater - type orbital basis sets are used the size of the matrix @xmath45 , and consequently the expansion space for the local potential in eq .
( [ vxspan ] ) , is much smaller than when using gtos .
this is the main point of this paper : when sto basis sets are used , the computational effort ( in memory size and in speed of the calculations ) for each self - consistent cycle is much lower , whereas the quality of the results is preserved .
the algorithm outlined in the previous section was implemented in the smiles suite of programs @xcite , which employs the sto basis sets in quantum chemical calculations .
we will compare the results of the xoep calculations obtained with the slater - type basis functions ( xoep - sto ) to both numerical exact solutions and xoep results obtained with the use of the gaussian - type basis functions ( xoep - gto ) .
the latter results were obtained with the use of the molpro2008.1 code @xcite , where the xoep formalism employing gaussian - type basis sets was recently implemented by some of us @xcite . in order to use comparable basis sets ,
the correlation - consistent basis sets of dunning ( cc - pvtz , cc - pvqz , and cc - pv5z ) were used in the xoep - gto calculations , and the sto of similar quality ( vb1 for cc - pvtz , vb2 for cc - pvqz and vb3 for cc - pv5z ) @xcite were selected for the xoep - sto calculations .
these sto and gto basis sets yield the total hf energies in close agreement ( see tables [ totalbe ] [ totalco ] below ) .
all the basis sets we have used ( xoep - sto and xoep - gto calculations ) were employed in their uncontracted form . due to the few exact numerical xoep solutions found in the literature , and to the small number of available stos for atomic and molecular computations , the calculations we present here were performed for the be and ne atoms and for the lih , bh , li@xmath48 and co molecules .
the numerical solutions were given in makmal et al .
@xcite and we have used the same internuclear distances ( in a.u . ) : @xmath49 for lih , @xmath50 for bh , @xmath51 for li@xmath48 and @xmath52 for co. this distance for the co molecule has also been used by heelmann et al .
@xcite in the xoep - gto solution , and we will also compare our results with theirs in section [ sec : results ] .
the dependence of the xoep - sto and the xoep - gto total energies on the size of the basis set and on the tsvd cutoff criterion @xmath53 for neglecting ( near ) zero eigenvalues of the matrix @xmath45 has been investigated .
the results are collected in tables [ totalbe ] , [ totallih ] and [ totalco ] . in the calculations , the tsvd cutoff criterion @xmath53 was varied in the range @xmath54 .
note that when @xmath55 is employed , the expansion space for the potential is very small because only a few eigenvalues of the matrix @xmath45 are greater than @xmath53 , yielding energies noticeably above the numeric xoep values .
for that reason , we have not reported that energies in the tables .
reducing the tsvd cutoff @xmath53 makes the expansion space bigger and , as a consequence , the total xoep - sto and xoep - gto energies decrease , approaching to the accurate numeric xoep values . for @xmath53 in the range @xmath56 ,
the total xoep energies remain constant to within a fraction of mha .
when the xoep - gto method is used with a very tight cutoff criteria ( @xmath57 ) , the iterative solution of the xoep equation ( [ nablaoep ] ) becomes unstable and the xoep total energy collapses towards the hf energy . as a matter of fact , the procedure breaks - down as the matrix @xmath45 becomes noninvertible .
on the other hand , the xoep - sto implementation shows somewhat greater stability and begins to break down at smaller values of @xmath53 , i. e. when @xmath58 .
this can be attributed to the fact that , with the use of sto functions , the expansion set of the potential is much smaller than with the more traditional gaussian - basis sets ( see below ) .
the number of the eigenfunctions of the matrix @xmath45 used for the expansion of the potential is also given in tables [ totalbe ] , [ totallih ] and [ totalco ] , as well as the total dimension of the matrix .
it is seen that the dimension of @xmath45 is considerably smaller in the xoep - sto method , and the dimension of the potential expansion space does not grow as fast as in the case of the xoep - gto method .
so , the xoep - sto implementation gives a noticeably memory savings and a greater stability with respect to the cutoff criterion of the tsvd procedure .
for all the systems in tables [ totalbe ] , [ totallih ] and [ totalco ] , the differences between the xoep - sto and the xoep - gto total energies are typically smaller than 1 mha .
when large basis sets are used the total xoep - sto and xoep - gto energies approach the exact numeric values with an accuracy better than 100 @xmath17 ha .
it is important to stress here that the xoep and hf total energies converge in a somewhat different way to respect to the basis set size ( remember that the basis sets we use in this paper were nor developed neither optimized for the xoep calculations ) so the differences @xmath59 oscillate .
table [ totalxoep ] summarizes the results of calculations for be and ne atoms and a number of diatomic molecules studied by makmal et al .
@xcite using the real space xoep method . for the sake of comparison , the xoep energies obtained by heelmann et al .
@xcite with the use of balanced auxiliary basis sets for the potential are also shown when available .
note that the xoep - sto energies are typically in a somewhat better agreement with the numerical values than the xoep - gto energies obtained with similar basis set ( the sto basis sets give results about @xmath60 mha below the energies obtained by heelmann et al . ) .
table [ orbitalxoep ] collects the energies of the occupied orbitals obtained with the use of the xoep - sto and xoep - gto methods for the be atoms and the lih and li@xmath48 molecules .
there is a good agreement between the numerical values of these orbital energies and the energies obtained with the two xoep methods .
furthermore , the orbital energies from the xoep - sto and from the xoep - gto calculations agree with each other to within a few mha . besides the total and orbital xoep energies , we have also studied the energy decomposition into the kinetic , nuclear - electron attraction , electron - electron repulsion and the exchange energies .
table [ edecomp ] presents the above components of the total xoep energy as obtained using the sto and gto basis sets for the be atom and the co molecule ( there are no numerical solutions available in this case for the xoep ) .
[ fig : ne ] and [ fig : co ] show the results of the xoep potential for the ne atom and the co molecule along the main axis . for the ne atom , note the close agreement between our result and that obtained with the exact numerical calculation by kurth and pittalis @xcite , thus yielding a smooth potential that shows only small deviations from the numerical potential ; in any case , these xoep potentials are also very similar to that evaluated with a gaussian basis set @xcite using the procedure presented in refs .
@xcite and @xcite . in the lower panel of fig .
[ fig : co ] the results we have obtained using the vb3 basis set for the co molecule are compared with the calculation by heelmann et al .
@xcite , using an auxiliary basis set within a gaussian representation ; the sto results show a good agreement with the xoep - gto potential . for the sake of completeness , in the upper panel we also present several other calculations for the internuclear region , using sto basis sets of different quality ( vb1 , vb2 and vb3 ) .
note that our xoep potentials do not present any unphysical wiggle as those found by staronerov et al .
@xcite and have a good agreement with both the numerical calculation and the xoep - gto solution by hessellmann et al .
@xcite with the use of auxiliary basis set . in summary ,
the previous results show that both xoep - sto and xoep - gto methods yield close results to the numerical exact ones for the total energies , the one - electron energies of the occupied orbitals and the xoep potentials .
the implementation of the xoep formalism with slater - type basis functions has been developed .
this procedure has been tested in a small group of closed shell atoms and diatomic molecules , for which both numerical xoep solutions and sto basis sets are available . when compared to the exact numerical solution of the xoep equations we have obtained very good results ; they are even a bit better than those given by the gaussian - type basis sets procedure . on the other hand , both xoep - sto and xoep - gto results obtained with the prescription proposed in refs . [ ] and [ ] give energies about @xmath60 mha below those obtained for xoep - gto by heelmann et al .
, and thus they are closer to the exact results @xcite .
the new method leads to a considerably more compact expansion space for the xoep local multiplicative potential , yielding noticeable savings in the computational effort to be done in each one of the cycles of the self - consistent procedure .
yet another advantage of using the slate - type basis sets is that , within the tsvd algorithm , fewer eigenvalues of the ( near ) singular matrix @xmath45 need to be employed , which leads to an increased numeric stability of the xoep - sto method as compared to the xoep - gto algorithm . as a final remark
, it is known that a more efficient xoep algorithm can be developed based on the use of the incomplete cholesky decomposition technique .
the use of this technique would facilitate the application of the xoep - sto method to larger molecules .
this implementation is currently in progress and will be reported elsewhere . on the other hand , this work is a first step to develop a local potential formalism for both exchange and correlation . due to the smaller memory requirements of the sto scheme we have presented here , it can be used to study bigger molecules than those that can be solved with the standard gto approach .
the implementation of the correlation part of the potential is under development .
the authors thank prof .
jaime fernndez rico and his research group for providing a copy of the smiles suite of programs and supporting this work .
the authors acknowledge the financial support from the spanish ministerio de ciencia e innovacin through the research project grant fis2010 - 21282-c02 - 02 .
.total energies ( hartree ) for the be atom using the hf and the xoep methods with the sto and the gto basis sets . in the first column
we indicate the value @xmath53 used as a threshold in the tsvd decomposition of the tsvd decomposition of the @xmath45 matrix in the xoep method . for the sto and gto xoep results
we have included , within brackets , the total number of occupied - virtual products ( first number in the brackets ) and the number of them that are used ( second number in the brackets ) in each calculation . [
cols="^,^,^,^,^",options="header " , ] a.u . ) . in the lower panel ,
the red line show the results of this paper and the green one reflects the xoep - gto results of heelmann et al .
@xcite . in the upper panel the region between the nuclei of the molecule
is depicted in more detail ; the xoep - gto results are shown again in green , and results for xoep - sto calculations with different sto basis sets @xcite are also plotted ( vb1 , blue dashed line ; vb2 , magenta dotted line ; vb3 , red thick line ) . , width=453 ] | the exchange - only optimized effective potential method is implemented with the use of slater - type basis functions , seeking for an alternative to the standard methods of solution with some computational advantages .
this procedure has been tested in a small group of closed shell atoms and diatomic molecules , for which numerical solutions are available .
the results obtained with this implementation have been compared to the exact numerical solutions and to the results obtained when the optimized effective equations are solved using the gaussian - type basis sets .
this slater - type basis approach leads to a more compact expansion space for representing the potential of the optimized effective method and to considerable computational savings when compared to both the numerical solution and the more traditional one in terms of the gaussian basis sets . |
gamma - ray bursts ( grbs ) are the most luminous objects in the universe , releasing @xmath2 in a few seconds .
the spectral features and lightcurves of their prompt and afterglow emission imply that they are produced in ultrarelativistic jets .
the central engines most likely to launch these jets are hyperaccreting black holes @xcite . in this model ,
a relativistic jet emerges from a system with a massive accretion disk ( @xmath3 ) surrounding a stellar - mass black hole , a system expected to form after a cataclysmic event such as the gravitational collapse of a massive star or a merger of a neutron star - neutron star binary @xcite .
the observed luminosity requires a very large mass accretion rate , @xmath4 .
in such a case the accretion flow is extremely optically thick with respect to photons and it can not cool efficiently via radiation .
however , since the density and temperature are very large ( @xmath5 , @xmath6 ) when the mass accretion rate is sufficiently high , neutrino cooling becomes efficient .
such disks are ` neutrino - dominated accretion flows ' ( ndafs ; @xcite ; see also chapter 10.6 of @xcite ) . the accreting system must launch a relativistic jet to produce a grb .
neutrino pair annihilation ( @xmath7 ; @xcite ) and magnetohydrodynamical mechanism such as blandford - znajek process @xcite are the two processes most discussed in the literature .
the goal of this work is to estimate the luminosity of the jets emerging from a hyperaccretion flow driven by an mhd mechanism , as well as the luminosity of thermal neutrinos emitted from the disk .
the first is estimated using the maximal magnetic field sustainable on the event horizon , which is expected to be limited by the pressure in the disk near the isco , the innermost stable circular orbit @xcite . the second we calculate directly from the density , pressure , and lepton number of the material in the disk . in section 2
we predict the jet power in the mhd model by solving for the structure of a surrounding ndaf and compare this luminosity to the energy deposition rate via neutrino pair annihilation above the accretion flow .
we discuss the parameter dependence of our model and its application to grb jets in section 3 .
we summarize our results in section 4 .
we consider a rotating black hole whose horizon is threaded by a large - scale poloidal magnetic field @xcite .
the poynting luminosity expected from such a system is @xmath8 , where @xmath9 is the poloidal magnetic field strength on the horizon and @xmath10 is the gravitational radius of the central black hole @xcite .
following @xcite we estimate the jet luminosity as @xmath11 where @xmath12 is an increasing function of @xmath13 whose exact form depends on the specific field configuration ( e.g. @xcite ) .
there have been some attempts to calculate @xmath12 analytically for specified field geometries @xcite , but these configurations were not required to be consistent with the dynamics of the surrounding accretion flow .
they were also relatively simple poloidal topologies , but as @xcite showed , high jet power requires that the vertical component of the magnetic field maintain the same sign for long times ; in other words , low - order poloidal topologies are the most favorable to strong jets .
consequently , it is possible for @xmath12 to vary over an extremely wide range , from very small numbers to @xmath14 . when the topology is optimal and @xmath15 , the @xmath12 arising from a dynamically self - consistent field structure
can be @xmath16 , and even larger for spin parameters closer to unity @xcite .
@xcite argued that the magnetic pressure near the horizon may be limited by the inner disk pressure .
the magnetic field energy may therefore be estimated ( or at least bounded ) by the disk pressure @xmath17 near the isco : @xmath18 where @xmath19 is the radius of the pressure reference point and @xmath20 is the ratio of the midplane pressure at @xmath19 to the magnetic pressure in the black hole s stretched horizon . to evaluate the midplane pressure of the hyperaccretion flow , we solve for the disk structure in the innermost region ( @xmath21 ) , where the jet is expected to be launched .
as we show below , the behavior of an hyperaccretion disk is mainly determined by its mass accretion rate @xmath22 . to highlight what we believe to be the principal physical mechanisms ( and
also eliminate mathematical complications ) , we adopt a highly simplified model for this structure in the expectation that the trends we uncover while exploring a factor of @xmath23 range in @xmath24 will be robust with respect to refinements in the model . in particular , we assume that both the disk dynamics and the gravitational potential are newtonian despite the fact that this near a black hole general relativistic effects are important .
we also assume that the system is stationary and axisymmetric .
we adopt the shakura - sunyaev formalism for the disk structure , which formally applies only to thin disks , assumes that the inflow speed is very small compared to the orbital speed , and parameterizes the @xmath25-@xmath26 component of the stress @xcite . when the disk becomes geometrically thick , in the advection - dominated regimes
, we use in our analytic estimates ( but not in the numerical solution ) a dimensional analysis .
remarkably the results are similar up to a numerical coefficient to those obtained using the shakura & sunyaev model .
we further set the correction factor @xmath27 appearing in equation [ angcons ] to always be unity , when it should in fact be somewhat less than that .
these are rough approximations , but we nonetheless expect our results to give the right qualitative picture in this scenario . here
we present the basic disk equations for the density @xmath28 , the temperature @xmath29 , and the scale height @xmath30 .
these are the expressions for mass conservation , angular momentum conservation , energy conservation , and hydrostatic balance applied at @xmath19 : @xmath31 where @xmath32 , @xmath33 , @xmath34 , and @xmath35 denote the surface density ( @xmath36 ) , radial velocity , angular velocity ( @xmath37 ) , and ratio of integrated stress to integrated pressure , respectively . in newtonian dynamics
, the factor @xmath38 accounts for the reduction in stress due to the net angular momentum flux through the disk ; although the net angular momentum flux per unit mass accreted has traditionally been identified with the specific angular momentum of orbits at the isco @xcite , it also depends upon mhd effects @xcite .
the pressure @xmath17 is the sum of the contributions from radiation , baryonic gas , degenerate electrons , and neutrinos ( if they are trapped ) : @xmath39 where @xmath40 is the radiation constant , @xmath41 is the boltzman constant , and @xmath42 is the mass of a proton . the first term on the right hand side represents the contribution from both photons and relativistic electron - positron pairs ; the coefficient is 11/12 when the temperature is higher than @xmath43 .
in the third term , which corresponds to degeneracy pressure , @xmath44 is the electron fraction , which should be evaluated from the @xmath45-equilibrium condition , and when neutrino emission is not efficient the electron fraction should be nearly equal to 0.5 . in the numerical calculations we make use of a bridging formula in evaluating @xmath44 to connect these two extreme regimes smoothly . ]
@xmath46 where @xmath47 is the difference between the neutron and proton masses , and @xmath48 is the chemical potential of the electrons ( we assume that the chemical potential of neutrinos can be ignored and that heavy nuclei are totally dissociated in the innermost region of the accretion flow ; see @xcite ) .
the last term represents trapped neutrino pressure ; @xmath49 is the neutrino energy density , defined below .
@xmath50 is the heating rate per unit area : @xmath51 @xmath52 is the cooling rate per unit area , including both neutrino and advective cooling : @xmath53 .
the advective cooling rate @xmath54 , which equals the inward flux of the gas energy along the flow , can be described as : @xmath55 here @xmath56 denotes the entropy per unit mass .
there are several significant neutrino emission processes expected in such a hyperaccretion flow ( see also @xcite and references therein ) : electron / positron capture on nuclei ( @xmath57 ; @xmath58 , also known as the urca process ) , electron - positron pair annihilation ( @xmath59 , where @xmath60 represents electron- , mu- , and tau - type neutrinos ) , nucleon - nucleon bremsstrahlung ( @xmath61 ) , and plasmon decay ( @xmath62 ) .
we should also take into account the neutrino optical depth in the accretion flow , which can be larger than unity when the mass accretion rate is sufficiently large .
here we adopt a prescription based on the two - stream approximation @xcite : @xmath63 where @xmath64 and the neutrino flavor is labeled by subscript @xmath65 . @xmath66 and @xmath67 represent absorptive and scattering optical depths respectively .
the detailed description of these neutrino opacities can be found in @xcite .
when the optical depth is so large that the neutrino diffusion timescale in the accretion flow is longer than the accretion timescale , neutrinos are completely trapped in the accretion flow . in that state ,
advective cooling dominates : @xmath68 .
the accretion flow would also be advection - dominated when the mass accretion rate is small enough that the density and temperature of the disk are so low that neutrinos are not efficiently emitted , but is still much larger than the eddington accretion rate .
we can divide the hyperaccretion flow into several regimes according to the dominant physical processes taking place .
in particular , the cooling processes ( advection , optically - thin neutrino emission and optically - thick neutrino emission ) and the pressure components ( radiation , baryonic gas , and degenerate electrons ) vary from one regime to another . in the rest of this section
, we solve the structure of a hyperaccretion flow in its innermost region as a function of the mass accretion rate and examine the dominant physical processes determining the behavior of the accretion flow in the different regimes .
we numerically solve the fundamental equations of an hyperaccretion flow , eqs .
( [ masscons])-([hydrostat ] ) , in the innermost region to find the density , temperature , and scale height as functions of the mass accretion rate @xmath22 .
in the course of doing so , we also identify the dominant physical mechanisms operating over a wide range of @xmath22 values . as the relevant ranges of parameters for the central engine of a grb , we consider @xmath69 , @xmath70 , and @xmath71 .
our fiducial parameters are @xmath72 , @xmath73 and @xmath74 .
figure [ f1 ] depicts the different cooling rates ( neutrino emission and advection ) , optical depths for different flavors of neutrinos , and pressure components ( radiation , baryonic gas , and degenerate electrons ) as functions of mass accretion rate .
we can see that neutrino cooling becomes dominant over advective cooling above @xmath75 .
figure [ f2 ] depicts the jet luminosity @xmath76 and efficiency @xmath77 , estimated from eq .
( [ jetlum ] ) . inspection of figures [ f1 ] and [ f2 ] reveals a sharp break in the nature of the disk and jet at @xmath78 , the transition from advection to neutrino cooling . in order to interpret this break ,
we have constructed an approximate analytical model for hyperaccretion disk structure .
this model can qualitatively reproduce not only the physical properties of the flow , but also the mass accretion rates at which one regime changes to another . when the mass accretion rate is not large enough for the disk to cool efficiently via neutrino emission , the accretion flow is advection - dominated , and the pressure is dominated by photons and relativistic electron / positron pairs . in this regime ,
the accretion flow is geometrically thick and the thin - disk approximation is not valid .
instead , we can estimate the disk pressure using dimensional analysis . because the aspect ratio of the accretion flow @xmath81 , the density at @xmath19 is @xmath82 ) .
as the gas does not cool efficiently , the ratio of pressure to density ( i.e. the square of the sound speed ) is close to the virial temperature @xmath83 . using the relation between @xmath33 and the circular orbital speed @xmath84 , @xmath85 ,
the pressure can be evaluated as : @xmath86 using this expression for @xmath17 we estimate the jet luminosity as @xmath87 the corresponding energy efficiency of the jet : @xmath88 is independent of the mass accretion rate .
the accretion flow is advection - dominated as long as the neutrino emissivity per unit volume is smaller than half the advective cooling rate . in this regime ,
neutrino emission is dominated by electron / positron capture , and the neutrino emissivity @xmath89 is given by @xmath90 where @xmath91 and @xmath92 ( popham , woosley & fryer 1999 ) .
the density and temperature of the gas are : @xmath93 substituting these values into the expression for neutrino emissivity , we obtain : @xmath94 on the other hand , the advective cooling rate per unit volume is @xmath95 here we use @xmath96 where @xmath97 is the adiabatic index of relativistic particles .
a transition takes place when @xmath98 , i.e. @xmath99 above this accretion rate the accretion flow is no longer advection - dominated , as neutrino cooling determines the structure of the flow .
in addition , the dominant pressure source is no longer radiation , but baryonic gas pressure . in the following , we call this transition mass accretion rate @xmath100 ( `` ign '' stands for ignition of neutrino emission ) . in this regime the disk is mainly cooled by neutrino emission via electron / positron capture , while the pressure is dominated by baryonic gas ( see @xcite and references therein ) . in this case , the pressure can be expressed as @xmath104 the neutrino cooling rate per unit surface area by electron / positron capture is given by eq .
( [ ecap ] ) . substituting @xmath105 in eq .
( [ energy ] ) , we solve the basic equations and derive the disk properties in this regime : @xmath106 the disk pressure is given by @xmath107 finally we estimate the jet luminosity : @xmath108 the jet luminosity in this regime depends linearly on the mass accretion rate .
the efficiency @xmath77 is : @xmath109 and is independent of the mass accretion rate .
the higher jet luminosity and efficiency in this regime is the result of more efficient neutrino cooling , which leads to a geometrically thinner accretion flow .
thinner disks accrete more slowly . in the end ,
the midplane pressure @xmath17 increases , permitting it to confine a stronger magnetic field on the black hole s event horizon .
this stronger magnetic field supports a more powerful jet .
the system is very optically thin to neutrinos at this stage : @xmath110 where @xmath111 and @xmath112 are the optical depth due to nucleon absorption and scattering , respectively ( see @xcite ) .
the total optical depth for electron neutrinos @xmath113 exceeds unity only when @xmath114 and then our expression for @xmath52 would not be valid .
this mass accretion rate marks the transition to the third regime , which is optically thick to @xmath102 . in the following we call this transition @xmath115 . in this regime
the innermost region of the disk is optically thick to electron neutrinos , and the opacity is dominated by absorption onto nucleons . while the disk is not so optically thick that the neutrinos are totally trapped in the flow , the cooling rate is : @xmath117 where @xmath118 is the neutrino optical depth .
first we consider the case in which the disk is optically thick only to @xmath102 and thin for @xmath119 and @xmath120 . with cooling only by electron neutrino emission ,
the cooling rate is @xmath121 which should be equated with the heating rate in the energy balance equation .
solving the fundamental equations of the disk we derive its density , temperature and scale height : @xmath122 the disk pressure in the midplane is @xmath123 and the jet luminosity is @xmath124 the dependence on the mass accretion rate is weaker than in the last regime .
the efficiency is : @xmath125 in order for the disk to cool by neutrinos efficiently , the neutrino diffusion time @xmath126 must be shorter than the accretion time @xmath127 .
this condition can be expressed as @xmath128 and when the mass accretion rate exceeds this value , electron neutrinos are no longer able to escape the disk .
we call it @xmath129 . for the other kinds of neutrinos ( @xmath119 and @xmath120 ) ,
the diffusion time is still shorter than the accretion time , so they can cool the disk efficiently .
this is the fourth regime , which we discuss in the next subsection .
even when the mass accretion rate is so large that @xmath102s are completely trapped in the disk , @xmath119s and @xmath120s emitted via the electron - positron pair annihilation process may escape the disk so long as their diffusion time is shorter than the accretion time .
for those neutrinos , the opacity is dominated by scattering on nucleons ( eq . [ taus ] ) .
then the cooling rate is @xmath132 where the factor 2 in the first line represents the number of flavors of neutrinos that we are interested in ( @xmath119/@xmath133 and @xmath120/@xmath134 ) . substituting this @xmath52 into eq .
( [ energy ] ) , we derive the disk properties : @xmath135 the disk pressure is then @xmath136 and the jet luminosity is @xmath137 the luminosity now depends on the mass accretion rate in the same way as in the last regime .
the efficiency is @xmath138 the timescale for @xmath139/@xmath140 neutrinos to escape the disk diffusively is @xmath141 and the accretion timescale is @xmath142 therefore , @xmath139/@xmath140 neutrinos can escape an accretion flow without being trapped as long as @xmath143 at this mass accretion rate , there is a transition to the fifth and last regime , in which the accretion flow is once more advection - dominated . when the mass accretion rate is so large that the all kinds of neutrinos are trapped , the accretion flow can not cool and the flow is advection - dominated . in this case
the pressure would be dominated by several kinds of relativistic particles , photons , relativistic electron - positron pairs , and all three flavors of neutrinos .
we can estimate the disk pressure in a way similar to sec .
2.2.1 to find @xmath147 the jet luminosity is then @xmath148 as a result , the efficiency is the same as in equation ( [ eff1 ] ) , @xmath149 in the discussion above , we have picked out the four mass accretion rates that divide the five states of an hyperaccretion flow .
however , the sequence we have presented does not always apply .
these mass accretion rates are functions of @xmath35 , @xmath150 , and @xmath19 , and in some regions of parameter space , the transition from advection - dominated flow to neutrino - cooled flow ( @xmath100 ) takes place at a higher accretion rate than the transition to optically thick flow ( with respect to @xmath102 ) .
this means that the second regime of optically thick neutrino cooling does not exist for these parameters .
for example , we show the transition mass accretion rates as functions of the black hole mass ( while the other parameters are fixed at @xmath151 and @xmath72 ) in figure [ f3 ] .
one can see that there is a critical black hole mass at which the first transition and the second transition change places , and also that for this choice of parameters the mass accretion rate range of the optically - thin ndaf regime is quite narrow , if it exists at all . in figure
[ f4]we show @xmath152 for various parameter sets as functions of the black hole mass . for @xmath153 and @xmath154 , the curve is well above @xmath152 , and the optically - thin ndaf regime exists for each of those parameter sets .
the general condition for such a change in the order of the transition points is : @xmath155 when this condition is satisfied , advection - dominated flow transforms directly to optically thick neutrino cooling at ( [ trans2 ] ) .
figure [ f5]displays the limited range of parameters satisfying this condition .
in other words , except in the case of extreme values of the parameters , the five states for hyperaccretion flows are ordered as we have described them .
from the analytic model presented above , we can estimate the jet luminosity and efficiency expected from a hyperaccreton flow as functions of mass accretion rate for various choices of parameters ( @xmath35 , @xmath150 , @xmath19 ) . figure [ f6 ] shows how the jet luminosity based on the bz scenario ( thick lines ) depends on the mass accretion rate .
the different lines show the results for different parameters .
we can see that with all parameter sets , the jet luminosity has a significant discontinuity at @xmath100 .
this jump , as explained already , arises from the transition between an advection - dominated accretion flow , in which neutrino cooling is not efficient , and a neutrino - dominated accretion flow . in the top panel
, we can see that for larger black hole mass the mass accretion rate at which the transition from the advection - dominated flow to the neutrino - cooled flow ( @xmath100 ) is larger .
the value of @xmath100 rises from @xmath156 s@xmath157 at @xmath158 to @xmath159 s@xmath157 at @xmath160 .
this is because the temperature of the accretion flow is lower for larger black hole mass when the radius ( in units of the gravitational radius ) and the mass accretion rate are fixed .
therefore , a larger mass accretion rate is needed for an accretion flow around a larger black hole to cool efficiently via neutrino emission .
a similar explanation also applies to the bottom panel , which shows the dependence of the jet luminosity on @xmath19 . here , for smaller @xmath19
the temperature of the accretion flow is higher , which results in a smaller @xmath100 and larger luminosity . over the range of plausible values of @xmath19 ,
the variation of @xmath100 is much smaller than for the dependence on @xmath161 : it changes by only a factor @xmath162 .
the middle panel shows that the jet luminosity depends quite strongly on the stress / pressure ratio in the sense that for larger @xmath35 , the luminosity is smaller ; equivalently , @xmath100 increases with increasing @xmath35 . over a span of a factor of 30 in @xmath35 , from @xmath163 to @xmath151
, @xmath100 increases by a factor of 300 . to interpret this behavior
, we note that the inflow rate is proportional to @xmath35 ( i.e. the accretion timescale is proportional to @xmath164 ) .
therefore , when @xmath35 is larger in the radiation - dominated advective regime , the midplane pressure is smaller ( eqn .
[ pres1 ] ) .
the temperature is also lower ( eqn .
[ rhotempadvect ] ) .
these two trends together diminish the neutrino cooling rate , forcing the transition accretion rate higher .
they are reinforced by the scalings on the opposite side of the transition , in which @xmath165 , while the temperature depends much more weakly on @xmath35 ( only @xmath166 ) .
thus , the jet luminosity is strongly reduced by a larger stress / pressure ratio .
figure [ f7 ] similarly shows how the jet efficiency depends on our three free parameters .
as we have previously emphasized , the principal dependence is on accretion regime : the jet efficiency is greatest when the flow is in the optically thin neutrino - cooling state .
although this state occurs over different ranges of accretion rate for different black hole masses , the efficiency achieved there varies rather little with @xmath150 . on the other hand , even after allowing for the different accretion rate ranges in which the flow is neutrino - cooled , it is quite sensitive to both @xmath35 and @xmath167 . in that favored regime
, @xmath168 is roughly @xmath169 .
still more dramatically , a factor of 3 decrease in @xmath167 leads to a factor of 30 increase in @xmath168 at the respective peaks of jet efficiency in the optically - thin neutrino - cooled state and this comparison is taken at fixed @xmath12 , whereas realistically one would expect smaller @xmath167 to be associated with more rapid spin and therefore a significant increase in @xmath12 .
note that in a small region of the lower panel @xmath170 . while possible , in principle , in a bz process , our newtonian approximation is not suitable to explore this regime .
neutrino - antineutrino annihilation into electrons and positrons ( hereafter neutrino pair annihilation " ) above the accretion flow @xcite is another mechanism proposed for the energy source of relativistic jets that power grbs .
it is worth comparing the luminosity and efficiency of such neutrino pair annihilation driven jets with the bz driven jets expected from our ndaf model . in estimating the energy deposition rate due to neutrino pair annihilation above a hyperaccretion flow
, we adopt a simplified formulation : the accretion flow is assumed to be geometrically thin ( i.e. neutrinos are emitted from the equatorial plane ) , and we neglect the bending of neutrino paths and the gravitational redshift . the energy deposition rate per unit volume above the accretion flow
can then be calculated from @xmath171 ( 1-\cos \theta)^2 \nonumber \\ & & + a_2\sum_k \frac{\delta l_{\nu_i}^k}{d_k^2}\sum_{k^{\prime}}\frac{\delta l_{\bar{\nu}_i}^{k^{\prime}}}{d_{k^{\prime}}^2}\frac{\langle\epsilon\rangle_{\nu_i}+\langle\epsilon\rangle_{\bar{\nu}_i}}{\langle\epsilon\rangle_{\nu_i}\langle\epsilon\rangle_{\bar{\nu}_i}}(1-\cos \theta ) , \nonumber \\\end{aligned}\ ] ] where @xmath172 and @xmath173 are the neutrino cross section constants for electron neutrinos ( ruffert et al .
1997 ) . here
@xmath174 and @xmath175 are the neutrino luminosity coming from the @xmath176-th cell in the accretion flow and the mean neutrino energy , respectively . for the neutrino emissivity distribution in the accretion flow ,
we adopt a simplified model : neutrinos and antineutrinos are emitted with a fermi - dirac distribution function with a zero - chemical potential and a temperature @xmath177 where @xmath178 . according to zalamea & beloborodov ( 2011 ) ,
such a simplification does not change the results significantly . integrating @xmath179 over the volume for @xmath180 , where @xmath181 is the angle from the rotation axis and @xmath182
, we obtain the total energy deposition rate via neutrino pair annihilation above the accretion flow .
we contrast the jet luminosity expected from the neutrino pair annihilation scenario with the jet luminosity generated by the bz mechanism in fig .
[ f6 ] ( thin lines ) . in all but the highest accretion rate state , the one in which all three neutrino flavors are very optically thick
, we see that the bz jet luminosity ( for our fiducial parameters ) is many orders of magnitude greater than the jet luminosity that would be produced by neutrinos . even in that state ,
the neutrino luminosity is generally two orders of magnitude smaller than the bz jet luminosity . in the accretion rate range
most favorably to the bz mechanism , the ratio is at least another order of magnitude larger .
thus , even if @xmath12 is as small as @xmath183 , the bz mechanism dominates at all accretion rates ; in its favored range , it can be as small as @xmath184 and still generate a larger luminosity than the neutrino mechanism . when comparing the jet peak luminosity with those observed in typical grbs , @xmath185 , after beaming corrections @xcite we find that with our fiducial parameters those require accretion rate larger than @xmath186 .
the drop by a factor of @xmath187 in the efficiency at the transition @xmath100 may be a mechanism that strongly modulates the jet power , leading to the observed variability in grbs .
we have examined the blandford - znajek jet luminosity and efficiency expected from hyperaccreting black holes , investigating the possibility that such a jet can account for observed grbs .
following @xcite , we estimated the poynting luminosity using a simple dimensional argument : we assumed that the intensity of the magnetic field responsible for launching the jet is linked to the pressure in the innermost region of the accretion flow . by using an ndaf solution at the innermost region of the accretion flow ,
we have demonstrated that the luminosity of this poynting - dominated jet can easily exceed the energy deposition rate expected from neutrino pair annihilation above the accretion flow .
this means that mhd processes around the black hole horizon are a plausible mechanism for the formation of the relativistic jets required from the observations of grbs .
in addition , we have derived the jet luminosity as a function of mass accretion rate and have shown that it has a step function - like behavior at a mass accretion rate corresponding to the transition between an advection - dominated accretion flow and a neutrino - dominated accretion flow .
this means that the jet power suddenly drops when the mass accretion rate decreases and crosses @xmath100 . in the context of grbs
, this jump may correspond to switching on and off the activity needed for internal shocks .
it may also explain the steep decay ( @xmath188 ) often observed after the prompt emission @xcite .
we have also shown that the mhd jet formation process is most efficient when the accretion flow efficiently cools via optically - thin neutrino emission . in this regime , the jet luminosity is proportional to the mass accretion rate @xmath22 , i.e. , the efficiency does not depend on @xmath22 .
we stress that the bz jet luminosity is larger than the energy deposition rate via neutrino pair annihilation whenever @xmath189 . in the optimal accretion range of accretion rates
, this lower bound is relaxed by at least another order of magnitude .
thus , for the bz mechanism to dominate the neutrion mechanism , the central black hole must spin at least moderately fast , and the magnetic field geometry must be reasonably well - organized in the poloidal sense .
however , as this lower bound on @xmath12 shows , these thresholds are far from extreme . | a hyperaccretion disk formed around a stellar mass black hole is a plausible model for the central engine that powers gamma - ray bursts ( grbs ) .
if the central black hole rotates and a poloidal magnetic field threads its horizon , a powerful relativistic jet may be driven by a process resembling the blandford - znajek mechanism .
we estimate the luminosity of such a jet assuming that the poloidal magnetic field strength is comparable to the inner accretion disk pressure . we show that the jet efficiency attains its maximal value when the accretion flow is cooled via optically - thin neutrino emission .
the jet luminosity is much larger than the energy deposition through neutrino - antineutrino annihilation ( @xmath0 ) provided that the black hole is spinning rapidly enough . when the accretion rate onto a rapidly spinning black hole is large enough ( @xmath1 ) ,
the predicted jet luminosity is sufficient to drive a grb . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.